diff --git a/.github/membrowse-targets.json b/.github/membrowse-targets.json
index e0d1595a0c..3fc387f04f 100644
--- a/.github/membrowse-targets.json
+++ b/.github/membrowse-targets.json
@@ -7,6 +7,7 @@
     "build_cmd": "test -f IDE/GCC-ARM/Header/user_settings.h && mkdir -p IDE/GCC-ARM/Header-gen && { cat IDE/GCC-ARM/Header/user_settings.h; printf '#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n'; } > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen CFLAGS_EXTRA='-Wno-cpp -DWOLFCRYPT_ONLY -DWOLFSSL_NO_SOCK' LDFLAGS='-mcpu=cortex-m4 -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20010000'",
     "elf": "IDE/GCC-ARM/Build/WolfCryptTest.elf",
     "ld": "IDE/GCC-ARM/linker.ld",
+    "map_file": "IDE/GCC-ARM/Build/WolfCryptTest.map",
     "linker_vars": ""
   },
   {
@@ -14,9 +15,10 @@
     "port": "gcc-arm",
     "board": "cortex-m4-min-ecc",
     "apt_packages": "gcc-arm-none-eabi libnewlib-arm-none-eabi libstdc++-arm-none-eabi-newlib",
-    "build_cmd": "test -f examples/configs/user_settings_min_ecc.h && mkdir -p IDE/GCC-ARM/Header-gen && { cat examples/configs/user_settings_min_ecc.h; printf '#define WOLFSSL_GENERAL_ALIGNMENT 4\\n#define SINGLE_THREADED\\n#define WOLFSSL_SMALL_STACK\\n#define NO_FILESYSTEM\\n#define NO_WRITEV\\n#define NO_MAIN_DRIVER\\n#define NO_DEV_RANDOM\\n#define BENCH_EMBEDDED\\n#define USE_CERT_BUFFERS_256\\n#define WOLFSSL_IGNORE_FILE_WARN\\n#define USE_WOLF_ARM_STARTUP\\n#define WOLFSSL_USER_CURRTIME\\n#define WOLFSSL_GMTIME\\n#define USER_TICKS\\nextern unsigned long my_time(unsigned long* timer);\\n#define XTIME my_time\\n#define CUSTOM_RAND_TYPE unsigned int\\nextern unsigned int my_rng_seed_gen(void);\\n#undef CUSTOM_RAND_GENERATE\\n#define CUSTOM_RAND_GENERATE my_rng_seed_gen\\n#define HAVE_HASHDRBG\\n#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n'; } > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen CFLAGS_EXTRA='-Wno-cpp -DWOLFSSL_NO_SOCK -DWOLFCRYPT_ONLY' LDFLAGS='-mcpu=cortex-m4 -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20010000'",
+    "build_cmd": "test -f examples/configs/user_settings_min_ecc.h && mkdir -p IDE/GCC-ARM/Header-gen && { cat examples/configs/user_settings_min_ecc.h; printf '#define WOLFSSL_GENERAL_ALIGNMENT 4\\n#define SINGLE_THREADED\\n#define WOLFSSL_SMALL_STACK\\n#define NO_FILESYSTEM\\n#define NO_WRITEV\\n#define NO_MAIN_DRIVER\\n#define NO_DEV_RANDOM\\n#define BENCH_EMBEDDED\\n#define USE_CERT_BUFFERS_256\\n#define WOLFSSL_IGNORE_FILE_WARN\\n#define WOLFSSL_USER_IO\\n#define WOLFSSL_USER_CURRTIME\\n#define TIME_OVERRIDES\\n#define USER_TICKS\\n#define XTIME my_time\\n#define XGMTIME my_gmtime\\n#define CUSTOM_RAND_TYPE unsigned int\\nextern unsigned int my_rng_seed_gen(void);\\n#undef CUSTOM_RAND_GENERATE\\n#define CUSTOM_RAND_GENERATE my_rng_seed_gen\\n#define HAVE_HASHDRBG\\n#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n'; } > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen CFLAGS_EXTRA='-Wno-cpp -DWOLFSSL_NO_SOCK -DWOLFCRYPT_ONLY' LDFLAGS='-mcpu=cortex-m4 -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20010000'",
     "elf": "IDE/GCC-ARM/Build/WolfCryptTest.elf",
     "ld": "IDE/GCC-ARM/linker.ld",
+    "map_file": "IDE/GCC-ARM/Build/WolfCryptTest.map",
     "linker_vars": ""
   },
   {
@@ -24,9 +26,10 @@
     "port": "gcc-arm",
     "board": "cortex-m4-tls12",
     "apt_packages": "gcc-arm-none-eabi libnewlib-arm-none-eabi libstdc++-arm-none-eabi-newlib",
-    "build_cmd": "test -f examples/configs/user_settings_tls12.h && mkdir -p IDE/GCC-ARM/Header-gen && { cat examples/configs/user_settings_tls12.h; printf '#define WOLFSSL_GENERAL_ALIGNMENT 4\\n#define SINGLE_THREADED\\n#define WOLFSSL_SMALL_STACK\\n#define NO_FILESYSTEM\\n#define NO_WRITEV\\n#define NO_MAIN_DRIVER\\n#define NO_DEV_RANDOM\\n#define BENCH_EMBEDDED\\n#define USE_CERT_BUFFERS_256\\n#define USE_CERT_BUFFERS_2048\\n#define WOLFSSL_IGNORE_FILE_WARN\\n#define USE_WOLF_ARM_STARTUP\\n#define WOLFSSL_USER_CURRTIME\\n#define WOLFSSL_GMTIME\\n#define USER_TICKS\\nextern unsigned long my_time(unsigned long* timer);\\n#define XTIME my_time\\n#define CUSTOM_RAND_TYPE unsigned int\\nextern unsigned int my_rng_seed_gen(void);\\n#undef CUSTOM_RAND_GENERATE\\n#define CUSTOM_RAND_GENERATE my_rng_seed_gen\\n#define HAVE_HASHDRBG\\n#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n'; } > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen CFLAGS_EXTRA='-Wno-cpp -DWOLFSSL_NO_SOCK' LDFLAGS='-mcpu=cortex-m4 -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20010000'",
+    "build_cmd": "test -f examples/configs/user_settings_tls12.h && mkdir -p IDE/GCC-ARM/Header-gen && { cat examples/configs/user_settings_tls12.h; printf '#define WOLFSSL_GENERAL_ALIGNMENT 4\\n#define SINGLE_THREADED\\n#define WOLFSSL_SMALL_STACK\\n#define NO_FILESYSTEM\\n#define NO_WRITEV\\n#define NO_MAIN_DRIVER\\n#define NO_DEV_RANDOM\\n#define BENCH_EMBEDDED\\n#define USE_CERT_BUFFERS_256\\n#define USE_CERT_BUFFERS_2048\\n#define WOLFSSL_IGNORE_FILE_WARN\\n#define WOLFSSL_USER_IO\\n#define WOLFSSL_USER_CURRTIME\\n#define TIME_OVERRIDES\\n#define USER_TICKS\\n#define XTIME my_time\\n#define XGMTIME my_gmtime\\n#define CUSTOM_RAND_TYPE unsigned int\\nextern unsigned int my_rng_seed_gen(void);\\n#undef CUSTOM_RAND_GENERATE\\n#define CUSTOM_RAND_GENERATE my_rng_seed_gen\\n#define HAVE_HASHDRBG\\n#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n'; } > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen CFLAGS_EXTRA='-Wno-cpp -DWOLFSSL_NO_SOCK' LDFLAGS='-mcpu=cortex-m4 -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20010000'",
     "elf": "IDE/GCC-ARM/Build/WolfCryptTest.elf",
     "ld": "IDE/GCC-ARM/linker.ld",
+    "map_file": "IDE/GCC-ARM/Build/WolfCryptTest.map",
     "linker_vars": ""
   },
   {
@@ -34,9 +37,185 @@
     "port": "gcc-arm",
     "board": "cortex-m4-baremetal",
     "apt_packages": "gcc-arm-none-eabi libnewlib-arm-none-eabi libstdc++-arm-none-eabi-newlib",
-    "build_cmd": "test -f examples/configs/user_settings_baremetal.h && mkdir -p IDE/GCC-ARM/Header-gen && { cat examples/configs/user_settings_baremetal.h; printf '#define WOLFSSL_GENERAL_ALIGNMENT 4\\n#define SINGLE_THREADED\\n#define WOLFSSL_SMALL_STACK\\n#define NO_FILESYSTEM\\n#define NO_WRITEV\\n#define NO_MAIN_DRIVER\\n#define NO_DEV_RANDOM\\n#define BENCH_EMBEDDED\\n#define USE_CERT_BUFFERS_256\\n#define USE_CERT_BUFFERS_2048\\n#define WOLFSSL_IGNORE_FILE_WARN\\n#define USE_WOLF_ARM_STARTUP\\n#define WOLFSSL_USER_CURRTIME\\n#define WOLFSSL_GMTIME\\n#define USER_TICKS\\nextern unsigned long my_time(unsigned long* timer);\\n#define XTIME my_time\\n#define CUSTOM_RAND_TYPE unsigned int\\nextern unsigned int my_rng_seed_gen(void);\\n#undef CUSTOM_RAND_GENERATE\\n#define CUSTOM_RAND_GENERATE my_rng_seed_gen\\n#define HAVE_HASHDRBG\\n#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n'; } > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen CFLAGS_EXTRA='-Wno-cpp -DWOLFSSL_NO_SOCK -DWOLFCRYPT_ONLY' LDFLAGS='-mcpu=cortex-m4 -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20010000'",
+    "build_cmd": "test -f examples/configs/user_settings_baremetal.h && mkdir -p IDE/GCC-ARM/Header-gen && { cat examples/configs/user_settings_baremetal.h; printf '#define WOLFSSL_GENERAL_ALIGNMENT 4\\n#define SINGLE_THREADED\\n#define WOLFSSL_SMALL_STACK\\n#define NO_FILESYSTEM\\n#define NO_WRITEV\\n#define NO_MAIN_DRIVER\\n#define NO_DEV_RANDOM\\n#define BENCH_EMBEDDED\\n#define USE_CERT_BUFFERS_256\\n#define USE_CERT_BUFFERS_2048\\n#define WOLFSSL_IGNORE_FILE_WARN\\n#define WOLFSSL_USER_IO\\n#define WOLFSSL_USER_CURRTIME\\n#define TIME_OVERRIDES\\n#define USER_TICKS\\n#define XTIME my_time\\n#define XGMTIME my_gmtime\\n#define CUSTOM_RAND_TYPE unsigned int\\nextern unsigned int my_rng_seed_gen(void);\\n#undef CUSTOM_RAND_GENERATE\\n#define CUSTOM_RAND_GENERATE my_rng_seed_gen\\n#define HAVE_HASHDRBG\\n#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n'; } > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen CFLAGS_EXTRA='-Wno-cpp -DWOLFSSL_NO_SOCK -DWOLFCRYPT_ONLY' LDFLAGS='-mcpu=cortex-m4 -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20010000'",
     "elf": "IDE/GCC-ARM/Build/WolfCryptTest.elf",
     "ld": "IDE/GCC-ARM/linker.ld",
+    "map_file": "IDE/GCC-ARM/Build/WolfCryptTest.map",
     "linker_vars": ""
+  },
+  {
+    "target_name": "gcc-arm-cortex-m0plus",
+    "port": "gcc-arm",
+    "board": "cortex-m0plus",
+    "apt_packages": "gcc-arm-none-eabi libnewlib-arm-none-eabi libstdc++-arm-none-eabi-newlib",
+    "build_cmd": "test -f examples/configs/user_settings_min_ecc.h && mkdir -p IDE/GCC-ARM/Header-gen && { cat examples/configs/user_settings_min_ecc.h; printf '#define WOLFSSL_GENERAL_ALIGNMENT 4\\n#define SINGLE_THREADED\\n#define WOLFSSL_SMALL_STACK\\n#define NO_FILESYSTEM\\n#define NO_WRITEV\\n#define NO_MAIN_DRIVER\\n#define NO_DEV_RANDOM\\n#define BENCH_EMBEDDED\\n#define USE_CERT_BUFFERS_256\\n#define WOLFSSL_IGNORE_FILE_WARN\\n#define WOLFSSL_USER_IO\\n#define WOLFSSL_USER_CURRTIME\\n#define TIME_OVERRIDES\\n#define USER_TICKS\\n#define XTIME my_time\\n#define XGMTIME my_gmtime\\n#define CUSTOM_RAND_TYPE unsigned int\\nextern unsigned int my_rng_seed_gen(void);\\n#undef CUSTOM_RAND_GENERATE\\n#define CUSTOM_RAND_GENERATE my_rng_seed_gen\\n#define HAVE_HASHDRBG\\n#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n'; } > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen ARCHFLAGS='-mcpu=cortex-m0plus -mthumb -mabi=aapcs -DUSE_WOLF_ARM_STARTUP' CFLAGS_EXTRA='-Wno-cpp -DWOLFSSL_NO_SOCK -DWOLFCRYPT_ONLY' LDFLAGS='-mcpu=cortex-m0plus -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20010000'",
+    "elf": "IDE/GCC-ARM/Build/WolfCryptTest.elf",
+    "ld": "IDE/GCC-ARM/linker.ld",
+    "map_file": "IDE/GCC-ARM/Build/WolfCryptTest.map",
+    "linker_vars": ""
+  },
+  {
+    "target_name": "gcc-arm-cortex-m3",
+    "port": "gcc-arm",
+    "board": "cortex-m3",
+    "apt_packages": "gcc-arm-none-eabi libnewlib-arm-none-eabi libstdc++-arm-none-eabi-newlib",
+    "build_cmd": "test -f examples/configs/user_settings_tls12.h && mkdir -p IDE/GCC-ARM/Header-gen && { cat examples/configs/user_settings_tls12.h; printf '#define WOLFSSL_GENERAL_ALIGNMENT 4\\n#define SINGLE_THREADED\\n#define WOLFSSL_SMALL_STACK\\n#define NO_FILESYSTEM\\n#define NO_WRITEV\\n#define NO_MAIN_DRIVER\\n#define NO_DEV_RANDOM\\n#define BENCH_EMBEDDED\\n#define USE_CERT_BUFFERS_256\\n#define USE_CERT_BUFFERS_2048\\n#define WOLFSSL_IGNORE_FILE_WARN\\n#define WOLFSSL_USER_IO\\n#define WOLFSSL_USER_CURRTIME\\n#define TIME_OVERRIDES\\n#define USER_TICKS\\n#define XTIME my_time\\n#define XGMTIME my_gmtime\\n#define CUSTOM_RAND_TYPE unsigned int\\nextern unsigned int my_rng_seed_gen(void);\\n#undef CUSTOM_RAND_GENERATE\\n#define CUSTOM_RAND_GENERATE my_rng_seed_gen\\n#define HAVE_HASHDRBG\\n#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n'; } > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen ARCHFLAGS='-mcpu=cortex-m3 -mthumb -mabi=aapcs -DUSE_WOLF_ARM_STARTUP' CFLAGS_EXTRA='-Wno-cpp -DWOLFSSL_NO_SOCK' LDFLAGS='-mcpu=cortex-m3 -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20010000'",
+    "elf": "IDE/GCC-ARM/Build/WolfCryptTest.elf",
+    "ld": "IDE/GCC-ARM/linker.ld",
+    "map_file": "IDE/GCC-ARM/Build/WolfCryptTest.map",
+    "linker_vars": ""
+  },
+  {
+    "target_name": "gcc-arm-cortex-m7",
+    "port": "gcc-arm",
+    "board": "cortex-m7",
+    "apt_packages": "gcc-arm-none-eabi libnewlib-arm-none-eabi libstdc++-arm-none-eabi-newlib",
+    "build_cmd": "test -f IDE/GCC-ARM/Header/user_settings.h && mkdir -p IDE/GCC-ARM/Header-gen && { cat IDE/GCC-ARM/Header/user_settings.h; printf '#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n'; } > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen ARCHFLAGS='-mcpu=cortex-m7 -mthumb -mabi=aapcs -DUSE_WOLF_ARM_STARTUP' CFLAGS_EXTRA='-Wno-cpp -DWOLFCRYPT_ONLY -DWOLFSSL_NO_SOCK' LDFLAGS='-mcpu=cortex-m7 -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20010000'",
+    "elf": "IDE/GCC-ARM/Build/WolfCryptTest.elf",
+    "ld": "IDE/GCC-ARM/linker.ld",
+    "map_file": "IDE/GCC-ARM/Build/WolfCryptTest.map",
+    "linker_vars": ""
+  },
+  {
+    "target_name": "gcc-arm-cortex-m4-tls13",
+    "port": "gcc-arm",
+    "board": "cortex-m4-tls13",
+    "apt_packages": "gcc-arm-none-eabi libnewlib-arm-none-eabi libstdc++-arm-none-eabi-newlib",
+    "build_cmd": "test -f examples/configs/user_settings_tls13.h && mkdir -p IDE/GCC-ARM/Header-gen && { cat examples/configs/user_settings_tls13.h; printf '#define WOLFSSL_GENERAL_ALIGNMENT 4\\n#define SINGLE_THREADED\\n#define WOLFSSL_SMALL_STACK\\n#define NO_FILESYSTEM\\n#define NO_WRITEV\\n#define NO_MAIN_DRIVER\\n#define NO_DEV_RANDOM\\n#define BENCH_EMBEDDED\\n#define USE_CERT_BUFFERS_256\\n#define USE_CERT_BUFFERS_2048\\n#define WOLFSSL_IGNORE_FILE_WARN\\n#define WOLFSSL_USER_IO\\n#define WOLFSSL_USER_CURRTIME\\n#define TIME_OVERRIDES\\n#define USER_TICKS\\n#define XTIME my_time\\n#define XGMTIME my_gmtime\\n#define CUSTOM_RAND_TYPE unsigned int\\nextern unsigned int my_rng_seed_gen(void);\\n#undef CUSTOM_RAND_GENERATE\\n#define CUSTOM_RAND_GENERATE my_rng_seed_gen\\n#define HAVE_HASHDRBG\\n#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n'; } > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen CFLAGS_EXTRA='-Wno-cpp -DWOLFSSL_NO_SOCK' LDFLAGS='-mcpu=cortex-m4 -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20010000'",
+    "elf": "IDE/GCC-ARM/Build/WolfCryptTest.elf",
+    "ld": "IDE/GCC-ARM/linker.ld",
+    "map_file": "IDE/GCC-ARM/Build/WolfCryptTest.map",
+    "linker_vars": ""
+  },
+  {
+    "target_name": "gcc-arm-cortex-m4-dtls13",
+    "port": "gcc-arm",
+    "board": "cortex-m4-dtls13",
+    "apt_packages": "gcc-arm-none-eabi libnewlib-arm-none-eabi libstdc++-arm-none-eabi-newlib",
+    "build_cmd": "test -f examples/configs/user_settings_dtls13.h && mkdir -p IDE/GCC-ARM/Header-gen && { cat examples/configs/user_settings_dtls13.h; printf '#define WOLFSSL_GENERAL_ALIGNMENT 4\\n#define SINGLE_THREADED\\n#define WOLFSSL_SMALL_STACK\\n#define NO_FILESYSTEM\\n#define NO_WRITEV\\n#define NO_MAIN_DRIVER\\n#define NO_DEV_RANDOM\\n#define BENCH_EMBEDDED\\n#define USE_CERT_BUFFERS_256\\n#define USE_CERT_BUFFERS_2048\\n#define WOLFSSL_IGNORE_FILE_WARN\\n#define WOLFSSL_USER_IO\\n#define WOLFSSL_USER_CURRTIME\\n#define TIME_OVERRIDES\\n#define USER_TICKS\\n#define XTIME my_time\\n#define XGMTIME my_gmtime\\n#define CUSTOM_RAND_TYPE unsigned int\\nextern unsigned int my_rng_seed_gen(void);\\n#undef CUSTOM_RAND_GENERATE\\n#define CUSTOM_RAND_GENERATE my_rng_seed_gen\\n#define HAVE_HASHDRBG\\n#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n'; } > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen SRC_LD=-T./linker_large.ld CFLAGS_EXTRA='-Wno-cpp -DWOLFSSL_NO_SOCK' LDFLAGS='-mcpu=cortex-m4 -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20040000'",
+    "elf": "IDE/GCC-ARM/Build/WolfCryptTest.elf",
+    "ld": "IDE/GCC-ARM/linker_large.ld",
+    "map_file": "IDE/GCC-ARM/Build/WolfCryptTest.map",
+    "linker_vars": ""
+  },
+  {
+    "target_name": "gcc-arm-cortex-m4-pq",
+    "port": "gcc-arm",
+    "board": "cortex-m4-pq",
+    "apt_packages": "gcc-arm-none-eabi libnewlib-arm-none-eabi libstdc++-arm-none-eabi-newlib",
+    "build_cmd": "test -f examples/configs/user_settings_pq.h && mkdir -p IDE/GCC-ARM/Header-gen && { cat examples/configs/user_settings_pq.h; printf '#define WOLFSSL_GENERAL_ALIGNMENT 4\\n#define SINGLE_THREADED\\n#define WOLFSSL_SMALL_STACK\\n#define NO_FILESYSTEM\\n#define NO_WRITEV\\n#define NO_MAIN_DRIVER\\n#define NO_DEV_RANDOM\\n#define BENCH_EMBEDDED\\n#define USE_CERT_BUFFERS_256\\n#define USE_CERT_BUFFERS_2048\\n#define WOLFSSL_IGNORE_FILE_WARN\\n#define WOLFSSL_USER_IO\\n#define WOLFSSL_USER_CURRTIME\\n#define TIME_OVERRIDES\\n#define USER_TICKS\\n#define XTIME my_time\\n#define XGMTIME my_gmtime\\n#define CUSTOM_RAND_TYPE unsigned int\\nextern unsigned int my_rng_seed_gen(void);\\n#undef CUSTOM_RAND_GENERATE\\n#define CUSTOM_RAND_GENERATE my_rng_seed_gen\\n#define HAVE_HASHDRBG\\n#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n'; } > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen SRC_LD=-T./linker_large.ld CFLAGS_EXTRA='-Wno-cpp -DWOLFSSL_NO_SOCK' LDFLAGS='-mcpu=cortex-m4 -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20040000'",
+    "elf": "IDE/GCC-ARM/Build/WolfCryptTest.elf",
+    "ld": "IDE/GCC-ARM/linker_large.ld",
+    "map_file": "IDE/GCC-ARM/Build/WolfCryptTest.map",
+    "linker_vars": ""
+  },
+  {
+    "target_name": "gcc-arm-cortex-m4-rsa-only",
+    "port": "gcc-arm",
+    "board": "cortex-m4-rsa-only",
+    "apt_packages": "gcc-arm-none-eabi libnewlib-arm-none-eabi libstdc++-arm-none-eabi-newlib",
+    "build_cmd": "test -f examples/configs/user_settings_rsa_only.h && mkdir -p IDE/GCC-ARM/Header-gen && { cat examples/configs/user_settings_rsa_only.h; printf '#define WOLFSSL_GENERAL_ALIGNMENT 4\\n#define SINGLE_THREADED\\n#define WOLFSSL_SMALL_STACK\\n#define NO_FILESYSTEM\\n#define NO_WRITEV\\n#define NO_MAIN_DRIVER\\n#define NO_DEV_RANDOM\\n#define BENCH_EMBEDDED\\n#define USE_CERT_BUFFERS_2048\\n#define WOLFSSL_IGNORE_FILE_WARN\\n#define WOLFSSL_USER_IO\\n#define WOLFSSL_USER_CURRTIME\\n#define TIME_OVERRIDES\\n#define USER_TICKS\\n#define XTIME my_time\\n#define XGMTIME my_gmtime\\n#define CUSTOM_RAND_TYPE unsigned int\\nextern unsigned int my_rng_seed_gen(void);\\n#undef CUSTOM_RAND_GENERATE\\n#define CUSTOM_RAND_GENERATE my_rng_seed_gen\\n#define HAVE_HASHDRBG\\n#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n'; } > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen SRC_LD=-T./linker_large.ld CFLAGS_EXTRA='-Wno-cpp -DWOLFSSL_NO_SOCK' LDFLAGS='-mcpu=cortex-m4 -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20040000'",
+    "elf": "IDE/GCC-ARM/Build/WolfCryptTest.elf",
+    "ld": "IDE/GCC-ARM/linker_large.ld",
+    "map_file": "IDE/GCC-ARM/Build/WolfCryptTest.map",
+    "linker_vars": ""
+  },
+  {
+    "target_name": "gcc-arm-cortex-m4-pkcs7",
+    "port": "gcc-arm",
+    "board": "cortex-m4-pkcs7",
+    "apt_packages": "gcc-arm-none-eabi libnewlib-arm-none-eabi libstdc++-arm-none-eabi-newlib",
+    "build_cmd": "test -f examples/configs/user_settings_pkcs7.h && mkdir -p IDE/GCC-ARM/Header-gen && { cat examples/configs/user_settings_pkcs7.h; printf '#define WOLFSSL_GENERAL_ALIGNMENT 4\\n#define SINGLE_THREADED\\n#define WOLFSSL_SMALL_STACK\\n#define NO_FILESYSTEM\\n#define NO_WRITEV\\n#define NO_MAIN_DRIVER\\n#define NO_DEV_RANDOM\\n#define BENCH_EMBEDDED\\n#define USE_CERT_BUFFERS_256\\n#define USE_CERT_BUFFERS_2048\\n#define WOLFSSL_IGNORE_FILE_WARN\\n#define WOLFSSL_USER_IO\\n#define CUSTOM_RAND_TYPE unsigned int\\nextern unsigned int my_rng_seed_gen(void);\\n#undef CUSTOM_RAND_GENERATE\\n#define CUSTOM_RAND_GENERATE my_rng_seed_gen\\n#define HAVE_HASHDRBG\\n#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n'; } > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen CFLAGS_EXTRA='-Wno-cpp -DWOLFSSL_NO_SOCK -DWOLFCRYPT_ONLY' LDFLAGS='-mcpu=cortex-m4 -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20010000'",
+    "elf": "IDE/GCC-ARM/Build/WolfCryptTest.elf",
+    "ld": "IDE/GCC-ARM/linker.ld",
+    "map_file": "IDE/GCC-ARM/Build/WolfCryptTest.map",
+    "linker_vars": ""
+  },
+  {
+    "target_name": "gcc-arm-cortex-m4-openssl-compat",
+    "port": "gcc-arm",
+    "board": "cortex-m4-openssl-compat",
+    "apt_packages": "gcc-arm-none-eabi libnewlib-arm-none-eabi libstdc++-arm-none-eabi-newlib",
+    "build_cmd": "test -f examples/configs/user_settings_openssl_compat.h && mkdir -p IDE/GCC-ARM/Header-gen && { cat examples/configs/user_settings_openssl_compat.h; printf '#define WOLFSSL_GENERAL_ALIGNMENT 4\\n#define SINGLE_THREADED\\n#define WOLFSSL_SMALL_STACK\\n#define SMALL_SESSION_CACHE\\n#undef HAVE_OCSP\\n#undef HAVE_CERTIFICATE_STATUS_REQUEST\\n#undef HAVE_CERTIFICATE_STATUS_REQUEST_V2\\n#define NO_FILESYSTEM\\n#define NO_WRITEV\\n#define NO_MAIN_DRIVER\\n#define NO_DEV_RANDOM\\n#define BENCH_EMBEDDED\\n#define USE_CERT_BUFFERS_256\\n#define USE_CERT_BUFFERS_2048\\n#define WOLFSSL_IGNORE_FILE_WARN\\n#define WOLFSSL_USER_IO\\n#define USER_TICKS\\n#define CUSTOM_RAND_TYPE unsigned int\\nextern unsigned int my_rng_seed_gen(void);\\n#undef CUSTOM_RAND_GENERATE\\n#define CUSTOM_RAND_GENERATE my_rng_seed_gen\\n#define HAVE_HASHDRBG\\n#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n'; } > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen SRC_LD=-T./linker_large.ld CFLAGS_EXTRA='-Wno-cpp -DWOLFSSL_NO_SOCK' LDFLAGS='-mcpu=cortex-m4 -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20040000'",
+    "elf": "IDE/GCC-ARM/Build/WolfCryptTest.elf",
+    "ld": "IDE/GCC-ARM/linker_large.ld",
+    "map_file": "IDE/GCC-ARM/Build/WolfCryptTest.map",
+    "linker_vars": ""
+  },
+  {
+    "target_name": "gcc-arm-cortex-m4-sp-math",
+    "port": "gcc-arm",
+    "board": "cortex-m4-sp-math",
+    "apt_packages": "gcc-arm-none-eabi libnewlib-arm-none-eabi libstdc++-arm-none-eabi-newlib",
+    "build_cmd": "test -f examples/configs/user_settings_min_ecc.h && mkdir -p IDE/GCC-ARM/Header-gen && { cat examples/configs/user_settings_min_ecc.h; printf '#define WOLFSSL_GENERAL_ALIGNMENT 4\\n#define SINGLE_THREADED\\n#define WOLFSSL_SMALL_STACK\\n#define NO_FILESYSTEM\\n#define NO_WRITEV\\n#define NO_MAIN_DRIVER\\n#define NO_DEV_RANDOM\\n#define BENCH_EMBEDDED\\n#define USE_CERT_BUFFERS_256\\n#define WOLFSSL_IGNORE_FILE_WARN\\n#define WOLFSSL_USER_IO\\n#define WOLFSSL_USER_CURRTIME\\n#define TIME_OVERRIDES\\n#define USER_TICKS\\n#define XTIME my_time\\n#define XGMTIME my_gmtime\\n#define CUSTOM_RAND_TYPE unsigned int\\nextern unsigned int my_rng_seed_gen(void);\\n#undef CUSTOM_RAND_GENERATE\\n#define CUSTOM_RAND_GENERATE my_rng_seed_gen\\n#define HAVE_HASHDRBG\\n#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n#define WOLFSSL_SP_MATH\\n#define WOLFSSL_SP_NO_ASM\\n'; } > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen CFLAGS_EXTRA='-Wno-cpp -DWOLFSSL_NO_SOCK -DWOLFCRYPT_ONLY' LDFLAGS='-mcpu=cortex-m4 -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20010000'",
+    "elf": "IDE/GCC-ARM/Build/WolfCryptTest.elf",
+    "ld": "IDE/GCC-ARM/linker.ld",
+    "map_file": "IDE/GCC-ARM/Build/WolfCryptTest.map",
+    "linker_vars": ""
+  },
+  {
+    "target_name": "gcc-arm-cortex-m4-crypto-only",
+    "port": "gcc-arm",
+    "board": "cortex-m4-crypto-only",
+    "apt_packages": "gcc-arm-none-eabi libnewlib-arm-none-eabi libstdc++-arm-none-eabi-newlib",
+    "build_cmd": "mkdir -p IDE/GCC-ARM/Header-gen && printf '#ifndef WOLFSSL_USER_SETTINGS_H\\n#define WOLFSSL_USER_SETTINGS_H\\n#define WOLFCRYPT_ONLY\\n#define WOLFSSL_GENERAL_ALIGNMENT 4\\n#define SINGLE_THREADED\\n#define WOLFSSL_SMALL_STACK\\n#define NO_FILESYSTEM\\n#define NO_WRITEV\\n#define NO_MAIN_DRIVER\\n#define NO_DEV_RANDOM\\n#define BENCH_EMBEDDED\\n#define USE_CERT_BUFFERS_256\\n#define USE_CERT_BUFFERS_2048\\n#define WOLFSSL_IGNORE_FILE_WARN\\n#define WOLFSSL_USER_IO\\n#define WOLFSSL_USER_CURRTIME\\n#define TIME_OVERRIDES\\n#define USER_TICKS\\n#define XTIME my_time\\n#define XGMTIME my_gmtime\\n#define CUSTOM_RAND_TYPE unsigned int\\nextern unsigned int my_rng_seed_gen(void);\\n#undef CUSTOM_RAND_GENERATE\\n#define CUSTOM_RAND_GENERATE my_rng_seed_gen\\n#define HAVE_HASHDRBG\\n#define HAVE_AESGCM\\n#define HAVE_AES_DECRYPT\\n#define HAVE_ECC\\n#define HAVE_CHACHA\\n#define HAVE_POLY1305\\n#define WOLFSSL_SHA512\\n#define WOLFSSL_SHA384\\n#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n#endif\\n' > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen CFLAGS_EXTRA='-Wno-cpp -DWOLFSSL_NO_SOCK -DWOLFCRYPT_ONLY' LDFLAGS='-mcpu=cortex-m4 -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20010000'",
+    "elf": "IDE/GCC-ARM/Build/WolfCryptTest.elf",
+    "ld": "IDE/GCC-ARM/linker.ld",
+    "map_file": "IDE/GCC-ARM/Build/WolfCryptTest.map",
+    "linker_vars": ""
+  },
+  {
+    "target_name": "gcc-arm-cortex-m7-tls13",
+    "port": "gcc-arm",
+    "board": "cortex-m7-tls13",
+    "apt_packages": "gcc-arm-none-eabi libnewlib-arm-none-eabi libstdc++-arm-none-eabi-newlib",
+    "build_cmd": "test -f examples/configs/user_settings_tls13.h && mkdir -p IDE/GCC-ARM/Header-gen && { cat examples/configs/user_settings_tls13.h; printf '#define WOLFSSL_GENERAL_ALIGNMENT 4\\n#define SINGLE_THREADED\\n#define WOLFSSL_SMALL_STACK\\n#define NO_FILESYSTEM\\n#define NO_WRITEV\\n#define NO_MAIN_DRIVER\\n#define NO_DEV_RANDOM\\n#define BENCH_EMBEDDED\\n#define USE_CERT_BUFFERS_256\\n#define USE_CERT_BUFFERS_2048\\n#define WOLFSSL_IGNORE_FILE_WARN\\n#define WOLFSSL_USER_IO\\n#define WOLFSSL_USER_CURRTIME\\n#define TIME_OVERRIDES\\n#define USER_TICKS\\n#define XTIME my_time\\n#define XGMTIME my_gmtime\\n#define CUSTOM_RAND_TYPE unsigned int\\nextern unsigned int my_rng_seed_gen(void);\\n#undef CUSTOM_RAND_GENERATE\\n#define CUSTOM_RAND_GENERATE my_rng_seed_gen\\n#define HAVE_HASHDRBG\\n#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n'; } > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen ARCHFLAGS='-mcpu=cortex-m7 -mthumb -mabi=aapcs -DUSE_WOLF_ARM_STARTUP' CFLAGS_EXTRA='-Wno-cpp -DWOLFSSL_NO_SOCK' LDFLAGS='-mcpu=cortex-m7 -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20010000'",
+    "elf": "IDE/GCC-ARM/Build/WolfCryptTest.elf",
+    "ld": "IDE/GCC-ARM/linker.ld",
+    "map_file": "IDE/GCC-ARM/Build/WolfCryptTest.map",
+    "linker_vars": ""
+  },
+  {
+    "target_name": "gcc-arm-cortex-m7-pq",
+    "port": "gcc-arm",
+    "board": "cortex-m7-pq",
+    "apt_packages": "gcc-arm-none-eabi libnewlib-arm-none-eabi libstdc++-arm-none-eabi-newlib",
+    "build_cmd": "test -f examples/configs/user_settings_pq.h && mkdir -p IDE/GCC-ARM/Header-gen && { cat examples/configs/user_settings_pq.h; printf '#define WOLFSSL_GENERAL_ALIGNMENT 4\\n#define SINGLE_THREADED\\n#define WOLFSSL_SMALL_STACK\\n#define NO_FILESYSTEM\\n#define NO_WRITEV\\n#define NO_MAIN_DRIVER\\n#define NO_DEV_RANDOM\\n#define BENCH_EMBEDDED\\n#define USE_CERT_BUFFERS_256\\n#define USE_CERT_BUFFERS_2048\\n#define WOLFSSL_IGNORE_FILE_WARN\\n#define WOLFSSL_USER_IO\\n#define WOLFSSL_USER_CURRTIME\\n#define TIME_OVERRIDES\\n#define USER_TICKS\\n#define XTIME my_time\\n#define XGMTIME my_gmtime\\n#define CUSTOM_RAND_TYPE unsigned int\\nextern unsigned int my_rng_seed_gen(void);\\n#undef CUSTOM_RAND_GENERATE\\n#define CUSTOM_RAND_GENERATE my_rng_seed_gen\\n#define HAVE_HASHDRBG\\n#define NO_CRYPT_TEST\\n#define NO_CRYPT_BENCHMARK\\n'; } > IDE/GCC-ARM/Header-gen/user_settings.h && cd IDE/GCC-ARM && make -f Makefile.test TOOLCHAIN=arm-none-eabi- FIPS=0 USER_SETTINGS_DIR=./Header-gen ARCHFLAGS='-mcpu=cortex-m7 -mthumb -mabi=aapcs -DUSE_WOLF_ARM_STARTUP' SRC_LD=-T./linker_large.ld CFLAGS_EXTRA='-Wno-cpp -DWOLFSSL_NO_SOCK' LDFLAGS='-mcpu=cortex-m7 -mthumb -mabi=aapcs --specs=nosys.specs --specs=nano.specs -Wl,-Map=./Build/WolfCryptTest.map -Wl,-ereset_handler -flto -Wl,--defsym=__stack_process_end__=0x20040000'",
+    "elf": "IDE/GCC-ARM/Build/WolfCryptTest.elf",
+    "ld": "IDE/GCC-ARM/linker_large.ld",
+    "map_file": "IDE/GCC-ARM/Build/WolfCryptTest.map",
+    "linker_vars": ""
+  },
+  {
+    "target_name": "stm32-sim-stm32h753",
+    "port": "stm32-sim",
+    "board": "stm32h753",
+    "apt_packages": "build-essential ca-certificates cmake ninja-build python3 git gcc-arm-none-eabi libnewlib-arm-none-eabi libstdc++-arm-none-eabi-newlib wget unzip",
+    "build_cmd": "if [ ! -d simulators ]; then git clone --depth 1 https://github.com/wolfSSL/simulators simulators; fi && sudo mkdir -p /opt && if [ ! -d /opt/cmsis-device-h7 ]; then sudo git clone --depth 1 https://github.com/STMicroelectronics/cmsis-device-h7.git /opt/cmsis-device-h7; fi && if [ ! -d /opt/CMSIS_5 ]; then sudo git clone --depth 1 https://github.com/ARM-software/CMSIS_5.git /opt/CMSIS_5; fi && if [ ! -d /opt/STM32CubeH7 ]; then (sudo git clone --depth 1 --branch v1.11.2 --recurse-submodules https://github.com/STMicroelectronics/STM32CubeH7.git /opt/STM32CubeH7 || (sudo git clone --depth 1 --branch v1.11.2 https://github.com/STMicroelectronics/STM32CubeH7.git /opt/STM32CubeH7 && cd /opt/STM32CubeH7 && sudo git submodule update --init --recursive --depth 1)); fi && sudo rm -rf /opt/firmware-stm32sim-h7 /opt/wolfssl-stm32sim && sudo mkdir -p /opt/firmware-stm32sim-h7 && sudo cp -r simulators/STM32Sim/firmware/wolfcrypt-test-h7/. /opt/firmware-stm32sim-h7/ && sudo cp /opt/firmware-stm32sim-h7/stm32h7xx_hal_conf.h /opt/STM32CubeH7/Drivers/STM32H7xx_HAL_Driver/Inc/ && sudo cp -r . /opt/wolfssl-stm32sim && sudo rm -f /opt/wolfssl-stm32sim/config.h && cd /opt/firmware-stm32sim-h7 && sudo cmake -G Ninja -DWOLFSSL_USER_SETTINGS=ON -DUSER_SETTINGS_FILE=/opt/firmware-stm32sim-h7/user_settings.h -DCMAKE_TOOLCHAIN_FILE=/opt/firmware-stm32sim-h7/toolchain-arm-none-eabi.cmake -DCMAKE_BUILD_TYPE=Release -DWOLFSSL_CRYPT_TESTS=OFF -DWOLFSSL_EXAMPLES=OFF -DWOLFSSL_ROOT=/opt/wolfssl-stm32sim -B /opt/firmware-stm32sim-h7/build -S /opt/firmware-stm32sim-h7 && sudo cmake --build /opt/firmware-stm32sim-h7/build && sudo cp /opt/firmware-stm32sim-h7/build/wolfcrypt_test.elf $GITHUB_WORKSPACE/wolfcrypt_test.elf",
+    "elf": "wolfcrypt_test.elf",
+    "ld": "simulators/STM32Sim/firmware/wolfcrypt-test-h7/stm32h753.ld",
+    "linker_vars": ""
+  },
+  {
+    "target_name": "linuxkm-standard",
+    "port": "linuxkm",
+    "board": "linux-kernel-module-standard",
+    "apt_packages": "build-essential autoconf automake libtool linux-headers-$(uname -r)",
+    "build_cmd": "./autogen.sh && ./configure --with-linux-source=/lib/modules/$(uname -r)/build EXTRA_CPPFLAGS=-Werror --enable-option-checking=fatal --enable-linuxkm --enable-linuxkm-lkcapi-register=all --enable-all --enable-kyber=yes,original --enable-lms --enable-xmss --enable-dilithium --enable-experimental --enable-dual-alg-certs --disable-qt --disable-quic --with-sys-crypto-policy=no --disable-testcert --enable-all-asm --enable-crypttests --enable-linuxkm-benchmarks CFLAGS='-Wframe-larger-than=2048 -Wstack-usage=4096 -DBENCH_EMBEDDED -DBENCH_MIN_RUNTIME_SEC=0.01 -DBENCH_NTIMES=1 -DBENCH_AGREETIMES=1' --with-max-rsa-bits=16384 && make -j$(nproc) KERNEL_EXTRA_CFLAGS_REMOVE=-pg FORCE_NO_MODULE_SIG=1",
+    "elf": "linuxkm/libwolfssl.ko",
+    "ld": "linuxkm/wolfcrypt.lds",
+    "linker_vars": "",
+    "apt_cache": "false"
+  },
+  {
+    "target_name": "linuxkm-pie",
+    "port": "linuxkm",
+    "board": "linux-kernel-module-pie",
+    "apt_packages": "build-essential autoconf automake libtool linux-headers-$(uname -r)",
+    "build_cmd": "./autogen.sh && ./configure --with-linux-source=/lib/modules/$(uname -r)/build EXTRA_CPPFLAGS=-Werror --enable-option-checking=fatal --enable-linuxkm --enable-linuxkm-pie --enable-reproducible-build --enable-linuxkm-lkcapi-register=all --enable-all-crypto --enable-cryptonly --enable-kyber=yes,original --enable-lms --enable-xmss --enable-dilithium --enable-experimental --disable-qt --disable-quic --with-sys-crypto-policy=no --disable-opensslextra --disable-testcert --enable-intelasm --disable-sp-asm --enable-crypttests --enable-linuxkm-benchmarks CFLAGS='-DWOLFSSL_LINUXKM_VERBOSE_DEBUG -DDEBUG_LINUXKM_PIE_SUPPORT -Wframe-larger-than=2048 -Wstack-usage=4096 -DBENCH_EMBEDDED -DBENCH_MIN_RUNTIME_SEC=0.01 -DBENCH_NTIMES=1 -DBENCH_AGREETIMES=1' --with-max-rsa-bits=16384 && make -j$(nproc) KERNEL_EXTRA_CFLAGS_REMOVE=-pg FORCE_NO_MODULE_SIG=1",
+    "elf": "linuxkm/libwolfssl.ko",
+    "ld": "linuxkm/wolfcrypt.lds",
+    "linker_vars": "",
+    "apt_cache": "false"
   }
 ]
diff --git a/.github/scripts/zephyr-4.x/zephyr-test.sh b/.github/scripts/zephyr-4.x/zephyr-test.sh
index c65e6950f7..fa0e827d4d 100755
--- a/.github/scripts/zephyr-4.x/zephyr-test.sh
+++ b/.github/scripts/zephyr-4.x/zephyr-test.sh
@@ -112,6 +112,9 @@ LOG_DIR="${SCRIPT_DIR}/logs"
 mkdir -p "${LOG_DIR}"
 TIMESTAMP=$(date +%Y%m%d_%H%M%S)
 LOG_FILE="${LOG_DIR}/${BOARD_SLUG}_${TIMESTAMP}.log"
+ARTIFACTS_DIR="${SCRIPT_DIR}/artifacts/${BOARD_SLUG}-${SAMPLE_NAME}"
+mkdir -p "${ARTIFACTS_DIR}"
+chmod 0755 "${ARTIFACTS_DIR}"
 
 echo "==> wolfSSL repo:   ${WOLFSSL_REPO}"
 echo "==> wolfSSL branch: ${WOLFSSL_BRANCH}"
@@ -256,6 +259,23 @@ else
     echo ""
     echo "==> [container] Build succeeded!"
 
+    # Stage Membrowse-relevant artifacts on the host-mounted volume.
+    # /artifacts is bind-mounted by the host wrapper; if it isn't writable
+    # (e.g. interactive runs without the mount), skip silently.
+    if [[ -d /artifacts && -w /artifacts ]]; then
+        BUILD_OUT="${WORKDIR}/zephyrproject/build/zephyr"
+        if [[ -f "${BUILD_OUT}/zephyr.elf" ]]; then
+            cp "${BUILD_OUT}/zephyr.elf" /artifacts/zephyr.elf
+        fi
+        if [[ -f "${BUILD_OUT}/linker.cmd" ]]; then
+            cp "${BUILD_OUT}/linker.cmd" /artifacts/linker.cmd
+        fi
+        # Map file enables Membrowse library/object attribution
+        if [[ -f "${BUILD_OUT}/zephyr.map" ]]; then
+            cp "${BUILD_OUT}/zephyr.map" /artifacts/zephyr.map
+        fi
+    fi
+
     # Run the app for emulator targets and watch for completion
     case "${BOARD_TARGET}" in
         native_sim*|qemu_*)
@@ -342,6 +362,7 @@ docker rm -f "${CONTAINER_NAME}" 2>/dev/null || true
 DOCKER_ARGS=(
     --name "${CONTAINER_NAME}"
     --rm
+    -v "${ARTIFACTS_DIR}:/artifacts"
 )
 
 if [[ "$INTERACTIVE" == "1" ]]; then
diff --git a/.github/workflows/linuxkm.yml b/.github/workflows/linuxkm.yml
index e785cf56cb..c2a7ff9242 100644
--- a/.github/workflows/linuxkm.yml
+++ b/.github/workflows/linuxkm.yml
@@ -19,7 +19,8 @@ jobs:
       matrix:
         config: [
           'EXTRA_CPPFLAGS=-Werror --enable-option-checking=fatal --enable-linuxkm --enable-linuxkm-lkcapi-register=all --enable-all --enable-kyber=yes,original --enable-lms --enable-xmss --enable-dilithium --enable-experimental --enable-dual-alg-certs --disable-qt --disable-quic --with-sys-crypto-policy=no --disable-testcert --enable-all-asm --enable-crypttests --enable-linuxkm-benchmarks CFLAGS="-Wframe-larger-than=2048 -Wstack-usage=4096 -DBENCH_EMBEDDED -DBENCH_MIN_RUNTIME_SEC=0.01 -DBENCH_NTIMES=1 -DBENCH_AGREETIMES=1" --with-max-rsa-bits=16384',
-          'EXTRA_CPPFLAGS=-Werror --enable-option-checking=fatal --enable-linuxkm --enable-linuxkm-pie --enable-reproducible-build --enable-linuxkm-lkcapi-register=all --enable-all-crypto --enable-cryptonly --enable-kyber=yes,original --enable-lms --enable-xmss --enable-dilithium --enable-experimental --disable-qt --disable-quic --with-sys-crypto-policy=no --disable-opensslextra --disable-testcert --enable-intelasm --disable-sp-asm --enable-crypttests --enable-linuxkm-benchmarks CFLAGS="-DWOLFSSL_LINUXKM_VERBOSE_DEBUG -DDEBUG_LINUXKM_PIE_SUPPORT -Wframe-larger-than=2048 -Wstack-usage=4096 -DBENCH_EMBEDDED -DBENCH_MIN_RUNTIME_SEC=0.01 -DBENCH_NTIMES=1 -DBENCH_AGREETIMES=1" --with-max-rsa-bits=16384'
+          'EXTRA_CPPFLAGS=-Werror --enable-option-checking=fatal --enable-linuxkm --enable-linuxkm-pie --enable-reproducible-build --enable-linuxkm-lkcapi-register=all --enable-all-crypto --enable-cryptonly --enable-kyber=yes,original --enable-lms --enable-xmss --enable-dilithium --enable-experimental --disable-qt --disable-quic --with-sys-crypto-policy=no --disable-opensslextra --disable-testcert --enable-intelasm --disable-sp-asm --enable-crypttests --enable-linuxkm-benchmarks CFLAGS="-DWOLFSSL_LINUXKM_VERBOSE_DEBUG -DDEBUG_LINUXKM_PIE_SUPPORT -Wframe-larger-than=2048 -Wstack-usage=4096 -DBENCH_EMBEDDED -DBENCH_MIN_RUNTIME_SEC=0.01 -DBENCH_NTIMES=1 -DBENCH_AGREETIMES=1" --with-max-rsa-bits=16384',
+          'EXTRA_CPPFLAGS=-Werror --enable-option-checking=fatal --enable-linuxkm --enable-tls13 --enable-dtls --enable-dtls13'
         ]
     name: build module
     if: ${{ (github.repository_owner == 'wolfssl') && (github.event_name != 'pull_request' || github.event.pull_request.draft == false) }}
@@ -55,3 +56,15 @@ jobs:
           make -j 4 KERNEL_EXTRA_CFLAGS_REMOVE=-pg FORCE_NO_MODULE_SIG=1 || $(exit 11)
           ls -l linuxkm/libwolfssl.ko || $(exit 12)
           echo "Successful linuxkm build."
+
+      - name: Verify DTLS 1.3 symbols are exported (when WOLFSSL_DTLS13 is configured)
+        run: |
+          if grep -q '^#define WOLFSSL_DTLS13' wolfssl/options.h; then
+            echo "WOLFSSL_DTLS13 defined; checking GPL exports in libwolfssl.ko..."
+            objdump -t linuxkm/libwolfssl.ko \
+              | grep -qE '__ksymtab_wolfDTLSv1_3_(client|server)_method$' \
+              || { echo "::error::DTLS 1.3 entry points not exported from libwolfssl.ko"; exit 13; }
+            echo "DTLS 1.3 export check: PASS"
+          else
+            echo "WOLFSSL_DTLS13 not defined for this matrix entry; skipping symbol check."
+          fi
diff --git a/.github/workflows/membrowse-comment.yml b/.github/workflows/membrowse-comment.yml
new file mode 100644
index 0000000000..1357bca7d9
--- /dev/null
+++ b/.github/workflows/membrowse-comment.yml
@@ -0,0 +1,32 @@
+name: Membrowse Comment
+
+on:
+  workflow_run:
+    workflows: [Membrowse Memory Report]
+    types:
+      - completed
+
+jobs:
+  post-comment:
+    runs-on: ubuntu-24.04
+    timeout-minutes: 10
+    # Run the comment job even if some of the builds fail
+    if: >
+      github.event.workflow_run.event == 'pull_request' &&
+      github.event.workflow_run.conclusion != 'cancelled'
+    permissions:
+      contents: read
+      pull-requests: write
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v5
+
+      - name: Post Membrowse PR comment
+        if: ${{ env.MEMBROWSE_API_KEY != '' }}
+        uses: membrowse/membrowse-action/comment-action@v1
+        with:
+          api_key: ${{ secrets.MEMBROWSE_API_KEY }}
+          commit: ${{ github.event.workflow_run.head_sha }}
+        env:
+          MEMBROWSE_API_KEY: ${{ secrets.MEMBROWSE_API_KEY }}
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
diff --git a/.github/workflows/membrowse-onboard.yml b/.github/workflows/membrowse-onboard.yml
index 503942b75f..978a23c132 100644
--- a/.github/workflows/membrowse-onboard.yml
+++ b/.github/workflows/membrowse-onboard.yml
@@ -11,6 +11,9 @@ on:
 
 jobs:
   load-targets:
+    # Only run from the wolfssl org to avoid burning forks' CI minutes
+    # and reporting fork builds to the membrowse backend.
+    if: github.repository_owner == 'wolfssl'
     runs-on: ubuntu-24.04
     timeout-minutes: 10
     outputs:
@@ -25,12 +28,13 @@ jobs:
 
   onboard:
     needs: load-targets
+    if: github.repository_owner == 'wolfssl'
     runs-on: ubuntu-24.04
-    timeout-minutes: 10
+    timeout-minutes: 30
     strategy:
       fail-fast: false
       matrix:
-        include: ${{ fromJson(needs.load-targets.outputs.matrix) }}
+        include: ${{ fromJson(needs.load-targets.outputs.matrix || '[]') }}
 
     steps:
       - name: Checkout repository
@@ -43,6 +47,7 @@ jobs:
         uses: ./.github/actions/install-apt-deps
         with:
           packages: ${{ matrix.apt_packages }}
+          cache: ${{ matrix.apt_cache || 'true' }}
 
       - name: Run Membrowse Onboard Action
         uses: membrowse/membrowse-action/onboard-action@v1
diff --git a/.github/workflows/membrowse-report.yml b/.github/workflows/membrowse-report.yml
index 40a0885183..201fbed2e8 100644
--- a/.github/workflows/membrowse-report.yml
+++ b/.github/workflows/membrowse-report.yml
@@ -1,20 +1,21 @@
 name: Membrowse Memory Report
 
-# Runs nightly instead of per-PR - the report is for trend tracking, not
-# gating individual PRs, and the build matrix is too heavy to run on every
-# push.  Use workflow_dispatch to trigger an ad-hoc run.
-
 on:
-  schedule:
-    - cron: '0 4 * * *'   # daily at 04:00 UTC
+  pull_request:
+    types: [opened, synchronize, reopened, ready_for_review]
+  push:
+    branches: [master]
   workflow_dispatch:
 
 concurrency:
-  group: ${{ github.workflow }}
-  cancel-in-progress: false
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: ${{ github.ref != 'refs/heads/master' }}
 
 jobs:
   load-targets:
+    # Only run from the wolfssl org to avoid burning forks' CI minutes
+    # and reporting fork builds to the membrowse backend.
+    if: github.repository_owner == 'wolfssl' && (github.event_name != 'pull_request' || github.event.pull_request.draft == false)
     runs-on: ubuntu-24.04
     timeout-minutes: 10
     outputs:
@@ -27,28 +28,68 @@ jobs:
         id: set-matrix
         run: echo "matrix=$(jq -c '.' .github/membrowse-targets.json)" >> $GITHUB_OUTPUT
 
-  analyze:
-    needs: load-targets
+  check-changes:
+    if: github.repository_owner == 'wolfssl' && (github.event_name != 'pull_request' || github.event.pull_request.draft == false)
     runs-on: ubuntu-24.04
-    timeout-minutes: 10
+    timeout-minutes: 5
+    outputs:
+      needs_build: ${{ steps.filter.outputs.code == 'true' || github.event_name == 'workflow_dispatch' }}
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v5
+
+      - name: Detect binary-affecting changes
+        id: filter
+        if: github.event_name == 'pull_request' || github.event_name == 'push'
+        uses: dorny/paths-filter@v4
+        with:
+          predicate-quantifier: every
+          filters: |
+            code:
+              - '!**.md'
+              - '!**/README*'
+              - '!doc/**'
+              - '!AUTHORS'
+              - '!COPYING*'
+              - '!LICENSE*'
+              - '!LICENSING'
+              - '!INSTALL'
+              - '!ChangeLog*'
+              - '!SCRIPTS-LIST'
+              - '!.gitignore'
+              - '!.editorconfig'
+              - '!.codespellexcludelines'
+              - '!.cyignore'
+              - '!.wolfssl_known_macro_extras'
+              - '!.github/ISSUE_TEMPLATE/**'
+              - '!.github/PULL_REQUEST_TEMPLATE.md'
+
+  analyze:
+    needs: [load-targets, check-changes]
+    if: github.repository_owner == 'wolfssl' && (github.event_name != 'pull_request' || needs.check-changes.outputs.needs_build == 'true')
+    runs-on: ubuntu-24.04
+    timeout-minutes: 30
     strategy:
       fail-fast: false
       matrix:
-        include: ${{ fromJson(needs.load-targets.outputs.matrix) }}
+        include: ${{ fromJson(needs.load-targets.outputs.matrix || '[]') }}
 
     steps:
       - name: Checkout repository
         uses: actions/checkout@v5
         with:
-          fetch-depth: 0
-          submodules: recursive
+          fetch-depth: 2
+          submodules: ${{ needs.check-changes.outputs.needs_build == 'true' && 'recursive' || 'false' }}
 
       - name: Install packages
+        if: needs.check-changes.outputs.needs_build == 'true'
         uses: ./.github/actions/install-apt-deps
         with:
           packages: ${{ matrix.apt_packages }}
+          cache: ${{ matrix.apt_cache || 'true' }}
 
       - name: Build firmware
+        if: needs.check-changes.outputs.needs_build == 'true'
         run: ${{ matrix.build_cmd }}
 
       - name: Run Membrowse PR Action
@@ -56,9 +97,11 @@ jobs:
         uses: membrowse/membrowse-action@v1
         with:
           target_name: ${{ matrix.target_name }}
-          elf: ${{ matrix.elf }}
-          ld: ${{ matrix.ld }}
+          elf: ${{ needs.check-changes.outputs.needs_build == 'true' && matrix.elf || '' }}
+          ld: ${{ needs.check-changes.outputs.needs_build == 'true' && matrix.ld || '' }}
+          map_file: ${{ needs.check-changes.outputs.needs_build == 'true' && matrix.map_file || '' }}
           linker_vars: ${{ matrix.linker_vars }}
           api_key: ${{ secrets.MEMBROWSE_API_KEY }}
           api_url: ${{ vars.MEMBROWSE_API_URL }}
+          identical: ${{ needs.check-changes.outputs.needs_build != 'true' }}
           verbose: INFO
diff --git a/.github/workflows/membrowse-zephyr.yml b/.github/workflows/membrowse-zephyr.yml
new file mode 100644
index 0000000000..993581edc0
--- /dev/null
+++ b/.github/workflows/membrowse-zephyr.yml
@@ -0,0 +1,115 @@
+name: Membrowse Zephyr Report
+
+# Triggered after the heavy Zephyr 4.x test workflow completes. Pulls the
+# pre-built zephyr.elf, zephyr.map and linker.cmd artifacts staged by
+# zephyr-test.sh and feeds them to the Membrowse memory-tracking service.
+# This avoids duplicating the (slow) Zephyr build inside the Membrowse
+# matrix.
+
+on:
+  workflow_run:
+    workflows: [Zephyr 4.x tests]
+    types:
+      - completed
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.event.workflow_run.head_sha }}
+  cancel-in-progress: true
+
+jobs:
+  analyze:
+    runs-on: ubuntu-24.04
+    timeout-minutes: 10
+    # Only run from the wolfssl org to avoid burning forks' CI minutes
+    # and reporting fork builds to the membrowse backend.
+    if: >
+      github.event.workflow_run.conclusion == 'success' &&
+      github.repository_owner == 'wolfssl'
+    permissions:
+      contents: read
+      actions: read
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - target_name: zephyr-native_sim
+            artifact: membrowse-zephyr-native_sim
+          - target_name: zephyr-frdm_rw612
+            artifact: membrowse-zephyr-frdm_rw612
+    steps:
+      # Check out the commit the Zephyr workflow actually built so Membrowse
+      # attributes the report to the right commit. Only the last 2 commits
+      # are needed (current + parent) to resolve the base for comparison.
+      - name: Checkout repository
+        uses: actions/checkout@v5
+        with:
+          ref: ${{ github.event.workflow_run.head_sha }}
+          fetch-depth: 2
+
+      - name: Download Zephyr build artifact
+        id: download
+        uses: actions/download-artifact@v4
+        with:
+          name: ${{ matrix.artifact }}
+          path: zephyr-artifacts/${{ matrix.target_name }}
+          run-id: ${{ github.event.workflow_run.id }}
+          github-token: ${{ secrets.GITHUB_TOKEN }}
+        continue-on-error: true
+
+      - name: Verify artifact present
+        id: verify
+        run: |
+          ELF="zephyr-artifacts/${{ matrix.target_name }}/zephyr.elf"
+          LD="zephyr-artifacts/${{ matrix.target_name }}/linker.cmd"
+          MAP="zephyr-artifacts/${{ matrix.target_name }}/zephyr.map"
+          if [[ -f "$ELF" && -f "$LD" ]]; then
+            echo "have_artifacts=true" >> "$GITHUB_OUTPUT"
+          else
+            echo "have_artifacts=false" >> "$GITHUB_OUTPUT"
+            echo "::warning::Membrowse artifact for ${{ matrix.target_name }} not found; the matching cell of zephyr-4.x.yml may have been skipped or excluded."
+          fi
+          # The map file is optional; it enables library/object attribution.
+          if [[ -f "$MAP" ]]; then
+            echo "map_file=$MAP" >> "$GITHUB_OUTPUT"
+          else
+            echo "map_file=" >> "$GITHUB_OUTPUT"
+          fi
+
+      - name: Run Membrowse PR Action
+        if: steps.verify.outputs.have_artifacts == 'true'
+        uses: membrowse/membrowse-action@v1
+        with:
+          target_name: ${{ matrix.target_name }}
+          elf: zephyr-artifacts/${{ matrix.target_name }}/zephyr.elf
+          ld: zephyr-artifacts/${{ matrix.target_name }}/linker.cmd
+          map_file: ${{ steps.verify.outputs.map_file }}
+          linker_vars: ""
+          api_key: ${{ secrets.MEMBROWSE_API_KEY }}
+          api_url: ${{ vars.MEMBROWSE_API_URL }}
+          verbose: INFO
+
+  # Refresh the consolidated Membrowse PR comment after the Zephyr targets
+  # have been submitted, mirroring membrowse-comment.yml.
+  post-comment:
+    needs: analyze
+    runs-on: ubuntu-24.04
+    timeout-minutes: 10
+    if: >
+      github.event.workflow_run.event == 'pull_request' &&
+      github.repository_owner == 'wolfssl'
+    permissions:
+      contents: read
+      pull-requests: write
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v5
+
+      - name: Post Membrowse PR comment
+        if: ${{ env.MEMBROWSE_API_KEY != '' }}
+        uses: membrowse/membrowse-action/comment-action@v1
+        with:
+          api_key: ${{ secrets.MEMBROWSE_API_KEY }}
+          commit: ${{ github.event.workflow_run.head_sha }}
+        env:
+          MEMBROWSE_API_KEY: ${{ secrets.MEMBROWSE_API_KEY }}
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
diff --git a/.github/workflows/os-check.yml b/.github/workflows/os-check.yml
index 5c1957e4da..1519bd0c88 100644
--- a/.github/workflows/os-check.yml
+++ b/.github/workflows/os-check.yml
@@ -117,14 +117,14 @@ jobs:
           # variants of mod_exp_<words>_nb / RSA / DH wrappers.
           '--enable-curve25519=nonblock --enable-ecc=nonblock --enable-rsa=nonblock --enable-dh=nonblock --enable-sp=yes,nonblock CPPFLAGS="-DWOLFSSL_PUBLIC_MP -DWOLFSSL_DEBUG_NONBLOCK -DRSA_LOW_MEM -DSP_WORD_SIZE=32"',
           '--enable-certreq --enable-certext --enable-certgen --disable-secure-renegotiation-info CPPFLAGS="-DNO_TLS"',
-          # Minimal DTLS 1.3 client-only build. The SHA-224/384/512/3
-          # disables are deliberately omitted: --disable-sha384 alone
-          # trips a pre-existing wolfSSL bug in
-          # test_tls13_duplicate_extension (reproducible on clean master).
+          # Minimal DTLS 1.3 client-only build with the SHA-224/384/512/3
+          # hash families disabled. SHA-256 (used by TLS_AES_128_GCM_SHA256)
+          # and SHA-1 remain enabled.
           '--enable-dtls13 --disable-tlsv12 --disable-oldtls --disable-rsa --disable-dh
             --disable-aescbc --disable-aesecb --disable-md5 --disable-chacha
             --disable-poly1305 --disable-errorstrings --disable-asn-print
             --disable-eccshamir --disable-base64encode --disable-coding --disable-sni
+            --disable-sha224 --disable-sha384 --disable-sha512 --disable-sha3
             --enable-aesgcm=small --enable-sp-math --enable-sp=smallec256 --disable-sp-asm
             CPPFLAGS=''-DNO_WOLFSSL_SERVER -DWOLFSSL_NO_TLS12 -DNO_SESSION_CACHE
             -DWOLFSSL_AES_NO_UNROLL -DUSE_SLOW_SHA256 -DWOLFSSL_NO_ASYNC_IO
diff --git a/.github/workflows/pic32mz-sim.yml b/.github/workflows/pic32mz-sim.yml
new file mode 100644
index 0000000000..51625cd43a
--- /dev/null
+++ b/.github/workflows/pic32mz-sim.yml
@@ -0,0 +1,86 @@
+name: PIC32MZ simulator test
+
+# START OF COMMON SECTION
+on:
+  push:
+    branches: [ 'master', 'main', 'release/**' ]
+  pull_request:
+    branches: [ '*' ]
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+# END OF COMMON SECTION
+
+# Build the PIC32MZ software simulator (https://github.com/wolfSSL/simulators,
+# PIC32MZSim/ subdirectory) and run the wolfCrypt test suite on emulated
+# PIC32MZ EC (no FPU, CE ignores OUT_SWAP) and EF (FPU + OUT_SWAP) parts,
+# through both the direct-register PIC32 port and the MPLAB Harmony 3 driver
+# port.
+#
+# Like stm32-sim.yml, the Dockerfiles read wolfSSL from /opt/wolfssl at
+# runtime via a bind mount, so no Dockerfile patching is required - the PR
+# checkout is mounted directly.
+
+jobs:
+  pic32mz_sim:
+    name: wolfCrypt on PIC32MZ ${{ matrix.chip_label }} (${{ matrix.port_label }})
+    if: github.repository_owner == 'wolfssl'
+    runs-on: ubuntu-24.04
+    timeout-minutes: 30
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - port_label: direct
+            chip_label: EC
+            dockerfile: Dockerfile.wolfcrypt-direct
+            image_tag: pic32mz-wolfcrypt-direct:ci
+            script: run-wolfcrypt-direct-ec.sh
+            cache_scope: pic32mz-direct
+          - port_label: direct
+            chip_label: EF
+            dockerfile: Dockerfile.wolfcrypt-direct
+            image_tag: pic32mz-wolfcrypt-direct:ci
+            script: run-wolfcrypt-direct-ef.sh
+            cache_scope: pic32mz-direct
+          - port_label: harmony
+            chip_label: EC
+            dockerfile: Dockerfile.wolfcrypt-harmony
+            image_tag: pic32mz-wolfcrypt-harmony:ci
+            script: run-wolfcrypt-harmony-ec.sh
+            cache_scope: pic32mz-harmony
+          - port_label: harmony
+            chip_label: EF
+            dockerfile: Dockerfile.wolfcrypt-harmony
+            image_tag: pic32mz-wolfcrypt-harmony:ci
+            script: run-wolfcrypt-harmony-ef.sh
+            cache_scope: pic32mz-harmony
+    steps:
+      - name: Checkout wolfSSL (PR source)
+        uses: actions/checkout@v4
+        with:
+          path: wolfssl
+
+      - name: Clone PIC32MZ simulator
+        run: git clone --depth 1 https://github.com/wolfSSL/simulators simulators
+
+      - uses: docker/setup-buildx-action@v3
+
+      - name: Build ${{ matrix.image_tag }} image
+        uses: docker/build-push-action@v5
+        with:
+          context: simulators/PIC32MZSim
+          file: simulators/PIC32MZSim/${{ matrix.dockerfile }}
+          push: false
+          load: true
+          tags: ${{ matrix.image_tag }}
+          cache-from: type=gha,scope=${{ matrix.cache_scope }}
+          cache-to: type=gha,mode=max,scope=${{ matrix.cache_scope }}
+
+      - name: Run wolfCrypt tests on PIC32MZ ${{ matrix.chip_label }} (${{ matrix.port_label }})
+        run: |
+          docker run --rm \
+            -v "${{ github.workspace }}/wolfssl:/opt/wolfssl:ro" \
+            ${{ matrix.image_tag }} \
+            /app/scripts/${{ matrix.script }}
diff --git a/.github/workflows/zephyr-4.x.yml b/.github/workflows/zephyr-4.x.yml
index ac5358dff7..afbbde5ebf 100644
--- a/.github/workflows/zephyr-4.x.yml
+++ b/.github/workflows/zephyr-4.x.yml
@@ -79,3 +79,22 @@ jobs:
           path: .github/scripts/zephyr-4.x/logs/
           retention-days: 5
           if-no-files-found: ignore
+
+      # Stage Membrowse artifacts only for the cells we want to track on the
+      # Membrowse dashboard. Keep this tight to avoid duplicate target_name
+      # entries on the dashboard and unnecessary artifact uploads.
+      - name: Upload Membrowse artifacts (selected cells only)
+        if: >
+          success() &&
+          matrix.zephyr-ref == 'v4.3.0' &&
+          matrix.sample == 'wolfssl_test' &&
+          matrix.extra-conf == ''
+        uses: actions/upload-artifact@v4
+        with:
+          name: membrowse-zephyr-${{ matrix.board == 'native_sim' && 'native_sim' || 'frdm_rw612' }}
+          path: |
+            .github/scripts/zephyr-4.x/artifacts/${{ matrix.board == 'native_sim' && 'native_sim' || 'frdm_rw612-rw612' }}-wolfssl_test/zephyr.elf
+            .github/scripts/zephyr-4.x/artifacts/${{ matrix.board == 'native_sim' && 'native_sim' || 'frdm_rw612-rw612' }}-wolfssl_test/linker.cmd
+            .github/scripts/zephyr-4.x/artifacts/${{ matrix.board == 'native_sim' && 'native_sim' || 'frdm_rw612-rw612' }}-wolfssl_test/zephyr.map
+          if-no-files-found: warn
+          retention-days: 1
diff --git a/.wolfssl_known_macro_extras b/.wolfssl_known_macro_extras
index c4a2620543..fe9146dca7 100644
--- a/.wolfssl_known_macro_extras
+++ b/.wolfssl_known_macro_extras
@@ -741,10 +741,10 @@ WOLFSSL_CONTIKI
 WOLFSSL_CRL_ALLOW_MISSING_CDP
 WOLFSSL_DISABLE_EARLY_SANITY_CHECKS
 WOLFSSL_DRBG_SHA256
+WOLFSSL_DTLS13_ECHO_LEGACY_SESSION_ID
 WOLFSSL_DTLS_DISALLOW_FUTURE
 WOLFSSL_DTLS_RECORDS_CAN_SPAN_DATAGRAMS
 WOLFSSL_DTLS_RESEND_ONLY_TIMEOUT
-WOLFSSL_DTLS13_ECHO_LEGACY_SESSION_ID
 WOLFSSL_DUMP_MEMIO_STREAM
 WOLFSSL_DUP_CERTPOL
 WOLFSSL_EARLY_DATA_NO_ANTI_REPLAY
@@ -869,6 +869,7 @@ WOLFSSL_PASSTHRU_ERR
 WOLFSSL_PB
 WOLFSSL_PEER_ADDRESS_CHANGES
 WOLFSSL_PKCS11_RW_TOKENS
+WOLFSSL_PPC64_ASM_AES_NO_HARDEN
 WOLFSSL_PRCONNECT_PRO
 WOLFSSL_PREFIX
 WOLFSSL_PSA_NO_AES
@@ -888,6 +889,7 @@ WOLFSSL_RENESAS_RZN2L
 WOLFSSL_RENESAS_TLS
 WOLFSSL_RENESAS_TSIP_IAREWRX
 WOLFSSL_REQUIRE_TCA
+WOLFSSL_RISCV_ASM_NO_UNALIGNED
 WOLFSSL_RNG_USE_FULL_SEED
 WOLFSSL_RSA_CHECK_D_ON_DECRYPT
 WOLFSSL_RSA_DECRYPT_TO_0_LEN
@@ -976,6 +978,7 @@ XSECURE_CACHE_DISABLE
 _ABI64
 _ABIO64
 _ARCH_PPC64
+_ARCH_PWR8
 _COMPILER_VERSION
 _INTPTR_T_DECLARED
 _LINUX_REFCOUNT_H
@@ -1004,6 +1007,7 @@ __32MZ2048ECM144__
 __32MZ2048EFM144__
 __ANDROID__
 __APPLE__
+__ARCH_PWR8
 __ARCH_STRCASECMP_NO_REDIRECT
 __ARCH_STRCMP_NO_REDIRECT
 __ARCH_STRNCASECMP_NO_REDIRECT
diff --git a/CMakeLists.txt b/CMakeLists.txt
index 4a2c0e9750..217a9e1574 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -2942,11 +2942,14 @@ if(WOLFSSL_EXAMPLES)
         tests/api/test_mldsa_legacy.c
         tests/api/test_slhdsa.c
         tests/api/test_signature.c
+        tests/api/test_lms_xmss.c
         tests/api/test_dtls.c
+        tests/api/test_dtls13.c
         tests/api/test_ocsp.c
         tests/api/test_evp.c
         tests/api/test_tls_ext.c
         tests/api/test_tls.c
+        tests/api/test_session.c
         tests/api/test_x509.c
         tests/api/test_asn.c
         tests/api/test_pkcs7.c
diff --git a/IDE/GCC-ARM/Header/user_settings.h b/IDE/GCC-ARM/Header/user_settings.h
index a3a1561b0e..971180c9bd 100644
--- a/IDE/GCC-ARM/Header/user_settings.h
+++ b/IDE/GCC-ARM/Header/user_settings.h
@@ -469,10 +469,10 @@ extern "C" {
 /* Override Current Time */
 /* Allows custom "custom_time()" function to be used for benchmark */
 #define WOLFSSL_USER_CURRTIME
-#define WOLFSSL_GMTIME
+#define TIME_OVERRIDES
 #define USER_TICKS
-extern unsigned long my_time(unsigned long* timer);
 #define XTIME my_time
+#define XGMTIME my_gmtime
 
 
 /* ------------------------------------------------------------------------- */
@@ -656,4 +656,3 @@ extern unsigned int my_rng_seed_gen(void);
 #endif
 
 #endif /* WOLFSSL_USER_SETTINGS_H */
-
diff --git a/IDE/GCC-ARM/Makefile.common b/IDE/GCC-ARM/Makefile.common
index 742b350df7..90603e215b 100644
--- a/IDE/GCC-ARM/Makefile.common
+++ b/IDE/GCC-ARM/Makefile.common
@@ -89,6 +89,8 @@ SRC_C += ../../src/sniffer.c
 SRC_C += ../../src/ssl.c
 SRC_C += ../../src/tls.c
 SRC_C += ../../src/tls13.c
+SRC_C += ../../src/dtls.c
+SRC_C += ../../src/dtls13.c
 SRC_C += ../../src/wolfio.c
 
 # wolfCrypt Core (FIPS)
@@ -185,6 +187,9 @@ SRC_C += ../../wolfcrypt/src/dsa.c
 SRC_C += ../../wolfcrypt/src/md2.c
 SRC_C += ../../wolfcrypt/src/md4.c
 SRC_C += ../../wolfcrypt/src/ripemd.c
+SRC_C += ../../wolfcrypt/src/wc_mldsa.c
+SRC_C += ../../wolfcrypt/src/wc_mlkem.c
+SRC_C += ../../wolfcrypt/src/wc_mlkem_poly.c
 
 
 FILENAMES_C = $(notdir $(SRC_C))
diff --git a/IDE/GCC-ARM/Source/wolf_main.c b/IDE/GCC-ARM/Source/wolf_main.c
index 57e5d224b7..8b6f05da40 100644
--- a/IDE/GCC-ARM/Source/wolf_main.c
+++ b/IDE/GCC-ARM/Source/wolf_main.c
@@ -27,6 +27,9 @@
 #include <stdio.h>
 #include <stdarg.h>
 #include <string.h>
+#ifdef NO_ASN_TIME
+    #include <time.h>
+#endif
 
 
 /* TIME CODE */
@@ -37,21 +40,103 @@ static int hw_get_time_sec(void)
 {
     #warning Must implement your own time source if validating certificates
 
-	return ++gTimeMs;
+    return ++gTimeMs;
+}
+
+static int IsLeapYear(int year)
+{
+    return ((year % 4) == 0 && ((year % 100) != 0 || (year % 400) == 0));
 }
 
 /* This is used by wolfCrypt asn.c for cert time checking */
-unsigned long my_time(unsigned long* timer)
+time_t my_time(time_t* timer)
 {
-    (void)timer;
-    return hw_get_time_sec();
+    time_t curTime = (time_t)hw_get_time_sec();
+
+    if (timer != NULL) {
+        *timer = curTime;
+    }
+
+    return curTime;
+}
+
+struct tm* my_gmtime(const time_t* timer, struct tm* tmp)
+{
+    static struct tm staticTime;
+    static const unsigned char daysPerMonth[] =
+        { 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 };
+    time_t curTime;
+    long days;
+    long rem;
+    int year;
+    int yearDays;
+    int month;
+    int monthDays;
+
+    if (tmp == NULL) {
+        tmp = &staticTime;
+    }
+
+    curTime = (timer != NULL) ? *timer : my_time(NULL);
+    if (curTime < 0) {
+        curTime = 0;
+    }
+
+    days = (long)(curTime / 86400);
+    rem = (long)(curTime % 86400);
+
+    tmp->tm_hour = (int)(rem / 3600);
+    rem %= 3600;
+    tmp->tm_min = (int)(rem / 60);
+    tmp->tm_sec = (int)(rem % 60);
+    tmp->tm_wday = (int)((days + 4) % 7);
+
+    year = 1970;
+    while (1) {
+        yearDays = IsLeapYear(year) ? 366 : 365;
+        if (days < yearDays) {
+            break;
+        }
+        days -= yearDays;
+        year++;
+    }
+
+    tmp->tm_year = year - 1900;
+    tmp->tm_yday = (int)days;
+
+    for (month = 0; month < 12; month++) {
+        monthDays = daysPerMonth[month];
+        if (month == 1 && IsLeapYear(year)) {
+            monthDays++;
+        }
+        if (days < monthDays) {
+            break;
+        }
+        days -= monthDays;
+    }
+
+    tmp->tm_mon = month;
+    tmp->tm_mday = (int)days + 1;
+    tmp->tm_isdst = 0;
+
+    return tmp;
 }
 
 #ifndef WOLFCRYPT_ONLY
 /* This is used by TLS only */
-unsigned int LowResTimer(void)
+word32 LowResTimer(void)
 {
-    return hw_get_time_sec();
+    return (word32)hw_get_time_sec();
+}
+
+/* This is used by TLS 1.3 ticket and PSK timeouts. */
+#ifdef WOLFSSL_32BIT_MILLI_TIME
+word32 TimeNowInMilliseconds(void)
+#else
+sword64 TimeNowInMilliseconds(void)
+#endif
+{
+    return (sword64)my_time(NULL) * 1000;
 }
 #endif
 
diff --git a/IDE/GCC-ARM/include.am b/IDE/GCC-ARM/include.am
index ca12aca5c2..61d6eeb19a 100644
--- a/IDE/GCC-ARM/include.am
+++ b/IDE/GCC-ARM/include.am
@@ -10,6 +10,7 @@ EXTRA_DIST+= IDE/GCC-ARM/Source/test_main.c
 EXTRA_DIST+= IDE/GCC-ARM/Source/tls_client.c
 EXTRA_DIST+= IDE/GCC-ARM/Source/tls_server.c
 EXTRA_DIST+= IDE/GCC-ARM/linker.ld
+EXTRA_DIST+= IDE/GCC-ARM/linker_large.ld
 EXTRA_DIST+= IDE/GCC-ARM/linker_fips.ld
 EXTRA_DIST+= IDE/GCC-ARM/Makefile
 EXTRA_DIST+= IDE/GCC-ARM/Makefile.bench
diff --git a/IDE/GCC-ARM/linker_large.ld b/IDE/GCC-ARM/linker_large.ld
new file mode 100644
index 0000000000..c4b3e4b43b
--- /dev/null
+++ b/IDE/GCC-ARM/linker_large.ld
@@ -0,0 +1,30 @@
+MEMORY
+{
+  FLASH (wx) : ORIGIN = 0x00000000, LENGTH = 1024K
+  RAM   (wx) : ORIGIN = 0x20000000, LENGTH = 256K
+}
+
+SECTIONS
+{
+    __vectors_start__ = .;
+    .vectors : { *(.vectors) } > FLASH
+    __vectors_end__ = __vectors_start__ + 0x400;
+    .sys    : { *(.sys*) }    > FLASH
+    .text   : { *(.text*) }   > FLASH
+    .rodata : { *(.rodata*) } > FLASH
+
+    __data_load_start__ = .;
+    __data_start__ = .;
+    .data   : { *(.data*) }   > RAM
+    __data_end__ = __data_start__ + SIZEOF(.data);
+
+    __bss_start__ = .;
+    .bss    : { *(.bss*)  }   > RAM
+    __bss_end__   = __bss_start__ + SIZEOF(.bss);
+
+    __heap_start__ = .;
+    .heap   : { *(.heap*)  }   > RAM
+    __heap_end__ = __heap_start__ + SIZEOF(.heap);
+
+    end = .;
+}
diff --git a/IDE/Renesas/e2studio/RA6M4/README.md b/IDE/Renesas/e2studio/RA6M4/README.md
index c576d8c1e2..e74b13b12b 100644
--- a/IDE/Renesas/e2studio/RA6M4/README.md
+++ b/IDE/Renesas/e2studio/RA6M4/README.md
@@ -85,6 +85,8 @@ The wolfssl Project Summary is listed below and is relevant for every project.
 + Add `FreeRTOS + TCP` stack to sce_tst_thread from `New Stack` -> `Networking` -> `FreeRTOS+TCP` and set properties
 + Add Ethernet Driver by clicking `Add Ethernet Driver` element and select `New` -> `Ethernet(r_ether)`
 + Increase Heap size of `RA Common`. Go to `BSP` tab and inclease `RA Common` -> `Heap size (bytes)` to 0x2000
+
+
 |Property|Value|
 |:--|:--|
 |Network Events call vApplicationIPNetworkEventHook|Disable|
@@ -130,7 +132,7 @@ The wolfssl Project Summary is listed below and is relevant for every project.
     SEGGER_RTT_printf.c
 
 + To connect RTT block, you can configure RTT viewer configuration based on where RTT block is in a map file.
-+ To place RTT block specific area, you can add the following line to `fsp.ld`:
++ To place RTT block specific area, you can add the following line to `fsp_gen.ld`:
 
 ```
     __ram_from_flash$$ :
@@ -161,9 +163,9 @@ SEGGER_RTT_CB _SEGGER_RTT __attribute__((section(".txt.rtt_block")));
                0x20000000       0xa8 ./src/SEGGER_RTT/SEGGER_RTT.o
                0x20000000                _SEGGER_RTT
    ````
-    you can specify "RTT control block" to 0x20023648 by Address
+    you can specify "RTT control block" to 0x20000000 by Address
     OR
-    you can specify "RTT control block" to 0x20023000 0x1000 by Search Range
+    you can specify "RTT control block" to 0x20000000 0x1000 by Search Range
 
 ## Run Client
 1.) Enable TLS_CLIENT definition in wolfssl_demo.h of test_RA6M4 project
@@ -174,7 +176,7 @@ SEGGER_RTT_CB _SEGGER_RTT __attribute__((section(".txt.rtt_block")));
 ```
 static const byte ucIPAddress[4]          = { 192, 168, 11, 241 };
 ```
-+ Client IP address can be changed by the following line in wolf_client.c.
++ Server IP address can be changed by the following line in wolfssl_demo.h.
 ```
 #define SERVER_IP    "192.168.11.40"
 ```
diff --git a/IDE/Renesas/e2studio/RA6M4/common/user_settings.h b/IDE/Renesas/e2studio/RA6M4/common/user_settings.h
index eea211f233..ba77b8dc0c 100644
--- a/IDE/Renesas/e2studio/RA6M4/common/user_settings.h
+++ b/IDE/Renesas/e2studio/RA6M4/common/user_settings.h
@@ -115,5 +115,4 @@
 #endif
 
 #define CUSTOM_RAND_GENERATE_BLOCK wc_fspsm_GenerateRandBlock
-/* use original asn parsing */
-#define WOLFSSL_ASN_ORIGINAL
+
diff --git a/IDE/Renesas/e2studio/RA6M4/test/key_data/key_data_sce.c b/IDE/Renesas/e2studio/RA6M4/test/key_data/key_data_sce.c
index 2212cbac91..ef297aa53a 100644
--- a/IDE/Renesas/e2studio/RA6M4/test/key_data/key_data_sce.c
+++ b/IDE/Renesas/e2studio/RA6M4/test/key_data/key_data_sce.c
@@ -72,32 +72,32 @@ const st_user_key_block_data_t g_key_block_data =
  * This is used for Root Certificate verify by SCE */
 const unsigned char ca_cert_der_sign[] =
 {
-    0xAF, 0x76, 0x00, 0x5E, 0x2C, 0x9C, 0xCB, 0xAF, 0x96, 0x9F,
-    0x7A, 0x0E, 0xA1, 0x41, 0x36, 0xBE, 0x72, 0x8C, 0xED, 0x86,
-    0x31, 0xEF, 0x03, 0x19, 0x2D, 0x68, 0xB0, 0x17, 0xD2, 0x3F,
-    0x15, 0xAC, 0x51, 0x00, 0x98, 0xE9, 0xE0, 0x07, 0x39, 0x00,
-    0x83, 0x63, 0xE3, 0x0F, 0xCA, 0xD5, 0x8A, 0xBD, 0xBB, 0xC5,
-    0x5C, 0xC7, 0xE4, 0xE0, 0xCB, 0x2F, 0x9F, 0x4E, 0xE8, 0xB0,
-    0xBF, 0xA6, 0x12, 0x95, 0xFF, 0x5C, 0xA8, 0xF7, 0x89, 0xD4,
-    0xA0, 0x84, 0x74, 0x50, 0x91, 0x11, 0x16, 0xEA, 0xEC, 0x59,
-    0x03, 0x63, 0x2C, 0xBC, 0x7B, 0x16, 0xD5, 0xAB, 0x2F, 0x7A,
-    0xF8, 0xC0, 0x51, 0xC2, 0x58, 0x9C, 0x2C, 0x89, 0xFB, 0x26,
-    0x88, 0x13, 0x3F, 0x77, 0xD8, 0x2F, 0x3F, 0x52, 0x5C, 0x65,
-    0x0F, 0x04, 0xF8, 0xAC, 0x54, 0x4F, 0x9C, 0x36, 0x80, 0x4A,
-    0xE7, 0x54, 0x5A, 0x86, 0x6E, 0xD5, 0x34, 0xC9, 0x31, 0x12,
-    0x9D, 0x64, 0x8F, 0x3D, 0xA3, 0x9D, 0x07, 0x87, 0xF5, 0xD7,
-    0x54, 0x6E, 0xFA, 0x53, 0xF0, 0x09, 0xCE, 0x0A, 0xD3, 0xCA,
-    0x63, 0x23, 0x0A, 0x07, 0xF5, 0xA2, 0x28, 0x67, 0xCD, 0xC7,
-    0x1E, 0x0C, 0x4F, 0x15, 0x6B, 0xBA, 0x50, 0xDE, 0x08, 0x7B,
-    0x56, 0x18, 0x8C, 0xA2, 0xE0, 0x9C, 0x5A, 0x31, 0xBA, 0xAC,
-    0xD3, 0x8B, 0xFC, 0x7F, 0xE1, 0x60, 0x26, 0x93, 0x47, 0xA6,
-    0x42, 0x67, 0x50, 0x0F, 0x0B, 0x90, 0x27, 0xE3, 0x53, 0x35,
-    0xFD, 0x20, 0x2A, 0xEF, 0xB1, 0x68, 0x81, 0xF8, 0x05, 0x19,
-    0xB2, 0x37, 0x57, 0xE8, 0x14, 0x5E, 0xD0, 0xC8, 0x10, 0xB0,
-    0xAA, 0x15, 0x27, 0xA3, 0xBF, 0x09, 0xC1, 0xC4, 0xAC, 0x5B,
-    0xB4, 0xE4, 0x9A, 0xD3, 0x3A, 0x59, 0xF0, 0x2A, 0x05, 0xFE,
-    0xA6, 0xA6, 0x86, 0x96, 0x9C, 0xF7, 0x6C, 0xA3, 0x3B, 0x3A,
-    0x54, 0xA0, 0x38, 0x6D, 0x84, 0x73
+    0x4D, 0x7E, 0xE5, 0x34, 0x04, 0x8F, 0xA6, 0x81, 0x22, 0x8E,
+    0x2A, 0x9E, 0xEA, 0xBE, 0x1D, 0x0A, 0x1C, 0xC3, 0xEC, 0xCD,
+    0x99, 0x1D, 0xE0, 0x2E, 0x79, 0x42, 0x8F, 0xEA, 0x79, 0x47,
+    0x13, 0xC3, 0x40, 0x47, 0x83, 0xF6, 0xFE, 0x32, 0x42, 0x52,
+    0x26, 0xED, 0xC4, 0xCF, 0x2E, 0x86, 0xF4, 0xEC, 0xFE, 0x8F,
+    0x6B, 0xF4, 0x3F, 0x0F, 0xA7, 0x1E, 0x9F, 0x09, 0x37, 0xAE,
+    0x5E, 0x35, 0x39, 0xCD, 0x35, 0xAC, 0x6F, 0x37, 0x98, 0x01,
+    0x6C, 0x8B, 0x01, 0x8A, 0xAE, 0xCC, 0xAF, 0x05, 0x6E, 0x05,
+    0x3A, 0x0A, 0x03, 0xA2, 0xEB, 0xD4, 0x4A, 0x63, 0x64, 0x68,
+    0xB6, 0xE4, 0xFC, 0xDE, 0xA5, 0x6B, 0xC4, 0x3C, 0x8F, 0xCF,
+    0xC2, 0x5A, 0x2F, 0xBB, 0xBE, 0x70, 0xDC, 0x19, 0x07, 0x4D,
+    0xC3, 0x5F, 0x3E, 0xD9, 0x1A, 0x8E, 0xB9, 0x36, 0x74, 0xF6,
+    0x38, 0xA6, 0x62, 0x7A, 0xBD, 0x53, 0xF5, 0x91, 0xCF, 0xFD,
+    0xD2, 0xCA, 0x4A, 0x20, 0x47, 0xD3, 0x78, 0x15, 0x82, 0x75,
+    0xDC, 0x3F, 0x2D, 0xAE, 0x18, 0xFE, 0x9F, 0x22, 0x99, 0x12,
+    0x4D, 0xF3, 0x3E, 0xE3, 0x47, 0x33, 0xED, 0x06, 0x79, 0x42,
+    0x12, 0xBD, 0x07, 0x39, 0x3F, 0x0D, 0x8A, 0x64, 0xA9, 0xDC,
+    0xF0, 0xCF, 0x87, 0x85, 0x4F, 0x09, 0x31, 0x26, 0x6D, 0x05,
+    0x2B, 0xBD, 0x52, 0xFB, 0x03, 0xD9, 0xD8, 0x0E, 0x34, 0x33,
+    0x72, 0xC0, 0x32, 0x69, 0x6F, 0x23, 0x2A, 0xF9, 0xD3, 0xBB,
+    0x55, 0xD4, 0xF8, 0x85, 0xF5, 0xF8, 0x5E, 0xAB, 0x57, 0x35,
+    0xAC, 0x3A, 0x33, 0x1F, 0xB8, 0x29, 0x1E, 0xFA, 0x30, 0xC1,
+    0x32, 0xEB, 0xAB, 0xD3, 0x51, 0x15, 0x6A, 0xAE, 0xE6, 0xD1,
+    0x42, 0x6C, 0x93, 0x3C, 0x48, 0xE1, 0xF8, 0xBC, 0x03, 0xFB,
+    0x38, 0x25, 0x18, 0x8F, 0x5B, 0x60, 0xE2, 0x86, 0x15, 0x88,
+    0xA3, 0x09, 0xFD, 0x29, 0xB7, 0xF5
 };
 
 const int sizeof_ca_cert_der_sign = sizeof(ca_cert_der_sign);
@@ -108,32 +108,32 @@ const int sizeof_ca_cert_der_sign = sizeof(ca_cert_der_sign);
  */
 const unsigned char ca_ecc_cert_der_sign[] =
 {
-    0x5A, 0x12, 0x05, 0x7F, 0x18, 0x16, 0x41, 0x18, 0x37, 0x11,
-    0x67, 0x22, 0xCE, 0x03, 0x92, 0xDB, 0xCB, 0xE2, 0xD2, 0x5B,
-    0x35, 0x3A, 0x2E, 0x35, 0xE8, 0x86, 0x6F, 0xB6, 0x3F, 0x86,
-    0x99, 0xFB, 0xB7, 0x7A, 0xB1, 0x38, 0xAE, 0x79, 0x7D, 0x98,
-    0x2D, 0x81, 0xBD, 0x74, 0xB1, 0x7D, 0xAD, 0xCE, 0x4F, 0x63,
-    0xD2, 0xB3, 0x06, 0xDF, 0x00, 0x91, 0x12, 0x21, 0x53, 0xAE,
-    0xEB, 0x0E, 0x4E, 0xDA, 0xF1, 0xD7, 0x29, 0x2E, 0xC8, 0xF2,
-    0x42, 0xEB, 0x90, 0xDE, 0xE3, 0xF2, 0x4B, 0x9A, 0xC0, 0x16,
-    0x51, 0x4C, 0xE3, 0x54, 0xCD, 0xE7, 0x33, 0xE6, 0x0A, 0xD4,
-    0x0F, 0x51, 0x6D, 0xA5, 0x69, 0x5E, 0xD7, 0xDD, 0x5A, 0xF9,
-    0x95, 0xE7, 0x89, 0x71, 0x56, 0x22, 0x37, 0x24, 0x2D, 0x39,
-    0x21, 0xE4, 0xC9, 0x4F, 0x22, 0x2E, 0x01, 0xA0, 0xA1, 0x2F,
-    0xDC, 0x75, 0xD9, 0xED, 0x25, 0x37, 0xB6, 0x7E, 0xAB, 0xF7,
-    0xFA, 0xCD, 0x68, 0x9B, 0x89, 0x1F, 0x21, 0x43, 0x8C, 0xAC,
-    0xE6, 0x55, 0xC2, 0xDE, 0xF5, 0xCE, 0xA9, 0x1B, 0x6D, 0x7E,
-    0x64, 0x3D, 0x91, 0x5A, 0x8D, 0x55, 0x8D, 0x32, 0x56, 0x0D,
-    0xE2, 0x65, 0x73, 0xBE, 0x21, 0x2E, 0x7D, 0xFF, 0x2C, 0xA4,
-    0x28, 0x0F, 0x37, 0x0F, 0xEA, 0xE7, 0xC4, 0xAC, 0x73, 0x7E,
-    0x5F, 0x35, 0x35, 0xD2, 0xE1, 0x76, 0x5B, 0xB3, 0x44, 0xA2,
-    0xD7, 0x8E, 0x99, 0xE5, 0x7E, 0xD1, 0x61, 0xAB, 0x78, 0xFC,
-    0x70, 0x61, 0x34, 0x99, 0x41, 0x1F, 0x83, 0xFF, 0x71, 0x8D,
-    0x83, 0xC5, 0x03, 0x3A, 0x64, 0x54, 0xD0, 0x68, 0x83, 0xC7,
-    0x3B, 0x75, 0x62, 0x86, 0x7B, 0xF3, 0x4B, 0xFE, 0xAC, 0x4B,
-    0x6F, 0x45, 0x70, 0xC4, 0x71, 0x03, 0x9F, 0x90, 0x32, 0x35,
-    0xA8, 0x44, 0x63, 0x2B, 0x89, 0xF4, 0xCE, 0x38, 0x40, 0x61,
-    0xC4, 0x9A, 0x2E, 0x84, 0xD5, 0x08
+    0x10, 0xDC, 0xA3, 0x0E, 0xD9, 0xA7, 0x3C, 0xBE, 0x37, 0x6E,
+    0x34, 0xC1, 0xDB, 0x4E, 0x90, 0x42, 0x05, 0xCC, 0x94, 0x2B,
+    0xAF, 0x0D, 0xE0, 0xC9, 0xE3, 0x91, 0x5F, 0x57, 0x58, 0x48,
+    0x44, 0x69, 0x96, 0x4A, 0x94, 0x24, 0x17, 0xE6, 0xDC, 0x1F,
+    0x6B, 0xCA, 0xC5, 0xD2, 0x66, 0x25, 0xF7, 0x6F, 0x1E, 0x67,
+    0xC0, 0x16, 0x7D, 0x59, 0x43, 0x55, 0x52, 0x6F, 0x1C, 0x23,
+    0x55, 0xAC, 0x5E, 0x2C, 0x1A, 0x33, 0x9A, 0x3F, 0x8D, 0x2E,
+    0xA9, 0x2B, 0x2A, 0x57, 0x90, 0xA9, 0x4B, 0x5F, 0x30, 0xD1,
+    0xB5, 0x76, 0x10, 0xAB, 0xDA, 0x0C, 0x6A, 0x0A, 0x42, 0xEF,
+    0x5C, 0x4C, 0xFD, 0x4E, 0x5B, 0x4B, 0x68, 0x30, 0x23, 0x11,
+    0x53, 0xA8, 0xBE, 0x01, 0x2E, 0x6F, 0xEB, 0x40, 0x60, 0x1D,
+    0xBC, 0xE7, 0xCC, 0xCC, 0x89, 0xA0, 0xD1, 0x2E, 0x08, 0x0F,
+    0xE5, 0x81, 0xB1, 0xCF, 0x32, 0x95, 0xA0, 0xF6, 0x0E, 0x59,
+    0x88, 0x10, 0x2F, 0xB2, 0x60, 0x59, 0xA4, 0x08, 0xBA, 0x81,
+    0x1A, 0x6C, 0x24, 0xB8, 0xBE, 0x2B, 0x4A, 0x64, 0x3E, 0x14,
+    0x88, 0x46, 0x22, 0xEC, 0xAA, 0x24, 0x6E, 0x1D, 0x5F, 0xA1,
+    0x89, 0xA3, 0xD0, 0x5B, 0x26, 0xC1, 0x63, 0xF4, 0xDC, 0x32,
+    0x7F, 0xED, 0xDB, 0x0F, 0xE9, 0x0D, 0xBB, 0x24, 0xB9, 0x25,
+    0x87, 0x11, 0xE4, 0x90, 0xE3, 0xAE, 0x4B, 0xEE, 0x3C, 0xFB,
+    0x49, 0xDD, 0x5E, 0xB0, 0xF3, 0xC9, 0x13, 0xA9, 0x73, 0x78,
+    0x23, 0xBB, 0x66, 0x29, 0x53, 0xF2, 0xBF, 0xC6, 0x92, 0xDB,
+    0x09, 0x21, 0x18, 0xAF, 0x76, 0x43, 0x4C, 0x7E, 0xD7, 0x4D,
+    0xCA, 0xB7, 0xAA, 0x90, 0xEF, 0x6E, 0xF5, 0x3F, 0x66, 0xCA,
+    0x40, 0x1A, 0xCE, 0x87, 0x20, 0x9F, 0x39, 0x36, 0x03, 0x53,
+    0x8E, 0x2C, 0xFD, 0x31, 0x14, 0xB8, 0x7E, 0x81, 0xC7, 0x84,
+    0x69, 0x1B, 0xBD, 0x1A, 0x0F, 0x3A
 };
 static const int sizeof_ca_ecc_cert_der_sign = sizeof(ca_ecc_cert_der_sign);
 #endif /*  USE_CERT_BUFFERS_256 */
diff --git a/IDE/iotsafe-raspberrypi/client-tls13.c b/IDE/iotsafe-raspberrypi/client-tls13.c
index ec148bc69d..c68a9e520f 100644
--- a/IDE/iotsafe-raspberrypi/client-tls13.c
+++ b/IDE/iotsafe-raspberrypi/client-tls13.c
@@ -170,39 +170,31 @@ int client_loop(const char *peer_ip, const char *peer_name,
     char         randombytes[16] = {0};
 
     /* Construct HTTP POST */
-
-    /* Header */
-    strcat(buff, "POST /iot/device HTTP/1.1\r\n");
-    strcat(buff, "Content-Type: application/json\r\n");
-    strcat(buff, "Content-Length: 1000\r\n");
-    strcat(buff, "Accept: */*\r\n");
-    strcat(buff, "Host: ");
-    strcat(buff, peer_name);
-    strcat(buff, ":");
-    strcat(buff, peer_port);
-    strcat(buff, "\r\n");
-
-    /* Delimiter */
-    strcat(buff, "\r\n");
-
-    /* Body */
     srand(time(NULL));
-    int devid    = rand() % 100;
-    char snum[5] = {0};
-    snprintf(snum, sizeof(snum), "%d", devid);
+    int devid = rand() % 100;
 
-    strcat(buff, "{");
-    strcat(buff, "\"deviceId\": \"");
-    strcat(buff, snum);
-    strcat(buff, "\",");
-    strcat(buff, "\"sensorType\": \"Temperature\",");
-    strcat(buff, "\"sensorValue\": \"");
-    strcat(buff, temperature);
-    strcat(buff, "\",");
-    strcat(buff, "\"sensorUnit\": \"Celsius\",");
-    strcat(buff, "\"sensorTime\": 1582181510");
-    strcat(buff, "}");
-    strcat(buff, "\r\n");
+    int n = snprintf(buff, sizeof(buff),
+        "POST /iot/device HTTP/1.1\r\n"
+        "Content-Type: application/json\r\n"
+        "Content-Length: 1000\r\n"
+        "Accept: */*\r\n"
+        "Host: %s:%s\r\n"
+        "\r\n"
+        "{"
+        "\"deviceId\": \"%d\","
+        "\"sensorType\": \"Temperature\","
+        "\"sensorValue\": \"%s\","
+        "\"sensorUnit\": \"Celsius\","
+        "\"sensorTime\": 1582181510"
+        "}"
+        "\r\n",
+        peer_name, peer_port, devid, temperature);
+
+    if (n < 0 || n >= (int)sizeof(buff)) {
+        fprintf(stderr, "ERROR: HTTP request too large for buffer\n");
+        ret = -1;
+        goto exit;
+    }
 
     printf("\n\nPOST REQUEST\n\n%s\n\n", buff);
 
diff --git a/IDE/iotsafe-raspberrypi/main.c b/IDE/iotsafe-raspberrypi/main.c
index 44042a16f5..873a17c1e5 100644
--- a/IDE/iotsafe-raspberrypi/main.c
+++ b/IDE/iotsafe-raspberrypi/main.c
@@ -203,27 +203,27 @@ int main(int argc, char** argv)
 
     if (argc == 11)
     {
-        if (strcmp(argv[1], "-ip") == 0)
-           strcpy((char*)&ip, argv[2]);
+        if (strcmp(argv[1], "-ip") == 0 && strlen(argv[2]) < sizeof(ip))
+            strcpy((char*)&ip, argv[2]);
         else
             show_usage(argv[0]);
 
-        if (strcmp(argv[3], "-h") == 0)
+        if (strcmp(argv[3], "-h") == 0 && strlen(argv[4]) < sizeof(name))
             strcpy((char*)&name, argv[4]);
         else
             show_usage(argv[0]);
 
-        if (strcmp(argv[5], "-p") == 0)
+        if (strcmp(argv[5], "-p") == 0 && strlen(argv[6]) < sizeof(port))
             strcpy((char*)&port, argv[6]);
         else
             show_usage(argv[0]);
 
-        if (strcmp(argv[7], "-t") == 0)
+        if (strcmp(argv[7], "-t") == 0 && strlen(argv[8]) < sizeof(temperature))
             strcpy((char*)&temperature, argv[8]);
         else
             show_usage(argv[0]);
 
-        if (strcmp(argv[9], "-d") == 0)
+        if (strcmp(argv[9], "-d") == 0 && strlen(argv[10]) < sizeof(device))
             strcpy((char*)&device, argv[10]);
         else
             show_usage(argv[0]);
diff --git a/bsdkm/bsdkm_wc_port.h b/bsdkm/bsdkm_wc_port.h
index fba3562212..01d564b0cc 100644
--- a/bsdkm/bsdkm_wc_port.h
+++ b/bsdkm/bsdkm_wc_port.h
@@ -36,6 +36,7 @@
 #ifndef CHAR_BIT
     #include <sys/limits.h>
 #endif /* !CHAR_BIT*/
+#include <sys/proc.h>
 
 #define NO_THREAD_LS
 #define NO_ATTRIBUTE_CONSTRUCTOR
@@ -80,20 +81,22 @@ extern struct malloc_type M_WOLFSSL[1];
 #if defined(WOLFSSL_BSDKM_MEMORY_DEBUG)
     #define XMALLOC(s, h, t) ({                                              \
         (void)(h); (void)(t);                                                \
-        void * _ptr = malloc(s, M_WOLFSSL, M_WAITOK | M_ZERO);               \
+        int _wait_flag = curthread->td_critnest == 0 ? M_WAITOK : M_NOWAIT;  \
+        void * _ptr = malloc((s), M_WOLFSSL, _wait_flag | M_ZERO);           \
         printf("info: malloc: %p, M_WOLFSSL, %zu\n", _ptr, (size_t) s);      \
         (void *)_ptr;                                                        \
     })
 
     #define XFREE(p, h, t) ({                                                \
         void* _xp; (void)(h); (void)(t); _xp = (p);                          \
-        printf("info: free: %p, M_WOLFSSL\n", p);                            \
+        printf("info: free: %p, M_WOLFSSL\n", _xp);                          \
         if(_xp) free(_xp, M_WOLFSSL);                                        \
     })
 #else
     #define XMALLOC(s, h, t) ({                                              \
         (void)(h); (void)(t);                                                \
-        void * _ptr = malloc(s, M_WOLFSSL, M_WAITOK | M_ZERO);               \
+        int _wait_flag = curthread->td_critnest == 0 ? M_WAITOK : M_NOWAIT;  \
+        void * _ptr = malloc((s), M_WOLFSSL, _wait_flag | M_ZERO);           \
         (void *)_ptr;                                                        \
     })
 
diff --git a/bsdkm/wolfkmod.c b/bsdkm/wolfkmod.c
index 84e8a01490..872d68589f 100644
--- a/bsdkm/wolfkmod.c
+++ b/bsdkm/wolfkmod.c
@@ -249,15 +249,14 @@ static int wolfkmod_load(void)
 
     error = wolfkmod_init();
     if (error != 0) {
-        return (ECANCELED);
+        goto wolfkmod_load_out;
     }
 
     #ifndef NO_CRYPT_TEST
     error = wolfcrypt_test(NULL);
     if (error != 0) {
         printf("error: wolfcrypt test failed: %d\n", error);
-        (void)wolfkmod_cleanup();
-        return (ECANCELED);
+        goto wolfkmod_load_out;
     }
     printf("info: wolfCrypt self-test passed.\n");
     #endif /* NO_CRYPT_TEST */
@@ -266,15 +265,19 @@ static int wolfkmod_load(void)
     error = benchmark_test(NULL);
     if (error != 0) {
         printf("error: wolfcrypt benchmark failed: %d\n", error);
-        (void)wolfkmod_cleanup();
-        return (ECANCELED);
+        goto wolfkmod_load_out;
     }
     printf("info: wolfCrypt benchmark passed.\n");
     #endif /* WOLFSSL_KERNEL_BENCHMARKS */
-
     printf("info: libwolfssl loaded\n");
 
-    return (0);
+wolfkmod_load_out:
+    if (error != 0) {
+        (void)wolfkmod_cleanup();
+        error = ECANCELED;
+    }
+
+    return (error);
 }
 
 static int wolfkmod_unload(void)
@@ -435,7 +438,8 @@ static int wolfkdriv_attach(device_t dev)
 
     ret = wolfkmod_init();
     if (ret != 0) {
-        return (ECANCELED);
+        error = ECANCELED;
+        goto attach_out;
     }
 
     /**
@@ -452,7 +456,8 @@ static int wolfkdriv_attach(device_t dev)
     if (softc->crid < 0) {
         device_printf(dev, "error: crypto_get_driverid failed: %d\n",
                softc->crid);
-        return (ENXIO);
+        error = ENXIO;
+        goto attach_out;
     }
 
     /*
@@ -487,7 +492,7 @@ static int wolfkdriv_attach(device_t dev)
 attach_out:
     if (error) {
         wolfkdriv_unregister(softc);
-        error = ENXIO;
+        (void)wolfkmod_cleanup();
     }
 
     return (error);
diff --git a/bsdkm/x86_vecreg.c b/bsdkm/x86_vecreg.c
index 973f17c13e..c96cc0afd1 100644
--- a/bsdkm/x86_vecreg.c
+++ b/bsdkm/x86_vecreg.c
@@ -29,13 +29,13 @@
 #include <machine/pcb.h>
 
 struct wolfkmod_fpu_state_t {
-    volatile lwpid_t td_tid;
-    volatile u_int   nest;
+    volatile lwpid_t td_tid; /* the thread currently using fpu. */
+    volatile u_int   nest;   /* the fpu nesting level. */
 };
 
 typedef struct wolfkmod_fpu_state_t wolfkmod_fpu_state_t;
 
-/* fpu_states array tracks thread id and nesting level of save/restore
+/* The fpu_states array tracks thread id and nesting level of save/restore
  * and push/pop vector registers macro calls. It is indexed by raw cpu id,
  * and only accessed after the thread calls fpu_kern_enter(), and before
  * calling fpu_kern_leave(), and only indexed by the thread's PCPU_GET(cpuid).
diff --git a/configure.ac b/configure.ac
index 7cd5cf6671..0c427d4ca7 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1432,6 +1432,11 @@ then
         AC_MSG_ERROR([--enable-all-asm is incompatible with --disable-ppc32-asm])
     fi
 
+    if test "$enable_ppc64_asm" = "no"
+    then
+        AC_MSG_ERROR([--enable-all-asm is incompatible with --disable-ppc64-asm])
+    fi
+
     case "$host_cpu" in
         *x86_64*|*amd64*)
             if test "$enable_intelasm" = ""
@@ -1462,6 +1467,10 @@ then
             fi
             ;;
         *powerpc64*)
+            if test "$enable_ppc64_asm" = ""
+            then
+               enable_ppc64_asm=yes
+            fi
             ;;
         *powerpc*)
             if test "$enable_ppc32_asm" = ""
@@ -1513,6 +1522,7 @@ then
     test "$enable_hkdf" = "" && enable_hkdf=yes
     test "$enable_eccencrypt" = "" && test "$enable_ecc" != "no" && enable_eccencrypt=yes
     test "$enable_fpecc" = "" && test "$enable_ecc" != "no" && enable_fpecc=yes
+    test "$enable_eccsi" = "" && test "$enable_ecc" != "no" && enable_eccsi=yes
     test "$enable_psk" = "" && enable_psk=yes
     test "$enable_cmac" = "" && enable_cmac=yes
     test "$enable_cmac_kdf" = "" && enable_cmac_kdf=yes
@@ -1580,13 +1590,12 @@ then
         test "$enable_aessiv" = "" && enable_aessiv=yes
         # AFALG lacks AES-EAX
         test "$enable_aeseax" = "" && test "$enable_afalg" != "yes" && enable_aeseax=yes
+        test "$enable_sakke" = "" && test "$enable_ecc" != "no" && enable_sakke=yes
 
         if test "$KERNEL_MODE_DEFAULTS" != "yes"
         then
             test "$enable_cryptocb" = "" && enable_cryptocb=yes
             test "$enable_pkcallbacks" = "" && enable_pkcallbacks=yes
-            test "$enable_eccsi" = "" && test "$enable_ecc" != "no" && enable_eccsi=yes
-            test "$enable_sakke" = "" && test "$enable_ecc" != "no" && enable_sakke=yes
         fi
     fi
 
@@ -3859,7 +3868,7 @@ do
     # FSL, FSR, FSRI, CMOV, CMIX - QEMU doesn't know about these instructions
     AM_CFLAGS="$AM_CFLAGS -DWOLFSSL_RISCV_BIT_MANIPULATION_TERNARY"
     ;;
-  zkn|zkned)
+  zkned)
     # AES encrypt/decrpyt, SHA-2
     ENABLED_RISCV_ASM=yes
     AM_CFLAGS="$AM_CFLAGS -DWOLFSSL_RISCV_SCALAR_CRYPTO_ASM"
@@ -3934,21 +3943,70 @@ then
     AM_CFLAGS="$AM_CFLAGS -DWOLFSSL_PPC32_ASM"
     AC_MSG_NOTICE([32-bit PowerPC assembly for SHA-256])
     ENABLED_PPC32_ASM=yes
+
+    if test "$ENABLED_PPC32_ASM_INLINE" = "yes" || test "$ENABLED_PPC32_ASM_INLINE_REG" = "yes"; then
+        AM_CFLAGS="$AM_CFLAGS -DWOLFSSL_PPC32_ASM_INLINE"
+    else
+        AM_CCASFLAGS="$AM_CCASFLAGS -DWOLFSSL_PPC32_ASM"
+    fi
+    if test "$ENABLED_PPC32_ASM_SMALL" = "yes"; then
+        AM_CFLAGS="$AM_CFLAGS -DWOLFSSL_PPC32_ASM_SMALL"
+        AM_CCASFLAGS="$AM_CCASFLAGS -DWOLFSSL_PPC32_ASM_SMALL"
+    fi
+    if test "$ENABLED_PPC32_ASM_SPE" = "yes"; then
+        AM_CFLAGS="$AM_CFLAGS -DWOLFSSL_PPC32_ASM_SPE"
+        AM_CCASFLAGS="$AM_CCASFLAGS -DWOLFSSL_PPC32_ASM_SPE"
+    fi
 fi
-if test "$ENABLED_PPC32_ASM_INLINE" = "yes" || test "$ENABLED_PPC32_ASM_INLINE_REG" = "yes"; then
-    AM_CFLAGS="$AM_CFLAGS -DWOLFSSL_PPC32_ASM_INLINE"
+
+# PPC64 Assembly
+AC_ARG_ENABLE([ppc64-asm],
+    [AS_HELP_STRING([--enable-ppc64-asm],[Enable wolfSSL PowerPC 64-bit ASM support (default: disabled).])],
+    [ ENABLED_PPC64_ASM=$enableval ],
+    [ ENABLED_PPC64_ASM=no ]
+    )
+
+
+if test "$ENABLED_PPC64_ASM" != "no" && test "$ENABLED_ASM" = "yes"
+then
+    ENABLED_PPC64_ASM_OPTS=$ENABLED_PPC64_ASM
+    for v in `echo $ENABLED_PPC64_ASM_OPTS | tr "," " "`
+    do
+      case $v in
+      yes)
+        ;;
+      inline)
+        ENABLED_PPC64_ASM_INLINE=yes
+        ;;
+      inline-reg)
+        ENABLED_PPC64_ASM_INLINE_REG=yes
+        ;;
+      small)
+        ENABLED_PPC64_ASM_SMALL=yes
+        ;;
+      *)
+        AC_MSG_ERROR([Invalid RISC-V option [yes,inline,small]: $ENABLED_PPC64_ASM.])
+        break
+        ;;
+      esac
+    done
+
+    AM_CFLAGS="$AM_CFLAGS -DWOLFSSL_PPC64_ASM"
+    AM_CCASFLAGS="$AM_CCASFLAGS -DEXTERNAL_OPTS_OPENVPN"
+    AC_MSG_NOTICE([64-bit PowerPC assembly for AES])
+    ENABLED_PPC64_ASM=yes
+fi
+if test "$ENABLED_PPC64_ASM_INLINE" = "yes" || test "$ENABLED_PPC64_ASM_INLINE_REG" = "yes"; then
+    AM_CFLAGS="$AM_CFLAGS -DWOLFSSL_PPC64_ASM_INLINE"
 else
-    AM_CCASFLAGS="$AM_CCASFLAGS -DWOLFSSL_PPC32_ASM"
+    AM_CCASFLAGS="$AM_CCASFLAGS -DWOLFSSL_PPC64_ASM"
 fi
-if test "$ENABLED_PPC32_ASM_SMALL" = "yes"; then
-    AM_CFLAGS="$AM_CFLAGS -DWOLFSSL_PPC32_ASM_SMALL"
-    AM_CCASFLAGS="$AM_CCASFLAGS -DWOLFSSL_PPC32_ASM_SMALL"
-fi
-if test "$ENABLED_PPC32_ASM_SPE" = "yes"; then
-    AM_CFLAGS="$AM_CFLAGS -DWOLFSSL_PPC32_ASM_SPE"
-    AM_CCASFLAGS="$AM_CCASFLAGS -DWOLFSSL_PPC32_ASM_SPE"
+if test "$ENABLED_PPC64_ASM_SMALL" = "yes"; then
+    AM_CFLAGS="$AM_CFLAGS -DWOLFSSL_PPC64_ASM_SMALL"
+    AM_CCASFLAGS="$AM_CCASFLAGS -DWOLFSSL_PPC64_ASM_SMALL"
 fi
 
+
 # Xilinx hardened crypto
 AC_ARG_ENABLE([xilinx],
     [AS_HELP_STRING([--enable-xilinx],[Enable wolfSSL support for Xilinx hardened crypto(default: disabled)])],
@@ -12219,6 +12277,9 @@ AM_CONDITIONAL([BUILD_RISCV_ASM],[test "x$ENABLED_RISCV_ASM" = "xyes"])
 AM_CONDITIONAL([BUILD_PPC32_ASM],[test "x$ENABLED_PPC32_ASM" = "xyes"])
 AM_CONDITIONAL([BUILD_PPC32_ASM_INLINE],[test "x$ENABLED_PPC32_ASM_INLINE" = "xyes"])
 AM_CONDITIONAL([BUILD_PPC32_ASM_INLINE_REG],[test "x$ENABLED_PPC32_ASM_INLINE_REG" = "xyes"])
+AM_CONDITIONAL([BUILD_PPC64_ASM],[test "x$ENABLED_PPC64_ASM" = "xyes"])
+AM_CONDITIONAL([BUILD_PPC64_ASM_INLINE],[test "x$ENABLED_PPC64_ASM_INLINE" = "xyes"])
+AM_CONDITIONAL([BUILD_PPC64_ASM_INLINE_REG],[test "x$ENABLED_PPC64_ASM_INLINE_REG" = "xyes"])
 AM_CONDITIONAL([BUILD_XILINX],[test "x$ENABLED_XILINX" = "xyes"])
 AM_CONDITIONAL([BUILD_AESNI],[test "x$ENABLED_AESNI" = "xyes"])
 AM_CONDITIONAL([BUILD_INTELASM],[test "x$ENABLED_INTELASM" = "xyes"])
@@ -12925,6 +12986,15 @@ then
     ENABLED_PPC32_ASM="inline C Reg"
 fi
 echo "   * PPC32 ASM                   $ENABLED_PPC32_ASM"
+if test "$ENABLED_PPC64_ASM_INLINE" = "yes"
+then
+    ENABLED_PPC64_ASM="inline C"
+fi
+if test "$ENABLED_PPC64_ASM_INLINE_REG" = "yes"
+then
+    ENABLED_PPC64_ASM="inline C Reg"
+fi
+echo "   * PPC64 ASM                   $ENABLED_PPC64_ASM"
 echo "   * Write duplicate:            $ENABLED_WRITEDUP"
 echo "   * Xilinx Hardware Acc.:       $ENABLED_XILINX"
 echo "   * C89:                        $ENABLED_C89"
diff --git a/examples/configs/user_settings_openssl_compat.h b/examples/configs/user_settings_openssl_compat.h
index b2557bb293..297e50a2c3 100644
--- a/examples/configs/user_settings_openssl_compat.h
+++ b/examples/configs/user_settings_openssl_compat.h
@@ -51,6 +51,23 @@ extern "C" {
 #define HAVE_WOLFSSL_SSL_H
 #define OPENSSL_COMPATIBLE_DEFAULTS
 
+/* No-socket embedded builds still compile IP SAN helpers through OPENSSL_ALL.
+ * Provide local address-conversion stubs when the platform has no socket API. */
+#ifdef WOLFSSL_NO_SOCK
+    #ifndef WOLFSSL_IP4
+        #define WOLFSSL_IP4 2
+    #endif
+    #ifndef WOLFSSL_IP6
+        #define WOLFSSL_IP6 10
+    #endif
+    #ifndef XINET_PTON
+        #define XINET_PTON(a,b,c) 0
+    #endif
+    #ifndef XINET_NTOP
+        #define XINET_NTOP(a,b,c,d) NULL
+    #endif
+#endif
+
 /* Avoid old name conflicts */
 #define NO_OLD_RNGNAME
 #define NO_OLD_WC_NAMES
diff --git a/linuxkm/README.md b/linuxkm/README.md
index dd65cec87d..43652e3c1c 100644
--- a/linuxkm/README.md
+++ b/linuxkm/README.md
@@ -57,6 +57,24 @@ $ sudo modprobe libwolfssl
 | `--enable-intelasm`                | x86/amd64 crypto acceleration            |
 | `--enable-cryptonly`               | Omit TLS/DTLS implementation (normally recommended) |
 
+### Enabling DTLS 1.3 in the kernel module
+
+`--enable-linuxkm` does not implicitly enable TLS 1.3 or DTLS, so the DTLS 1.3
+configure check (`configure.ac:5634-5636`) requires all three flags to be
+passed explicitly:
+
+```sh
+./configure --enable-linuxkm \
+            --enable-tls13 --enable-dtls --enable-dtls13 \
+            --with-linux-source=/lib/modules/$(uname -r)/build
+make -j$(nproc) module
+```
+
+The resulting `linuxkm/libwolfssl.ko` exports the DTLS 1.3 entry points
+(`wolfDTLSv1_3_client_method`, `wolfDTLSv1_3_server_method`, etc.) as GPL
+kernel symbols, available to other in-kernel consumers via
+`EXPORT_SYMBOL_GPL`.
+
 ### Additional configuration options for verification, performance evaluation, and troubleshooting
 
 | option                             | description                              |
diff --git a/linuxkm/linuxkm_wc_port.h b/linuxkm/linuxkm_wc_port.h
index 892f323a8a..83e35dc692 100644
--- a/linuxkm/linuxkm_wc_port.h
+++ b/linuxkm/linuxkm_wc_port.h
@@ -167,13 +167,14 @@
     #ifndef WC_LINUXKM_INTR_SIGNALS
         #define WC_LINUXKM_INTR_SIGNALS { SIGKILL, SIGABRT, SIGHUP, SIGINT }
     #endif
-    extern int wc_linuxkm_sig_ignore_begin(void);
-    extern int wc_linuxkm_sig_ignore_end(void);
-    extern int wc_linuxkm_check_for_intr_signals(void);
+    WOLFSSL_API int wc_linuxkm_can_block(void);
+    WOLFSSL_API int wc_linuxkm_sig_ignore_begin(void);
+    WOLFSSL_API int wc_linuxkm_sig_ignore_end(void);
+    WOLFSSL_API int wc_linuxkm_check_for_intr_signals(void);
     #ifndef WC_LINUXKM_MAX_NS_WITHOUT_YIELD
         #define WC_LINUXKM_MAX_NS_WITHOUT_YIELD 1000000000
     #endif
-    extern void wc_linuxkm_relax_long_loop(void);
+    WOLFSSL_API void wc_linuxkm_relax_long_loop(void);
 
     #ifndef WC_SIG_IGNORE_BEGIN
         #define WC_SIG_IGNORE_BEGIN() wc_linuxkm_sig_ignore_begin()
@@ -1221,6 +1222,7 @@
         typeof(wc_lkm_LockMutex) *wc_lkm_LockMutex;
         #endif
 
+        typeof(wc_linuxkm_can_block) *wc_linuxkm_can_block;
         typeof(wc_linuxkm_sig_ignore_begin) *wc_linuxkm_sig_ignore_begin;
         typeof(wc_linuxkm_sig_ignore_end) *wc_linuxkm_sig_ignore_end;
         typeof(wc_linuxkm_check_for_intr_signals) *wc_linuxkm_check_for_intr_signals;
@@ -1521,8 +1523,9 @@
         wolfssl_spin_unlock_irqrestore_rt((lock), (flags))
     #endif
 
-    #define wc_linuxkm_sig_ignore_begin WC_PIE_INDIRECT_SYM(wc_linuxkm_sig_ignore_begin);
-    #define wc_linuxkm_sig_ignore_end WC_PIE_INDIRECT_SYM(wc_linuxkm_sig_ignore_end);
+    #define wc_linuxkm_can_block WC_PIE_INDIRECT_SYM(wc_linuxkm_can_block)
+    #define wc_linuxkm_sig_ignore_begin WC_PIE_INDIRECT_SYM(wc_linuxkm_sig_ignore_begin)
+    #define wc_linuxkm_sig_ignore_end WC_PIE_INDIRECT_SYM(wc_linuxkm_sig_ignore_end)
     #define wc_linuxkm_check_for_intr_signals WC_PIE_INDIRECT_SYM(wc_linuxkm_check_for_intr_signals)
     #define wc_linuxkm_relax_long_loop WC_PIE_INDIRECT_SYM(wc_linuxkm_relax_long_loop)
 
@@ -1905,21 +1908,21 @@
         #endif
     #else /* !WC_LINUXKM_USE_HEAP_WRAPPERS */
     #ifdef USE_KVMALLOC
-        #define malloc(size) kvmalloc_node(WC_LINUXKM_ROUND_UP_P_OF_2(size), (preempt_count() == 0 ? GFP_KERNEL : GFP_ATOMIC), NUMA_NO_NODE)
+        #define malloc(size) kvmalloc_node(WC_LINUXKM_ROUND_UP_P_OF_2(size), (wc_linuxkm_can_block() ? GFP_KERNEL : GFP_ATOMIC), NUMA_NO_NODE)
         #if LINUX_VERSION_CODE >= KERNEL_VERSION(7, 2, 0)
-            #define free(ptr) (preempt_count() == 0 ? kvfree(ptr) : kvfree_atomic(ptr))
+            #define free(ptr) (wc_linuxkm_can_block() ? kvfree(ptr) : kvfree_atomic(ptr))
         #else
             #define free(ptr) kvfree(ptr)
         #endif
         #ifdef USE_KVREALLOC
-            #define realloc(ptr, newsize) kvrealloc(ptr, WC_LINUXKM_ROUND_UP_P_OF_2(newsize), (preempt_count() == 0 ? GFP_KERNEL : GFP_ATOMIC))
+            #define realloc(ptr, newsize) kvrealloc(ptr, WC_LINUXKM_ROUND_UP_P_OF_2(newsize), (wc_linuxkm_can_block() ? GFP_KERNEL : GFP_ATOMIC))
         #else
             #define realloc(ptr, newsize) ((void)(ptr), (void)(newsize), NULL)
         #endif
     #else
-        #define malloc(size) kmalloc(WC_LINUXKM_ROUND_UP_P_OF_2(size), (preempt_count() == 0 ? GFP_KERNEL : GFP_ATOMIC))
+        #define malloc(size) kmalloc(WC_LINUXKM_ROUND_UP_P_OF_2(size), (wc_linuxkm_can_block() ? GFP_KERNEL : GFP_ATOMIC))
         #define free(ptr) kfree(ptr)
-        #define realloc(ptr, newsize) krealloc(ptr, WC_LINUXKM_ROUND_UP_P_OF_2(newsize), (preempt_count() == 0 ? GFP_KERNEL : GFP_ATOMIC))
+        #define realloc(ptr, newsize) krealloc(ptr, WC_LINUXKM_ROUND_UP_P_OF_2(newsize), (wc_linuxkm_can_block() ? GFP_KERNEL : GFP_ATOMIC))
     #endif
     #endif /* !WC_LINUXKM_USE_HEAP_WRAPPERS */
 
diff --git a/linuxkm/lkcapi_dh_glue.c b/linuxkm/lkcapi_dh_glue.c
index 6caabf06ac..581776b230 100644
--- a/linuxkm/lkcapi_dh_glue.c
+++ b/linuxkm/lkcapi_dh_glue.c
@@ -89,37 +89,40 @@
 #include <crypto/dh.h>
 
 #define WOLFKM_DH_NAME    ("dh")
-#define WOLFKM_DH_DRIVER  ("dh" WOLFKM_DRIVER_FIPS \
-                           "-wolfcrypt")
+
+#if defined(WOLFSSL_SP_X86_64_ASM) && !defined(NO_AVX2_SUPPORT)
+    #define WOLFKM_DH_DRIVER_ISA_EXT "-avx2"
+#else
+    #define WOLFKM_DH_DRIVER_ISA_EXT ""
+#endif
+#define WOLFKM_DH_DRIVER_SUFFIX WOLFKM_DH_DRIVER_ISA_EXT \
+                           WOLFKM_DRIVER_SUFFIX_BASE
+
+#define WOLFKM_DH_DRIVER  ("dh" WOLFKM_DH_DRIVER_SUFFIX)
 
 #ifdef HAVE_FFDHE_2048
     #define WOLFKM_FFDHE2048_NAME   ("ffdhe2048(dh)")
-    #define WOLFKM_FFDHE2048_DRIVER ("ffdhe2048" WOLFKM_DRIVER_FIPS \
-                                     "-wolfcrypt")
+    #define WOLFKM_FFDHE2048_DRIVER ("ffdhe2048" WOLFKM_DH_DRIVER_SUFFIX)
 #endif /* HAVE_FFDHE_2048 */
 
 #ifdef HAVE_FFDHE_3072
     #define WOLFKM_FFDHE3072_NAME   ("ffdhe3072(dh)")
-    #define WOLFKM_FFDHE3072_DRIVER ("ffdhe3072" WOLFKM_DRIVER_FIPS \
-                                     "-wolfcrypt")
+    #define WOLFKM_FFDHE3072_DRIVER ("ffdhe3072" WOLFKM_DH_DRIVER_SUFFIX)
 #endif /* HAVE_FFDHE_3072 */
 
 #ifdef HAVE_FFDHE_4096
     #define WOLFKM_FFDHE4096_NAME   ("ffdhe4096(dh)")
-    #define WOLFKM_FFDHE4096_DRIVER ("ffdhe4096" WOLFKM_DRIVER_FIPS \
-                                     "-wolfcrypt")
+    #define WOLFKM_FFDHE4096_DRIVER ("ffdhe4096" WOLFKM_DH_DRIVER_SUFFIX)
 #endif /* HAVE_FFDHE_4096 */
 
 #ifdef HAVE_FFDHE_6144
     #define WOLFKM_FFDHE6144_NAME   ("ffdhe6144(dh)")
-    #define WOLFKM_FFDHE6144_DRIVER ("ffdhe6144" WOLFKM_DRIVER_FIPS \
-                                     "-wolfcrypt")
+    #define WOLFKM_FFDHE6144_DRIVER ("ffdhe6144" WOLFKM_DH_DRIVER_SUFFIX)
 #endif /* HAVE_FFDHE_6144 */
 
 #ifdef HAVE_FFDHE_8192
     #define WOLFKM_FFDHE8192_NAME   ("ffdhe8192(dh)")
-    #define WOLFKM_FFDHE8192_DRIVER ("ffdhe8192" WOLFKM_DRIVER_FIPS \
-                                     "-wolfcrypt")
+    #define WOLFKM_FFDHE8192_DRIVER ("ffdhe8192" WOLFKM_DH_DRIVER_SUFFIX)
 #endif /* HAVE_FFDHE_8192 */
 
 static int linuxkm_test_kpp_driver(const char * driver,
diff --git a/linuxkm/lkcapi_ecdh_glue.c b/linuxkm/lkcapi_ecdh_glue.c
index 2f88bd2037..24755917a5 100644
--- a/linuxkm/lkcapi_ecdh_glue.c
+++ b/linuxkm/lkcapi_ecdh_glue.c
@@ -63,19 +63,24 @@
 #include <wolfssl/wolfcrypt/ecc.h>
 #include <crypto/ecdh.h>
 
-#define WOLFKM_ECDH_DRIVER       ("ecdh-wolfcrypt")
+#if defined(WOLFSSL_SP_X86_64_ASM) && !defined(NO_AVX2_SUPPORT)
+    #define WOLFKM_ECDH_DRIVER_ISA_EXT "-avx2"
+#else
+    #define WOLFKM_ECDH_DRIVER_ISA_EXT ""
+#endif
+#define WOLFKM_ECDH_DRIVER_SUFFIX WOLFKM_ECDH_DRIVER_ISA_EXT \
+                           WOLFKM_DRIVER_SUFFIX_BASE
+
+#define WOLFKM_ECDH_DRIVER       ("ecdh" WOLFKM_ECDH_DRIVER_SUFFIX)
 
 #define WOLFKM_ECDH_P192_NAME    ("ecdh-nist-p192")
-#define WOLFKM_ECDH_P192_DRIVER  ("ecdh-nist-p192" WOLFKM_DRIVER_FIPS \
-                                   "-wolfcrypt")
+#define WOLFKM_ECDH_P192_DRIVER  ("ecdh-nist-p192" WOLFKM_ECDH_DRIVER_SUFFIX)
 
 #define WOLFKM_ECDH_P256_NAME    ("ecdh-nist-p256")
-#define WOLFKM_ECDH_P256_DRIVER  ("ecdh-nist-p256" WOLFKM_DRIVER_FIPS \
-                                   "-wolfcrypt")
+#define WOLFKM_ECDH_P256_DRIVER  ("ecdh-nist-p256" WOLFKM_ECDH_DRIVER_SUFFIX)
 
 #define WOLFKM_ECDH_P384_NAME    ("ecdh-nist-p384")
-#define WOLFKM_ECDH_P384_DRIVER  ("ecdh-nist-p384" WOLFKM_DRIVER_FIPS \
-                                   "-wolfcrypt")
+#define WOLFKM_ECDH_P384_DRIVER  ("ecdh-nist-p384" WOLFKM_ECDH_DRIVER_SUFFIX)
 
 static int linuxkm_test_ecdh_nist_driver(const char * driver,
                                          const byte * b_pub,
diff --git a/linuxkm/lkcapi_ecdsa_glue.c b/linuxkm/lkcapi_ecdsa_glue.c
index 46469131e5..13a0e90be5 100644
--- a/linuxkm/lkcapi_ecdsa_glue.c
+++ b/linuxkm/lkcapi_ecdsa_glue.c
@@ -82,23 +82,27 @@
 #include <wolfssl/wolfcrypt/asn.h>
 #include <wolfssl/wolfcrypt/ecc.h>
 
-#define WOLFKM_ECDSA_DRIVER       ("ecdsa-wolfcrypt")
+#if defined(WOLFSSL_SP_X86_64_ASM) && !defined(NO_AVX2_SUPPORT)
+    #define WOLFKM_ECDSA_DRIVER_ISA_EXT "-avx2"
+#else
+    #define WOLFKM_ECDSA_DRIVER_ISA_EXT ""
+#endif
+#define WOLFKM_ECDSA_DRIVER_SUFFIX WOLFKM_ECDSA_DRIVER_ISA_EXT \
+                           WOLFKM_DRIVER_SUFFIX_BASE
+
+#define WOLFKM_ECDSA_DRIVER       ("ecdsa" WOLFKM_ECDSA_DRIVER_SUFFIX)
 
 #define WOLFKM_ECDSA_P192_NAME    ("ecdsa-nist-p192")
-#define WOLFKM_ECDSA_P192_DRIVER  ("ecdsa-nist-p192" WOLFKM_DRIVER_FIPS \
-                                   "-wolfcrypt")
+#define WOLFKM_ECDSA_P192_DRIVER  ("ecdsa-nist-p192" WOLFKM_ECDSA_DRIVER_SUFFIX)
 
 #define WOLFKM_ECDSA_P256_NAME    ("ecdsa-nist-p256")
-#define WOLFKM_ECDSA_P256_DRIVER  ("ecdsa-nist-p256" WOLFKM_DRIVER_FIPS \
-                                   "-wolfcrypt")
+#define WOLFKM_ECDSA_P256_DRIVER  ("ecdsa-nist-p256" WOLFKM_ECDSA_DRIVER_SUFFIX)
 
 #define WOLFKM_ECDSA_P384_NAME    ("ecdsa-nist-p384")
-#define WOLFKM_ECDSA_P384_DRIVER  ("ecdsa-nist-p384" WOLFKM_DRIVER_FIPS \
-                                   "-wolfcrypt")
+#define WOLFKM_ECDSA_P384_DRIVER  ("ecdsa-nist-p384" WOLFKM_ECDSA_DRIVER_SUFFIX)
 
 #define WOLFKM_ECDSA_P521_NAME    ("ecdsa-nist-p521")
-#define WOLFKM_ECDSA_P521_DRIVER  ("ecdsa-nist-p521" WOLFKM_DRIVER_FIPS \
-                                   "-wolfcrypt")
+#define WOLFKM_ECDSA_P521_DRIVER  ("ecdsa-nist-p521" WOLFKM_ECDSA_DRIVER_SUFFIX)
 
 
 static int  linuxkm_test_ecdsa_nist_driver(const char * driver,
diff --git a/linuxkm/lkcapi_rsa_glue.c b/linuxkm/lkcapi_rsa_glue.c
index e38dfaf28c..ac7f8ff5c7 100644
--- a/linuxkm/lkcapi_rsa_glue.c
+++ b/linuxkm/lkcapi_rsa_glue.c
@@ -99,45 +99,46 @@
 #include <wolfssl/wolfcrypt/rsa.h>
 
 #define WOLFKM_RSA_NAME      ("rsa")
-#define WOLFKM_RSA_DRIVER    ("rsa" WOLFKM_DRIVER_FIPS "-wolfcrypt")
+
+#if defined(WOLFSSL_SP_X86_64_ASM) && !defined(NO_AVX2_SUPPORT)
+    #define WOLFKM_RSA_DRIVER_ISA_EXT "-avx2"
+#else
+    #define WOLFKM_RSA_DRIVER_ISA_EXT ""
+#endif
+#define WOLFKM_RSA_DRIVER_SUFFIX WOLFKM_RSA_DRIVER_ISA_EXT \
+                           WOLFKM_DRIVER_SUFFIX_BASE
+
+#define WOLFKM_RSA_DRIVER    ("rsa" WOLFKM_RSA_DRIVER_SUFFIX)
 
 #if defined(LINUXKM_AKCIPHER_NO_SIGNVERIFY)
     /* the akcipher alg */
-    #define WOLFKM_PKCS1PAD_NAME   ("pkcs1pad(rsa)")
-    #define WOLFKM_PKCS1PAD_DRIVER ("pkcs1pad(rsa" WOLFKM_DRIVER_FIPS \
-                                        "-wolfcrypt)")
+    #define WOLFKM_PKCS1PAD_NAME   "pkcs1pad(rsa)"
+    #define WOLFKM_PKCS1PAD_DRIVER "pkcs1pad-rsa" WOLFKM_RSA_DRIVER_SUFFIX
 #endif /* LINUXKM_AKCIPHER_NO_SIGNVERIFY */
 
 /*
  * pkcs1 sign verify alg names
  * */
 #define WOLFKM_PKCS1_SHA224_NAME   (PKCS1_NAME "(rsa,sha224)")
-#define WOLFKM_PKCS1_SHA224_DRIVER (PKCS1_NAME "(rsa" WOLFKM_DRIVER_FIPS \
-                                    "-wolfcrypt,sha224)")
+#define WOLFKM_PKCS1_SHA224_DRIVER ("pkcs1pad-rsa-sha224" WOLFKM_RSA_DRIVER_SUFFIX)
 
 #define WOLFKM_PKCS1_SHA256_NAME   (PKCS1_NAME "(rsa,sha256)")
-#define WOLFKM_PKCS1_SHA256_DRIVER (PKCS1_NAME "(rsa" WOLFKM_DRIVER_FIPS \
-                                    "-wolfcrypt,sha256)")
+#define WOLFKM_PKCS1_SHA256_DRIVER ("pkcs1pad-rsa-sha256" WOLFKM_RSA_DRIVER_SUFFIX)
 
 #define WOLFKM_PKCS1_SHA384_NAME   (PKCS1_NAME "(rsa,sha384)")
-#define WOLFKM_PKCS1_SHA384_DRIVER (PKCS1_NAME "(rsa" WOLFKM_DRIVER_FIPS \
-                                    "-wolfcrypt,sha384)")
+#define WOLFKM_PKCS1_SHA384_DRIVER ("pkcs1pad-rsa-sha384" WOLFKM_RSA_DRIVER_SUFFIX)
 
 #define WOLFKM_PKCS1_SHA512_NAME   (PKCS1_NAME "(rsa,sha512)")
-#define WOLFKM_PKCS1_SHA512_DRIVER (PKCS1_NAME "(rsa" WOLFKM_DRIVER_FIPS \
-                                    "-wolfcrypt,sha512)")
+#define WOLFKM_PKCS1_SHA512_DRIVER ("pkcs1pad-rsa-sha512" WOLFKM_RSA_DRIVER_SUFFIX)
 
 #define WOLFKM_PKCS1_SHA3_256_NAME   (PKCS1_NAME "(rsa,sha3-256)")
-#define WOLFKM_PKCS1_SHA3_256_DRIVER (PKCS1_NAME "(rsa" WOLFKM_DRIVER_FIPS \
-                                      "-wolfcrypt,sha3-256)")
+#define WOLFKM_PKCS1_SHA3_256_DRIVER ("pkcs1pad-rsa-sha3-256" WOLFKM_RSA_DRIVER_SUFFIX)
 
 #define WOLFKM_PKCS1_SHA3_384_NAME   (PKCS1_NAME "(rsa,sha3-384)")
-#define WOLFKM_PKCS1_SHA3_384_DRIVER (PKCS1_NAME "(rsa" WOLFKM_DRIVER_FIPS \
-                                      "-wolfcrypt,sha3-384)")
+#define WOLFKM_PKCS1_SHA3_384_DRIVER ("pkcs1pad-rsa-sha3-384" WOLFKM_RSA_DRIVER_SUFFIX)
 
 #define WOLFKM_PKCS1_SHA3_512_NAME   (PKCS1_NAME "(rsa,sha3-512)")
-#define WOLFKM_PKCS1_SHA3_512_DRIVER (PKCS1_NAME "(rsa" WOLFKM_DRIVER_FIPS \
-                                      "-wolfcrypt,sha3-512)")
+#define WOLFKM_PKCS1_SHA3_512_DRIVER ("pkcs1pad-rsa-sha3-512" WOLFKM_RSA_DRIVER_SUFFIX)
 
 #if defined(WOLFSSL_KEY_GEN)
     #if defined(LINUXKM_DIRECT_RSA)
diff --git a/linuxkm/lkcapi_sha_glue.c b/linuxkm/lkcapi_sha_glue.c
index 39754a6424..9a80247bc3 100644
--- a/linuxkm/lkcapi_sha_glue.c
+++ b/linuxkm/lkcapi_sha_glue.c
@@ -1009,7 +1009,7 @@ static volatile int wc_linuxkm_rng_initing_default_bank_flag = 0;
 
 static int linuxkm_affinity_lock(void *arg) {
     (void)arg;
-    if (preempt_count() != 0)
+    if (! wc_linuxkm_can_block())
         return ALREADY_E;
 #if defined(CONFIG_SMP) && (LINUX_VERSION_CODE >= KERNEL_VERSION(5, 7, 0))
     migrate_disable(); /* this actually makes irq_count() nonzero, so that
@@ -1118,7 +1118,7 @@ static struct wc_rng_bank_inst *linuxkm_get_drbg(struct wc_rng_bank *ctx) {
         WC_RNG_BANK_FLAG_CAN_WAIT |
         WC_RNG_BANK_FLAG_PREFER_AFFINITY_INST;
 
-    if (preempt_count() == 0)
+    if (wc_linuxkm_can_block())
         flags |= WC_RNG_BANK_FLAG_AFFINITY_LOCK;
     else
         flags |= WC_RNG_BANK_FLAG_NO_VECTOR_OPS;
@@ -1487,7 +1487,7 @@ static int wc_mix_pool_bytes(const void *buf, size_t len) {
     struct wc_rng_bank *ctx;
     size_t i;
     int n;
-    int can_sleep = (preempt_count() == 0);
+    int can_sleep = wc_linuxkm_can_block();
 
     if (len == 0)
         return 0;
@@ -1545,7 +1545,7 @@ static int wc_mix_pool_bytes(const void *buf, size_t len) {
 
 static int wc_crng_reseed(void) {
     struct wc_rng_bank *ctx;
-    int can_sleep = (preempt_count() == 0);
+    int can_sleep = wc_linuxkm_can_block();
     int ret = wc_rng_bank_default_checkout(&ctx);
 
     if (ret) {
diff --git a/linuxkm/module_exports.c.template b/linuxkm/module_exports.c.template
index f37eea1e49..8209c6277f 100644
--- a/linuxkm/module_exports.c.template
+++ b/linuxkm/module_exports.c.template
@@ -91,6 +91,12 @@
 #ifdef HAVE_ECC
     #include <wolfssl/wolfcrypt/ecc.h>
 #endif
+#ifdef WOLFCRYPT_HAVE_ECCSI
+    #include <wolfssl/wolfcrypt/eccsi.h>
+#endif
+#ifdef WOLFCRYPT_HAVE_SAKKE
+    #include <wolfssl/wolfcrypt/sakke.h>
+#endif
 #ifdef HAVE_HPKE
     #include <wolfssl/wolfcrypt/hpke.h>
 #endif
diff --git a/linuxkm/module_hooks.c b/linuxkm/module_hooks.c
index be1fd01d7b..e5b94ba278 100644
--- a/linuxkm/module_hooks.c
+++ b/linuxkm/module_hooks.c
@@ -316,6 +316,13 @@ MODULE_PARM_DESC(rodata_dump_path,
     #include "linuxkm/lkcapi_glue.c"
 #endif
 
+int wc_linuxkm_can_block(void) {
+    /* We can't use preemptible() for this, because we need an accurate test
+     * even in !CONFIG_PREEMPT_COUNT configs where preemptible() is always 0.
+     */
+    return (preempt_count() == 0) && (! irqs_disabled());
+}
+
 /* for simplicity, we use a global count to suspend signal processing while any
  * thread is running fipsEntry(), wolfCrypt_IntegrityTest_fips(),
  * linuxkm_lkcapi_register(), or linuxkm_lkcapi_unregister().  This only affects
@@ -336,7 +343,7 @@ int wc_linuxkm_sig_ignore_end(void) {
 
 int wc_linuxkm_check_for_intr_signals(void) {
     static const int intr_signals[] = WC_LINUXKM_INTR_SIGNALS;
-    if (preempt_count() != 0)
+    if (! wc_linuxkm_can_block())
         return 0;
     if (signal_pending(current)) {
         int i;
@@ -365,7 +372,7 @@ int wc_linuxkm_check_for_intr_signals(void) {
 
 void wc_linuxkm_relax_long_loop(void) {
     #if WC_LINUXKM_MAX_NS_WITHOUT_YIELD >= 0
-    if (preempt_count() == 0) {
+    if (wc_linuxkm_can_block()) {
         #if (WC_LINUXKM_MAX_NS_WITHOUT_YIELD == 0) || !defined(CONFIG_SCHED_INFO)
         cond_resched();
         #else
@@ -662,6 +669,11 @@ static int wolfssl_init(void)
     }
 #endif
 
+#if defined(WC_LINUXKM_USE_HEAP_WRAPPERS) && defined(CONFIG_HAVE_KPROBES)
+    /* cache the function pointer to find_vm_area(). */
+    (void)wc_linuxkm_malloc_usable_size(NULL);
+#endif
+
 #ifdef WOLFCRYPT_FIPS_CORE_DYNAMIC_HASH_VALUE
 #ifdef CONFIG_MODULE_SIG
     if (THIS_MODULE->sig_ok == false) {
@@ -1283,13 +1295,13 @@ static const struct wc_reloc_table_segments seg_map = {
 
 void *wc_linuxkm_malloc(size_t size)
 {
-    return kvmalloc_node(WC_LINUXKM_ROUND_UP_P_OF_2(size), (preempt_count() == 0 ? GFP_KERNEL : GFP_ATOMIC), NUMA_NO_NODE);
+    return kvmalloc_node(WC_LINUXKM_ROUND_UP_P_OF_2(size), (wc_linuxkm_can_block() ? GFP_KERNEL : GFP_ATOMIC), NUMA_NO_NODE);
 }
 
 void wc_linuxkm_free(void *ptr)
 {
 #if LINUX_VERSION_CODE >= KERNEL_VERSION(7, 2, 0)
-    if (preempt_count() == 0)
+    if (wc_linuxkm_can_block())
         kvfree(ptr);
     else
         kvfree_atomic(ptr);
@@ -1300,12 +1312,41 @@ void wc_linuxkm_free(void *ptr)
 
 void *wc_linuxkm_realloc(void *ptr, size_t newsize)
 {
-    return kvrealloc(ptr, WC_LINUXKM_ROUND_UP_P_OF_2(newsize), (preempt_count() == 0 ? GFP_KERNEL : GFP_ATOMIC));
+    return kvrealloc(ptr, WC_LINUXKM_ROUND_UP_P_OF_2(newsize), (wc_linuxkm_can_block() ? GFP_KERNEL : GFP_ATOMIC));
 }
 
+#ifdef CONFIG_HAVE_KPROBES
+#include <linux/vmalloc.h>
+#endif
+
 size_t wc_linuxkm_malloc_usable_size(void *ptr)
 {
-    return ksize(ptr);
+    if ((ptr == NULL) || is_vmalloc_addr(ptr)) {
+#ifdef CONFIG_HAVE_KPROBES
+        static typeof(find_vm_area) *find_vm_area_ptr = NULL;
+        if (find_vm_area_ptr == NULL) {
+            if (! wc_linuxkm_can_block())
+                return 0;
+            find_vm_area_ptr = my_kallsyms_lookup_name("find_vm_area");
+        }
+        if (find_vm_area_ptr == NULL)
+            return 0;
+        else if (ptr == NULL)
+            return 0;
+        else {
+            struct vm_struct *vm = find_vm_area_ptr(ptr);
+            if (vm)
+                return get_vm_area_size(vm);
+            else
+                return 0;
+        }
+#else
+        return 0;
+#endif
+    }
+    else {
+        return ksize(ptr);
+    }
 }
 
 #endif /* WC_LINUXKM_USE_HEAP_WRAPPERS */
@@ -1678,6 +1719,7 @@ static int set_up_wolfssl_linuxkm_pie_redirect_table(void) {
 #endif
 #endif
 
+    wolfssl_linuxkm_pie_redirect_table.wc_linuxkm_can_block = wc_linuxkm_can_block;
     wolfssl_linuxkm_pie_redirect_table.wc_linuxkm_sig_ignore_begin = wc_linuxkm_sig_ignore_begin;
     wolfssl_linuxkm_pie_redirect_table.wc_linuxkm_sig_ignore_end = wc_linuxkm_sig_ignore_end;
     wolfssl_linuxkm_pie_redirect_table.wc_linuxkm_check_for_intr_signals = wc_linuxkm_check_for_intr_signals;
diff --git a/src/include.am b/src/include.am
index 563a6fa3e8..df6d59c8f7 100644
--- a/src/include.am
+++ b/src/include.am
@@ -247,6 +247,14 @@ endif !BUILD_ARMASM_INLINE
 endif BUILD_ARMASM
 endif !BUILD_ARMASM_NEON
 
+if BUILD_PPC64_ASM
+if BUILD_PPC64_ASM_INLINE
+src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/port/ppc64/ppc64-aes-asm_c.c
+else
+src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/port/ppc64/ppc64-aes-asm.S
+endif !BUILD_PPC64_ASM_INLINE
+endif BUILD_PPC64_ASM
+
 if BUILD_AESNI
 src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/aes_asm.S
 if BUILD_X86_ASM
@@ -512,6 +520,14 @@ endif BUILD_ARMASM
 endif !BUILD_ARMASM_NEON
 endif BUILD_AES
 
+if BUILD_PPC64_ASM
+if BUILD_PPC64_ASM_INLINE
+src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/port/ppc64/ppc64-aes-asm_c.c
+else
+src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/port/ppc64/ppc64-aes-asm.S
+endif !BUILD_PPC64_ASM_INLINE
+endif BUILD_PPC64_ASM
+
 if BUILD_AESNI
 src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/aes_asm.S
 if BUILD_X86_ASM
@@ -1390,6 +1406,14 @@ endif !BUILD_ARMASM_INLINE
 endif BUILD_ARMASM
 endif !BUILD_ARMASM_NEON
 
+if BUILD_PPC64_ASM
+if BUILD_PPC64_ASM_INLINE
+src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/port/ppc64/ppc64-aes-asm_c.c
+else
+src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/port/ppc64/ppc64-aes-asm.S
+endif !BUILD_PPC64_ASM_INLINE
+endif BUILD_PPC64_ASM
+
 if BUILD_AFALG
 src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/port/af_alg/afalg_aes.c
 endif BUILD_AFALG
diff --git a/src/internal.c b/src/internal.c
index 5f09158039..4ca81cfcb2 100644
--- a/src/internal.c
+++ b/src/internal.c
@@ -1423,7 +1423,9 @@ static int ExportOptions(WOLFSSL* ssl, byte* exp, word32 len, byte ver,
     exp[idx++] = options->acceptState;
     exp[idx++] = options->asyncState;
 
-    if (type == WOLFSSL_EXPORT_TLS) {
+    /* Encrypt-Then-MAC state. Historically only serialized for TLS; export
+     * version 6 and later also serialize it for DTLS. */
+    if (type == WOLFSSL_EXPORT_TLS || ver > WOLFSSL_EXPORT_VERSION_5) {
 #ifdef HAVE_ENCRYPT_THEN_MAC
         exp[idx++] = options->disallowEncThenMac;
         exp[idx++] = options->encThenMac;
@@ -1502,6 +1504,13 @@ static int ImportOptions(WOLFSSL* ssl, const byte* exp, word32 len, byte ver,
             }
             break;
 
+        case WOLFSSL_EXPORT_VERSION_5:
+            if (len < DTLS_EXPORT_OPT_SZ_5) {
+                WOLFSSL_MSG("Sanity check on buffer size failed");
+                return BAD_FUNC_ARG;
+            }
+            break;
+
         case WOLFSSL_EXPORT_VERSION_4:
             if (len < DTLS_EXPORT_OPT_SZ_4) {
                 WOLFSSL_MSG("Sanity check on buffer size failed");
@@ -1628,7 +1637,9 @@ static int ImportOptions(WOLFSSL* ssl, const byte* exp, word32 len, byte ver,
     options->acceptState    = exp[idx++];
     options->asyncState     = exp[idx++];
 
-    if (type == WOLFSSL_EXPORT_TLS) {
+    /* Encrypt-Then-MAC state. Historically only serialized for TLS; export
+     * version 6 and later also serialize it for DTLS. */
+    if (type == WOLFSSL_EXPORT_TLS || ver > WOLFSSL_EXPORT_VERSION_5) {
 #ifdef HAVE_ENCRYPT_THEN_MAC
         options->disallowEncThenMac = exp[idx++];
         options->encThenMac         = exp[idx++];
@@ -1723,8 +1734,8 @@ static int ImportPeerInfo(WOLFSSL* ssl, const byte* buf, word32 len, byte ver)
     word16 port;
     char   ip[MAX_EXPORT_IP];
 
-    if (ver != WOLFSSL_EXPORT_VERSION && ver != WOLFSSL_EXPORT_VERSION_4 &&
-            ver != WOLFSSL_EXPORT_VERSION_3) {
+    if (ver != WOLFSSL_EXPORT_VERSION && ver != WOLFSSL_EXPORT_VERSION_5 &&
+            ver != WOLFSSL_EXPORT_VERSION_4 && ver != WOLFSSL_EXPORT_VERSION_3) {
         WOLFSSL_MSG("Export version not supported");
         return BAD_FUNC_ARG;
     }
@@ -1982,6 +1993,15 @@ int wolfSSL_session_import_internal(WOLFSSL* ssl, const unsigned char* buf,
                 }
                 break;
 
+            case WOLFSSL_EXPORT_VERSION_5:
+                if (type == WOLFSSL_EXPORT_DTLS) {
+                    optSz = DTLS_EXPORT_OPT_SZ_5;
+                }
+                else {
+                    optSz = TLS_EXPORT_OPT_SZ_5;
+                }
+                break;
+
             case WOLFSSL_EXPORT_VERSION_4:
                 if (type == WOLFSSL_EXPORT_DTLS) {
                     optSz = DTLS_EXPORT_OPT_SZ_4;
@@ -13696,6 +13716,16 @@ int CheckForAltNames(DecodedCert* dCert, const char* domain, word32 domainLen,
             continue;
         }
 
+        /* RFC 9525 Sec. 6.3: a DNS-ID reference identifier is matched only
+         * against dNSName SAN entries, never uniformResourceIdentifier
+         * (even when the URI value resembles a ostname). URI-ID matching
+         * requires scheme and host parsing (RFC 9525 Sec. 6.5, Sec. 7.2). */
+        if (!isIP && altName->type == ASN_URI_TYPE) {
+            WOLFSSL_MSG("\tAltName is uniformResourceIdentifier, "
+                        "skipping for DNS hostname");
+            continue;
+        }
+
         if (MatchDomainName(buf, (int)len, domain, domainLen, flags)) {
             match = 1;
             if (checkCN != NULL) {
diff --git a/src/ocsp.c b/src/ocsp.c
index b90fcc8af9..c84eebc5e5 100644
--- a/src/ocsp.c
+++ b/src/ocsp.c
@@ -776,7 +776,8 @@ int wolfSSL_OCSP_resp_find_status(WOLFSSL_OCSP_BASICRESP *bs,
     single = bs->single;
     while (single != NULL) {
         if (single->status != NULL && id->status != NULL &&
-            (XMEMCMP(single->status->serial, id->status->serial,
+            (single->status->serialSz == id->status->serialSz)
+         && (XMEMCMP(single->status->serial, id->status->serial,
                      (size_t)single->status->serialSz) == 0)
          && (XMEMCMP(single->issuerHash, id->issuerHash, OCSP_DIGEST_SIZE) == 0)
          && (XMEMCMP(single->issuerKeyHash, id->issuerKeyHash, OCSP_DIGEST_SIZE) == 0)) {
@@ -1270,7 +1271,7 @@ OcspResponse* wolfSSL_d2i_OCSP_RESPONSE(OcspResponse** response,
     int length = 0;
     int ret;
 
-    if (data == NULL)
+    if (data == NULL || *data == NULL || len <= 0)
         return NULL;
 
     if (response != NULL)
@@ -1283,30 +1284,27 @@ OcspResponse* wolfSSL_d2i_OCSP_RESPONSE(OcspResponse** response,
         XMEMSET(resp, 0, sizeof(OcspResponse));
     }
 
+    if (resp->source != NULL)
+        XFREE(resp->source, NULL, DYNAMIC_TYPE_TMP_BUFFER);
     resp->source = (byte*)XMALLOC((size_t)len, NULL, DYNAMIC_TYPE_TMP_BUFFER);
-    if (resp->source == NULL) {
-        XFREE(resp, NULL, DYNAMIC_TYPE_OCSP_REQUEST);
-        return NULL;
+    if (resp->source == NULL)
+        goto error;
+
+    if (resp->single != NULL) {
+        FreeOcspEntry(resp->single, NULL);
+        XFREE(resp->single, NULL, DYNAMIC_TYPE_OCSP_ENTRY);
     }
     resp->single = (OcspEntry*)XMALLOC(sizeof(OcspEntry), NULL,
                                       DYNAMIC_TYPE_OCSP_ENTRY);
-    if (resp->single == NULL) {
-        XFREE(resp->source, NULL, DYNAMIC_TYPE_TMP_BUFFER);
-        XFREE(resp, NULL, DYNAMIC_TYPE_OCSP_REQUEST);
-        return NULL;
-    }
+    if (resp->single == NULL)
+        goto error;
     XMEMSET(resp->single, 0, sizeof(OcspEntry));
     resp->single->status = (CertStatus*)XMALLOC(sizeof(CertStatus), NULL,
                                       DYNAMIC_TYPE_OCSP_STATUS);
+    if (resp->single->status == NULL)
+        goto error;
     resp->single->ownStatus = 1;
-    if (resp->single->status == NULL) {
-        XFREE(resp->source, NULL, DYNAMIC_TYPE_TMP_BUFFER);
-        XFREE(resp->single, NULL, DYNAMIC_TYPE_OCSP_ENTRY);
-        XFREE(resp, NULL, DYNAMIC_TYPE_OCSP_REQUEST);
-        return NULL;
-    }
     XMEMSET(resp->single->status, 0, sizeof(CertStatus));
-
     XMEMCPY(resp->source, *data, (size_t)len);
     resp->maxIdx = (word32)len;
 
@@ -1314,8 +1312,7 @@ OcspResponse* wolfSSL_d2i_OCSP_RESPONSE(OcspResponse** response,
     if (ret != 0 && ret != WC_NO_ERR_TRACE(ASN_OCSP_CONFIRM_E)) {
         /* for just converting from a DER to an internal structure the CA may
          * not yet be known to this function for signature verification */
-        wolfSSL_OCSP_RESPONSE_free(resp);
-        return NULL;
+        goto error;
     }
 
     if (GetSequence(*data, &idx, &length, (word32)len) >= 0)
@@ -1325,6 +1322,12 @@ OcspResponse* wolfSSL_d2i_OCSP_RESPONSE(OcspResponse** response,
         *response = resp;
 
     return resp;
+
+error:
+    wolfSSL_OCSP_RESPONSE_free(resp);
+    if (response != NULL && *response == resp)
+        *response = NULL;
+    return NULL;
 }
 
 int wolfSSL_i2d_OCSP_RESPONSE(OcspResponse* response,
diff --git a/src/pk.c b/src/pk.c
index 410cce3a95..131e1367e0 100644
--- a/src/pk.c
+++ b/src/pk.c
@@ -292,11 +292,10 @@ static int der_write_to_bio_as_pem(const unsigned char* der, int derSz,
 #endif
 #endif
 
-#if defined(OPENSSL_EXTRA) && \
-    ((!defined(NO_RSA) && defined(WOLFSSL_KEY_GEN)) || \
-     (!defined(NO_DH) && defined(WOLFSSL_DH_EXTRA)) || \
-     (defined(HAVE_ECC) && defined(WOLFSSL_KEY_GEN)))
-#if !defined(NO_FILESYSTEM)
+#if !defined(NO_FILESYSTEM) && \
+    ((defined(OPENSSL_EXTRA) && !defined(NO_CERTS) && !defined(NO_ASN) && \
+      !defined(NO_PWDBASED)) || \
+     defined(WOLFSSL_DH_EXTRA))
 /* Write the DER data as PEM into file pointer.
  *
  * @param [in] der    Buffer containing DER data.
@@ -326,8 +325,9 @@ static int der_write_to_file_as_pem(const unsigned char* der, int derSz,
     XFREE(pem, NULL, DYNAMIC_TYPE_TMP_BUFFER);
     return ret;
 }
-#endif
-#endif
+#endif /* !NO_FILESYSTEM &&
+        * ((OPENSSL_EXTRA && !NO_CERTS && !NO_ASN && !NO_PWDBASED) ||
+        *  WOLFSSL_DH_EXTRA) */
 
 #if defined(OPENSSL_EXTRA) && defined(WOLFSSL_KEY_GEN) && \
     defined(WOLFSSL_PEM_TO_DER)
@@ -6282,6 +6282,166 @@ int wolfSSL_PEM_write_bio_PrivateKey(WOLFSSL_BIO* bio, WOLFSSL_EVP_PKEY* key,
 }
 #endif /* !NO_BIO */
 
+#if !defined(NO_FILESYSTEM) && !defined(NO_CERTS) && defined(OPENSSL_EXTRA) && \
+    !defined(NO_ASN) && !defined(NO_PWDBASED)
+/* Writes a public key to a file pointer encoded in PEM format.
+ *
+ * @param [in] fp   File pointer to write to.
+ * @param [in] key  Public key to write in PEM format.
+ * @return  1 on success.
+ * @return  0 on failure.
+ */
+int wolfSSL_PEM_write_PUBKEY(XFILE fp, WOLFSSL_EVP_PKEY* key)
+{
+    int err = 0;
+    unsigned char* derBuf = NULL;
+    int derSz = 0;
+
+    WOLFSSL_ENTER("wolfSSL_PEM_write_PUBKEY");
+
+    if ((fp == XBADFILE) || (key == NULL)) {
+        WOLFSSL_MSG("Bad Function Arguments");
+        err = 1;
+    }
+
+    if (!err) {
+        derSz = wolfSSL_i2d_PUBKEY(key, NULL);
+        if (derSz <= 0) {
+            WOLFSSL_MSG("Failed to get DER size for key");
+            err = 1;
+        }
+    }
+
+    if (!err) {
+        unsigned char* tmp;
+        derBuf = (unsigned char*)XMALLOC((size_t)derSz, NULL,
+            DYNAMIC_TYPE_TMP_BUFFER);
+        if (derBuf == NULL) {
+            WOLFSSL_MSG("Failed to allocate DER buffer");
+            err = 1;
+        }
+        else {
+            tmp = derBuf;
+            if (wolfSSL_i2d_PUBKEY(key, &tmp) <= 0) {
+                WOLFSSL_MSG("Failed to convert key to DER");
+                err = 1;
+            }
+        }
+    }
+
+    /* Write DER buffer to file as PEM. */
+    if ((!err) && (der_write_to_file_as_pem(derBuf, derSz, fp,
+            PUBLICKEY_TYPE, NULL) != 1)) {
+        WOLFSSL_MSG("Failed to write DER to file as PEM");
+        err = 1;
+    }
+
+    /* Dispose of the DER encoding. */
+    XFREE(derBuf, NULL, DYNAMIC_TYPE_TMP_BUFFER);
+
+    WOLFSSL_LEAVE("wolfSSL_PEM_write_PUBKEY", err);
+    return !err;
+}
+
+/* Writes a private key to a file pointer encoded in PEM format.
+ *
+ * @param [in] fp      File pointer to write to.
+ * @param [in] key     Private key to write in PEM format.
+ * @param [in] cipher  Encryption cipher to use. May be NULL.
+ * @param [in] passwd  Password to use when encrypting. May be NULL.
+ * @param [in] len     Length of password.
+ * @param [in] cb      Password callback.
+ * @param [in] arg     Password callback argument.
+ * @return  1 on success.
+ * @return  0 on failure.
+ */
+int wolfSSL_PEM_write_PrivateKey(XFILE fp, WOLFSSL_EVP_PKEY* key,
+    const WOLFSSL_EVP_CIPHER* cipher, unsigned char* passwd, int len,
+    wc_pem_password_cb* cb, void* arg)
+{
+    int err = 0;
+    int type = 0;
+    unsigned char* derBuf = NULL;
+    int derSz = 0;
+
+    (void)cipher;
+    (void)passwd;
+    (void)len;
+    (void)cb;
+    (void)arg;
+
+    WOLFSSL_ENTER("wolfSSL_PEM_write_PrivateKey");
+
+    /* Validate parameters. */
+    if ((fp == XBADFILE) || (key == NULL)) {
+        WOLFSSL_MSG("Bad Function Arguments");
+        err = 1;
+    }
+
+    /* Determine PEM type from key type, mirroring wolfSSL_PEM_read_PrivateKey's
+     * keyFormat switch. */
+    if (!err) {
+        switch (key->type) {
+            case WC_EVP_PKEY_RSA:
+                type = PRIVATEKEY_TYPE;
+                break;
+            case WC_EVP_PKEY_DSA:
+                type = DSA_PRIVATEKEY_TYPE;
+                break;
+            case WC_EVP_PKEY_EC:
+                type = ECC_PRIVATEKEY_TYPE;
+                break;
+            case WC_EVP_PKEY_DH:
+                type = DH_PRIVATEKEY_TYPE;
+                break;
+            default:
+                WOLFSSL_MSG("Unknown key type");
+                err = 1;
+                break;
+        }
+    }
+
+    if (!err) {
+        derSz = wolfSSL_i2d_PrivateKey(key, NULL);
+        if (derSz <= 0) {
+            WOLFSSL_MSG("Failed to get DER size for private key");
+            err = 1;
+        }
+    }
+
+    if (!err) {
+        unsigned char* tmp;
+        derBuf = (unsigned char*)XMALLOC((size_t)derSz, NULL,
+            DYNAMIC_TYPE_TMP_BUFFER);
+        if (derBuf == NULL) {
+            WOLFSSL_MSG("Failed to allocate DER buffer");
+            err = 1;
+        }
+        else {
+            tmp = derBuf;
+            if (wolfSSL_i2d_PrivateKey(key, &tmp) <= 0) {
+                WOLFSSL_MSG("Error encoding private key as DER");
+                err = 1;
+            }
+        }
+    }
+
+    /* Write DER buffer to file as PEM. */
+    if ((!err) && (der_write_to_file_as_pem(derBuf, derSz, fp, type,
+            NULL) != 1)) {
+        WOLFSSL_MSG("Error writing DER to file as PEM");
+        err = 1;
+    }
+
+    /* Dispose of the DER encoding. */
+    XFREE(derBuf, NULL, DYNAMIC_TYPE_TMP_BUFFER);
+
+    WOLFSSL_LEAVE("wolfSSL_PEM_write_PrivateKey", err);
+    return !err;
+}
+#endif /* !NO_FILESYSTEM && !NO_CERTS && OPENSSL_EXTRA && !NO_ASN &&
+        * !NO_PWDBASED */
+
 #ifndef NO_BIO
 /* Create a private key object from the data in the BIO.
  *
diff --git a/src/ssl.c b/src/ssl.c
index 2669f37aa1..b7fe52ddde 100644
--- a/src/ssl.c
+++ b/src/ssl.c
@@ -12978,44 +12978,62 @@ int wolfSSL_set_tlsext_max_fragment_length(WOLFSSL *s, unsigned char mode)
 size_t wolfSSL_get_finished(const WOLFSSL *ssl, void *buf, size_t count)
 {
     byte len = 0;
+    byte const * src;
 
     WOLFSSL_ENTER("wolfSSL_get_finished");
 
-    if (!ssl || !buf || count < TLS_FINISHED_SZ) {
+    if (!ssl || !buf) {
         WOLFSSL_MSG("Bad parameter");
         return WOLFSSL_FAILURE;
     }
 
     if (ssl->options.side == WOLFSSL_SERVER_END) {
+        src = ssl->serverFinished;
         len = ssl->serverFinished_len;
-        XMEMCPY(buf, ssl->serverFinished, len);
     }
     else {
+        src = ssl->clientFinished;
         len = ssl->clientFinished_len;
-        XMEMCPY(buf, ssl->clientFinished, len);
     }
+
+    if (count < len) {
+        WOLFSSL_MSG("Buffer too small");
+        return WOLFSSL_FAILURE;
+    }
+
+    XMEMCPY(buf, src, len);
+
     return len;
 }
 
 size_t wolfSSL_get_peer_finished(const WOLFSSL *ssl, void *buf, size_t count)
 {
     byte len = 0;
+    byte const * src;
+
     WOLFSSL_ENTER("wolfSSL_get_peer_finished");
 
-    if (!ssl || !buf || count < TLS_FINISHED_SZ) {
+    if (!ssl || !buf) {
         WOLFSSL_MSG("Bad parameter");
         return WOLFSSL_FAILURE;
     }
 
     if (ssl->options.side == WOLFSSL_CLIENT_END) {
+        src = ssl->serverFinished;
         len = ssl->serverFinished_len;
-        XMEMCPY(buf, ssl->serverFinished, len);
     }
     else {
+        src = ssl->clientFinished;
         len = ssl->clientFinished_len;
-        XMEMCPY(buf, ssl->clientFinished, len);
     }
 
+    if (count < len) {
+        WOLFSSL_MSG("Buffer too small");
+        return WOLFSSL_FAILURE;
+    }
+
+    XMEMCPY(buf, src, len);
+
     return len;
 }
 #endif /* WOLFSSL_HAVE_TLS_UNIQUE */
diff --git a/src/x509.c b/src/x509.c
index 73f4c3c92e..11b74af7fd 100644
--- a/src/x509.c
+++ b/src/x509.c
@@ -945,10 +945,20 @@ WOLFSSL_X509_EXTENSION* wolfSSL_X509_set_ext(WOLFSSL_X509* x509, int loc)
                     WC_FREE_VAR_EX(cert, NULL, DYNAMIC_TYPE_DCERT);
                     return NULL;
                 }
-                a->length = (int)x509->pathLength;
+                /* RFC 5280 4.2.1.9: only populate pathLen when the
+                 * pathLenConstraint field was actually present. When it is
+                 * absent no limit is imposed, and pathlen must remain NULL so
+                 * it is distinguishable from a pathLenConstraint of 0. */
+                if (x509->basicConstPlSet) {
+                    a->length = (int)x509->pathLength;
 
-                /* Save ASN1_INTEGER in x509 extension */
-                ext->obj->pathlen = a;
+                    /* Save ASN1_INTEGER in x509 extension */
+                    ext->obj->pathlen = a;
+                }
+                else {
+                    wolfSSL_ASN1_INTEGER_free(a);
+                    ext->obj->pathlen = NULL;
+                }
 
                 ext->obj->ca = x509->isCa;
                 break;
@@ -1234,6 +1244,136 @@ static int asn1_string_copy_to_buffer(WOLFSSL_ASN1_STRING* str, byte** buf,
     return WOLFSSL_SUCCESS;
 }
 
+/* Handle WC_NID_subject_alt_name for wolfSSL_X509_add_ext: iterate the
+ * GENERAL_NAME stack on the extension and add each entry to x509->altNames.
+ * Only types whose union member is WOLFSSL_ASN1_STRING* (DNS/RFC822/URI/IP/
+ * IA5) are read via gn->d.ia5; ASN_OTHER_TYPE is handled via SetOthername.
+ * Other types (DIRNAME/RID/X400/EDIPARTY) are rejected to avoid the
+ * type-confused union access that would XMEMCPY from a wild pointer in
+ * wolfSSL_X509_add_altname_ex. */
+static int wolfssl_x509_add_subj_alt_name_ext(WOLFSSL_X509 *x509,
+    WOLFSSL_X509_EXTENSION *ext)
+{
+    WOLFSSL_GENERAL_NAMES* gns = ext->ext_sk;
+    while (gns) {
+        WOLFSSL_GENERAL_NAME* gn = gns->data.gn;
+        if ((gn != NULL) && (gn->type == ASN_OTHER_TYPE)) {
+            char *buf = NULL;
+            int ret = 0;
+            word32 len = 0;
+
+            len = SetOthername(gn->d.otherName, NULL);
+            if (len == WC_NO_ERR_TRACE(WOLFSSL_FAILURE)) {
+                return WOLFSSL_FAILURE;
+            }
+
+            buf = (char*)XMALLOC(len, x509->heap, DYNAMIC_TYPE_X509_EXT);
+            if (buf == NULL) {
+                WOLFSSL_MSG("Couldn't allocate memory for othername");
+                return WOLFSSL_FAILURE;
+            }
+
+            /* SetOthername() cannot fail; already passed above. */
+            SetOthername(gn->d.otherName, (byte*)buf);
+
+            ret = wolfSSL_X509_add_altname_ex(x509, buf, len,
+                                              ASN_OTHER_TYPE);
+            XFREE(buf, x509->heap, DYNAMIC_TYPE_X509_EXT);
+            if (ret == WC_NO_ERR_TRACE(WOLFSSL_FAILURE)) {
+                 WOLFSSL_MSG("wolfSSL_X509_add_altname_ex() failed");
+                 return WOLFSSL_FAILURE;
+            }
+        }
+        /* Only types whose union member are WOLFSSL_ASN1_STRING. */
+        else if (gn == NULL) {
+            WOLFSSL_MSG("Subject alternative name missing");
+            return WOLFSSL_FAILURE;
+        }
+        else {
+            if ((gn->type != ASN_RFC822_TYPE && gn->type != ASN_DNS_TYPE &&
+                 gn->type != ASN_URI_TYPE && gn->type != ASN_IP_TYPE &&
+                 gn->type != WOLFSSL_GEN_IA5) || (gn->d.ia5 == NULL)) {
+                WOLFSSL_MSG("Subject alternative name unsupported "
+                            "GeneralName type or missing name");
+                return WOLFSSL_FAILURE;
+            }
+            if (wolfSSL_X509_add_altname_ex(x509, gn->d.ia5->data,
+                         gn->d.ia5->length, gn->type) != WOLFSSL_SUCCESS) {
+                WOLFSSL_MSG("Failed to add subject alternative name");
+                return WOLFSSL_FAILURE;
+            }
+        }
+        gns = gns->next;
+    }
+    x509->subjAltNameSet = 1;
+    x509->subjAltNameCrit = (byte)ext->crit;
+    return WOLFSSL_SUCCESS;
+}
+
+#ifdef WOLFSSL_CUSTOM_OID
+/* Handle the default (unrecognized NID) case of wolfSSL_X509_add_ext when
+ * custom-OID extensions are enabled: copy the extension OID text and value
+ * into the next free slot in x509->custom_exts, taking ownership of the
+ * allocations on success. */
+static int wolfssl_x509_add_custom_ext(WOLFSSL_X509 *x509,
+    WOLFSSL_X509_EXTENSION *ext)
+{
+    char *oid = NULL;
+    byte *val = NULL;
+    int err = 0;
+
+    if ((ext->obj == NULL) || (ext->value.length == 0) ||
+        (ext->value.data == NULL)) {
+        WOLFSSL_MSG("Extension has insufficient information.");
+        return WOLFSSL_FAILURE;
+    }
+
+    if ((x509->customExtCount < 0) ||
+        (x509->customExtCount >= NUM_CUSTOM_EXT)) {
+        WOLFSSL_MSG("Bad value for customExtCount.");
+        return WOLFSSL_FAILURE;
+    }
+
+    /* This is a viable custom extension. */
+    oid = (char*)XMALLOC(MAX_OID_STRING_SZ, x509->heap,
+        DYNAMIC_TYPE_X509_EXT);
+    val = (byte*)XMALLOC(ext->value.length, x509->heap,
+        DYNAMIC_TYPE_X509_EXT);
+    if ((oid == NULL) || (val == NULL)) {
+        WOLFSSL_MSG("Memory allocation failure.\n");
+        err = 1;
+    }
+
+    if (err == 0) {
+        XMEMCPY(val, ext->value.data, ext->value.length);
+        if (wolfSSL_OBJ_obj2txt(oid, MAX_OID_STRING_SZ, ext->obj, 1) < 0) {
+            err = 1;
+        }
+    }
+
+    if (err == 1) {
+        XFREE(val, x509->heap, DYNAMIC_TYPE_X509_EXT);
+        XFREE(oid, x509->heap, DYNAMIC_TYPE_X509_EXT);
+        return WOLFSSL_FAILURE;
+    }
+
+    /* ext->crit is WOLFSSL_ASN1_BOOLEAN */
+    if (ext->crit != 0 && ext->crit != -1) {
+        XFREE(val, x509->heap, DYNAMIC_TYPE_X509_EXT);
+        XFREE(oid, x509->heap, DYNAMIC_TYPE_X509_EXT);
+        return WOLFSSL_FAILURE;
+    }
+
+    /* x509->custom_exts now owns the buffers and they must be managed. */
+    x509->custom_exts[x509->customExtCount].oid = oid;
+    x509->custom_exts[x509->customExtCount].crit = (byte)ext->crit;
+    x509->custom_exts[x509->customExtCount].val = val;
+    x509->custom_exts[x509->customExtCount].valSz = ext->value.length;
+    x509->customExtCount++;
+    return WOLFSSL_SUCCESS;
+}
+#endif /* WOLFSSL_CUSTOM_OID */
+
 int wolfSSL_X509_add_ext(WOLFSSL_X509 *x509, WOLFSSL_X509_EXTENSION *ext,
     int loc)
 {
@@ -1272,49 +1412,10 @@ int wolfSSL_X509_add_ext(WOLFSSL_X509 *x509, WOLFSSL_X509_EXTENSION *ext,
         x509->subjKeyIdCrit = (byte)ext->crit;
         break;
     case WC_NID_subject_alt_name:
-    {
-        WOLFSSL_GENERAL_NAMES* gns = ext->ext_sk;
-        while (gns) {
-            WOLFSSL_GENERAL_NAME* gn = gns->data.gn;
-            if ((gn != NULL) && (gn->type == ASN_OTHER_TYPE)) {
-                char *buf = NULL;
-                int ret = 0;
-                word32 len = 0;
-
-                len = SetOthername(gn->d.otherName, NULL);
-                if (len == WC_NO_ERR_TRACE(WOLFSSL_FAILURE)) {
-                    return WOLFSSL_FAILURE;
-                }
-
-                buf = (char*)XMALLOC(len, x509->heap, DYNAMIC_TYPE_X509_EXT);
-                if (buf == NULL) {
-                    WOLFSSL_MSG("Couldn't allocate memory for othername");
-                    return WOLFSSL_FAILURE;
-                }
-
-                /* SetOthername() cannot fail; already passed above. */
-                SetOthername(gn->d.otherName, (byte*)buf);
-
-                ret = wolfSSL_X509_add_altname_ex(x509, buf, len,
-                                                  ASN_OTHER_TYPE);
-                XFREE(buf, x509->heap, DYNAMIC_TYPE_X509_EXT);
-                if (ret == WC_NO_ERR_TRACE(WOLFSSL_FAILURE)) {
-                     WOLFSSL_MSG("wolfSSL_X509_add_altname_ex() failed");
-                     return WOLFSSL_FAILURE;
-                }
-            }
-            else if (!gn || !gn->d.ia5 ||
-                wolfSSL_X509_add_altname_ex(x509, gn->d.ia5->data,
-                    gn->d.ia5->length, gn->type) != WOLFSSL_SUCCESS) {
-                WOLFSSL_MSG("Subject alternative name missing extension");
-                return WOLFSSL_FAILURE;
-            }
-            gns = gns->next;
+        if (wolfssl_x509_add_subj_alt_name_ext(x509, ext) != WOLFSSL_SUCCESS) {
+            return WOLFSSL_FAILURE;
         }
-        x509->subjAltNameSet = 1;
-        x509->subjAltNameCrit = (byte)ext->crit;
         break;
-    }
     case WC_NID_key_usage:
         if (ext && ext->value.data) {
             if (ext->value.length == sizeof(word16)) {
@@ -1376,60 +1477,10 @@ int wolfSSL_X509_add_ext(WOLFSSL_X509 *x509, WOLFSSL_X509_EXTENSION *ext,
         break;
     default:
 #ifdef WOLFSSL_CUSTOM_OID
-    {
-        char *oid = NULL;
-        byte *val = NULL;
-        int err = 0;
-
-        if ((ext->obj == NULL) || (ext->value.length == 0)) {
-            WOLFSSL_MSG("Extension has insufficient information.");
+        if (wolfssl_x509_add_custom_ext(x509, ext) != WOLFSSL_SUCCESS) {
             return WOLFSSL_FAILURE;
         }
-
-        if ((x509->customExtCount < 0) ||
-            (x509->customExtCount >= NUM_CUSTOM_EXT)) {
-            WOLFSSL_MSG("Bad value for customExtCount.");
-            return WOLFSSL_FAILURE;
-        }
-
-        /* This is a viable custom extension. */
-        oid = (char*)XMALLOC(MAX_OID_STRING_SZ, x509->heap,
-            DYNAMIC_TYPE_X509_EXT);
-        val = (byte*)XMALLOC(ext->value.length, x509->heap,
-            DYNAMIC_TYPE_X509_EXT);
-        if ((oid == NULL) || (val == NULL)) {
-            WOLFSSL_MSG("Memory allocation failure.\n");
-            err = 1;
-        }
-
-        if (err == 0) {
-            XMEMCPY(val, ext->value.data, ext->value.length);
-            if (wolfSSL_OBJ_obj2txt(oid, MAX_OID_STRING_SZ, ext->obj, 1) < 0) {
-                err = 1;
-            }
-        }
-
-        if (err == 1) {
-            XFREE(val, x509->heap, DYNAMIC_TYPE_X509_EXT);
-            XFREE(oid, x509->heap, DYNAMIC_TYPE_X509_EXT);
-            return WOLFSSL_FAILURE;
-        }
-
-        /* ext->crit is WOLFSSL_ASN1_BOOLEAN */
-        if (ext->crit != 0 && ext->crit != -1) {
-            XFREE(val, x509->heap, DYNAMIC_TYPE_X509_EXT);
-            XFREE(oid, x509->heap, DYNAMIC_TYPE_X509_EXT);
-            return WOLFSSL_FAILURE;
-        }
-
-        /* x509->custom_exts now owns the buffers and they must be managed. */
-        x509->custom_exts[x509->customExtCount].oid = oid;
-        x509->custom_exts[x509->customExtCount].crit = (byte)ext->crit;
-        x509->custom_exts[x509->customExtCount].val = val;
-        x509->custom_exts[x509->customExtCount].valSz = ext->value.length;
-        x509->customExtCount++;
         break;
-    }
 #else
         WOLFSSL_MSG("Unsupported extension to add");
         return WOLFSSL_FAILURE;
@@ -1529,9 +1580,9 @@ int wolfSSL_X509V3_EXT_print(WOLFSSL_BIO *out, WOLFSSL_X509_EXTENSION *ext,
                         WOLFSSL_MSG("Memory error");
                         return rc;
                     }
-                    valLen = XSNPRINTF(val, (size_t)len, "%*s%s", indent, "",
-                            str->strData);
-                    if ((valLen < 0) || (valLen >= len)
+                    valLen = XSNPRINTF(val, (size_t)len + indent,
+                            "%*s%.*s", indent, "", str->length, str->data);
+                    if ((valLen < 0) || (valLen >= len + indent)
                             || ((tmpLen + valLen) >= tmpSz)) {
                         XFREE(val, NULL, DYNAMIC_TYPE_TMP_BUFFER);
                         return rc;
@@ -1549,6 +1600,10 @@ int wolfSSL_X509V3_EXT_print(WOLFSSL_BIO *out, WOLFSSL_X509_EXTENSION *ext,
         {
             char* asn1str;
             asn1str = wolfSSL_i2s_ASN1_STRING(NULL, str);
+            if (asn1str == NULL) {
+                WOLFSSL_MSG("wolfSSL_i2s_ASN1_STRING returned NULL");
+                return rc;
+            }
             tmpLen = XSNPRINTF(tmp, tmpSz, "%*s%s", indent, "", asn1str);
             XFREE(asn1str, NULL, DYNAMIC_TYPE_TMP_BUFFER);
             if (tmpLen >= tmpSz)
@@ -1961,7 +2016,13 @@ void* wolfSSL_X509V3_EXT_d2i(WOLFSSL_X509_EXTENSION* ext)
             }
             /* Copy pathlen and CA into BASIC_CONSTRAINTS from object */
             bc->ca = object->ca;
-            if (object->pathlen != NULL && object->pathlen->length > 0) {
+            /* RFC 5280 4.2.1.9:
+             *   "A pathLenConstraint of zero indicates that no non-self-issued
+             *    intermediate CA certificates may follow in a valid
+             *    certification path."
+             * Check for length of 0 or greater.
+             */
+            if ((object->pathlen != NULL) && (object->pathlen->length >= 0)) {
                 bc->pathlen = wolfSSL_ASN1_INTEGER_dup(object->pathlen);
                 if (bc->pathlen == NULL) {
                     WOLFSSL_MSG("Failed to duplicate ASN1_INTEGER");
@@ -2433,27 +2494,31 @@ void* wolfSSL_X509_get_ext_d2i(const WOLFSSL_X509* x509, int nid, int* c,
     switch (nid) {
         case BASIC_CA_OID:
             if (x509->basicConstSet) {
-                WOLFSSL_ASN1_INTEGER* a;
-
                 bc = wolfSSL_BASIC_CONSTRAINTS_new();
                 if (!bc) {
                     WOLFSSL_MSG("wolfSSL_BASIC_CONSTRAINTS_new error");
                     return NULL;
                 }
 
-                a = wolfSSL_ASN1_INTEGER_new();
-                if (!a) {
-                    WOLFSSL_MSG("wolfSSL_ASN1_INTEGER_new error");
-                    wolfSSL_BASIC_CONSTRAINTS_free(bc);
-                    return NULL;
-                }
-                a->length = (int)x509->pathLength;
-
 #if defined(OPENSSL_ALL) || defined(WOLFSSL_QT) || \
         defined(WOLFSSL_APACHE_HTTPD)
                 bc->ca = x509->isCa;
 #endif
-                bc->pathlen = a;
+                /* RFC 5280 4.2.1.9: only populate pathLen when the
+                 * pathLenConstraint field was present. When absent, no limit
+                 * is imposed and pathlen must remain NULL so it is
+                 * distinguishable from a pathLenConstraint of 0. */
+                if (x509->basicConstPlSet) {
+                    WOLFSSL_ASN1_INTEGER* a = wolfSSL_ASN1_INTEGER_new();
+                    if (a == NULL) {
+                        WOLFSSL_MSG("wolfSSL_ASN1_INTEGER_new error");
+                        wolfSSL_BASIC_CONSTRAINTS_free(bc);
+                        return NULL;
+                    }
+                    a->length = (int)x509->pathLength;
+                    bc->pathlen = a;
+                }
+
                 if (c != NULL) {
                     *c = x509->basicConstCrit;
                 }
@@ -2648,14 +2713,15 @@ void* wolfSSL_X509_get_ext_d2i(const WOLFSSL_X509* x509, int nid, int* c,
                     WOLFSSL_MSG("wolfSSL_sk_GENERAL_NAME_push error");
                     goto err;
                 }
+                /* gn now owned by stack. */
+                gn = NULL;
 
                 /* push DIST_POINT onto stack */
                 if (wolfSSL_sk_DIST_POINT_push(sk, dp) <= 0) {
                     WOLFSSL_MSG("Error pushing DIST_POINT onto stack");
                     goto err;
                 }
-
-                gn = NULL;
+                /* dp now owned by stack. */
                 dp = NULL;
 
             }
@@ -5871,8 +5937,7 @@ int wolfSSL_GENERAL_NAME_print(WOLFSSL_BIO* out, WOLFSSL_GENERAL_NAME* gen)
     case GEN_EMAIL:
         ret = wolfSSL_BIO_printf(out, "email:");
         ret = (ret > 0) ? WOLFSSL_SUCCESS : WOLFSSL_FAILURE;
-        if (ret == WOLFSSL_SUCCESS)
-        {
+        if (ret == WOLFSSL_SUCCESS) {
             ret = wolfSSL_ASN1_STRING_print(out, gen->d.rfc822Name);
         }
         break;
@@ -5881,8 +5946,7 @@ int wolfSSL_GENERAL_NAME_print(WOLFSSL_BIO* out, WOLFSSL_GENERAL_NAME* gen)
         ret = wolfSSL_BIO_printf(out, "DNS:");
         ret = (ret > 0) ? WOLFSSL_SUCCESS : WOLFSSL_FAILURE;
         if (ret == WOLFSSL_SUCCESS) {
-            ret = wolfSSL_BIO_printf(out, "%s", gen->d.dNSName->strData);
-            ret = (ret > 0) ? WOLFSSL_SUCCESS : WOLFSSL_FAILURE;
+            ret = wolfSSL_ASN1_STRING_print(out, gen->d.dNSName);
         }
         break;
 
@@ -5893,6 +5957,7 @@ int wolfSSL_GENERAL_NAME_print(WOLFSSL_BIO* out, WOLFSSL_GENERAL_NAME* gen)
 
     case GEN_DIRNAME:
         ret = wolfSSL_BIO_printf(out, "DirName:");
+        ret = (ret > 0) ? WOLFSSL_SUCCESS : WOLFSSL_FAILURE;
         if (ret == WOLFSSL_SUCCESS) {
             ret = wolfSSL_X509_NAME_print_ex(out, gen->d.directoryName, 0,
                                                          XN_FLAG_ONELINE);
@@ -7094,7 +7159,8 @@ static int X509PrintExtendedKeyUsage(WOLFSSL_BIO* bio, WOLFSSL_X509* x509,
         EXTKEYUSE_EMAILPROT,
         EXTKEYUSE_CODESIGN,
         EXTKEYUSE_CLIENT_AUTH,
-        EXTKEYUSE_SERVER_AUTH
+        EXTKEYUSE_SERVER_AUTH,
+        EXTKEYUSE_ANY
     };
     const char* usageStrs[] = {
         "OCSP Signing",
@@ -7102,7 +7168,8 @@ static int X509PrintExtendedKeyUsage(WOLFSSL_BIO* bio, WOLFSSL_X509* x509,
         "E-mail Protection",
         "Code Signing",
         "TLS Web Client Authentication",
-        "TLS Web Server Authentication"
+        "TLS Web Server Authentication",
+        "Any Extended Key Usage"
     };
 
     if (bio == NULL || x509 == NULL) {
@@ -7483,12 +7550,22 @@ static int X509PrintExtensions(WOLFSSL_BIO* bio, WOLFSSL_X509* x509, int indent)
                     ret = WOLFSSL_FAILURE;
                     break;
                 }
-                if ((scratchLen = XSNPRINTF(
-                         scratch, scratchSz,
-                         "%*sCA:%s\n",
-                         indent + 8, "", (x509->isCa)? "TRUE": "FALSE"))
-                    >= scratchSz)
-                {
+                /* Match OpenSSL output: print "CA:TRUE"/"CA:FALSE" and append
+                 * ", pathlen:N" only when a pathLenConstraint is present
+                 * (RFC 5280 4.2.1.9). An absent constraint imposes no limit
+                 * and must not be shown as pathlen:0. */
+                if (x509->basicConstPlSet) {
+                    scratchLen = XSNPRINTF(scratch, scratchSz,
+                        "%*sCA:%s, pathlen:%u\n",
+                        indent + 8, "", (x509->isCa) ? "TRUE" : "FALSE",
+                        (unsigned int)x509->pathLength);
+                }
+                else {
+                    scratchLen = XSNPRINTF(scratch, scratchSz,
+                        "%*sCA:%s\n",
+                        indent + 8, "", (x509->isCa) ? "TRUE" : "FALSE");
+                }
+                if (scratchLen >= scratchSz) {
                     ret = WOLFSSL_FAILURE;
                     break;
                 }
diff --git a/tests/api.c b/tests/api.c
index 58a9781239..beb1d1b0f5 100644
--- a/tests/api.c
+++ b/tests/api.c
@@ -234,12 +234,15 @@
 #include <tests/api/test_slhdsa.h>
 #include <tests/api/test_signature.h>
 #include <tests/api/test_dtls.h>
+#include <tests/api/test_dtls13.h>
 #include <tests/api/test_ocsp.h>
 #include <tests/api/test_evp.h>
 #include <tests/api/test_tls_ext.h>
 #include <tests/api/test_tls.h>
+#include <tests/api/test_session.h>
 #include <tests/api/test_x509.h>
 #include <tests/api/test_asn.h>
+#include <tests/api/test_lms_xmss.h>
 #include <tests/api/test_pkcs7.h>
 #include <tests/api/test_pkcs12.h>
 #include <tests/api/test_pwdbased.h>
@@ -6466,7 +6469,7 @@ done:
 #if defined(ENABLE_TLS_CALLBACK_TEST) || \
     (defined(WOLFSSL_DTLS) && defined(WOLFSSL_SESSION_EXPORT))
 /* TLS server for API unit testing - generic */
-static THREAD_RETURN WOLFSSL_THREAD run_wolfssl_server(void* args)
+THREAD_RETURN WOLFSSL_THREAD run_wolfssl_server(void* args)
 {
     callback_functions* callbacks = ((func_args*)args)->callbacks;
 
@@ -6723,7 +6726,7 @@ cleanup:
 }
 
 /* TLS Client for API unit testing - generic */
-static void run_wolfssl_client(void* args)
+void run_wolfssl_client(void* args)
 {
     callback_functions* callbacks = ((func_args*)args)->callbacks;
 
@@ -7279,1152 +7282,162 @@ static int test_wolfSSL_get_finished(void)
     return EXPECT_RESULT();
 }
 
-#if defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_EXT_CACHE) && \
-    !defined(SINGLE_THREADED) && defined(WOLFSSL_TLS13) && \
-    !defined(NO_SESSION_CACHE)
-
-/* Sessions to restore/store */
-static WOLFSSL_SESSION* test_wolfSSL_CTX_add_session_client_sess;
-static WOLFSSL_SESSION* test_wolfSSL_CTX_add_session_server_sess;
-static WOLFSSL_CTX*     test_wolfSSL_CTX_add_session_server_ctx;
-
-static void test_wolfSSL_CTX_add_session_ctx_ready(WOLFSSL_CTX* ctx)
-{
-    /* Don't store sessions. Lookup is still enabled. */
-    AssertIntEQ(wolfSSL_CTX_set_session_cache_mode(ctx,
-            WOLFSSL_SESS_CACHE_NO_INTERNAL_STORE), WOLFSSL_SUCCESS);
-#ifdef OPENSSL_EXTRA
-    AssertIntEQ(wolfSSL_CTX_get_session_cache_mode(ctx) &
-            WOLFSSL_SESS_CACHE_NO_INTERNAL_STORE,
-            WOLFSSL_SESS_CACHE_NO_INTERNAL_STORE);
-#endif
-    /* Require both peers to provide certs */
-    wolfSSL_CTX_set_verify(ctx, WOLFSSL_VERIFY_PEER, NULL);
-}
-
-static void test_wolfSSL_CTX_add_session_on_result(WOLFSSL* ssl)
-{
-    WOLFSSL_SESSION** sess;
-#ifdef WOLFSSL_MUTEX_INITIALIZER
-    static wolfSSL_Mutex m = WOLFSSL_MUTEX_INITIALIZER(m);
-
-    (void)wc_LockMutex(&m);
-#endif
-    if (wolfSSL_is_server(ssl))
-        sess = &test_wolfSSL_CTX_add_session_server_sess;
-    else
-        sess = &test_wolfSSL_CTX_add_session_client_sess;
-    if (*sess == NULL) {
-#ifdef NO_SESSION_CACHE_REF
-        *sess = wolfSSL_get1_session(ssl);
-        AssertNotNull(*sess);
-#else
-        /* Test for backwards compatibility */
-        if (wolfSSL_is_server(ssl)) {
-            *sess = wolfSSL_get1_session(ssl);
-            AssertNotNull(*sess);
-        }
-        else {
-            *sess = wolfSSL_get_session(ssl);
-            AssertNotNull(*sess);
-        }
-#endif
-        /* Now save the session in the internal store to make it available
-         * for lookup. For TLS 1.3, we can't save the session without
-         * WOLFSSL_TICKET_HAVE_ID because there is no way to retrieve the
-         * session from cache. */
-        if (wolfSSL_is_server(ssl)
-#ifndef WOLFSSL_TICKET_HAVE_ID
-                && wolfSSL_version(ssl) != TLS1_3_VERSION
-#endif
-                )
-            AssertIntEQ(wolfSSL_CTX_add_session(wolfSSL_get_SSL_CTX(ssl),
-                    *sess), WOLFSSL_SUCCESS);
-    }
-    else {
-        /* If we have a session retrieved then remaining connections should be
-         * resuming on that session */
-        AssertIntEQ(wolfSSL_session_reused(ssl), 1);
-    }
-#ifdef WOLFSSL_MUTEX_INITIALIZER
-    wc_UnLockMutex(&m);
-#endif
-
-    /* Save CTX to be able to decrypt tickets */
-    if (wolfSSL_is_server(ssl) &&
-            test_wolfSSL_CTX_add_session_server_ctx == NULL) {
-        test_wolfSSL_CTX_add_session_server_ctx = wolfSSL_get_SSL_CTX(ssl);
-        AssertNotNull(test_wolfSSL_CTX_add_session_server_ctx);
-        AssertIntEQ(wolfSSL_CTX_up_ref(wolfSSL_get_SSL_CTX(ssl)),
-                WOLFSSL_SUCCESS);
-    }
-#if defined(SESSION_CERTS) && !defined(WOLFSSL_NO_CLIENT_AUTH)
-#ifndef WOLFSSL_TICKET_HAVE_ID
-    if (wolfSSL_version(ssl) != TLS1_3_VERSION &&
-            wolfSSL_session_reused(ssl))
-#endif
-    {
-        /* With WOLFSSL_TICKET_HAVE_ID the peer certs should be available
-         * for all connections. TLS 1.3 only has tickets so if we don't
-         * include the session id in the ticket then the certificates
-         * will not be available on resumption. */
-    #ifdef KEEP_PEER_CERT
-        WOLFSSL_X509* peer = wolfSSL_get_peer_certificate(ssl);
-        AssertNotNull(peer);
-        wolfSSL_X509_free(peer);
-    #endif
-        AssertNotNull(wolfSSL_SESSION_get_peer_chain(*sess));
-    #ifdef OPENSSL_EXTRA
-        AssertNotNull(SSL_SESSION_get0_peer(*sess));
-    #endif
-    }
-#endif /* SESSION_CERTS && !WOLFSSL_NO_CLIENT_AUTH */
-}
-
-static void test_wolfSSL_CTX_add_session_ssl_ready(WOLFSSL* ssl)
-{
-    /* Set the session to reuse for the client */
-    AssertIntEQ(wolfSSL_set_session(ssl,
-            test_wolfSSL_CTX_add_session_client_sess), WOLFSSL_SUCCESS);
-}
-#endif
-
-static int test_wolfSSL_CTX_add_session(void)
+#if defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES) && \
+    defined(WOLFSSL_HAVE_TLS_UNIQUE) && defined(WOLFSSL_TLS13)
+static int test_wolfSSL_get_finished_overrun_on_handshake(WOLFSSL_CTX* ctx,
+    WOLFSSL* ssl)
 {
     EXPECT_DECLS;
-#if defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_EXT_CACHE) && \
-    !defined(SINGLE_THREADED) && defined(WOLFSSL_TLS13) && \
-    !defined(NO_SESSION_CACHE)
-    tcp_ready ready;
-    func_args client_args;
-    func_args server_args;
-    THREAD_TYPE serverThread;
-    callback_functions client_cb;
-    callback_functions server_cb;
-    method_provider methods[][2] = {
-#if !defined(NO_OLD_TLS) && ((!defined(NO_AES) && !defined(NO_AES_CBC)) || \
-        !defined(NO_DES3))
-        /* Without AES there are almost no ciphersuites available. This leads
-         * to no ciphersuites being available and an error. */
-        { wolfTLSv1_1_client_method, wolfTLSv1_1_server_method },
-#endif
-#ifndef WOLFSSL_NO_TLS12
-        { wolfTLSv1_2_client_method, wolfTLSv1_2_server_method },
-#endif
-        /* Needs the default ticket callback since it is tied to the
-         * connection context and this makes it easy to carry over the ticket
-         * crypto context between connections */
-#if defined(WOLFSSL_TLS13) && !defined(WOLFSSL_NO_DEF_TICKET_ENC_CB) && \
-    defined(HAVE_SESSION_TICKET)
-        { wolfTLSv1_3_client_method, wolfTLSv1_3_server_method },
-#endif
-    };
-    const size_t methodsLen = sizeof(methods)/sizeof(*methods);
-    size_t i, j;
+    /* Destination buffer immediately followed by a guard region. Both are
+     * byte arrays (alignment 1) so there is no padding between them and the
+     * guard bytes sit directly after buf. The caller promises only
+     * TLS_FINISHED_SZ writable bytes via the count argument, so wolfSSL must
+     * not write past buf[] into the guard region. */
+    struct {
+        byte buf[TLS_FINISHED_SZ];
+        byte guard[WC_MAX_DIGEST_SIZE];
+    } probe;
+    byte   full[WC_MAX_DIGEST_SIZE];
+    size_t msg_len;
+    word32 i;
 
-    for (i = 0; i < methodsLen; i++) {
-        /* First run creates a connection while the second+ run will attempt
-         * to resume the connection. The trick is that the internal cache
-         * is turned off. wolfSSL_CTX_add_session should put the session in
-         * the cache anyway. */
-        test_wolfSSL_CTX_add_session_client_sess = NULL;
-        test_wolfSSL_CTX_add_session_server_sess = NULL;
-        test_wolfSSL_CTX_add_session_server_ctx = NULL;
-
-#ifdef NO_SESSION_CACHE_REF
-        for (j = 0; j < 4; j++) {
-#else
-        /* The session may be overwritten in this case. Do only one resumption
-         * to stop this test from failing intermittently. */
-        for (j = 0; j < 2; j++) {
-#endif
-#ifdef WOLFSSL_TIRTOS
-            fdOpenSession(Task_self());
-#endif
-
-            StartTCP();
-            InitTcpReady(&ready);
-
-            XMEMSET(&client_args, 0, sizeof(func_args));
-            XMEMSET(&server_args, 0, sizeof(func_args));
-
-            XMEMSET(&client_cb, 0, sizeof(callback_functions));
-            XMEMSET(&server_cb, 0, sizeof(callback_functions));
-            client_cb.method  = methods[i][0];
-            server_cb.method  = methods[i][1];
-
-            server_args.signal    = &ready;
-            server_args.callbacks = &server_cb;
-            client_args.signal    = &ready;
-            client_args.callbacks = &client_cb;
-
-            if (test_wolfSSL_CTX_add_session_server_ctx != NULL) {
-                server_cb.ctx = test_wolfSSL_CTX_add_session_server_ctx;
-                server_cb.isSharedCtx = 1;
-            }
-            server_cb.ctx_ready = test_wolfSSL_CTX_add_session_ctx_ready;
-            client_cb.ctx_ready = test_wolfSSL_CTX_add_session_ctx_ready;
-            if (j != 0)
-                client_cb.ssl_ready = test_wolfSSL_CTX_add_session_ssl_ready;
-            server_cb.on_result = test_wolfSSL_CTX_add_session_on_result;
-            client_cb.on_result = test_wolfSSL_CTX_add_session_on_result;
-            server_cb.ticNoInit = 1; /* Use default builtin */
-
-            start_thread(test_server_nofail, &server_args, &serverThread);
-            wait_tcp_ready(&server_args);
-            test_client_nofail(&client_args, NULL);
-            join_thread(serverThread);
-
-            ExpectTrue(client_args.return_code);
-            ExpectTrue(server_args.return_code);
-
-            FreeTcpReady(&ready);
-
-            if (EXPECT_FAIL())
-                break;
-        }
-        wolfSSL_SESSION_free(test_wolfSSL_CTX_add_session_client_sess);
-        wolfSSL_SESSION_free(test_wolfSSL_CTX_add_session_server_sess);
-        wolfSSL_CTX_free(test_wolfSSL_CTX_add_session_server_ctx);
-
-        if (EXPECT_FAIL())
-            break;
-    }
-#endif
-
-    return EXPECT_RESULT();
-}
-#if defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES) && defined(HAVE_EXT_CACHE) && \
-     defined(WOLFSSL_TLS13) && !defined(NO_SESSION_CACHE) && \
-     defined(OPENSSL_EXTRA) && defined(SESSION_CERTS) && \
-     defined(HAVE_SESSION_TICKET) && \
-    !defined(TITAN_SESSION_CACHE) && \
-    !defined(HUGE_SESSION_CACHE) && \
-    !defined(BIG_SESSION_CACHE) && \
-    !defined(MEDIUM_SESSION_CACHE)
-
-/* twcase - prefix for test_wolfSSL_CTX_add_session_ext */
-/* Sessions to restore/store */
-static WOLFSSL_SESSION* twcase_server_first_session_ptr;
-static WOLFSSL_SESSION* twcase_client_first_session_ptr;
-static WOLFSSL_CTX*     twcase_server_current_ctx_ptr;
-static int twcase_new_session_called    = 0;
-static int twcase_remove_session_called = 0;
-static int twcase_get_session_called    = 0;
-
-/* Test default, SESSIONS_PER_ROW*SESSION_ROWS = 3*11, see ssl.c */
-#define SESSION_CACHE_SIZE 33
-
-typedef struct {
-    const byte* key;  /* key, altSessionID, session ID, NULL if empty */
-    WOLFSSL_SESSION* value;
-} hashTable_entry;
-
-typedef struct {
-    hashTable_entry entries[SESSION_CACHE_SIZE];  /* hash slots */
-    size_t capacity;                     /* size of entries */
-    size_t length;                       /* number of items in the hash table */
-    wolfSSL_Mutex htLock;                /* lock */
-}hashTable;
-
-static hashTable server_sessionCache;
-
-static int twcase_new_sessionCb(WOLFSSL *ssl, WOLFSSL_SESSION *sess)
-{
-    int i;
-    unsigned int len;
-    (void)ssl;
-
-    /*
-     * This example uses a hash table.
-     * Steps you should take for a non-demo code:
-     * - acquire a lock for the file named according to the session id
-     * - open the file
-     * - encrypt and write the SSL_SESSION object to the file
-     * - release the lock
-     *
-     * Return:
-     *  0: The callback does not wish to hold a reference of the sess
-     *  1: The callback wants to hold a reference of the sess. The callback is
-     *     now also responsible for calling wolfSSL_SESSION_free() on sess.
-     */
-    if (sess == NULL)
-        return 0;
-
-    if (wc_LockMutex(&server_sessionCache.htLock) != 0) {
-        return 0;
-    }
-    for (i = 0; i < SESSION_CACHE_SIZE; i++) {
-        if (server_sessionCache.entries[i].value == NULL) {
-            server_sessionCache.entries[i].key = SSL_SESSION_get_id(sess, &len);
-            server_sessionCache.entries[i].value = sess;
-            server_sessionCache.length++;
-            break;
-        }
-    }
-    ++twcase_new_session_called;
-    wc_UnLockMutex(&server_sessionCache.htLock);
-    fprintf(stderr, "\t\ttwcase_new_session_called %d\n",
-            twcase_new_session_called);
-    return 1;
-}
-
-static void twcase_remove_sessionCb(WOLFSSL_CTX *ctx, WOLFSSL_SESSION *sess)
-{
-    int i;
     (void)ctx;
-    (void)sess;
 
-    if (sess == NULL)
-        return;
-    /*
-     * This example uses a hash table.
-     * Steps you should take for a non-demo code:
-     * - acquire a lock for the file named according to the session id
-     * - remove the file
-     * - release the lock
-     */
-    if (wc_LockMutex(&server_sessionCache.htLock) != 0) {
-        return;
+    XMEMSET(&probe, 0, sizeof(probe));
+    XMEMSET(probe.guard, 0x5A, sizeof(probe.guard));
+
+    /* For a TLS 1.3 connection the Finished message is the handshake hash
+     * size (32 bytes for SHA-256, 48 for SHA-384), which is larger than the
+     * count of TLS_FINISHED_SZ (12) we pass here. The buffer is too small to
+     * hold the Finished message, so the call must fail and write nothing
+     * rather than copying the full length and overrunning buf[]. */
+    msg_len = wolfSSL_get_finished(ssl, probe.buf, TLS_FINISHED_SZ);
+    ExpectIntEQ(msg_len, WOLFSSL_FAILURE);
+
+    /* Nothing must have been written: buf[] stays zero and the guard bytes
+     * that follow it stay 0x5A. */
+    for (i = 0; i < (word32)sizeof(probe.buf); i++) {
+        ExpectIntEQ(probe.buf[i], 0x00);
     }
-    for (i = 0; i < SESSION_CACHE_SIZE; i++) {
-        if (server_sessionCache.entries[i].key != NULL &&
-           XMEMCMP(server_sessionCache.entries[i].key,
-                   sess->sessionID, SSL_MAX_SSL_SESSION_ID_LENGTH) == 0) {
-            wolfSSL_SESSION_free(server_sessionCache.entries[i].value);
-            server_sessionCache.entries[i].value = NULL;
-            server_sessionCache.entries[i].key = NULL;
-            server_sessionCache.length--;
-            break;
-        }
-    }
-    ++twcase_remove_session_called;
-    wc_UnLockMutex(&server_sessionCache.htLock);
-    fprintf(stderr, "\t\ttwcase_remove_session_called %d\n",
-            twcase_remove_session_called);
-}
-
-static WOLFSSL_SESSION *twcase_get_sessionCb(WOLFSSL *ssl,
-                                  const unsigned char *id, int len, int *ref)
-{
-    int i;
-    (void)ssl;
-    (void)id;
-    (void)len;
-
-    /*
-     * This example uses a hash table.
-     * Steps you should take for a non-demo code:
-     * - acquire a lock for the file named according to the session id in the
-     *   2nd arg
-     * - read and decrypt contents of file and create a new SSL_SESSION
-     * - object release the lock
-     * - return the new session object
-     */
-    fprintf(stderr, "\t\ttwcase_get_session_called %d\n",
-            ++twcase_get_session_called);
-    /* This callback want to retain a copy of the object. If we want wolfSSL to
-     * be responsible for the pointer then set to 0. */
-    *ref = 1;
-
-    for (i = 0; i < SESSION_CACHE_SIZE; i++) {
-        if (server_sessionCache.entries[i].key != NULL &&
-           XMEMCMP(server_sessionCache.entries[i].key, id,
-                   SSL_MAX_SSL_SESSION_ID_LENGTH) == 0) {
-           return server_sessionCache.entries[i].value;
-        }
-    }
-    return NULL;
-}
-static int twcase_get_sessionCb_cleanup(void)
-{
-    int i;
-    int cnt = 0;
-
-    /* If  twcase_get_sessionCb sets *ref = 1, the application is responsible
-     * for freeing sessions */
-
-    for (i = 0; i < SESSION_CACHE_SIZE; i++) {
-        if (server_sessionCache.entries[i].value != NULL) {
-            wolfSSL_SESSION_free(server_sessionCache.entries[i].value);
-            cnt++;
-        }
+    for (i = 0; i < (word32)sizeof(probe.guard); i++) {
+        ExpectIntEQ(probe.guard[i], 0x5A);
     }
 
-    fprintf(stderr, "\t\ttwcase_get_sessionCb_cleanup freed %d sessions\n",
-            cnt);
-
-    return TEST_SUCCESS;
-}
-
-static int twcase_cache_intOff_extOff(WOLFSSL_CTX* ctx)
-{
-    EXPECT_DECLS;
-    /* off - Disable internal cache */
-    ExpectIntEQ(wolfSSL_CTX_set_session_cache_mode(ctx,
-            WOLFSSL_SESS_CACHE_NO_INTERNAL_STORE), WOLFSSL_SUCCESS);
-#ifdef OPENSSL_EXTRA
-    ExpectIntEQ(wolfSSL_CTX_get_session_cache_mode(ctx) &
-            WOLFSSL_SESS_CACHE_NO_INTERNAL_STORE,
-            WOLFSSL_SESS_CACHE_NO_INTERNAL_STORE);
-#endif
-    /* off - Do not setup external cache */
-
-    /* Require both peers to provide certs */
-    wolfSSL_CTX_set_verify(ctx, WOLFSSL_VERIFY_PEER, NULL);
-    return EXPECT_RESULT();
-}
-
-static int twcase_cache_intOn_extOff(WOLFSSL_CTX* ctx)
-{
-    /* on - internal cache is on by default */
-    /* off - Do not setup external cache */
-    /* Require both peers to provide certs */
-    wolfSSL_CTX_set_verify(ctx, WOLFSSL_VERIFY_PEER, NULL);
-    return TEST_SUCCESS;
-}
-
-static int twcase_cache_intOff_extOn(WOLFSSL_CTX* ctx)
-{
-    EXPECT_DECLS;
-    /* off - Disable internal cache */
-    ExpectIntEQ(wolfSSL_CTX_set_session_cache_mode(ctx,
-            WOLFSSL_SESS_CACHE_NO_INTERNAL_STORE), WOLFSSL_SUCCESS);
-#ifdef OPENSSL_EXTRA
-    ExpectIntEQ(wolfSSL_CTX_get_session_cache_mode(ctx) &
-            WOLFSSL_SESS_CACHE_NO_INTERNAL_STORE,
-            WOLFSSL_SESS_CACHE_NO_INTERNAL_STORE);
-#endif
-    /* on - Enable external cache */
-    wolfSSL_CTX_sess_set_new_cb(ctx, twcase_new_sessionCb);
-    wolfSSL_CTX_sess_set_remove_cb(ctx, twcase_remove_sessionCb);
-    wolfSSL_CTX_sess_set_get_cb(ctx, twcase_get_sessionCb);
-
-    /* Require both peers to provide certs */
-    wolfSSL_CTX_set_verify(ctx, WOLFSSL_VERIFY_PEER, NULL);
-    return EXPECT_RESULT();
-}
-
-static int twcase_cache_intOn_extOn(WOLFSSL_CTX* ctx)
-{
-    /* on - internal cache is on by default */
-    /* on - Enable external cache */
-    wolfSSL_CTX_sess_set_new_cb(ctx, twcase_new_sessionCb);
-    wolfSSL_CTX_sess_set_remove_cb(ctx, twcase_remove_sessionCb);
-    wolfSSL_CTX_sess_set_get_cb(ctx, twcase_get_sessionCb);
-
-    /* Require both peers to provide certs */
-    wolfSSL_CTX_set_verify(ctx, WOLFSSL_VERIFY_PEER, NULL);
-    return TEST_SUCCESS;
-}
-static int twcase_cache_intOn_extOn_noTicket(WOLFSSL_CTX* ctx)
-{
-    /* on - internal cache is on by default */
-    /* on - Enable external cache */
-    wolfSSL_CTX_sess_set_new_cb(ctx, twcase_new_sessionCb);
-    wolfSSL_CTX_sess_set_remove_cb(ctx, twcase_remove_sessionCb);
-    wolfSSL_CTX_sess_set_get_cb(ctx, twcase_get_sessionCb);
-
-    wolfSSL_CTX_set_options(ctx, WOLFSSL_OP_NO_TICKET);
-    /* Require both peers to provide certs */
-    wolfSSL_CTX_set_verify(ctx, WOLFSSL_VERIFY_PEER, NULL);
-    return TEST_SUCCESS;
-}
-static int twcase_server_sess_ctx_pre_shutdown(WOLFSSL* ssl)
-{
-    EXPECT_DECLS;
-    WOLFSSL_SESSION** sess;
-    if (wolfSSL_is_server(ssl))
-        sess = &twcase_server_first_session_ptr;
-    else
-        return TEST_SUCCESS;
-
-    if (*sess == NULL) {
-        ExpectNotNull(*sess = wolfSSL_get1_session(ssl));
-        /* Now save the session in the internal store to make it available
-         * for lookup. For TLS 1.3, we can't save the session without
-         * WOLFSSL_TICKET_HAVE_ID because there is no way to retrieve the
-         * session from cache. */
-        if (wolfSSL_is_server(ssl)
-#ifndef WOLFSSL_TICKET_HAVE_ID
-                && wolfSSL_version(ssl) != TLS1_3_VERSION
-                && wolfSSL_version(ssl) != DTLS1_3_VERSION
-#endif
-                ) {
-            ExpectIntEQ(wolfSSL_CTX_add_session(wolfSSL_get_SSL_CTX(ssl),
-                    *sess), WOLFSSL_SUCCESS);
-        }
-    }
-    /* Save CTX to be able to decrypt tickets */
-    if (twcase_server_current_ctx_ptr == NULL) {
-        ExpectNotNull(twcase_server_current_ctx_ptr = wolfSSL_get_SSL_CTX(ssl));
-        ExpectIntEQ(wolfSSL_CTX_up_ref(wolfSSL_get_SSL_CTX(ssl)),
-                    WOLFSSL_SUCCESS);
-    }
-#if defined(SESSION_CERTS) && !defined(WOLFSSL_NO_CLIENT_AUTH)
-#ifndef WOLFSSL_TICKET_HAVE_ID
-    if (wolfSSL_version(ssl) != TLS1_3_VERSION &&
-            wolfSSL_session_reused(ssl))
-#endif
-    {
-        /* With WOLFSSL_TICKET_HAVE_ID the peer certs should be available
-         * for all connections. TLS 1.3 only has tickets so if we don't
-         * include the session id in the ticket then the certificates
-         * will not be available on resumption. */
-    #ifdef KEEP_PEER_CERT
-        WOLFSSL_X509* peer = NULL;
-        ExpectNotNull(peer = wolfSSL_get_peer_certificate(ssl));
-        wolfSSL_X509_free(peer);
-    #endif
-        ExpectNotNull(wolfSSL_SESSION_get_peer_chain(*sess));
-    }
-#endif
-    return EXPECT_RESULT();
-}
-
-static int twcase_client_sess_ctx_pre_shutdown(WOLFSSL* ssl)
-{
-    EXPECT_DECLS;
-    WOLFSSL_SESSION** sess;
-    sess = &twcase_client_first_session_ptr;
-    if (*sess == NULL) {
-        ExpectNotNull(*sess = wolfSSL_get1_session(ssl));
-    }
-    else {
-        /* If we have a session retrieved then remaining connections should be
-         * resuming on that session */
-        ExpectIntEQ(wolfSSL_session_reused(ssl), 1);
-    }
-
-#if defined(SESSION_CERTS) && !defined(WOLFSSL_NO_CLIENT_AUTH)
-#ifndef WOLFSSL_TICKET_HAVE_ID
-    if (wolfSSL_version(ssl) != TLS1_3_VERSION &&
-            wolfSSL_session_reused(ssl))
-#endif
-    {
-    #ifdef KEEP_PEER_CERT
-        WOLFSSL_X509* peer = wolfSSL_get_peer_certificate(ssl);
-        ExpectNotNull(peer);
-        wolfSSL_X509_free(peer);
-    #endif
-        ExpectNotNull(wolfSSL_SESSION_get_peer_chain(*sess));
-#ifdef OPENSSL_EXTRA
-        ExpectNotNull(wolfSSL_SESSION_get0_peer(*sess));
-#endif
-    }
-#endif
-    return EXPECT_RESULT();
-}
-static int twcase_client_set_sess_ssl_ready(WOLFSSL* ssl)
-{
-    EXPECT_DECLS;
-    /* Set the session to reuse for the client */
-    ExpectNotNull(ssl);
-    ExpectNotNull(twcase_client_first_session_ptr);
-    ExpectIntEQ(wolfSSL_set_session(ssl,twcase_client_first_session_ptr),
-                WOLFSSL_SUCCESS);
-    return EXPECT_RESULT();
-}
-
-struct test_add_session_ext_params {
-    method_provider client_meth;
-    method_provider server_meth;
-    const char* tls_version;
-};
-
-static int test_wolfSSL_CTX_add_session_ext(
-    struct test_add_session_ext_params* param)
-{
-    EXPECT_DECLS;
-    /* Test the default 33 sessions */
-    int j;
-
-    /* Clear cache before starting */
-    wolfSSL_CTX_flush_sessions(NULL, -1);
-
-    XMEMSET(&server_sessionCache, 0, sizeof(hashTable));
-    if (wc_InitMutex(&server_sessionCache.htLock) != 0)
-        return BAD_MUTEX_E;
-    server_sessionCache.capacity = SESSION_CACHE_SIZE;
-
-    fprintf(stderr, "\tBegin %s\n", param->tls_version);
-    for (j = 0; j < 5; j++) {
-        int tls13 = XSTRSTR(param->tls_version, "TLSv1_3") != NULL;
-        int dtls = XSTRSTR(param->tls_version, "DTLS") != NULL;
-        test_ssl_cbf client_cb;
-        test_ssl_cbf server_cb;
-
-        (void)dtls;
-
-        /* Test five cache configurations */
-        twcase_client_first_session_ptr = NULL;
-        twcase_server_first_session_ptr = NULL;
-        twcase_server_current_ctx_ptr = NULL;
-        twcase_new_session_called    = 0;
-        twcase_remove_session_called = 0;
-        twcase_get_session_called    = 0;
-
-        /* connection 1 - first connection */
-        fprintf(stderr, "\tconnect: %s: j=%d\n", param->tls_version, j);
-
-        XMEMSET(&client_cb, 0, sizeof(client_cb));
-        XMEMSET(&server_cb, 0, sizeof(server_cb));
-        client_cb.method  = param->client_meth;
-        server_cb.method  = param->server_meth;
-
-        if (dtls)
-            client_cb.doUdp = server_cb.doUdp = 1;
-
-        /* Setup internal and external cache */
-        switch (j) {
-            case 0:
-                /* SSL_OP_NO_TICKET stateful ticket case */
-                server_cb.ctx_ready = twcase_cache_intOn_extOn_noTicket;
-                break;
-            case 1:
-                server_cb.ctx_ready = twcase_cache_intOn_extOn;
-                break;
-            case 2:
-                server_cb.ctx_ready = twcase_cache_intOff_extOn;
-                break;
-            case 3:
-                server_cb.ctx_ready = twcase_cache_intOn_extOff;
-                break;
-            case 4:
-                server_cb.ctx_ready = twcase_cache_intOff_extOff;
-                break;
-        }
-        client_cb.ctx_ready = twcase_cache_intOff_extOff;
-
-        /* Add session to internal cache and save SSL session for testing */
-        server_cb.on_result = twcase_server_sess_ctx_pre_shutdown;
-        /* Save client SSL session for testing */
-        client_cb.on_result = twcase_client_sess_ctx_pre_shutdown;
-        server_cb.ticNoInit = 1; /* Use default builtin */
-        /* Don't free/release ctx */
-        server_cb.ctx = twcase_server_current_ctx_ptr;
-        server_cb.isSharedCtx = 1;
-
-        ExpectIntEQ(test_wolfSSL_client_server_nofail_memio(&client_cb,
-            &server_cb, NULL), TEST_SUCCESS);
-
-        ExpectIntEQ(twcase_get_session_called, 0);
-        if (EXPECT_FAIL()) {
-            wolfSSL_SESSION_free(twcase_client_first_session_ptr);
-            wolfSSL_SESSION_free(twcase_server_first_session_ptr);
-            wolfSSL_CTX_free(twcase_server_current_ctx_ptr);
-            break;
-        }
-
-        switch (j) {
-            case 0:
-            case 1:
-            case 2:
-                /* cache cannot be searched with out a connection */
-                /* Add a new session */
-                ExpectIntEQ(twcase_new_session_called, 1);
-                /* In twcase_server_sess_ctx_pre_shutdown
-                 * wolfSSL_CTX_add_session which evicts the existing session
-                 * in cache and adds it back in */
-                ExpectIntLE(twcase_remove_session_called, 1);
-                break;
-            case 3:
-            case 4:
-                /* no external cache  */
-                ExpectIntEQ(twcase_new_session_called, 0);
-                ExpectIntEQ(twcase_remove_session_called, 0);
-                break;
-        }
-
-        /* connection 2 - session resume */
-        fprintf(stderr, "\tresume: %s: j=%d\n", param->tls_version, j);
-        twcase_new_session_called    = 0;
-        twcase_remove_session_called = 0;
-        twcase_get_session_called    = 0;
-        server_cb.on_result = 0;
-        client_cb.on_result = 0;
-        server_cb.ticNoInit = 1; /* Use default builtin */
-
-        server_cb.ctx = twcase_server_current_ctx_ptr;
-
-        /* try session resumption */
-        client_cb.ssl_ready = twcase_client_set_sess_ssl_ready;
-
-        ExpectIntEQ(test_wolfSSL_client_server_nofail_memio(&client_cb,
-            &server_cb, NULL), TEST_SUCCESS);
-
-        /* Clear cache before checking */
-        wolfSSL_CTX_flush_sessions(NULL, -1);
-
-        switch (j) {
-            case 0:
-                if (tls13) {
-                    /* (D)TLSv1.3 stateful case */
-                    /* cache hit */
-                    /* DTLS accesses cache once for stateless parsing and
-                     * once for stateful parsing */
-                    ExpectIntEQ(twcase_get_session_called, !dtls ? 1 : 2);
-
-                    /* (D)TLSv1.3 creates a new ticket,
-                     * updates both internal and external cache */
-                    ExpectIntEQ(twcase_new_session_called, 1);
-                    /* A new session ID is created for a new ticket */
-                    ExpectIntEQ(twcase_remove_session_called, 2);
-
-                }
-                else {
-                    /* non (D)TLSv1.3 case, no update */
-                    /* DTLS accesses cache once for stateless parsing and
-                     * once for stateful parsing */
-#ifdef WOLFSSL_DTLS_NO_HVR_ON_RESUME
-                    ExpectIntEQ(twcase_get_session_called, !dtls ? 1 : 2);
-#else
-                    ExpectIntEQ(twcase_get_session_called, 1);
-#endif
-                    ExpectIntEQ(twcase_new_session_called, 0);
-                    /* Called on session added in
-                     * twcase_server_sess_ctx_pre_shutdown */
-                    ExpectIntEQ(twcase_remove_session_called, 1);
-                }
-                break;
-            case 1:
-                if (tls13) {
-                    /* (D)TLSv1.3 case */
-                    /* cache hit */
-                    ExpectIntEQ(twcase_get_session_called, 1);
-                    /* (D)TLSv1.3 creates a new ticket,
-                     * updates both internal and external cache */
-                    ExpectIntEQ(twcase_new_session_called, 1);
-                    /* Called on session added in
-                     * twcase_server_sess_ctx_pre_shutdown and by wolfSSL */
-                    ExpectIntEQ(twcase_remove_session_called, 1);
-                }
-                else {
-                    /* non (D)TLSv1.3 case */
-                    /* cache hit */
-                    /* DTLS accesses cache once for stateless parsing and
-                     * once for stateful parsing */
-#ifdef WOLFSSL_DTLS_NO_HVR_ON_RESUME
-                    ExpectIntEQ(twcase_get_session_called, !dtls ? 1 : 2);
-#else
-                    ExpectIntEQ(twcase_get_session_called, 1);
-#endif
-                    ExpectIntEQ(twcase_new_session_called, 0);
-                    /* Called on session added in
-                     * twcase_server_sess_ctx_pre_shutdown */
-                    ExpectIntEQ(twcase_remove_session_called, 1);
-                }
-                break;
-            case 2:
-                if (tls13) {
-                    /* (D)TLSv1.3 case */
-                    /* cache hit */
-                    ExpectIntEQ(twcase_get_session_called, 1);
-                    /* (D)TLSv1.3 creates a new ticket,
-                     * updates both internal and external cache */
-                    ExpectIntEQ(twcase_new_session_called, 1);
-                    /* Called on session added in
-                     * twcase_server_sess_ctx_pre_shutdown and by wolfSSL */
-                    ExpectIntEQ(twcase_remove_session_called, 1);
-                }
-                else {
-                    /* non (D)TLSv1.3 case */
-                    /* cache hit */
-                    /* DTLS accesses cache once for stateless parsing and
-                     * once for stateful parsing */
-#ifdef WOLFSSL_DTLS_NO_HVR_ON_RESUME
-                    ExpectIntEQ(twcase_get_session_called, !dtls ? 1 : 2);
-#else
-                    ExpectIntEQ(twcase_get_session_called, 1);
-#endif
-                    ExpectIntEQ(twcase_new_session_called, 0);
-                    /* Called on session added in
-                     * twcase_server_sess_ctx_pre_shutdown */
-                    ExpectIntEQ(twcase_remove_session_called, 1);
-                }
-                break;
-            case 3:
-            case 4:
-                /* no external cache */
-                ExpectIntEQ(twcase_get_session_called, 0);
-                ExpectIntEQ(twcase_new_session_called, 0);
-                ExpectIntEQ(twcase_remove_session_called, 0);
-                break;
-        }
-        wolfSSL_SESSION_free(twcase_client_first_session_ptr);
-        wolfSSL_SESSION_free(twcase_server_first_session_ptr);
-        wolfSSL_CTX_free(twcase_server_current_ctx_ptr);
-
-        if (EXPECT_FAIL())
-            break;
-    }
-    twcase_get_sessionCb_cleanup();
-    XMEMSET(&server_sessionCache.entries, 0,
-            sizeof(server_sessionCache.entries));
-    fprintf(stderr, "\tEnd %s\n", param->tls_version);
-
-    wc_FreeMutex(&server_sessionCache.htLock);
+    /* A sufficiently large buffer must still succeed and report the actual
+     * Finished length, which is between TLS_FINISHED_SZ and the max digest
+     * size. */
+    XMEMSET(full, 0, sizeof(full));
+    msg_len = wolfSSL_get_finished(ssl, full, sizeof(full));
+    ExpectIntGE(msg_len, TLS_FINISHED_SZ);
+    ExpectIntLE(msg_len, WC_MAX_DIGEST_SIZE);
 
     return EXPECT_RESULT();
 }
-#endif
+#endif /* HAVE_SSL_MEMIO_TESTS_DEPENDENCIES && WOLFSSL_HAVE_TLS_UNIQUE &&
+        * WOLFSSL_TLS13 */
 
-static int test_wolfSSL_CTX_add_session_ext_tls13(void)
+static int test_wolfSSL_get_finished_overrun(void)
 {
     EXPECT_DECLS;
-#if defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_EXT_CACHE) && \
-     defined(WOLFSSL_TLS13) && !defined(NO_SESSION_CACHE) && \
-     defined(OPENSSL_EXTRA) && defined(SESSION_CERTS) && \
-     defined(HAVE_SESSION_TICKET) && \
-    !defined(TITAN_SESSION_CACHE) && \
-    !defined(HUGE_SESSION_CACHE) && \
-    !defined(BIG_SESSION_CACHE) && \
-    !defined(MEDIUM_SESSION_CACHE)
-#if defined(WOLFSSL_TLS13) && !defined(WOLFSSL_NO_DEF_TICKET_ENC_CB) && \
-    defined(HAVE_SESSION_TICKET) && defined(WOLFSSL_TICKET_HAVE_ID)
-    struct test_add_session_ext_params param[1] =  {
-        { wolfTLSv1_3_client_method, wolfTLSv1_3_server_method, "TLSv1_3" }
-    };
-    ExpectIntEQ(test_wolfSSL_CTX_add_session_ext(param), TEST_SUCCESS);
-#endif
-#endif
-    return EXPECT_RESULT();
-}
-static int test_wolfSSL_CTX_add_session_ext_dtls13(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_EXT_CACHE) && \
-     defined(WOLFSSL_TLS13) && !defined(NO_SESSION_CACHE) && \
-     defined(OPENSSL_EXTRA) && defined(SESSION_CERTS) && \
-     defined(HAVE_SESSION_TICKET) && \
-    !defined(TITAN_SESSION_CACHE) && \
-    !defined(HUGE_SESSION_CACHE) && \
-    !defined(BIG_SESSION_CACHE) && \
-    !defined(MEDIUM_SESSION_CACHE)
-#if defined(WOLFSSL_TLS13) && !defined(WOLFSSL_NO_DEF_TICKET_ENC_CB) && \
-    defined(HAVE_SESSION_TICKET) && defined(WOLFSSL_TICKET_HAVE_ID)
-#ifdef WOLFSSL_DTLS13
-    struct test_add_session_ext_params param[1] =  {
-        { wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method, "DTLSv1_3" }
-    };
-    ExpectIntEQ(test_wolfSSL_CTX_add_session_ext(param), TEST_SUCCESS);
-#endif
-#endif
-#endif
-    return EXPECT_RESULT();
-}
-static int test_wolfSSL_CTX_add_session_ext_tls12(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_EXT_CACHE) && \
-     defined(WOLFSSL_TLS13) && !defined(NO_SESSION_CACHE) && \
-     defined(OPENSSL_EXTRA) && defined(SESSION_CERTS) && \
-     defined(HAVE_SESSION_TICKET) && \
-    !defined(TITAN_SESSION_CACHE) && \
-    !defined(HUGE_SESSION_CACHE) && \
-    !defined(BIG_SESSION_CACHE) && \
-    !defined(MEDIUM_SESSION_CACHE)
-#ifndef WOLFSSL_NO_TLS12
-    struct test_add_session_ext_params param[1] =  {
-        { wolfTLSv1_2_client_method, wolfTLSv1_2_server_method, "TLSv1_2" }
-    };
-    ExpectIntEQ(test_wolfSSL_CTX_add_session_ext(param), TEST_SUCCESS);
-#endif
-#endif
-    return EXPECT_RESULT();
-}
-static int test_wolfSSL_CTX_add_session_ext_dtls12(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_EXT_CACHE) && \
-     defined(WOLFSSL_TLS13) && !defined(NO_SESSION_CACHE) && \
-     defined(OPENSSL_EXTRA) && defined(SESSION_CERTS) && \
-     defined(HAVE_SESSION_TICKET) && \
-    !defined(TITAN_SESSION_CACHE) && \
-    !defined(HUGE_SESSION_CACHE) && \
-    !defined(BIG_SESSION_CACHE) && \
-    !defined(MEDIUM_SESSION_CACHE)
-#ifndef WOLFSSL_NO_TLS12
-#ifdef WOLFSSL_DTLS
-    struct test_add_session_ext_params param[1] =  {
-        { wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method, "DTLSv1_2" }
-    };
-    ExpectIntEQ(test_wolfSSL_CTX_add_session_ext(param), TEST_SUCCESS);
-#endif
-#endif
-#endif
-    return EXPECT_RESULT();
-}
-static int test_wolfSSL_CTX_add_session_ext_tls11(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_EXT_CACHE) && \
-     defined(WOLFSSL_TLS13) && !defined(NO_SESSION_CACHE) && \
-     defined(OPENSSL_EXTRA) && defined(SESSION_CERTS) && \
-     defined(HAVE_SESSION_TICKET) && \
-    !defined(TITAN_SESSION_CACHE) && \
-    !defined(HUGE_SESSION_CACHE) && \
-    !defined(BIG_SESSION_CACHE) && \
-    !defined(MEDIUM_SESSION_CACHE)
-#if !defined(NO_OLD_TLS) && ((!defined(NO_AES) && !defined(NO_AES_CBC)) || \
-        !defined(NO_DES3))
-    struct test_add_session_ext_params param[1] =  {
-        { wolfTLSv1_1_client_method, wolfTLSv1_1_server_method, "TLSv1_1" }
-    };
-    ExpectIntEQ(test_wolfSSL_CTX_add_session_ext(param), TEST_SUCCESS);
-#endif
-#endif
-    return EXPECT_RESULT();
-}
-static int test_wolfSSL_CTX_add_session_ext_dtls1(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_EXT_CACHE) && \
-     defined(WOLFSSL_TLS13) && !defined(NO_SESSION_CACHE) && \
-     defined(OPENSSL_EXTRA) && defined(SESSION_CERTS) && \
-     defined(HAVE_SESSION_TICKET) && \
-    !defined(TITAN_SESSION_CACHE) && \
-    !defined(HUGE_SESSION_CACHE) && \
-    !defined(BIG_SESSION_CACHE) && \
-    !defined(MEDIUM_SESSION_CACHE)
-#if !defined(NO_OLD_TLS) && ((!defined(NO_AES) && !defined(NO_AES_CBC)) || \
-        !defined(NO_DES3))
-#ifdef WOLFSSL_DTLS
-    struct test_add_session_ext_params param[1] =  {
-        { wolfDTLSv1_client_method, wolfDTLSv1_server_method, "DTLSv1_0" }
-    };
-    ExpectIntEQ(test_wolfSSL_CTX_add_session_ext(param), TEST_SUCCESS);
-#endif
-#endif
-#endif
+#if defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES) && \
+    defined(WOLFSSL_HAVE_TLS_UNIQUE) && defined(WOLFSSL_TLS13)
+    test_ssl_cbf client_cbf;
+    test_ssl_cbf server_cbf;
+
+    XMEMSET(&client_cbf, 0, sizeof(client_cbf));
+    XMEMSET(&server_cbf, 0, sizeof(server_cbf));
+
+    /* Force TLS 1.3 so the stored Finished length exceeds TLS_FINISHED_SZ. */
+    client_cbf.method = wolfTLSv1_3_client_method;
+    server_cbf.method = wolfTLSv1_3_server_method;
+
+    ExpectIntEQ(test_wolfSSL_client_server_nofail_memio(&client_cbf,
+        &server_cbf, test_wolfSSL_get_finished_overrun_on_handshake),
+        TEST_SUCCESS);
+#endif /* HAVE_SSL_MEMIO_TESTS_DEPENDENCIES && WOLFSSL_HAVE_TLS_UNIQUE &&
+        * WOLFSSL_TLS13 */
+
     return EXPECT_RESULT();
 }
 
-#if defined(WOLFSSL_DTLS) && defined(WOLFSSL_SESSION_EXPORT)
-/* canned export of a session using older version 3 */
-static unsigned char version_3[] = {
-    0xA5, 0xA3, 0x01, 0x88, 0x00, 0x3c, 0x00, 0x01,
-    0x00, 0x00, 0x00, 0x80, 0x0C, 0x00, 0x00, 0x00,
-    0x00, 0x80, 0x00, 0x1C, 0x00, 0x00, 0x00, 0x00,
-    0x00, 0x01, 0x01, 0x01, 0x01, 0x00, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00,
-    0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xC0, 0x30,
-    0x05, 0x09, 0x0A, 0x01, 0x01, 0x00, 0x0D, 0x05,
-    0xFE, 0xFD, 0x01, 0x25, 0x00, 0x00, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x01, 0x00, 0x01, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-    0x00, 0x06, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00,
-    0x00, 0x06, 0x00, 0x01, 0x00, 0x07, 0x00, 0x00,
-    0x00, 0x30, 0x00, 0x00, 0x00, 0x10, 0x01, 0x01,
-    0x00, 0x02, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x3F,
-    0x00, 0x00, 0x00, 0x00, 0x00, 0x30, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x20, 0x05,
-    0x12, 0xCF, 0x22, 0xA1, 0x9F, 0x1C, 0x39, 0x1D,
-    0x31, 0x11, 0x12, 0x1D, 0x11, 0x18, 0x0D, 0x0B,
-    0xF3, 0xE1, 0x4D, 0xDC, 0xB1, 0xF1, 0x39, 0x98,
-    0x91, 0x6C, 0x48, 0xE5, 0xED, 0x11, 0x12, 0xA0,
-    0x00, 0xF2, 0x25, 0x4C, 0x09, 0x26, 0xD1, 0x74,
-    0xDF, 0x23, 0x40, 0x15, 0x6A, 0x42, 0x2A, 0x26,
-    0xA5, 0xAC, 0x56, 0xD5, 0x4A, 0x20, 0xB7, 0xE9,
-    0xEF, 0xEB, 0xAF, 0xA8, 0x1E, 0x23, 0x7C, 0x04,
-    0xAA, 0xA1, 0x6D, 0x92, 0x79, 0x7B, 0xFA, 0x80,
-    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01,
-    0x0C, 0x79, 0x7B, 0xFA, 0x80, 0x00, 0x00, 0x00,
-    0x00, 0x00, 0x00, 0x00, 0x00, 0xAA, 0xA1, 0x6D,
-    0x92, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-    0x00, 0x00, 0x10, 0x00, 0x20, 0x00, 0x04, 0x00,
-    0x10, 0x00, 0x10, 0x08, 0x02, 0x05, 0x08, 0x01,
-    0x30, 0x28, 0x00, 0x00, 0x0F, 0x00, 0x02, 0x00,
-    0x09, 0x31, 0x32, 0x37, 0x2E, 0x30, 0x2E, 0x30,
-    0x2E, 0x31, 0xED, 0x4F
-};
-#endif /* defined(WOLFSSL_DTLS) && defined(WOLFSSL_SESSION_EXPORT) */
-
-static int test_wolfSSL_dtls_export(void)
+#if defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES) && \
+    defined(WOLFSSL_HAVE_TLS_UNIQUE) && defined(WOLFSSL_TLS13)
+static int test_wolfSSL_get_peer_finished_overrun_on_handshake(WOLFSSL_CTX* ctx,
+    WOLFSSL* ssl)
 {
     EXPECT_DECLS;
-#if defined(WOLFSSL_DTLS) && defined(WOLFSSL_SESSION_EXPORT)
-    tcp_ready ready;
-    func_args client_args;
-    func_args server_args;
-    THREAD_TYPE serverThread;
-    callback_functions server_cbf;
-    callback_functions client_cbf;
-#ifdef WOLFSSL_TIRTOS
-    fdOpenSession(Task_self());
-#endif
+    /* Destination buffer immediately followed by a guard region. Both are
+     * byte arrays (alignment 1) so there is no padding between them and the
+     * guard bytes sit directly after buf. The caller promises only
+     * TLS_FINISHED_SZ writable bytes via the count argument, so wolfSSL must
+     * not write past buf[] into the guard region. */
+    struct {
+        byte buf[TLS_FINISHED_SZ];
+        byte guard[WC_MAX_DIGEST_SIZE];
+    } probe;
+    byte   full[WC_MAX_DIGEST_SIZE];
+    size_t msg_len;
+    word32 i;
 
-    InitTcpReady(&ready);
+    (void)ctx;
 
-    /* set using dtls */
-    XMEMSET(&client_args, 0, sizeof(func_args));
-    XMEMSET(&server_args, 0, sizeof(func_args));
-    XMEMSET(&server_cbf, 0, sizeof(callback_functions));
-    XMEMSET(&client_cbf, 0, sizeof(callback_functions));
-    server_cbf.method = wolfDTLSv1_2_server_method;
-    client_cbf.method = wolfDTLSv1_2_client_method;
-    server_args.callbacks = &server_cbf;
-    client_args.callbacks = &client_cbf;
+    XMEMSET(&probe, 0, sizeof(probe));
+    XMEMSET(probe.guard, 0x5A, sizeof(probe.guard));
 
-    server_args.signal = &ready;
-    client_args.signal = &ready;
+    /* For a TLS 1.3 connection the peer Finished message is the handshake
+     * hash size (32 bytes for SHA-256, 48 for SHA-384), which is larger than
+     * the count of TLS_FINISHED_SZ (12) we pass here. The buffer is too small
+     * to hold the Finished message, so the call must fail and write nothing
+     * rather than copying the full length and overrunning buf[]. */
+    msg_len = wolfSSL_get_peer_finished(ssl, probe.buf, TLS_FINISHED_SZ);
+    ExpectIntEQ(msg_len, WOLFSSL_FAILURE);
 
-    start_thread(run_wolfssl_server, &server_args, &serverThread);
-    wait_tcp_ready(&server_args);
-    run_wolfssl_client(&client_args);
-    join_thread(serverThread);
-
-    ExpectTrue(client_args.return_code);
-    ExpectTrue(server_args.return_code);
-
-    FreeTcpReady(&ready);
-
-#ifdef WOLFSSL_TIRTOS
-    fdOpenSession(Task_self());
-#endif
-
-    if (EXPECT_SUCCESS()) {
-        SOCKET_T sockfd = 0;
-        WOLFSSL_CTX* ctx = NULL;
-        WOLFSSL*     ssl = NULL;
-        char msg[64] = "hello wolfssl!";
-        char reply[1024];
-        int  msgSz = (int)XSTRLEN(msg);
-        byte *session, *window;
-        unsigned int sessionSz = 0;
-        unsigned int windowSz = 0;
-
-#ifndef TEST_IPV6
-        struct sockaddr_in peerAddr;
-#else
-        struct sockaddr_in6 peerAddr;
-#endif /* TEST_IPV6 */
-
-        int i;
-
-
-        /* Set ctx to DTLS 1.2 */
-        ExpectNotNull(ctx = wolfSSL_CTX_new(wolfDTLSv1_2_server_method()));
-        ExpectNotNull(ssl = wolfSSL_new(ctx));
-
-        /* test importing version 3 */
-        ExpectIntGE(wolfSSL_dtls_import(ssl, version_3, sizeof(version_3)), 0);
-
-        /* test importing bad length and bad version */
-        version_3[2]++;
-        ExpectIntLT(wolfSSL_dtls_import(ssl, version_3, sizeof(version_3)), 0);
-        version_3[2]--; version_3[1] = 0XA0;
-        ExpectIntLT(wolfSSL_dtls_import(ssl, version_3, sizeof(version_3)), 0);
-        wolfSSL_free(ssl);
-        wolfSSL_CTX_free(ctx);
-
-
-    /* check storing client state after connection and storing window only */
-#ifdef WOLFSSL_TIRTOS
-    fdOpenSession(Task_self());
-#endif
-
-    InitTcpReady(&ready);
-
-    /* set using dtls */
-    XMEMSET(&server_args, 0, sizeof(func_args));
-    XMEMSET(&server_cbf, 0, sizeof(callback_functions));
-    server_cbf.method = wolfDTLSv1_2_server_method;
-    server_cbf.doUdp = 1;
-    server_args.callbacks = &server_cbf;
-    server_args.argc = 3; /* set loop_count to 3 */
-
-
-    server_args.signal = &ready;
-    start_thread(test_server_nofail, &server_args, &serverThread);
-    wait_tcp_ready(&server_args);
-
-    /* create and connect with client */
-    ExpectNotNull(ctx = wolfSSL_CTX_new(wolfDTLSv1_2_client_method()));
-    ExpectIntEQ(WOLFSSL_SUCCESS,
-            wolfSSL_CTX_load_verify_locations(ctx, caCertFile, 0));
-    ExpectIntEQ(WOLFSSL_SUCCESS,
-          wolfSSL_CTX_use_certificate_file(ctx, cliCertFile, SSL_FILETYPE_PEM));
-    ExpectIntEQ(WOLFSSL_SUCCESS,
-            wolfSSL_CTX_use_PrivateKey_file(ctx, cliKeyFile, SSL_FILETYPE_PEM));
-    tcp_connect(&sockfd, wolfSSLIP, server_args.signal->port, 1, 0, NULL);
-    ExpectNotNull(ssl = wolfSSL_new(ctx));
-    ExpectIntEQ(wolfSSL_set_fd(ssl, sockfd), WOLFSSL_SUCCESS);
-
-    /* store server information connected too */
-    XMEMSET(&peerAddr, 0, sizeof(peerAddr));
-#ifndef TEST_IPV6
-    peerAddr.sin_family = AF_INET;
-    ExpectIntEQ(XINET_PTON(AF_INET, wolfSSLIP, &peerAddr.sin_addr),1);
-    peerAddr.sin_port = XHTONS(server_args.signal->port);
-#else
-    peerAddr.sin6_family = AF_INET6;
-    ExpectIntEQ(
-        XINET_PTON(AF_INET6, wolfSSLIP, &peerAddr.sin6_addr),1);
-    peerAddr.sin6_port = XHTONS(server_args.signal->port);
-#endif
-
-    ExpectIntEQ(wolfSSL_dtls_set_peer(ssl, &peerAddr, sizeof(peerAddr)),
-                    WOLFSSL_SUCCESS);
-
-    ExpectIntEQ(wolfSSL_connect(ssl), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_dtls_export(ssl, NULL, &sessionSz), 0);
-    session = (byte*)XMALLOC(sessionSz, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
-    ExpectIntGT(wolfSSL_dtls_export(ssl, session, &sessionSz), 0);
-    ExpectIntEQ(wolfSSL_write(ssl, msg, msgSz), msgSz);
-    ExpectIntGT(wolfSSL_read(ssl, reply, sizeof(reply)), 0);
-    ExpectIntEQ(wolfSSL_dtls_export_state_only(ssl, NULL, &windowSz), 0);
-    window = (byte*)XMALLOC(windowSz, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
-    ExpectIntGT(wolfSSL_dtls_export_state_only(ssl, window, &windowSz), 0);
-    wolfSSL_free(ssl);
-
-    for (i = 1; EXPECT_SUCCESS() && i < server_args.argc; i++) {
-        /* restore state */
-        ExpectNotNull(ssl = wolfSSL_new(ctx));
-        ExpectIntGT(wolfSSL_dtls_import(ssl, session, sessionSz), 0);
-        ExpectIntGT(wolfSSL_dtls_import(ssl, window, windowSz), 0);
-        ExpectIntEQ(wolfSSL_set_fd(ssl, sockfd), WOLFSSL_SUCCESS);
-        ExpectIntEQ(wolfSSL_dtls_set_peer(ssl, &peerAddr, sizeof(peerAddr)),
-                    WOLFSSL_SUCCESS);
-        ExpectIntEQ(wolfSSL_write(ssl, msg, msgSz), msgSz);
-        ExpectIntGE(wolfSSL_read(ssl, reply, sizeof(reply)), 0);
-        ExpectIntGT(wolfSSL_dtls_export_state_only(ssl, window, &windowSz), 0);
-        wolfSSL_free(ssl);
+    /* Nothing must have been written: buf[] stays zero and the guard bytes
+     * that follow it stay 0x5A. */
+    for (i = 0; i < (word32)sizeof(probe.buf); i++) {
+        ExpectIntEQ(probe.buf[i], 0x00);
     }
-    XFREE(session, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
-    XFREE(window, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
-    wolfSSL_CTX_free(ctx);
-
-    fprintf(stderr, "done and waiting for server\n");
-    join_thread(serverThread);
-    ExpectIntEQ(server_args.return_code, TEST_SUCCESS);
-
-    FreeTcpReady(&ready);
-
-#ifdef WOLFSSL_TIRTOS
-    fdOpenSession(Task_self());
-#endif
+    for (i = 0; i < (word32)sizeof(probe.guard); i++) {
+        ExpectIntEQ(probe.guard[i], 0x5A);
     }
-#endif
+
+    /* A sufficiently large buffer must still succeed and report the actual
+     * Finished length, which is between TLS_FINISHED_SZ and the max digest
+     * size. */
+    XMEMSET(full, 0, sizeof(full));
+    msg_len = wolfSSL_get_peer_finished(ssl, full, sizeof(full));
+    ExpectIntGE(msg_len, TLS_FINISHED_SZ);
+    ExpectIntLE(msg_len, WC_MAX_DIGEST_SIZE);
 
     return EXPECT_RESULT();
 }
+#endif /* HAVE_SSL_MEMIO_TESTS_DEPENDENCIES && WOLFSSL_HAVE_TLS_UNIQUE &&
+        * WOLFSSL_TLS13 */
+
+static int test_wolfSSL_get_peer_finished_overrun(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES) && \
+    defined(WOLFSSL_HAVE_TLS_UNIQUE) && defined(WOLFSSL_TLS13)
+    test_ssl_cbf client_cbf;
+    test_ssl_cbf server_cbf;
+
+    XMEMSET(&client_cbf, 0, sizeof(client_cbf));
+    XMEMSET(&server_cbf, 0, sizeof(server_cbf));
+
+    /* Force TLS 1.3 so the stored Finished length exceeds TLS_FINISHED_SZ. */
+    client_cbf.method = wolfTLSv1_3_client_method;
+    server_cbf.method = wolfTLSv1_3_server_method;
+
+    ExpectIntEQ(test_wolfSSL_client_server_nofail_memio(&client_cbf,
+        &server_cbf, test_wolfSSL_get_peer_finished_overrun_on_handshake),
+        TEST_SUCCESS);
+#endif /* HAVE_SSL_MEMIO_TESTS_DEPENDENCIES && WOLFSSL_HAVE_TLS_UNIQUE &&
+        * WOLFSSL_TLS13 */
+
+    return EXPECT_RESULT();
+}
+
 
 #if defined(WOLFSSL_SESSION_EXPORT) && !defined(WOLFSSL_NO_TLS12)
 #ifdef WOLFSSL_TLS13
@@ -9353,223 +8366,6 @@ static int test_wolfSSL_SNI_GetFromBuffer(void)
 #endif /* HAVE_IO_TESTS_DEPENDENCIES */
 
 
-#if defined(WOLFSSL_DTLS) && defined(WOLFSSL_SESSION_EXPORT) && \
-    defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES)
-/* Dummy peer functions to satisfy the exporter/importer */
-static int test_wolfSSL_dtls_export_peers_get_peer(WOLFSSL* ssl, char* ip,
-        int* ipSz, unsigned short* port, int* fam)
-{
-    (void)ssl;
-    ip[0] = -1;
-    *ipSz = 1;
-    *port = 1;
-    *fam = 2;
-    return 1;
-}
-
-static int test_wolfSSL_dtls_export_peers_set_peer(WOLFSSL* ssl, char* ip,
-        int ipSz, unsigned short port, int fam)
-{
-    (void)ssl;
-    if (ip[0] != -1 || ipSz != 1 || port != 1 || fam != 2)
-        return 0;
-    return 1;
-}
-
-static int test_wolfSSL_dtls_export_peers_on_handshake(WOLFSSL_CTX **ctx,
-        WOLFSSL **ssl)
-{
-    EXPECT_DECLS;
-    unsigned char* sessionBuf = NULL;
-    unsigned int sessionSz = 0;
-    void* ioWriteCtx = wolfSSL_GetIOWriteCtx(*ssl);
-    void* ioReadCtx = wolfSSL_GetIOReadCtx(*ssl);
-
-    wolfSSL_CTX_SetIOGetPeer(*ctx, test_wolfSSL_dtls_export_peers_get_peer);
-    wolfSSL_CTX_SetIOSetPeer(*ctx, test_wolfSSL_dtls_export_peers_set_peer);
-    ExpectIntGE(wolfSSL_dtls_export(*ssl, NULL, &sessionSz), 0);
-    ExpectNotNull(sessionBuf =
-        (unsigned char*)XMALLOC(sessionSz, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER));
-    ExpectIntGE(wolfSSL_dtls_export(*ssl, sessionBuf, &sessionSz), 0);
-    wolfSSL_free(*ssl);
-    *ssl = NULL;
-    ExpectNotNull(*ssl = wolfSSL_new(*ctx));
-    ExpectIntGE(wolfSSL_dtls_import(*ssl, sessionBuf, sessionSz), 0);
-    wolfSSL_SetIOWriteCtx(*ssl, ioWriteCtx);
-    wolfSSL_SetIOReadCtx(*ssl, ioReadCtx);
-
-    XFREE(sessionBuf, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
-    return EXPECT_RESULT();
-}
-#endif
-
-static int test_wolfSSL_dtls_export_peers(void)
-{
-    EXPECT_DECLS;
-#if defined(WOLFSSL_DTLS) && defined(WOLFSSL_SESSION_EXPORT) && \
-    defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES)
-    test_ssl_cbf client_cbf;
-    test_ssl_cbf server_cbf;
-    size_t i, j;
-    struct test_params {
-        method_provider client_meth;
-        method_provider server_meth;
-        const char* dtls_version;
-    } params[] = {
-#ifndef NO_OLD_TLS
-        {wolfDTLSv1_client_method, wolfDTLSv1_server_method, "1.0"},
-#endif
-        {wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method, "1.2"},
-        /* TODO DTLS 1.3 exporting not supported
-#ifdef WOLFSSL_DTLS13
-        {wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method, "1.3"},
-#endif
-         */
-    };
-
-    for (i = 0; i < sizeof(params)/sizeof(*params); i++) {
-        for (j = 0; j <= 3; j++) {
-            XMEMSET(&client_cbf, 0, sizeof(client_cbf));
-            XMEMSET(&server_cbf, 0, sizeof(server_cbf));
-
-            printf("\n\tTesting DTLS %s connection;", params[i].dtls_version);
-
-            client_cbf.method = params[i].client_meth;
-            server_cbf.method = params[i].server_meth;
-
-            if (j & 0x1) {
-                client_cbf.on_handshake =
-                        test_wolfSSL_dtls_export_peers_on_handshake;
-                printf(" With client export;");
-            }
-            if (j & 0x2) {
-                server_cbf.on_handshake =
-                        test_wolfSSL_dtls_export_peers_on_handshake;
-                printf(" With server export;");
-            }
-
-            printf("\n");
-
-            ExpectIntEQ(test_wolfSSL_client_server_nofail_memio(&client_cbf,
-                &server_cbf, NULL), TEST_SUCCESS);
-            if (!EXPECT_SUCCESS())
-                break;
-        }
-    }
-#endif
-    return EXPECT_RESULT();
-}
-
-/* Test that ImportKeyState correctly skips extra window words when importing
- * state from a peer compiled with a larger WOLFSSL_DTLS_WINDOW_WORDS. */
-static int test_wolfSSL_dtls_import_state_extra_window_words(void)
-{
-    EXPECT_DECLS;
-#if defined(WOLFSSL_DTLS) && defined(WOLFSSL_SESSION_EXPORT)
-    WOLFSSL_CTX* ctx = NULL;
-    WOLFSSL*     ssl = NULL;
-    unsigned int stateSz = 0;
-    byte*        state = NULL;
-    byte*        modified = NULL;
-    unsigned int modifiedSz;
-    word16       origKeyLen;
-    word16       origTotalLen;
-    /* Offset from start of key state data to the first wordCount field.
-     * Layout: 4 sequence numbers (16 bytes) + DTLS-specific fields (42 bytes) +
-     * encryptSz(4) + padSz(4) + encryptionOn(1) + decryptedCur(1) = 68 */
-    const int keyStateWindowOffset = 68;
-    /* Buffer header: 2 proto + 2 total_len + 2 key_len = 6 */
-    const int headerSz = 6;
-    int idx, modIdx;
-    int extraPerWindow = 2 * (int)sizeof(word32); /* 8 bytes extra per window */
-    int totalExtra = extraPerWindow * 2; /* 16 bytes extra total */
-
-    /* Create DTLS context and SSL object */
-    ExpectNotNull(ctx = wolfSSL_CTX_new(wolfDTLSv1_2_server_method()));
-    ExpectNotNull(ssl = wolfSSL_new(ctx));
-
-    /* Get required buffer size and export state-only */
-    ExpectIntEQ(wolfSSL_dtls_export_state_only(ssl, NULL, &stateSz), 0);
-    ExpectIntGT((int)stateSz, 0);
-    state = (byte*)XMALLOC(stateSz, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
-    ExpectNotNull(state);
-    ExpectIntGT(wolfSSL_dtls_export_state_only(ssl, state, &stateSz), 0);
-
-    /* Build a modified buffer that simulates a peer with
-     * WOLFSSL_DTLS_WINDOW_WORDS = WOLFSSL_DTLS_WINDOW_WORDS + 2.
-     * Each window section gets 2 extra word32 values (8 bytes).
-     * Two windows => 16 extra bytes total. */
-    modifiedSz = stateSz + totalExtra;
-    modified = (byte*)XMALLOC(modifiedSz, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
-    ExpectNotNull(modified);
-
-    if (EXPECT_SUCCESS()) {
-        int windowWords = WOLFSSL_DTLS_WINDOW_WORDS;
-        int windowDataSz = windowWords * (int)sizeof(word32);
-
-        XMEMSET(modified, 0, modifiedSz);
-
-        /* Copy protocol/version bytes (first 2 bytes) */
-        XMEMCPY(modified, state, 2);
-
-        /* Read original total length and key state length */
-        ato16(state + 2, &origTotalLen);
-        ato16(state + 4, &origKeyLen);
-
-        /* Write updated total length and key state length */
-        c16toa((word16)(origTotalLen + totalExtra), modified + 2);
-        c16toa((word16)(origKeyLen + totalExtra), modified + 4);
-
-        /* Copy key state data up to first window section */
-        idx = headerSz;
-        modIdx = headerSz;
-        XMEMCPY(modified + modIdx, state + idx, keyStateWindowOffset);
-        idx += keyStateWindowOffset;
-        modIdx += keyStateWindowOffset;
-
-        /* First window: write increased wordCount */
-        c16toa((word16)(windowWords + 2), modified + modIdx);
-        idx += OPAQUE16_LEN;
-        modIdx += OPAQUE16_LEN;
-
-        /* Copy original window data */
-        XMEMCPY(modified + modIdx, state + idx, windowDataSz);
-        idx += windowDataSz;
-        modIdx += windowDataSz;
-
-        /* Insert 2 extra word32 padding values */
-        XMEMSET(modified + modIdx, 0, extraPerWindow);
-        modIdx += extraPerWindow;
-
-        /* Second window (prevWindow): same transformation */
-        c16toa((word16)(windowWords + 2), modified + modIdx);
-        idx += OPAQUE16_LEN;
-        modIdx += OPAQUE16_LEN;
-
-        XMEMCPY(modified + modIdx, state + idx, windowDataSz);
-        idx += windowDataSz;
-        modIdx += windowDataSz;
-
-        XMEMSET(modified + modIdx, 0, extraPerWindow);
-        modIdx += extraPerWindow;
-
-        /* Copy remainder of key state (after both windows) */
-        XMEMCPY(modified + modIdx, state + idx, stateSz - idx);
-    }
-
-    /* Import the modified state - should succeed with the fix */
-    wolfSSL_free(ssl);
-    ssl = NULL;
-    ExpectNotNull(ssl = wolfSSL_new(ctx));
-    ExpectIntGT(wolfSSL_dtls_import(ssl, modified, modifiedSz), 0);
-
-    XFREE(state, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
-    XFREE(modified, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
-    wolfSSL_free(ssl);
-    wolfSSL_CTX_free(ctx);
-#endif
-    return EXPECT_RESULT();
-}
 
 static int test_wolfSSL_UseTrustedCA(void)
 {
@@ -11469,8 +10265,7 @@ static int test_wolfSSL_URI(void)
     wolfSSL_FreeX509(x509);
     x509 = NULL;
 
-#if !defined(IGNORE_NAME_CONSTRAINTS) && !defined(WOLFSSL_NO_ASN_STRICT) \
-    && !defined(WOLFSSL_FPKI)
+#if !defined(IGNORE_NAME_CONSTRAINTS) && !defined(WOLFSSL_NO_ASN_STRICT)
     ExpectNull(x509 = wolfSSL_X509_load_certificate_file(badUri,
         WOLFSSL_FILETYPE_PEM));
 #else
@@ -17050,61 +15845,6 @@ static int test_wolfSSL_either_side(void)
     return EXPECT_RESULT();
 }
 
-static int test_wolfSSL_DTLS_either_side(void)
-{
-    EXPECT_DECLS;
-#if (defined(OPENSSL_EXTRA) || defined(WOLFSSL_EITHER_SIDE)) && \
-    defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS)
-    test_ssl_cbf client_cb;
-    test_ssl_cbf server_cb;
-
-    XMEMSET(&client_cb, 0, sizeof(client_cb));
-    XMEMSET(&server_cb, 0, sizeof(server_cb));
-
-    /* Use different CTX for client and server */
-    client_cb.ctx = wolfSSL_CTX_new(wolfDTLS_method());
-    ExpectNotNull(client_cb.ctx);
-    server_cb.ctx = wolfSSL_CTX_new(wolfDTLS_method());
-    ExpectNotNull(server_cb.ctx);
-    /* we are responsible for free'ing WOLFSSL_CTX */
-    server_cb.isSharedCtx = client_cb.isSharedCtx = 1;
-
-    ExpectIntEQ(test_wolfSSL_client_server_nofail_memio(&client_cb,
-        &server_cb, NULL), TEST_SUCCESS);
-
-    wolfSSL_CTX_free(client_cb.ctx);
-    wolfSSL_CTX_free(server_cb.ctx);
-#endif
-    return EXPECT_RESULT();
-}
-
-static int test_generate_cookie(void)
-{
-    EXPECT_DECLS;
-#if defined(WOLFSSL_DTLS) && defined(OPENSSL_EXTRA) && defined(USE_WOLFSSL_IO)
-    SSL_CTX* ctx = NULL;
-    SSL* ssl = NULL;
-    byte    buf[FOURK_BUF] = {0};
-
-    ExpectNotNull(ctx = wolfSSL_CTX_new(wolfDTLS_method()));
-    ExpectNotNull(ssl = SSL_new(ctx));
-
-    /* Test unconnected */
-    ExpectIntEQ(EmbedGenerateCookie(ssl, buf, FOURK_BUF, NULL), WC_NO_ERR_TRACE(GEN_COOKIE_E));
-
-    wolfSSL_CTX_SetGenCookie(ctx, EmbedGenerateCookie);
-
-    wolfSSL_SetCookieCtx(ssl, ctx);
-
-    ExpectNotNull(wolfSSL_GetCookieCtx(ssl));
-
-    ExpectNull(wolfSSL_GetCookieCtx(NULL));
-
-    SSL_free(ssl);
-    SSL_CTX_free(ctx);
-#endif
-    return EXPECT_RESULT();
-}
 
 static int test_wolfSSL_set_options(void)
 {
@@ -19469,563 +18209,6 @@ static int test_wolfSSL_d2i_SSL_SESSION_bounds_check(void)
     return EXPECT_RESULT();
 }
 
-static int test_wolfSSL_SESSION(void)
-{
-    EXPECT_DECLS;
-#if !defined(NO_FILESYSTEM) && !defined(NO_CERTS) && \
-    !defined(NO_RSA) && !defined(NO_SHA256) && \
-    defined(HAVE_IO_TESTS_DEPENDENCIES) && !defined(NO_SESSION_CACHE)
-    WOLFSSL*     ssl = NULL;
-    WOLFSSL_CTX* ctx = NULL;
-    WOLFSSL_SESSION* sess = NULL;
-    WOLFSSL_SESSION* sess_copy = NULL;
-#ifdef OPENSSL_EXTRA
-#ifdef HAVE_EXT_CACHE
-    unsigned char* sessDer = NULL;
-    unsigned char* ptr     = NULL;
-    int sz = 0;
-#endif
-    const unsigned char context[] = "user app context";
-    unsigned int contextSz = (unsigned int)sizeof(context);
-#endif
-    int ret = 0, err = 0;
-    SOCKET_T sockfd;
-    tcp_ready ready;
-    func_args server_args;
-    THREAD_TYPE serverThread;
-    char msg[80];
-    const char* sendGET = "GET";
-
-    /* TLS v1.3 requires session tickets */
-    /* CHACHA and POLY1305 required for myTicketEncCb */
-#if !defined(WOLFSSL_NO_TLS12) && (!defined(WOLFSSL_TLS13) || \
-    !(defined(HAVE_SESSION_TICKET) && ((defined(HAVE_CHACHA) && \
-            defined(HAVE_POLY1305)) || defined(HAVE_AESGCM))))
-    ExpectNotNull(ctx = wolfSSL_CTX_new(wolfTLSv1_2_client_method()));
-#else
-    ExpectNotNull(ctx = wolfSSL_CTX_new(wolfSSLv23_client_method()));
-#endif
-
-    ExpectTrue(wolfSSL_CTX_use_certificate_file(ctx, cliCertFile,
-        CERT_FILETYPE));
-    ExpectTrue(wolfSSL_CTX_use_PrivateKey_file(ctx, cliKeyFile,
-        CERT_FILETYPE));
-    ExpectIntEQ(wolfSSL_CTX_load_verify_locations(ctx, caCertFile, 0),
-        WOLFSSL_SUCCESS);
-#ifdef WOLFSSL_ENCRYPTED_KEYS
-    wolfSSL_CTX_set_default_passwd_cb(ctx, PasswordCallBack);
-#endif
-#ifdef HAVE_SESSION_TICKET
-    /* Use session tickets, for ticket tests below */
-    ExpectIntEQ(wolfSSL_CTX_UseSessionTicket(ctx), WOLFSSL_SUCCESS);
-#endif
-
-    XMEMSET(&server_args, 0, sizeof(func_args));
-#ifdef WOLFSSL_TIRTOS
-    fdOpenSession(Task_self());
-#endif
-
-    StartTCP();
-    InitTcpReady(&ready);
-
-    server_args.signal = &ready;
-    start_thread(test_server_nofail, &server_args, &serverThread);
-    wait_tcp_ready(&server_args);
-
-    /* client connection */
-    ExpectNotNull(ssl = wolfSSL_new(ctx));
-    tcp_connect(&sockfd, wolfSSLIP, ready.port, 0, 0, ssl);
-    ExpectIntEQ(wolfSSL_set_fd(ssl, sockfd), WOLFSSL_SUCCESS);
-
-    WOLFSSL_ASYNC_WHILE_PENDING(ret = wolfSSL_connect(ssl),
-                                ret != WOLFSSL_SUCCESS);
-    ExpectIntEQ(ret, WOLFSSL_SUCCESS);
-
-    WOLFSSL_ASYNC_WHILE_PENDING(
-        ret = wolfSSL_write(ssl, sendGET, (int)XSTRLEN(sendGET)),
-        ret <= 0);
-    ExpectIntEQ(ret, (int)XSTRLEN(sendGET));
-
-    WOLFSSL_ASYNC_WHILE_PENDING(ret = wolfSSL_read(ssl, msg, sizeof(msg)),
-                                ret != 23);
-    ExpectIntEQ(ret, 23);
-
-    ExpectPtrNE((sess = wolfSSL_get1_session(ssl)), NULL); /* ref count 1 */
-    ExpectPtrNE((sess_copy = wolfSSL_get1_session(ssl)), NULL); /* ref count 2 */
-    ExpectIntEQ(wolfSSL_SessionIsSetup(sess), 1);
-#ifdef HAVE_EXT_CACHE
-    ExpectPtrEq(sess, sess_copy); /* they should be the same pointer but without
-                                   * HAVE_EXT_CACHE we get new objects each time */
-#endif
-    wolfSSL_SESSION_free(sess_copy); sess_copy = NULL;
-    wolfSSL_SESSION_free(sess);      sess = NULL; /* free session ref */
-
-    sess = wolfSSL_get_session(ssl);
-
-#ifdef OPENSSL_EXTRA
-    ExpectIntEQ(SSL_SESSION_is_resumable(NULL), 0);
-    ExpectIntEQ(SSL_SESSION_is_resumable(sess), 1);
-
-    ExpectIntEQ(wolfSSL_SESSION_has_ticket(NULL), 0);
-    ExpectIntEQ(wolfSSL_SESSION_get_ticket_lifetime_hint(NULL), 0);
-    #ifdef HAVE_SESSION_TICKET
-    ExpectIntEQ(wolfSSL_SESSION_has_ticket(sess), 1);
-    ExpectIntEQ(wolfSSL_SESSION_get_ticket_lifetime_hint(sess),
-                SESSION_TICKET_HINT_DEFAULT);
-    #else
-    ExpectIntEQ(wolfSSL_SESSION_has_ticket(sess), 0);
-    #endif
-#else
-    (void)sess;
-#endif /* OPENSSL_EXTRA */
-
-    /* Retain copy of the session for later testing */
-    ExpectNotNull(sess = wolfSSL_get1_session(ssl));
-
-    wolfSSL_shutdown(ssl);
-    wolfSSL_free(ssl); ssl = NULL;
-
-    CloseSocket(sockfd);
-
-    join_thread(serverThread);
-
-    FreeTcpReady(&ready);
-
-#ifdef WOLFSSL_TIRTOS
-    fdOpenSession(Task_self());
-#endif
-
-#if defined(SESSION_CERTS) && defined(OPENSSL_EXTRA)
-    {
-        X509 *x509 = NULL;
-        char buf[30];
-        int  bufSz = 0;
-
-        ExpectNotNull(x509 = SSL_SESSION_get0_peer(sess));
-        ExpectIntGT((bufSz = X509_NAME_get_text_by_NID(
-            X509_get_subject_name(x509), NID_organizationalUnitName, buf,
-            sizeof(buf))), 0);
-        ExpectIntNE((bufSz == 7 || bufSz == 16), 0); /* should be one of these*/
-        if (bufSz == 7) {
-            ExpectIntEQ(XMEMCMP(buf, "Support", bufSz), 0);
-        }
-        if (bufSz == 16) {
-            ExpectIntEQ(XMEMCMP(buf, "Programming-2048", bufSz), 0);
-        }
-    }
-#endif
-
-#ifdef HAVE_EXT_CACHE
-    ExpectNotNull(sess_copy = wolfSSL_SESSION_dup(sess));
-    wolfSSL_SESSION_free(sess_copy); sess_copy = NULL;
-    sess_copy = NULL;
-#endif
-
-#if defined(OPENSSL_EXTRA) && defined(HAVE_EXT_CACHE)
-    /* get session from DER and update the timeout */
-    ExpectIntEQ(wolfSSL_i2d_SSL_SESSION(NULL, &sessDer), WC_NO_ERR_TRACE(BAD_FUNC_ARG));
-    ExpectIntGT((sz = wolfSSL_i2d_SSL_SESSION(sess, &sessDer)), 0);
-    wolfSSL_SESSION_free(sess); sess = NULL;
-    sess = NULL;
-    ptr = sessDer;
-    ExpectNull(sess = wolfSSL_d2i_SSL_SESSION(NULL, NULL, sz));
-    ExpectNotNull(sess = wolfSSL_d2i_SSL_SESSION(NULL,
-                (const unsigned char**)&ptr, sz));
-    XFREE(sessDer, NULL, DYNAMIC_TYPE_OPENSSL);
-    sessDer = NULL;
-
-    ExpectIntGT(wolfSSL_SESSION_get_time(sess), 0);
-    ExpectIntEQ(wolfSSL_SSL_SESSION_set_timeout(sess, 500), SSL_SUCCESS);
-#endif
-
-    /* successful set session test */
-    ExpectNotNull(ssl = wolfSSL_new(ctx));
-    ExpectIntEQ(wolfSSL_set_session(ssl, sess), WOLFSSL_SUCCESS);
-
-#ifdef HAVE_SESSION_TICKET
-    /* Test set/get session ticket */
-    {
-        const char* ticket = "This is a session ticket";
-        char buf[64] = {0};
-        word32 bufSz = (word32)sizeof(buf);
-        word32 retSz = 0;
-
-        ExpectIntEQ(WOLFSSL_SUCCESS,
-            wolfSSL_set_SessionTicket(ssl, (byte *)ticket,
-                (word32)XSTRLEN(ticket)));
-        ExpectIntEQ(WOLFSSL_SUCCESS,
-            wolfSSL_get_SessionTicket(ssl, (byte *)buf, &bufSz));
-        ExpectStrEQ(ticket, buf);
-
-        /* return ticket length if buffer parameter is null */
-        wolfSSL_get_SessionTicket(ssl, NULL, &retSz);
-        ExpectIntEQ(bufSz, retSz);
-    }
-#endif
-
-#ifdef OPENSSL_EXTRA
-    /* session timeout case */
-    /* make the session to be expired */
-    ExpectIntEQ(SSL_SESSION_set_timeout(sess,1), SSL_SUCCESS);
-    XSLEEP_MS(1200);
-
-    /* SSL_set_session should reject specified session but return success
-     * if WOLFSSL_ERROR_CODE_OPENSSL macro is defined for OpenSSL compatibility.
-     */
-#if defined(WOLFSSL_ERROR_CODE_OPENSSL)
-    ExpectIntEQ(wolfSSL_set_session(ssl,sess), SSL_SUCCESS);
-#else
-    ExpectIntEQ(wolfSSL_set_session(ssl,sess), WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
-#endif
-    ExpectIntEQ(wolfSSL_SSL_SESSION_set_timeout(sess, 500), SSL_SUCCESS);
-
-#ifdef WOLFSSL_SESSION_ID_CTX
-    /* fail case with miss match session context IDs (use compatibility API) */
-    ExpectIntEQ(SSL_set_session_id_context(ssl, context, contextSz),
-            SSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_set_session(ssl, sess), WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
-    wolfSSL_free(ssl); ssl = NULL;
-
-    ExpectIntEQ(SSL_CTX_set_session_id_context(NULL, context, contextSz),
-            WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
-    ExpectIntEQ(SSL_CTX_set_session_id_context(ctx, context, contextSz),
-            SSL_SUCCESS);
-    ExpectNotNull(ssl = wolfSSL_new(ctx));
-    ExpectIntEQ(wolfSSL_set_session(ssl, sess), WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
-#endif
-#endif /* OPENSSL_EXTRA */
-
-    wolfSSL_free(ssl);
-    wolfSSL_SESSION_free(sess);
-    wolfSSL_CTX_free(ctx);
-#endif
-    return EXPECT_RESULT();
-}
-
-#if !defined(NO_FILESYSTEM) && !defined(NO_CERTS) && \
-    !defined(NO_RSA) && defined(HAVE_IO_TESTS_DEPENDENCIES) && \
-    !defined(NO_SESSION_CACHE) && defined(OPENSSL_EXTRA) && \
-    !defined(WOLFSSL_NO_TLS12)
-static WOLFSSL_SESSION* test_wolfSSL_SESSION_expire_sess = NULL;
-
-static void test_wolfSSL_SESSION_expire_downgrade_ctx_ready(WOLFSSL_CTX* ctx)
-{
-    #ifdef WOLFSSL_ERROR_CODE_OPENSSL
-    /* returns previous timeout value */
-    AssertIntEQ(wolfSSL_CTX_set_timeout(ctx, 1), 500);
-    #else
-    AssertIntEQ(wolfSSL_CTX_set_timeout(ctx, 1), WOLFSSL_SUCCESS);
-    #endif
-}
-
-
-/* set the session to timeout in a second */
-static void test_wolfSSL_SESSION_expire_downgrade_ssl_ready(WOLFSSL* ssl)
-{
-    AssertIntEQ(wolfSSL_set_timeout(ssl, 2), 1);
-}
-
-
-/* store the client side session from the first successful connection */
-static void test_wolfSSL_SESSION_expire_downgrade_ssl_result(WOLFSSL* ssl)
-{
-    AssertPtrNE((test_wolfSSL_SESSION_expire_sess = wolfSSL_get1_session(ssl)),
-        NULL); /* ref count 1 */
-}
-
-
-/* wait till session is expired then set it in the WOLFSSL struct for use */
-static void test_wolfSSL_SESSION_expire_downgrade_ssl_ready_wait(WOLFSSL* ssl)
-{
-    AssertIntEQ(wolfSSL_set_timeout(ssl, 1), 1);
-    AssertIntEQ(wolfSSL_set_session(ssl, test_wolfSSL_SESSION_expire_sess),
-        WOLFSSL_SUCCESS);
-    XSLEEP_MS(2000); /* wait 2 seconds for session to expire */
-}
-
-
-/* set expired session in the WOLFSSL struct for use */
-static void test_wolfSSL_SESSION_expire_downgrade_ssl_ready_set(WOLFSSL* ssl)
-{
-    XSLEEP_MS(1200); /* wait a second for session to expire */
-
-    /* set the expired session, call to set session fails but continuing on
-       after failure should be handled here */
-#if defined(OPENSSL_EXTRA) && defined(WOLFSSL_ERROR_CODE_OPENSSL)
-    AssertIntEQ(wolfSSL_set_session(ssl, test_wolfSSL_SESSION_expire_sess),
-        WOLFSSL_SUCCESS);
-#else
-    AssertIntNE(wolfSSL_set_session(ssl, test_wolfSSL_SESSION_expire_sess),
-        WOLFSSL_SUCCESS);
-#endif
-}
-
-
-/* check that the expired session was not reused */
-static void test_wolfSSL_SESSION_expire_downgrade_ssl_result_reuse(WOLFSSL* ssl)
-{
-    /* since the session has expired it should not have been reused */
-    AssertIntEQ(wolfSSL_session_reused(ssl), 0);
-}
-#endif
-
-static int test_wolfSSL_SESSION_expire_downgrade(void)
-{
-    EXPECT_DECLS;
-#if !defined(NO_FILESYSTEM) && !defined(NO_CERTS) && \
-    !defined(NO_RSA) && defined(HAVE_IO_TESTS_DEPENDENCIES) && \
-    !defined(NO_SESSION_CACHE) && defined(OPENSSL_EXTRA) && \
-    !defined(WOLFSSL_NO_TLS12)
-    callback_functions server_cbf, client_cbf;
-
-    XMEMSET(&server_cbf, 0, sizeof(callback_functions));
-    XMEMSET(&client_cbf, 0, sizeof(callback_functions));
-
-    /* force server side to use TLS 1.2 */
-    server_cbf.method = wolfTLSv1_2_server_method;
-
-    client_cbf.method = wolfSSLv23_client_method;
-    server_cbf.ctx_ready = test_wolfSSL_SESSION_expire_downgrade_ctx_ready;
-    client_cbf.ssl_ready = test_wolfSSL_SESSION_expire_downgrade_ssl_ready;
-    client_cbf.on_result = test_wolfSSL_SESSION_expire_downgrade_ssl_result;
-
-    test_wolfSSL_client_server_nofail(&client_cbf, &server_cbf);
-    ExpectIntEQ(client_cbf.return_code, TEST_SUCCESS);
-    ExpectIntEQ(server_cbf.return_code, TEST_SUCCESS);
-
-    client_cbf.method = wolfSSLv23_client_method;
-    server_cbf.ctx_ready = test_wolfSSL_SESSION_expire_downgrade_ctx_ready;
-    client_cbf.ssl_ready = test_wolfSSL_SESSION_expire_downgrade_ssl_ready_wait;
-    client_cbf.on_result =
-        test_wolfSSL_SESSION_expire_downgrade_ssl_result_reuse;
-
-    test_wolfSSL_client_server_nofail(&client_cbf, &server_cbf);
-    ExpectIntEQ(client_cbf.return_code, TEST_SUCCESS);
-    ExpectIntEQ(server_cbf.return_code, TEST_SUCCESS);
-
-    client_cbf.method = wolfSSLv23_client_method;
-    server_cbf.ctx_ready = test_wolfSSL_SESSION_expire_downgrade_ctx_ready;
-    client_cbf.ssl_ready = test_wolfSSL_SESSION_expire_downgrade_ssl_ready_set;
-    client_cbf.on_result =
-        test_wolfSSL_SESSION_expire_downgrade_ssl_result_reuse;
-
-    test_wolfSSL_client_server_nofail(&client_cbf, &server_cbf);
-    ExpectIntEQ(client_cbf.return_code, TEST_SUCCESS);
-    ExpectIntEQ(server_cbf.return_code, TEST_SUCCESS);
-
-    wolfSSL_SESSION_free(test_wolfSSL_SESSION_expire_sess);
-#endif
-    return EXPECT_RESULT();
-}
-
-#if defined(OPENSSL_EXTRA) && defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES) && \
-    defined(HAVE_EX_DATA) && !defined(NO_SESSION_CACHE)
-#ifdef WOLFSSL_ATOMIC_OPS
-    typedef wolfSSL_Atomic_Int SessRemCounter_t;
-#else
-    typedef int SessRemCounter_t;
-#endif
-static SessRemCounter_t clientSessRemCountMalloc;
-static SessRemCounter_t serverSessRemCountMalloc;
-static SessRemCounter_t clientSessRemCountFree;
-static SessRemCounter_t serverSessRemCountFree;
-
-static WOLFSSL_CTX* serverSessCtx = NULL;
-static WOLFSSL_SESSION* serverSess = NULL;
-#if (defined(WOLFSSL_TLS13) && defined(HAVE_SESSION_TICKET)) || \
-        !defined(NO_SESSION_CACHE_REF)
-static WOLFSSL_CTX* clientSessCtx = NULL;
-static WOLFSSL_SESSION* clientSess = NULL;
-#endif
-static int serverSessRemIdx = 3;
-static int sessRemCtx_Server = WOLFSSL_SERVER_END;
-static int sessRemCtx_Client = WOLFSSL_CLIENT_END;
-
-static void SessRemCtxCb(WOLFSSL_CTX *ctx, WOLFSSL_SESSION *sess)
-{
-    int* side;
-
-    (void)ctx;
-
-    side = (int*)SSL_SESSION_get_ex_data(sess, serverSessRemIdx);
-    if (side != NULL) {
-        if (*side == WOLFSSL_CLIENT_END)
-            (void)wolfSSL_Atomic_Int_FetchAdd(&clientSessRemCountFree, 1);
-        else
-            (void)wolfSSL_Atomic_Int_FetchAdd(&serverSessRemCountFree, 1);
-
-        SSL_SESSION_set_ex_data(sess, serverSessRemIdx, NULL);
-    }
-}
-
-static int SessRemCtxSetupCb(WOLFSSL_CTX* ctx)
-{
-    SSL_CTX_sess_set_remove_cb(ctx, SessRemCtxCb);
-#if defined(WOLFSSL_TLS13) && !defined(HAVE_SESSION_TICKET) && \
-        !defined(NO_SESSION_CACHE_REF)
-    {
-        EXPECT_DECLS;
-        /* Allow downgrade, set min version, and disable TLS 1.3.
-         * Do this because without NO_SESSION_CACHE_REF we will want to return a
-         * reference to the session cache. But with WOLFSSL_TLS13 and without
-         * HAVE_SESSION_TICKET we won't have a session ID to be able to place
-         * the session in the cache. In this case we need to downgrade to
-         * previous versions to just use the legacy session ID field. */
-        ExpectIntEQ(SSL_CTX_set_min_proto_version(ctx, SSL3_VERSION),
-            SSL_SUCCESS);
-        ExpectIntEQ(SSL_CTX_set_max_proto_version(ctx, TLS1_2_VERSION),
-            SSL_SUCCESS);
-        return EXPECT_RESULT();
-    }
-#else
-    return TEST_SUCCESS;
-#endif
-}
-
-static int SessRemSslSetupCb(WOLFSSL* ssl)
-{
-    EXPECT_DECLS;
-    int* side;
-
-    if (SSL_is_server(ssl)) {
-        side = &sessRemCtx_Server;
-        (void)wolfSSL_Atomic_Int_FetchAdd(&serverSessRemCountMalloc, 1);
-        ExpectNotNull(serverSess = SSL_get1_session(ssl));
-        ExpectIntEQ(SSL_CTX_up_ref(serverSessCtx = SSL_get_SSL_CTX(ssl)),
-                SSL_SUCCESS);
-    }
-    else {
-        side = &sessRemCtx_Client;
-        (void)wolfSSL_Atomic_Int_FetchAdd(&clientSessRemCountMalloc, 1);
-#if (defined(WOLFSSL_TLS13) && defined(HAVE_SESSION_TICKET)) || \
-    !defined(NO_SESSION_CACHE_REF)
-        ExpectNotNull(clientSess = SSL_get1_session(ssl));
-        ExpectIntEQ(SSL_CTX_up_ref(clientSessCtx = SSL_get_SSL_CTX(ssl)),
-                SSL_SUCCESS);
-#endif
-    }
-    ExpectIntEQ(SSL_SESSION_set_ex_data(SSL_get_session(ssl),
-        serverSessRemIdx, side), SSL_SUCCESS);
-
-    return EXPECT_RESULT();
-}
-#endif
-
-static int test_wolfSSL_CTX_sess_set_remove_cb(void)
-{
-    EXPECT_DECLS;
-#if defined(OPENSSL_EXTRA) && defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES) && \
-    defined(HAVE_EX_DATA) && !defined(NO_SESSION_CACHE)
-    /* Check that the remove callback gets called for external data in a
-     * session object */
-    test_ssl_cbf func_cb;
-
-    wolfSSL_Atomic_Int_Init(&clientSessRemCountMalloc, 0);
-    wolfSSL_Atomic_Int_Init(&serverSessRemCountMalloc, 0);
-    wolfSSL_Atomic_Int_Init(&clientSessRemCountFree, 0);
-    wolfSSL_Atomic_Int_Init(&serverSessRemCountFree, 0);
-
-    XMEMSET(&func_cb, 0, sizeof(func_cb));
-    func_cb.ctx_ready = SessRemCtxSetupCb;
-    func_cb.on_result = SessRemSslSetupCb;
-
-    ExpectIntEQ(test_wolfSSL_client_server_nofail_memio(&func_cb, &func_cb,
-        NULL), TEST_SUCCESS);
-
-    /* Both should have been allocated */
-    ExpectIntEQ(clientSessRemCountMalloc, 1);
-    ExpectIntEQ(serverSessRemCountMalloc, 1);
-
-    /* This should not be called yet. Session wasn't evicted from cache yet. */
-    ExpectIntEQ(clientSessRemCountFree, 0);
-#if (defined(WOLFSSL_TLS13) && defined(HAVE_SESSION_TICKET)) || \
-        !defined(NO_SESSION_CACHE_REF)
-    /* Force a cache lookup */
-    ExpectNotNull(SSL_SESSION_get_ex_data(clientSess, serverSessRemIdx));
-    /* Force a cache update */
-    ExpectNotNull(SSL_SESSION_set_ex_data(clientSess, serverSessRemIdx - 1, 0));
-    /* This should set the timeout to 0 and call the remove callback from within
-     * the session cache. Returns 1 per OpenSSL semantics (session was
-     * present in the cache and removed). */
-    ExpectIntEQ(SSL_CTX_remove_session(clientSessCtx, clientSess), 1);
-    ExpectNull(SSL_SESSION_get_ex_data(clientSess, serverSessRemIdx));
-    ExpectIntEQ(clientSessRemCountFree, 1);
-#endif
-    /* Server session is in the cache so ex_data isn't free'd with the SSL
-     * object */
-    ExpectIntEQ(serverSessRemCountFree, 0);
-    /* Force a cache lookup */
-    ExpectNotNull(SSL_SESSION_get_ex_data(serverSess, serverSessRemIdx));
-    /* Force a cache update */
-    ExpectNotNull(SSL_SESSION_set_ex_data(serverSess, serverSessRemIdx - 1, 0));
-    /* This should set the timeout to 0 and call the remove callback from within
-     * the session cache. Returns 1 per OpenSSL semantics (session was
-     * present in the cache and removed). */
-    ExpectIntEQ(SSL_CTX_remove_session(serverSessCtx, serverSess), 1);
-    ExpectNull(SSL_SESSION_get_ex_data(serverSess, serverSessRemIdx));
-    ExpectIntEQ(serverSessRemCountFree, 1);
-    /* Need to free the references that we kept */
-    SSL_CTX_free(serverSessCtx);
-    SSL_SESSION_free(serverSess);
-#if (defined(WOLFSSL_TLS13) && defined(HAVE_SESSION_TICKET)) || \
-        !defined(NO_SESSION_CACHE_REF)
-    SSL_CTX_free(clientSessCtx);
-    SSL_SESSION_free(clientSess);
-#endif
-#endif
-    return EXPECT_RESULT();
-}
-
-static int test_wolfSSL_ticket_keys(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_SESSION_TICKET) && !defined(WOLFSSL_NO_DEF_TICKET_ENC_CB) && \
-    !defined(NO_WOLFSSL_SERVER) && !defined(NO_TLS)
-    WOLFSSL_CTX* ctx = NULL;
-    byte keys[WOLFSSL_TICKET_KEYS_SZ];
-
-    ExpectNotNull(ctx = wolfSSL_CTX_new(wolfSSLv23_server_method()));
-
-    ExpectIntEQ(wolfSSL_CTX_get_tlsext_ticket_keys(NULL, NULL, 0),
-                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
-    ExpectIntEQ(wolfSSL_CTX_get_tlsext_ticket_keys(ctx, NULL, 0),
-                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
-    ExpectIntEQ(wolfSSL_CTX_get_tlsext_ticket_keys(ctx, keys, 0),
-                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
-    ExpectIntEQ(wolfSSL_CTX_get_tlsext_ticket_keys(NULL, keys, 0),
-                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
-    ExpectIntEQ(wolfSSL_CTX_get_tlsext_ticket_keys(NULL, NULL, sizeof(keys)),
-                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
-    ExpectIntEQ(wolfSSL_CTX_get_tlsext_ticket_keys(ctx, NULL, sizeof(keys)),
-                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
-    ExpectIntEQ(wolfSSL_CTX_get_tlsext_ticket_keys(NULL, keys, sizeof(keys)),
-                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
-
-    ExpectIntEQ(wolfSSL_CTX_set_tlsext_ticket_keys(NULL, NULL, 0),
-                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
-    ExpectIntEQ(wolfSSL_CTX_set_tlsext_ticket_keys(ctx, NULL, 0),
-                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
-    ExpectIntEQ(wolfSSL_CTX_set_tlsext_ticket_keys(ctx, keys, 0),
-                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
-    ExpectIntEQ(wolfSSL_CTX_set_tlsext_ticket_keys(NULL, keys, 0),
-                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
-    ExpectIntEQ(wolfSSL_CTX_set_tlsext_ticket_keys(NULL, NULL, sizeof(keys)),
-                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
-    ExpectIntEQ(wolfSSL_CTX_set_tlsext_ticket_keys(ctx, NULL, sizeof(keys)),
-                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
-    ExpectIntEQ(wolfSSL_CTX_set_tlsext_ticket_keys(NULL, keys, sizeof(keys)),
-                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
-
-    ExpectIntEQ(wolfSSL_CTX_get_tlsext_ticket_keys(ctx, keys, sizeof(keys)),
-                WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_CTX_set_tlsext_ticket_keys(ctx, keys, sizeof(keys)),
-                WOLFSSL_SUCCESS);
-
-    wolfSSL_CTX_free(ctx);
-#endif
-    return EXPECT_RESULT();
-}
-
 #ifndef NO_BIO
 
 #if defined(OPENSSL_EXTRA) && defined(WOLFSSL_HAVE_MLDSA)
@@ -20650,6 +18833,7 @@ static int test_wolfSSL_GENERAL_NAME_print(void)
     ACCESS_DESCRIPTION* ad = NULL;
     ASN1_IA5STRING *dnsname = NULL;
     ASN1_OBJECT* ridObj = NULL;
+    X509_NAME* dirName = NULL;
 
     const unsigned char v4Addr[] = {192,168,53,1};
     const unsigned char v6Addr[] =
@@ -20904,22 +19088,35 @@ static int test_wolfSSL_GENERAL_NAME_print(void)
 
     /* test for GEN_DIRNAME */
     ExpectNotNull(gn = wolfSSL_GENERAL_NAME_new());
+    /* Build a real directoryName (X509_NAME) so the print path exercises
+     * wolfSSL_X509_NAME_print_ex on a valid object. Forcing the type without
+     * setting d.directoryName would leave it aliasing the default IA5 string
+     * and cause an out-of-bounds read when printed. */
+    ExpectNotNull(dirName = X509_NAME_new());
+    ExpectIntEQ(X509_NAME_add_entry_by_NID(dirName, NID_commonName,
+        MBSTRING_UTF8, (unsigned char*)"wolfSSLDirNameTest", -1, -1, 0), 1);
     if (gn != NULL) {
+        /* Replace the default IA5 string allocated by GENERAL_NAME_new with
+         * the directoryName and take ownership of it. */
+        wolfSSL_ASN1_STRING_free(gn->d.ia5);
         gn->type = GEN_DIRNAME;
+        gn->d.directoryName = dirName;
+        dirName = NULL; /* gn owns it now; freed by GENERAL_NAME_free */
     }
     ExpectIntEQ(GENERAL_NAME_print(out, gn), 1);
     XMEMSET(outbuf,0,sizeof(outbuf));
     ExpectIntGT(BIO_read(out, outbuf, sizeof(outbuf)), 0);
+    /* Output must start with the label and contain the directory name. */
     ExpectIntEQ(XSTRNCMP((const char*)outbuf, dirNameStr, XSTRLEN(dirNameStr)),
         0);
+    ExpectNotNull(XSTRSTR((const char*)outbuf, "wolfSSLDirNameTest"));
     /* Duplicating GEN_DIRNAME not supported. */
     ExpectNull(dup_gn = GENERAL_NAME_dup(gn));
-    /* Restore to GEN_IA5 (default) to avoid memory leak. */
-    if (gn != NULL) {
-        gn->type = GEN_IA5;
-    }
     GENERAL_NAME_free(gn);
     gn = NULL;
+    /* Only freed here if ownership was not transferred (e.g. gn alloc failed). */
+    X509_NAME_free(dirName);
+    dirName = NULL;
 
     /* test for GEN_RID */
     p = ridData;
@@ -24563,6 +22760,172 @@ static int test_NameConstraints_OtherName(void)
     return EXPECT_RESULT();
 }
 
+#if defined(WOLFSSL_ASN_TEMPLATE) && \
+    defined(WOLFSSL_CERT_REQ) && !defined(NO_ASN_TIME) && \
+    defined(WOLFSSL_CERT_GEN) && defined(HAVE_ECC) && \
+    defined(WOLFSSL_CERT_EXT) && !defined(NO_CERTS) && \
+    defined(WOLFSSL_ALT_NAMES) && defined(WOLFSSL_CUSTOM_OID) && \
+    defined(HAVE_OID_ENCODING) && !defined(IGNORE_NAME_CONSTRAINTS)
+/* Build a SubjectAltName extension value with a single GeneralName of the
+ * given context tag (0x82 dnsName, 0x86 URI) carrying `val`. */
+static word32 build_simple_san(byte* out, word32 outSz, byte gnTag,
+    const char* val)
+{
+    word32 vlen = (word32)XSTRLEN(val);
+    if (vlen > 0x7F || outSz < vlen + 4)
+        return 0;
+    out[0] = 0x30;                              /* SEQUENCE */
+    out[1] = (byte)(vlen + 2);
+    out[2] = gnTag;                             /* [tag] GeneralName */
+    out[3] = (byte)vlen;
+    XMEMCPY(out + 4, val, vlen);
+    return vlen + 4;
+}
+
+/* Build a NameConstraints extension value with a single subtree ([0]
+ * permitted or [1] excluded) carrying a GeneralName of context tag `gnTag`
+ * with value `val`. */
+static word32 build_simple_nameConstraints(byte* out, word32 outSz,
+    int excluded, byte gnTag, const char* val)
+{
+    word32 vlen = (word32)XSTRLEN(val);
+    word32 n3 = vlen + 2;       /* GeneralSubtree content: the base GN */
+    word32 n2 = n3 + 2;         /* subtrees list content: one GeneralSubtree */
+    word32 n1 = n2 + 2;         /* SEQUENCE content: the subtrees list */
+    if (vlen > 0x7F || outSz < n1 + 2)
+        return 0;
+    out[0] = 0x30;                              /* SEQUENCE */
+    out[1] = (byte)n1;
+    out[2] = excluded ? 0xA1 : 0xA0;            /* [1] excluded / [0] permitted */
+    out[3] = (byte)n2;
+    out[4] = 0x30;                              /* GeneralSubtree */
+    out[5] = (byte)n3;
+    out[6] = gnTag;                             /* base GeneralName */
+    out[7] = (byte)vlen;
+    XMEMCPY(out + 8, val, vlen);
+    return n1 + 2;
+}
+#endif
+
+/* End-to-end enforcement of DNS and URI nameConstraints against wildcard and
+ * sub-host leaf SANs, exercised through the real chain-verification path.
+ *
+ * DNS subtree semantics (RFC 5280 4.2.1.10) plus wildcard reasoning:
+ *   - excluded DNS:foo.example.com must reject a leaf SAN *.example.com,
+ *     because the wildcard can expand to the excluded host (security gap).
+ *   - permitted DNS:example.com must accept *.example.com (contained), while
+ *     permitted DNS:foo.example.com must reject it (the wildcard can escape).
+ * URI semantics: a constraint without a leading dot is an EXACT host match,
+ * not a subtree, so permitted URI:host.com must reject https://www.host.com/.
+ */
+static int test_NameConstraints_DnsUriWildcard(void)
+{
+    EXPECT_DECLS;
+#if defined(WOLFSSL_ASN_TEMPLATE) && \
+    defined(WOLFSSL_CERT_REQ) && !defined(NO_ASN_TIME) && \
+    defined(WOLFSSL_CERT_GEN) && defined(HAVE_ECC) && \
+    defined(WOLFSSL_CERT_EXT) && !defined(NO_CERTS) && \
+    defined(WOLFSSL_ALT_NAMES) && defined(WOLFSSL_CUSTOM_OID) && \
+    defined(HAVE_OID_ENCODING) && !defined(IGNORE_NAME_CONSTRAINTS)
+    byte nc[64];
+    byte san[64];
+    word32 ncSz, sanSz;
+    const byte DNS = 0x82;  /* [2] dnsName */
+    const byte URI = 0x86;  /* [6] uniformResourceIdentifier */
+
+    /* --- DNS, excluded subtree --- */
+
+    /* (1) Excluded foo.example.com vs wildcard SAN that can reach it: the
+     *     pre-fix byte-length guard let this through. Must reject. */
+    ncSz  = build_simple_nameConstraints(nc, sizeof(nc), 1, DNS,
+                "foo.example.com");
+    sanSz = build_simple_san(san, sizeof(san), DNS, "*.example.com");
+    ExpectIntGT((int)ncSz, 0);
+    ExpectIntGT((int)sanSz, 0);
+    ExpectIntEQ(verify_with_otherName_chain(nc, ncSz, 1, san, sanSz),
+        WC_NO_ERR_TRACE(ASN_NAME_INVALID_E));
+
+    /* (2) Positive control: same exclusion, wildcard in a different domain
+     *     cannot reach foo.example.com -> accept. */
+    sanSz = build_simple_san(san, sizeof(san), DNS, "*.other.com");
+    ExpectIntGT((int)sanSz, 0);
+    ExpectIntEQ(verify_with_otherName_chain(nc, ncSz, 1, san, sanSz), 0);
+
+    /* (3) Regression: literal SAN inside the excluded subtree still rejects. */
+    sanSz = build_simple_san(san, sizeof(san), DNS, "foo.example.com");
+    ExpectIntGT((int)sanSz, 0);
+    ExpectIntEQ(verify_with_otherName_chain(nc, ncSz, 1, san, sanSz),
+        WC_NO_ERR_TRACE(ASN_NAME_INVALID_E));
+
+    /* (4) Regression: literal SAN outside the excluded subtree accepts. */
+    sanSz = build_simple_san(san, sizeof(san), DNS, "bar.other.com");
+    ExpectIntGT((int)sanSz, 0);
+    ExpectIntEQ(verify_with_otherName_chain(nc, ncSz, 1, san, sanSz), 0);
+
+    /* --- DNS, permitted subtree --- */
+
+    /* (5) Permitted example.com accepts *.example.com (fully contained). */
+    ncSz  = build_simple_nameConstraints(nc, sizeof(nc), 0, DNS, "example.com");
+    sanSz = build_simple_san(san, sizeof(san), DNS, "*.example.com");
+    ExpectIntGT((int)ncSz, 0);
+    ExpectIntGT((int)sanSz, 0);
+    ExpectIntEQ(verify_with_otherName_chain(nc, ncSz, 1, san, sanSz), 0);
+
+    /* (6) Permitted foo.example.com rejects *.example.com (the wildcard can
+     *     expand outside the permitted subtree). */
+    ncSz  = build_simple_nameConstraints(nc, sizeof(nc), 0, DNS,
+                "foo.example.com");
+    sanSz = build_simple_san(san, sizeof(san), DNS, "*.example.com");
+    ExpectIntGT((int)ncSz, 0);
+    ExpectIntGT((int)sanSz, 0);
+    ExpectIntEQ(verify_with_otherName_chain(nc, ncSz, 1, san, sanSz),
+        WC_NO_ERR_TRACE(ASN_NAME_INVALID_E));
+
+    /* --- URI, exact-host vs subtree semantics --- */
+
+    /* (7) Permitted URI host.com (no leading dot) is an EXACT host match:
+     *     a sub-host URI is NOT contained -> reject. Pre-fix this wrongly
+     *     subtree-matched and accepted. */
+    ncSz  = build_simple_nameConstraints(nc, sizeof(nc), 0, URI, "host.com");
+    sanSz = build_simple_san(san, sizeof(san), URI, "https://www.host.com/");
+    ExpectIntGT((int)ncSz, 0);
+    ExpectIntGT((int)sanSz, 0);
+    ExpectIntEQ(verify_with_otherName_chain(nc, ncSz, 1, san, sanSz),
+        WC_NO_ERR_TRACE(ASN_NAME_INVALID_E));
+
+    /* (8) Permitted URI host.com accepts the exact host. */
+    sanSz = build_simple_san(san, sizeof(san), URI, "https://host.com/path");
+    ExpectIntGT((int)sanSz, 0);
+    ExpectIntEQ(verify_with_otherName_chain(nc, ncSz, 1, san, sanSz), 0);
+
+    /* (9) Permitted URI .host.com (leading dot) is a subtree: the sub-host is
+     *     accepted but the bare host is not. */
+    ncSz  = build_simple_nameConstraints(nc, sizeof(nc), 0, URI, ".host.com");
+    sanSz = build_simple_san(san, sizeof(san), URI, "https://www.host.com/");
+    ExpectIntGT((int)ncSz, 0);
+    ExpectIntGT((int)sanSz, 0);
+    ExpectIntEQ(verify_with_otherName_chain(nc, ncSz, 1, san, sanSz), 0);
+
+    sanSz = build_simple_san(san, sizeof(san), URI, "https://host.com/");
+    ExpectIntGT((int)sanSz, 0);
+    ExpectIntEQ(verify_with_otherName_chain(nc, ncSz, 1, san, sanSz),
+        WC_NO_ERR_TRACE(ASN_NAME_INVALID_E));
+
+    /* (10) Excluded URI host.com (exact): bare host excluded, sub-host not. */
+    ncSz  = build_simple_nameConstraints(nc, sizeof(nc), 1, URI, "host.com");
+    sanSz = build_simple_san(san, sizeof(san), URI, "https://host.com/");
+    ExpectIntGT((int)ncSz, 0);
+    ExpectIntGT((int)sanSz, 0);
+    ExpectIntEQ(verify_with_otherName_chain(nc, ncSz, 1, san, sanSz),
+        WC_NO_ERR_TRACE(ASN_NAME_INVALID_E));
+
+    sanSz = build_simple_san(san, sizeof(san), URI, "https://www.host.com/");
+    ExpectIntGT((int)sanSz, 0);
+    ExpectIntEQ(verify_with_otherName_chain(nc, ncSz, 1, san, sanSz), 0);
+#endif
+    return EXPECT_RESULT();
+}
+
 static int test_MakeCertWithCaFalse(void)
 {
     EXPECT_DECLS;
@@ -27058,6 +25421,111 @@ static int test_wolfSSL_X509_print(void)
     return EXPECT_RESULT();
 }
 
+static int test_wolfSSL_X509_print_basic_constraints(void)
+{
+    EXPECT_DECLS;
+#if defined(OPENSSL_EXTRA) && !defined(NO_FILESYSTEM) && \
+   !defined(NO_RSA) && defined(XSNPRINTF) && !defined(WC_DISABLE_RADIX_ZERO_PAD)
+    /* X509_print must match OpenSSL's Basic Constraints output: "CA:TRUE"
+     * alone when no pathLenConstraint is present, and "CA:TRUE, pathlen:N"
+     * when one is (including the meaningful value 0). */
+    struct {
+        const char* file;
+        const char* expect; /* substring that must be present */
+        const char* absent; /* substring that must be absent, or NULL */
+    } cases[] = {
+        { "./certs/ca-cert.pem", "CA:TRUE", "pathlen" },
+        { "./certs/intermediate/ca-int-cert.pem", "CA:TRUE, pathlen:1", NULL },
+        { "./certs/test-pathlen/chainG-ICA1-pathlen0.pem",
+              "CA:TRUE, pathlen:0", NULL },
+    };
+    size_t i;
+
+    for (i = 0; i < sizeof(cases) / sizeof(cases[0]); i++) {
+        X509* x509 = NULL;
+        BIO*  bio  = NULL;
+        char* data = NULL;
+        int   len  = 0;
+        char  buf[8192];
+
+        ExpectNotNull(x509 = X509_load_certificate_file(cases[i].file,
+            WOLFSSL_FILETYPE_PEM));
+        ExpectNotNull(bio = BIO_new(BIO_s_mem()));
+        ExpectIntEQ(X509_print(bio, x509), SSL_SUCCESS);
+        /* Memory BIO data is not NUL-terminated; copy into a bounded buffer. */
+        ExpectIntGT((len = BIO_get_mem_data(bio, &data)), 0);
+        ExpectIntLT(len, (int)sizeof(buf));
+        if ((data != NULL) && (len > 0) && (len < (int)sizeof(buf))) {
+            XMEMCPY(buf, data, (size_t)len);
+            buf[len] = '\0';
+            ExpectNotNull(XSTRSTR(buf, cases[i].expect));
+            if (cases[i].absent != NULL) {
+                ExpectNull(XSTRSTR(buf, cases[i].absent));
+            }
+        }
+        BIO_free(bio);
+        X509_free(x509);
+    }
+#endif
+    return EXPECT_RESULT();
+}
+
+static int test_wolfSSL_X509_print_ext_key_usage(void)
+{
+    EXPECT_DECLS;
+#if defined(OPENSSL_EXTRA) && !defined(NO_FILESYSTEM) && \
+   !defined(NO_RSA) && defined(XSNPRINTF) && !defined(WC_DISABLE_RADIX_ZERO_PAD)
+    /* Self-signed RSA cert (CN="wolfSSL anyEKU test", valid until 2126) whose
+     * only Extended Key Usage is anyExtendedKeyUsage (OID 2.5.29.37.0).
+     * X509_print must render it as "Any Extended Key Usage" to match OpenSSL,
+     * rather than omitting it. */
+    static const char anyEkuCertPem[] =
+        "-----BEGIN CERTIFICATE-----\n"
+        "MIIDMDCCAhigAwIBAgIUGFipYIiuuQZHNjsi0R8t6nZrFVowDQYJKoZIhvcNAQEL\n"
+        "BQAwHjEcMBoGA1UEAwwTd29sZlNTTCBhbnlFS1UgdGVzdDAgFw0yNjA1MjgwMTI5\n"
+        "MjBaGA8yMTI2MDUwNDAxMjkyMFowHjEcMBoGA1UEAwwTd29sZlNTTCBhbnlFS1Ug\n"
+        "dGVzdDCCASIwDQYJKoZIhvcNAQEBBQADggEPADCCAQoCggEBAKIXS15QjpWlqpqk\n"
+        "sUTD/mgm7akIZfp2DVoweJsf8BC/tnwMX1noWUFjBC8SLOHyhx4bmWxVBzFXv/uu\n"
+        "ICoUCq63pgnxj3rQbPAa1pwAX4UsYS1PS/2mO2o7lRLRdDLIaBXuUClVHGjti9x5\n"
+        "x5++4FFhOhKNt7CkYVAfXasTFKdZqSnKdlYX2rM8LoCjP3YHIC5jEIgjyNUEqzup\n"
+        "Ls02dIaAly5O8yzasUtULwE76E/UpyhE+o2dkhxnc6ukqU5CBLeHNOUJgvThW1lS\n"
+        "SlGXl8Kd5uvBvDflXb6y3TQBEY/hb40JzYeH+hHf4YIqCtvfy6PMA+Rcu2CKlDpP\n"
+        "FVf85pUCAwEAAaNkMGIwHQYDVR0OBBYEFAkmMlYp9K3sbeQ8/RluOviXZ3JcMB8G\n"
+        "A1UdIwQYMBaAFAkmMlYp9K3sbeQ8/RluOviXZ3JcMA8GA1UdEwEB/wQFMAMBAf8w\n"
+        "DwYDVR0lBAgwBgYEVR0lADANBgkqhkiG9w0BAQsFAAOCAQEAHX5wAZDAyFSmmsn4\n"
+        "Mu7TayCx+VbcBgvL4ZdxxJYXexslzI8OviKSgpC56LaVB5JpNqZtFS4pZZikG7nv\n"
+        "kYCoiU33XN82/gggq1bv8rKO740V6kGev3gfEeJuTX30fCPZ18znE1fU8+VQba1L\n"
+        "bPn0h746Ivom47/1VMg6y7CbTJg90+lloPWcTWujYfm5jWychqWurhfZmAYcUlBH\n"
+        "Ksk0l4kiEJA6lSnPp3MqBS0GzwsCixSqYc1W1TlwNGNg38cLP2Z5jerJZlazuVws\n"
+        "SeEbYO6TIhsy6QJy0Pd7hu9DUOKRxp+OQubL6WgWpjrGl1LUzH5sI5pyWueEAEBu\n"
+        "e74xbw==\n"
+        "-----END CERTIFICATE-----\n";
+    X509* x509 = NULL;
+    BIO*  bio  = NULL;
+    char* data = NULL;
+    int   len  = 0;
+    char  buf[8192];
+
+    ExpectNotNull(x509 = wolfSSL_X509_load_certificate_buffer(
+        (const unsigned char*)anyEkuCertPem, (int)XSTRLEN(anyEkuCertPem),
+        WOLFSSL_FILETYPE_PEM));
+    ExpectNotNull(bio = BIO_new(BIO_s_mem()));
+    ExpectIntEQ(X509_print(bio, x509), SSL_SUCCESS);
+    /* Memory BIO data is not NUL-terminated; copy into a bounded buffer. */
+    ExpectIntGT((len = BIO_get_mem_data(bio, &data)), 0);
+    ExpectIntLT(len, (int)sizeof(buf));
+    if ((data != NULL) && (len > 0) && (len < (int)sizeof(buf))) {
+        XMEMCPY(buf, data, (size_t)len);
+        buf[len] = '\0';
+        ExpectNotNull(XSTRSTR(buf, "X509v3 Extended Key Usage"));
+        ExpectNotNull(XSTRSTR(buf, "Any Extended Key Usage"));
+    }
+    BIO_free(bio);
+    X509_free(x509);
+#endif
+    return EXPECT_RESULT();
+}
+
 static int test_wolfSSL_X509_CRL_print(void)
 {
     EXPECT_DECLS;
@@ -27669,1188 +26137,6 @@ static int test_wolfSSL_CTX_get0_privatekey(void)
     return EXPECT_RESULT();
 }
 
-static int test_wolfSSL_dtls_set_mtu(void)
-{
-    EXPECT_DECLS;
-#if (defined(WOLFSSL_DTLS_MTU) || defined(WOLFSSL_SCTP)) && \
-    !defined(NO_WOLFSSL_SERVER) && defined(WOLFSSL_DTLS) && \
-    !defined(WOLFSSL_NO_TLS12)
-    WOLFSSL_CTX* ctx = NULL;
-    WOLFSSL*     ssl = NULL;
-    const char* testCertFile;
-    const char* testKeyFile;
-
-    ExpectNotNull(ctx = wolfSSL_CTX_new(wolfDTLSv1_2_server_method()));
-#ifndef NO_RSA
-        testCertFile = svrCertFile;
-        testKeyFile = svrKeyFile;
-#elif defined(HAVE_ECC)
-        testCertFile = eccCertFile;
-        testKeyFile = eccKeyFile;
-#endif
-    if  (testCertFile != NULL && testKeyFile != NULL) {
-        ExpectTrue(wolfSSL_CTX_use_certificate_file(ctx, testCertFile,
-                                                    WOLFSSL_FILETYPE_PEM));
-        ExpectTrue(wolfSSL_CTX_use_PrivateKey_file(ctx, testKeyFile,
-                                                   WOLFSSL_FILETYPE_PEM));
-    }
-    ExpectNotNull(ssl = wolfSSL_new(ctx));
-
-    ExpectIntEQ(wolfSSL_CTX_dtls_set_mtu(NULL, 1488), WC_NO_ERR_TRACE(BAD_FUNC_ARG));
-    ExpectIntEQ(wolfSSL_dtls_set_mtu(NULL, 1488), WC_NO_ERR_TRACE(BAD_FUNC_ARG));
-    ExpectIntEQ(wolfSSL_CTX_dtls_set_mtu(ctx, 20000), WC_NO_ERR_TRACE(BAD_FUNC_ARG));
-    ExpectIntEQ(wolfSSL_dtls_set_mtu(ssl, 20000), WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
-    ExpectIntEQ(wolfSSL_get_error(ssl, WC_NO_ERR_TRACE(WOLFSSL_FAILURE)), WC_NO_ERR_TRACE(BAD_FUNC_ARG));
-    ExpectIntEQ(wolfSSL_CTX_dtls_set_mtu(ctx, 1488), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_dtls_set_mtu(ssl, 1488), WOLFSSL_SUCCESS);
-
-#ifdef OPENSSL_EXTRA
-    ExpectIntEQ(SSL_set_mtu(ssl, 1488), WOLFSSL_SUCCESS);
-#endif
-
-    wolfSSL_free(ssl);
-    wolfSSL_CTX_free(ctx);
-#endif
-
-    return EXPECT_RESULT();
-}
-
-#if defined(HAVE_IO_TESTS_DEPENDENCIES) && !defined(SINGLE_THREADED) && \
-    defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12)
-
-static WC_INLINE void generateDTLSMsg(byte* out, int outSz, word32 seq,
-        enum HandShakeType hsType, word16 length)
-{
-    size_t idx = 0;
-    byte* l;
-
-    /* record layer */
-    /* handshake type */
-    out[idx++] = handshake;
-    /* protocol version */
-    out[idx++] = 0xfe;
-    out[idx++] = 0xfd; /* DTLS 1.2 */
-    /* epoch 0 */
-    XMEMSET(out + idx, 0, 2);
-    idx += 2;
-    /* sequence number */
-    XMEMSET(out + idx, 0, 6);
-    c32toa(seq, out + idx + 2);
-    idx += 6;
-    /* length in BE */
-    if (length)
-        c16toa(length, out + idx);
-    else
-        c16toa(outSz - idx - 2, out + idx);
-    idx += 2;
-
-    /* handshake layer */
-    /* handshake type */
-    out[idx++] = (byte)hsType;
-    /* length */
-    l = out + idx;
-    idx += 3;
-    /* message seq */
-    c16toa(0, out + idx);
-    idx += 2;
-    /* frag offset */
-    c32to24(0, out + idx);
-    idx += 3;
-    /* frag length */
-    c32to24((word32)outSz - (word32)idx - 3, l);
-    c32to24((word32)outSz - (word32)idx - 3, out + idx);
-    idx += 3;
-    XMEMSET(out + idx, 0, outSz - idx);
-}
-
-static void test_wolfSSL_dtls_plaintext_server(WOLFSSL* ssl)
-{
-    byte msg[] = "This is a msg for the client";
-    byte reply[40];
-    AssertIntGT(wolfSSL_read(ssl, reply, sizeof(reply)),0);
-    reply[sizeof(reply) - 1] = '\0';
-    fprintf(stderr, "Client message: %s\n", reply);
-    AssertIntEQ(wolfSSL_write(ssl, msg, sizeof(msg)), sizeof(msg));
-}
-
-static void test_wolfSSL_dtls_plaintext_client(WOLFSSL* ssl)
-{
-    byte ch[50];
-    int fd = wolfSSL_get_wfd(ssl);
-    byte msg[] = "This is a msg for the server";
-    byte reply[40];
-
-    AssertIntGE(fd, 0);
-    generateDTLSMsg(ch, sizeof(ch), 20, client_hello, 0);
-    /* Server should ignore this datagram */
-    AssertIntEQ(send(fd, (MESSAGE_TYPE_CAST)ch, sizeof(ch), 0), sizeof(ch));
-    generateDTLSMsg(ch, sizeof(ch), 20, client_hello, 10000);
-    /* Server should ignore this datagram */
-    AssertIntEQ(send(fd, (MESSAGE_TYPE_CAST)ch, sizeof(ch), 0), sizeof(ch));
-
-    AssertIntEQ(wolfSSL_write(ssl, msg, sizeof(msg)), sizeof(msg));
-    AssertIntGT(wolfSSL_read(ssl, reply, sizeof(reply)),0);
-    reply[sizeof(reply) - 1] = '\0';
-    fprintf(stderr, "Server response: %s\n", reply);
-}
-
-static int test_wolfSSL_dtls_plaintext(void)
-{
-    callback_functions func_cb_client;
-    callback_functions func_cb_server;
-    size_t i;
-    struct test_params {
-        method_provider client_meth;
-        method_provider server_meth;
-        ssl_callback on_result_server;
-        ssl_callback on_result_client;
-    } params[] = {
-        {wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method,
-                test_wolfSSL_dtls_plaintext_server,
-                test_wolfSSL_dtls_plaintext_client},
-    };
-
-    for (i = 0; i < sizeof(params)/sizeof(*params); i++) {
-        XMEMSET(&func_cb_client, 0, sizeof(callback_functions));
-        XMEMSET(&func_cb_server, 0, sizeof(callback_functions));
-
-        func_cb_client.doUdp = func_cb_server.doUdp = 1;
-        func_cb_server.method = params[i].server_meth;
-        func_cb_client.method = params[i].client_meth;
-        func_cb_client.on_result = params[i].on_result_client;
-        func_cb_server.on_result = params[i].on_result_server;
-
-        test_wolfSSL_client_server_nofail(&func_cb_client, &func_cb_server);
-
-        if (!func_cb_client.return_code)
-            return TEST_FAIL;
-        if (!func_cb_server.return_code)
-            return TEST_FAIL;
-    }
-
-    return TEST_RES_CHECK(1);
-}
-#else
-static int test_wolfSSL_dtls_plaintext(void)
-{
-    return TEST_SKIPPED;
-}
-#endif
-
-#if defined(HAVE_IO_TESTS_DEPENDENCIES) && !defined(SINGLE_THREADED) && \
-    defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12)
-
-static void test_wolfSSL_dtls12_fragments_spammer(WOLFSSL* ssl)
-{
-    byte b[1100]; /* buffer for the messages to send */
-    size_t idx = 0;
-    size_t seq_offset = 0;
-    size_t msg_offset = 0;
-    int i;
-    int fd = wolfSSL_get_wfd(ssl);
-    int ret = wolfSSL_connect_cert(ssl); /* This gets us past the cookie */
-    word32 seq_number = 100; /* start high so server definitely reads this */
-    word16 msg_number = 50; /* start high so server has to buffer this */
-    AssertIntEQ(ret, 1);
-    /* Now let's start spamming the peer with fragments it needs to store */
-    XMEMSET(b, -1, sizeof(b));
-
-    /* record layer */
-    /* handshake type */
-    b[idx++] = 22;
-    /* protocol version */
-    b[idx++] = 0xfe;
-    b[idx++] = 0xfd; /* DTLS 1.2 */
-    /* epoch 0 */
-    XMEMSET(b + idx, 0, 2);
-    idx += 2;
-    /* sequence number */
-    XMEMSET(b + idx, 0, 6);
-    seq_offset = idx + 2; /* increment only the low 32 bits */
-    idx += 6;
-    /* static length in BE */
-    c16toa(42, b + idx);
-    idx += 2;
-
-    /* handshake layer */
-    /* cert type */
-    b[idx++] = 11;
-    /* length */
-    c32to24(1000, b + idx);
-    idx += 3;
-    /* message seq */
-    c16toa(0, b + idx);
-    msg_offset = idx;
-    idx += 2;
-    /* frag offset */
-    c32to24(500, b + idx);
-    idx += 3;
-    /* frag length */
-    c32to24(30, b + idx);
-    idx += 3;
-    (void)idx; /* inhibit clang-analyzer-deadcode.DeadStores */
-
-    for (i = 0; i < DTLS_POOL_SZ * 2 && ret > 0;
-            seq_number++, msg_number++, i++) {
-        struct timespec delay;
-        XMEMSET(&delay, 0, sizeof(delay));
-        delay.tv_nsec = 10000000; /* wait 0.01 seconds */
-        c32toa(seq_number, b + seq_offset);
-        c16toa(msg_number, b + msg_offset);
-        ret = (int)send(fd, (MESSAGE_TYPE_CAST)b, 55, 0);
-        nanosleep(&delay, NULL);
-    }
-}
-
-#ifdef WOLFSSL_DTLS13
-static void test_wolfSSL_dtls13_fragments_spammer(WOLFSSL* ssl)
-{
-    const word16 sendCountMax = 100;
-    byte b[150]; /* buffer for the messages to send */
-    size_t idx = 0;
-    size_t msg_offset = 0;
-    int fd = wolfSSL_get_wfd(ssl);
-    word16 msg_number = 10; /* start high so server has to buffer this */
-    int ret = wolfSSL_connect_cert(ssl); /* This gets us past the cookie */
-    AssertIntEQ(ret, 1);
-    /* Now let's start spamming the peer with fragments it needs to store */
-    XMEMSET(b, -1, sizeof(b));
-
-    /* handshake type */
-    b[idx++] = 11;
-    /* length */
-    c32to24(10000, b + idx);
-    idx += 3;
-    /* message_seq */
-    msg_offset = idx;
-    idx += 2;
-    /* fragment_offset */
-    c32to24(5000, b + idx);
-    idx += 3;
-    /* fragment_length */
-    c32to24(100, b + idx);
-    idx += 3;
-    /* fragment contents */
-    idx += 100;
-
-    for (; ret > 0 && msg_number < sendCountMax; msg_number++) {
-        byte sendBuf[150];
-        int sendSz = sizeof(sendBuf);
-        struct timespec delay;
-        XMEMSET(&delay, 0, sizeof(delay));
-        delay.tv_nsec = 10000000; /* wait 0.01 seconds */
-        c16toa(msg_number, b + msg_offset);
-        ret = sendSz = BuildTls13Message(ssl, sendBuf, sendSz, b,
-            (int)idx, handshake, 0, 0, 0);
-        if (sendSz > 0)
-            ret = (int)send(fd, (MESSAGE_TYPE_CAST)sendBuf, (size_t)sendSz, 0);
-        nanosleep(&delay, NULL);
-    }
-}
-#endif
-
-static int test_wolfSSL_dtls_fragments(void)
-{
-    EXPECT_DECLS;
-    callback_functions func_cb_client;
-    callback_functions func_cb_server;
-    size_t i;
-    struct test_params {
-        method_provider client_meth;
-        method_provider server_meth;
-        ssl_callback spammer;
-    } params[] = {
-#if !defined(WOLFSSL_NO_TLS12)
-        {wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method,
-                test_wolfSSL_dtls12_fragments_spammer},
-#endif
-#ifdef WOLFSSL_DTLS13
-        {wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method,
-                test_wolfSSL_dtls13_fragments_spammer},
-#endif
-    };
-
-    for (i = 0; i < sizeof(params)/sizeof(*params); i++) {
-        XMEMSET(&func_cb_client, 0, sizeof(callback_functions));
-        XMEMSET(&func_cb_server, 0, sizeof(callback_functions));
-
-        func_cb_client.doUdp = func_cb_server.doUdp = 1;
-        func_cb_server.method = params[i].server_meth;
-        func_cb_client.method = params[i].client_meth;
-        func_cb_client.ssl_ready = params[i].spammer;
-
-        test_wolfSSL_client_server_nofail(&func_cb_client, &func_cb_server);
-
-        /* If the client failed, check that the error it encountered was from
-         * the server aborting, resulting in a socket error, fatal error or
-         * reading a close notify alert.
-         *
-         * Under slow execution (e.g. valgrind + noasm), the server may
-         * still be processing fragments when the client completes its
-         * handshake and write, so the client may succeed -- in that
-         * case return_code is TEST_SUCCESS and these checks don't apply.
-         */
-        if (func_cb_client.return_code == TEST_FAIL) {
-            if (func_cb_client.last_err != WC_NO_ERR_TRACE(SOCKET_ERROR_E) &&
-                    func_cb_client.last_err != WOLFSSL_ERROR_ZERO_RETURN &&
-                    func_cb_client.last_err != WC_NO_ERR_TRACE(FATAL_ERROR)) {
-                ExpectIntEQ(func_cb_client.last_err, WC_NO_ERR_TRACE(SOCKET_ERROR_E));
-            }
-        }
-        /* Check the server returned an error indicating the msg buffer
-         * was full.
-         *
-         * Under slow execution (e.g. valgrind + noasm), the real handshake
-         * from wolfSSL_negotiate() may complete before enough spam fragments
-         * accumulate to trigger DTLS_TOO_MANY_FRAGMENTS_E. Accept both
-         * outcomes: server hit the fragment limit, or completed normally.
-         */
-        if (func_cb_server.return_code == TEST_FAIL) {
-            ExpectIntEQ(func_cb_server.last_err, WC_NO_ERR_TRACE(DTLS_TOO_MANY_FRAGMENTS_E));
-        }
-
-        if (EXPECT_FAIL())
-            break;
-    }
-
-    return EXPECT_RESULT();
-}
-
-static void test_wolfSSL_dtls_send_alert(WOLFSSL* ssl)
-{
-    int fd, ret;
-    byte alert_msg[] = {
-        0x15, /* alert type */
-        0xfe, 0xfd, /* version */
-        0x00, 0x00, /* epoch */
-        0x00, 0x00, 0x00, 0x00, 0x00, 0x01, /* seq number */
-        0x00, 0x02, /* length */
-        0x02, /* level: fatal */
-        0x46 /* protocol version */
-    };
-
-    fd = wolfSSL_get_wfd(ssl);
-    AssertIntGE(fd, 0);
-    ret = (int)send(fd, (MESSAGE_TYPE_CAST)alert_msg, sizeof(alert_msg), 0);
-    AssertIntGT(ret, 0);
-}
-
-static int _test_wolfSSL_ignore_alert_before_cookie(byte version12)
-{
-    callback_functions client_cbs, server_cbs;
-
-    XMEMSET(&client_cbs, 0, sizeof(client_cbs));
-    XMEMSET(&server_cbs, 0, sizeof(server_cbs));
-    client_cbs.doUdp = server_cbs.doUdp = 1;
-    if (version12) {
-#if !defined(WOLFSSL_NO_TLS12)
-        client_cbs.method = wolfDTLSv1_2_client_method;
-        server_cbs.method = wolfDTLSv1_2_server_method;
-#else
-        return TEST_SKIPPED;
-#endif
-    }
-    else
-    {
-#ifdef WOLFSSL_DTLS13
-        client_cbs.method = wolfDTLSv1_3_client_method;
-        server_cbs.method = wolfDTLSv1_3_server_method;
-#else
-        return TEST_SKIPPED;
-#endif /* WOLFSSL_DTLS13 */
-    }
-
-    client_cbs.ssl_ready = test_wolfSSL_dtls_send_alert;
-    test_wolfSSL_client_server_nofail(&client_cbs, &server_cbs);
-
-    if (!client_cbs.return_code)
-        return TEST_FAIL;
-    if (!server_cbs.return_code)
-        return TEST_FAIL;
-
-    return TEST_SUCCESS;
-}
-
-static int test_wolfSSL_ignore_alert_before_cookie(void)
-{
-    int ret;
-    ret =_test_wolfSSL_ignore_alert_before_cookie(0);
-    if (ret != 0)
-        return ret;
-    ret =_test_wolfSSL_ignore_alert_before_cookie(1);
-    if (ret != 0)
-        return ret;
-    return 0;
-}
-
-static void test_wolfSSL_send_bad_record(WOLFSSL* ssl)
-{
-    int ret;
-    int fd;
-
-    byte bad_msg[] = {
-        0x17, /* app data  */
-        0xaa, 0xfd, /* bad version */
-        0x00, 0x01, /* epoch 1 */
-        0x00, 0x00, 0x00, 0x00, 0x00, 0x55, /* not seen seq number */
-        0x00, 0x26, /* length: 38 bytes */
-        0xae, 0x30, 0x31, 0xb1, 0xf1, 0xb9, 0x6f, 0xda, 0x17, 0x19, 0xd9, 0x57,
-        0xa9, 0x9d, 0x5c, 0x51, 0x9b, 0x53, 0x63, 0xa5, 0x24, 0x70, 0xa1,
-        0xae, 0xdf, 0x1c, 0xb9, 0xfc, 0xe3, 0xd7, 0x77, 0x6d, 0xb6, 0x89, 0x0f,
-        0x03, 0x18, 0x72
-    };
-
-    fd = wolfSSL_get_wfd(ssl);
-    AssertIntGE(fd, 0);
-    ret = (int)send(fd, (MESSAGE_TYPE_CAST)bad_msg, sizeof(bad_msg), 0);
-    AssertIntEQ(ret, sizeof(bad_msg));
-    ret = wolfSSL_write(ssl, "badrecordtest", sizeof("badrecordtest"));
-    AssertIntEQ(ret, sizeof("badrecordtest"));
-}
-
-static void test_wolfSSL_read_string(WOLFSSL* ssl)
-{
-    byte buf[100];
-    int ret;
-
-    ret = wolfSSL_read(ssl, buf, sizeof(buf));
-    AssertIntGT(ret, 0);
-    AssertIntEQ(strcmp((char*)buf, "badrecordtest"), 0);
-}
-
-static int _test_wolfSSL_dtls_bad_record(
-    method_provider client_method, method_provider server_method)
-{
-    callback_functions client_cbs, server_cbs;
-
-    XMEMSET(&client_cbs, 0, sizeof(client_cbs));
-    XMEMSET(&server_cbs, 0, sizeof(server_cbs));
-    client_cbs.doUdp = server_cbs.doUdp = 1;
-    client_cbs.method = client_method;
-    server_cbs.method = server_method;
-
-    client_cbs.on_result = test_wolfSSL_send_bad_record;
-    server_cbs.on_result = test_wolfSSL_read_string;
-
-    test_wolfSSL_client_server_nofail(&client_cbs, &server_cbs);
-
-    if (!client_cbs.return_code)
-        return TEST_FAIL;
-    if (!server_cbs.return_code)
-        return TEST_FAIL;
-
-    return TEST_SUCCESS;
-}
-
-static int test_wolfSSL_dtls_bad_record(void)
-{
-    int ret = TEST_SUCCESS;
-#if !defined(WOLFSSL_NO_TLS12)
-    ret = _test_wolfSSL_dtls_bad_record(wolfDTLSv1_2_client_method,
-        wolfDTLSv1_2_server_method);
-#endif
-#ifdef WOLFSSL_DTLS13
-    if (ret == TEST_SUCCESS) {
-        ret = _test_wolfSSL_dtls_bad_record(wolfDTLSv1_3_client_method,
-        wolfDTLSv1_3_server_method);
-    }
-#endif /* WOLFSSL_DTLS13 */
-    return ret;
-
-}
-
-#else
-static int test_wolfSSL_dtls_fragments(void)
-{
-    return TEST_SKIPPED;
-}
-static int test_wolfSSL_ignore_alert_before_cookie(void)
-{
-    return TEST_SKIPPED;
-}
-static int test_wolfSSL_dtls_bad_record(void)
-{
-    return TEST_SKIPPED;
-}
-#endif
-
-#if defined(WOLFSSL_DTLS13) && !defined(WOLFSSL_TLS13_IGNORE_AEAD_LIMITS) && \
-    !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER) && \
-    defined(HAVE_IO_TESTS_DEPENDENCIES)
-static volatile int test_AEAD_seq_num = 0;
-#ifdef WOLFSSL_NO_ATOMICS
-static volatile int test_AEAD_done = 0;
-#else
-wolfSSL_Atomic_Int test_AEAD_done = WOLFSSL_ATOMIC_INITIALIZER(0);
-#endif
-#ifdef WOLFSSL_MUTEX_INITIALIZER
-static wolfSSL_Mutex test_AEAD_mutex = WOLFSSL_MUTEX_INITIALIZER(test_AEAD_mutex);
-#endif
-
-static int test_AEAD_fail_decryption = 0;
-static int test_AEAD_cbiorecv(WOLFSSL *ssl, char *buf, int sz, void *ctx)
-{
-    int fd = wolfSSL_get_fd(ssl);
-    int ret = -1;
-    if (fd >= 0 && (ret = (int)recv(fd, buf, sz, 0)) > 0) {
-        if (test_AEAD_fail_decryption) {
-            /* Modify the packet to trigger a decryption failure */
-            buf[ret/2] ^= 0xFF;
-            if (test_AEAD_fail_decryption == 1)
-                test_AEAD_fail_decryption = 0;
-        }
-    }
-    (void)ctx;
-    return ret;
-}
-
-static void test_AEAD_get_limits(WOLFSSL* ssl, w64wrapper* hardLimit,
-        w64wrapper* keyUpdateLimit, w64wrapper* sendLimit)
-{
-    if (sendLimit)
-        w64Zero(sendLimit);
-    switch (ssl->specs.bulk_cipher_algorithm) {
-        case wolfssl_aes_gcm:
-            if (sendLimit)
-                *sendLimit = AEAD_AES_LIMIT;
-            FALL_THROUGH;
-        case wolfssl_chacha:
-            if (hardLimit)
-                *hardLimit = DTLS_AEAD_AES_GCM_CHACHA_FAIL_LIMIT;
-            if (keyUpdateLimit)
-                *keyUpdateLimit = DTLS_AEAD_AES_GCM_CHACHA_FAIL_KU_LIMIT;
-            break;
-        case wolfssl_aes_ccm:
-            if (sendLimit)
-                *sendLimit = DTLS_AEAD_AES_CCM_LIMIT;
-            if (ssl->specs.aead_mac_size == AES_CCM_8_AUTH_SZ) {
-                if (hardLimit)
-                    *hardLimit = DTLS_AEAD_AES_CCM_8_FAIL_LIMIT;
-                if (keyUpdateLimit)
-                    *keyUpdateLimit = DTLS_AEAD_AES_CCM_8_FAIL_KU_LIMIT;
-            }
-            else {
-                if (hardLimit)
-                    *hardLimit = DTLS_AEAD_AES_CCM_FAIL_LIMIT;
-                if (keyUpdateLimit)
-                    *keyUpdateLimit = DTLS_AEAD_AES_CCM_FAIL_KU_LIMIT;
-            }
-            break;
-        default:
-            fprintf(stderr, "Unrecognized bulk cipher");
-            AssertFalse(1);
-            break;
-    }
-}
-
-static void test_AEAD_limit_client(WOLFSSL* ssl)
-{
-    int ret;
-    int i;
-    int didReKey = 0;
-    char msgBuf[20];
-    w64wrapper hardLimit;
-    w64wrapper keyUpdateLimit;
-    w64wrapper counter;
-    w64wrapper sendLimit;
-
-    test_AEAD_get_limits(ssl, &hardLimit, &keyUpdateLimit, &sendLimit);
-
-    w64Zero(&counter);
-    AssertTrue(w64Equal(Dtls13GetEpoch(ssl, ssl->dtls13Epoch)->dropCount, counter));
-
-    wolfSSL_SSLSetIORecv(ssl, test_AEAD_cbiorecv);
-
-    for (i = 0; i < 10; i++) {
-        /* Test some failed decryptions */
-        test_AEAD_fail_decryption = 1;
-        w64Increment(&counter);
-        ret = wolfSSL_read(ssl, msgBuf, sizeof(msgBuf));
-        /* Should succeed since decryption failures are dropped */
-        AssertIntGT(ret, 0);
-        AssertTrue(w64Equal(Dtls13GetEpoch(ssl, ssl->dtls13PeerEpoch)->dropCount, counter));
-    }
-
-    test_AEAD_fail_decryption = 1;
-    Dtls13GetEpoch(ssl, ssl->dtls13PeerEpoch)->dropCount = keyUpdateLimit;
-    w64Increment(&Dtls13GetEpoch(ssl, ssl->dtls13PeerEpoch)->dropCount);
-    /* 100 read calls should be enough to complete the key update */
-    w64Zero(&counter);
-    for (i = 0; i < 100; i++) {
-        /* Key update should be sent and negotiated */
-        ret = wolfSSL_read(ssl, msgBuf, sizeof(msgBuf));
-        AssertIntGT(ret, 0);
-        /* Epoch after one key update is 4 */
-        if (w64Equal(ssl->dtls13PeerEpoch, w64From32(0, 4)) &&
-                w64Equal(Dtls13GetEpoch(ssl, ssl->dtls13PeerEpoch)->dropCount, counter)) {
-            didReKey = 1;
-            break;
-        }
-    }
-    AssertTrue(didReKey);
-
-    if (!w64IsZero(sendLimit)) {
-        /* Test the sending limit for AEAD ciphers */
-#ifdef WOLFSSL_MUTEX_INITIALIZER
-        (void)wc_LockMutex(&test_AEAD_mutex);
-#endif
-        Dtls13GetEpoch(ssl, ssl->dtls13Epoch)->nextSeqNumber = sendLimit;
-        test_AEAD_seq_num = 1;
-        XMEMSET(msgBuf, 0, sizeof(msgBuf));
-        ret = wolfSSL_write(ssl, msgBuf, sizeof(msgBuf));
-        AssertIntGT(ret, 0);
-        didReKey = 0;
-        w64Zero(&counter);
-#ifdef WOLFSSL_MUTEX_INITIALIZER
-        wc_UnLockMutex(&test_AEAD_mutex);
-#endif
-        /* 100 read calls should be enough to complete the key update */
-        for (i = 0; i < 100; i++) {
-            /* Key update should be sent and negotiated */
-            ret = wolfSSL_read(ssl, msgBuf, sizeof(msgBuf));
-            AssertIntGT(ret, 0);
-            /* Epoch after another key update is 5 */
-            if (w64Equal(ssl->dtls13Epoch, w64From32(0, 5)) &&
-                    w64Equal(Dtls13GetEpoch(ssl, ssl->dtls13Epoch)->dropCount, counter)) {
-                didReKey = 1;
-                break;
-            }
-        }
-        AssertTrue(didReKey);
-    }
-
-    test_AEAD_fail_decryption = 2;
-    Dtls13GetEpoch(ssl, ssl->dtls13PeerEpoch)->dropCount = hardLimit;
-    w64Decrement(&Dtls13GetEpoch(ssl, ssl->dtls13PeerEpoch)->dropCount);
-    /* Connection should fail with a DECRYPT_ERROR */
-    ret = wolfSSL_read(ssl, msgBuf, sizeof(msgBuf));
-    AssertIntEQ(ret, WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR));
-    AssertIntEQ(wolfSSL_get_error(ssl, ret), WC_NO_ERR_TRACE(DECRYPT_ERROR));
-
-#ifdef WOLFSSL_ATOMIC_INITIALIZER
-    WOLFSSL_ATOMIC_STORE(test_AEAD_done, 1);
-#else
-    test_AEAD_done = 1;
-#endif
-}
-
-int counter = 0;
-static void test_AEAD_limit_server(WOLFSSL* ssl)
-{
-    char msgBuf[] = "Sending data";
-    int ret = WOLFSSL_SUCCESS;
-    w64wrapper sendLimit;
-    SOCKET_T fd = wolfSSL_get_fd(ssl);
-    struct timespec delay;
-    XMEMSET(&delay, 0, sizeof(delay));
-    delay.tv_nsec = 100000000; /* wait 0.1 seconds */
-    tcp_set_nonblocking(&fd); /* So that read doesn't block */
-    wolfSSL_dtls_set_using_nonblock(ssl, 1);
-    test_AEAD_get_limits(ssl, NULL, NULL, &sendLimit);
-    while (!
-    #ifdef WOLFSSL_ATOMIC_INITIALIZER
-           WOLFSSL_ATOMIC_LOAD(test_AEAD_done)
-    #else
-           test_AEAD_done
-    #endif
-           && ret > 0)
-    {
-        counter++;
-#ifdef WOLFSSL_MUTEX_INITIALIZER
-        (void)wc_LockMutex(&test_AEAD_mutex);
-#endif
-        if (test_AEAD_seq_num) {
-            /* We need to update the seq number so that we can understand the
-             * peer. Otherwise we will incorrectly interpret the seq number. */
-            Dtls13Epoch* e = Dtls13GetEpoch(ssl, ssl->dtls13PeerEpoch);
-            AssertNotNull(e);
-            e->nextPeerSeqNumber = sendLimit;
-            test_AEAD_seq_num = 0;
-        }
-#ifdef WOLFSSL_MUTEX_INITIALIZER
-        wc_UnLockMutex(&test_AEAD_mutex);
-#endif
-        (void)wolfSSL_read(ssl, msgBuf, sizeof(msgBuf));
-        ret = wolfSSL_write(ssl, msgBuf, sizeof(msgBuf));
-        nanosleep(&delay, NULL);
-    }
-}
-
-static int test_wolfSSL_dtls_AEAD_limit(void)
-{
-    callback_functions func_cb_client;
-    callback_functions func_cb_server;
-    XMEMSET(&func_cb_client, 0, sizeof(callback_functions));
-    XMEMSET(&func_cb_server, 0, sizeof(callback_functions));
-
-    func_cb_client.doUdp = func_cb_server.doUdp = 1;
-    func_cb_server.method = wolfDTLSv1_3_server_method;
-    func_cb_client.method = wolfDTLSv1_3_client_method;
-    func_cb_server.on_result = test_AEAD_limit_server;
-    func_cb_client.on_result = test_AEAD_limit_client;
-
-    test_wolfSSL_client_server_nofail(&func_cb_client, &func_cb_server);
-
-    if (!func_cb_client.return_code)
-        return TEST_FAIL;
-    if (!func_cb_server.return_code)
-        return TEST_FAIL;
-
-    return TEST_SUCCESS;
-}
-#else
-static int test_wolfSSL_dtls_AEAD_limit(void)
-{
-    return TEST_SKIPPED;
-}
-#endif
-
-#if defined(WOLFSSL_DTLS) && \
-    defined(HAVE_IO_TESTS_DEPENDENCIES) && !defined(SINGLE_THREADED) && \
-    !defined(DEBUG_VECTOR_REGISTER_ACCESS_FUZZING)
-static void test_wolfSSL_dtls_send_ch(WOLFSSL* ssl)
-{
-    int fd, ret;
-    byte ch_msg[] = {
-        0x16, 0xfe, 0xfd, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01,
-        0xfa, 0x01, 0x00, 0x01, 0xee, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01,
-        0xee, 0xfe, 0xfd, 0xc0, 0xca, 0xb5, 0x6f, 0x3d, 0x23, 0xcc, 0x53, 0x9a,
-        0x67, 0x17, 0x70, 0xd3, 0xfb, 0x23, 0x16, 0x9e, 0x4e, 0xd6, 0x7e, 0x29,
-        0xab, 0xfa, 0x4c, 0xa5, 0x84, 0x95, 0xc3, 0xdb, 0x21, 0x9a, 0x52, 0x00,
-        0x00, 0x00, 0x36, 0x13, 0x01, 0x13, 0x02, 0x13, 0x03, 0xc0, 0x2c, 0xc0,
-        0x2b, 0xc0, 0x30, 0xc0, 0x2f, 0x00, 0x9f, 0x00, 0x9e, 0xcc, 0xa9, 0xcc,
-        0xa8, 0xcc, 0xaa, 0xc0, 0x27, 0xc0, 0x23, 0xc0, 0x28, 0xc0, 0x24, 0xc0,
-        0x0a, 0xc0, 0x09, 0xc0, 0x14, 0xc0, 0x13, 0x00, 0x6b, 0x00, 0x67, 0x00,
-        0x39, 0x00, 0x33, 0xcc, 0x14, 0xcc, 0x13, 0xcc, 0x15, 0x01, 0x00, 0x01,
-        0x8e, 0x00, 0x2b, 0x00, 0x03, 0x02, 0xfe, 0xfc, 0x00, 0x0d, 0x00, 0x20,
-        0x00, 0x1e, 0x06, 0x03, 0x05, 0x03, 0x04, 0x03, 0x02, 0x03, 0x08, 0x06,
-        0x08, 0x0b, 0x08, 0x05, 0x08, 0x0a, 0x08, 0x04, 0x08, 0x09, 0x06, 0x01,
-        0x05, 0x01, 0x04, 0x01, 0x03, 0x01, 0x02, 0x01, 0x00, 0x0a, 0x00, 0x0c,
-        0x00, 0x0a, 0x00, 0x19, 0x00, 0x18, 0x00, 0x17, 0x00, 0x15, 0x01, 0x00,
-        0x00, 0x16, 0x00, 0x00, 0x00, 0x33, 0x01, 0x4b, 0x01, 0x49, 0x00, 0x17,
-        0x00, 0x41, 0x04, 0x96, 0xcb, 0x2e, 0x4e, 0xd9, 0x88, 0x71, 0xc7, 0xf3,
-        0x1a, 0x16, 0xdd, 0x7a, 0x7c, 0xf7, 0x67, 0x8a, 0x5d, 0x9a, 0x55, 0xa6,
-        0x4a, 0x90, 0xd9, 0xfb, 0xc7, 0xfb, 0xbe, 0x09, 0xa9, 0x8a, 0xb5, 0x7a,
-        0xd1, 0xde, 0x83, 0x74, 0x27, 0x31, 0x1c, 0xaa, 0xae, 0xef, 0x58, 0x43,
-        0x13, 0x7d, 0x15, 0x4d, 0x7f, 0x68, 0xf6, 0x8a, 0x38, 0xef, 0x0e, 0xb3,
-        0xcf, 0xb8, 0x4a, 0xa9, 0xb4, 0xd7, 0xcb, 0x01, 0x00, 0x01, 0x00, 0x1d,
-        0x0a, 0x22, 0x8a, 0xd1, 0x78, 0x85, 0x1e, 0x5a, 0xe1, 0x1d, 0x1e, 0xb7,
-        0x2d, 0xbc, 0x5f, 0x52, 0xbc, 0x97, 0x5d, 0x8b, 0x6a, 0x8b, 0x9d, 0x1e,
-        0xb1, 0xfc, 0x8a, 0xb2, 0x56, 0xcd, 0xed, 0x4b, 0xfb, 0x66, 0x3f, 0x59,
-        0x3f, 0x15, 0x5d, 0x09, 0x9e, 0x2f, 0x60, 0x5b, 0x31, 0x81, 0x27, 0xf0,
-        0x1c, 0xda, 0xcd, 0x48, 0x66, 0xc6, 0xbb, 0x25, 0xf0, 0x5f, 0xda, 0x4c,
-        0xcf, 0x1d, 0x88, 0xc8, 0xda, 0x1b, 0x53, 0xea, 0xbd, 0xce, 0x6d, 0xf6,
-        0x4a, 0x76, 0xdb, 0x75, 0x99, 0xaf, 0xcf, 0x76, 0x4a, 0xfb, 0xe3, 0xef,
-        0xb2, 0xcb, 0xae, 0x4a, 0xc0, 0xe8, 0x63, 0x1f, 0xd6, 0xe8, 0xe6, 0x45,
-        0xf9, 0xea, 0x0d, 0x06, 0x19, 0xfc, 0xb1, 0xfd, 0x5d, 0x92, 0x89, 0x7b,
-        0xc7, 0x9f, 0x1a, 0xb3, 0x2b, 0xc7, 0xad, 0x0e, 0xfb, 0x13, 0x41, 0x83,
-        0x84, 0x58, 0x3a, 0x25, 0xb9, 0x49, 0x35, 0x1c, 0x23, 0xcb, 0xd6, 0xe7,
-        0xc2, 0x8c, 0x4b, 0x2a, 0x73, 0xa1, 0xdf, 0x4f, 0x73, 0x9b, 0xb3, 0xd2,
-        0xb2, 0x95, 0x00, 0x3c, 0x26, 0x09, 0x89, 0x71, 0x05, 0x39, 0xc8, 0x98,
-        0x8f, 0xed, 0x32, 0x15, 0x78, 0xcd, 0xd3, 0x7e, 0xfb, 0x5a, 0x78, 0x2a,
-        0xdc, 0xca, 0x20, 0x09, 0xb5, 0x14, 0xf9, 0xd4, 0x58, 0xf6, 0x69, 0xf8,
-        0x65, 0x9f, 0xb7, 0xe4, 0x93, 0xf1, 0xa3, 0x84, 0x7e, 0x1b, 0x23, 0x5d,
-        0xea, 0x59, 0x3e, 0x4d, 0xca, 0xfd, 0xa5, 0x55, 0xdd, 0x99, 0xb5, 0x02,
-        0xf8, 0x0d, 0xe5, 0xf4, 0x06, 0xb0, 0x43, 0x9e, 0x2e, 0xbf, 0x05, 0x33,
-        0x65, 0x7b, 0x13, 0x8c, 0xf9, 0x16, 0x4d, 0xc5, 0x15, 0x0b, 0x40, 0x2f,
-        0x66, 0x94, 0xf2, 0x43, 0x95, 0xe7, 0xa9, 0xb6, 0x39, 0x99, 0x73, 0xb3,
-        0xb0, 0x06, 0xfe, 0x52, 0x9e, 0x57, 0xba, 0x75, 0xfd, 0x76, 0x7b, 0x20,
-        0x31, 0x68, 0x4c
-    };
-
-    fd = wolfSSL_get_wfd(ssl);
-    AssertIntGE(fd, 0);
-    ret = (int)send(fd, (MESSAGE_TYPE_CAST)ch_msg, sizeof(ch_msg), 0);
-    AssertIntGT(ret, 0);
-    /* consume the HRR otherwise handshake will fail */
-    ret = (int)recv(fd, (MESSAGE_TYPE_CAST)ch_msg, sizeof(ch_msg), 0);
-    AssertIntGT(ret, 0);
-}
-
-#if defined(WOLFSSL_DTLS13) && defined(WOLFSSL_SEND_HRR_COOKIE)
-static void test_wolfSSL_dtls_send_ch_with_invalid_cookie(WOLFSSL* ssl)
-{
-    int fd, ret;
-    byte ch_msh_invalid_cookie[] = {
-      0x16, 0xfe, 0xfd, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x02,
-      0x4e, 0x01, 0x00, 0x02, 0x42, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x02,
-      0x42, 0xfe, 0xfd, 0x69, 0xca, 0x77, 0x60, 0x6f, 0xfc, 0xd1, 0x5b, 0x60,
-      0x5d, 0xf1, 0xa6, 0x5c, 0x44, 0x71, 0xae, 0xca, 0x62, 0x19, 0x0c, 0xb6,
-      0xf7, 0x2c, 0xa6, 0xd5, 0xd2, 0x99, 0x9d, 0x18, 0xae, 0xac, 0x11, 0x00,
-      0x00, 0x00, 0x36, 0x13, 0x01, 0x13, 0x02, 0x13, 0x03, 0xc0, 0x2c, 0xc0,
-      0x2b, 0xc0, 0x30, 0xc0, 0x2f, 0x00, 0x9f, 0x00, 0x9e, 0xcc, 0xa9, 0xcc,
-      0xa8, 0xcc, 0xaa, 0xc0, 0x27, 0xc0, 0x23, 0xc0, 0x28, 0xc0, 0x24, 0xc0,
-      0x0a, 0xc0, 0x09, 0xc0, 0x14, 0xc0, 0x13, 0x00, 0x6b, 0x00, 0x67, 0x00,
-      0x39, 0x00, 0x33, 0xcc, 0x14, 0xcc, 0x13, 0xcc, 0x15, 0x01, 0x00, 0x01,
-      0xe2, 0x00, 0x2b, 0x00, 0x03, 0x02, 0xfe, 0xfc, 0x00, 0x0d, 0x00, 0x20,
-      0x00, 0x1e, 0x06, 0x03, 0x05, 0x03, 0x04, 0x03, 0x02, 0x03, 0x08, 0x06,
-      0x08, 0x0b, 0x08, 0x05, 0x08, 0x0a, 0x08, 0x04, 0x08, 0x09, 0x06, 0x01,
-      0x05, 0x01, 0x04, 0x01, 0x03, 0x01, 0x02, 0x01, 0x00, 0x2c, 0x00, 0x45,
-      0x00, 0x43, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
-      0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
-      0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
-      0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
-      0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
-      0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00, 0x2d, 0x00,
-      0x03, 0x02, 0x00, 0x01, 0x00, 0x0a, 0x00, 0x0c, 0x00, 0x0a, 0x00, 0x19,
-      0x00, 0x18, 0x00, 0x17, 0x00, 0x15, 0x01, 0x00, 0x00, 0x16, 0x00, 0x00,
-      0x00, 0x33, 0x01, 0x4b, 0x01, 0x49, 0x00, 0x17, 0x00, 0x41, 0x04, 0x7c,
-      0x5a, 0xc2, 0x5a, 0xfd, 0xcd, 0x2b, 0x08, 0xb2, 0xeb, 0x8e, 0xc0, 0x02,
-      0x03, 0x9d, 0xb1, 0xc1, 0x0d, 0x7b, 0x7f, 0x46, 0x43, 0xdf, 0xf3, 0xee,
-      0x2b, 0x78, 0x0e, 0x29, 0x8c, 0x42, 0x11, 0x2c, 0xde, 0xd7, 0x41, 0x0f,
-      0x28, 0x94, 0x80, 0x41, 0x70, 0xc4, 0x17, 0xfd, 0x6d, 0xfa, 0xee, 0x9a,
-      0xf2, 0xc4, 0x15, 0x4c, 0x5f, 0x54, 0xb6, 0x78, 0x6e, 0xf9, 0x63, 0x27,
-      0x33, 0xb8, 0x7b, 0x01, 0x00, 0x01, 0x00, 0xd4, 0x46, 0x62, 0x9c, 0xbf,
-      0x8f, 0x1b, 0x65, 0x9b, 0xf0, 0x29, 0x64, 0xd8, 0x50, 0x0e, 0x74, 0xf1,
-      0x58, 0x10, 0xc9, 0xd9, 0x82, 0x5b, 0xd9, 0xbe, 0x14, 0xdf, 0xde, 0x86,
-      0xb4, 0x2e, 0x15, 0xee, 0x4f, 0xf6, 0x74, 0x9e, 0x59, 0x11, 0x36, 0x2d,
-      0xb9, 0x67, 0xaa, 0x5a, 0x09, 0x9b, 0x45, 0xf1, 0x01, 0x4c, 0x4e, 0xf6,
-      0xda, 0x6a, 0xae, 0xa7, 0x73, 0x7b, 0x2e, 0xb6, 0x24, 0x89, 0x99, 0xb7,
-      0x52, 0x16, 0x62, 0x0a, 0xab, 0x58, 0xf8, 0x3f, 0x10, 0x5b, 0x83, 0xfd,
-      0x7b, 0x81, 0x77, 0x81, 0x8d, 0xef, 0x24, 0x56, 0x6d, 0xba, 0x49, 0xd4,
-      0x8b, 0xb5, 0xa0, 0xb1, 0xc9, 0x8c, 0x32, 0x95, 0x1c, 0x5e, 0x0a, 0x4b,
-      0xf6, 0x00, 0x50, 0x0a, 0x87, 0x99, 0x59, 0xcf, 0x6f, 0x9d, 0x02, 0xd0,
-      0x1b, 0xa1, 0x96, 0x45, 0x28, 0x76, 0x40, 0x33, 0x28, 0xc9, 0xa1, 0xfd,
-      0x46, 0xab, 0x2c, 0x9e, 0x5e, 0xc6, 0x74, 0x19, 0x9a, 0xf5, 0x9b, 0x51,
-      0x11, 0x4f, 0xc8, 0xb9, 0x99, 0x6b, 0x4e, 0x3e, 0x31, 0x64, 0xb4, 0x92,
-      0xf4, 0x0d, 0x41, 0x4b, 0x2c, 0x65, 0x23, 0xf7, 0x47, 0xe3, 0xa5, 0x2e,
-      0xe4, 0x9c, 0x2b, 0xc9, 0x41, 0x22, 0x83, 0x8a, 0x23, 0xef, 0x29, 0x7e,
-      0x4f, 0x3f, 0xa3, 0xbf, 0x73, 0x2b, 0xd7, 0xcc, 0xc8, 0xc6, 0xe9, 0xbc,
-      0x01, 0xb7, 0x32, 0x63, 0xd4, 0x7e, 0x7f, 0x9a, 0xaf, 0x5f, 0x05, 0x31,
-      0x53, 0xd6, 0x1f, 0xa2, 0xd0, 0xdf, 0x67, 0x56, 0xf1, 0x9c, 0x4a, 0x9d,
-      0x83, 0xb4, 0xef, 0xb3, 0xf2, 0xcc, 0xf1, 0x91, 0x6c, 0x47, 0xc3, 0x8b,
-      0xd0, 0x92, 0x79, 0x3d, 0xa0, 0xc0, 0x3a, 0x57, 0x26, 0x6d, 0x0a, 0xad,
-      0x5f, 0xad, 0xb4, 0x74, 0x48, 0x4a, 0x51, 0xe1, 0xb5, 0x82, 0x0a, 0x4c,
-      0x4f, 0x9d, 0xaf, 0xee, 0x5a, 0xa2, 0x4d, 0x4d, 0x5f, 0xe0, 0x17, 0x00,
-      0x23, 0x00, 0x00
-    };
-    byte alert_reply[50];
-    byte expected_alert_reply[] = {
-        0x15, 0xfe, 0xfd, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00,
-        0x02, 0x02, 0x2f
-    };
-
-    fd = wolfSSL_get_wfd(ssl);
-    if (fd >= 0) {
-        ret = (int)send(fd, (MESSAGE_TYPE_CAST)ch_msh_invalid_cookie,
-                sizeof(ch_msh_invalid_cookie), 0);
-        AssertIntGT(ret, 0);
-        /* should reply with an illegal_parameter reply */
-        ret = (int)recv(fd, (MESSAGE_TYPE_CAST)alert_reply, sizeof(alert_reply), 0);
-        AssertIntEQ(ret, sizeof(expected_alert_reply));
-        AssertIntEQ(XMEMCMP(alert_reply, expected_alert_reply,
-                sizeof(expected_alert_reply)), 0);
-    }
-}
-#endif
-
-static word32 test_wolfSSL_dtls_stateless_HashWOLFSSL(const WOLFSSL* ssl)
-{
-#ifndef NO_MD5
-    enum wc_HashType hashType = WC_HASH_TYPE_MD5;
-#elif !defined(NO_SHA)
-    enum wc_HashType hashType = WC_HASH_TYPE_SHA;
-#elif !defined(NO_SHA256)
-    enum wc_HashType hashType = WC_HASH_TYPE_SHA256;
-#else
-    #error "We need a digest to hash the WOLFSSL object"
-#endif
-    byte hashBuf[WC_MAX_DIGEST_SIZE];
-    wc_HashAlg hash;
-    const TLSX* exts = ssl->extensions;
-    WOLFSSL sslCopy; /* Use a copy to omit certain fields */
-#ifndef WOLFSSL_SMALL_STACK_CACHE
-    HS_Hashes* hsHashes = ssl->hsHashes; /* Is re-allocated in
-                                          * InitHandshakeHashes */
-#endif
-
-    XMEMCPY(&sslCopy, ssl, sizeof(*ssl));
-    XMEMSET(hashBuf, 0, sizeof(hashBuf));
-
-    /* Following fields are not important to compare */
-    XMEMSET(sslCopy.buffers.inputBuffer.staticBuffer, 0, STATIC_BUFFER_LEN);
-    sslCopy.buffers.inputBuffer.buffer = NULL;
-    sslCopy.buffers.inputBuffer.bufferSize = 0;
-    sslCopy.buffers.inputBuffer.dynamicFlag = 0;
-    sslCopy.buffers.inputBuffer.offset = 0;
-    XMEMSET(sslCopy.buffers.outputBuffer.staticBuffer, 0, STATIC_BUFFER_LEN);
-    sslCopy.buffers.outputBuffer.buffer = NULL;
-    sslCopy.buffers.outputBuffer.bufferSize = 0;
-    sslCopy.buffers.outputBuffer.dynamicFlag = 0;
-    sslCopy.buffers.outputBuffer.offset = 0;
-    sslCopy.error = 0;
-    sslCopy.curSize = 0;
-    sslCopy.curStartIdx = 0;
-    sslCopy.keys.curSeq_lo = 0;
-    XMEMSET(&sslCopy.curRL, 0, sizeof(sslCopy.curRL));
-#ifdef WOLFSSL_DTLS13
-    XMEMSET(&sslCopy.keys.curSeq, 0, sizeof(sslCopy.keys.curSeq));
-    sslCopy.dtls13FastTimeout = 0;
-#endif
-    sslCopy.keys.dtls_peer_handshake_number = 0;
-    XMEMSET(&sslCopy.alert_history, 0, sizeof(sslCopy.alert_history));
-    sslCopy.hsHashes = NULL;
-#if !defined(WOLFSSL_NO_CLIENT_AUTH) && \
-    ((defined(WOLFSSL_SM2) && defined(WOLFSSL_SM3)) || \
-     (defined(HAVE_ED25519) && !defined(NO_ED25519_CLIENT_AUTH)) || \
-     (defined(HAVE_ED448) && !defined(NO_ED448_CLIENT_AUTH)))
-    sslCopy.options.cacheMessages = 0;
-#endif
-#ifdef WOLFSSL_ASYNC_IO
-#ifdef WOLFSSL_ASYNC_CRYPT
-    sslCopy.asyncDev = NULL;
-#endif
-    sslCopy.async = NULL;
-#endif
-
-    AssertIntEQ(wc_HashInit(&hash, hashType), 0);
-    AssertIntEQ(wc_HashUpdate(&hash, hashType, (byte*)&sslCopy, sizeof(sslCopy)), 0);
-    /* hash extension list */
-    while (exts != NULL) {
-        AssertIntEQ(wc_HashUpdate(&hash, hashType, (byte*)exts, sizeof(*exts)), 0);
-        exts = exts->next;
-    }
-    /* Hash suites */
-    if (sslCopy.suites != NULL) {
-        AssertIntEQ(wc_HashUpdate(&hash, hashType, (byte*)sslCopy.suites,
-                sizeof(struct Suites)), 0);
-    }
-
-#ifdef WOLFSSL_SMALL_STACK_CACHE
-    /* with WOLFSSL_SMALL_STACK_CACHE, the SHA-2 objects always differ after
-     * initialization because of cached W and (for SHA512) X buffers.
-     */
-#else
-    /* Hash hsHashes */
-    AssertIntEQ(wc_HashUpdate(&hash, hashType, (byte*)hsHashes,
-            sizeof(*hsHashes)), 0);
-#endif
-
-    AssertIntEQ(wc_HashFinal(&hash, hashType, hashBuf), 0);
-    AssertIntEQ(wc_HashFree(&hash, hashType), 0);
-
-    return MakeWordFromHash(hashBuf);
-}
-
-static CallbackIORecv test_wolfSSL_dtls_compare_stateless_cb;
-static int test_wolfSSL_dtls_compare_stateless_cb_call_once;
-static int test_wolfSSL_dtls_compare_stateless_read_cb_once(WOLFSSL *ssl,
-        char *buf, int sz, void *ctx)
-{
-    if (test_wolfSSL_dtls_compare_stateless_cb_call_once) {
-        test_wolfSSL_dtls_compare_stateless_cb_call_once = 0;
-        return test_wolfSSL_dtls_compare_stateless_cb(ssl, buf, sz, ctx);
-    }
-    else {
-        return WOLFSSL_CBIO_ERR_WANT_READ;
-    }
-}
-
-static void test_wolfSSL_dtls_compare_stateless(WOLFSSL* ssl)
-{
-    /* Compare the ssl object before and after one ClientHello msg */
-    SOCKET_T fd = wolfSSL_get_fd(ssl);
-    int res;
-    int err;
-    word32 initHash;
-
-    test_wolfSSL_dtls_compare_stateless_cb = ssl->CBIORecv;
-    test_wolfSSL_dtls_compare_stateless_cb_call_once = 1;
-    wolfSSL_dtls_set_using_nonblock(ssl, 1);
-    ssl->CBIORecv = test_wolfSSL_dtls_compare_stateless_read_cb_once;
-
-    initHash = test_wolfSSL_dtls_stateless_HashWOLFSSL(ssl);
-    (void)initHash;
-
-    res = tcp_select(fd, 5);
-    /* We are expecting a msg. A timeout indicates failure. */
-    AssertIntEQ(res, TEST_RECV_READY);
-
-    res = wolfSSL_accept(ssl);
-    err = wolfSSL_get_error(ssl, res);
-    AssertIntEQ(res, WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR));
-    AssertIntEQ(err, WOLFSSL_ERROR_WANT_READ);
-
-    AssertIntEQ(initHash, test_wolfSSL_dtls_stateless_HashWOLFSSL(ssl));
-
-    wolfSSL_dtls_set_using_nonblock(ssl, 0);
-    ssl->CBIORecv = test_wolfSSL_dtls_compare_stateless_cb;
-
-}
-
-#if defined(WOLFSSL_DTLS13) && defined(WOLFSSL_SEND_HRR_COOKIE)
-static void test_wolfSSL_dtls_enable_hrrcookie(WOLFSSL* ssl)
-{
-    int ret;
-    ret = wolfSSL_send_hrr_cookie(ssl, NULL, 0);
-    AssertIntEQ(ret, WOLFSSL_SUCCESS);
-    test_wolfSSL_dtls_compare_stateless(ssl);
-}
-#endif
-
-static int test_wolfSSL_dtls_stateless(void)
-{
-    callback_functions client_cbs, server_cbs;
-    size_t i;
-    struct {
-        method_provider client_meth;
-        method_provider server_meth;
-        ssl_callback client_ssl_ready;
-        ssl_callback server_ssl_ready;
-    } test_params[] = {
-#if !defined(WOLFSSL_NO_TLS12)
-        {wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method,
-                test_wolfSSL_dtls_send_ch, test_wolfSSL_dtls_compare_stateless},
-#endif
-#if defined(WOLFSSL_DTLS13) && defined(WOLFSSL_SEND_HRR_COOKIE)
-        {wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method,
-                test_wolfSSL_dtls_send_ch, test_wolfSSL_dtls_enable_hrrcookie},
-        {wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method,
-                test_wolfSSL_dtls_send_ch_with_invalid_cookie, test_wolfSSL_dtls_enable_hrrcookie},
-#endif
-    };
-
-    if (0 == sizeof(test_params)){
-        return TEST_SKIPPED;
-    }
-
-    for (i = 0; i < sizeof(test_params)/sizeof(*test_params); i++) {
-        XMEMSET(&client_cbs, 0, sizeof(client_cbs));
-        XMEMSET(&server_cbs, 0, sizeof(server_cbs));
-        client_cbs.doUdp = server_cbs.doUdp = 1;
-        client_cbs.method = test_params[i].client_meth;
-        server_cbs.method = test_params[i].server_meth;
-
-        client_cbs.ssl_ready = test_params[i].client_ssl_ready;
-        server_cbs.ssl_ready = test_params[i].server_ssl_ready;
-        test_wolfSSL_client_server_nofail(&client_cbs, &server_cbs);
-
-        if (!client_cbs.return_code)
-            return TEST_FAIL;
-        if (!server_cbs.return_code)
-            return TEST_FAIL;
-    }
-
-    return TEST_SUCCESS;
-}
-
-/* DTLS stateless API handling multiple CHs with different HRR groups */
-static int test_wolfSSL_dtls_stateless_hrr_group(void)
-{
-    EXPECT_DECLS;
-#if defined(WOLFSSL_SEND_HRR_COOKIE)
-    size_t i;
-    word32 initHash;
-    struct {
-        method_provider client_meth;
-        method_provider server_meth;
-    } params[] = {
-#if defined(WOLFSSL_TLS13) && defined(WOLFSSL_DTLS13)
-        { wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method },
-#endif
-#if !defined(WOLFSSL_NO_TLS12) && defined(WOLFSSL_DTLS)
-        { wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method },
-#endif
-    };
-    for (i = 0; i < XELEM_CNT(params) && !EXPECT_FAIL(); i++) {
-        WOLFSSL_CTX *ctx_s = NULL, *ctx_c = NULL;
-        WOLFSSL *ssl_s = NULL, *ssl_c = NULL, *ssl_c2 = NULL;
-        struct test_memio_ctx test_ctx;
-        int groups_1[] = {
-            WOLFSSL_ECC_SECP256R1,
-            WOLFSSL_ECC_SECP384R1,
-            WOLFSSL_ECC_SECP521R1
-        };
-        int groups_2[] = {
-            WOLFSSL_ECC_SECP384R1,
-            WOLFSSL_ECC_SECP521R1
-        };
-        char hrrBuf[1000];
-        int hrrSz = sizeof(hrrBuf);
-
-        XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-
-        ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-                params[i].client_meth, params[i].server_meth), 0);
-
-        ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, NULL, &ssl_c2, NULL,
-                params[i].client_meth, params[i].server_meth), 0);
-
-
-        wolfSSL_SetLoggingPrefix("server");
-        wolfSSL_dtls_set_using_nonblock(ssl_s, 1);
-
-        initHash = test_wolfSSL_dtls_stateless_HashWOLFSSL(ssl_s);
-
-        /* Set groups and disable key shares. This ensures that only the given
-         * groups are in the SupportedGroups extension and that an empty key
-         * share extension is sent in the initial ClientHello of each session.
-         * This triggers the server to send a HelloRetryRequest with the first
-         * group in the SupportedGroups extension selected. */
-        wolfSSL_SetLoggingPrefix("client1");
-        ExpectIntEQ(wolfSSL_set_groups(ssl_c, groups_1, 3), WOLFSSL_SUCCESS);
-        ExpectIntEQ(wolfSSL_NoKeyShares(ssl_c), WOLFSSL_SUCCESS);
-
-        wolfSSL_SetLoggingPrefix("client2");
-        ExpectIntEQ(wolfSSL_set_groups(ssl_c2, groups_2, 2), WOLFSSL_SUCCESS);
-        ExpectIntEQ(wolfSSL_NoKeyShares(ssl_c2), WOLFSSL_SUCCESS);
-
-        /* Start handshake, send first ClientHello */
-        wolfSSL_SetLoggingPrefix("client1");
-        ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
-        ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-
-        /* Read first ClientHello, send HRR with WOLFSSL_ECC_SECP256R1 */
-        wolfSSL_SetLoggingPrefix("server");
-        ExpectIntEQ(wolfDTLS_accept_stateless(ssl_s), 0);
-        ExpectIntEQ(test_memio_copy_message(&test_ctx, 1, hrrBuf, &hrrSz, 0), 0);
-        ExpectIntGT(hrrSz, 0);
-        ExpectIntEQ(initHash, test_wolfSSL_dtls_stateless_HashWOLFSSL(ssl_s));
-        test_memio_clear_buffer(&test_ctx, 1);
-
-        /* Send second ClientHello */
-        wolfSSL_SetLoggingPrefix("client2");
-        ExpectIntEQ(wolfSSL_connect(ssl_c2), -1);
-        ExpectIntEQ(wolfSSL_get_error(ssl_c2, -1), WOLFSSL_ERROR_WANT_READ);
-
-        /* Read second ClientHello, send HRR now with WOLFSSL_ECC_SECP384R1 */
-        wolfSSL_SetLoggingPrefix("server");
-        ExpectIntEQ(wolfDTLS_accept_stateless(ssl_s), 0);
-        ExpectIntEQ(initHash, test_wolfSSL_dtls_stateless_HashWOLFSSL(ssl_s));
-        test_memio_clear_buffer(&test_ctx, 1);
-
-        /* Complete first handshake with WOLFSSL_ECC_SECP256R1 */
-        wolfSSL_SetLoggingPrefix("client1");
-        ExpectIntEQ(test_memio_inject_message(&test_ctx, 1, hrrBuf, hrrSz), 0);
-        ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
-        ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-
-        wolfSSL_SetLoggingPrefix("server");
-        ExpectIntEQ(wolfDTLS_accept_stateless(ssl_s), WOLFSSL_SUCCESS);
-
-        ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-
-        wolfSSL_free(ssl_s);
-        wolfSSL_free(ssl_c);
-        wolfSSL_free(ssl_c2);
-        wolfSSL_CTX_free(ctx_s);
-        wolfSSL_CTX_free(ctx_c);
-    }
-#endif /* WOLFSSL_SEND_HRR_COOKIE */
-    return EXPECT_RESULT();
-}
-#else
-static int test_wolfSSL_dtls_stateless(void)
-{
-    return TEST_SKIPPED;
-}
-
-static int test_wolfSSL_dtls_stateless_hrr_group(void)
-{
-    return TEST_SKIPPED;
-}
-#endif /* WOLFSSL_DTLS13 && WOLFSSL_SEND_HRR_COOKIE &&
-        * HAVE_IO_TESTS_DEPENDENCIES && !SINGLE_THREADED */
 
 #if defined(HAVE_KEYING_MATERIAL) && defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES)
 static int test_export_keying_material_cb(WOLFSSL_CTX *ctx, WOLFSSL *ssl)
@@ -30447,92 +27733,6 @@ static int test_wolfSSL_CRYPTO_get_ex_new_index(void)
     return EXPECT_RESULT();
 }
 
-#if defined(HAVE_EX_DATA_CRYPTO) && defined(OPENSSL_EXTRA)
-
-#define SESSION_NEW_IDX_LONG 0xDEADBEEF
-#define SESSION_NEW_IDX_VAL  ((void*)0xAEADAEAD)
-#define SESSION_DUP_IDX_VAL  ((void*)0xDEDEDEDE)
-#define SESSION_NEW_IDX_PTR  "Testing"
-
-static void test_wolfSSL_SESSION_get_ex_new_index_new_cb(void* p, void* ptr,
-        CRYPTO_EX_DATA* a, int idx, long argValue, void* arg)
-{
-    AssertNotNull(p);
-    AssertNull(ptr);
-    AssertIntEQ(CRYPTO_set_ex_data(a, idx, SESSION_NEW_IDX_VAL), SSL_SUCCESS);
-    AssertIntEQ(argValue, SESSION_NEW_IDX_LONG);
-    AssertStrEQ(arg, SESSION_NEW_IDX_PTR);
-}
-
-static int test_wolfSSL_SESSION_get_ex_new_index_dup_cb(CRYPTO_EX_DATA* out,
-        const CRYPTO_EX_DATA* in, void* inPtr, int idx, long argV,
-        void* arg)
-{
-    EXPECT_DECLS;
-
-    ExpectNotNull(out);
-    ExpectNotNull(in);
-    ExpectPtrEq(*(void**)inPtr, SESSION_NEW_IDX_VAL);
-    ExpectPtrEq(CRYPTO_get_ex_data(in, idx), SESSION_NEW_IDX_VAL);
-    ExpectPtrEq(CRYPTO_get_ex_data(out, idx), SESSION_NEW_IDX_VAL);
-    ExpectIntEQ(argV, SESSION_NEW_IDX_LONG);
-    ExpectStrEQ(arg, SESSION_NEW_IDX_PTR);
-    *(void**)inPtr = SESSION_DUP_IDX_VAL;
-    if (EXPECT_SUCCESS()) {
-        return SSL_SUCCESS;
-    }
-    else {
-        return SSL_FAILURE;
-    }
-}
-
-static int test_wolfSSL_SESSION_get_ex_new_index_free_cb_called = 0;
-static void test_wolfSSL_SESSION_get_ex_new_index_free_cb(void* p, void* ptr,
-        CRYPTO_EX_DATA* a, int idx, long argValue, void* arg)
-{
-    EXPECT_DECLS;
-
-    ExpectNotNull(p);
-    ExpectNull(ptr);
-    ExpectPtrNE(CRYPTO_get_ex_data(a, idx), 0);
-    ExpectIntEQ(argValue, SESSION_NEW_IDX_LONG);
-    ExpectStrEQ(arg, SESSION_NEW_IDX_PTR);
-    if (EXPECT_SUCCESS()) {
-        test_wolfSSL_SESSION_get_ex_new_index_free_cb_called++;
-    }
-}
-
-static int test_wolfSSL_SESSION_get_ex_new_index(void)
-{
-    EXPECT_DECLS;
-    int idx = SSL_SESSION_get_ex_new_index(SESSION_NEW_IDX_LONG,
-                (void*)SESSION_NEW_IDX_PTR,
-                test_wolfSSL_SESSION_get_ex_new_index_new_cb,
-                test_wolfSSL_SESSION_get_ex_new_index_dup_cb,
-                test_wolfSSL_SESSION_get_ex_new_index_free_cb);
-    SSL_SESSION* s = SSL_SESSION_new();
-    SSL_SESSION* d = NULL;
-
-    ExpectNotNull(s);
-    ExpectPtrEq(SSL_SESSION_get_ex_data(s, idx), SESSION_NEW_IDX_VAL);
-    ExpectNotNull(d = SSL_SESSION_dup(s));
-    ExpectPtrEq(SSL_SESSION_get_ex_data(d, idx), SESSION_DUP_IDX_VAL);
-    SSL_SESSION_free(s);
-    ExpectIntEQ(test_wolfSSL_SESSION_get_ex_new_index_free_cb_called, 1);
-    SSL_SESSION_free(d);
-    ExpectIntEQ(test_wolfSSL_SESSION_get_ex_new_index_free_cb_called, 2);
-
-    crypto_ex_cb_free(crypto_ex_cb_ctx_session);
-    crypto_ex_cb_ctx_session = NULL;
-    return EXPECT_RESULT();
-}
-#else
-static int test_wolfSSL_SESSION_get_ex_new_index(void)
-{
-    return TEST_SKIPPED;
-}
-#endif
-
 static int test_wolfSSL_set_psk_use_session_callback(void)
 {
     EXPECT_DECLS;
@@ -32179,560 +29379,6 @@ static int test_wolfSSL_FIPS_mode(void)
     return EXPECT_RESULT();
 }
 
-#ifdef WOLFSSL_DTLS
-
-/* Prints out the current window */
-static void DUW_TEST_print_window_binary(word32 h, word32 l, word32* w) {
-#ifdef WOLFSSL_DEBUG_DTLS_WINDOW
-    int i;
-    for (i = WOLFSSL_DTLS_WINDOW_WORDS - 1; i >= 0; i--) {
-        word32 b = w[i];
-        int j;
-        /* Prints out a 32 bit binary number in big endian order */
-        for (j = 0; j < 32; j++, b <<= 1) {
-            if (b & (((word32)1) << 31))
-                fprintf(stderr, "1");
-            else
-                fprintf(stderr, "0");
-        }
-        fprintf(stderr, " ");
-    }
-    fprintf(stderr, "cur_hi %u cur_lo %u\n", h, l);
-#else
-    (void)h;
-    (void)l;
-    (void)w;
-#endif
-}
-
-/* a - cur_hi
- * b - cur_lo
- * c - next_hi
- * d - next_lo
- * e - window
- * f - expected next_hi
- * g - expected next_lo
- * h - expected window[1]
- * i - expected window[0]
- */
-#define DUW_TEST(a,b,c,d,e,f,g,h,i) do { \
-    ExpectIntEQ(wolfSSL_DtlsUpdateWindow((a), (b), &(c), &(d), (e)), 1); \
-    DUW_TEST_print_window_binary((a), (b), (e)); \
-    ExpectIntEQ((c), (f)); \
-    ExpectIntEQ((d), (g)); \
-    ExpectIntEQ((e)[1], (h)); \
-    ExpectIntEQ((e)[0], (i)); \
-} while (0)
-
-static int test_wolfSSL_DtlsUpdateWindow(void)
-{
-    EXPECT_DECLS;
-    word32 window[WOLFSSL_DTLS_WINDOW_WORDS];
-    word32 next_lo = 0;
-    word16 next_hi = 0;
-
-#ifdef WOLFSSL_DEBUG_DTLS_WINDOW
-    fprintf(stderr, "\n");
-#endif
-
-    XMEMSET(window, 0, sizeof window);
-    DUW_TEST(0, 0, next_hi, next_lo, window, 0, 1, 0, 0x01);
-    DUW_TEST(0, 1, next_hi, next_lo, window, 0, 2, 0, 0x03);
-    DUW_TEST(0, 5, next_hi, next_lo, window, 0, 6, 0, 0x31);
-    DUW_TEST(0, 4, next_hi, next_lo, window, 0, 6, 0, 0x33);
-    DUW_TEST(0, 100, next_hi, next_lo, window, 0, 101, 0, 0x01);
-    DUW_TEST(0, 101, next_hi, next_lo, window, 0, 102, 0, 0x03);
-    DUW_TEST(0, 133, next_hi, next_lo, window, 0, 134, 0x03, 0x01);
-    DUW_TEST(0, 200, next_hi, next_lo, window, 0, 201, 0, 0x01);
-    DUW_TEST(0, 264, next_hi, next_lo, window, 0, 265, 0, 0x01);
-    DUW_TEST(0, 0xFFFFFFFF, next_hi, next_lo, window, 1, 0, 0, 0x01);
-    DUW_TEST(0, 0xFFFFFFFD, next_hi, next_lo, window, 1, 0, 0, 0x05);
-    DUW_TEST(0, 0xFFFFFFFE, next_hi, next_lo, window, 1, 0, 0, 0x07);
-    DUW_TEST(1, 3, next_hi, next_lo, window, 1, 4, 0, 0x71);
-    DUW_TEST(1, 0, next_hi, next_lo, window, 1, 4, 0, 0x79);
-    DUW_TEST(1, 0xFFFFFFFF, next_hi, next_lo, window, 2, 0, 0, 0x01);
-    DUW_TEST(2, 3, next_hi, next_lo, window, 2, 4, 0, 0x11);
-    DUW_TEST(2, 0, next_hi, next_lo, window, 2, 4, 0, 0x19);
-    DUW_TEST(2, 25, next_hi, next_lo, window, 2, 26, 0, 0x6400001);
-    DUW_TEST(2, 27, next_hi, next_lo, window, 2, 28, 0, 0x19000005);
-    DUW_TEST(2, 29, next_hi, next_lo, window, 2, 30, 0, 0x64000015);
-    DUW_TEST(2, 33, next_hi, next_lo, window, 2, 34, 6, 0x40000151);
-    DUW_TEST(2, 60, next_hi, next_lo, window, 2, 61, 0x3200000A, 0x88000001);
-    DUW_TEST(1, 0xFFFFFFF0, next_hi, next_lo, window, 2, 61, 0x3200000A, 0x88000001);
-    DUW_TEST(2, 0xFFFFFFFD, next_hi, next_lo, window, 2, 0xFFFFFFFE, 0, 0x01);
-    DUW_TEST(3, 1, next_hi, next_lo, window, 3, 2, 0, 0x11);
-    DUW_TEST(99, 66, next_hi, next_lo, window, 99, 67, 0, 0x01);
-    DUW_TEST(50, 66, next_hi, next_lo, window, 99, 67, 0, 0x01);
-    DUW_TEST(100, 68, next_hi, next_lo, window, 100, 69, 0, 0x01);
-    DUW_TEST(99, 50, next_hi, next_lo, window, 100, 69, 0, 0x01);
-    DUW_TEST(99, 0xFFFFFFFF, next_hi, next_lo, window, 100, 69, 0, 0x01);
-    DUW_TEST(150, 0xFFFFFFFF, next_hi, next_lo, window, 151, 0, 0, 0x01);
-    DUW_TEST(152, 0xFFFFFFFF, next_hi, next_lo, window, 153, 0, 0, 0x01);
-
-    return EXPECT_RESULT();
-}
-#endif /* WOLFSSL_DTLS */
-
-#ifdef WOLFSSL_DTLS
-static int DFB_TEST(WOLFSSL* ssl, word32 seq, word32 len, word32 f_offset,
-        word32 f_len, word32 f_count, byte ready, word32 bytesReceived)
-{
-    EXPECT_DECLS;
-    DtlsMsg* cur = NULL;
-    static byte msg[100];
-    static byte msgInit = 0;
-
-    if (!msgInit) {
-        int i;
-        for (i = 0; i < 100; i++)
-            msg[i] = i + 1;
-        msgInit = 1;
-    }
-
-    /* Sanitize test parameters */
-    ExpectIntLE(len, sizeof(msg));
-    ExpectIntLE(f_offset + f_len, sizeof(msg));
-
-    if (EXPECT_SUCCESS())
-        DtlsMsgStore(ssl, 0, seq, msg + f_offset, len, certificate, f_offset, f_len, NULL);
-
-    ExpectNotNull(ssl->dtls_rx_msg_list);
-
-    ExpectNotNull(cur = DtlsMsgFind(ssl->dtls_rx_msg_list, 0, seq));
-    ExpectIntEQ(cur->fragBucketListCount, f_count);
-    ExpectIntEQ(cur->ready, ready);
-    ExpectIntEQ(cur->bytesReceived, bytesReceived);
-    if (ready) {
-        ExpectNull(cur->fragBucketList);
-        ExpectBufEQ(cur->fullMsg, msg, cur->sz);
-    }
-    else {
-        DtlsFragBucket* fb;
-        ExpectNotNull(cur->fragBucketList);
-        for (fb = cur != NULL ? cur->fragBucketList : NULL;
-                EXPECT_SUCCESS() && fb != NULL; fb = fb->m.m.next)
-            ExpectBufEQ(fb->buf, msg + fb->m.m.offset, fb->m.m.sz);
-    }
-    if (EXPECT_FAIL()) {
-        printf("Test parameters: seq %u len %u f_offset %u f_len %u f_count %u ready %u bytesReceived %u\n",
-                                 seq, len, f_offset, f_len, f_count, ready, bytesReceived);
-    }
-    return EXPECT_RESULT();
-}
-
-static int test_wolfSSL_DTLS_fragment_buckets(void)
-{
-    EXPECT_DECLS;
-    WOLFSSL ssl[1];
-
-    XMEMSET(ssl, 0, sizeof(*ssl));
-
-    EXPECT_TEST(DFB_TEST(ssl, 0, 100, 0, 100, 0, 1, 100)); /*  0-100 */
-
-    EXPECT_TEST(DFB_TEST(ssl, 1, 100,  0, 20, 1, 0,  20)); /*  0-20  */
-    EXPECT_TEST(DFB_TEST(ssl, 1, 100, 20, 20, 1, 0,  40)); /* 20-40  */
-    EXPECT_TEST(DFB_TEST(ssl, 1, 100, 40, 20, 1, 0,  60)); /* 40-60  */
-    EXPECT_TEST(DFB_TEST(ssl, 1, 100, 60, 20, 1, 0,  80)); /* 60-80  */
-    EXPECT_TEST(DFB_TEST(ssl, 1, 100, 80, 20, 0, 1, 100)); /* 80-100 */
-
-    /* Test all permutations of 3 regions */
-    /* 1 2 3 */
-    EXPECT_TEST(DFB_TEST(ssl, 2, 100,  0, 30, 1, 0,  30)); /*  0-30  */
-    EXPECT_TEST(DFB_TEST(ssl, 2, 100, 30, 30, 1, 0,  60)); /* 30-60  */
-    EXPECT_TEST(DFB_TEST(ssl, 2, 100, 60, 40, 0, 1, 100)); /* 60-100 */
-    /* 1 3 2 */
-    EXPECT_TEST(DFB_TEST(ssl, 3, 100,  0, 30, 1, 0,  30)); /*  0-30  */
-    EXPECT_TEST(DFB_TEST(ssl, 3, 100, 60, 40, 2, 0,  70)); /* 60-100 */
-    EXPECT_TEST(DFB_TEST(ssl, 3, 100, 30, 30, 0, 1, 100)); /* 30-60  */
-    /* 2 1 3 */
-    EXPECT_TEST(DFB_TEST(ssl, 4, 100, 30, 30, 1, 0,  30)); /* 30-60  */
-    EXPECT_TEST(DFB_TEST(ssl, 4, 100,  0, 30, 1, 0,  60)); /*  0-30  */
-    EXPECT_TEST(DFB_TEST(ssl, 4, 100, 60, 40, 0, 1, 100)); /* 60-100 */
-    /* 2 3 1 */
-    EXPECT_TEST(DFB_TEST(ssl, 5, 100, 30, 30, 1, 0,  30)); /* 30-60  */
-    EXPECT_TEST(DFB_TEST(ssl, 5, 100, 60, 40, 1, 0,  70)); /* 60-100 */
-    EXPECT_TEST(DFB_TEST(ssl, 5, 100,  0, 30, 0, 1, 100)); /*  0-30  */
-    /* 3 1 2 */
-    EXPECT_TEST(DFB_TEST(ssl, 6, 100, 60, 40, 1, 0,  40)); /* 60-100 */
-    EXPECT_TEST(DFB_TEST(ssl, 6, 100,  0, 30, 2, 0,  70)); /*  0-30  */
-    EXPECT_TEST(DFB_TEST(ssl, 6, 100, 30, 30, 0, 1, 100)); /* 30-60  */
-    /* 3 2 1 */
-    EXPECT_TEST(DFB_TEST(ssl, 7, 100, 60, 40, 1, 0,  40)); /* 60-100 */
-    EXPECT_TEST(DFB_TEST(ssl, 7, 100, 30, 30, 1, 0,  70)); /* 30-60  */
-    EXPECT_TEST(DFB_TEST(ssl, 7, 100,  0, 30, 0, 1, 100)); /*  0-30  */
-
-    /* Test overlapping regions */
-    EXPECT_TEST(DFB_TEST(ssl, 8, 100,  0, 30, 1, 0,  30)); /*  0-30  */
-    EXPECT_TEST(DFB_TEST(ssl, 8, 100, 20, 10, 1, 0,  30)); /* 20-30  */
-    EXPECT_TEST(DFB_TEST(ssl, 8, 100, 70, 10, 2, 0,  40)); /* 70-80  */
-    EXPECT_TEST(DFB_TEST(ssl, 8, 100, 20, 30, 2, 0,  60)); /* 20-50  */
-    EXPECT_TEST(DFB_TEST(ssl, 8, 100, 40, 60, 0, 1, 100)); /* 40-100 */
-
-    /* Test overlapping multiple regions */
-    EXPECT_TEST(DFB_TEST(ssl, 9, 100,  0, 20, 1, 0,  20)); /*  0-20  */
-    EXPECT_TEST(DFB_TEST(ssl, 9, 100, 30,  5, 2, 0,  25)); /* 30-35  */
-    EXPECT_TEST(DFB_TEST(ssl, 9, 100, 40,  5, 3, 0,  30)); /* 40-45  */
-    EXPECT_TEST(DFB_TEST(ssl, 9, 100, 50,  5, 4, 0,  35)); /* 50-55  */
-    EXPECT_TEST(DFB_TEST(ssl, 9, 100, 60,  5, 5, 0,  40)); /* 60-65  */
-    EXPECT_TEST(DFB_TEST(ssl, 9, 100, 70,  5, 6, 0,  45)); /* 70-75  */
-    EXPECT_TEST(DFB_TEST(ssl, 9, 100, 30, 25, 4, 0,  55)); /* 30-55  */
-    EXPECT_TEST(DFB_TEST(ssl, 9, 100, 55, 15, 2, 0,  65)); /* 55-70  */
-    EXPECT_TEST(DFB_TEST(ssl, 9, 100, 75, 25, 2, 0,  90)); /* 75-100 */
-    EXPECT_TEST(DFB_TEST(ssl, 9, 100, 10, 25, 0, 1, 100)); /* 10-35 */
-
-    EXPECT_TEST(DFB_TEST(ssl,10, 100,  0, 20, 1, 0,  20)); /*  0-20  */
-    EXPECT_TEST(DFB_TEST(ssl,10, 100, 30, 20, 2, 0,  40)); /* 30-50  */
-    EXPECT_TEST(DFB_TEST(ssl,10, 100,  0, 40, 1, 0,  50)); /*  0-40  */
-    EXPECT_TEST(DFB_TEST(ssl,10, 100, 50, 50, 0, 1, 100)); /* 50-100 */
-
-    /* Test region between other regions */
-    EXPECT_TEST(DFB_TEST(ssl,11, 100,  0, 20, 1, 0,  20)); /*  0-20  */
-    EXPECT_TEST(DFB_TEST(ssl,11, 100, 80, 20, 2, 0,  40)); /* 80-100 */
-    EXPECT_TEST(DFB_TEST(ssl,11, 100, 40, 20, 3, 0,  60)); /* 40-60  */
-    EXPECT_TEST(DFB_TEST(ssl,11, 100, 20, 20, 2, 0,  80)); /* 20-40  */
-    EXPECT_TEST(DFB_TEST(ssl,11, 100, 60, 20, 0, 1, 100)); /* 60-80  */
-
-    /* Test gap before first bucket (prev==NULL in gap-before branch) */
-    EXPECT_TEST(DFB_TEST(ssl,12, 100, 50, 20, 1, 0,  20)); /* 50-70  */
-    EXPECT_TEST(DFB_TEST(ssl,12, 100,  0, 20, 2, 0,  40)); /*  0-20  gap before first */
-    EXPECT_TEST(DFB_TEST(ssl,12, 100, 20, 30, 1, 0,  70)); /* 20-50  bridges gap */
-    EXPECT_TEST(DFB_TEST(ssl,12, 100, 70, 30, 0, 1, 100)); /* 70-100 */
-
-    /* Test fragment after message is already complete (ready early return) */
-    EXPECT_TEST(DFB_TEST(ssl,13, 100,  0,100, 0, 1, 100)); /*  0-100 complete */
-    EXPECT_TEST(DFB_TEST(ssl,13, 100,  0, 50, 0, 1, 100)); /*  0-50  dup on ready */
-
-    /* Test combine where next bucket is larger than cur (chosenBucket=&next) */
-    EXPECT_TEST(DFB_TEST(ssl,14, 100,  0, 10, 1, 0,  10)); /*  0-10  */
-    EXPECT_TEST(DFB_TEST(ssl,14, 100, 30, 50, 2, 0,  60)); /* 30-80  */
-    EXPECT_TEST(DFB_TEST(ssl,14, 100,  5, 30, 1, 0,  80)); /*  5-35  next>cur */
-    EXPECT_TEST(DFB_TEST(ssl,14, 100, 80, 20, 0, 1, 100)); /* 80-100 */
-
-    /* Test super fragment covering all existing buckets */
-    EXPECT_TEST(DFB_TEST(ssl,15, 100, 10, 10, 1, 0,  10)); /* 10-20  */
-    EXPECT_TEST(DFB_TEST(ssl,15, 100, 30, 10, 2, 0,  20)); /* 30-40  */
-    EXPECT_TEST(DFB_TEST(ssl,15, 100, 60, 10, 3, 0,  30)); /* 60-70  */
-    EXPECT_TEST(DFB_TEST(ssl,15, 100,  0,100, 0, 1, 100)); /*  0-100 super frag */
-
-    /* Test exact duplicate fragment */
-    EXPECT_TEST(DFB_TEST(ssl,16, 100, 20, 40, 1, 0,  40)); /* 20-60  */
-    EXPECT_TEST(DFB_TEST(ssl,16, 100, 20, 40, 1, 0,  40)); /* 20-60  exact dup */
-    EXPECT_TEST(DFB_TEST(ssl,16, 100,  0, 20, 1, 0,  60)); /*  0-20  */
-    EXPECT_TEST(DFB_TEST(ssl,16, 100, 60, 40, 0, 1, 100)); /* 60-100 */
-
-    /* Test combine bridging two buckets (combineNext, cur->data) */
-    EXPECT_TEST(DFB_TEST(ssl,17, 100,  0, 30, 1, 0,  30)); /*  0-30  */
-    EXPECT_TEST(DFB_TEST(ssl,17, 100, 60, 20, 2, 0,  50)); /* 60-80  */
-    EXPECT_TEST(DFB_TEST(ssl,17, 100, 20, 45, 1, 0,  80)); /* 20-65  bridge */
-    EXPECT_TEST(DFB_TEST(ssl,17, 100, 80, 20, 0, 1, 100)); /* 80-100 */
-
-    /* Test progressive left-extension with partial overlaps */
-    EXPECT_TEST(DFB_TEST(ssl,18, 100, 70, 30, 1, 0,  30)); /* 70-100 */
-    EXPECT_TEST(DFB_TEST(ssl,18, 100, 50, 30, 1, 0,  50)); /* 50-80  extend left */
-    EXPECT_TEST(DFB_TEST(ssl,18, 100, 30, 30, 1, 0,  70)); /* 30-60  extend left */
-    EXPECT_TEST(DFB_TEST(ssl,18, 100,  0, 40, 0, 1, 100)); /*  0-40  complete left */
-
-    DtlsMsgListDelete(ssl->dtls_rx_msg_list, ssl->heap);
-    ssl->dtls_rx_msg_list = NULL;
-    ssl->dtls_rx_msg_list_sz = 0;
-    return EXPECT_RESULT();
-}
-
-#endif
-
-
-#if !defined(NO_FILESYSTEM) && \
-     defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) && \
-     defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && !defined(NO_RSA)
-
-static int test_wolfSSL_dtls_stateless2(void)
-{
-    EXPECT_DECLS;
-    WOLFSSL *ssl_c = NULL;
-    WOLFSSL *ssl_c2 = NULL;
-    WOLFSSL *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-    WOLFSSL_CTX *ctx_c = NULL;
-    WOLFSSL_CTX *ctx_s = NULL;
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method), 0);
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, NULL, &ssl_c2, NULL,
-        wolfDTLSv1_2_client_method, NULL), 0);
-    ExpectFalse(wolfSSL_is_stateful(ssl_s));
-    /* send CH */
-    ExpectTrue((wolfSSL_connect(ssl_c2) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
-        (ssl_c2->error == WC_NO_ERR_TRACE(WANT_READ)));
-    ExpectIntEQ(wolfDTLS_accept_stateless(ssl_s), WOLFSSL_FAILURE);
-    ExpectFalse(wolfSSL_is_stateful(ssl_s));
-    ExpectIntNE(test_ctx.c_len, 0);
-    /* consume HRR */
-    test_memio_clear_buffer(&test_ctx, 1);
-    /* send CH1 */
-    ExpectIntEQ(wolfSSL_connect(ssl_c), WOLFSSL_FATAL_ERROR);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, WOLFSSL_FATAL_ERROR),
-            WOLFSSL_ERROR_WANT_READ);
-    /* send HRR */
-    ExpectIntEQ(wolfDTLS_accept_stateless(ssl_s), WOLFSSL_FAILURE);
-    /* send CH2 */
-    ExpectIntEQ(wolfSSL_connect(ssl_c), WOLFSSL_FATAL_ERROR);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, WOLFSSL_FATAL_ERROR),
-            WOLFSSL_ERROR_WANT_READ);
-    /* send HRR */
-    ExpectIntEQ(wolfDTLS_accept_stateless(ssl_s), WOLFSSL_SUCCESS);
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-    ExpectTrue(wolfSSL_is_stateful(ssl_s));
-
-    wolfSSL_free(ssl_c2);
-    wolfSSL_free(ssl_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_CTX_free(ctx_s);
-    return EXPECT_RESULT();
-}
-
-#ifdef HAVE_MAX_FRAGMENT
-static int test_wolfSSL_dtls_stateless_maxfrag(void)
-{
-    EXPECT_DECLS;
-    WOLFSSL *ssl_c = NULL;
-    WOLFSSL *ssl_c2 = NULL;
-    WOLFSSL *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-    WOLFSSL_CTX *ctx_c = NULL;
-    WOLFSSL_CTX *ctx_s = NULL;
-    word16 max_fragment = 0;
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method), 0);
-    ExpectNotNull(ssl_s);
-    ExpectNotNull(ssl_c2 = wolfSSL_new(ctx_c));
-    ExpectIntEQ(wolfSSL_UseMaxFragment(ssl_c2, WOLFSSL_MFL_2_8),
-        WOLFSSL_SUCCESS);
-    wolfSSL_SetIOWriteCtx(ssl_c2, &test_ctx);
-    wolfSSL_SetIOReadCtx(ssl_c2, &test_ctx);
-    if (EXPECT_SUCCESS()) {
-        max_fragment = ssl_s->max_fragment;
-    }
-    /* send CH */
-    ExpectTrue((wolfSSL_connect(ssl_c2) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
-        (ssl_c2->error == WC_NO_ERR_TRACE(WANT_READ)));
-    ExpectTrue((wolfSSL_accept(ssl_s) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
-        (ssl_s->error == WC_NO_ERR_TRACE(WANT_READ)));
-    /* CH without cookie shouldn't change state */
-    ExpectIntEQ(ssl_s->max_fragment, max_fragment);
-    ExpectIntNE(test_ctx.c_len, 0);
-
-    /* consume HRR from buffer */
-    test_memio_clear_buffer(&test_ctx, 1);
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-
-    wolfSSL_free(ssl_c2);
-    wolfSSL_free(ssl_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_CTX_free(ctx_s);
-    return EXPECT_RESULT();
-}
-#endif /* HAVE_MAX_FRAGMENT */
-
-#if defined(WOLFSSL_DTLS_NO_HVR_ON_RESUME)
-#define ROUNDS_WITH_HVR 4
-#define ROUNDS_WITHOUT_HVR 2
-#define HANDSHAKE_TYPE_OFFSET DTLS_RECORD_HEADER_SZ
-static int buf_is_hvr(const byte *data, int len)
-{
-    if (len < DTLS_RECORD_HEADER_SZ + DTLS_HANDSHAKE_HEADER_SZ)
-        return 0;
-    return data[HANDSHAKE_TYPE_OFFSET] == hello_verify_request;
-}
-
-static int _test_wolfSSL_dtls_stateless_resume(byte useticket, byte bad)
-{
-    EXPECT_DECLS;
-    struct test_memio_ctx test_ctx;
-    WOLFSSL_CTX *ctx_c = NULL;
-    WOLFSSL_CTX *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL;
-    WOLFSSL *ssl_s = NULL;
-    WOLFSSL_SESSION *sess = NULL;
-    int round_trips;
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c,
-        &ssl_s, wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method), 0);
-#ifdef HAVE_SESSION_TICKET
-    if (useticket) {
-        ExpectIntEQ(wolfSSL_UseSessionTicket(ssl_c), WOLFSSL_SUCCESS);
-    }
-#endif
-    round_trips = ROUNDS_WITH_HVR;
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, round_trips,
-        &round_trips), 0);
-    ExpectIntEQ(round_trips, ROUNDS_WITH_HVR);
-    ExpectNotNull(sess = wolfSSL_get1_session(ssl_c));
-    wolfSSL_shutdown(ssl_c);
-    wolfSSL_shutdown(ssl_s);
-    wolfSSL_free(ssl_c);
-    ssl_c = NULL;
-    wolfSSL_free(ssl_s);
-    ssl_s = NULL;
-
-    test_memio_clear_buffer(&test_ctx, 1);
-    test_memio_clear_buffer(&test_ctx, 0);
-    /* make resumption invalid */
-    if (bad && (sess != NULL)) {
-        if (useticket) {
-#ifdef HAVE_SESSION_TICKET
-            if (sess->ticket != NULL) {
-                sess->ticket[0] = !sess->ticket[0];
-            }
-#endif /* HAVE_SESSION_TICKET */
-        }
-        else {
-            sess->sessionID[0] = !sess->sessionID[0];
-        }
-    }
-    ExpectNotNull(ssl_c = wolfSSL_new(ctx_c));
-    ExpectNotNull(ssl_s = wolfSSL_new(ctx_s));
-    wolfSSL_SetIOWriteCtx(ssl_c, &test_ctx);
-    wolfSSL_SetIOReadCtx(ssl_c, &test_ctx);
-    wolfSSL_SetIOWriteCtx(ssl_s, &test_ctx);
-    wolfSSL_SetIOReadCtx(ssl_s, &test_ctx);
-    ExpectIntEQ(wolfSSL_set_session(ssl_c, sess), WOLFSSL_SUCCESS);
-    ExpectTrue((wolfSSL_connect(ssl_c) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
-        (ssl_c->error == WC_NO_ERR_TRACE(WANT_READ)));
-    ExpectTrue((wolfSSL_accept(ssl_s) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
-        (ssl_s->error == WC_NO_ERR_TRACE(WANT_READ)));
-    ExpectFalse(bad && !buf_is_hvr(test_ctx.c_buff, test_ctx.c_len));
-    ExpectFalse(!bad && buf_is_hvr(test_ctx.c_buff, test_ctx.c_len));
-    if (!useticket) {
-        ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, &round_trips), 0);
-        ExpectFalse(bad && round_trips != ROUNDS_WITH_HVR - 1);
-        ExpectFalse(!bad && round_trips != ROUNDS_WITHOUT_HVR - 1);
-    }
-    wolfSSL_SESSION_free(sess);
-    wolfSSL_free(ssl_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_CTX_free(ctx_s);
-    return EXPECT_RESULT();
-}
-
-static int test_wolfSSL_dtls_stateless_resume(void)
-{
-    EXPECT_DECLS;
-#ifdef HAVE_SESSION_TICKET
-    ExpectIntEQ(_test_wolfSSL_dtls_stateless_resume(1, 0), TEST_SUCCESS);
-    ExpectIntEQ(_test_wolfSSL_dtls_stateless_resume(1, 1), TEST_SUCCESS);
-#endif /* HAVE_SESION_TICKET */
-    ExpectIntEQ(_test_wolfSSL_dtls_stateless_resume(0, 0), TEST_SUCCESS);
-    ExpectIntEQ(_test_wolfSSL_dtls_stateless_resume(0, 1), TEST_SUCCESS);
-    return EXPECT_RESULT();
-}
-#endif /* WOLFSSL_DTLS_NO_HVR_ON_RESUME */
-
-#if !defined(NO_OLD_TLS)
-static int test_wolfSSL_dtls_stateless_downgrade(void)
-{
-    EXPECT_DECLS;
-    WOLFSSL_CTX *ctx_c = NULL;
-    WOLFSSL_CTX *ctx_c2 = NULL;
-    WOLFSSL_CTX *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL;
-    WOLFSSL *ssl_c2 = NULL;
-    WOLFSSL *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method), 0);
-    ExpectIntEQ(wolfSSL_CTX_SetMinVersion(ctx_s, WOLFSSL_DTLSV1),
-        WOLFSSL_SUCCESS);
-    ExpectNotNull(ctx_c2 = wolfSSL_CTX_new(wolfDTLSv1_client_method()));
-    wolfSSL_SetIORecv(ctx_c2, test_memio_read_cb);
-    wolfSSL_SetIOSend(ctx_c2, test_memio_write_cb);
-    ExpectNotNull(ssl_c2 = wolfSSL_new(ctx_c2));
-    wolfSSL_SetIOWriteCtx(ssl_c2, &test_ctx);
-    wolfSSL_SetIOReadCtx(ssl_c2, &test_ctx);
-    /* send CH */
-    ExpectTrue((wolfSSL_connect(ssl_c2) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
-        (ssl_c2->error == WC_NO_ERR_TRACE(WANT_READ)));
-    ExpectTrue((wolfSSL_accept(ssl_s) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
-        (ssl_s->error == WC_NO_ERR_TRACE(WANT_READ)));
-    ExpectIntNE(test_ctx.c_len, 0);
-    /* consume HRR */
-    test_memio_clear_buffer(&test_ctx, 1);
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-
-    wolfSSL_free(ssl_c2);
-    wolfSSL_free(ssl_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_CTX_free(ctx_c2);
-    wolfSSL_CTX_free(ctx_s);
-
-    return EXPECT_RESULT();
-}
-#endif /* !defined(NO_OLD_TLS) */
-
-#endif /* defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) && \
-    !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER)*/
-
-#if defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) && \
-    !defined(NO_OLD_TLS) && defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES)
-static int test_WOLFSSL_dtls_version_alert(void)
-{
-    EXPECT_DECLS;
-    struct test_memio_ctx test_ctx;
-    WOLFSSL_CTX *ctx_c = NULL;
-    WOLFSSL_CTX *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL;
-    WOLFSSL *ssl_s = NULL;
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_2_client_method, wolfDTLSv1_server_method), 0);
-
-    /* client hello */
-    ExpectTrue((wolfSSL_connect(ssl_c) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
-        (ssl_c->error == WC_NO_ERR_TRACE(WANT_READ)));
-    /* hrr */
-    ExpectTrue((wolfSSL_accept(ssl_s) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
-        (ssl_s->error == WC_NO_ERR_TRACE(WANT_READ)));
-    /* client hello 1 */
-    ExpectTrue((wolfSSL_connect(ssl_c) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
-        (ssl_c->error == WC_NO_ERR_TRACE(WANT_READ)));
-    /* server hello */
-    ExpectTrue((wolfSSL_accept(ssl_s) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
-        (ssl_s->error == WC_NO_ERR_TRACE(WANT_READ)));
-    /* should fail */
-    ExpectTrue((wolfSSL_connect(ssl_c) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
-        (ssl_c->error == WC_NO_ERR_TRACE(VERSION_ERROR)));
-    /* shuould fail */
-    ExpectTrue((wolfSSL_accept(ssl_s) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
-        (ssl_s->error == WC_NO_ERR_TRACE(VERSION_ERROR) || ssl_s->error == WC_NO_ERR_TRACE(FATAL_ERROR)));
-
-    wolfSSL_free(ssl_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_CTX_free(ctx_s);
-
-    return EXPECT_RESULT();
-}
-#else
-static int test_WOLFSSL_dtls_version_alert(void)
-{
-    return TEST_SKIPPED;
-}
-#endif /* defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) &&
-        * !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER) &&
-        * !defined(NO_OLD_TLS) && !defined(NO_RSA)
-        */
 
 
 #if defined(WOLFSSL_TICKET_NONCE_MALLOC) && defined(HAVE_SESSION_TICKET)       \
@@ -33557,7 +30203,8 @@ static int test_extra_alerts_wrong_cs(void)
 #endif
 
 #if defined(WOLFSSL_TLS13) && !defined(WOLFSSL_NO_TLS12) &&   \
-    defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_AES_256)
+    defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_AES_256) && \
+    defined(BUILD_TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384)
 
 #define TEST_CS_DOWNGRADE_CLIENT "ECDHE-RSA-AES256-GCM-SHA384"
 
@@ -34358,71 +31005,6 @@ static int test_rpk_set_xxx_cert_type(void)
     return EXPECT_RESULT();
 }
 
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
-
-
-static int test_dtls13_bad_epoch_ch(void)
-{
-    EXPECT_DECLS;
-    WOLFSSL_CTX *ctx_c = NULL;
-    WOLFSSL_CTX *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL;
-    WOLFSSL *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-    const int EPOCH_OFF = 3;
-    int groups[] = {
-        WOLFSSL_ECC_SECP256R1,
-        WOLFSSL_ECC_SECP384R1,
-    };
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
-
-    /* disable hrr cookie so we can later check msgsReceived.got_client_hello
-     *  with just one message */
-    ExpectIntEQ(wolfSSL_disable_hrr_cookie(ssl_s), WOLFSSL_SUCCESS);
-
-    /* Set client groups to traditional only to avoid CH fragmentation */
-    ExpectIntEQ(wolfSSL_set_groups(ssl_c, groups, 2), WOLFSSL_SUCCESS);
-
-    ExpectIntNE(wolfSSL_connect(ssl_c), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)),
-        WOLFSSL_ERROR_WANT_READ);
-
-    ExpectIntGE(test_ctx.s_len, EPOCH_OFF + 2);
-
-    /* first CH should use epoch 0x0 */
-    ExpectTrue((test_ctx.s_buff[EPOCH_OFF] == 0x0) &&
-        (test_ctx.s_buff[EPOCH_OFF + 1] == 0x0));
-
-    /* change epoch to 2 */
-    test_ctx.s_buff[EPOCH_OFF + 1] = 0x2;
-
-    ExpectIntNE(wolfSSL_accept(ssl_s), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)),
-        WOLFSSL_ERROR_WANT_READ);
-
-    ExpectIntNE(ssl_s->msgsReceived.got_client_hello, 1);
-
-    /* resend the CH */
-    ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
-
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-
-    wolfSSL_free(ssl_c);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_s);
-
-    return EXPECT_RESULT();
-}
-#else
-static int test_dtls13_bad_epoch_ch(void)
-{
-    return TEST_SKIPPED;
-}
-#endif
 
 #if ((defined(WOLFSSL_TLS13) && !defined(WOLFSSL_NO_DEF_TICKET_ENC_CB) && \
       defined(HAVE_SESSION_TICKET) && defined(WOLFSSL_TICKET_HAVE_ID) && \
@@ -34511,232 +31093,6 @@ static int test_short_session_id(void)
 }
 #endif
 
-#if defined(HAVE_NULL_CIPHER) && defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) \
-    && defined(WOLFSSL_DTLS13)
-static byte* test_find_string(const char *string,
-    byte *buf, int buf_size)
-{
-    int string_size, i;
-
-    string_size = (int)XSTRLEN(string);
-    for (i = 0; i < buf_size - string_size - 1; i++) {
-        if (XSTRCMP((char*)&buf[i], string) == 0)
-            return &buf[i];
-    }
-    return NULL;
-}
-
-static int test_wolfSSL_dtls13_null_cipher(void)
-{
-    EXPECT_DECLS;
-    WOLFSSL_CTX *ctx_c = NULL;
-    WOLFSSL_CTX *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL;
-    WOLFSSL *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-    const char *test_str = "test";
-    int test_str_size;
-    byte buf[255], *ptr = NULL;
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-    test_ctx.c_ciphers = test_ctx.s_ciphers = "TLS13-SHA256-SHA256";
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-    test_str_size = XSTRLEN("test") + 1;
-    ExpectIntEQ(wolfSSL_write(ssl_c, test_str, test_str_size), test_str_size);
-    ExpectIntEQ(wolfSSL_read(ssl_s, buf, sizeof(buf)), test_str_size);
-    ExpectIntEQ(XSTRCMP((char*)buf, test_str), 0);
-
-    ExpectIntEQ(wolfSSL_write(ssl_c, test_str, test_str_size), test_str_size);
-
-    /* check that the packet was sent cleartext */
-    ExpectNotNull(ptr = test_find_string(test_str, test_ctx.s_buff,
-        test_ctx.s_len));
-    if (ptr != NULL) {
-        /* modify the message */
-        *ptr = 'H';
-        /* bad messages should be ignored in DTLS */
-        ExpectIntEQ(wolfSSL_read(ssl_s, buf, sizeof(buf)), -1);
-        ExpectIntEQ(ssl_s->error, WC_NO_ERR_TRACE(WANT_READ));
-    }
-
-    ExpectIntEQ(wolfSSL_shutdown(ssl_c), WOLFSSL_SHUTDOWN_NOT_DONE);
-    ExpectIntEQ(wolfSSL_shutdown(ssl_s), WOLFSSL_SHUTDOWN_NOT_DONE);
-    ExpectIntEQ(wolfSSL_shutdown(ssl_c), 1);
-    ExpectIntEQ(wolfSSL_shutdown(ssl_s), 1);
-
-    wolfSSL_free(ssl_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_CTX_free(ctx_s);
-    return TEST_SUCCESS;
-}
-#else
-static int test_wolfSSL_dtls13_null_cipher(void)
-{
-    return TEST_SKIPPED;
-}
-#endif
-#if defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) &&          \
-    !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER) &&   \
-    !defined(SINGLE_THREADED) && !defined(NO_RSA)
-
-static int test_dtls_msg_get_connected_port(int fd, word16 *port)
-{
-    SOCKADDR_S peer;
-    XSOCKLENT len;
-    int ret;
-
-    XMEMSET((byte*)&peer, 0, sizeof(peer));
-    len = sizeof(peer);
-    ret = getpeername(fd,  (SOCKADDR*)&peer, &len);
-    if (ret != 0 || len > (XSOCKLENT)sizeof(peer))
-        return -1;
-    switch (peer.ss_family) {
-#ifdef WOLFSSL_IPV6
-    case WOLFSSL_IP6: {
-        *port = ntohs(((SOCKADDR_IN6*)&peer)->sin6_port);
-        break;
-    }
-#endif /* WOLFSSL_IPV6 */
-    case WOLFSSL_IP4:
-        *port = ntohs(((SOCKADDR_IN*)&peer)->sin_port);
-        break;
-    default:
-        return -1;
-    }
-    return 0;
-}
-
-static int test_dtls_msg_from_other_peer_cb(WOLFSSL_CTX *ctx, WOLFSSL *ssl)
-{
-    char buf[1] = {'t'};
-    SOCKADDR_IN_T addr;
-    int sock_fd;
-    word16 port;
-    int err;
-
-    (void)ssl;
-    (void)ctx;
-
-    if (ssl == NULL)
-        return -1;
-
-    err = test_dtls_msg_get_connected_port(wolfSSL_get_fd(ssl), &port);
-    if (err != 0)
-        return -1;
-
-    sock_fd = socket(AF_INET_V, SOCK_DGRAM, 0);
-    if (sock_fd == -1)
-        return -1;
-    build_addr(&addr, wolfSSLIP, port, 1, 0);
-
-    /* send a packet to the server. Being another socket, the kernel will ensure
-     * the source port will be different. */
-    err = (int)sendto(sock_fd, buf, sizeof(buf), 0, (SOCKADDR*)&addr,
-        sizeof(addr));
-
-    close(sock_fd);
-    if (err == -1)
-        return -1;
-
-    return 0;
-}
-
-/* setup a SSL session but just after the handshake send a packet to the server
- * with a source address different than the one of the connected client. The I/O
- * callback EmbedRecvFrom should just ignore the packet. Sending of the packet
- * is done in test_dtls_msg_from_other_peer_cb */
-static int test_dtls_msg_from_other_peer(void)
-{
-    EXPECT_DECLS;
-    callback_functions client_cbs;
-    callback_functions server_cbs;
-
-    XMEMSET((byte*)&client_cbs, 0, sizeof(client_cbs));
-    XMEMSET((byte*)&server_cbs, 0, sizeof(server_cbs));
-
-    client_cbs.method = wolfDTLSv1_2_client_method;
-    server_cbs.method = wolfDTLSv1_2_server_method;
-    client_cbs.doUdp = 1;
-    server_cbs.doUdp = 1;
-
-    test_wolfSSL_client_server_nofail_ex(&client_cbs, &server_cbs,
-        test_dtls_msg_from_other_peer_cb);
-
-    ExpectIntEQ(client_cbs.return_code, WOLFSSL_SUCCESS);
-    ExpectIntEQ(server_cbs.return_code, WOLFSSL_SUCCESS);
-
-    return EXPECT_RESULT();
-}
-#else
-static int test_dtls_msg_from_other_peer(void)
-{
-    return TEST_SKIPPED;
-}
-#endif /* defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) &&          \
-        *  !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER) &&  \
-        *  !defined(SINGLE_THREADED) && !defined(NO_RSA) */
-#if defined(WOLFSSL_DTLS) && !defined(WOLFSSL_IPV6) &&               \
-    !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER) &&   \
-    defined(HAVE_IO_TESTS_DEPENDENCIES) && !defined(WOLFSSL_NO_TLS12) \
-    && !defined(USE_WINDOWS_API)
-static int test_dtls_ipv6_check(void)
-{
-    EXPECT_DECLS;
-    WOLFSSL_CTX *ctx_c = NULL;
-    WOLFSSL_CTX *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL;
-    WOLFSSL *ssl_s = NULL;
-    SOCKADDR_IN fake_addr6;
-    int sockfd = -1;
-
-    ExpectNotNull(ctx_c = wolfSSL_CTX_new(wolfDTLSv1_2_client_method()));
-    ExpectNotNull(ssl_c  = wolfSSL_new(ctx_c));
-    ExpectNotNull(ctx_s = wolfSSL_CTX_new(wolfDTLSv1_2_server_method()));
-    ExpectIntEQ(wolfSSL_CTX_use_PrivateKey_file(ctx_s, svrKeyFile,
-        WOLFSSL_FILETYPE_PEM), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_CTX_use_certificate_file(ctx_s, svrCertFile,
-        WOLFSSL_FILETYPE_PEM), WOLFSSL_SUCCESS);
-    ExpectNotNull(ssl_s  = wolfSSL_new(ctx_s));
-    XMEMSET((byte*)&fake_addr6, 0, sizeof(fake_addr6));
-    /* mimic a sockaddr_in6 struct, this way we can't test without
-     *  WOLFSSL_IPV6 */
-    fake_addr6.sin_family = WOLFSSL_IP6;
-    ExpectIntNE(sockfd = socket(AF_INET, SOCK_DGRAM, 0), -1);
-    ExpectIntEQ(wolfSSL_set_fd(ssl_c, sockfd), WOLFSSL_SUCCESS);
-    /* can't return error here, as the peer is opaque for wolfssl library at
-     * this point */
-    ExpectIntEQ(wolfSSL_dtls_set_peer(ssl_c, &fake_addr6, sizeof(fake_addr6)),
-        WOLFSSL_SUCCESS);
-    ExpectIntNE(fcntl(sockfd, F_SETFL, O_NONBLOCK), -1);
-    wolfSSL_dtls_set_using_nonblock(ssl_c, 1);
-    ExpectIntNE(wolfSSL_connect(ssl_c), WOLFSSL_SUCCESS);
-    ExpectIntEQ(ssl_c->error, WC_NO_ERR_TRACE(SOCKET_ERROR_E));
-
-    ExpectIntEQ(wolfSSL_dtls_set_peer(ssl_s, &fake_addr6, sizeof(fake_addr6)),
-        WOLFSSL_SUCCESS);
-    /* reuse the socket */
-    ExpectIntEQ(wolfSSL_set_fd(ssl_c, sockfd), WOLFSSL_SUCCESS);
-    wolfSSL_dtls_set_using_nonblock(ssl_s, 1);
-    ExpectIntNE(wolfSSL_accept(ssl_s), WOLFSSL_SUCCESS);
-    ExpectIntEQ(ssl_s->error, WC_NO_ERR_TRACE(SOCKET_ERROR_E));
-    if (sockfd != -1)
-        close(sockfd);
-
-    wolfSSL_free(ssl_c);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_s);
-    return EXPECT_RESULT();
-}
-#else
-static int test_dtls_ipv6_check(void)
-{
-    return TEST_SKIPPED;
-}
-#endif
 
 #if !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER) &&   \
     defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_SECURE_RENEGOTIATION)
@@ -34830,96 +31186,6 @@ static int test_wolfSSL_configure_args(void)
     return EXPECT_RESULT();
 }
 
-static int test_dtls_no_extensions(void)
-{
-    EXPECT_DECLS;
-#if defined(WOLFSSL_DTLS) && defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && \
-    !defined(WOLFSSL_NO_TLS12)
-    WOLFSSL *ssl_s = NULL;
-    WOLFSSL_CTX *ctx_s = NULL;
-    struct test_memio_ctx test_ctx;
-    const byte chNoExtensions[] = {
-        /* Handshake type */
-        0x16,
-        /* Version */
-        0xfe, 0xff,
-        /* Epoch */
-        0x00, 0x00,
-        /* Seq number */
-        0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-        /* Length */
-        0x00, 0x40,
-        /* CH type */
-        0x01,
-        /* Length */
-        0x00, 0x00, 0x34,
-        /* Msg Seq */
-        0x00, 0x00,
-        /* Frag offset */
-        0x00, 0x00, 0x00,
-        /* Frag length */
-        0x00, 0x00, 0x34,
-        /* Version */
-        0xfe, 0xff,
-        /* Random */
-        0x62, 0xfe, 0xbc, 0xfe, 0x2b, 0xfe, 0x3f, 0xeb, 0x03, 0xc4, 0xea, 0x37,
-        0xe7, 0x47, 0x7e, 0x8a, 0xd9, 0xbf, 0x77, 0x0f, 0x6c, 0xb6, 0x77, 0x0b,
-        0x03, 0x3f, 0x82, 0x2b, 0x21, 0x64, 0x57, 0x1d,
-        /* Session Length */
-        0x00,
-        /* Cookie Length */
-        0x00,
-        /* CS Length */
-        0x00, 0x0c,
-        /* CS */
-        0xc0, 0x0a, 0xc0, 0x09, 0xc0, 0x14, 0xc0, 0x13, 0x00, 0x39, 0x00, 0x33,
-        /* Comp Meths Length */
-        0x01,
-        /* Comp Meths */
-        0x00
-        /* And finally... no extensions */
-    };
-    int i;
-#ifdef OPENSSL_EXTRA
-    int repeats = 2;
-#else
-    int repeats = 1;
-#endif
-
-    for (i = 0; i < repeats; i++) {
-        XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-        ssl_s = NULL;
-        ctx_s = NULL;
-
-        ExpectIntEQ(test_memio_setup(&test_ctx, NULL, &ctx_s, NULL, &ssl_s,
-            NULL, wolfDTLS_server_method), 0);
-
-        test_memio_clear_buffer(&test_ctx, 0);
-        ExpectIntEQ(
-            test_memio_inject_message(&test_ctx, 1,
-                (const char *)chNoExtensions, sizeof(chNoExtensions)), 0);
-
-
-#ifdef OPENSSL_EXTRA
-        if (i > 0) {
-            ExpectIntEQ(wolfSSL_set_max_proto_version(ssl_s, DTLS1_2_VERSION),
-                        WOLFSSL_SUCCESS);
-        }
-#endif
-
-        ExpectIntEQ(wolfSSL_accept(ssl_s), -1);
-        ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-
-        /* Expecting a handshake msg. Either HVR or SH. */
-        ExpectIntGT(test_ctx.c_len, 0);
-        ExpectIntEQ(test_ctx.c_buff[0], 0x16);
-
-        wolfSSL_free(ssl_s);
-        wolfSSL_CTX_free(ctx_s);
-    }
-#endif
-    return EXPECT_RESULT();
-}
 
 static int test_tls_alert_no_server_hello(void)
 {
@@ -35044,42 +31310,6 @@ static int test_TLSX_CA_NAMES_bad_extension(void)
     return EXPECT_RESULT();
 }
 
-#if defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) && \
-    defined(HAVE_IO_TESTS_DEPENDENCIES)
-static void test_dtls_1_0_hvr_downgrade_ctx_ready(WOLFSSL_CTX* ctx)
-{
-    AssertIntEQ(wolfSSL_CTX_SetMinVersion(ctx, WOLFSSL_DTLSV1_2),
-                WOLFSSL_SUCCESS);
-}
-
-static int test_dtls_1_0_hvr_downgrade(void)
-{
-    EXPECT_DECLS;
-    callback_functions func_cb_client;
-    callback_functions func_cb_server;
-
-    XMEMSET(&func_cb_client, 0, sizeof(callback_functions));
-    XMEMSET(&func_cb_server, 0, sizeof(callback_functions));
-
-    func_cb_client.doUdp = func_cb_server.doUdp = 1;
-    func_cb_client.method = wolfDTLS_client_method;
-    func_cb_server.method = wolfDTLSv1_2_server_method;
-    func_cb_client.ctx_ready = test_dtls_1_0_hvr_downgrade_ctx_ready;
-
-    test_wolfSSL_client_server_nofail(&func_cb_client, &func_cb_server);
-
-    ExpectIntEQ(func_cb_client.return_code, TEST_SUCCESS);
-    ExpectIntEQ(func_cb_server.return_code, TEST_SUCCESS);
-
-    return EXPECT_RESULT();
-}
-#else
-static int test_dtls_1_0_hvr_downgrade(void)
-{
-    EXPECT_DECLS;
-    return EXPECT_RESULT();
-}
-#endif
 
 #if defined(HAVE_IO_TESTS_DEPENDENCIES) && !defined(WOLFSSL_NO_TLS12) && \
         defined(HAVE_SESSION_TICKET)
@@ -35236,499 +31466,12 @@ static int test_session_ticket_hs_update(void)
     return EXPECT_RESULT();
 }
 
-#if defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) && \
-    defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_SECURE_RENEGOTIATION)
-static void test_dtls_downgrade_scr_server_ctx_ready_server(WOLFSSL_CTX* ctx)
-{
-    AssertIntEQ(wolfSSL_CTX_SetMinVersion(ctx, WOLFSSL_DTLSV1_2),
-                WOLFSSL_SUCCESS);
-    AssertIntEQ(wolfSSL_CTX_UseSecureRenegotiation(ctx), WOLFSSL_SUCCESS);
-}
 
-static void test_dtls_downgrade_scr_server_ctx_ready(WOLFSSL_CTX* ctx)
-{
-    AssertIntEQ(wolfSSL_CTX_UseSecureRenegotiation(ctx), WOLFSSL_SUCCESS);
-}
 
-static void test_dtls_downgrade_scr_server_on_result(WOLFSSL* ssl)
-{
-    char testMsg[] = "Message after SCR";
-    char msgBuf[sizeof(testMsg)];
-    if (wolfSSL_is_server(ssl)) {
-        AssertIntEQ(wolfSSL_Rehandshake(ssl), WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR));
-        AssertIntEQ(wolfSSL_get_error(ssl, -1), WC_NO_ERR_TRACE(APP_DATA_READY));
-        AssertIntEQ(wolfSSL_read(ssl, msgBuf, sizeof(msgBuf)), sizeof(msgBuf));
-        AssertIntEQ(wolfSSL_Rehandshake(ssl), WOLFSSL_SUCCESS);
-        AssertIntEQ(wolfSSL_write(ssl, testMsg, sizeof(testMsg)),
-                    sizeof(testMsg));
-    }
-    else {
-        AssertIntEQ(wolfSSL_write(ssl, testMsg, sizeof(testMsg)),
-                    sizeof(testMsg));
-        AssertIntEQ(wolfSSL_read(ssl, msgBuf, sizeof(msgBuf)), sizeof(msgBuf));
-    }
-}
 
-static int test_dtls_downgrade_scr_server(void)
-{
-    EXPECT_DECLS;
-    callback_functions func_cb_client;
-    callback_functions func_cb_server;
-
-    XMEMSET(&func_cb_client, 0, sizeof(callback_functions));
-    XMEMSET(&func_cb_server, 0, sizeof(callback_functions));
-
-    func_cb_client.doUdp = func_cb_server.doUdp = 1;
-    func_cb_client.method = wolfDTLSv1_2_client_method;
-    func_cb_server.method = wolfDTLS_server_method;
-    func_cb_client.ctx_ready = test_dtls_downgrade_scr_server_ctx_ready;
-    func_cb_server.ctx_ready = test_dtls_downgrade_scr_server_ctx_ready_server;
-    func_cb_client.on_result = test_dtls_downgrade_scr_server_on_result;
-    func_cb_server.on_result = test_dtls_downgrade_scr_server_on_result;
-
-    test_wolfSSL_client_server_nofail(&func_cb_client, &func_cb_server);
-
-    ExpectIntEQ(func_cb_client.return_code, TEST_SUCCESS);
-    ExpectIntEQ(func_cb_server.return_code, TEST_SUCCESS);
-
-    return EXPECT_RESULT();
-}
-#else
-static int test_dtls_downgrade_scr_server(void)
-{
-    EXPECT_DECLS;
-    return EXPECT_RESULT();
-}
-#endif
-
-#if defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) && \
-    defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_SECURE_RENEGOTIATION)
-static void test_dtls_downgrade_scr_ctx_ready(WOLFSSL_CTX* ctx)
-{
-    AssertIntEQ(wolfSSL_CTX_SetMinVersion(ctx, WOLFSSL_DTLSV1_2),
-                WOLFSSL_SUCCESS);
-    AssertIntEQ(wolfSSL_CTX_UseSecureRenegotiation(ctx), WOLFSSL_SUCCESS);
-}
-
-static void test_dtls_downgrade_scr_on_result(WOLFSSL* ssl)
-{
-    char testMsg[] = "Message after SCR";
-    char msgBuf[sizeof(testMsg)];
-    if (wolfSSL_is_server(ssl)) {
-        AssertIntEQ(wolfSSL_Rehandshake(ssl), WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR));
-        AssertIntEQ(wolfSSL_get_error(ssl, -1), WC_NO_ERR_TRACE(APP_DATA_READY));
-        AssertIntEQ(wolfSSL_read(ssl, msgBuf, sizeof(msgBuf)), sizeof(msgBuf));
-        AssertIntEQ(wolfSSL_Rehandshake(ssl), WOLFSSL_SUCCESS);
-        AssertIntEQ(wolfSSL_write(ssl, testMsg, sizeof(testMsg)),
-                    sizeof(testMsg));
-    }
-    else {
-        AssertIntEQ(wolfSSL_write(ssl, testMsg, sizeof(testMsg)),
-                    sizeof(testMsg));
-        AssertIntEQ(wolfSSL_read(ssl, msgBuf, sizeof(msgBuf)), sizeof(msgBuf));
-    }
-}
-
-static int test_dtls_downgrade_scr(void)
-{
-    EXPECT_DECLS;
-    callback_functions func_cb_client;
-    callback_functions func_cb_server;
-
-    XMEMSET(&func_cb_client, 0, sizeof(callback_functions));
-    XMEMSET(&func_cb_server, 0, sizeof(callback_functions));
-
-    func_cb_client.doUdp = func_cb_server.doUdp = 1;
-    func_cb_client.method = wolfDTLS_client_method;
-    func_cb_server.method = wolfDTLSv1_2_server_method;
-    func_cb_client.ctx_ready = test_dtls_downgrade_scr_ctx_ready;
-    func_cb_client.on_result = test_dtls_downgrade_scr_on_result;
-    func_cb_server.on_result = test_dtls_downgrade_scr_on_result;
-
-    test_wolfSSL_client_server_nofail(&func_cb_client, &func_cb_server);
-
-    ExpectIntEQ(func_cb_client.return_code, TEST_SUCCESS);
-    ExpectIntEQ(func_cb_server.return_code, TEST_SUCCESS);
-
-    return EXPECT_RESULT();
-}
-#else
-static int test_dtls_downgrade_scr(void)
-{
-    EXPECT_DECLS;
-    return EXPECT_RESULT();
-}
-#endif
-
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13) \
-    && !defined(WOLFSSL_NO_TLS12)
-
-static int test_dtls_client_hello_timeout_downgrade_read_cb(WOLFSSL *ssl,
-        char *data, int sz, void *ctx)
-{
-    static int call_counter = 0;
-    call_counter++;
-    (void)ssl;
-    (void)data;
-    (void)sz;
-    (void)ctx;
-    switch (call_counter) {
-        case 1:
-        case 2:
-            return WOLFSSL_CBIO_ERR_TIMEOUT;
-        case 3:
-            return WOLFSSL_CBIO_ERR_WANT_READ;
-        default:
-            AssertIntLE(call_counter, 3);
-            return -1;
-    }
-}
-#endif
-
-/* Make sure we don't send acks before getting a server hello */
-static int test_dtls_client_hello_timeout_downgrade(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13) \
-    && !defined(WOLFSSL_NO_TLS12)
-
-    WOLFSSL_CTX *ctx_c = NULL;
-    WOLFSSL_CTX *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL;
-    WOLFSSL *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-    DtlsRecordLayerHeader* dtlsRH;
-    size_t len;
-    byte sequence_number[8];
-    int i;
-
-    for (i = 0; i < 2; i++) {
-        XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-
-        ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-            wolfDTLS_client_method, wolfDTLSv1_2_server_method), 0);
-
-        if (i == 0) {
-            /* First time simulate timeout in IO layer */
-            /* CH1 */
-            ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-            ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-            /* HVR */
-            ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-            ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-            /* CH2 */
-            ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-            ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-            /* SH flight */
-            ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-            ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-            /* Drop the SH */
-            if (EXPECT_SUCCESS()) {
-                ExpectIntEQ(test_memio_drop_message(&test_ctx, 1, 0), 0);
-            }
-            /* Read the remainder of the flight */
-            ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-            ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-            wolfSSL_SSLSetIORecv(ssl_c,
-                    test_dtls_client_hello_timeout_downgrade_read_cb);
-            /* CH3 */
-            ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-            ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-            wolfSSL_SSLSetIORecv(ssl_c, test_memio_read_cb);
-        }
-        else {
-            /* Second time call wolfSSL_dtls_got_timeout */
-            /* CH1 */
-            ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-            ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-            /* HVR */
-            ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-            ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-            /* CH2 */
-            ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-            ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-            /* SH flight */
-            ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-            ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-            /* Drop the SH */
-            if (EXPECT_SUCCESS()) {
-                ExpectIntEQ(test_memio_drop_message(&test_ctx, 1, 0), 0);
-            }
-            /* Read the remainder of the flight */
-            ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-            ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-            /* Quick timeout should be set as we received at least one msg */
-            ExpectIntEQ(wolfSSL_dtls13_use_quick_timeout(ssl_c), 1);
-            ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
-            /* Quick timeout should be cleared after a quick timeout */
-            /* CH3 */
-            ExpectIntEQ(wolfSSL_dtls13_use_quick_timeout(ssl_c), 0);
-            ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
-        }
-
-        /* Parse out to make sure we got exactly one ClientHello message */
-        XMEMSET(&sequence_number, 0, sizeof(sequence_number));
-        /* Second ClientHello after HVR */
-        sequence_number[7] = 2;
-        dtlsRH = (DtlsRecordLayerHeader*)test_ctx.s_buff;
-        ExpectIntEQ(dtlsRH->type, handshake);
-        ExpectIntEQ(dtlsRH->pvMajor, DTLS_MAJOR);
-        ExpectIntEQ(dtlsRH->pvMinor, DTLSv1_2_MINOR);
-        ExpectIntEQ(XMEMCMP(sequence_number, dtlsRH->sequence_number,
-                sizeof(sequence_number)), 0);
-        len = (size_t)((dtlsRH->length[0] << 8) | dtlsRH->length[1]);
-        ExpectIntEQ(sizeof(DtlsRecordLayerHeader) + len, test_ctx.s_len);
-
-        /* Connection should be able to continue */
-        ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-
-        wolfSSL_free(ssl_c);
-        wolfSSL_free(ssl_s);
-        wolfSSL_CTX_free(ctx_c);
-        wolfSSL_CTX_free(ctx_s);
-        ssl_c = NULL;
-        ssl_s = NULL;
-        ctx_c = NULL;
-        ctx_s = NULL;
-        if (!EXPECT_SUCCESS())
-            break;
-    }
-
-#endif
-    return EXPECT_RESULT();
-}
-
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
-static int test_dtls_client_hello_timeout_read_cb(WOLFSSL *ssl, char *data,
-        int sz, void *ctx)
-{
-    static int call_counter = 0;
-    call_counter++;
-    (void)ssl;
-    (void)data;
-    (void)sz;
-    (void)ctx;
-    switch (call_counter) {
-        case 1:
-            return WOLFSSL_CBIO_ERR_TIMEOUT;
-        case 2:
-            return WOLFSSL_CBIO_ERR_WANT_READ;
-        default:
-            AssertIntLE(call_counter, 2);
-            return -1;
-    }
-}
-#endif
-
-/* Make sure we don't send acks before getting a server hello */
-static int test_dtls_client_hello_timeout(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
-    WOLFSSL *ssl_c = NULL;
-    WOLFSSL_CTX *ctx_c = NULL;
-    struct test_memio_ctx test_ctx;
-    DtlsRecordLayerHeader* dtlsRH;
-    size_t idx;
-    size_t len;
-    byte sequence_number[8];
-    int i;
-
-    for (i = 0; i < 2; i++) {
-        XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-
-        ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, NULL, &ssl_c, NULL,
-            wolfDTLSv1_3_client_method, NULL), 0);
-
-        if (i == 0) {
-            /* First time simulate timeout in IO layer */
-            wolfSSL_SSLSetIORecv(ssl_c, test_dtls_client_hello_timeout_read_cb);
-            ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
-            ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-        }
-        else {
-            /* Second time call wolfSSL_dtls_got_timeout */
-            ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
-            ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-            ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
-        }
-
-        /* Parse out to make sure we got exactly two ClientHello messages */
-        idx = 0;
-        XMEMSET(&sequence_number, 0, sizeof(sequence_number));
-        /* First ClientHello */
-        dtlsRH = (DtlsRecordLayerHeader*)(test_ctx.s_buff + idx);
-        ExpectIntEQ(dtlsRH->type, handshake);
-        ExpectIntEQ(dtlsRH->pvMajor, DTLS_MAJOR);
-        ExpectIntEQ(dtlsRH->pvMinor, DTLSv1_2_MINOR);
-        ExpectIntEQ(XMEMCMP(sequence_number, dtlsRH->sequence_number,
-                sizeof(sequence_number)), 0);
-        len = (size_t)((dtlsRH->length[0] << 8) | dtlsRH->length[1]);
-        ExpectIntLT(idx + sizeof(DtlsRecordLayerHeader) + len, test_ctx.s_len);
-        idx += sizeof(DtlsRecordLayerHeader) + len;
-        /* Second ClientHello */
-        sequence_number[7] = 1;
-        dtlsRH = (DtlsRecordLayerHeader*)(test_ctx.s_buff + idx);
-        ExpectIntEQ(dtlsRH->type, handshake);
-        ExpectIntEQ(dtlsRH->pvMajor, DTLS_MAJOR);
-        ExpectIntEQ(dtlsRH->pvMinor, DTLSv1_2_MINOR);
-        ExpectIntEQ(XMEMCMP(sequence_number, dtlsRH->sequence_number,
-                sizeof(sequence_number)), 0);
-        len = (size_t)((dtlsRH->length[0] << 8) | dtlsRH->length[1]);
-        ExpectIntEQ(idx + sizeof(DtlsRecordLayerHeader) + len, test_ctx.s_len);
-
-        wolfSSL_free(ssl_c);
-        wolfSSL_CTX_free(ctx_c);
-        ssl_c = NULL;
-        ctx_c = NULL;
-        if (!EXPECT_SUCCESS())
-            break;
-    }
-
-#endif
-    return EXPECT_RESULT();
-}
 
 /* DTLS test when dropping the changed cipher spec message */
-static int test_dtls_dropped_ccs(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS) \
-    && !defined(WOLFSSL_NO_TLS12)
 
-    WOLFSSL_CTX *ctx_c = NULL;
-    WOLFSSL_CTX *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL;
-    WOLFSSL *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-    DtlsRecordLayerHeader* dtlsRH;
-    size_t len;
-    byte data[1];
-
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method), 0);
-
-    /* CH1 */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* HVR */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    /* CH2 */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Server first flight */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Client flight */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Server ccs + finished */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), 1);
-
-    /* Drop the ccs */
-    dtlsRH = (DtlsRecordLayerHeader*)test_ctx.c_buff;
-    len = (size_t)((dtlsRH->length[0] << 8) | dtlsRH->length[1]);
-    ExpectIntEQ(len, 1);
-    ExpectIntEQ(dtlsRH->type, change_cipher_spec);
-    if (EXPECT_SUCCESS()) {
-        ExpectIntEQ(test_memio_drop_message(&test_ctx, 1, 0), 0);
-    }
-
-    /* Client rtx flight */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
-    /* Server ccs + finished rtx */
-    ExpectIntEQ(wolfSSL_read(ssl_s, data, sizeof(data)), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Client processes finished */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), 1);
-
-    wolfSSL_free(ssl_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_CTX_free(ctx_s);
-#endif
-    return EXPECT_RESULT();
-}
-
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS) \
-    && !defined(WOLFSSL_NO_TLS12)
-static int test_dtls_seq_num_downgrade_check_num(byte* ioBuf, int ioBufLen,
-        byte seq_num)
-{
-    EXPECT_DECLS;
-    DtlsRecordLayerHeader* dtlsRH;
-    byte sequence_number[8];
-
-    XMEMSET(&sequence_number, 0, sizeof(sequence_number));
-
-    ExpectIntGE(ioBufLen, sizeof(*dtlsRH));
-    dtlsRH = (DtlsRecordLayerHeader*)ioBuf;
-    ExpectIntEQ(dtlsRH->type, handshake);
-    ExpectIntEQ(dtlsRH->pvMajor, DTLS_MAJOR);
-    ExpectIntEQ(dtlsRH->pvMinor, DTLSv1_2_MINOR);
-    sequence_number[7] = seq_num;
-    ExpectIntEQ(XMEMCMP(sequence_number, dtlsRH->sequence_number,
-            sizeof(sequence_number)), 0);
-
-    return EXPECT_RESULT();
-}
-#endif
-
-/*
- * Make sure that we send the correct sequence number after a HelloVerifyRequest
- * and after a HelloRetryRequest. This is testing the server side as it is
- * operating statelessly and should copy the sequence number of the ClientHello.
- */
-static int test_dtls_seq_num_downgrade(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS) \
-    && !defined(WOLFSSL_NO_TLS12)
-    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_2_client_method, wolfDTLS_server_method), 0);
-
-    /* CH1 */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    ExpectIntEQ(test_dtls_seq_num_downgrade_check_num(test_ctx.s_buff,
-            test_ctx.s_len, 0), TEST_SUCCESS);
-    /* HVR */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    ExpectIntEQ(test_dtls_seq_num_downgrade_check_num(test_ctx.c_buff,
-            test_ctx.c_len, 0), TEST_SUCCESS);
-    /* CH2 */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    ExpectIntEQ(test_dtls_seq_num_downgrade_check_num(test_ctx.s_buff,
-            test_ctx.s_len, 1), TEST_SUCCESS);
-    /* Server first flight */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    ExpectIntEQ(test_dtls_seq_num_downgrade_check_num(test_ctx.c_buff,
-            test_ctx.c_len, 1), TEST_SUCCESS);
-
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-
-    wolfSSL_free(ssl_c);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_s);
-#endif
-    return EXPECT_RESULT();
-}
 
 /**
  * Make sure we don't send RSA Signature Hash Algorithms in the
@@ -35956,626 +31699,13 @@ static int test_revoked_loaded_int_cert(void)
     return EXPECT_RESULT();
 }
 
-static int test_dtls13_frag_ch_pq(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13) \
-    && defined(WOLFSSL_DTLS_CH_FRAG) && defined(WOLFSSL_HAVE_MLKEM)
-    WOLFSSL_CTX *ctx_c = NULL;
-    WOLFSSL_CTX *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL;
-    WOLFSSL *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-    const char *test_str = "test";
-    int test_str_size;
-    byte buf[255];
-#if defined(WOLFSSL_MLKEM_KYBER)
-    #if !defined(WOLFSSL_NO_KYBER1024)
-    int group = WOLFSSL_KYBER_LEVEL5;
-    const char *group_name = "KYBER_LEVEL5";
-    #elif !defined(WOLFSSL_NO_KYBER768)
-    int group = WOLFSSL_KYBER_LEVEL3;
-    const char *group_name = "KYBER_LEVEL3";
-    #else
-    int group = WOLFSSL_KYBER_LEVEL1;
-    const char *group_name = "KYBER_LEVEL1";
-    #endif
-#elif !defined(WOLFSSL_NO_ML_KEM) && !defined(WOLFSSL_TLS_NO_MLKEM_STANDALONE)
-    #if !defined(WOLFSSL_NO_ML_KEM_1024)
-    int group = WOLFSSL_ML_KEM_1024;
-    const char *group_name = "ML_KEM_1024";
-    #elif !defined(WOLFSSL_NO_ML_KEM_768)
-    int group = WOLFSSL_ML_KEM_768;
-    const char *group_name = "ML_KEM_768";
-    #else
-    int group = WOLFSSL_ML_KEM_512;
-    const char *group_name = "ML_KEM_512";
-    #endif
-#elif defined(WOLFSSL_PQC_HYBRIDS)
-    #if defined(HAVE_CURVE25519) && !defined(WOLFSSL_NO_ML_KEM_768)
-    int group = WOLFSSL_X25519MLKEM768;
-    const char *group_name = "X25519MLKEM768";
-    #elif !defined(WOLFSSL_NO_ML_KEM_768)
-    int group = WOLFSSL_SECP256R1MLKEM768;
-    const char *group_name = "SecP256r1MLKEM768";
-    #else
-    int group = WOLFSSL_SECP384R1MLKEM1024;
-    const char *group_name = "SecP384r1MLKEM1024";
-    #endif
-#endif
 
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
-    /* Add in a large post-quantum key share to make the CH long. */
-    ExpectIntEQ(wolfSSL_set_groups(ssl_c, &group, 1), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_UseKeyShare(ssl_c, group), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_dtls13_allow_ch_frag(ssl_s, 1), WOLFSSL_SUCCESS);
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-    ExpectStrEQ(wolfSSL_get_curve_name(ssl_c), group_name);
-    ExpectStrEQ(wolfSSL_get_curve_name(ssl_s), group_name);
-    test_str_size = XSTRLEN("test") + 1;
-    ExpectIntEQ(wolfSSL_write(ssl_c, test_str, test_str_size), test_str_size);
-    ExpectIntEQ(wolfSSL_read(ssl_s, buf, sizeof(buf)), test_str_size);
-    ExpectIntEQ(XSTRCMP((char*)buf, test_str), 0);
-    ExpectIntEQ(wolfSSL_write(ssl_c, test_str, test_str_size), test_str_size);
-    wolfSSL_free(ssl_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_CTX_free(ctx_s);
-#endif
-    return EXPECT_RESULT();
-}
 
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS) \
-    && defined(WOLFSSL_DTLS_MTU) && defined(WOLFSSL_DTLS_CH_FRAG) && \
-    defined(WOLFSSL_AES_256)
-static int test_dtls_frag_ch_count_records(byte* b, int len)
-{
-    DtlsRecordLayerHeader* dtlsRH;
-    int records = 0;
-    size_t recordLen;
-    while (len > 0) {
-        records++;
-        dtlsRH = (DtlsRecordLayerHeader*)b;
-        recordLen = (dtlsRH->length[0] << 8) | dtlsRH->length[1];
-        if (recordLen > (size_t)len)
-            break;
-        b += sizeof(DtlsRecordLayerHeader) + recordLen;
-        len -= sizeof(DtlsRecordLayerHeader) + recordLen;
-    }
-    return records;
-}
-#endif
 
-static int test_dtls_frag_ch(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13) \
-    && defined(WOLFSSL_DTLS_MTU) && defined(WOLFSSL_DTLS_CH_FRAG) && \
-    defined(WOLFSSL_AES_256)
-    WOLFSSL_CTX *ctx_c = NULL;
-    WOLFSSL_CTX *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL;
-    WOLFSSL *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-    static unsigned int DUMMY_MTU = 256;
-    unsigned int len;
-    unsigned char four_frag_CH[] = {
-      0x16, 0xfe, 0xfd, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-      0xda, 0x01, 0x00, 0x02, 0xdc, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-      0xce, 0xfe, 0xfd, 0xf3, 0x94, 0x01, 0x33, 0x2c, 0xcf, 0x2c, 0x47, 0xb1,
-      0xe5, 0xa1, 0x7b, 0x19, 0x3e, 0xac, 0x68, 0xdd, 0xe6, 0x17, 0x6b, 0x85,
-      0xad, 0x5f, 0xfc, 0x7f, 0x6e, 0xf0, 0xb9, 0xe0, 0x2e, 0xca, 0x47, 0x00,
-      0x00, 0x00, 0x36, 0x13, 0x01, 0x13, 0x02, 0x13, 0x03, 0xc0, 0x2c, 0xc0,
-      0x2b, 0xc0, 0x30, 0xc0, 0x2f, 0x00, 0x9f, 0x00, 0x9e, 0xcc, 0xa9, 0xcc,
-      0xa8, 0xcc, 0xaa, 0xc0, 0x27, 0xc0, 0x23, 0xc0, 0x28, 0xc0, 0x24, 0xc0,
-      0x0a, 0xc0, 0x09, 0xc0, 0x14, 0xc0, 0x13, 0x00, 0x6b, 0x00, 0x67, 0x00,
-      0x39, 0x00, 0x33, 0xcc, 0x14, 0xcc, 0x13, 0xcc, 0x15, 0x01, 0x00, 0x02,
-      0x7c, 0x00, 0x2b, 0x00, 0x03, 0x02, 0xfe, 0xfc, 0x00, 0x0d, 0x00, 0x20,
-      0x00, 0x1e, 0x06, 0x03, 0x05, 0x03, 0x04, 0x03, 0x02, 0x03, 0x08, 0x06,
-      0x08, 0x0b, 0x08, 0x05, 0x08, 0x0a, 0x08, 0x04, 0x08, 0x09, 0x06, 0x01,
-      0x05, 0x01, 0x04, 0x01, 0x03, 0x01, 0x02, 0x01, 0x00, 0x0a, 0x00, 0x0c,
-      0x00, 0x0a, 0x00, 0x19, 0x00, 0x18, 0x00, 0x17, 0x00, 0x15, 0x01, 0x00,
-      0x00, 0x16, 0x00, 0x00, 0x00, 0x33, 0x02, 0x39, 0x02, 0x37, 0x00, 0x17,
-      0x00, 0x41, 0x04, 0x94, 0xdf, 0x36, 0xd7, 0xb3, 0x90, 0x6d, 0x01, 0xa1,
-      0xe6, 0xed, 0x67, 0xf4, 0xd9, 0x9d, 0x2c, 0xac, 0x57, 0x74, 0xff, 0x19,
-      0xbe, 0x5a, 0xc9, 0x30, 0x11, 0xb7, 0x2b, 0x59, 0x47, 0x80, 0x7c, 0xa9,
-      0xb7, 0x31, 0x8c, 0x16, 0xfe, 0xfd, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-      0x00, 0x01, 0x00, 0xda, 0x01, 0x00, 0x02, 0xdc, 0x00, 0x00, 0x00, 0x00,
-      0xce, 0x00, 0x00, 0xce, 0x9e, 0x13, 0x74, 0x3b, 0x86, 0xba, 0x69, 0x1f,
-      0x12, 0xf7, 0xcd, 0x78, 0x53, 0xe8, 0x50, 0x4d, 0x71, 0x3f, 0x4b, 0x4e,
-      0xeb, 0x3e, 0xe5, 0x43, 0x54, 0x78, 0x17, 0x6d, 0x00, 0x18, 0x00, 0x61,
-      0x04, 0xd1, 0x99, 0x66, 0x4f, 0xda, 0xc7, 0x12, 0x3b, 0xff, 0xb2, 0xd6,
-      0x2f, 0x35, 0xb6, 0x17, 0x1f, 0xb3, 0xd0, 0xb6, 0x52, 0xff, 0x97, 0x8b,
-      0x01, 0xe8, 0xd9, 0x68, 0x71, 0x40, 0x02, 0xd5, 0x68, 0x3a, 0x58, 0xb2,
-      0x5d, 0xee, 0xa4, 0xe9, 0x5f, 0xf4, 0xaf, 0x3e, 0x30, 0x9c, 0x3e, 0x2b,
-      0xda, 0x61, 0x43, 0x99, 0x02, 0x35, 0x33, 0x9f, 0xcf, 0xb5, 0xd3, 0x28,
-      0x19, 0x9d, 0x1c, 0xbe, 0x69, 0x07, 0x9e, 0xfc, 0xe4, 0x8e, 0xcd, 0x86,
-      0x4a, 0x1b, 0xf0, 0xfc, 0x17, 0x94, 0x66, 0x53, 0xda, 0x24, 0x5e, 0xaf,
-      0xce, 0xec, 0x62, 0x4c, 0x06, 0xb4, 0x52, 0x94, 0xb1, 0x4a, 0x7a, 0x8c,
-      0x4f, 0x00, 0x19, 0x00, 0x85, 0x04, 0x00, 0x27, 0xeb, 0x99, 0x49, 0x7f,
-      0xcb, 0x2c, 0x46, 0x54, 0x2d, 0x93, 0x5d, 0x25, 0x92, 0x58, 0x5e, 0x06,
-      0xc3, 0x7c, 0xfb, 0x9a, 0xa7, 0xec, 0xcd, 0x9f, 0xe1, 0x6b, 0x2d, 0x78,
-      0xf5, 0x16, 0xa9, 0x20, 0x52, 0x48, 0x19, 0x0f, 0x1a, 0xd0, 0xce, 0xd8,
-      0x68, 0xb1, 0x4e, 0x7f, 0x33, 0x03, 0x7d, 0x0c, 0x39, 0xdb, 0x9c, 0x4b,
-      0xf4, 0xe7, 0xc2, 0xf5, 0xdd, 0x51, 0x9b, 0x03, 0xa8, 0x53, 0x2b, 0xe6,
-      0x00, 0x15, 0x4b, 0xff, 0xd2, 0xa0, 0x16, 0xfe, 0xfd, 0x00, 0x00, 0x00,
-      0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0xda, 0x01, 0x00, 0x02, 0xdc, 0x00,
-      0x00, 0x00, 0x01, 0x9c, 0x00, 0x00, 0xce, 0x58, 0x30, 0x10, 0x3d, 0x46,
-      0xcc, 0xca, 0x1a, 0x44, 0xc8, 0x58, 0x9b, 0x27, 0x17, 0x67, 0x31, 0x96,
-      0x8a, 0x66, 0x39, 0xf4, 0xcc, 0xc1, 0x9f, 0x12, 0x1f, 0x01, 0x30, 0x50,
-      0x16, 0xd6, 0x89, 0x97, 0xa3, 0x66, 0xd7, 0x99, 0x50, 0x09, 0x6e, 0x80,
-      0x87, 0xe4, 0xa2, 0x88, 0xae, 0xb4, 0x23, 0x57, 0x2f, 0x12, 0x60, 0xe7,
-      0x7d, 0x44, 0x2d, 0xad, 0xbe, 0xe9, 0x0d, 0x01, 0x00, 0x01, 0x00, 0xd5,
-      0xdd, 0x62, 0xee, 0xf3, 0x0e, 0xd9, 0x30, 0x0e, 0x38, 0xf3, 0x48, 0xf4,
-      0xc9, 0x8f, 0x8c, 0x20, 0xf7, 0xd3, 0xa8, 0xb3, 0x87, 0x3c, 0x98, 0x5d,
-      0x70, 0xc5, 0x03, 0x76, 0xb7, 0xd5, 0x0b, 0x7b, 0x23, 0x97, 0x6b, 0xe3,
-      0xb5, 0x18, 0xeb, 0x64, 0x55, 0x18, 0xb2, 0x8a, 0x90, 0x1a, 0x8f, 0x0e,
-      0x15, 0xda, 0xb1, 0x8e, 0x7f, 0xee, 0x1f, 0xe0, 0x3b, 0xb9, 0xed, 0xfc,
-      0x4e, 0x3f, 0x78, 0x16, 0x39, 0x95, 0x5f, 0xb7, 0xcb, 0x65, 0x55, 0x72,
-      0x7b, 0x7d, 0x86, 0x2f, 0x8a, 0xe5, 0xee, 0xf7, 0x57, 0x40, 0xf3, 0xc4,
-      0x96, 0x4f, 0x11, 0x4d, 0x85, 0xf9, 0x56, 0xfa, 0x3d, 0xf0, 0xc9, 0xa4,
-      0xec, 0x1e, 0xaa, 0x47, 0x90, 0x53, 0xdf, 0xe1, 0xb7, 0x78, 0x18, 0xeb,
-      0xdd, 0x0d, 0x89, 0xb7, 0xf6, 0x15, 0x0e, 0x55, 0x12, 0xb3, 0x23, 0x17,
-      0x0b, 0x59, 0x6f, 0x83, 0x05, 0x6b, 0xa6, 0xf8, 0x6c, 0x3a, 0x9b, 0x1b,
-      0x50, 0x93, 0x51, 0xea, 0x95, 0x2d, 0x99, 0x96, 0x38, 0x16, 0xfe, 0xfd,
-      0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03, 0x00, 0x7e, 0x01, 0x00,
-      0x02, 0xdc, 0x00, 0x00, 0x00, 0x02, 0x6a, 0x00, 0x00, 0x72, 0x2d, 0x66,
-      0x3e, 0xf2, 0x36, 0x5a, 0xf2, 0x23, 0x8f, 0x28, 0x09, 0xa9, 0x55, 0x8c,
-      0x8f, 0xc0, 0x0d, 0x61, 0x98, 0x33, 0x56, 0x87, 0x7a, 0xfd, 0xa7, 0x50,
-      0x71, 0x84, 0x2e, 0x41, 0x58, 0x00, 0x87, 0xd9, 0x27, 0xe5, 0x7b, 0xf4,
-      0x6d, 0x84, 0x4e, 0x2e, 0x0c, 0x80, 0x0c, 0xf3, 0x8a, 0x02, 0x4b, 0x99,
-      0x3a, 0x1f, 0x9f, 0x18, 0x7d, 0x1c, 0xec, 0xad, 0x60, 0x54, 0xa6, 0xa3,
-      0x2c, 0x82, 0x5e, 0xf8, 0x8f, 0xae, 0xe1, 0xc4, 0x82, 0x7e, 0x43, 0x43,
-      0xc5, 0x99, 0x49, 0x05, 0xd3, 0xf6, 0xdf, 0xa1, 0xb5, 0x2d, 0x0c, 0x13,
-      0x2f, 0x1e, 0xb6, 0x28, 0x7c, 0x5c, 0xa1, 0x02, 0x6b, 0x8d, 0xa3, 0xeb,
-      0xd4, 0x58, 0xe6, 0xa0, 0x7e, 0x6b, 0xaa, 0x09, 0x43, 0x67, 0x71, 0x87,
-      0xa5, 0xcb, 0x68, 0xf3
-    };
 
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
 
-    /* Fragment msgs */
-    ExpectIntEQ(wolfSSL_dtls_set_mtu(ssl_c, DUMMY_MTU), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_dtls_set_mtu(ssl_s, DUMMY_MTU), WOLFSSL_SUCCESS);
 
-    /* Add in some key shares to make the CH long */
-    ExpectIntEQ(wolfSSL_UseKeyShare(ssl_c, WOLFSSL_ECC_SECP256R1),
-            WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_UseKeyShare(ssl_c, WOLFSSL_ECC_SECP384R1),
-            WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_UseKeyShare(ssl_c, WOLFSSL_ECC_SECP521R1),
-            WOLFSSL_SUCCESS);
-#ifdef HAVE_FFDHE_2048
-    ExpectIntEQ(wolfSSL_UseKeyShare(ssl_c, WOLFSSL_FFDHE_2048),
-            WOLFSSL_SUCCESS);
-#endif
-#ifdef HAVE_FFDHE_3072
-    ExpectIntEQ(wolfSSL_UseKeyShare(ssl_c, WOLFSSL_FFDHE_3072),
-            WOLFSSL_SUCCESS);
-#endif
-#ifdef HAVE_FFDHE_4096
-    ExpectIntEQ(wolfSSL_UseKeyShare(ssl_c, WOLFSSL_FFDHE_4096),
-            WOLFSSL_SUCCESS);
-#endif
 
-    ExpectIntEQ(wolfSSL_dtls13_allow_ch_frag(ssl_s, 1), WOLFSSL_SUCCESS);
-
-    /* Reject fragmented first CH */
-    ExpectIntEQ(test_dtls_frag_ch_count_records(four_frag_CH,
-            sizeof(four_frag_CH)), 4);
-    len = sizeof(four_frag_CH);
-    test_memio_clear_buffer(&test_ctx, 0);
-    while (len > 0 && EXPECT_SUCCESS()) {
-        unsigned int inj_len = len > DUMMY_MTU ? DUMMY_MTU : len;
-        unsigned char *idx = four_frag_CH + sizeof(four_frag_CH) - len;
-        ExpectIntEQ(test_memio_inject_message(&test_ctx, 0, (const char *)idx,
-            inj_len), 0);
-        len -= inj_len;
-    }
-    ExpectIntEQ(test_ctx.s_len, sizeof(four_frag_CH));
-    while (test_ctx.s_len > 0 && EXPECT_SUCCESS()) {
-        int s_len = test_ctx.s_len;
-        ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-        ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-        /* Fail if we didn't advance the buffer to avoid infinite loops */
-        ExpectIntLT(test_ctx.s_len, s_len);
-    }
-    /* Expect all fragments read */
-    ExpectIntEQ(test_ctx.s_len, 0);
-    /* Expect quietly dropping fragmented first CH */
-    ExpectIntEQ(test_ctx.c_len, 0);
-
-#if defined(WOLFSSL_TLS13) && defined(HAVE_ECH)
-    /* Disable ECH as it pushes it over our MTU */
-    wolfSSL_SetEchEnable(ssl_c, 0);
-#endif
-
-    /* Limit options to make the CH a fixed length */
-    /* See wolfSSL_parse_cipher_list for reason why we provide 1.3 AND 1.2
-     * ciphersuite. This is only necessary when building with OPENSSL_EXTRA. */
-#ifdef OPENSSL_EXTRA
-    ExpectTrue(wolfSSL_set_cipher_list(ssl_c, "TLS13-AES256-GCM-SHA384"
-                                       ":DHE-RSA-AES256-GCM-SHA384"));
-#else
-    ExpectTrue(wolfSSL_set_cipher_list(ssl_c, "TLS13-AES256-GCM-SHA384"));
-#endif
-
-    /* CH1 */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Count records. Expect 1 unfragmented CH */
-    ExpectIntEQ(test_dtls_frag_ch_count_records(test_ctx.s_buff,
-            test_ctx.s_len), 1);
-    /* HRR */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    /* CH2 */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Count records. Expect fragmented CH */
-    ExpectIntGT(test_dtls_frag_ch_count_records(test_ctx.s_buff,
-            test_ctx.s_len), 1);
-
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-
-    wolfSSL_free(ssl_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_CTX_free(ctx_s);
-    ssl_c = ssl_s = NULL;
-    ctx_c = ctx_s = NULL;
-#endif
-    return EXPECT_RESULT();
-}
-
-static int test_dtls_empty_keyshare_with_cookie(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
-    WOLFSSL_CTX *ctx_s = NULL;
-    WOLFSSL *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-    unsigned char ch_empty_keyshare_with_cookie[] = {
-        0x16, 0xfe, 0xfd, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01,
-        0x12, 0x01, 0x00, 0x01, 0x06, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x01,
-        0x06, 0xfe, 0xfd, 0xfb, 0x8c, 0x9b, 0x28, 0xae, 0x50, 0x1c, 0x4d, 0xf3,
-        0xb8, 0xcf, 0x4d, 0xd8, 0x7e, 0x93, 0x13, 0x7b, 0x9e, 0xd9, 0xeb, 0xe9,
-        0x13, 0x4b, 0x0d, 0x7f, 0x2e, 0x43, 0x62, 0x8c, 0xe4, 0x57, 0x79, 0x00,
-        0x00, 0x00, 0x36, 0x13, 0x01, 0x13, 0x02, 0x13, 0x03, 0xc0, 0x2c, 0xc0,
-        0x2b, 0xc0, 0x30, 0xc0, 0x2f, 0x00, 0x9f, 0x00, 0x9e, 0xcc, 0xa9, 0xcc,
-        0xa8, 0xcc, 0xaa, 0xc0, 0x27, 0xc0, 0x23, 0xc0, 0x28, 0xc0, 0x24, 0xc0,
-        0x0a, 0xc0, 0x09, 0xc0, 0x14, 0xc0, 0x13, 0x00, 0x6b, 0x00, 0x67, 0x00,
-        0x39, 0x00, 0x33, 0xcc, 0x14, 0xcc, 0x13, 0xcc, 0x15, 0x01, 0x00, 0x00,
-        0xa6, 0x00, 0x2b, 0x00, 0x03, 0x02, 0xfe, 0xfc, 0x00, 0x2c, 0x00, 0x47,
-        0x00, 0x45, 0x20, 0xee, 0x4b, 0x17, 0x70, 0x63, 0xa0, 0x4c, 0x82, 0xbf,
-        0x43, 0x01, 0x7d, 0x8d, 0xc1, 0x1b, 0x4e, 0x9b, 0xa0, 0x3c, 0x53, 0x1f,
-        0xb7, 0xd1, 0x10, 0x81, 0xa8, 0xdf, 0xdf, 0x8c, 0x7f, 0xf3, 0x11, 0x13,
-        0x01, 0x02, 0x3d, 0x3b, 0x7d, 0x14, 0x2c, 0x31, 0xb3, 0x60, 0x72, 0x4d,
-        0xe5, 0x1a, 0xb2, 0xa3, 0x61, 0x77, 0x73, 0x03, 0x40, 0x0e, 0x5f, 0xc5,
-        0x61, 0x38, 0x43, 0x56, 0x21, 0x4a, 0x95, 0xd5, 0x35, 0xa8, 0x0d, 0x00,
-        0x0d, 0x00, 0x2a, 0x00, 0x28, 0x06, 0x03, 0x05, 0x03, 0x04, 0x03, 0x02,
-        0x03, 0xfe, 0x0b, 0xfe, 0x0e, 0xfe, 0xa0, 0xfe, 0xa3, 0xfe, 0xa5, 0x08,
-        0x06, 0x08, 0x0b, 0x08, 0x05, 0x08, 0x0a, 0x08, 0x04, 0x08, 0x09, 0x06,
-        0x01, 0x05, 0x01, 0x04, 0x01, 0x03, 0x01, 0x02, 0x01, 0x00, 0x0a, 0x00,
-        0x18, 0x00, 0x16, 0x00, 0x19, 0x00, 0x18, 0x00, 0x17, 0x00, 0x15, 0x01,
-        0x00, 0x02, 0x3a, 0x02, 0x3c, 0x02, 0x3d, 0x2f, 0x3a, 0x2f, 0x3c, 0x2f,
-        0x3d, 0x00, 0x16, 0x00, 0x00, 0x00, 0x33, 0x00, 0x02, 0x00, 0x00
-    };
-    DtlsRecordLayerHeader* dtlsRH;
-    byte sequence_number[8];
-
-    XMEMSET(&sequence_number, 0, sizeof(sequence_number));
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-    ExpectIntEQ(test_memio_setup(&test_ctx, NULL, &ctx_s, NULL, &ssl_s,
-        NULL, wolfDTLSv1_3_server_method), 0);
-    ExpectIntEQ(test_memio_inject_message(&test_ctx, 0,
-            (const char *)ch_empty_keyshare_with_cookie,
-            sizeof(ch_empty_keyshare_with_cookie)), 0);
-
-    /* CH1 */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Expect an alert. A plaintext alert should be exactly 15 bytes. */
-    ExpectIntEQ(test_ctx.c_len, 15);
-    dtlsRH = (DtlsRecordLayerHeader*)test_ctx.c_buff;
-    ExpectIntEQ(dtlsRH->type, alert);
-    ExpectIntEQ(dtlsRH->pvMajor, DTLS_MAJOR);
-    ExpectIntEQ(dtlsRH->pvMinor, DTLSv1_2_MINOR);
-    sequence_number[7] = 1;
-    ExpectIntEQ(XMEMCMP(sequence_number, dtlsRH->sequence_number,
-            sizeof(sequence_number)), 0);
-    ExpectIntEQ(dtlsRH->length[0], 0);
-    ExpectIntEQ(dtlsRH->length[1], 2);
-    ExpectIntEQ(test_ctx.c_buff[13], alert_fatal);
-    ExpectIntEQ(test_ctx.c_buff[14], illegal_parameter);
-
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_s);
-#endif
-    return EXPECT_RESULT();
-}
-
-static int test_dtls_old_seq_number(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS) && \
-    !defined(WOLFSSL_NO_TLS12)
-    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method), 0);
-
-    /* CH1 */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* HVR */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    /* CH2 */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Server first flight */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Client second flight */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Modify the sequence number */
-    {
-        DtlsRecordLayerHeader* dtlsRH = (DtlsRecordLayerHeader*)test_ctx.s_buff;
-        XMEMSET(dtlsRH->sequence_number, 0, sizeof(dtlsRH->sequence_number));
-    }
-    /* Server second flight */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Server should not do anything as a pkt was dropped */
-    ExpectIntEQ(test_ctx.c_len, 0);
-    ExpectIntEQ(test_ctx.s_len, 0);
-    /* Trigger rtx */
-    ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
-
-    /* Complete connection */
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-
-    wolfSSL_free(ssl_c);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_s);
-#endif
-    return EXPECT_RESULT();
-}
-
-static int test_dtls12_missing_finished(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS) && \
-    !defined(WOLFSSL_NO_TLS12)
-    WOLFSSL_CTX *ctx_c = NULL;
-    WOLFSSL_CTX *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL;
-    WOLFSSL *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-    const char test_str[] = "test string";
-    char test_buf[sizeof(test_str)];
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method), 0);
-
-    /* CH1 */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* HVR */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    /* CH2 */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Server first flight */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Client second flight with finished */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Server second flight with finished */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), 1);
-    /* Let's clear the output */
-    test_memio_clear_buffer(&test_ctx, 1);
-    /* Let's send some app data */
-    ExpectIntEQ(wolfSSL_write(ssl_s, test_str, sizeof(test_str)),
-                sizeof(test_str));
-    /* Client should not error out on a missing finished */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Server rtx second flight with finished */
-    ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_s), 1);
-    /* Client process rest of handshake */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), 1);
-
-    /* Let's send some app data */
-    ExpectIntEQ(wolfSSL_write(ssl_s, test_str, sizeof(test_str)),
-                sizeof(test_str));
-    ExpectIntEQ(wolfSSL_read(ssl_c, test_buf, sizeof(test_buf)),
-                sizeof(test_str));
-    ExpectBufEQ(test_buf, test_str, sizeof(test_str));
-
-    wolfSSL_free(ssl_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_CTX_free(ctx_s);
-#endif
-    return EXPECT_RESULT();
-}
-
-static int test_dtls13_missing_finished_client(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
-    WOLFSSL_CTX *ctx_c = NULL;
-    WOLFSSL_CTX *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL;
-    WOLFSSL *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-    const char test_str[] = "test string";
-    char test_buf[sizeof(test_str)];
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
-
-    /* CH1 */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* HRR */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    /* CH2 */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Server first flight with finished */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Let's clear the output */
-    test_memio_clear_buffer(&test_ctx, 1);
-    /* Let's send some app data */
-    ExpectIntEQ(wolfSSL_write(ssl_s, test_str, sizeof(test_str)),
-                sizeof(test_str));
-    /* Client second flight with finished */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Server should not error out on a missing finished */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Client rtx second flight with finished */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Server */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), 1);
-    /* Client */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), 1);
-    /* Let's send some app data */
-    ExpectIntEQ(wolfSSL_write(ssl_s, test_str, sizeof(test_str)),
-                sizeof(test_str));
-    ExpectIntEQ(wolfSSL_read(ssl_c, test_buf, sizeof(test_buf)),
-                sizeof(test_str));
-    ExpectBufEQ(test_buf, test_str, sizeof(test_str));
-
-    wolfSSL_free(ssl_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_CTX_free(ctx_s);
-#endif
-    return EXPECT_RESULT();
-}
-
-static int test_dtls13_missing_finished_server(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
-    WOLFSSL_CTX *ctx_c = NULL;
-    WOLFSSL_CTX *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL;
-    WOLFSSL *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-    const char test_str[] = "test string";
-    char test_buf[sizeof(test_str)];
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
-
-    /* CH1 */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* HRR */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    /* CH2 */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Server first flight with finished */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Client second flight with finished */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Let's clear the output */
-    test_memio_clear_buffer(&test_ctx, 0);
-    ExpectFalse(wolfSSL_is_init_finished(ssl_c));
-    /* Let's send some app data */
-    ExpectIntEQ(wolfSSL_write(ssl_c, test_str, sizeof(test_str)),
-                sizeof(test_str));
-    /* Server should not error out on a missing finished */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Client rtx second flight with finished */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Server first flight with finished */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), 1);
-    /* Let's send some app data */
-    ExpectIntEQ(wolfSSL_write(ssl_c, test_str, sizeof(test_str)),
-                sizeof(test_str));
-    ExpectIntEQ(wolfSSL_read(ssl_s, test_buf, sizeof(test_buf)),
-                sizeof(test_str));
-    ExpectBufEQ(test_buf, test_str, sizeof(test_str));
-
-    wolfSSL_free(ssl_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_CTX_free(ctx_s);
-#endif
-    return EXPECT_RESULT();
-}
-
-static int test_dtls13_finished_send_error_propagation(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
-    WOLFSSL_CTX *ctx_c = NULL;
-    WOLFSSL_CTX *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL;
-    WOLFSSL *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
-
-    /* CH1 */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* HRR */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    /* CH2 */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Server first flight with finished */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    /* Client second flight with finished - block sends to force error */
-    test_ctx.s_len = TEST_MEMIO_BUF_SZ;
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    /* Verify the error is propagated, not silently swallowed as success */
-    ExpectIntNE(wolfSSL_get_error(ssl_c, -1), 0);
-
-    wolfSSL_free(ssl_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_CTX_free(ctx_s);
-#endif
-    return EXPECT_RESULT();
-}
 
 
 #if !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER)
@@ -38110,909 +33240,6 @@ static int test_wolfSSL_inject(void)
 int testAll = 1;
 int stopOnFail = 0;
 
-/*----------------------------------------------------------------------------*/
-/* LMS tests                                                                  */
-/*----------------------------------------------------------------------------*/
-int test_wc_LmsKey_sign_verify(void);
-int test_wc_LmsKey_reload_cache(void);
-int test_rfc9802_lms_x509_verify(void);
-int test_rfc9802_xmss_x509_verify(void);
-
-#if defined(WOLFSSL_HAVE_LMS) && !defined(WOLFSSL_LMS_VERIFY_ONLY)
-
-#include <wolfssl/wolfcrypt/wc_lms.h>
-
-#define LMS_TEST_PRIV_KEY_FILE "/tmp/wolfssl_test_lms.key"
-
-static int test_lms_write_key(const byte* priv, word32 privSz, void* context)
-{
-    FILE* f = fopen((const char*)context, "wb");
-    int ret = WC_LMS_RC_SAVED_TO_NV_MEMORY;
-    if (f == NULL)
-        return -1;
-    if (fwrite(priv, 1, privSz, f) != privSz)
-        ret = -1;
-    fclose(f);
-    return ret;
-}
-
-static int test_lms_read_key(byte* priv, word32 privSz, void* context)
-{
-    FILE* f = fopen((const char*)context, "rb");
-    if (f == NULL)
-        return -1;
-    if (fread(priv, 1, privSz, f) == 0) {
-        fclose(f);
-        return -1;
-    }
-    fclose(f);
-    return WC_LMS_RC_READ_TO_MEMORY;
-}
-
-/* Helper: init an LMS key with callbacks and L1-H10-W8 params */
-static int test_lms_init_key(LmsKey* key, WC_RNG* rng)
-{
-    int ret;
-
-    ret = wc_LmsKey_Init(key, NULL, INVALID_DEVID);
-    if (ret != 0) return ret;
-
-#if !defined(WOLFSSL_LMS_MAX_HEIGHT) || (WOLFSSL_LMS_MAX_HEIGHT >= 10)
-    ret = wc_LmsKey_SetParameters(key, 1, 10, 8);
-#else
-    ret = wc_LmsKey_SetParameters(key, 1, 5, 8);
-#endif
-    if (ret != 0) return ret;
-
-    ret = wc_LmsKey_SetWriteCb(key, test_lms_write_key);
-    if (ret != 0) return ret;
-
-    ret = wc_LmsKey_SetReadCb(key, test_lms_read_key);
-    if (ret != 0) return ret;
-
-    ret = wc_LmsKey_SetContext(key, (void*)LMS_TEST_PRIV_KEY_FILE);
-    if (ret != 0) return ret;
-
-    (void)rng;
-    return 0;
-}
-
-#endif /* WOLFSSL_HAVE_LMS && !WOLFSSL_LMS_VERIFY_ONLY */
-
-/*
- * Test basic LMS sign/verify with multiple signings.
- * Uses L1-H10-W8 (1024 total signatures, 32-entry leaf cache).
- */
-int test_wc_LmsKey_sign_verify(void)
-{
-    EXPECT_DECLS;
-#if defined(WOLFSSL_HAVE_LMS) && !defined(WOLFSSL_LMS_VERIFY_ONLY)
-    LmsKey  key;
-    WC_RNG  rng;
-    byte    msg[] = "test message for LMS signing";
-    byte    sig[2048];
-    word32  sigSz;
-    int     i;
-    int     numSigs = 5;
-
-    ExpectIntEQ(wc_InitRng(&rng), 0);
-
-    remove(LMS_TEST_PRIV_KEY_FILE);
-    ExpectIntEQ(test_lms_init_key(&key, &rng), 0);
-    ExpectIntEQ(wc_LmsKey_MakeKey(&key, &rng), 0);
-
-    for (i = 0; i < numSigs; i++) {
-        sigSz = sizeof(sig);
-        ExpectIntEQ(wc_LmsKey_Sign(&key, sig, &sigSz, msg, sizeof(msg)), 0);
-        ExpectIntEQ(wc_LmsKey_Verify(&key, sig, sigSz, msg, sizeof(msg)), 0);
-    }
-
-    wc_LmsKey_Free(&key);
-    wc_FreeRng(&rng);
-    remove(LMS_TEST_PRIV_KEY_FILE);
-#endif
-    return EXPECT_RESULT();
-}
-
-/*
- * Test LMS key reload after advancing past the leaf cache window.
- *
- * Reproduces a heap-buffer-overflow bug in wc_lms_treehash_init() where the
- * leaf cache write uses (i * hash_len) instead of ((i - leaf->idx) * hash_len).
- * When q > max_cb (default 32), wc_LmsKey_Reload calls wc_hss_init_auth_path
- * which calls wc_lms_treehash_init with q > 0, causing writes past the end of
- * the leaf cache buffer.
- *
- * Reproduction steps:
- *   1. Generate L1-H10-W8 key (cacheBits=5, max_cb=32)
- *   2. Sign 33 times to advance q past the cache window
- *   3. Free the key and reload from persisted state
- *   4. Sign and verify after reload
- *
- * Without the fix: heap-buffer-overflow at wc_lms_impl.c:1965
- * With the fix:    all operations succeed, signatures verify
- */
-int test_wc_LmsKey_reload_cache(void)
-{
-    EXPECT_DECLS;
-#if defined(WOLFSSL_HAVE_LMS) && !defined(WOLFSSL_LMS_VERIFY_ONLY) && \
-    (!defined(WOLFSSL_LMS_MAX_HEIGHT) || (WOLFSSL_LMS_MAX_HEIGHT >= 10))
-    LmsKey  key;
-    LmsKey  vkey;
-    WC_RNG  rng;
-    byte    msg[] = "test message for LMS signing";
-    byte    sig[2048];
-    word32  sigSz;
-    byte    pub[64];
-    word32  pubSz = sizeof(pub);
-    int     i;
-    /* Sign 33 times to advance q past the 32-entry cache window. */
-    int     preSigs = 33;
-
-    ExpectIntEQ(wc_InitRng(&rng), 0);
-
-    /* Phase 1: Generate key and sign past cache window */
-    remove(LMS_TEST_PRIV_KEY_FILE);
-    ExpectIntEQ(test_lms_init_key(&key, &rng), 0);
-    ExpectIntEQ(wc_LmsKey_MakeKey(&key, &rng), 0);
-
-    for (i = 0; i < preSigs; i++) {
-        sigSz = sizeof(sig);
-        ExpectIntEQ(wc_LmsKey_Sign(&key, sig, &sigSz, msg, sizeof(msg)), 0);
-    }
-
-    /* Save public key for verification after reload */
-    ExpectIntEQ(wc_LmsKey_ExportPubRaw(&key, pub, &pubSz), 0);
-
-    wc_LmsKey_Free(&key);
-
-    /* Phase 2: Reload key. Triggers wc_lms_treehash_init with q=33 */
-    ExpectIntEQ(test_lms_init_key(&key, &rng), 0);
-    ExpectIntEQ(wc_LmsKey_Reload(&key), 0);
-
-    /* Phase 3: Sign after reload and verify with separate verify-only key */
-    sigSz = sizeof(sig);
-    ExpectIntEQ(wc_LmsKey_Sign(&key, sig, &sigSz, msg, sizeof(msg)), 0);
-
-    ExpectIntEQ(wc_LmsKey_Init(&vkey, NULL, INVALID_DEVID), 0);
-#if !defined(WOLFSSL_LMS_MAX_HEIGHT) || (WOLFSSL_LMS_MAX_HEIGHT >= 10)
-    ExpectIntEQ(wc_LmsKey_SetParameters(&vkey, 1, 10, 8), 0);
-#else
-    ExpectIntEQ(wc_LmsKey_SetParameters(&vkey, 1, 5, 8), 0);
-#endif
-    ExpectIntEQ(wc_LmsKey_ImportPubRaw(&vkey, pub, pubSz), 0);
-    ExpectIntEQ(wc_LmsKey_Verify(&vkey, sig, sigSz, msg, sizeof(msg)), 0);
-
-    wc_LmsKey_Free(&vkey);
-    wc_LmsKey_Free(&key);
-    wc_FreeRng(&rng);
-    remove(LMS_TEST_PRIV_KEY_FILE);
-#endif
-    return EXPECT_RESULT();
-}
-
-/*----------------------------------------------------------------------------*/
-/* RFC 9802 (HSS/LMS and XMSS/XMSS^MT in X.509) tests                         */
-/*----------------------------------------------------------------------------*/
-
-/* For every committed self-signed test certificate confirm:
- *   - wc_ParseCert succeeds on the RFC 9802 AlgorithmIdentifier encoding
- *     (OID-only SEQUENCE, no NULL parameters)
- *   - keyOID and signatureOID are set to the expected values
- *   - loading as a trust anchor and verifying the same bytes through
- *     wolfSSL_CertManagerVerifyBuffer exercises the ConfirmSignature
- *     path and succeeds on a valid cert
- *   - flipping a byte in the signature AND flipping a byte in the
- *     TBSCertificate both cause verification to fail.
- *
- * Test vectors are in certs/lms/ and certs/xmss/, generated with Bouncy
- * Castle 1.81. BC's default XMSS / XMSS^MT X.509 encoding uses pre-
- * standard ISARA OIDs and wraps the raw RFC 8391 pub key in an OCTET
- * STRING, so the fixtures were produced with a small generator that
- * overrides the AlgorithmIdentifier and SPKI to match RFC 9802. */
-#if (defined(WOLFSSL_HAVE_LMS) || defined(WOLFSSL_HAVE_XMSS)) && \
-    !defined(NO_FILESYSTEM) && !defined(NO_CERTS)
-/* Sanity bound on a test fixture cert. The largest BC-generated
- * fixture we ship (XMSS^MT 40/8) is ~19 KiB; 1 MiB is well above
- * any realistic RFC 9802 cert and catches a wild XFTELL. Typed as
- * long to match XFTELL's return so the size comparison below isn't
- * a mixed long-vs-int compare. */
-#define RFC9802_TEST_MAX_CERT_SIZE ((long)(1L << 20))
-
-/* Load a whole file into a freshly-allocated buffer. Caller frees. */
-static int rfc9802_load_file(const char* path, byte** out, int* outLen)
-{
-    EXPECT_DECLS;
-    XFILE  f = XBADFILE;
-    long   sz = 0;
-    size_t got = 0;
-    byte*  buf = NULL;
-
-    *out = NULL;
-    *outLen = 0;
-    ExpectTrue((f = XFOPEN(path, "rb")) != XBADFILE);
-    if (f == XBADFILE)
-        return TEST_FAIL;
-    if (XFSEEK(f, 0, XSEEK_END) == 0)
-        sz = XFTELL(f);
-    (void)XFSEEK(f, 0, XSEEK_SET);
-    ExpectIntGT(sz, 0);
-    ExpectIntLT(sz, RFC9802_TEST_MAX_CERT_SIZE);
-    /* Hard-fail before XMALLOC if XFSEEK / XFTELL produced an unusable
-     * size: ExpectInt* records the failure but doesn't short-circuit,
-     * so without this guard a -1 from XFTELL would cast to a multi-GiB
-     * (size_t) allocation, and a 0 would request a zero-byte malloc. */
-    if (sz <= 0 || sz >= RFC9802_TEST_MAX_CERT_SIZE) {
-        XFCLOSE(f);
-        return TEST_FAIL;
-    }
-    ExpectNotNull(buf = (byte*)XMALLOC((size_t)sz, NULL,
-        DYNAMIC_TYPE_TMP_BUFFER));
-    if (buf != NULL) {
-        got = XFREAD(buf, 1, (size_t)sz, f);
-        ExpectIntEQ(got, (size_t)sz);
-        /* On a short read the caller would otherwise proceed with a
-         * partially-initialized buffer and produce cascading parse
-         * failures driven by the uninitialized tail. Free here so the
-         * caller's `if (buf == NULL) return TEST_FAIL;` short-circuits
-         * cleanly with a single recorded failure. */
-        if (got != (size_t)sz) {
-            XFREE(buf, NULL, DYNAMIC_TYPE_TMP_BUFFER);
-            buf = NULL;
-            sz = 0;
-        }
-    }
-    XFCLOSE(f);
-    *out = buf;
-    *outLen = (int)sz;
-    return EXPECT_RESULT();
-}
-
-static int rfc9802_verify_one_cert(const char* path, word32 expectedKeyOID,
-    word32 expectedSigOID)
-{
-    EXPECT_DECLS;
-    byte*                 buf = NULL;
-    byte*                 tampered = NULL;
-    int                   bytes = 0;
-    DecodedCert           cert;
-    WOLFSSL_CERT_MANAGER* cm = NULL;
-    word32                certBegin = 0;
-    word32                sigIndex = 0;
-
-    ExpectIntEQ(rfc9802_load_file(path, &buf, &bytes), TEST_SUCCESS);
-    if (buf == NULL)
-        return TEST_FAIL;
-
-    /* Parse + check OIDs, capture certBegin and sigIndex for later tamper. */
-    wc_InitDecodedCert(&cert, buf, (word32)bytes, NULL);
-    ExpectIntEQ(wc_ParseCert(&cert, CERT_TYPE, NO_VERIFY, NULL), 0);
-    ExpectIntEQ((int)cert.keyOID, (int)expectedKeyOID);
-    ExpectIntEQ((int)cert.signatureOID, (int)expectedSigOID);
-    certBegin = cert.certBegin;
-    sigIndex  = cert.sigIndex;
-    wc_FreeDecodedCert(&cert);
-
-    /* Full verify against a self-installed trust anchor. */
-    ExpectNotNull(cm = wolfSSL_CertManagerNew());
-    ExpectIntEQ(wolfSSL_CertManagerLoadCABuffer(cm, buf, (long)bytes,
-        WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_CertManagerVerifyBuffer(cm, buf, (long)bytes,
-        WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
-    if (cm != NULL) {
-        wolfSSL_CertManagerFree(cm);
-        cm = NULL;
-    }
-
-    ExpectNotNull(tampered = (byte*)XMALLOC((size_t)bytes, NULL,
-        DYNAMIC_TYPE_TMP_BUFFER));
-
-    /* Negative 1: flip a byte inside the signatureValue BIT STRING.
-     * Everything after sigIndex is the signatureAlgorithm + the BIT
-     * STRING payload, so flipping the last byte is always inside the
-     * signature content. */
-    if (tampered != NULL) {
-        XMEMCPY(tampered, buf, (size_t)bytes);
-        tampered[bytes - 1] ^= 0x01;
-        ExpectNotNull(cm = wolfSSL_CertManagerNew());
-        ExpectIntEQ(wolfSSL_CertManagerLoadCABuffer(cm, buf, (long)bytes,
-            WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
-        ExpectIntNE(wolfSSL_CertManagerVerifyBuffer(cm, tampered,
-            (long)bytes, WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
-        if (cm != NULL) {
-            wolfSSL_CertManagerFree(cm);
-            cm = NULL;
-        }
-    }
-
-    /* Negative 2: flip a byte at the midpoint of the TBSCertificate. The
-     * TBS is the first element of the outer Certificate SEQUENCE and
-     * its bytes lie between (certBegin + outerSeqHeader) and sigIndex.
-     * Picking the midpoint ensures we're inside TBS regardless of the
-     * fixture's DN / extensions layout. */
-    if (tampered != NULL && sigIndex > certBegin + 8U) {
-        word32 midTbs = certBegin + 8 + ((sigIndex - (certBegin + 8)) / 2);
-        XMEMCPY(tampered, buf, (size_t)bytes);
-        tampered[midTbs] ^= 0x01;
-        ExpectNotNull(cm = wolfSSL_CertManagerNew());
-        ExpectIntEQ(wolfSSL_CertManagerLoadCABuffer(cm, buf, (long)bytes,
-            WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
-        ExpectIntNE(wolfSSL_CertManagerVerifyBuffer(cm, tampered,
-            (long)bytes, WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
-        if (cm != NULL) {
-            wolfSSL_CertManagerFree(cm);
-            cm = NULL;
-        }
-    }
-
-    /* The fixtures MUST carry a KeyUsage extension with at least one of
-     * digitalSignature / nonRepudiation / keyCertSign / cRLSign set per
-     * RFC 9802 sec 3. Re-parse and assert that wolfSSL recorded a non-
-     * empty set of KeyUsage bits from one of those values. */
-    wc_InitDecodedCert(&cert, buf, (word32)bytes, NULL);
-    ExpectIntEQ(wc_ParseCert(&cert, CERT_TYPE, NO_VERIFY, NULL), 0);
-    ExpectIntEQ(cert.extKeyUsageSet, 1);
-    ExpectIntNE(cert.extKeyUsage & (KEYUSE_DIGITAL_SIG | KEYUSE_CONTENT_COMMIT |
-        KEYUSE_KEY_CERT_SIGN | KEYUSE_CRL_SIGN), 0);
-    wc_FreeDecodedCert(&cert);
-
-    XFREE(tampered, NULL, DYNAMIC_TYPE_TMP_BUFFER);
-    XFREE(buf, NULL, DYNAMIC_TYPE_TMP_BUFFER);
-    return EXPECT_RESULT();
-}
-#endif
-
-/* Direct wolfCrypt-level negative tests for the parameter-derivation
- * helpers used by the RFC 9802 parse path. These exercise failure modes
- * (unknown algorithm bytes, truncated inputs, mismatches) that a real
- * cert body wouldn't easily reach. */
-#if defined(WOLFSSL_HAVE_LMS)
-static int rfc9802_lms_import_negative(void)
-{
-    EXPECT_DECLS;
-    LmsKey key;
-    /* 60-byte buffer matches HSS_PUBLIC_KEY_LEN(32), just like a valid
-     * SHA-256/M32/H5 key; the algorithm-type bytes are junk so param
-     * derivation must fail cleanly. */
-    byte   junk[60];
-
-    XMEMSET(junk, 0, sizeof(junk));
-    /* levels=1, lmsType=0xFFFFFFFF, lmOtsType=0xFFFFFFFF. */
-    junk[3] = 1;
-    XMEMSET(junk + 4, 0xFF, 4);
-    XMEMSET(junk + 8, 0xFF, 4);
-
-    /* Unknown algorithm types must be rejected. */
-    ExpectIntEQ(wc_LmsKey_Init(&key, NULL, INVALID_DEVID), 0);
-    ExpectIntEQ(wc_LmsKey_ImportPubRaw(&key, junk, sizeof(junk)),
-        WC_NO_ERR_TRACE(NOT_COMPILED_IN));
-    wc_LmsKey_Free(&key);
-
-    /* Too-short buffer: only L + lmsType, no lmOtsType. */
-    ExpectIntEQ(wc_LmsKey_Init(&key, NULL, INVALID_DEVID), 0);
-    ExpectIntEQ(wc_LmsKey_ImportPubRaw(&key, junk, 8),
-        WC_NO_ERR_TRACE(BUFFER_E));
-    wc_LmsKey_Free(&key);
-
-#if !defined(WOLFSSL_NO_LMS_SHA256_256)
-    /* The two cases below pin specific SHA-256/M32 parameter codes
-     * (L1_H5_W8, L1_H5_W4, L1_H10_W2). Skip them in builds where the
-     * SHA-256/M32 family is disabled -- the family-agnostic checks
-     * above (junk algorithm types, too-short buffer, GetSigLen on
-     * unconfigured key) still cover the universal invariants. */
-
-    /* Pre-set params that disagree with the raw key's algorithm bytes:
-     * configure H=5/W=8 but feed buffer that claims H=10 / W=2. */
-    XMEMSET(junk, 0, sizeof(junk));
-    junk[3] = 1;       /* levels=1     */
-    junk[7] = 6;       /* lmsType = LMS_SHA256_M32_H10 = 6 */
-    junk[11] = 2;      /* lmOtsType = LMOTS_SHA256_N32_W2 = 2 */
-    ExpectIntEQ(wc_LmsKey_Init(&key, NULL, INVALID_DEVID), 0);
-    ExpectIntEQ(wc_LmsKey_SetParameters(&key, 1, 5, 8), 0);
-    ExpectIntEQ(wc_LmsKey_ImportPubRaw(&key, junk, sizeof(junk)),
-        WC_NO_ERR_TRACE(BAD_FUNC_ARG));
-    wc_LmsKey_Free(&key);
-#endif /* !WOLFSSL_NO_LMS_SHA256_256 */
-
-    /* GetSigLen on a key with no params set must not NULL-deref the
-     * params pointer; it must return BAD_FUNC_ARG instead. */
-    {
-        word32 sigLen = 0;
-        ExpectIntEQ(wc_LmsKey_Init(&key, NULL, INVALID_DEVID), 0);
-        ExpectIntEQ(wc_LmsKey_GetSigLen(&key, &sigLen),
-            WC_NO_ERR_TRACE(BAD_FUNC_ARG));
-        wc_LmsKey_Free(&key);
-    }
-
-#if !defined(WOLFSSL_NO_LMS_SHA256_256)
-    /* Partial-write invariant: a length mismatch after a successful
-     * auto-derive must leave key->params NULL. Build a buffer whose
-     * leading u32str(L) || lmsType || lmOtsType identifies a known
-     * parameter set, but truncate to one byte less than the real pub
-     * key length so the post-derive length check fails. */
-    {
-        byte truncated[59];   /* HSS_PUBLIC_KEY_LEN(32) is 60 */
-        XMEMSET(truncated, 0, sizeof(truncated));
-        truncated[3] = 1;     /* L = 1 */
-        truncated[7] = 5;     /* lmsType = LMS_SHA256_M32_H5 */
-        truncated[11] = 4;    /* lmOtsType = LMOTS_SHA256_N32_W4 */
-        ExpectIntEQ(wc_LmsKey_Init(&key, NULL, INVALID_DEVID), 0);
-        ExpectNull(key.params);
-        ExpectIntEQ(wc_LmsKey_ImportPubRaw(&key, truncated,
-            sizeof(truncated)), WC_NO_ERR_TRACE(BUFFER_E));
-        ExpectNull(key.params);
-        wc_LmsKey_Free(&key);
-    }
-#endif /* !WOLFSSL_NO_LMS_SHA256_256 */
-
-    return EXPECT_RESULT();
-}
-#endif
-
-#if defined(WOLFSSL_HAVE_XMSS)
-static int rfc9802_xmss_import_negative(void)
-{
-    EXPECT_DECLS;
-    XmssKey key;
-    byte    junk[8];
-
-    XMEMSET(junk, 0, sizeof(junk));
-
-    /* Too-short buffer. */
-    ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
-    ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, junk, 2, 0),
-        WC_NO_ERR_TRACE(BUFFER_E));
-    wc_XmssKey_Free(&key);
-
-    /* Unknown OID (all-zero) for both XMSS and XMSS^MT. */
-    ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
-    ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, junk, sizeof(junk), 0),
-        WC_NO_ERR_TRACE(NOT_COMPILED_IN));
-    wc_XmssKey_Free(&key);
-    ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
-    ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, junk, sizeof(junk), 1),
-        WC_NO_ERR_TRACE(NOT_COMPILED_IN));
-    wc_XmssKey_Free(&key);
-
-    /* NULL key / input. */
-    ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(NULL, junk, sizeof(junk), 0),
-        WC_NO_ERR_TRACE(BAD_FUNC_ARG));
-    ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
-    ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, NULL, 8, 0),
-        WC_NO_ERR_TRACE(BAD_FUNC_ARG));
-    wc_XmssKey_Free(&key);
-
-    /* GetSigLen on a key with no params set must not NULL-deref the
-     * params pointer; it must return BAD_FUNC_ARG instead. */
-    {
-        word32 sigLen = 0;
-        ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
-        ExpectIntEQ(wc_XmssKey_GetSigLen(&key, &sigLen),
-            WC_NO_ERR_TRACE(BAD_FUNC_ARG));
-        wc_XmssKey_Free(&key);
-    }
-
-    /* Once params have been configured (state != INITED), the OID
-     * prefix in the raw key MUST match key->oid and is_xmssmt MUST
-     * match key->is_xmssmt. Set XMSS-SHA2_10_256 and feed a valid-
-     * sized buffer whose 4-byte OID prefix is bogus -> BAD_FUNC_ARG. */
-    {
-        byte mismatch[XMSS_SHA256_PUBLEN];
-        ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
-        ExpectIntEQ(wc_XmssKey_SetParamStr(&key, "XMSS-SHA2_10_256"), 0);
-        XMEMSET(mismatch, 0, sizeof(mismatch));
-        mismatch[3] = 0x77; /* nonsense OID */
-        ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, mismatch,
-            sizeof(mismatch), 0), WC_NO_ERR_TRACE(BAD_FUNC_ARG));
-        /* Same buffer with the correct OID, but is_xmssmt hint
-         * contradicts the configured family -> BAD_FUNC_ARG. */
-        mismatch[3] = 0x01; /* WC_XMSS_OID_SHA2_10_256 */
-        ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, mismatch,
-            sizeof(mismatch), 1), WC_NO_ERR_TRACE(BAD_FUNC_ARG));
-        wc_XmssKey_Free(&key);
-    }
-
-    /* Partial-write invariant: a length mismatch after a successful
-     * auto-derive must leave the key in its INITED state, with
-     * key->params NULL. */
-    {
-        byte truncated[XMSS_SHA256_PUBLEN - 1];
-        XMEMSET(truncated, 0, sizeof(truncated));
-        truncated[3] = 0x01;
-        ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
-        ExpectNull(key.params);
-        ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, truncated,
-            sizeof(truncated), 0), WC_NO_ERR_TRACE(BUFFER_E));
-        ExpectNull(key.params);
-        wc_XmssKey_Free(&key);
-    }
-
-    /* is_xmssmt disambiguation: XMSS oid=1 and XMSS^MT oid=1 share
-     * the wire-numeric value but resolve to different parameter sets.
-     * Importing the same 68-byte buffer with hint=0 vs hint=1 must
-     * land in different tables and produce distinct is_xmssmt. */
-    {
-        byte buf[XMSS_SHA256_PUBLEN];
-        XMEMSET(buf, 0, sizeof(buf));
-        buf[3] = 0x01;
-
-        ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
-        ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, buf, sizeof(buf), 0), 0);
-        ExpectIntEQ((int)key.is_xmssmt, 0);
-        wc_XmssKey_Free(&key);
-
-        ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
-        ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, buf, sizeof(buf), 1), 0);
-        ExpectIntEQ((int)key.is_xmssmt, 1);
-        wc_XmssKey_Free(&key);
-    }
-
-    /* Lenient state: re-importing the same pub key into a VERIFYONLY
-     * key (params set, no private material) succeeds. The second
-     * call exercises the lenient-state branch. */
-    {
-        byte buf[XMSS_SHA256_PUBLEN];
-        XMEMSET(buf, 0, sizeof(buf));
-        buf[3] = 0x01;
-
-        ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
-        ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, buf, sizeof(buf), 0), 0);
-        ExpectIntEQ((int)key.state, (int)WC_XMSS_STATE_VERIFYONLY);
-        ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, buf, sizeof(buf), 0), 0);
-        ExpectIntEQ((int)key.state, (int)WC_XMSS_STATE_VERIFYONLY);
-        wc_XmssKey_Free(&key);
-    }
-
-    /* Strict signature-length check: wc_XmssKey_Verify rejects any
-     * sigLen != key->params->sig_len. This guards every consumer
-     * (RFC 9802 X.509, PKCS#7, CMS, ...) against a longer wrapper that
-     * happens to start with a valid signature. Construct a key in
-     * VERIFYONLY state, then verify with sig_len + 1 and sig_len - 1
-     * byte buffers; both must fail with BUFFER_E before any crypto
-     * runs. The buffer contents are irrelevant since the length check
-     * fires first. */
-    {
-        byte    pub[XMSS_SHA256_PUBLEN];
-        byte*   sigBuf = NULL;
-        word32  sigLen = 0;
-        const byte msg[1] = { 0 };
-
-        XMEMSET(pub, 0, sizeof(pub));
-        pub[3] = 0x01;
-        ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
-        ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, pub, sizeof(pub), 0), 0);
-        ExpectIntEQ((int)key.state, (int)WC_XMSS_STATE_VERIFYONLY);
-        ExpectIntEQ(wc_XmssKey_GetSigLen(&key, &sigLen), 0);
-        ExpectIntGT(sigLen, 0);
-        ExpectNotNull(sigBuf = (byte*)XMALLOC((size_t)sigLen + 1, NULL,
-            DYNAMIC_TYPE_TMP_BUFFER));
-        if (sigBuf != NULL) {
-            XMEMSET(sigBuf, 0, (size_t)sigLen + 1);
-            ExpectIntEQ(wc_XmssKey_Verify(&key, sigBuf, sigLen + 1,
-                msg, (int)sizeof(msg)), WC_NO_ERR_TRACE(BUFFER_E));
-            ExpectIntEQ(wc_XmssKey_Verify(&key, sigBuf, sigLen - 1,
-                msg, (int)sizeof(msg)), WC_NO_ERR_TRACE(BUFFER_E));
-            XFREE(sigBuf, NULL, DYNAMIC_TYPE_TMP_BUFFER);
-        }
-        wc_XmssKey_Free(&key);
-    }
-
-    /* BAD_STATE_E branch: WC_XMSS_STATE_OK must be rejected. Reaching
-     * OK normally requires a successful private-key Reload / sign,
-     * which is unavailable in WOLFSSL_XMSS_VERIFY_ONLY builds. Force
-     * the state directly to exercise the rejection without coupling
-     * this helper to the signing test fixture; sk stays NULL so Free
-     * is still safe. */
-    {
-        byte pub[XMSS_SHA256_PUBLEN];
-
-        XMEMSET(pub, 0, sizeof(pub));
-        pub[3] = 0x01;
-        ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
-        ExpectIntEQ(wc_XmssKey_SetParamStr(&key, "XMSS-SHA2_10_256"), 0);
-        key.state = WC_XMSS_STATE_OK;
-        ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, pub, sizeof(pub), 0),
-            WC_NO_ERR_TRACE(BAD_STATE_E));
-        wc_XmssKey_Free(&key);
-    }
-
-    return EXPECT_RESULT();
-}
-#endif
-
-/* Walk the AlgorithmIdentifier SEQUENCE that begins at sigIndex and
- * locate the byte offset of the last byte of its OID content. Handles
- * both short-form (length < 128) and long-form DER length encodings,
- * so a future fixture-regenerator that emits longer OIDs / SEQUENCEs
- * still drives this test rather than tripping the loud-fail branch.
- *
- * Returns 0 on success with *oidLastByte set; returns -1 on any DER
- * shape mismatch. */
-#if defined(WOLFSSL_HAVE_XMSS) && !defined(NO_FILESYSTEM) && !defined(NO_CERTS)
-static int rfc9802_find_sig_alg_oid_last_byte(const byte* buf, word32 bufLen,
-    word32 sigIndex, word32* oidLastByte)
-{
-    word32 idx = sigIndex;
-    word32 oidContentLen = 0;
-
-    /* AlgorithmIdentifier ::= SEQUENCE { algorithm OID, ... } */
-    if (idx >= bufLen || buf[idx] != 0x30)
-        return -1;
-    idx++;
-    /* Skip SEQUENCE length (short or long form). */
-    if (idx >= bufLen)
-        return -1;
-    if (buf[idx] < 0x80) {
-        idx++;
-    }
-    else {
-        word32 nbytes = (word32)(buf[idx] & 0x7F);
-        if (nbytes == 0 || nbytes > 4 || idx + 1 + nbytes > bufLen)
-            return -1;
-        idx += 1 + nbytes;
-    }
-    /* algorithm OID tag. */
-    if (idx >= bufLen || buf[idx] != 0x06)
-        return -1;
-    idx++;
-    /* OID length (short or long form). */
-    if (idx >= bufLen)
-        return -1;
-    if (buf[idx] < 0x80) {
-        oidContentLen = buf[idx];
-        idx++;
-    }
-    else {
-        word32 nbytes = (word32)(buf[idx] & 0x7F);
-        word32 i;
-        if (nbytes == 0 || nbytes > 4 || idx + 1 + nbytes > bufLen)
-            return -1;
-        for (i = 0; i < nbytes; i++)
-            oidContentLen = (oidContentLen << 8) | buf[idx + 1 + i];
-        idx += 1 + nbytes;
-    }
-    if (oidContentLen == 0 || idx + oidContentLen > bufLen)
-        return -1;
-    *oidLastByte = idx + oidContentLen - 1;
-    return 0;
-}
-
-/* Helper: load fixture, locate last byte of outer signatureAlgorithm
- * OID, patch it from `expected` to `swap`, and assert that verifying
- * the patched cert against itself as a trust anchor fails. */
-static int rfc9802_assert_oid_patch_breaks_verify(const char* path,
-    byte expectedLastByte, byte patchedLastByte)
-{
-    EXPECT_DECLS;
-    byte*                 buf = NULL;
-    int                   bytes = 0;
-    DecodedCert           cert;
-    WOLFSSL_CERT_MANAGER* cm = NULL;
-    word32                sigIndex = 0;
-    word32                lastOidByte = 0;
-
-    ExpectIntEQ(rfc9802_load_file(path, &buf, &bytes), TEST_SUCCESS);
-    if (buf == NULL)
-        return TEST_FAIL;
-
-    wc_InitDecodedCert(&cert, buf, (word32)bytes, NULL);
-    ExpectIntEQ(wc_ParseCert(&cert, CERT_TYPE, NO_VERIFY, NULL), 0);
-    sigIndex = cert.sigIndex;
-    wc_FreeDecodedCert(&cert);
-
-    ExpectIntEQ(rfc9802_find_sig_alg_oid_last_byte(buf, (word32)bytes,
-        sigIndex, &lastOidByte), 0);
-    /* Sanity-check the fixture matches the family the caller asserted,
-     * so a future regenerator swapping fixtures fails loudly here
-     * rather than silently testing the wrong direction. */
-    ExpectIntEQ((int)buf[lastOidByte], (int)expectedLastByte);
-
-    if (lastOidByte < (word32)bytes &&
-            buf[lastOidByte] == expectedLastByte) {
-        buf[lastOidByte] = patchedLastByte;
-        ExpectNotNull(cm = wolfSSL_CertManagerNew());
-        /* After the patch the cert's outer signatureAlgorithm and SPKI
-         * disagree. Verification must fail somewhere (at parse, at
-         * load, or at ConfirmSignature). The load is best-effort -
-         * some shape changes get caught there, others only at verify. */
-        (void)wolfSSL_CertManagerLoadCABuffer(cm, buf, (long)bytes,
-            WOLFSSL_FILETYPE_ASN1);
-        ExpectIntNE(wolfSSL_CertManagerVerifyBuffer(cm, buf,
-            (long)bytes, WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
-        if (cm != NULL) {
-            wolfSSL_CertManagerFree(cm);
-            cm = NULL;
-        }
-    }
-
-    XFREE(buf, NULL, DYNAMIC_TYPE_TMP_BUFFER);
-    return EXPECT_RESULT();
-}
-
-/* X.509-level negative: swap the outer signatureAlgorithm OID byte so
- * the cert declares XMSS where the SPKI is XMSS^MT, and vice versa.
- * SigOidMatchesKeyOid must reject both directions before any crypto. */
-static int rfc9802_xmss_sig_oid_mismatch(void)
-{
-    EXPECT_DECLS;
-    /* XMSS sigOID ends 0x22; XMSS^MT sigOID ends 0x23. Patch each
-     * direction so the asymmetric-key path is exercised both ways -
-     * a regression that only stripped the check from one branch of
-     * SigOidMatchesKeyOid would otherwise be missed. */
-    ExpectIntEQ(rfc9802_assert_oid_patch_breaks_verify(
-        "./certs/xmss/bc_xmss_sha2_10_256_root.der",
-        /* expected XMSS */ 0x22, /* patched to XMSS^MT */ 0x23),
-        TEST_SUCCESS);
-    ExpectIntEQ(rfc9802_assert_oid_patch_breaks_verify(
-        "./certs/xmss/bc_xmssmt_sha2_20_2_256_root.der",
-        /* expected XMSS^MT */ 0x23, /* patched to XMSS */ 0x22),
-        TEST_SUCCESS);
-    return EXPECT_RESULT();
-}
-#endif
-
-/* Exercise a real CA -> leaf certificate chain, not just self-signed.
- * Loads the CA as a trust anchor and verifies the leaf against it. */
-#if defined(WOLFSSL_HAVE_LMS) && !defined(NO_FILESYSTEM) && !defined(NO_CERTS)
-static int rfc9802_lms_chain_verify(void)
-{
-    EXPECT_DECLS;
-    byte*                 caBuf   = NULL;
-    byte*                 leafBuf = NULL;
-    int                   caLen   = 0;
-    int                   leafLen = 0;
-    WOLFSSL_CERT_MANAGER* cm      = NULL;
-
-    ExpectIntEQ(rfc9802_load_file("./certs/lms/bc_lms_chain_ca.der",
-        &caBuf, &caLen), TEST_SUCCESS);
-    ExpectIntEQ(rfc9802_load_file("./certs/lms/bc_lms_chain_leaf.der",
-        &leafBuf, &leafLen), TEST_SUCCESS);
-
-    ExpectNotNull(cm = wolfSSL_CertManagerNew());
-    /* Only the CA is a trust anchor; the leaf is verified against it. */
-    ExpectIntEQ(wolfSSL_CertManagerLoadCABuffer(cm, caBuf, (long)caLen,
-        WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_CertManagerVerifyBuffer(cm, leafBuf, (long)leafLen,
-        WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
-
-    /* Without loading the CA the leaf must NOT verify. */
-    if (cm != NULL) {
-        wolfSSL_CertManagerFree(cm);
-        cm = NULL;
-    }
-    ExpectNotNull(cm = wolfSSL_CertManagerNew());
-    ExpectIntNE(wolfSSL_CertManagerVerifyBuffer(cm, leafBuf, (long)leafLen,
-        WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
-    if (cm != NULL) {
-        wolfSSL_CertManagerFree(cm);
-        cm = NULL;
-    }
-
-    XFREE(leafBuf, NULL, DYNAMIC_TYPE_TMP_BUFFER);
-    XFREE(caBuf, NULL, DYNAMIC_TYPE_TMP_BUFFER);
-    return EXPECT_RESULT();
-}
-#endif
-
-/* Mirror of rfc9802_lms_chain_verify but for an XMSS CA -> leaf pair. */
-#if defined(WOLFSSL_HAVE_XMSS) && !defined(NO_FILESYSTEM) && !defined(NO_CERTS)
-static int rfc9802_xmss_chain_verify(void)
-{
-    EXPECT_DECLS;
-    byte*                 caBuf   = NULL;
-    byte*                 leafBuf = NULL;
-    int                   caLen   = 0;
-    int                   leafLen = 0;
-    WOLFSSL_CERT_MANAGER* cm      = NULL;
-
-    ExpectIntEQ(rfc9802_load_file("./certs/xmss/bc_xmss_chain_ca.der",
-        &caBuf, &caLen), TEST_SUCCESS);
-    ExpectIntEQ(rfc9802_load_file("./certs/xmss/bc_xmss_chain_leaf.der",
-        &leafBuf, &leafLen), TEST_SUCCESS);
-
-    ExpectNotNull(cm = wolfSSL_CertManagerNew());
-    ExpectIntEQ(wolfSSL_CertManagerLoadCABuffer(cm, caBuf, (long)caLen,
-        WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_CertManagerVerifyBuffer(cm, leafBuf, (long)leafLen,
-        WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
-
-    if (cm != NULL) {
-        wolfSSL_CertManagerFree(cm);
-        cm = NULL;
-    }
-    ExpectNotNull(cm = wolfSSL_CertManagerNew());
-    ExpectIntNE(wolfSSL_CertManagerVerifyBuffer(cm, leafBuf, (long)leafLen,
-        WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
-    if (cm != NULL) {
-        wolfSSL_CertManagerFree(cm);
-        cm = NULL;
-    }
-
-    XFREE(leafBuf, NULL, DYNAMIC_TYPE_TMP_BUFFER);
-    XFREE(caBuf, NULL, DYNAMIC_TYPE_TMP_BUFFER);
-    return EXPECT_RESULT();
-}
-#endif
-
-int test_rfc9802_lms_x509_verify(void)
-{
-    EXPECT_DECLS;
-#if defined(WOLFSSL_HAVE_LMS)
-#if !defined(NO_FILESYSTEM) && !defined(NO_CERTS) && \
-    !defined(WOLFSSL_NO_LMS_SHA256_256)
-    /* Mixed single-level LMS and multi-level HSS fixtures. The HSS
-     * public key carries only the top-level LMS/LM-OTS types, so
-     * wc_LmsKey_ImportPubRaw's auto-derive path searches the map
-     * by (levels, lmsType, lmOtsType). The bc_lms_native_bc_root
-     * fixture is generated through Bouncy Castle's stock
-     * JcaContentSignerBuilder("LMS") + JcaX509v3CertificateBuilder
-     * with no overrides; including it here is the cross-impl interop
-     * gate (BC's native LMS X.509 path is RFC 9802-compliant for HSS/
-     * LMS, so wolfSSL must accept it end-to-end).
-     *
-     * All fixtures use the SHA-256/M32 family, so the whole block
-     * is gated on that family being compiled in. Truncated SHA-256/192
-     * or SHAKE-only builds skip this block. */
-    static const char* const lmsFiles[] = {
-        "./certs/lms/bc_lms_sha256_h5_w4_root.der",
-#if !defined(WOLFSSL_LMS_MAX_HEIGHT) || (WOLFSSL_LMS_MAX_HEIGHT >= 10)
-        "./certs/lms/bc_lms_sha256_h10_w8_root.der",
-#endif
-#if !defined(WOLFSSL_LMS_MAX_LEVELS) || (WOLFSSL_LMS_MAX_LEVELS >= 2)
-        "./certs/lms/bc_hss_L2_H5_W8_root.der",
-#endif
-#if !defined(WOLFSSL_LMS_MAX_LEVELS) || (WOLFSSL_LMS_MAX_LEVELS >= 3)
-        "./certs/lms/bc_hss_L3_H5_W4_root.der",
-#endif
-        "./certs/lms/bc_lms_native_bc_root.der",
-    };
-    size_t i;
-    for (i = 0; i < sizeof(lmsFiles) / sizeof(lmsFiles[0]); i++) {
-        ExpectIntEQ(rfc9802_verify_one_cert(lmsFiles[i],
-            HSS_LMSk, CTC_HSS_LMS), TEST_SUCCESS);
-    }
-    ExpectIntEQ(rfc9802_lms_chain_verify(), TEST_SUCCESS);
-#endif /* !NO_FILESYSTEM && !NO_CERTS && !WOLFSSL_NO_LMS_SHA256_256 */
-    /* Pure wolfCrypt-level negative tests don't need filesystem or cert
-     * support, so they run for any LMS-enabled build. */
-    ExpectIntEQ(rfc9802_lms_import_negative(), TEST_SUCCESS);
-#endif
-    return EXPECT_RESULT();
-}
-
-int test_rfc9802_xmss_x509_verify(void)
-{
-    EXPECT_DECLS;
-#if defined(WOLFSSL_HAVE_XMSS)
-#if !defined(NO_FILESYSTEM) && !defined(NO_CERTS)
-    static const char* const xmssFiles[] = {
-        "./certs/xmss/bc_xmss_sha2_10_256_root.der",
-        "./certs/xmss/bc_xmss_sha2_16_256_root.der",
-    };
-    static const char* const xmssmtFiles[] = {
-        "./certs/xmss/bc_xmssmt_sha2_20_2_256_root.der",
-        "./certs/xmss/bc_xmssmt_sha2_20_4_256_root.der",
-        "./certs/xmss/bc_xmssmt_sha2_40_8_256_root.der",
-    };
-    size_t i;
-    for (i = 0; i < sizeof(xmssFiles) / sizeof(xmssFiles[0]); i++) {
-        ExpectIntEQ(rfc9802_verify_one_cert(xmssFiles[i],
-            XMSSk, CTC_XMSS), TEST_SUCCESS);
-    }
-    for (i = 0; i < sizeof(xmssmtFiles) / sizeof(xmssmtFiles[0]); i++) {
-        ExpectIntEQ(rfc9802_verify_one_cert(xmssmtFiles[i],
-            XMSSMTk, CTC_XMSSMT), TEST_SUCCESS);
-    }
-    ExpectIntEQ(rfc9802_xmss_sig_oid_mismatch(), TEST_SUCCESS);
-    ExpectIntEQ(rfc9802_xmss_chain_verify(), TEST_SUCCESS);
-#endif /* !NO_FILESYSTEM && !NO_CERTS */
-    /* Pure wolfCrypt-level negative tests don't need filesystem or cert
-     * support, so they run for any XMSS-enabled build. */
-    ExpectIntEQ(rfc9802_xmss_import_negative(), TEST_SUCCESS);
-#endif
-    return EXPECT_RESULT();
-}
-
 #if defined(WOLFSSL_SNIFFER) && defined(WOLFSSL_SNIFFER_CHAIN_INPUT)
 static int test_sniffer_chain_input_overflow(void)
 {
@@ -39128,47 +33355,6 @@ static int test_DhAgree_rejects_p_minus_1(void)
     return EXPECT_RESULT();
 }
 
-
-/* ML-DSA HashML-DSA verify must reject hashLen > WC_MAX_DIGEST_SIZE */
-static int test_mldsa_verify_hash(void)
-{
-    EXPECT_DECLS;
-#if defined(WOLFSSL_HAVE_MLDSA) && \
-    !defined(WOLFSSL_MLDSA_NO_MAKE_KEY) && \
-    !defined(WOLFSSL_MLDSA_NO_VERIFY)
-    wc_MlDsaKey key;
-    WC_RNG rng;
-    int res = 0;
-    byte sig[4000];
-    byte hash[4096]; /* larger than WC_MAX_DIGEST_SIZE */
-
-    XMEMSET(&key, 0, sizeof(key));
-    XMEMSET(&rng, 0, sizeof(rng));
-    XMEMSET(sig, 0x41, sizeof(sig));
-    XMEMSET(hash, 'A', sizeof(hash));
-
-    ExpectIntEQ(wc_InitRng(&rng), 0);
-    ExpectIntEQ(wc_MlDsaKey_Init(&key, NULL, INVALID_DEVID), 0);
-#ifndef WOLFSSL_NO_ML_DSA_65
-    ExpectIntEQ(wc_MlDsaKey_SetParams(&key, WC_ML_DSA_65), 0);
-#elif !defined(WOLFSSL_NO_ML_DSA_44)
-    ExpectIntEQ(wc_MlDsaKey_SetParams(&key, WC_ML_DSA_44), 0);
-#else
-    ExpectIntEQ(wc_MlDsaKey_SetParams(&key, WC_ML_DSA_87), 0);
-#endif
-    ExpectIntEQ(wc_MlDsaKey_MakeKey(&key, &rng), 0);
-
-    /* hashLen=4096 must be rejected, not overflow the stack */
-    ExpectIntEQ(wc_MlDsaKey_VerifyCtxHash(&key, sig, sizeof(sig), NULL, 0,
-        hash, sizeof(hash), WC_HASH_TYPE_SHA256, &res),
-        WC_NO_ERR_TRACE(BAD_LENGTH_E));
-
-    wc_MlDsaKey_Free(&key);
-    DoExpectIntEQ(wc_FreeRng(&rng), 0);
-#endif
-    return EXPECT_RESULT();
-}
-
 /* Test: Ed448 must reject identity public key (0,1) */
 static int test_ed448_rejects_identity_key(void)
 {
@@ -40047,44 +34233,6 @@ static int test_pkcs7_enveloped_content_size_overflow(void)
     return EXPECT_RESULT();
 }
 
-/* Dilithium verify_ctx_msg must reject absurdly large msgLen */
-static int test_dilithium_hash(void)
-{
-    EXPECT_DECLS;
-#if defined(WOLFSSL_HAVE_MLDSA) && \
-    !defined(WOLFSSL_MLDSA_NO_MAKE_KEY) && \
-    !defined(WOLFSSL_MLDSA_NO_VERIFY)
-    wc_MlDsaKey key;
-    WC_RNG rng;
-    int res = 0;
-    byte sig[4000];
-    byte msg[64];
-
-    XMEMSET(&key, 0, sizeof(key));
-    XMEMSET(&rng, 0, sizeof(rng));
-    XMEMSET(sig, 0, sizeof(sig));
-    XMEMSET(msg, 'A', sizeof(msg));
-
-    ExpectIntEQ(wc_InitRng(&rng), 0);
-    ExpectIntEQ(wc_MlDsaKey_Init(&key, NULL, INVALID_DEVID), 0);
-#ifndef WOLFSSL_NO_ML_DSA_65
-    ExpectIntEQ(wc_MlDsaKey_SetParams(&key, WC_ML_DSA_65), 0);
-#elif !defined(WOLFSSL_NO_ML_DSA_44)
-    ExpectIntEQ(wc_MlDsaKey_SetParams(&key, WC_ML_DSA_44), 0);
-#else
-    ExpectIntEQ(wc_MlDsaKey_SetParams(&key, WC_ML_DSA_87), 0);
-#endif
-    ExpectIntEQ(wc_MlDsaKey_MakeKey(&key, &rng), 0);
-
-    ExpectIntEQ(wc_MlDsaKey_VerifyCtx(&key, sig, sizeof(sig), NULL, 0,
-        msg, 0xFFFFFFC0, &res), WC_NO_ERR_TRACE(BAD_FUNC_ARG));
-
-    wc_MlDsaKey_Free(&key);
-    DoExpectIntEQ(wc_FreeRng(&rng), 0);
-#endif
-    return EXPECT_RESULT();
-}
-
 /* PKCS7 CBC must validate all padding bytes */
 static int test_pkcs7_padding(void)
 {
@@ -40283,13 +34431,8 @@ TEST_CASE testCases[] = {
     /* ASN */
     TEST_ASN_DECLS,
 
-    /* LMS */
-    TEST_DECL_GROUP("lms", test_wc_LmsKey_sign_verify),
-    TEST_DECL_GROUP("lms", test_wc_LmsKey_reload_cache),
-
-    /* RFC 9802 (HSS/LMS and XMSS/XMSS^MT in X.509) */
-    TEST_DECL_GROUP("lms", test_rfc9802_lms_x509_verify),
-    TEST_DECL_GROUP("xmss", test_rfc9802_xmss_x509_verify),
+    /* LMS, and RFC 9802 (HSS/LMS and XMSS/XMSS^MT in X.509) */
+    TEST_LMS_XMSS_DECLS,
 
     /* PEM and DER APIs. */
     TEST_DECL(test_wc_PemToDer),
@@ -40323,6 +34466,7 @@ TEST_CASE testCases[] = {
     TEST_DECL(test_PathLenSelfIssuedAllowed),
     TEST_DECL(test_PathLenNoKeyUsage),
     TEST_DECL(test_NameConstraints_OtherName),
+    TEST_DECL(test_NameConstraints_DnsUriWildcard),
     TEST_DECL(test_ParseSerial0FixtureMatrix),
     TEST_DECL(test_MakeCertWithCaFalse),
 #ifdef WOLFSSL_CERT_SIGN_CB
@@ -40454,6 +34598,8 @@ TEST_CASE testCases[] = {
     TEST_DECL(test_wolfSSL_X509_CRL),
 #ifndef NO_BIO
     TEST_DECL(test_wolfSSL_X509_print),
+    TEST_DECL(test_wolfSSL_X509_print_basic_constraints),
+    TEST_DECL(test_wolfSSL_X509_print_ext_key_usage),
     TEST_DECL(test_wolfSSL_X509_CRL_print),
 #endif
 
@@ -40531,10 +34677,6 @@ TEST_CASE testCases[] = {
     TEST_DECL(test_wolfSSL_sigalg_info),
     /* Can't memory test as tcp_connect aborts. */
     TEST_DECL(test_wolfSSL_d2i_SSL_SESSION_bounds_check),
-    TEST_DECL(test_wolfSSL_SESSION),
-    TEST_DECL(test_wolfSSL_SESSION_expire_downgrade),
-    TEST_DECL(test_wolfSSL_CTX_sess_set_remove_cb),
-    TEST_DECL(test_wolfSSL_ticket_keys),
     TEST_DECL(test_wolfSSL_sk_GENERAL_NAME),
     TEST_DECL(test_wolfSSL_GENERAL_NAME_print),
     TEST_DECL(test_wolfSSL_sk_DIST_POINT),
@@ -40572,7 +34714,6 @@ TEST_CASE testCases[] = {
 
     TEST_DECL(test_wolfSSL_CRYPTO_memcmp),
     TEST_DECL(test_wolfSSL_CRYPTO_get_ex_new_index),
-    TEST_DECL(test_wolfSSL_SESSION_get_ex_new_index),
     TEST_DECL(test_CRYPTO_set_dynlock_xxx),
     TEST_DECL(test_CRYPTO_THREADID_xxx),
     TEST_DECL(test_ENGINE_cleanup),
@@ -40685,20 +34826,8 @@ TEST_CASE testCases[] = {
 
 #ifdef HAVE_IO_TESTS_DEPENDENCIES
     TEST_DECL(test_wolfSSL_get_finished),
-    /* Uses Assert in handshake callback. */
-    TEST_DECL(test_wolfSSL_CTX_add_session),
-    /* Large number of memory allocations. */
-    TEST_DECL(test_wolfSSL_CTX_add_session_ext_tls13),
-    /* Large number of memory allocations. */
-    TEST_DECL(test_wolfSSL_CTX_add_session_ext_dtls13),
-    /* Large number of memory allocations. */
-    TEST_DECL(test_wolfSSL_CTX_add_session_ext_tls12),
-    /* Large number of memory allocations. */
-    TEST_DECL(test_wolfSSL_CTX_add_session_ext_dtls12),
-    /* Large number of memory allocations. */
-    TEST_DECL(test_wolfSSL_CTX_add_session_ext_tls11),
-    /* Large number of memory allocations. */
-    TEST_DECL(test_wolfSSL_CTX_add_session_ext_dtls1),
+    TEST_DECL(test_wolfSSL_get_finished_overrun),
+    TEST_DECL(test_wolfSSL_get_peer_finished_overrun),
 #endif
     TEST_DECL(test_SSL_CIPHER_get_xxx),
     TEST_DECL(test_wolfSSL_ERR_strings),
@@ -40751,12 +34880,8 @@ TEST_CASE testCases[] = {
     TEST_DECL(test_SetTmpEC_DHE_Sz),
     TEST_DECL(test_wolfSSL_CTX_get0_privatekey),
 #ifdef WOLFSSL_DTLS
-    TEST_DECL(test_wolfSSL_DtlsUpdateWindow),
-    TEST_DECL(test_wolfSSL_DTLS_fragment_buckets),
 #endif
-    TEST_DECL(test_wolfSSL_dtls_set_mtu),
     /* Uses Assert in handshake callback. */
-    TEST_DECL(test_wolfSSL_dtls_plaintext),
 #if !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER) && \
     defined(HAVE_IO_TESTS_DEPENDENCIES)
     TEST_DECL(test_wolfSSL_read_write),
@@ -40769,12 +34894,9 @@ TEST_CASE testCases[] = {
     TEST_DECL(test_wolfSSL_CTX_verifyDepth_ServerClient_3),
     TEST_DECL(test_wolfSSL_CTX_set_cipher_list),
     /* Can't memory test as server hangs. */
-    TEST_DECL(test_wolfSSL_dtls_export),
     /* Uses Assert in handshake callback. */
     TEST_DECL(test_wolfSSL_tls_export),
 #endif
-    TEST_DECL(test_wolfSSL_dtls_export_peers),
-    TEST_DECL(test_wolfSSL_dtls_import_state_extra_window_words),
     TEST_DECL(test_wolfSSL_SetMinVersion),
     TEST_DECL(test_wolfSSL_CTX_SetMinVersion),
 
@@ -40896,19 +35018,11 @@ TEST_CASE testCases[] = {
 
     TEST_DECL(test_wolfSSL_msgCb),
     TEST_DECL(test_wolfSSL_either_side),
-    TEST_DECL(test_wolfSSL_DTLS_either_side),
     /* Uses Assert in handshake callback. */
-    TEST_DECL(test_wolfSSL_dtls_fragments),
     /* Uses Assert in handshake callback. */
-    TEST_DECL(test_wolfSSL_dtls_AEAD_limit),
     /* Uses Assert in handshake callback. */
-    TEST_DECL(test_wolfSSL_ignore_alert_before_cookie),
     /* Uses Assert in handshake callback. */
-    TEST_DECL(test_wolfSSL_dtls_bad_record),
     /* Uses Assert in handshake callback. */
-    TEST_DECL(test_wolfSSL_dtls_stateless),
-    TEST_DECL(test_wolfSSL_dtls_stateless_hrr_group),
-    TEST_DECL(test_generate_cookie),
 
 #ifndef NO_BIO
     TEST_OSSL_BIO_TLS_DECLS,
@@ -40935,15 +35049,11 @@ TEST_CASE testCases[] = {
      defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) &&                    \
     !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER)
 #ifdef WOLFSSL_DTLS_NO_HVR_ON_RESUME
-    TEST_DECL(test_wolfSSL_dtls_stateless_resume),
 #endif /* WOLFSSL_DTLS_NO_HVR_ON_RESUME */
 #ifdef HAVE_MAX_FRAGMENT
-    TEST_DECL(test_wolfSSL_dtls_stateless_maxfrag),
 #endif /* HAVE_MAX_FRAGMENT */
 #ifndef NO_RSA
-    TEST_DECL(test_wolfSSL_dtls_stateless2),
 #if !defined(NO_OLD_TLS)
-    TEST_DECL(test_wolfSSL_dtls_stateless_downgrade),
 #endif /* !defined(NO_OLD_TLS) */
 #endif /* ! NO_RSA */
 #endif /* defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) &&     \
@@ -40951,7 +35061,6 @@ TEST_CASE testCases[] = {
     TEST_DECL(test_wolfSSL_CTX_set_ciphersuites),
     TEST_DECL(test_wolfSSL_CRL_CERT_REVOKED_alert),
     TEST_DECL(test_TLS_13_ticket_different_ciphers),
-    TEST_DECL(test_WOLFSSL_dtls_version_alert),
 
 #if defined(WOLFSSL_TICKET_NONCE_MALLOC) && defined(HAVE_SESSION_TICKET)       \
     && defined(WOLFSSL_TLS13) &&                                               \
@@ -40970,36 +35079,17 @@ TEST_CASE testCases[] = {
     TEST_DECL(test_harden_no_secure_renegotiation),
     TEST_DECL(test_override_alt_cert_chain),
     TEST_DECL(test_rpk_set_xxx_cert_type),
-    TEST_DECL(test_dtls13_bad_epoch_ch),
     TEST_DECL(test_short_session_id),
-    TEST_DECL(test_wolfSSL_dtls13_null_cipher),
     /* Can't memory test as client/server hangs. */
-    TEST_DECL(test_dtls_msg_from_other_peer),
-    TEST_DECL(test_dtls_ipv6_check),
     TEST_DECL(test_wolfSSL_SCR_after_resumption),
-    TEST_DECL(test_dtls_no_extensions),
     TEST_DECL(test_tls_alert_no_server_hello),
     TEST_DECL(test_TLSX_CA_NAMES_bad_extension),
-    TEST_DECL(test_dtls_1_0_hvr_downgrade),
     TEST_DECL(test_session_ticket_no_id),
     TEST_DECL(test_session_ticket_hs_update),
-    TEST_DECL(test_dtls_downgrade_scr_server),
-    TEST_DECL(test_dtls_downgrade_scr),
-    TEST_DECL(test_dtls_client_hello_timeout_downgrade),
-    TEST_DECL(test_dtls_client_hello_timeout),
-    TEST_DECL(test_dtls_dropped_ccs),
-    TEST_DECL(test_dtls_seq_num_downgrade),
     TEST_DECL(test_certreq_sighash_algos),
     TEST_DECL(test_revoked_loaded_int_cert),
-    TEST_DECL(test_dtls_frag_ch),
-    TEST_DECL(test_dtls13_frag_ch_pq),
-    TEST_DECL(test_dtls_empty_keyshare_with_cookie),
-    TEST_DECL(test_dtls_old_seq_number),
-    TEST_DECL(test_dtls12_missing_finished),
-    TEST_DECL(test_dtls13_missing_finished_client),
-    TEST_DECL(test_dtls13_missing_finished_server),
-    TEST_DECL(test_dtls13_finished_send_error_propagation),
     TEST_DTLS_DECLS,
+    TEST_DTLS13_DECLS,
     TEST_DECL(test_tls_multi_handshakes_one_record),
     TEST_DECL(test_write_dup),
     TEST_DECL(test_write_dup_want_write),
@@ -41019,12 +35109,14 @@ TEST_CASE testCases[] = {
     TEST_DECL(test_ocsp_response_parsing),
     TEST_DECL(test_ocsp_certid_enc_dec),
     TEST_DECL(test_ocsp_certid_dup),
+    TEST_DECL(test_ocsp_resp_find_status_serial_prefix),
     TEST_DECL(test_ocsp_tls_cert_cb),
     TEST_DECL(test_ocsp_cert_unknown_crl_fallback),
     TEST_DECL(test_ocsp_cert_unknown_crl_fallback_nonleaf),
     TEST_DECL(test_tls13_nonblock_ocsp_low_mfl),
     TEST_DECL(test_ocsp_responder),
     TEST_TLS_DECLS,
+    TEST_SESSION_DECLS,
     TEST_DECL(test_wc_DhSetNamedKey),
     TEST_DECL(test_DhAgree_rejects_p_minus_1),
     TEST_DECL(test_ed448_rejects_identity_key),
@@ -41043,8 +35135,6 @@ TEST_CASE testCases[] = {
     TEST_DECL(test_sniffer_chain_input_overflow),
 #endif
 
-    TEST_DECL(test_mldsa_verify_hash),
-    TEST_DECL(test_dilithium_hash),
     /* This test needs to stay at the end to clean up any caches allocated. */
     TEST_DECL(test_wolfSSL_Cleanup)
 };
diff --git a/tests/api/include.am b/tests/api/include.am
index f9d4dfe3f1..7ad64d3e63 100644
--- a/tests/api/include.am
+++ b/tests/api/include.am
@@ -51,13 +51,16 @@ tests_unit_test_SOURCES += tests/api/test_mldsa.c
 tests_unit_test_SOURCES += tests/api/test_mldsa_legacy.c
 tests_unit_test_SOURCES += tests/api/test_slhdsa.c
 tests_unit_test_SOURCES += tests/api/test_signature.c
+tests_unit_test_SOURCES += tests/api/test_lms_xmss.c
 # TLS Protocol
 tests_unit_test_SOURCES += tests/api/test_dtls.c
+tests_unit_test_SOURCES += tests/api/test_dtls13.c
 # TLS Feature
 tests_unit_test_SOURCES += tests/api/test_ocsp.c
 tests_unit_test_SOURCES += tests/api/test_evp.c
 tests_unit_test_SOURCES += tests/api/test_tls_ext.c
 tests_unit_test_SOURCES += tests/api/test_tls.c
+tests_unit_test_SOURCES += tests/api/test_session.c
 # Certs
 tests_unit_test_SOURCES += tests/api/test_x509.c
 # ASN
@@ -157,13 +160,16 @@ EXTRA_DIST += tests/api/test_mlkem.h
 EXTRA_DIST += tests/api/test_mldsa.h
 EXTRA_DIST += tests/api/test_slhdsa.h
 EXTRA_DIST += tests/api/test_signature.h
+EXTRA_DIST += tests/api/test_lms_xmss.h
 EXTRA_DIST += tests/api/test_dtls.h
+EXTRA_DIST += tests/api/test_dtls13.h
 EXTRA_DIST += tests/api/test_ocsp.h
 EXTRA_DIST += tests/api/test_ocsp_test_blobs.h
 EXTRA_DIST += tests/api/create_ocsp_test_blobs.py
 EXTRA_DIST += tests/api/test_evp.h
 EXTRA_DIST += tests/api/test_tls_ext.h
 EXTRA_DIST += tests/api/test_tls.h
+EXTRA_DIST += tests/api/test_session.h
 EXTRA_DIST += tests/api/test_x509.h
 EXTRA_DIST += tests/api/test_asn.h
 EXTRA_DIST += tests/api/test_pkcs7.h
diff --git a/tests/api/test_asn.c b/tests/api/test_asn.c
index 23f391c938..ce4eea1487 100644
--- a/tests/api/test_asn.c
+++ b/tests/api/test_asn.c
@@ -31,6 +31,12 @@
 #ifdef HAVE_ED25519
     #include <wolfssl/wolfcrypt/ed25519.h>
 #endif
+#ifdef HAVE_ED448
+    #include <wolfssl/wolfcrypt/ed448.h>
+#endif
+#ifdef HAVE_DILITHIUM
+    #include <wolfssl/wolfcrypt/dilithium.h>
+#endif
 
 #if defined(WC_ENABLE_ASYM_KEY_EXPORT) && defined(HAVE_ED25519)
 static int test_SetAsymKeyDer_once(byte* privKey, word32 privKeySz, byte* pubKey,
@@ -843,6 +849,19 @@ int test_wolfssl_local_MatchBaseName(void)
                 "sub.domain.com", 14, ".domain.com", 11), 1);
     ExpectIntEQ(wolfssl_local_MatchBaseName(ASN_DNS_TYPE,
                 "a.b.domain.com", 14, ".domain.com", 11), 1);
+    /* Trailing-dot normalization: absolute DNS form is equivalent. */
+    ExpectIntEQ(wolfssl_local_MatchBaseName(ASN_DNS_TYPE,
+                "domain.com.", (int)XSTRLEN("domain.com."),
+                "domain.com", (int)XSTRLEN("domain.com")), 1);
+    ExpectIntEQ(wolfssl_local_MatchBaseName(ASN_DNS_TYPE,
+                "domain.com", (int)XSTRLEN("domain.com"),
+                "domain.com.", (int)XSTRLEN("domain.com.")), 1);
+    ExpectIntEQ(wolfssl_local_MatchBaseName(ASN_DNS_TYPE,
+                "domain.com.", (int)XSTRLEN("domain.com."),
+                "domain.com.", (int)XSTRLEN("domain.com.")), 1);
+    ExpectIntEQ(wolfssl_local_MatchBaseName(ASN_DNS_TYPE,
+                "sub.domain.com.", (int)XSTRLEN("sub.domain.com."),
+                ".domain.com.", (int)XSTRLEN(".domain.com.")), 1);
 
     /* Negative tests - should NOT match */
     /* Bug #3: fakedomain.com should NOT match domain.com (no dot boundary) */
@@ -864,6 +883,10 @@ int test_wolfssl_local_MatchBaseName(void)
     /* Name starting with dot */
     ExpectIntEQ(wolfssl_local_MatchBaseName(ASN_DNS_TYPE,
                 ".domain.com", 11, "domain.com", 10), 0);
+    /* More than one trailing dot leaves an empty label after normalization. */
+    ExpectIntEQ(wolfssl_local_MatchBaseName(ASN_DNS_TYPE,
+                "domain.com..", (int)XSTRLEN("domain.com.."),
+                "domain.com", (int)XSTRLEN("domain.com")), 0);
 
     /*
      * Tests for email type (ASN_RFC822_TYPE = 0x01)
@@ -963,6 +986,210 @@ int test_wolfssl_local_MatchBaseName(void)
     return EXPECT_RESULT();
 }
 
+#if !defined(NO_CERTS) && !defined(NO_ASN) && !defined(IGNORE_NAME_CONSTRAINTS)
+/* Convenience wrappers so the cases below read as (name, base) pairs and the
+ * string lengths can't drift out of sync with the literals. */
+static int dnsWildPermitted(const char* name, const char* base)
+{
+    return wolfssl_local_MatchDnsConstraintWildcard(name, (int)XSTRLEN(name),
+        base, (int)XSTRLEN(base), 1);
+}
+static int dnsWildExcluded(const char* name, const char* base)
+{
+    return wolfssl_local_MatchDnsConstraintWildcard(name, (int)XSTRLEN(name),
+        base, (int)XSTRLEN(base), 0);
+}
+static int uriNC(const char* uri, const char* base)
+{
+    return wolfssl_local_MatchUriNameConstraint(uri, (int)XSTRLEN(uri), base,
+        (int)XSTRLEN(base));
+}
+#endif
+
+/*
+ * Tests label-aware matching of a wildcard DNS SAN against a name-constraint
+ * subtree. The permitted variant must prove containment (every expansion of
+ * the wildcard stays inside the subtree); the excluded variant must detect
+ * intersection (some expansion falls inside the subtree). A '*' never crosses
+ * a label boundary, so the comparison is by label from the right.
+ */
+int test_wolfssl_local_MatchDnsConstraintWildcard(void)
+{
+    EXPECT_DECLS;
+
+#if !defined(NO_CERTS) && !defined(NO_ASN) && !defined(IGNORE_NAME_CONSTRAINTS)
+    /*
+     * PERMITTED subtree -- containment. Accept only when EVERY expansion of
+     * the wildcard is inside the base subtree.
+     */
+
+    /* Wildcard is an extra label to the left of the base: always contained. */
+    ExpectIntEQ(dnsWildPermitted("*.example.com",     "example.com"),  1);
+    ExpectIntEQ(dnsWildPermitted("*.sub.example.com", "example.com"),  1);
+    ExpectIntEQ(dnsWildPermitted("foo*.example.com",  "example.com"),  1);
+    ExpectIntEQ(dnsWildPermitted("a*b.example.com",   "example.com"),  1);
+    /* Case-insensitive on the literal tail labels. */
+    ExpectIntEQ(dnsWildPermitted("*.EXAMPLE.CoM",      "example.com"),  1);
+    /* Single-label base; the matched tail "com" is literal. */
+    ExpectIntEQ(dnsWildPermitted("*.example.com",     "com"),          1);
+    /* Leading-dot base requires at least one label before it -- the wildcard
+     * label satisfies that. */
+    ExpectIntEQ(dnsWildPermitted("*.example.com",     ".example.com"), 1);
+    ExpectIntEQ(dnsWildPermitted("*.sub.example.com", ".example.com"), 1);
+    /* Trailing-dot normalization: absolute DNS form is equivalent. */
+    ExpectIntEQ(dnsWildPermitted("*.example.com.",    "example.com"),  1);
+    ExpectIntEQ(dnsWildPermitted("*.example.com",     "example.com."), 1);
+    ExpectIntEQ(dnsWildPermitted("*.example.com.",    "example.com."), 1);
+    ExpectIntEQ(dnsWildPermitted("*.example.com.",    ".example.com."), 1);
+
+    /* Wildcard lands on a label that must equal the base: NOT provably
+     * contained, because the label can expand to something else. */
+    ExpectIntEQ(dnsWildPermitted("*.example.com",     "foo.example.com"), 0);
+    ExpectIntEQ(dnsWildPermitted("*.example.com.",    "foo.example.com"), 0);
+    ExpectIntEQ(dnsWildPermitted("*.example.com",     "foo.example.com."), 0);
+    ExpectIntEQ(dnsWildPermitted("ex*.com",           "example.com"),     0);
+    ExpectIntEQ(dnsWildPermitted("foo.exa*ple.com",   "example.com"),     0);
+    /* Tail labels do not match the base at all. */
+    ExpectIntEQ(dnsWildPermitted("*.example.com",     "example.org"),     0);
+    ExpectIntEQ(dnsWildPermitted("*.evil.com",        "example.com"),     0);
+    /* Leading-dot base, but wildcard would have to equal an interior base
+     * label. */
+    ExpectIntEQ(dnsWildPermitted("*.example.com",     ".sub.example.com"), 0);
+    /* A bare '*' cannot be proven inside any multi-label-or-single subtree. */
+    ExpectIntEQ(dnsWildPermitted("*",                 "com"),             0);
+
+    /*
+     * EXCLUDED subtree -- intersection. Reject when SOME expansion of the
+     * wildcard falls inside the base subtree. A wildcard label is
+     * conservatively treated as able to match any single base label.
+     */
+
+    ExpectIntEQ(dnsWildExcluded("*.example.com",      "foo.example.com"), 1);
+    ExpectIntEQ(dnsWildExcluded("*.example.com.",     "foo.example.com"), 1);
+    ExpectIntEQ(dnsWildExcluded("*.example.com",      "foo.example.com."), 1);
+    ExpectIntEQ(dnsWildExcluded("*.example.com.",     "foo.example.com."), 1);
+    /* Wildcard adds a label on top of the excluded subtree. */
+    ExpectIntEQ(dnsWildExcluded("*.example.com",      "example.com"),     1);
+    ExpectIntEQ(dnsWildExcluded("*.example.com",      "com"),             1);
+    ExpectIntEQ(dnsWildExcluded("*.example.com",      ".example.com"),    1);
+    /* Wildcard in a non-left label still intersects. */
+    ExpectIntEQ(dnsWildExcluded("foo.*.example.com",  "bar.example.com"), 1);
+    /* Partial-label wildcard: conservatively excluded even though "ex*"
+     * cannot actually expand to "foo" (over-rejection, safe). */
+    ExpectIntEQ(dnsWildExcluded("ex*.example.com",    "foo.example.com"), 1);
+    /* A bare '*' can expand to the apex label of a single-label subtree. */
+    ExpectIntEQ(dnsWildExcluded("*",                  "com"),             1);
+
+    /* No intersection: literal tail labels differ from the base. */
+    ExpectIntEQ(dnsWildExcluded("*.example.com",      "foo.other.com"),   0);
+    ExpectIntEQ(dnsWildExcluded("*.other.com",        "example.com"),     0);
+    ExpectIntEQ(dnsWildExcluded("*.example.com",      "example.org"),     0);
+    /* Leading-dot excluded base needs a label before it; the wildcard SAN has
+     * no room for one, so no expansion reaches the proper subtree. */
+    ExpectIntEQ(dnsWildExcluded("*.example.com",      ".foo.example.com"), 0);
+    /* Same arity: '*' can expand to the apex label of the base, so the
+     * wildcard intersects (*.com can be example.com, which is excluded). */
+    ExpectIntEQ(dnsWildExcluded("*.com",              "example.com"),     1);
+    /* But a base with MORE labels than the name cannot be reached. */
+    ExpectIntEQ(dnsWildExcluded("*.com",              "a.example.com"),   0);
+
+    /*
+     * Error / degenerate inputs (both flags reject).
+     */
+    ExpectIntEQ(wolfssl_local_MatchDnsConstraintWildcard(NULL, 5,
+                "com", 3, 1), 0);
+    ExpectIntEQ(wolfssl_local_MatchDnsConstraintWildcard("*.com", 5,
+                NULL, 3, 1), 0);
+    ExpectIntEQ(wolfssl_local_MatchDnsConstraintWildcard("*.com", 0,
+                "com", 3, 1), 0);
+    ExpectIntEQ(wolfssl_local_MatchDnsConstraintWildcard("*.com", 5,
+                "com", 0, 1), 0);
+    /* Name beginning with a dot is invalid. */
+    ExpectIntEQ(dnsWildPermitted(".x.com",            "com"),             0);
+    ExpectIntEQ(dnsWildExcluded(".x.com",             "com"),             0);
+    /* Base that is only dots collapses to nothing. */
+    ExpectIntEQ(dnsWildExcluded("*.example.com",      "."),               0);
+    ExpectIntEQ(dnsWildExcluded("*.example.com",      ".."),              0);
+    /* SAN has an empty interior label ("*..com"), but only the right-most
+     * "com" label overlaps the base "com" -- the empty label sits outside the
+     * compared suffix, and '*' can expand to any label, so the matcher
+     * conservatively reports intersection. */
+    ExpectIntEQ(dnsWildExcluded("*..com",             "com"),             1);
+
+#endif /* !NO_CERTS && !NO_ASN && !IGNORE_NAME_CONSTRAINTS */
+
+    return EXPECT_RESULT();
+}
+
+/*
+ * Tests URI name-constraint matching (RFC 5280 4.2.1.10): the constraint
+ * applies to the host portion of the URI. A constraint that does NOT begin
+ * with a dot is an exact host match; one that begins with a dot matches any
+ * host with one or more additional leading labels (the bare host is excluded).
+ */
+int test_wolfssl_local_MatchUriNameConstraint(void)
+{
+    EXPECT_DECLS;
+
+#if !defined(NO_CERTS) && !defined(NO_ASN) && !defined(IGNORE_NAME_CONSTRAINTS)
+    /*
+     * Exact host match (no leading dot in the constraint).
+     */
+    ExpectIntEQ(uriNC("https://host.com/path",        "host.com"), 1);
+    ExpectIntEQ(uriNC("https://host.com",             "host.com"), 1);
+    ExpectIntEQ(uriNC("https://host.com:8443/x",      "host.com"), 1);
+    ExpectIntEQ(uriNC("ftp://user@host.com/x",        "host.com"), 1);
+    ExpectIntEQ(uriNC("https://HOST.COM",             "host.com"), 1);
+    ExpectIntEQ(uriNC("https://host.com?q=1",         "host.com"), 1);
+    ExpectIntEQ(uriNC("https://host.com#frag",        "host.com"), 1);
+
+    /* The bug this fix closes: an exact-host constraint must NOT subtree-match
+     * a sub-host. */
+    ExpectIntEQ(uriNC("https://www.host.com/",        "host.com"), 0);
+    ExpectIntEQ(uriNC("https://a.b.host.com",         "host.com"), 0);
+    /* Suffix that does not respect a label boundary. */
+    ExpectIntEQ(uriNC("https://xhost.com",            "host.com"), 0);
+    /* host.com is a prefix of the URI host but not the whole host. */
+    ExpectIntEQ(uriNC("https://host.com.evil.com",    "host.com"), 0);
+    ExpectIntEQ(uriNC("https://other.com",            "host.com"), 0);
+
+    /*
+     * Leading-dot constraint: proper subtree of hosts (apex excluded).
+     */
+    ExpectIntEQ(uriNC("https://www.host.com/",        ".host.com"), 1);
+    ExpectIntEQ(uriNC("https://a.b.host.com",         ".host.com"), 1);
+    ExpectIntEQ(uriNC("https://www.host.com:443",     ".host.com"), 1);
+    /* The bare host is NOT in the leading-dot subtree. */
+    ExpectIntEQ(uriNC("https://host.com",             ".host.com"), 0);
+    ExpectIntEQ(uriNC("https://evilhost.com",         ".host.com"), 0);
+
+    /*
+     * IPv6 literal host extraction ([..]) then exact match.
+     */
+    ExpectIntEQ(uriNC("https://[2001:db8::1]:443/x",  "2001:db8::1"), 1);
+    ExpectIntEQ(uriNC("https://[2001:db8::1]",        "2001:db8::2"), 0);
+
+    /*
+     * Malformed / degenerate URIs and inputs (reject).
+     */
+    ExpectIntEQ(uriNC("no-scheme-host.com",           "host.com"), 0);
+    ExpectIntEQ(uriNC("https://",                     "host.com"), 0);
+    /* double literal to abide source-check thinking it's a c++ comment */
+    ExpectIntEQ(uriNC("https://" "/path",             "host.com"), 0);
+    ExpectIntEQ(wolfssl_local_MatchUriNameConstraint(NULL, 10,
+                "host.com", 8), 0);
+    ExpectIntEQ(wolfssl_local_MatchUriNameConstraint("https://host.com", 16,
+                NULL, 8), 0);
+    ExpectIntEQ(wolfssl_local_MatchUriNameConstraint("https://host.com", 0,
+                "host.com", 8), 0);
+    ExpectIntEQ(wolfssl_local_MatchUriNameConstraint("https://host.com", 16,
+                "host.com", 0), 0);
+
+#endif /* !NO_CERTS && !NO_ASN && !IGNORE_NAME_CONSTRAINTS */
+
+    return EXPECT_RESULT();
+}
+
 /*
  * Testing wc_DecodeRsaPssParams with known DER byte arrays.
  * Exercises both WOLFSSL_ASN_TEMPLATE and non-template paths.
@@ -1363,3 +1590,440 @@ int test_wc_DecodeObjectId(void)
 
     return EXPECT_RESULT();
 }
+
+#if defined(HAVE_PKCS8) && !defined(NO_ASN) && \
+    (defined(WOLFSSL_TEST_CERT) || defined(OPENSSL_EXTRA) || \
+     defined(OPENSSL_EXTRA_X509_SMALL) || defined(WOLFSSL_PUBLIC_ASN)) && \
+    (defined(HAVE_ED25519) || \
+     (defined(HAVE_ED448) && defined(HAVE_ED448_KEY_EXPORT) && \
+      defined(WOLFSSL_KEY_GEN)) || \
+     (defined(HAVE_DILITHIUM) && \
+      !defined(WOLFSSL_DILITHIUM_NO_MAKE_KEY) && \
+      !defined(WOLFSSL_DILITHIUM_NO_ASN1)))
+/* Run ToTraditional_ex() on a copy of der and assert the algId, returned
+ * length, and the inner OCTET STRING tag/length at the start of the
+ * (in-place rewritten) buffer. */
+static int test_ToTraditional_ex_once(const byte* der, word32 derSz,
+    word32 expectAlgId, word32 expectPrivKeySz)
+{
+    EXPECT_DECLS;
+    byte* copy = NULL;
+    word32 algId = 0;
+    int    ret;
+
+    copy = (byte*)XMALLOC(derSz, NULL, DYNAMIC_TYPE_TMP_BUFFER);
+    ExpectNotNull(copy);
+    if (copy != NULL) {
+        XMEMCPY(copy, der, derSz);
+        ret = ToTraditional_ex(copy, derSz, &algId);
+        ExpectIntGT(ret, 0);
+        ExpectIntEQ(algId, expectAlgId);
+        if (ret > 0) {
+            /* wolfSSL writes nested OCTET STRING, but accept raw bytes
+             * too per RFC 5958. */
+            if (copy[0] == ASN_OCTET_STRING) {
+                if (expectPrivKeySz < 0x80) {
+                    ExpectIntEQ(copy[1], (byte)expectPrivKeySz);
+                }
+                else if (expectPrivKeySz < 0x100) {
+                    ExpectIntEQ(copy[1], 0x81);
+                    ExpectIntEQ(copy[2], (byte)expectPrivKeySz);
+                }
+                else {
+                    ExpectIntEQ(copy[1], 0x82);
+                    ExpectIntEQ(((word32)copy[2] << 8) | copy[3],
+                        expectPrivKeySz);
+                }
+            }
+            else {
+                ExpectIntEQ(ret, (int)expectPrivKeySz);
+            }
+        }
+    }
+    XFREE(copy, NULL, DYNAMIC_TYPE_TMP_BUFFER);
+
+    return EXPECT_RESULT();
+}
+#endif
+
+/* Hand crafted PKCS#8 v0 and v1 Ed25519 buffers to test parser directly. */
+int test_ToTraditional_ex_handcrafted(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_PKCS8) && defined(HAVE_ED25519) && \
+    (defined(WOLFSSL_TEST_CERT) || defined(OPENSSL_EXTRA) || \
+     defined(OPENSSL_EXTRA_X509_SMALL) || defined(WOLFSSL_PUBLIC_ASN))
+    /* Ed25519 algorithm OID body (1.3.101.112). */
+    static const byte algId[] = { 43, 101, 112 };
+    const word32 privKeySz = ED25519_KEY_SIZE;
+    const word32 pubKeySz  = ED25519_PUB_KEY_SIZE;
+    byte der[128];
+    word32 sz;
+    word32 outerLenIdx;
+    /* Filler bytes for the dummy private/public key bodies */
+    const byte keyPat = 0xCC;
+    const byte pubPat = 0xDD;
+
+    /* v0: SEQ { INTEGER 0, SEQ { OID }, OCTET STRING { OCTET STRING priv } } */
+    sz = 0;
+    der[sz++] = ASN_SEQUENCE | ASN_CONSTRUCTED;
+    outerLenIdx = sz;
+    der[sz++] = 0;  /* outer length, filled in below */
+    der[sz++] = ASN_INTEGER;
+    der[sz++] = 1;
+    der[sz++] = 0x00;
+    der[sz++] = ASN_SEQUENCE | ASN_CONSTRUCTED;
+    der[sz++] = (byte)(sizeof(algId) + 2);
+    der[sz++] = ASN_OBJECT_ID;
+    der[sz++] = (byte)sizeof(algId);
+    XMEMCPY(der + sz, algId, sizeof(algId)); sz += sizeof(algId);
+    der[sz++] = ASN_OCTET_STRING;
+    der[sz++] = (byte)(privKeySz + 2);
+    der[sz++] = ASN_OCTET_STRING;
+    der[sz++] = (byte)privKeySz;
+    XMEMSET(der + sz, keyPat, privKeySz); sz += privKeySz;
+    der[outerLenIdx] = (byte)(sz - outerLenIdx - 1);
+
+    EXPECT_TEST(test_ToTraditional_ex_once(der, sz, ED25519k, privKeySz));
+
+    /* v1: same plus [1] publicKey trailer. */
+    sz = 0;
+    der[sz++] = ASN_SEQUENCE | ASN_CONSTRUCTED;
+    outerLenIdx = sz;
+    der[sz++] = 0;
+    der[sz++] = ASN_INTEGER;
+    der[sz++] = 1;
+    der[sz++] = 0x01;
+    der[sz++] = ASN_SEQUENCE | ASN_CONSTRUCTED;
+    der[sz++] = (byte)(sizeof(algId) + 2);
+    der[sz++] = ASN_OBJECT_ID;
+    der[sz++] = (byte)sizeof(algId);
+    XMEMCPY(der + sz, algId, sizeof(algId)); sz += sizeof(algId);
+    der[sz++] = ASN_OCTET_STRING;
+    der[sz++] = (byte)(privKeySz + 2);
+    der[sz++] = ASN_OCTET_STRING;
+    der[sz++] = (byte)privKeySz;
+    XMEMSET(der + sz, keyPat, privKeySz); sz += privKeySz;
+    /* [1] publicKey trailer */
+    der[sz++] = ASN_CONTEXT_SPECIFIC | ASN_ASYMKEY_PUBKEY;
+    der[sz++] = (byte)pubKeySz;
+    XMEMSET(der + sz, pubPat, pubKeySz); sz += pubKeySz;
+    der[outerLenIdx] = (byte)(sz - outerLenIdx - 1);
+
+    EXPECT_TEST(test_ToTraditional_ex_once(der, sz, ED25519k, privKeySz));
+
+    /* v1 without publicKey: should still accept per RFC 5958. */
+    sz = 0;
+    der[sz++] = ASN_SEQUENCE | ASN_CONSTRUCTED;
+    outerLenIdx = sz;
+    der[sz++] = 0;
+    der[sz++] = ASN_INTEGER;
+    der[sz++] = 1;
+    der[sz++] = 0x01;
+    der[sz++] = ASN_SEQUENCE | ASN_CONSTRUCTED;
+    der[sz++] = (byte)(sizeof(algId) + 2);
+    der[sz++] = ASN_OBJECT_ID;
+    der[sz++] = (byte)sizeof(algId);
+    XMEMCPY(der + sz, algId, sizeof(algId)); sz += sizeof(algId);
+    der[sz++] = ASN_OCTET_STRING;
+    der[sz++] = (byte)(privKeySz + 2);
+    der[sz++] = ASN_OCTET_STRING;
+    der[sz++] = (byte)privKeySz;
+    XMEMSET(der + sz, keyPat, privKeySz); sz += privKeySz;
+    der[outerLenIdx] = (byte)(sz - outerLenIdx - 1);
+
+    EXPECT_TEST(test_ToTraditional_ex_once(der, sz, ED25519k, privKeySz));
+#endif /* HAVE_PKCS8 && HAVE_ED25519 */
+    return EXPECT_RESULT();
+}
+
+/* Encoder/parser round trip: ToTraditional_ex() must accept both forms created
+ * by SetAsymKeyDer() (v0 with PrivateKeyToDer, v1 with KeyToDer). */
+int test_ToTraditional_ex_roundtrip(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_PKCS8) && \
+    (defined(WOLFSSL_TEST_CERT) || defined(OPENSSL_EXTRA) || \
+     defined(OPENSSL_EXTRA_X509_SMALL) || defined(WOLFSSL_PUBLIC_ASN))
+
+#if defined(HAVE_ED25519) && defined(HAVE_ED25519_KEY_EXPORT) && \
+    defined(WOLFSSL_KEY_GEN)
+    {
+        ed25519_key key;
+        WC_RNG rng;
+        byte der[256];
+        int  derSz = 0;
+
+        XMEMSET(&key, 0, sizeof(key));
+        XMEMSET(&rng, 0, sizeof(rng));
+        ExpectIntEQ(wc_InitRng(&rng), 0);
+        ExpectIntEQ(wc_ed25519_init(&key), 0);
+        ExpectIntEQ(wc_ed25519_make_key(&rng, ED25519_KEY_SIZE, &key), 0);
+
+        if (EXPECT_SUCCESS()) {
+            ExpectIntGT(derSz = wc_Ed25519KeyToDer(&key, der, sizeof(der)), 0);
+            EXPECT_TEST(test_ToTraditional_ex_once(der, (word32)derSz, ED25519k,
+                ED25519_KEY_SIZE));
+
+            derSz = wc_Ed25519PrivateKeyToDer(&key, der, sizeof(der));
+            ExpectIntGT(derSz, 0);
+            EXPECT_TEST(test_ToTraditional_ex_once(der, (word32)derSz, ED25519k,
+                ED25519_KEY_SIZE));
+        }
+
+        wc_ed25519_free(&key);
+        wc_FreeRng(&rng);
+    }
+#endif /* HAVE_ED25519 */
+
+#if defined(HAVE_ED448) && defined(HAVE_ED448_KEY_EXPORT) && \
+    defined(WOLFSSL_KEY_GEN)
+    {
+        ed448_key key;
+        WC_RNG rng;
+        byte der[256];
+        int  derSz = 0;
+
+        XMEMSET(&key, 0, sizeof(key));
+        XMEMSET(&rng, 0, sizeof(rng));
+        ExpectIntEQ(wc_InitRng(&rng), 0);
+        ExpectIntEQ(wc_ed448_init(&key), 0);
+        ExpectIntEQ(wc_ed448_make_key(&rng, ED448_KEY_SIZE, &key), 0);
+
+        if (EXPECT_SUCCESS()) {
+            ExpectIntGT(derSz = wc_Ed448KeyToDer(&key, der, sizeof(der)), 0);
+            EXPECT_TEST(test_ToTraditional_ex_once(der, (word32)derSz, ED448k,
+                ED448_KEY_SIZE));
+
+            derSz = wc_Ed448PrivateKeyToDer(&key, der, sizeof(der));
+            ExpectIntGT(derSz, 0);
+            EXPECT_TEST(test_ToTraditional_ex_once(der, (word32)derSz, ED448k,
+                ED448_KEY_SIZE));
+        }
+
+        wc_ed448_free(&key);
+        wc_FreeRng(&rng);
+    }
+#endif /* HAVE_ED448 */
+
+#if defined(HAVE_DILITHIUM) && \
+    !defined(WOLFSSL_DILITHIUM_NO_MAKE_KEY) && \
+    !defined(WOLFSSL_DILITHIUM_NO_ASN1) && \
+    (!defined(WOLFSSL_NO_ML_DSA_44) || !defined(WOLFSSL_NO_ML_DSA_65) || \
+     !defined(WOLFSSL_NO_ML_DSA_87))
+    {
+        static const struct {
+            int   wcLevel;
+            word32 oidSum;
+            word32 privKeySz;
+        } variants[] = {
+        #ifndef WOLFSSL_NO_ML_DSA_44
+            { WC_ML_DSA_44, ML_DSA_LEVEL2k, ML_DSA_LEVEL2_KEY_SIZE },
+        #endif
+        #ifndef WOLFSSL_NO_ML_DSA_65
+            { WC_ML_DSA_65, ML_DSA_LEVEL3k, ML_DSA_LEVEL3_KEY_SIZE },
+        #endif
+        #ifndef WOLFSSL_NO_ML_DSA_87
+            { WC_ML_DSA_87, ML_DSA_LEVEL5k, ML_DSA_LEVEL5_KEY_SIZE },
+        #endif
+        };
+
+        const word32 derMaxSz = DILITHIUM_MAX_BOTH_KEY_DER_SIZE;
+        byte* der = NULL;
+        WC_RNG rng;
+        size_t i;
+        int derSz;
+
+        XMEMSET(&rng, 0, sizeof(rng));
+        ExpectIntEQ(wc_InitRng(&rng), 0);
+        ExpectNotNull(der = (byte*)XMALLOC(derMaxSz, NULL,
+            DYNAMIC_TYPE_TMP_BUFFER));
+
+        for (i = 0; i < sizeof(variants) / sizeof(variants[0]); i++) {
+            dilithium_key key;
+
+            XMEMSET(&key, 0, sizeof(key));
+            ExpectIntEQ(wc_dilithium_init(&key), 0);
+            ExpectIntEQ(wc_dilithium_set_level(&key, variants[i].wcLevel), 0);
+            ExpectIntEQ(wc_dilithium_make_key(&key, &rng), 0);
+
+            if (EXPECT_SUCCESS()) {
+                ExpectIntGT(derSz = wc_Dilithium_KeyToDer(&key, der, derMaxSz),
+                    0);
+                EXPECT_TEST(test_ToTraditional_ex_once(der, (word32)derSz,
+                    variants[i].oidSum, variants[i].privKeySz));
+
+                derSz = wc_Dilithium_PrivateKeyToDer(&key, der, derMaxSz);
+                ExpectIntGT(derSz, 0);
+                EXPECT_TEST(test_ToTraditional_ex_once(der, (word32)derSz,
+                    variants[i].oidSum, variants[i].privKeySz));
+            }
+
+            wc_dilithium_free(&key);
+        }
+
+        XFREE(der, NULL, DYNAMIC_TYPE_TMP_BUFFER);
+        wc_FreeRng(&rng);
+    }
+#endif /* HAVE_DILITHIUM */
+
+#endif /* HAVE_PKCS8 */
+    return EXPECT_RESULT();
+}
+
+/* Trailing garbage that is neither [0] attributes nor [1] publicKey must
+ * still be rejected. */
+int test_ToTraditional_ex_negative(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_PKCS8) && defined(HAVE_ED25519) && \
+    defined(HAVE_ED25519_KEY_EXPORT) && defined(WOLFSSL_KEY_GEN) && \
+    defined(WOLFSSL_ASN_TEMPLATE) && \
+    (defined(WOLFSSL_TEST_CERT) || defined(OPENSSL_EXTRA) || \
+     defined(OPENSSL_EXTRA_X509_SMALL) || defined(WOLFSSL_PUBLIC_ASN))
+    ed25519_key key;
+    WC_RNG rng;
+    byte der[256];
+    byte copy[256];
+    int  derSz = 0;
+    word32 algId;
+
+    XMEMSET(&key, 0, sizeof(key));
+    XMEMSET(&rng, 0, sizeof(rng));
+    ExpectIntEQ(wc_InitRng(&rng), 0);
+    ExpectIntEQ(wc_ed25519_init(&key), 0);
+    ExpectIntEQ(wc_ed25519_make_key(&rng, ED25519_KEY_SIZE, &key), 0);
+    ExpectIntGT(derSz = wc_Ed25519PrivateKeyToDer(&key, der, sizeof(der)), 0);
+
+    if (EXPECT_SUCCESS() && (derSz > 0) &&
+        ((size_t)derSz + 1 <= sizeof(copy))) {
+        /* Append one byte of trailing data, grow outer SEQ length to cover.
+         * Ed25519 PKCS#8 outer SEQ is under 128 bytes, expect DER short form
+         * so the negative path is always exercised. */
+        XMEMCPY(copy, der, (size_t)derSz);
+        ExpectTrue(copy[1] < 0x80);
+        if (EXPECT_SUCCESS() && copy[1] < 0x80) {
+            copy[1] = (byte)(copy[1] + 1);
+            copy[derSz] = 0x05;
+            algId = 0;
+            ExpectIntLT(ToTraditional_ex(copy, (word32)(derSz + 1), &algId), 0);
+        }
+    }
+
+    /* publicKey trailer is permitted only when version == v1 */
+    if (EXPECT_SUCCESS() && (derSz > 0) &&
+        ((size_t)derSz + 2 + ED25519_PUB_KEY_SIZE <= sizeof(copy))) {
+        word32 trailerSz = 2 + ED25519_PUB_KEY_SIZE;
+        XMEMCPY(copy, der, (size_t)derSz);
+        ExpectTrue(copy[1] < (byte)(0x80 - trailerSz));
+        if (EXPECT_SUCCESS() && copy[1] < (byte)(0x80 - trailerSz)) {
+            copy[1] = (byte)(copy[1] + trailerSz);
+            copy[derSz] = ASN_CONTEXT_SPECIFIC | ASN_ASYMKEY_PUBKEY;
+            copy[derSz + 1] = ED25519_PUB_KEY_SIZE;
+            XMEMSET(copy + derSz + 2, 0xDD, ED25519_PUB_KEY_SIZE);
+            algId = 0;
+            ExpectIntLT(ToTraditional_ex(copy,
+                (word32)(derSz + (int)trailerSz), &algId), 0);
+        }
+    }
+
+    /* v1 buffer (with publicKey) plus extra trailing garbage. */
+    ExpectIntGT(derSz = wc_Ed25519KeyToDer(&key, der, sizeof(der)), 0);
+    if (EXPECT_SUCCESS() && (derSz > 0) &&
+        ((size_t)derSz + 1 <= sizeof(copy))) {
+        XMEMCPY(copy, der, (size_t)derSz);
+        ExpectTrue(copy[1] < 0x80);
+        if (EXPECT_SUCCESS() && copy[1] < 0x80) {
+            copy[1] = (byte)(copy[1] + 1);
+            copy[derSz] = 0x05;
+            algId = 0;
+            ExpectIntLT(ToTraditional_ex(copy, (word32)(derSz + 1), &algId), 0);
+        }
+    }
+
+    wc_ed25519_free(&key);
+    wc_FreeRng(&rng);
+#endif
+    return EXPECT_RESULT();
+}
+
+/* ML-DSA AlgorithmIdentifier has no parameters per FIPS 204. Verify
+ * ToTraditional_ex() rejects a PKCS#8 whose algoSeq carries trailing NULL
+ * or OBJECT_ID parameters. Template parser only (legacy is lenient). */
+int test_ToTraditional_ex_mldsa_bad_params(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_PKCS8) && defined(HAVE_DILITHIUM) && \
+    defined(WOLFSSL_ASN_TEMPLATE) && \
+    (defined(WOLFSSL_TEST_CERT) || defined(OPENSSL_EXTRA) || \
+     defined(OPENSSL_EXTRA_X509_SMALL) || defined(WOLFSSL_PUBLIC_ASN))
+    /* ML-DSA-65 OID body: 2.16.840.1.101.3.4.3.18 */
+    static const byte mldsaOid[] = { 0x60, 0x86, 0x48, 0x01, 0x65, 0x03,
+                                     0x04, 0x03, 0x12 };
+    /* Single-arc OID body, used only to occupy the OBJECT_ID slot. */
+    static const byte extraOid[] = { 0x01 };
+    byte der[64];
+    byte copy[64];
+    word32 sz;
+    word32 outerLenIdx;
+    word32 algId;
+    const word32 privKeySz = 4;
+    const byte   privBody  = 0xAA;
+
+    /* Bad case, algoSeq = { OID, NULL } */
+    sz = 0;
+    der[sz++] = ASN_SEQUENCE | ASN_CONSTRUCTED;
+    outerLenIdx = sz;
+    der[sz++] = 0;  /* outer length, filled in below */
+    der[sz++] = ASN_INTEGER;
+    der[sz++] = 1;
+    der[sz++] = 0x00;
+    der[sz++] = ASN_SEQUENCE | ASN_CONSTRUCTED;
+    der[sz++] = (byte)(sizeof(mldsaOid) + 2 + 2);
+    der[sz++] = ASN_OBJECT_ID;
+    der[sz++] = (byte)sizeof(mldsaOid);
+    XMEMCPY(der + sz, mldsaOid, sizeof(mldsaOid)); sz += sizeof(mldsaOid);
+    /* Disallowed, NULL parameter after the ML-DSA OID. */
+    der[sz++] = ASN_TAG_NULL;
+    der[sz++] = 0;
+    der[sz++] = ASN_OCTET_STRING;
+    der[sz++] = (byte)(privKeySz + 2);
+    der[sz++] = ASN_OCTET_STRING;
+    der[sz++] = (byte)privKeySz;
+    XMEMSET(der + sz, privBody, privKeySz); sz += privKeySz;
+    der[outerLenIdx] = (byte)(sz - outerLenIdx - 1);
+
+    XMEMCPY(copy, der, sz);
+    algId = 0;
+    ExpectIntLT(ToTraditional_ex(copy, sz, &algId), 0);
+
+    /* Bad case, algoSeq = { OID, OBJECT_ID } */
+    sz = 0;
+    der[sz++] = ASN_SEQUENCE | ASN_CONSTRUCTED;
+    outerLenIdx = sz;
+    der[sz++] = 0;
+    der[sz++] = ASN_INTEGER;
+    der[sz++] = 1;
+    der[sz++] = 0x00;
+    der[sz++] = ASN_SEQUENCE | ASN_CONSTRUCTED;
+    der[sz++] = (byte)(sizeof(mldsaOid) + 2 + sizeof(extraOid) + 2);
+    der[sz++] = ASN_OBJECT_ID;
+    der[sz++] = (byte)sizeof(mldsaOid);
+    XMEMCPY(der + sz, mldsaOid, sizeof(mldsaOid)); sz += sizeof(mldsaOid);
+    /* Disallowed, OBJECT_ID parameter after the ML-DSA OID. */
+    der[sz++] = ASN_OBJECT_ID;
+    der[sz++] = (byte)sizeof(extraOid);
+    XMEMCPY(der + sz, extraOid, sizeof(extraOid)); sz += sizeof(extraOid);
+    der[sz++] = ASN_OCTET_STRING;
+    der[sz++] = (byte)(privKeySz + 2);
+    der[sz++] = ASN_OCTET_STRING;
+    der[sz++] = (byte)privKeySz;
+    XMEMSET(der + sz, privBody, privKeySz); sz += privKeySz;
+    der[outerLenIdx] = (byte)(sz - outerLenIdx - 1);
+
+    XMEMCPY(copy, der, sz);
+    algId = 0;
+    ExpectIntLT(ToTraditional_ex(copy, sz, &algId), 0);
+#endif
+    return EXPECT_RESULT();
+}
diff --git a/tests/api/test_asn.h b/tests/api/test_asn.h
index 598c255f6b..2654e3e858 100644
--- a/tests/api/test_asn.h
+++ b/tests/api/test_asn.h
@@ -30,10 +30,16 @@ int test_DecodeAsymKey_negative(void);
 int test_GetSetShortInt(void);
 int test_wc_IndexSequenceOf(void);
 int test_wolfssl_local_MatchBaseName(void);
+int test_wolfssl_local_MatchDnsConstraintWildcard(void);
+int test_wolfssl_local_MatchUriNameConstraint(void);
 int test_wc_DecodeRsaPssParams(void);
 int test_SerialNumber0_RootCA(void);
 int test_DecodeAltNames_length_underflow(void);
 int test_wc_DecodeObjectId(void);
+int test_ToTraditional_ex_handcrafted(void);
+int test_ToTraditional_ex_roundtrip(void);
+int test_ToTraditional_ex_negative(void);
+int test_ToTraditional_ex_mldsa_bad_params(void);
 
 #define TEST_ASN_DECLS                                              \
     TEST_DECL_GROUP("asn", test_SetAsymKeyDer),                     \
@@ -42,9 +48,15 @@ int test_wc_DecodeObjectId(void);
     TEST_DECL_GROUP("asn", test_GetSetShortInt),                    \
     TEST_DECL_GROUP("asn", test_wc_IndexSequenceOf),                \
     TEST_DECL_GROUP("asn", test_wolfssl_local_MatchBaseName),       \
+    TEST_DECL_GROUP("asn", test_wolfssl_local_MatchDnsConstraintWildcard), \
+    TEST_DECL_GROUP("asn", test_wolfssl_local_MatchUriNameConstraint), \
     TEST_DECL_GROUP("asn", test_wc_DecodeRsaPssParams),             \
     TEST_DECL_GROUP("asn", test_SerialNumber0_RootCA),              \
     TEST_DECL_GROUP("asn", test_DecodeAltNames_length_underflow),   \
-    TEST_DECL_GROUP("asn", test_wc_DecodeObjectId)
+    TEST_DECL_GROUP("asn", test_wc_DecodeObjectId),                 \
+    TEST_DECL_GROUP("asn", test_ToTraditional_ex_handcrafted),      \
+    TEST_DECL_GROUP("asn", test_ToTraditional_ex_roundtrip),        \
+    TEST_DECL_GROUP("asn", test_ToTraditional_ex_negative),         \
+    TEST_DECL_GROUP("asn", test_ToTraditional_ex_mldsa_bad_params)
 
 #endif /* WOLFCRYPT_TEST_ASN_H */
diff --git a/tests/api/test_certman.c b/tests/api/test_certman.c
index f611010565..1c73e64a73 100644
--- a/tests/api/test_certman.c
+++ b/tests/api/test_certman.c
@@ -2109,6 +2109,70 @@ int test_wolfSSL_X509_check_host_IP_only_SAN_CN_fallback(void)
     return EXPECT_RESULT();
 }
 
+int test_wolfSSL_X509_check_host_URI_SAN_not_DNS_match(void)
+{
+    EXPECT_DECLS;
+#if !defined(NO_FILESYSTEM) && !defined(NO_CERTS) && !defined(NO_RSA) && \
+    defined(OPENSSL_EXTRA) && defined(WOLFSSL_CERT_GEN) && \
+    defined(WOLFSSL_CERT_EXT) && defined(WOLFSSL_ALT_NAMES) && \
+    !defined(NO_SHA256)
+    /* RFC 6125 Sec. 6.4 / RFC 9525 Sec. 6.3: DNS-ID reference identifiers
+     * must be matched only against dNSName SANs, not uniformResourceIdentifier.
+     * wolfSSL_X509_add_altname() is used to attach a bare-hostname URI SAN
+     * (the misissue shape that can reach altNames when certificate parsing is
+     * built without strict URI checks). URI SAN presence still suppresses CN
+     * fallback per RFC 6125 Sec. 6.4.4. */
+    WOLFSSL_EVP_PKEY *priv = NULL;
+    WOLFSSL_X509_NAME* name = NULL;
+    const char* server_cert = "./certs/test/server-goodcn.pem";
+    const char hostName[] = "cnhost.local";
+    const char uriSan[] = "http://cnhost.local";
+    byte    *pt;
+    WOLFSSL_X509 *leafUri = NULL;
+    WOLFSSL_X509 *leafUriDns = NULL;
+
+    pt = (byte*)server_key_der_2048;
+    ExpectNotNull(priv = wolfSSL_d2i_PrivateKey(EVP_PKEY_RSA, NULL,
+                (const unsigned char**)&pt, sizeof_server_key_der_2048));
+
+    ExpectNotNull(leafUri = wolfSSL_X509_load_certificate_file(server_cert,
+                WOLFSSL_FILETYPE_PEM));
+    ExpectNotNull(name = X509_NAME_new());
+    ExpectIntEQ(X509_NAME_add_entry_by_txt(name, "commonName", MBSTRING_UTF8,
+                (byte*)hostName, (int)XSTRLEN(hostName), -1, 0), SSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_X509_set_subject_name(leafUri, name), WOLFSSL_SUCCESS);
+    X509_NAME_free(name);
+    name = NULL;
+    ExpectIntEQ(wolfSSL_X509_add_altname(leafUri, uriSan, ASN_URI_TYPE),
+            WOLFSSL_SUCCESS);
+    ExpectIntGT(wolfSSL_X509_sign(leafUri, priv, EVP_sha256()), 0);
+    ExpectIntEQ(wolfSSL_X509_check_host(leafUri, hostName, XSTRLEN(hostName),
+                0, NULL), WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+
+    ExpectNotNull(leafUriDns = wolfSSL_X509_load_certificate_file(server_cert,
+                WOLFSSL_FILETYPE_PEM));
+    ExpectNotNull(name = X509_NAME_new());
+    ExpectIntEQ(X509_NAME_add_entry_by_txt(name, "commonName", MBSTRING_UTF8,
+                (byte*)hostName, (int)XSTRLEN(hostName), -1, 0), SSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_X509_set_subject_name(leafUriDns, name),
+            WOLFSSL_SUCCESS);
+    X509_NAME_free(name);
+    name = NULL;
+    ExpectIntEQ(wolfSSL_X509_add_altname(leafUriDns, uriSan, ASN_URI_TYPE),
+            WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_X509_add_altname(leafUriDns, hostName, ASN_DNS_TYPE),
+            WOLFSSL_SUCCESS);
+    ExpectIntGT(wolfSSL_X509_sign(leafUriDns, priv, EVP_sha256()), 0);
+    ExpectIntEQ(wolfSSL_X509_check_host(leafUriDns, hostName,
+                XSTRLEN(hostName), 0, NULL), WOLFSSL_SUCCESS);
+
+    wolfSSL_X509_free(leafUri);
+    wolfSSL_X509_free(leafUriDns);
+    wolfSSL_EVP_PKEY_free(priv);
+#endif
+    return EXPECT_RESULT();
+}
+
 int test_wolfSSL_CertManagerCRL(void)
 {
     EXPECT_DECLS;
diff --git a/tests/api/test_certman.h b/tests/api/test_certman.h
index 52bef8e1aa..173c0f7e30 100644
--- a/tests/api/test_certman.h
+++ b/tests/api/test_certman.h
@@ -40,6 +40,7 @@ int test_wolfSSL_CertManagerNameConstraint_IP_SAN(void);
 int test_wolfSSL_CertManagerNameConstraint_RID_SAN(void);
 int test_wolfSSL_X509_get_ext_d2i_RID_SAN(void);
 int test_wolfSSL_X509_check_host_IP_only_SAN_CN_fallback(void);
+int test_wolfSSL_X509_check_host_URI_SAN_not_DNS_match(void);
 int test_wolfSSL_CertManagerCRL(void);
 int test_wolfSSL_CRL_reason_extensions_cleanup(void);
 int test_wolfSSL_CRL_static_revoked_list(void);
@@ -70,6 +71,8 @@ int test_wolfSSL_X509_V_ERR_strings(void);
     TEST_DECL_GROUP("certman", test_wolfSSL_X509_get_ext_d2i_RID_SAN),       \
     TEST_DECL_GROUP("certman",                                              \
         test_wolfSSL_X509_check_host_IP_only_SAN_CN_fallback),              \
+    TEST_DECL_GROUP("certman",                                              \
+        test_wolfSSL_X509_check_host_URI_SAN_not_DNS_match),                \
     TEST_DECL_GROUP("certman", test_wolfSSL_CertManagerCRL),                \
     TEST_DECL_GROUP("certman", test_wolfSSL_CRL_reason_extensions_cleanup), \
     TEST_DECL_GROUP("certman", test_wolfSSL_CRL_static_revoked_list),      \
diff --git a/tests/api/test_curve25519.c b/tests/api/test_curve25519.c
index 4108ad0d34..33c51253f7 100644
--- a/tests/api/test_curve25519.c
+++ b/tests/api/test_curve25519.c
@@ -398,6 +398,99 @@ int test_wc_curve25519_shared_secret_zero_check(void)
     return EXPECT_RESULT();
 } /* END test_wc_curve25519_shared_secret_zero_check */
 
+/*
+ * Known-answer tests for wc_curve25519_shared_secret_ex.
+ *
+ * Both vectors share one private scalar and produce a shared secret that is a
+ * small canonical value (9 and 16, little-endian). Because the result is close
+ * to a multiple of the field prime, these exercise the final modular reduction
+ * of the X25519 computation: a result that was only reduced mod 2^256 (or left
+ * in [p, 2^255)) instead of fully reduced mod 2^255-19 would not match.
+ * All values are 32-byte little-endian encodings per RFC 7748.
+ */
+int test_wc_curve25519_shared_secret_ex_kat(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_CURVE25519) && defined(HAVE_CURVE25519_KEY_IMPORT)
+    /* Private scalar shared by both vectors. */
+    static const byte kPriv[CURVE25519_KEYSIZE] = {
+        0x60, 0xa3, 0xa4, 0xf1, 0x30, 0xb9, 0x8a, 0x5b,
+        0xe4, 0xb1, 0xce, 0xdb, 0x7c, 0xb8, 0x55, 0x84,
+        0xa3, 0x52, 0x0e, 0x14, 0x2d, 0x47, 0x4d, 0xc9,
+        0xcc, 0xb9, 0x09, 0xa0, 0x73, 0xa9, 0x76, 0x7f
+    };
+    /* Vector 1 public value, expected shared secret == 9. */
+    static const byte kPub1[CURVE25519_KEYSIZE] = {
+        0x3b, 0x18, 0xdf, 0x1e, 0x50, 0xb8, 0x99, 0xeb,
+        0xd5, 0x88, 0xc3, 0x16, 0x1c, 0xbd, 0x3b, 0xf9,
+        0x8e, 0xbc, 0xc2, 0xc1, 0xf7, 0xdf, 0x53, 0xb8,
+        0x11, 0xbd, 0x0e, 0x91, 0xb4, 0xd5, 0x15, 0x3d
+    };
+    static const byte kExpected1[CURVE25519_KEYSIZE] = {
+        0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
+    };
+    /* Vector 2 public value, expected shared secret == 16. */
+    static const byte kPub2[CURVE25519_KEYSIZE] = {
+        0xca, 0xb6, 0xf9, 0xe7, 0xd8, 0xce, 0x00, 0xdf,
+        0xce, 0xa9, 0xbb, 0xd8, 0xf0, 0x69, 0xef, 0x7f,
+        0xb2, 0xac, 0x50, 0x4a, 0xbf, 0x83, 0xb8, 0x7d,
+        0xb6, 0x01, 0xb5, 0xae, 0x0a, 0x7f, 0x76, 0x15
+    };
+    static const byte kExpected2[CURVE25519_KEYSIZE] = {
+        0x10, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
+    };
+    /* Table-driven so both vectors run through the identical code path. */
+    struct {
+        const byte* pub;
+        const byte* expected;
+    } vec[2];
+    curve25519_key private_key;
+    curve25519_key public_key;
+    WC_RNG         rng;
+    byte           out[CURVE25519_KEYSIZE];
+    word32         outLen;
+    int i;
+
+    vec[0].pub = kPub1; vec[0].expected = kExpected1;
+    vec[1].pub = kPub2; vec[1].expected = kExpected2;
+
+    XMEMSET(&rng, 0, sizeof(WC_RNG));
+    ExpectIntEQ(wc_InitRng(&rng), 0);
+
+    for (i = 0; i < 2; i++) {
+        XMEMSET(&private_key, 0, sizeof(private_key));
+        XMEMSET(&public_key, 0, sizeof(public_key));
+        ExpectIntEQ(wc_curve25519_init(&private_key), 0);
+        ExpectIntEQ(wc_curve25519_init(&public_key), 0);
+    #ifdef WOLFSSL_CURVE25519_BLINDING
+        ExpectIntEQ(wc_curve25519_set_rng(&private_key, &rng), 0);
+    #endif
+        ExpectIntEQ(wc_curve25519_import_private_ex(kPriv, sizeof(kPriv),
+            &private_key, EC25519_LITTLE_ENDIAN), 0);
+        ExpectIntEQ(wc_curve25519_import_public_ex(vec[i].pub,
+            CURVE25519_KEYSIZE, &public_key, EC25519_LITTLE_ENDIAN), 0);
+
+        outLen = sizeof(out);
+        ExpectIntEQ(wc_curve25519_shared_secret_ex(&private_key, &public_key,
+            out, &outLen, EC25519_LITTLE_ENDIAN), 0);
+        ExpectIntEQ(outLen, CURVE25519_KEYSIZE);
+        ExpectIntEQ(XMEMCMP(out, vec[i].expected, CURVE25519_KEYSIZE), 0);
+
+        wc_curve25519_free(&private_key);
+        wc_curve25519_free(&public_key);
+    }
+
+    DoExpectIntEQ(wc_FreeRng(&rng), 0);
+#endif
+    return EXPECT_RESULT();
+} /* END test_wc_curve25519_shared_secret_ex_kat */
+
 /*
  * Testing wc_curve25519_make_pub
  */
diff --git a/tests/api/test_curve25519.h b/tests/api/test_curve25519.h
index c65ccb34f8..d6b749bed6 100644
--- a/tests/api/test_curve25519.h
+++ b/tests/api/test_curve25519.h
@@ -31,6 +31,7 @@ int test_wc_curve25519_export_key_raw_ex(void);
 int test_wc_curve25519_make_key(void);
 int test_wc_curve25519_shared_secret_ex(void);
 int test_wc_curve25519_shared_secret_zero_check(void);
+int test_wc_curve25519_shared_secret_ex_kat(void);
 int test_wc_curve25519_make_pub(void);
 int test_wc_curve25519_export_public_ex(void);
 int test_wc_curve25519_export_private_raw_ex(void);
@@ -47,6 +48,7 @@ int test_wc_Curve25519KeyToDer_oneasymkey_version(void);
     TEST_DECL_GROUP("curve25519", test_wc_curve25519_make_key),                \
     TEST_DECL_GROUP("curve25519", test_wc_curve25519_shared_secret_ex),        \
     TEST_DECL_GROUP("curve25519", test_wc_curve25519_shared_secret_zero_check),\
+    TEST_DECL_GROUP("curve25519", test_wc_curve25519_shared_secret_ex_kat),    \
     TEST_DECL_GROUP("curve25519", test_wc_curve25519_make_pub),                \
     TEST_DECL_GROUP("curve25519", test_wc_curve25519_export_public_ex),        \
     TEST_DECL_GROUP("curve25519", test_wc_curve25519_export_private_raw_ex),   \
diff --git a/tests/api/test_dtls.c b/tests/api/test_dtls.c
index d72a9395c5..572dda6c94 100644
--- a/tests/api/test_dtls.c
+++ b/tests/api/test_dtls.c
@@ -32,8 +32,19 @@
 #include <wolfssl/internal.h>
 
 #include <tests/utils.h>
+#include <tests/api/api.h>
 #include <tests/api/test_dtls.h>
 
+/* Cast used to make send()/recv() buffer arguments portable between
+ * Windows (char*) and POSIX (void*). Mirrors the private macro in
+ * tests/api.c so the DTLS plaintext/fragments tests moved out of api.c
+ * still build here. */
+#ifdef USE_WINDOWS_API
+    #define MESSAGE_TYPE_CAST char*
+#else
+    #define MESSAGE_TYPE_CAST void*
+#endif
+
 int test_dtls12_basic_connection_id(void)
 {
     EXPECT_DECLS;
@@ -408,165 +419,6 @@ loop_exit:
     return EXPECT_RESULT();
 }
 
-int test_dtls13_basic_connection_id(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13) \
-    && defined(WOLFSSL_DTLS_CID)
-    unsigned char client_cid[] = { 9, 8, 7, 6, 5, 4, 3, 2, 1, 0 };
-    unsigned char server_cid[] = { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 };
-    unsigned char readBuf[50];
-    void *        cid = NULL;
-    const char* params[] = {
-#ifndef NO_SHA256
-#ifdef WOLFSSL_AES_128
-#ifdef HAVE_AESGCM
-        "TLS13-AES128-GCM-SHA256",
-#endif
-#if defined(HAVE_CHACHA) && defined(HAVE_POLY1305)
-        "TLS13-CHACHA20-POLY1305-SHA256",
-#endif
-#ifdef HAVE_AESCCM
-        "TLS13-AES128-CCM-8-SHA256",
-        "TLS13-AES128-CCM-SHA256",
-#endif
-#endif
-#ifdef HAVE_NULL_CIPHER
-        "TLS13-SHA256-SHA256",
-#endif
-#endif
-    };
-    size_t i;
-
-    /* We check if the side included the CID in their output */
-#define CLIENT_CID() mymemmem(test_ctx.s_buff, test_ctx.s_len, \
-                              client_cid, sizeof(client_cid))
-#define SERVER_CID() mymemmem(test_ctx.c_buff, test_ctx.c_len, \
-                              server_cid, sizeof(server_cid))
-#define RESET_CID(cid) if ((cid) != NULL) { \
-                           ((char*)(cid))[0] = -1; \
-                       }
-
-
-    printf("\n");
-    for (i = 0; i < XELEM_CNT(params) && EXPECT_SUCCESS(); i++) {
-        WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
-        WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
-        struct test_memio_ctx test_ctx;
-
-        printf("Testing %s ... ", params[i]);
-
-        XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-
-        ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-            wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
-
-        ExpectIntEQ(wolfSSL_set_cipher_list(ssl_c, params[i]), WOLFSSL_SUCCESS);
-        ExpectIntEQ(wolfSSL_set_cipher_list(ssl_s, params[i]), WOLFSSL_SUCCESS);
-
-        ExpectIntEQ(wolfSSL_dtls_cid_use(ssl_c), 1);
-        ExpectIntEQ(wolfSSL_dtls_cid_set(ssl_c, server_cid, sizeof(server_cid)),
-                1);
-        ExpectIntEQ(wolfSSL_dtls_cid_use(ssl_s), 1);
-        ExpectIntEQ(wolfSSL_dtls_cid_set(ssl_s, client_cid, sizeof(client_cid)),
-                1);
-
-        /* CH1 */
-        ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-        ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-        ExpectNull(CLIENT_CID());
-        /* HRR */
-        ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-        ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-        ExpectNull(SERVER_CID());
-        /* CH2 */
-        ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-        ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-        ExpectNull(CLIENT_CID());
-        /* Server first flight */
-        ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-        ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-        ExpectNotNull(SERVER_CID());
-        /* Client second flight */
-        ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-        ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-        ExpectNotNull(CLIENT_CID());
-        /* Server process flight */
-        ExpectIntEQ(wolfSSL_negotiate(ssl_s), 1);
-        /* Client process flight */
-        ExpectIntEQ(wolfSSL_negotiate(ssl_c), 1);
-
-        /* Write some data */
-        ExpectIntEQ(wolfSSL_write(ssl_c, params[i], (int)XSTRLEN(params[i])),
-                XSTRLEN(params[i]));
-        ExpectNotNull(CLIENT_CID());
-        ExpectIntEQ(wolfSSL_write(ssl_s, params[i], (int)XSTRLEN(params[i])),
-                XSTRLEN(params[i]));
-        ExpectNotNull(SERVER_CID());
-        /* Read the data */
-        XMEMSET(readBuf, 0, sizeof(readBuf));
-        ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)),
-                XSTRLEN(params[i]));
-        ExpectStrEQ(readBuf, params[i]);
-        XMEMSET(readBuf, 0, sizeof(readBuf));
-        ExpectIntEQ(wolfSSL_read(ssl_s, readBuf, sizeof(readBuf)),
-                XSTRLEN(params[i]));
-        ExpectStrEQ(readBuf, params[i]);
-        /* Write short data */
-        ExpectIntEQ(wolfSSL_write(ssl_c, params[i], 1), 1);
-        ExpectNotNull(CLIENT_CID());
-        ExpectIntEQ(wolfSSL_write(ssl_s, params[i], 1), 1);
-        ExpectNotNull(SERVER_CID());
-        /* Read the short data */
-        XMEMSET(readBuf, 0, sizeof(readBuf));
-        ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), 1);
-        ExpectIntEQ(readBuf[0], params[i][0]);
-        XMEMSET(readBuf, 0, sizeof(readBuf));
-        ExpectIntEQ(wolfSSL_read(ssl_s, readBuf, sizeof(readBuf)), 1);
-        ExpectIntEQ(readBuf[0], params[i][0]);
-        /* Write some data but with wrong CID */
-        ExpectIntEQ(wolfSSL_write(ssl_c, params[i], (int)XSTRLEN(params[i])),
-                XSTRLEN(params[i]));
-        /* Reset client cid. */
-        ExpectNotNull(cid = CLIENT_CID());
-        RESET_CID(cid);
-        ExpectIntEQ(wolfSSL_write(ssl_s, params[i], (int)XSTRLEN(params[i])),
-                XSTRLEN(params[i]));
-        /* Reset server cid. */
-        ExpectNotNull(cid = SERVER_CID());
-        RESET_CID(cid);
-        /* Try to read the data but it shouldn't be there */
-        ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), -1);
-        ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-        ExpectIntEQ(wolfSSL_read(ssl_s, readBuf, sizeof(readBuf)), -1);
-        ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-
-        /* Close connection */
-        ExpectIntEQ(wolfSSL_shutdown(ssl_c), WOLFSSL_SHUTDOWN_NOT_DONE);
-        ExpectNotNull(CLIENT_CID());
-        ExpectIntEQ(wolfSSL_shutdown(ssl_s), WOLFSSL_SHUTDOWN_NOT_DONE);
-        ExpectNotNull(SERVER_CID());
-        ExpectIntEQ(wolfSSL_shutdown(ssl_c), 1);
-        ExpectIntEQ(wolfSSL_shutdown(ssl_s), 1);
-
-        if (EXPECT_SUCCESS())
-            printf("ok\n");
-        else
-            printf("failed\n");
-
-        wolfSSL_free(ssl_c);
-        wolfSSL_CTX_free(ctx_c);
-        wolfSSL_free(ssl_s);
-        wolfSSL_CTX_free(ctx_s);
-    }
-
-#undef CLIENT_CID
-#undef SERVER_CID
-#undef RESET_CID
-
-#endif
-    return EXPECT_RESULT();
-}
 
 /** Test DTLS 1.3 behavior when server hits WANT_WRITE during HRR
  * The test sets up a DTLS 1.3 connection where the server is forced to
@@ -574,74 +426,7 @@ int test_dtls13_basic_connection_id(void)
  * application data is exchanged in both directions to verify the connection
  * works as expected.
  */
-int test_dtls13_hrr_want_write(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
-    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
-    const char msg[] = "hello";
-    const int msgLen = sizeof(msg);
-    struct test_memio_ctx test_ctx;
-    char readBuf[sizeof(msg)];
 
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-                    wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method),
-        0);
-
-    /* Client sends first ClientHello */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-
-    /* Force server to hit WANT_WRITE when producing the HRR */
-    test_memio_simulate_want_write(&test_ctx, 0, 1);
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_WRITE);
-
-    /* Allow the server to flush the HRR and proceed */
-    test_memio_simulate_want_write(&test_ctx, 0, 0);
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-
-    /* Resume the DTLS 1.3 handshake */
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-
-    /* Verify post-handshake application data in both directions */
-    XMEMSET(readBuf, 0, sizeof(readBuf));
-    ExpectIntEQ(wolfSSL_write(ssl_c, msg, msgLen), msgLen);
-    ExpectIntEQ(wolfSSL_read(ssl_s, readBuf, sizeof(readBuf)), msgLen);
-    ExpectStrEQ(readBuf, msg);
-
-    XMEMSET(readBuf, 0, sizeof(readBuf));
-    ExpectIntEQ(wolfSSL_write(ssl_s, msg, msgLen), msgLen);
-    ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), msgLen);
-    ExpectStrEQ(readBuf, msg);
-
-    wolfSSL_free(ssl_c);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_s);
-#endif
-    return EXPECT_RESULT();
-}
-
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
-struct test_dtls13_wwrite_ctx {
-    int want_write;
-    struct test_memio_ctx *text_ctx;
-};
-static int test_dtls13_want_write_send_cb(WOLFSSL *ssl, char *data, int sz, void *ctx)
-{
-    struct test_dtls13_wwrite_ctx *wwctx = (struct test_dtls13_wwrite_ctx *)ctx;
-    wwctx->want_write = !wwctx->want_write;
-    if (wwctx->want_write) {
-        return WOLFSSL_CBIO_ERR_WANT_WRITE;
-    }
-    return test_memio_write_cb(ssl, data, sz, wwctx->text_ctx);
-}
-#endif
 /** Test DTLS 1.3 behavior when every other write returns WANT_WRITE
  * The test sets up a DTLS 1.3 connection where both client and server
  * alternate between WANT_WRITE and successful writes. After the handshake,
@@ -651,82 +436,6 @@ static int test_dtls13_want_write_send_cb(WOLFSSL *ssl, char *data, int sz, void
  * Data exchanged after the handshake is also tested with simulated WANT_WRITE
  * conditions to ensure the connection remains functional.
  */
-int test_dtls13_every_write_want_write(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
-    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-    const char msg[] = "want-write";
-    const int msgLen = sizeof(msg);
-    char readBuf[sizeof(msg)];
-    struct test_dtls13_wwrite_ctx wwctx_c;
-    struct test_dtls13_wwrite_ctx wwctx_s;
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-                    wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method),
-        0);
-
-    wwctx_c.want_write = 0;
-    wwctx_c.text_ctx = &test_ctx;
-    wolfSSL_SetIOWriteCtx(ssl_c, &wwctx_c);
-    wolfSSL_SSLSetIOSend(ssl_c, test_dtls13_want_write_send_cb);
-    wwctx_s.want_write = 0;
-    wwctx_s.text_ctx = &test_ctx;
-    wolfSSL_SetIOWriteCtx(ssl_s, &wwctx_s);
-    wolfSSL_SSLSetIOSend(ssl_s, test_dtls13_want_write_send_cb);
-
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 20, NULL), 0);
-
-    ExpectTrue(wolfSSL_is_init_finished(ssl_c));
-    ExpectTrue(wolfSSL_is_init_finished(ssl_s));
-
-    test_memio_simulate_want_write(&test_ctx, 0, 0);
-    test_memio_simulate_want_write(&test_ctx, 1, 0);
-
-    wolfSSL_SetIOWriteCtx(ssl_c, &test_ctx);
-    wolfSSL_SSLSetIOSend(ssl_c, test_memio_write_cb);
-    wolfSSL_SetIOWriteCtx(ssl_s, &test_ctx);
-    wolfSSL_SSLSetIOSend(ssl_s, test_memio_write_cb);
-
-    XMEMSET(readBuf, 0, sizeof(readBuf));
-    ExpectIntEQ(wolfSSL_write(ssl_c, msg, msgLen), msgLen);
-    ExpectIntEQ(wolfSSL_read(ssl_s, readBuf, sizeof(readBuf)), msgLen);
-    ExpectStrEQ(readBuf, msg);
-
-    XMEMSET(readBuf, 0, sizeof(readBuf));
-    ExpectIntEQ(wolfSSL_write(ssl_s, msg, msgLen), msgLen);
-    ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), msgLen);
-    ExpectStrEQ(readBuf, msg);
-
-    test_memio_simulate_want_write(&test_ctx, 0, 1);
-    XMEMSET(readBuf, 0, sizeof(readBuf));
-    ExpectIntEQ(wolfSSL_write(ssl_s, msg, msgLen), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_WRITE);
-    test_memio_simulate_want_write(&test_ctx, 0, 0);
-    ExpectIntEQ(wolfSSL_write(ssl_s, msg, msgLen), msgLen);
-    ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), msgLen);
-    ExpectStrEQ(readBuf, msg);
-
-    XMEMSET(readBuf, 0, sizeof(readBuf));
-    test_memio_simulate_want_write(&test_ctx, 1, 1);
-    ExpectIntEQ(wolfSSL_write(ssl_c, msg, msgLen), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_WRITE);
-    test_memio_simulate_want_write(&test_ctx, 1, 0);
-    ExpectIntEQ(wolfSSL_write(ssl_c, msg, msgLen), msgLen);
-    ExpectIntEQ(wolfSSL_read(ssl_s, readBuf, sizeof(readBuf)), msgLen);
-    ExpectStrEQ(readBuf, msg);
-
-    wolfSSL_free(ssl_c);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_s);
-#endif
-    return EXPECT_RESULT();
-}
 
 int test_wolfSSL_dtls_cid_parse(void)
 {
@@ -809,253 +518,9 @@ int test_wolfSSL_dtls_set_pending_peer(void)
     return EXPECT_RESULT();
 }
 
-int test_dtls13_epochs(void) {
-    EXPECT_DECLS;
-#if defined(WOLFSSL_DTLS13) && !defined(NO_WOLFSSL_CLIENT)
-    WOLFSSL_CTX* ctx = NULL;
-    WOLFSSL* ssl = NULL;
-    byte input[20];
-    word32 inOutIdx = 0;
 
-    XMEMSET(input, 0, sizeof(input));
 
-    ExpectNotNull(ctx = wolfSSL_CTX_new(wolfDTLSv1_3_client_method()));
-    ExpectNotNull(ssl = wolfSSL_new(ctx));
-    /* Some manual setup to enter the epoch check */
-    ExpectTrue(ssl->options.tls1_3 = 1);
 
-    inOutIdx = 0;
-    if (ssl != NULL) ssl->keys.curEpoch64 = w64From32(0x0, 0x0);
-    ExpectIntEQ(DoApplicationData(ssl, input, &inOutIdx, 0), SANITY_MSG_E);
-    inOutIdx = 0;
-    if (ssl != NULL) ssl->keys.curEpoch64 = w64From32(0x0, 0x2);
-    ExpectIntEQ(DoApplicationData(ssl, input, &inOutIdx, 0), SANITY_MSG_E);
-
-    if (ssl != NULL) ssl->keys.curEpoch64 = w64From32(0x0, 0x1);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, client_hello), SANITY_MSG_E);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, server_hello), SANITY_MSG_E);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, hello_verify_request), SANITY_MSG_E);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, hello_retry_request), SANITY_MSG_E);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, hello_request), SANITY_MSG_E);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, encrypted_extensions), SANITY_MSG_E);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, server_key_exchange), SANITY_MSG_E);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, server_hello_done), SANITY_MSG_E);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, client_key_exchange), SANITY_MSG_E);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, certificate_request), SANITY_MSG_E);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, certificate), SANITY_MSG_E);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, certificate_verify), SANITY_MSG_E);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, finished), SANITY_MSG_E);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, certificate_status), SANITY_MSG_E);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, change_cipher_hs), SANITY_MSG_E);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, key_update), SANITY_MSG_E);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, session_ticket), SANITY_MSG_E);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, end_of_early_data), SANITY_MSG_E);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, message_hash), SANITY_MSG_E);
-    ExpectIntEQ(Dtls13CheckEpoch(ssl, no_shake), SANITY_MSG_E);
-
-    wolfSSL_CTX_free(ctx);
-    wolfSSL_free(ssl);
-#endif
-    return EXPECT_RESULT();
-}
-
-int test_dtls13_ack_order(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
-    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-    unsigned char readBuf[50];
-    word32 length = 0;
-    /* struct {
-     *     uint64 epoch;
-     *     uint64 sequence_number;
-     * } RecordNumber;
-     * Big endian */
-    static const unsigned char expected_output[] = {
-        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02,
-        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02,
-        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01,
-        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02,
-        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02,
-        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03,
-        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
-        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03,
-        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01,
-        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03,
-        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02,
-        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03,
-        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04,
-        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03,
-        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x06,
-    };
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-
-    /* Get a populated DTLS object */
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-    ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), -1);
-    /* Clear the buffer of any extra messages */
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    ExpectIntEQ(wolfSSL_read(ssl_s, readBuf, sizeof(readBuf)), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    ExpectIntEQ(test_ctx.c_len, 0);
-    ExpectIntEQ(test_ctx.s_len, 0);
-
-    /* Add seen records */
-    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 3), w64From32(0, 2)), 0);
-    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 3), w64From32(0, 0)), 0);
-    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 3), w64From32(0, 1)), 0);
-    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 3), w64From32(0, 4)), 0);
-    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 2), w64From32(0, 0)), 0);
-    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 3), w64From32(0, 6)), 0);
-    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 3), w64From32(0, 6)), 0);
-    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 2), w64From32(0, 1)), 0);
-    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 2), w64From32(0, 2)), 0);
-    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 2), w64From32(0, 2)), 0);
-    ExpectIntEQ(Dtls13WriteAckMessage(ssl_c, ssl_c->dtls13Rtx.seenRecords,
-            ssl_c->dtls13Rtx.seenRecordsCount, &length), 0);
-
-    /* must zero the span reserved for the header to avoid read of uninited
-     * data.
-     */
-    XMEMSET(ssl_c->buffers.outputBuffer.buffer, 0,
-            5 /* DTLS13_UNIFIED_HEADER_SIZE */);
-    /* N * RecordNumber + 2 extra bytes for length */
-    ExpectIntEQ(length, sizeof(expected_output) + 2);
-    ExpectNotNull(mymemmem(ssl_c->buffers.outputBuffer.buffer,
-            ssl_c->buffers.outputBuffer.bufferSize, expected_output,
-            sizeof(expected_output)));
-
-    wolfSSL_free(ssl_c);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_s);
-#endif
-    return EXPECT_RESULT();
-}
-
-int test_dtls13_ack_overflow(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
-    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-    unsigned char readBuf[50];
-    word32 length = 0;
-    int i;
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-    ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    ExpectIntEQ(wolfSSL_read(ssl_s, readBuf, sizeof(readBuf)), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-
-    /* Edge case 1: one below limit - all inserts must succeed */
-    for (i = 0; i < DTLS13_ACK_MAX_RECORDS - 1; i++) {
-        ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 0),
-                                    w64From32(0, (word32)i)), 0);
-    }
-    ExpectIntEQ(ssl_c->dtls13Rtx.seenRecordsCount, DTLS13_ACK_MAX_RECORDS - 1);
-
-    /* Edge case 2: insert the last allowed record - must succeed */
-    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 0),
-                    w64From32(0, (word32)(DTLS13_ACK_MAX_RECORDS - 1))), 0);
-    ExpectIntEQ(ssl_c->dtls13Rtx.seenRecordsCount, DTLS13_ACK_MAX_RECORDS);
-
-    /* Writing a full-but-valid list must succeed */
-    ExpectIntEQ(Dtls13WriteAckMessage(ssl_c, ssl_c->dtls13Rtx.seenRecords,
-                    ssl_c->dtls13Rtx.seenRecordsCount, &length), 0);
-
-    /* Edge case 3: one over limit - must be silently dropped */
-    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 0),
-                    w64From32(0, (word32)DTLS13_ACK_MAX_RECORDS)), 0);
-    ExpectIntEQ(ssl_c->dtls13Rtx.seenRecordsCount, DTLS13_ACK_MAX_RECORDS);
-
-    /* Bypass the insert guard to force the list one element over the limit,
-     * then verify Dtls13WriteAckMessage errors out instead of overflowing */
-    ssl_c->dtls13Rtx.seenRecordsCount = 0;
-    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 1),
-                    w64From32(0, (word32)DTLS13_ACK_MAX_RECORDS)), 0);
-    ssl_c->dtls13Rtx.seenRecordsCount = (word16)(DTLS13_ACK_MAX_RECORDS + 1);
-    ExpectIntEQ(Dtls13WriteAckMessage(ssl_c, ssl_c->dtls13Rtx.seenRecords,
-                    ssl_c->dtls13Rtx.seenRecordsCount, &length), BUFFER_E);
-
-    wolfSSL_free(ssl_c);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_s);
-#endif
-    return EXPECT_RESULT();
-}
-
-int test_dtls13_ack_dup_write_counter(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13) \
-    && defined(HAVE_WRITE_DUP)
-    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
-    WOLFSSL *ssl_c2 = NULL;
-    struct test_memio_ctx test_ctx;
-    unsigned char readBuf[50];
-    int i;
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-    /* Drain any post-handshake messages */
-    ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    ExpectIntEQ(wolfSSL_read(ssl_s, readBuf, sizeof(readBuf)), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-
-    /* Split ssl_c: ssl_c becomes READ_DUP_SIDE, ssl_c2 becomes WRITE_DUP_SIDE */
-    ExpectNotNull(ssl_c2 = wolfSSL_write_dup(ssl_c));
-
-    /* Cycle 1: add records, trigger handoff, verify counter is reset to 0 */
-    for (i = 0; i < 5; i++)
-        ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 3),
-                                    w64From32(0, (word32)i)), 0);
-    ExpectIntEQ(ssl_c->dtls13Rtx.seenRecordsCount, 5);
-    ssl_c->dtls13Rtx.sendAcks = 1;
-    ExpectIntEQ(Dtls13DoScheduledWork(ssl_c), 0);
-    /* seenRecords ownership was transferred to dupWrite->sendAckList;
-     * seenRecordsCount must be reset to 0,  not left at 5. */
-    ExpectNull(ssl_c->dtls13Rtx.seenRecords);
-    ExpectIntEQ(ssl_c->dtls13Rtx.seenRecordsCount, 0);
-
-    /* Cycle 2 (different epoch to avoid the dup-filter): verify the counter
-     * did not accumulate across the previous transfer.  Without the fix,
-     * seenRecordsCount would now be 10 after this second batch. */
-    for (i = 0; i < 5; i++)
-        ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 4),
-                                    w64From32(0, (word32)i)), 0);
-    ExpectIntEQ(ssl_c->dtls13Rtx.seenRecordsCount, 5);
-    ssl_c->dtls13Rtx.sendAcks = 1;
-    ExpectIntEQ(Dtls13DoScheduledWork(ssl_c), 0);
-    ExpectNull(ssl_c->dtls13Rtx.seenRecords);
-    ExpectIntEQ(ssl_c->dtls13Rtx.seenRecordsCount, 0);
-
-    wolfSSL_free(ssl_c);
-    wolfSSL_free(ssl_c2);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_s);
-#endif
-    return EXPECT_RESULT();
-}
 
 int test_dtls_version_checking(void)
 {
@@ -1836,225 +1301,8 @@ int test_dtls_rtx_across_epoch_change(void)
     return EXPECT_RESULT();
 }
 
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) &&                           \
-    defined(WOLFSSL_DTLS13) && defined(WOLFSSL_DTLS)
-static int test_dtls13_get_message_seq(const char* msg, int msgSz,
-    word16* msgSeq)
-{
-    int hsOff = DTLS_RECORD_HEADER_SZ;
 
-    if (msg == NULL || msgSeq == NULL ||
-            msgSz < DTLS_RECORD_HEADER_SZ + DTLS_HANDSHAKE_HEADER_SZ) {
-        return BAD_FUNC_ARG;
-    }
 
-    *msgSeq = ((word16)(byte)msg[hsOff + 4] << 8) |
-              (word16)(byte)msg[hsOff + 5];
-
-    return WOLFSSL_SUCCESS;
-}
-#endif
-
-int test_dtls13_ch2_rtx_no_ch1(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) &&                           \
-    defined(WOLFSSL_DTLS13) && defined(WOLFSSL_DTLS)
-    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-    const char* msg = NULL;
-    int msgSz = 0;
-    word16 ch1Seq = 0;
-    int i;
-    int foundCh1Seq = 0;
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-                    wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method),
-        0);
-
-    /* To force HRR */
-    ExpectIntEQ(wolfSSL_NoKeyShares(ssl_c), WOLFSSL_SUCCESS);
-
-    /* CH1 */
-    ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    ExpectIntEQ(test_memio_get_message(&test_ctx, 0, &msg, &msgSz, 0), 0);
-    ExpectIntGE(msgSz, DTLS_RECORD_HEADER_SZ + DTLS_HANDSHAKE_HEADER_SZ);
-    ExpectIntEQ(test_dtls13_get_message_seq(msg, msgSz, &ch1Seq),
-        WOLFSSL_SUCCESS);
-
-    /* HRR */
-    ExpectIntEQ(wolfSSL_accept(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    ExpectIntGT(test_ctx.c_msg_count, 0);
-
-    /* CH2 */
-    ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-    ExpectIntGT(test_ctx.s_msg_count, 0);
-
-    /* Drop CH2 and trigger the client retransmission timeout. */
-    test_memio_clear_buffer(&test_ctx, 0);
-    if (wolfSSL_dtls13_use_quick_timeout(ssl_c))
-        ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
-    ExpectIntGT(test_ctx.s_msg_count, 0);
-
-    for (i = 0; i < test_ctx.s_msg_count && EXPECT_SUCCESS(); i++) {
-        int hsOff = DTLS_RECORD_HEADER_SZ;
-        word16 msgSeq = 0;
-
-        ExpectIntEQ(test_memio_get_message(&test_ctx, 0, &msg, &msgSz, i), 0);
-        /* memio stores one DTLS record per message in this handshake path. */
-        if (msgSz >= DTLS_RECORD_HEADER_SZ + DTLS_HANDSHAKE_HEADER_SZ &&
-                (byte)msg[0] == handshake && msg[hsOff] == client_hello) {
-            ExpectIntEQ(test_dtls13_get_message_seq(msg, msgSz, &msgSeq),
-                WOLFSSL_SUCCESS);
-            if (msgSeq == ch1Seq)
-                foundCh1Seq = 1;
-        }
-    }
-
-    ExpectIntEQ(foundCh1Seq, 0);
-
-    wolfSSL_free(ssl_c);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_s);
-#endif
-    return EXPECT_RESULT();
-}
-
-int test_dtls13_frag_ch2_with_ch1_rtx(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) &&                           \
-    defined(WOLFSSL_DTLS13) && defined(WOLFSSL_DTLS) &&                        \
-    defined(WOLFSSL_DTLS_MTU) && defined(WOLFSSL_DTLS_CH_FRAG) &&              \
-    (!defined(HAVE_FIPS) || FIPS_VERSION3_GE(7,0,0))
-    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-    char hrr[TEST_MEMIO_BUF_SZ];
-    int hrrSz = (int)sizeof(hrr);
-    char ch1Rtx[TEST_MEMIO_BUF_SZ];
-    int ch1RtxSz = (int)sizeof(ch1Rtx);
-    char ch2[TEST_MEMIO_BUF_SZ];
-    int ch2Sz = 0;
-    int ch2MsgCount = 0;
-    int ch2MsgSizes[TEST_MEMIO_MAX_MSGS] = {0};
-    /* The DTLS record sequence number occupies the last 8 bytes of the
-     * record header. */
-    int recordSeqOff = DTLS_RECORD_HEADER_SZ - 8;
-    int ch2Seq = 0;
-    int ch1RtxSeq = 0;
-    int off;
-    int i;
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-                    wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method),
-        0);
-
-    /* To force HRR */
-    ExpectIntEQ(wolfSSL_NoKeyShares(ssl_c), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_dtls13_allow_ch_frag(ssl_s, 1), WOLFSSL_SUCCESS);
-
-    /* CH1 */
-    ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-
-    /* HRR */
-    ExpectIntEQ(wolfSSL_accept(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    ExpectIntEQ(test_memio_copy_message(&test_ctx, 1, hrr, &hrrSz, 0), 0);
-
-    /* Drop HRR, trigger CH1 retransmission, copy and drop it */
-    test_memio_clear_buffer(&test_ctx, 1);
-    if (wolfSSL_dtls13_use_quick_timeout(ssl_c))
-        ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
-    ExpectIntEQ(test_memio_copy_message(&test_ctx, 0, ch1Rtx, &ch1RtxSz, 0), 0);
-    test_memio_clear_buffer(&test_ctx, 0);
-
-    /* Force CH2 fragmentation. MTU must be small enough to fragment but large
-     * enough that the cookie extension lands in the first fragment, otherwise
-     * the server can't validate it statelessly and the test scenario (server
-     * stateful after frag 1) does not hold. With --enable-all (PQ groups in
-     * supported_groups), the cookie extension can sit ~400 bytes into CH2; 600
-     * gives margin while still producing multiple fragments (CH2 is ~2KB). */
-    ExpectIntEQ(wolfSSL_dtls_set_mtu(ssl_c, 600), WOLFSSL_SUCCESS);
-
-    /* Forward HRR and let the client create fragmented CH2 */
-    ExpectIntEQ(test_memio_inject_message(&test_ctx, 1, hrr, hrrSz), 0);
-    ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-
-    ExpectIntGT(test_ctx.s_msg_count, 1);
-    ExpectIntLE(test_ctx.s_msg_count, TEST_MEMIO_MAX_MSGS);
-    ExpectIntLE(test_ctx.s_len, (int)sizeof(ch2));
-    if (EXPECT_SUCCESS()) {
-        ch2Sz = test_ctx.s_len;
-        ch2MsgCount = test_ctx.s_msg_count;
-        XMEMCPY(ch2, test_ctx.s_buff, ch2Sz);
-        XMEMCPY(ch2MsgSizes, test_ctx.s_msg_sizes,
-            sizeof(ch2MsgSizes[0]) * (size_t)ch2MsgCount);
-
-        ch2Seq = ((byte)ch2[recordSeqOff + 4] << 8) |
-                 (byte)ch2[recordSeqOff + 5];
-        ch1RtxSeq = ch2Seq + ch2MsgCount;
-
-        /* Synthesize a CH1 retransmission that can pass the replay window after
-         * the first CH2 fragment makes the server stateful. The handshake
-         * message_seq remains the original CH1 value; only the DTLS record
-         * sequence is moved past the fragmented CH2 flight */
-        ch1Rtx[recordSeqOff + 0] = 0;
-        ch1Rtx[recordSeqOff + 1] = 0;
-        ch1Rtx[recordSeqOff + 2] = 0;
-        ch1Rtx[recordSeqOff + 3] = 0;
-        ch1Rtx[recordSeqOff + 4] = (byte)(ch1RtxSeq >> 8);
-        ch1Rtx[recordSeqOff + 5] = (byte)ch1RtxSeq;
-    }
-
-    test_memio_clear_buffer(&test_ctx, 0);
-
-    /* Deliver CH2 first fragment only. Now server is stateful */
-    ExpectIntEQ(test_memio_inject_message(&test_ctx, 0, ch2, ch2MsgSizes[0]), 0);
-    ExpectIntEQ(wolfSSL_accept(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-
-    /* Deliver the retransmitted CH1 between CH2 fragments, it should be
-     * discarded as rtx */
-    ExpectIntEQ(test_memio_inject_message(&test_ctx, 0, ch1Rtx, ch1RtxSz), 0);
-    ExpectIntEQ(wolfSSL_accept(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-    test_memio_clear_buffer(&test_ctx, 1);
-
-    /* Deliver the rest of CH2 */
-    off = ch2MsgSizes[0];
-    for (i = 1; i < ch2MsgCount && EXPECT_SUCCESS(); i++) {
-        ExpectIntEQ(test_memio_inject_message(&test_ctx, 0, ch2 + off,
-            ch2MsgSizes[i]), 0);
-        off += ch2MsgSizes[i];
-    }
-
-    /* Restore MTU so the client's input buffer can hold the full server
-     * flight (e.g. an SH carrying a hybrid PQC key share). */
-    ExpectIntEQ(wolfSSL_dtls_set_mtu(ssl_c, 1500), WOLFSSL_SUCCESS);
-
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-
-    wolfSSL_free(ssl_c);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_s);
-#endif
-    return EXPECT_RESULT();
-}
 
 int test_dtls_drop_client_ack(void)
 {
@@ -2262,43 +1510,6 @@ int test_dtls_replay(void)
     return EXPECT_RESULT();
 }
 
-#if defined(WOLFSSL_DTLS13) && defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES) && \
-    defined(WOLFSSL_SRTP)
-static int test_dtls_srtp_ctx_ready(WOLFSSL_CTX* ctx)
-{
-    EXPECT_DECLS;
-    ExpectIntEQ(wolfSSL_CTX_set_tlsext_use_srtp(ctx, "SRTP_AEAD_AES_256_GCM:"
-         "SRTP_AEAD_AES_128_GCM:SRTP_AES128_CM_SHA1_80:SRTP_AES128_CM_SHA1_32"),
-          0);
-    return EXPECT_RESULT();
-}
-
-int test_dtls_srtp(void)
-{
-    EXPECT_DECLS;
-    test_ssl_cbf client_cbf;
-    test_ssl_cbf server_cbf;
-
-    XMEMSET(&client_cbf, 0, sizeof(client_cbf));
-    XMEMSET(&server_cbf, 0, sizeof(server_cbf));
-
-    client_cbf.method = wolfDTLSv1_3_client_method;
-    client_cbf.ctx_ready = test_dtls_srtp_ctx_ready;
-    server_cbf.method = wolfDTLSv1_3_server_method;
-    server_cbf.ctx_ready = test_dtls_srtp_ctx_ready;
-
-    ExpectIntEQ(test_wolfSSL_client_server_nofail_memio(&client_cbf,
-        &server_cbf, NULL), TEST_SUCCESS);
-
-    return EXPECT_RESULT();
-}
-#else
-int test_dtls_srtp(void)
-{
-    EXPECT_DECLS;
-    return EXPECT_RESULT();
-}
-#endif
 
 int test_dtls_timeout(void)
 {
@@ -2876,287 +2087,17 @@ int test_dtls_mtu_split_messages(void)
  * This tests relies on timing of the retransmission logic so it may be
  * flaky on very slow systems.
  */
-int test_dtls13_min_rtx_interval(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && \
-    defined(WOLFSSL_DTLS13) && !defined(DTLS13_MIN_RTX_INTERVAL) && \
-    !defined(NO_ASN_TIME)
-    /* We don't want to test when DTLS13_MIN_RTX_INTERVAL is defined because
-     * it may be too low to trigger reliably in a test. The default value is
-     * 1 second which is sufficient for testing here. */
-    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
-    struct test_memio_ctx test_ctx;
-    int c_msg_count = 0;
-
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-
-    /* Setup DTLS 1.3 contexts */
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-                    wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
-
-    /* CH0 */
-    ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), SSL_ERROR_WANT_READ);
-
-    /* HRR */
-    ExpectIntEQ(wolfSSL_accept(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), SSL_ERROR_WANT_READ);
-
-    /* CH1 */
-    ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), SSL_ERROR_WANT_READ);
-
-    /* SH ... FINISHED */
-    ExpectIntEQ(wolfSSL_accept(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), SSL_ERROR_WANT_READ);
-
-    /* We should have SH ... FINISHED messages in the buffer */
-    ExpectIntGE(test_ctx.c_msg_count, 2);
-
-    /* Drop everything */
-    test_memio_clear_buffer(&test_ctx, 1);
-
-    /* First timeout. This one should trigger a retransmission */
-    if (wolfSSL_dtls13_use_quick_timeout(ssl_s))
-        ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_s), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_s), WOLFSSL_SUCCESS);
-    /* Save the message count to make sure no new messages are sent */
-    ExpectIntGE(test_ctx.c_msg_count, 2);
-    c_msg_count = test_ctx.c_msg_count;
-
-    /* Second timeout. This one should not trigger a retransmission */
-    if (wolfSSL_dtls13_use_quick_timeout(ssl_s))
-        ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_s), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_s), WOLFSSL_SUCCESS);
-    /* This is the critical check. The message count should not increase
-     * after the second timeout. DTLS13_MIN_RTX_INTERVAL should have blocked
-     * retransmission here. */
-    ExpectIntEQ(c_msg_count, test_ctx.c_msg_count);
-
-    /* Now complete the handshake. We didn't clear the first retransmission
-     * so the handshake should proceed without issues. */
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-
-    /* Cleanup */
-    wolfSSL_free(ssl_c);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_s);
-#endif
-    return EXPECT_RESULT();
-}
 
 /* RFC 9147 Section 5.3: DTLS 1.3 ServerHello must have empty
  * legacy_session_id_echo, even if the ClientHello had a non-empty
  * legacy_session_id. */
-int test_dtls13_no_session_id_echo(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13) && \
-    defined(HAVE_SESSION_TICKET) && defined(HAVE_ECC) && \
-    !defined(WOLFSSL_DTLS13_ECHO_LEGACY_SESSION_ID)
-    struct test_memio_ctx test_ctx;
-    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
-    WOLFSSL_SESSION *sess = NULL;
-    char readBuf[1];
-    /* Pin to SECP256R1 to avoid a PQ-induced key-share HRR */
-    int groups[] = { WOLFSSL_ECC_SECP256R1 };
-
-    /* First connection: complete a DTLS 1.3 handshake to get a session */
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
-    ExpectIntEQ(wolfSSL_set_groups(ssl_c, groups, 1), WOLFSSL_SUCCESS);
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-
-    /* Read to process any NewSessionTicket */
-    ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-
-    ExpectNotNull(sess = wolfSSL_get1_session(ssl_c));
-
-    /* Ensure the session has a non-empty session ID so the ClientHello
-     * will have a populated legacy_session_id field (which is legal per
-     * RFC 9147). */
-    if (sess != NULL && sess->sessionIDSz == 0) {
-        sess->sessionIDSz = ID_LEN;
-        XMEMSET(sess->sessionID, 0x42, ID_LEN);
-    }
-
-    wolfSSL_free(ssl_c); ssl_c = NULL;
-    wolfSSL_free(ssl_s); ssl_s = NULL;
-    wolfSSL_CTX_free(ctx_c); ctx_c = NULL;
-    wolfSSL_CTX_free(ctx_s); ctx_s = NULL;
-
-    /* Second connection: set the session on the client so the ClientHello
-     * contains a non-empty legacy_session_id. Verify the server does NOT
-     * echo it in the ServerHello. */
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
-    ExpectIntEQ(wolfSSL_set_session(ssl_c, sess), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_set_groups(ssl_c, groups, 1), WOLFSSL_SUCCESS);
-    /* Disable HRR cookie so the server directly sends a ServerHello */
-    ExpectIntEQ(wolfSSL_disable_hrr_cookie(ssl_s), WOLFSSL_SUCCESS);
-
-    /* Client sends ClientHello (with non-empty legacy_session_id) */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-
-    /* Server processes ClientHello and sends ServerHello + flight */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-
-    /* Verify the ServerHello on the wire.
-     * Layout: DTLS Record Header (13) + DTLS Handshake Header (12) +
-     *         ProtocolVersion (2) + Random (32) = offset 59 for
-     *         legacy_session_id_echo length byte. */
-    ExpectIntGE(test_ctx.c_len, 60);
-    ExpectIntEQ(test_ctx.c_buff[0], handshake);
-    ExpectIntEQ(test_ctx.c_buff[DTLS_RECORD_HEADER_SZ], server_hello);
-    ExpectIntEQ(test_ctx.c_buff[DTLS_RECORD_HEADER_SZ +
-        DTLS_HANDSHAKE_HEADER_SZ + OPAQUE16_LEN + RAN_LEN], 0);
-
-    /* Complete the handshake */
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-
-    wolfSSL_SESSION_free(sess);
-    wolfSSL_free(ssl_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_CTX_free(ctx_s);
-#endif
-    return EXPECT_RESULT();
-}
 
 /* Test that a server built with WOLFSSL_DTLS13_ECHO_LEGACY_SESSION_ID echoes the
  * client's legacy_session_id in both the direct ServerHello path and the
  * stateless HRR path (which also exercises RestartHandshakeHashWithCookie). */
-int test_dtls13_5_9_0_compat(void)
-{
-    EXPECT_DECLS;
-#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13) && \
-    defined(HAVE_SESSION_TICKET) && defined(WOLFSSL_DTLS13_ECHO_LEGACY_SESSION_ID) && \
-    defined(HAVE_ECC)
-    struct test_memio_ctx test_ctx;
-    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
-    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
-    WOLFSSL_SESSION *sess = NULL;
-    char readBuf[1];
-    /* Pin to SECP256R1 to avoid a PQ-induced key-share HRR */
-    int groups[] = { WOLFSSL_ECC_SECP256R1 };
-    /* RFC 8446 Section 4.1.3: an HRR is a ServerHello carrying this magic
-     * random. Used to assert sub-test 1 is a real ServerHello, not an HRR. */
-    static const byte hrrRandom[RAN_LEN] = {
-        0xCF, 0x21, 0xAD, 0x74, 0xE5, 0x9A, 0x61, 0x11,
-        0xBE, 0x1D, 0x8C, 0x02, 0x1E, 0x65, 0xB8, 0x91,
-        0xC2, 0xA2, 0x11, 0x16, 0x7A, 0xBB, 0x8C, 0x5E,
-        0x07, 0x9E, 0x09, 0xE2, 0xC8, 0xA8, 0x33, 0x9C
-    };
 
-    /* --- initial connection: get a real session to carry the session ID --- */
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
-    ExpectIntEQ(wolfSSL_set_groups(ssl_c, groups, 1), WOLFSSL_SUCCESS);
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-
-    /* drain any NewSessionTicket before calling get1_session */
-    ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-
-    ExpectNotNull(sess = wolfSSL_get1_session(ssl_c));
-
-    /* Force a non-zero session ID - simulates a wolfSSL <=v5.9.0 client that
-     * mistakenly sends 32 bytes as legacy_session_id in DTLS 1.3. */
-    if (sess != NULL && sess->sessionIDSz == 0) {
-        sess->sessionIDSz = ID_LEN;
-        XMEMSET(sess->sessionID, 0x42, ID_LEN);
-    }
-
-    wolfSSL_free(ssl_c); ssl_c = NULL;
-    wolfSSL_free(ssl_s); ssl_s = NULL;
-    wolfSSL_CTX_free(ctx_c); ctx_c = NULL;
-    wolfSSL_CTX_free(ctx_s); ctx_s = NULL;
-
-    /* --- sub-test 1: direct ServerHello (HRR cookie disabled) ---
-     * Exercises DoTls13ClientHello (change 1) and
-     * SendTls13ServerHello (change 2). */
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
-    ExpectIntEQ(wolfSSL_set_session(ssl_c, sess), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_set_groups(ssl_c, groups, 1), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_disable_hrr_cookie(ssl_s), WOLFSSL_SUCCESS);
-
-    /* Client sends CH1 with non-empty legacy_session_id */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-
-    /* Server processes CH1 and sends ServerHello */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-
-    /* Verify that the ServerHello on the wire echoes the session ID.
-     * Layout: DTLS Record Header (13) + DTLS Handshake Header (12) +
-     *         ProtocolVersion (2) + Random (32) = byte 59 for
-     *         legacy_session_id_echo length. */
-    ExpectIntGE(test_ctx.c_len, 60);
-    ExpectIntEQ(test_ctx.c_buff[0], handshake);
-    ExpectIntEQ(test_ctx.c_buff[DTLS_RECORD_HEADER_SZ], server_hello);
-    /* Confirm it is a real ServerHello, not an HRR (also encoded as a
-     * ServerHello but bearing the HelloRetryRequest magic random). */
-    ExpectIntNE(XMEMCMP(&test_ctx.c_buff[DTLS_RECORD_HEADER_SZ +
-        DTLS_HANDSHAKE_HEADER_SZ + OPAQUE16_LEN], hrrRandom, RAN_LEN), 0);
-    ExpectIntEQ(test_ctx.c_buff[DTLS_RECORD_HEADER_SZ +
-        DTLS_HANDSHAKE_HEADER_SZ + OPAQUE16_LEN + RAN_LEN], ID_LEN);
-
-    /* Complete the handshake - Finished MAC validates the transcript */
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-
-    wolfSSL_free(ssl_c); ssl_c = NULL;
-    wolfSSL_free(ssl_s); ssl_s = NULL;
-    wolfSSL_CTX_free(ctx_c); ctx_c = NULL;
-    wolfSSL_CTX_free(ctx_s); ctx_s = NULL;
-
-    /* --- sub-test 2: stateless HRR (HRR cookie enabled by default) ---
-     * Exercises SendStatelessReplyDtls13 (change 4) and
-     * RestartHandshakeHashWithCookie (change 3). */
-    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
-    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
-        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
-    ExpectIntEQ(wolfSSL_set_session(ssl_c, sess), WOLFSSL_SUCCESS);
-    ExpectIntEQ(wolfSSL_set_groups(ssl_c, groups, 1), WOLFSSL_SUCCESS);
-
-    /* Client sends CH1 */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
-
-    /* Server sends stateless HRR (SendStatelessReplyDtls13) */
-    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
-    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
-
-    /* Verify the HRR echoes the session ID at the same wire offset */
-    ExpectIntGE(test_ctx.c_len, 60);
-    ExpectIntEQ(test_ctx.c_buff[0], handshake);
-    ExpectIntEQ(test_ctx.c_buff[DTLS_RECORD_HEADER_SZ], server_hello);
-    ExpectIntEQ(test_ctx.c_buff[DTLS_RECORD_HEADER_SZ +
-        DTLS_HANDSHAKE_HEADER_SZ + OPAQUE16_LEN + RAN_LEN], ID_LEN);
-
-    /* Complete the handshake - Finished MAC validates RestartHandshakeHashWithCookie */
-    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
-
-    wolfSSL_SESSION_free(sess);
-    wolfSSL_free(ssl_c);
-    wolfSSL_free(ssl_s);
-    wolfSSL_CTX_free(ctx_c);
-    wolfSSL_CTX_free(ctx_s);
-#endif
-    return EXPECT_RESULT();
-}
+/* Test that a DTLS 1.3 handshake with an oversized certificate chain does
+ * not crash or cause out-of-bounds access in SendTls13Certificate. */
 
 /* DTLS counterpart to test_tls_set_session_min_downgrade. Exercises the
  * inverted DTLS minor-version comparison (DTLS 1.2 minor 0xFD is "below"
@@ -3199,3 +2140,3301 @@ int test_dtls_set_session_min_downgrade(void)
 #endif
     return EXPECT_RESULT();
 }
+
+/*----------------------------------------------------------------------------*/
+/* DTLS session export / import                                               */
+/*----------------------------------------------------------------------------*/
+
+#if defined(WOLFSSL_DTLS) && defined(WOLFSSL_SESSION_EXPORT)
+/* canned export of a session using older version 3 */
+static unsigned char version_3[] = {
+    0xA5, 0xA3, 0x01, 0x88, 0x00, 0x3c, 0x00, 0x01,
+    0x00, 0x00, 0x00, 0x80, 0x0C, 0x00, 0x00, 0x00,
+    0x00, 0x80, 0x00, 0x1C, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x01, 0x01, 0x01, 0x01, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00,
+    0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xC0, 0x30,
+    0x05, 0x09, 0x0A, 0x01, 0x01, 0x00, 0x0D, 0x05,
+    0xFE, 0xFD, 0x01, 0x25, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x01, 0x00, 0x01, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x06, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x06, 0x00, 0x01, 0x00, 0x07, 0x00, 0x00,
+    0x00, 0x30, 0x00, 0x00, 0x00, 0x10, 0x01, 0x01,
+    0x00, 0x02, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x3F,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x30, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x20, 0x05,
+    0x12, 0xCF, 0x22, 0xA1, 0x9F, 0x1C, 0x39, 0x1D,
+    0x31, 0x11, 0x12, 0x1D, 0x11, 0x18, 0x0D, 0x0B,
+    0xF3, 0xE1, 0x4D, 0xDC, 0xB1, 0xF1, 0x39, 0x98,
+    0x91, 0x6C, 0x48, 0xE5, 0xED, 0x11, 0x12, 0xA0,
+    0x00, 0xF2, 0x25, 0x4C, 0x09, 0x26, 0xD1, 0x74,
+    0xDF, 0x23, 0x40, 0x15, 0x6A, 0x42, 0x2A, 0x26,
+    0xA5, 0xAC, 0x56, 0xD5, 0x4A, 0x20, 0xB7, 0xE9,
+    0xEF, 0xEB, 0xAF, 0xA8, 0x1E, 0x23, 0x7C, 0x04,
+    0xAA, 0xA1, 0x6D, 0x92, 0x79, 0x7B, 0xFA, 0x80,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01,
+    0x0C, 0x79, 0x7B, 0xFA, 0x80, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x00, 0x00, 0x00, 0xAA, 0xA1, 0x6D,
+    0x92, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+    0x00, 0x00, 0x10, 0x00, 0x20, 0x00, 0x04, 0x00,
+    0x10, 0x00, 0x10, 0x08, 0x02, 0x05, 0x08, 0x01,
+    0x30, 0x28, 0x00, 0x00, 0x0F, 0x00, 0x02, 0x00,
+    0x09, 0x31, 0x32, 0x37, 0x2E, 0x30, 0x2E, 0x30,
+    0x2E, 0x31, 0xED, 0x4F
+};
+#endif /* defined(WOLFSSL_DTLS) && defined(WOLFSSL_SESSION_EXPORT) */
+
+int test_wolfSSL_dtls_export(void)
+{
+    EXPECT_DECLS;
+#if defined(WOLFSSL_DTLS) && defined(WOLFSSL_SESSION_EXPORT)
+    tcp_ready ready;
+    func_args client_args;
+    func_args server_args;
+    THREAD_TYPE serverThread;
+    callback_functions server_cbf;
+    callback_functions client_cbf;
+#ifdef WOLFSSL_TIRTOS
+    fdOpenSession(Task_self());
+#endif
+
+    InitTcpReady(&ready);
+
+    /* set using dtls */
+    XMEMSET(&client_args, 0, sizeof(func_args));
+    XMEMSET(&server_args, 0, sizeof(func_args));
+    XMEMSET(&server_cbf, 0, sizeof(callback_functions));
+    XMEMSET(&client_cbf, 0, sizeof(callback_functions));
+    server_cbf.method = wolfDTLSv1_2_server_method;
+    client_cbf.method = wolfDTLSv1_2_client_method;
+    server_args.callbacks = &server_cbf;
+    client_args.callbacks = &client_cbf;
+
+    server_args.signal = &ready;
+    client_args.signal = &ready;
+
+    start_thread(run_wolfssl_server, &server_args, &serverThread);
+    wait_tcp_ready(&server_args);
+    run_wolfssl_client(&client_args);
+    join_thread(serverThread);
+
+    ExpectTrue(client_args.return_code);
+    ExpectTrue(server_args.return_code);
+
+    FreeTcpReady(&ready);
+
+#ifdef WOLFSSL_TIRTOS
+    fdOpenSession(Task_self());
+#endif
+
+    if (EXPECT_SUCCESS()) {
+        SOCKET_T sockfd = 0;
+        WOLFSSL_CTX* ctx = NULL;
+        WOLFSSL*     ssl = NULL;
+        char msg[64] = "hello wolfssl!";
+        char reply[1024];
+        int  msgSz = (int)XSTRLEN(msg);
+        byte *session, *window;
+        unsigned int sessionSz = 0;
+        unsigned int windowSz = 0;
+
+#ifndef TEST_IPV6
+        struct sockaddr_in peerAddr;
+#else
+        struct sockaddr_in6 peerAddr;
+#endif /* TEST_IPV6 */
+
+        int i;
+
+
+        /* Set ctx to DTLS 1.2 */
+        ExpectNotNull(ctx = wolfSSL_CTX_new(wolfDTLSv1_2_server_method()));
+        ExpectNotNull(ssl = wolfSSL_new(ctx));
+
+        /* test importing version 3 */
+        ExpectIntGE(wolfSSL_dtls_import(ssl, version_3, sizeof(version_3)), 0);
+
+        /* test importing bad length and bad version */
+        version_3[2]++;
+        ExpectIntLT(wolfSSL_dtls_import(ssl, version_3, sizeof(version_3)), 0);
+        version_3[2]--; version_3[1] = 0XA0;
+        ExpectIntLT(wolfSSL_dtls_import(ssl, version_3, sizeof(version_3)), 0);
+        wolfSSL_free(ssl);
+        wolfSSL_CTX_free(ctx);
+
+
+    /* check storing client state after connection and storing window only */
+#ifdef WOLFSSL_TIRTOS
+    fdOpenSession(Task_self());
+#endif
+
+    InitTcpReady(&ready);
+
+    /* set using dtls */
+    XMEMSET(&server_args, 0, sizeof(func_args));
+    XMEMSET(&server_cbf, 0, sizeof(callback_functions));
+    server_cbf.method = wolfDTLSv1_2_server_method;
+    server_cbf.doUdp = 1;
+    server_args.callbacks = &server_cbf;
+    server_args.argc = 3; /* set loop_count to 3 */
+
+
+    server_args.signal = &ready;
+    start_thread(test_server_nofail, &server_args, &serverThread);
+    wait_tcp_ready(&server_args);
+
+    /* create and connect with client */
+    ExpectNotNull(ctx = wolfSSL_CTX_new(wolfDTLSv1_2_client_method()));
+    ExpectIntEQ(WOLFSSL_SUCCESS,
+            wolfSSL_CTX_load_verify_locations(ctx, caCertFile, 0));
+    ExpectIntEQ(WOLFSSL_SUCCESS,
+          wolfSSL_CTX_use_certificate_file(ctx, cliCertFile, SSL_FILETYPE_PEM));
+    ExpectIntEQ(WOLFSSL_SUCCESS,
+            wolfSSL_CTX_use_PrivateKey_file(ctx, cliKeyFile, SSL_FILETYPE_PEM));
+    tcp_connect(&sockfd, wolfSSLIP, server_args.signal->port, 1, 0, NULL);
+    ExpectNotNull(ssl = wolfSSL_new(ctx));
+    ExpectIntEQ(wolfSSL_set_fd(ssl, sockfd), WOLFSSL_SUCCESS);
+
+    /* store server information connected too */
+    XMEMSET(&peerAddr, 0, sizeof(peerAddr));
+#ifndef TEST_IPV6
+    peerAddr.sin_family = AF_INET;
+    ExpectIntEQ(XINET_PTON(AF_INET, wolfSSLIP, &peerAddr.sin_addr),1);
+    peerAddr.sin_port = XHTONS(server_args.signal->port);
+#else
+    peerAddr.sin6_family = AF_INET6;
+    ExpectIntEQ(
+        XINET_PTON(AF_INET6, wolfSSLIP, &peerAddr.sin6_addr),1);
+    peerAddr.sin6_port = XHTONS(server_args.signal->port);
+#endif
+
+    ExpectIntEQ(wolfSSL_dtls_set_peer(ssl, &peerAddr, sizeof(peerAddr)),
+                    WOLFSSL_SUCCESS);
+
+    ExpectIntEQ(wolfSSL_connect(ssl), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_dtls_export(ssl, NULL, &sessionSz), 0);
+    session = (byte*)XMALLOC(sessionSz, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
+    ExpectIntGT(wolfSSL_dtls_export(ssl, session, &sessionSz), 0);
+    ExpectIntEQ(wolfSSL_write(ssl, msg, msgSz), msgSz);
+    ExpectIntGT(wolfSSL_read(ssl, reply, sizeof(reply)), 0);
+    ExpectIntEQ(wolfSSL_dtls_export_state_only(ssl, NULL, &windowSz), 0);
+    window = (byte*)XMALLOC(windowSz, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
+    ExpectIntGT(wolfSSL_dtls_export_state_only(ssl, window, &windowSz), 0);
+    wolfSSL_free(ssl);
+
+    for (i = 1; EXPECT_SUCCESS() && i < server_args.argc; i++) {
+        /* restore state */
+        ExpectNotNull(ssl = wolfSSL_new(ctx));
+        ExpectIntGT(wolfSSL_dtls_import(ssl, session, sessionSz), 0);
+        ExpectIntGT(wolfSSL_dtls_import(ssl, window, windowSz), 0);
+        ExpectIntEQ(wolfSSL_set_fd(ssl, sockfd), WOLFSSL_SUCCESS);
+        ExpectIntEQ(wolfSSL_dtls_set_peer(ssl, &peerAddr, sizeof(peerAddr)),
+                    WOLFSSL_SUCCESS);
+        ExpectIntEQ(wolfSSL_write(ssl, msg, msgSz), msgSz);
+        ExpectIntGE(wolfSSL_read(ssl, reply, sizeof(reply)), 0);
+        ExpectIntGT(wolfSSL_dtls_export_state_only(ssl, window, &windowSz), 0);
+        wolfSSL_free(ssl);
+    }
+    XFREE(session, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
+    XFREE(window, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
+    wolfSSL_CTX_free(ctx);
+
+    fprintf(stderr, "done and waiting for server\n");
+    join_thread(serverThread);
+    ExpectIntEQ(server_args.return_code, TEST_SUCCESS);
+
+    FreeTcpReady(&ready);
+
+#ifdef WOLFSSL_TIRTOS
+    fdOpenSession(Task_self());
+#endif
+    }
+#endif
+
+    return EXPECT_RESULT();
+}
+
+#if defined(WOLFSSL_DTLS) && defined(WOLFSSL_SESSION_EXPORT) && \
+    defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES)
+/* Dummy peer functions to satisfy the exporter/importer */
+static int test_wolfSSL_dtls_export_peers_get_peer(WOLFSSL* ssl, char* ip,
+        int* ipSz, unsigned short* port, int* fam)
+{
+    (void)ssl;
+    ip[0] = -1;
+    *ipSz = 1;
+    *port = 1;
+    *fam = 2;
+    return 1;
+}
+
+static int test_wolfSSL_dtls_export_peers_set_peer(WOLFSSL* ssl, char* ip,
+        int ipSz, unsigned short port, int fam)
+{
+    (void)ssl;
+    if (ip[0] != -1 || ipSz != 1 || port != 1 || fam != 2)
+        return 0;
+    return 1;
+}
+
+static int test_wolfSSL_dtls_export_peers_on_handshake(WOLFSSL_CTX **ctx,
+        WOLFSSL **ssl)
+{
+    EXPECT_DECLS;
+    unsigned char* sessionBuf = NULL;
+    unsigned int sessionSz = 0;
+    void* ioWriteCtx = wolfSSL_GetIOWriteCtx(*ssl);
+    void* ioReadCtx = wolfSSL_GetIOReadCtx(*ssl);
+
+    wolfSSL_CTX_SetIOGetPeer(*ctx, test_wolfSSL_dtls_export_peers_get_peer);
+    wolfSSL_CTX_SetIOSetPeer(*ctx, test_wolfSSL_dtls_export_peers_set_peer);
+    ExpectIntGE(wolfSSL_dtls_export(*ssl, NULL, &sessionSz), 0);
+    ExpectNotNull(sessionBuf =
+        (unsigned char*)XMALLOC(sessionSz, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER));
+    ExpectIntGE(wolfSSL_dtls_export(*ssl, sessionBuf, &sessionSz), 0);
+    wolfSSL_free(*ssl);
+    *ssl = NULL;
+    ExpectNotNull(*ssl = wolfSSL_new(*ctx));
+    ExpectIntGE(wolfSSL_dtls_import(*ssl, sessionBuf, sessionSz), 0);
+    wolfSSL_SetIOWriteCtx(*ssl, ioWriteCtx);
+    wolfSSL_SetIOReadCtx(*ssl, ioReadCtx);
+
+    XFREE(sessionBuf, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
+    return EXPECT_RESULT();
+}
+#endif
+
+int test_wolfSSL_dtls_export_peers(void)
+{
+    EXPECT_DECLS;
+#if defined(WOLFSSL_DTLS) && defined(WOLFSSL_SESSION_EXPORT) && \
+    defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES)
+    test_ssl_cbf client_cbf;
+    test_ssl_cbf server_cbf;
+    size_t i, j;
+    struct test_params {
+        method_provider client_meth;
+        method_provider server_meth;
+        const char* dtls_version;
+    } params[] = {
+#ifndef NO_OLD_TLS
+        {wolfDTLSv1_client_method, wolfDTLSv1_server_method, "1.0"},
+#endif
+        {wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method, "1.2"},
+        /* TODO DTLS 1.3 exporting not supported
+#ifdef WOLFSSL_DTLS13
+        {wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method, "1.3"},
+#endif
+         */
+    };
+
+    for (i = 0; i < sizeof(params)/sizeof(*params); i++) {
+        for (j = 0; j <= 3; j++) {
+            XMEMSET(&client_cbf, 0, sizeof(client_cbf));
+            XMEMSET(&server_cbf, 0, sizeof(server_cbf));
+
+            printf("\n\tTesting DTLS %s connection;", params[i].dtls_version);
+
+            client_cbf.method = params[i].client_meth;
+            server_cbf.method = params[i].server_meth;
+
+            if (j & 0x1) {
+                client_cbf.on_handshake =
+                        test_wolfSSL_dtls_export_peers_on_handshake;
+                printf(" With client export;");
+            }
+            if (j & 0x2) {
+                server_cbf.on_handshake =
+                        test_wolfSSL_dtls_export_peers_on_handshake;
+                printf(" With server export;");
+            }
+
+            printf("\n");
+
+            ExpectIntEQ(test_wolfSSL_client_server_nofail_memio(&client_cbf,
+                &server_cbf, NULL), TEST_SUCCESS);
+            if (!EXPECT_SUCCESS())
+                break;
+        }
+    }
+#endif
+    return EXPECT_RESULT();
+}
+
+/* Test that ImportKeyState correctly skips extra window words when importing
+ * state from a peer compiled with a larger WOLFSSL_DTLS_WINDOW_WORDS. */
+int test_wolfSSL_dtls_import_state_extra_window_words(void)
+{
+    EXPECT_DECLS;
+#if defined(WOLFSSL_DTLS) && defined(WOLFSSL_SESSION_EXPORT)
+    WOLFSSL_CTX* ctx = NULL;
+    WOLFSSL*     ssl = NULL;
+    unsigned int stateSz = 0;
+    byte*        state = NULL;
+    byte*        modified = NULL;
+    unsigned int modifiedSz;
+    word16       origKeyLen;
+    word16       origTotalLen;
+    /* Offset from start of key state data to the first wordCount field.
+     * Layout: 4 sequence numbers (16 bytes) + DTLS-specific fields (42 bytes) +
+     * encryptSz(4) + padSz(4) + encryptionOn(1) + decryptedCur(1) = 68 */
+    const int keyStateWindowOffset = 68;
+    /* Buffer header: 2 proto + 2 total_len + 2 key_len = 6 */
+    const int headerSz = 6;
+    int idx, modIdx;
+    int extraPerWindow = 2 * (int)sizeof(word32); /* 8 bytes extra per window */
+    int totalExtra = extraPerWindow * 2; /* 16 bytes extra total */
+
+    /* Create DTLS context and SSL object */
+    ExpectNotNull(ctx = wolfSSL_CTX_new(wolfDTLSv1_2_server_method()));
+    ExpectNotNull(ssl = wolfSSL_new(ctx));
+
+    /* Get required buffer size and export state-only */
+    ExpectIntEQ(wolfSSL_dtls_export_state_only(ssl, NULL, &stateSz), 0);
+    ExpectIntGT((int)stateSz, 0);
+    state = (byte*)XMALLOC(stateSz, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
+    ExpectNotNull(state);
+    ExpectIntGT(wolfSSL_dtls_export_state_only(ssl, state, &stateSz), 0);
+
+    /* Build a modified buffer that simulates a peer with
+     * WOLFSSL_DTLS_WINDOW_WORDS = WOLFSSL_DTLS_WINDOW_WORDS + 2.
+     * Each window section gets 2 extra word32 values (8 bytes).
+     * Two windows => 16 extra bytes total. */
+    modifiedSz = stateSz + totalExtra;
+    modified = (byte*)XMALLOC(modifiedSz, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
+    ExpectNotNull(modified);
+
+    if (EXPECT_SUCCESS()) {
+        int windowWords = WOLFSSL_DTLS_WINDOW_WORDS;
+        int windowDataSz = windowWords * (int)sizeof(word32);
+
+        XMEMSET(modified, 0, modifiedSz);
+
+        /* Copy protocol/version bytes (first 2 bytes) */
+        XMEMCPY(modified, state, 2);
+
+        /* Read original total length and key state length */
+        ato16(state + 2, &origTotalLen);
+        ato16(state + 4, &origKeyLen);
+
+        /* Write updated total length and key state length */
+        c16toa((word16)(origTotalLen + totalExtra), modified + 2);
+        c16toa((word16)(origKeyLen + totalExtra), modified + 4);
+
+        /* Copy key state data up to first window section */
+        idx = headerSz;
+        modIdx = headerSz;
+        XMEMCPY(modified + modIdx, state + idx, keyStateWindowOffset);
+        idx += keyStateWindowOffset;
+        modIdx += keyStateWindowOffset;
+
+        /* First window: write increased wordCount */
+        c16toa((word16)(windowWords + 2), modified + modIdx);
+        idx += OPAQUE16_LEN;
+        modIdx += OPAQUE16_LEN;
+
+        /* Copy original window data */
+        XMEMCPY(modified + modIdx, state + idx, windowDataSz);
+        idx += windowDataSz;
+        modIdx += windowDataSz;
+
+        /* Insert 2 extra word32 padding values */
+        XMEMSET(modified + modIdx, 0, extraPerWindow);
+        modIdx += extraPerWindow;
+
+        /* Second window (prevWindow): same transformation */
+        c16toa((word16)(windowWords + 2), modified + modIdx);
+        idx += OPAQUE16_LEN;
+        modIdx += OPAQUE16_LEN;
+
+        XMEMCPY(modified + modIdx, state + idx, windowDataSz);
+        idx += windowDataSz;
+        modIdx += windowDataSz;
+
+        XMEMSET(modified + modIdx, 0, extraPerWindow);
+        modIdx += extraPerWindow;
+
+        /* Copy remainder of key state (after both windows) */
+        XMEMCPY(modified + modIdx, state + idx, stateSz - idx);
+    }
+
+    /* Import the modified state - should succeed with the fix */
+    wolfSSL_free(ssl);
+    ssl = NULL;
+    ExpectNotNull(ssl = wolfSSL_new(ctx));
+    ExpectIntGT(wolfSSL_dtls_import(ssl, modified, modifiedSz), 0);
+
+    XFREE(state, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
+    XFREE(modified, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
+    wolfSSL_free(ssl);
+    wolfSSL_CTX_free(ctx);
+#endif
+    return EXPECT_RESULT();
+}
+
+/*----------------------------------------------------------------------------*/
+/* DTLS either-side method and cookie generation                              */
+/*----------------------------------------------------------------------------*/
+
+int test_wolfSSL_DTLS_either_side(void)
+{
+    EXPECT_DECLS;
+#if (defined(OPENSSL_EXTRA) || defined(WOLFSSL_EITHER_SIDE)) && \
+    defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS)
+    test_ssl_cbf client_cb;
+    test_ssl_cbf server_cb;
+
+    XMEMSET(&client_cb, 0, sizeof(client_cb));
+    XMEMSET(&server_cb, 0, sizeof(server_cb));
+
+    /* Use different CTX for client and server */
+    client_cb.ctx = wolfSSL_CTX_new(wolfDTLS_method());
+    ExpectNotNull(client_cb.ctx);
+    server_cb.ctx = wolfSSL_CTX_new(wolfDTLS_method());
+    ExpectNotNull(server_cb.ctx);
+    /* we are responsible for free'ing WOLFSSL_CTX */
+    server_cb.isSharedCtx = client_cb.isSharedCtx = 1;
+
+    ExpectIntEQ(test_wolfSSL_client_server_nofail_memio(&client_cb,
+        &server_cb, NULL), TEST_SUCCESS);
+
+    wolfSSL_CTX_free(client_cb.ctx);
+    wolfSSL_CTX_free(server_cb.ctx);
+#endif
+    return EXPECT_RESULT();
+}
+
+int test_generate_cookie(void)
+{
+    EXPECT_DECLS;
+#if defined(WOLFSSL_DTLS) && defined(OPENSSL_EXTRA) && defined(USE_WOLFSSL_IO)
+    SSL_CTX* ctx = NULL;
+    SSL* ssl = NULL;
+    byte    buf[FOURK_BUF] = {0};
+
+    ExpectNotNull(ctx = wolfSSL_CTX_new(wolfDTLS_method()));
+    ExpectNotNull(ssl = SSL_new(ctx));
+
+    /* Test unconnected */
+    ExpectIntEQ(EmbedGenerateCookie(ssl, buf, FOURK_BUF, NULL), WC_NO_ERR_TRACE(GEN_COOKIE_E));
+
+    wolfSSL_CTX_SetGenCookie(ctx, EmbedGenerateCookie);
+
+    wolfSSL_SetCookieCtx(ssl, ctx);
+
+    ExpectNotNull(wolfSSL_GetCookieCtx(ssl));
+
+    ExpectNull(wolfSSL_GetCookieCtx(NULL));
+
+    SSL_free(ssl);
+    SSL_CTX_free(ctx);
+#endif
+    return EXPECT_RESULT();
+}
+
+/*----------------------------------------------------------------------------*/
+/* DTLS handshake: MTU, plaintext, fragments, bad records, AEAD, stateless    */
+/*----------------------------------------------------------------------------*/
+
+int test_wolfSSL_dtls_set_mtu(void)
+{
+    EXPECT_DECLS;
+#if (defined(WOLFSSL_DTLS_MTU) || defined(WOLFSSL_SCTP)) && \
+    !defined(NO_WOLFSSL_SERVER) && defined(WOLFSSL_DTLS) && \
+    !defined(WOLFSSL_NO_TLS12)
+    WOLFSSL_CTX* ctx = NULL;
+    WOLFSSL*     ssl = NULL;
+    const char* testCertFile;
+    const char* testKeyFile;
+
+    ExpectNotNull(ctx = wolfSSL_CTX_new(wolfDTLSv1_2_server_method()));
+#ifndef NO_RSA
+        testCertFile = svrCertFile;
+        testKeyFile = svrKeyFile;
+#elif defined(HAVE_ECC)
+        testCertFile = eccCertFile;
+        testKeyFile = eccKeyFile;
+#endif
+    if  (testCertFile != NULL && testKeyFile != NULL) {
+        ExpectTrue(wolfSSL_CTX_use_certificate_file(ctx, testCertFile,
+                                                    WOLFSSL_FILETYPE_PEM));
+        ExpectTrue(wolfSSL_CTX_use_PrivateKey_file(ctx, testKeyFile,
+                                                   WOLFSSL_FILETYPE_PEM));
+    }
+    ExpectNotNull(ssl = wolfSSL_new(ctx));
+
+    ExpectIntEQ(wolfSSL_CTX_dtls_set_mtu(NULL, 1488), WC_NO_ERR_TRACE(BAD_FUNC_ARG));
+    ExpectIntEQ(wolfSSL_dtls_set_mtu(NULL, 1488), WC_NO_ERR_TRACE(BAD_FUNC_ARG));
+    ExpectIntEQ(wolfSSL_CTX_dtls_set_mtu(ctx, 20000), WC_NO_ERR_TRACE(BAD_FUNC_ARG));
+    ExpectIntEQ(wolfSSL_dtls_set_mtu(ssl, 20000), WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+    ExpectIntEQ(wolfSSL_get_error(ssl, WC_NO_ERR_TRACE(WOLFSSL_FAILURE)), WC_NO_ERR_TRACE(BAD_FUNC_ARG));
+    ExpectIntEQ(wolfSSL_CTX_dtls_set_mtu(ctx, 1488), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_dtls_set_mtu(ssl, 1488), WOLFSSL_SUCCESS);
+
+#ifdef OPENSSL_EXTRA
+    ExpectIntEQ(SSL_set_mtu(ssl, 1488), WOLFSSL_SUCCESS);
+#endif
+
+    wolfSSL_free(ssl);
+    wolfSSL_CTX_free(ctx);
+#endif
+
+    return EXPECT_RESULT();
+}
+
+#if defined(HAVE_IO_TESTS_DEPENDENCIES) && !defined(SINGLE_THREADED) && \
+    defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12)
+
+static WC_INLINE void generateDTLSMsg(byte* out, int outSz, word32 seq,
+        enum HandShakeType hsType, word16 length)
+{
+    size_t idx = 0;
+    byte* l;
+
+    /* record layer */
+    /* handshake type */
+    out[idx++] = handshake;
+    /* protocol version */
+    out[idx++] = 0xfe;
+    out[idx++] = 0xfd; /* DTLS 1.2 */
+    /* epoch 0 */
+    XMEMSET(out + idx, 0, 2);
+    idx += 2;
+    /* sequence number */
+    XMEMSET(out + idx, 0, 6);
+    c32toa(seq, out + idx + 2);
+    idx += 6;
+    /* length in BE */
+    if (length)
+        c16toa(length, out + idx);
+    else
+        c16toa(outSz - idx - 2, out + idx);
+    idx += 2;
+
+    /* handshake layer */
+    /* handshake type */
+    out[idx++] = (byte)hsType;
+    /* length */
+    l = out + idx;
+    idx += 3;
+    /* message seq */
+    c16toa(0, out + idx);
+    idx += 2;
+    /* frag offset */
+    c32to24(0, out + idx);
+    idx += 3;
+    /* frag length */
+    c32to24((word32)outSz - (word32)idx - 3, l);
+    c32to24((word32)outSz - (word32)idx - 3, out + idx);
+    idx += 3;
+    XMEMSET(out + idx, 0, outSz - idx);
+}
+
+static void test_wolfSSL_dtls_plaintext_server(WOLFSSL* ssl)
+{
+    byte msg[] = "This is a msg for the client";
+    byte reply[40];
+    AssertIntGT(wolfSSL_read(ssl, reply, sizeof(reply)),0);
+    reply[sizeof(reply) - 1] = '\0';
+    fprintf(stderr, "Client message: %s\n", reply);
+    AssertIntEQ(wolfSSL_write(ssl, msg, sizeof(msg)), sizeof(msg));
+}
+
+static void test_wolfSSL_dtls_plaintext_client(WOLFSSL* ssl)
+{
+    byte ch[50];
+    int fd = wolfSSL_get_wfd(ssl);
+    byte msg[] = "This is a msg for the server";
+    byte reply[40];
+
+    AssertIntGE(fd, 0);
+    generateDTLSMsg(ch, sizeof(ch), 20, client_hello, 0);
+    /* Server should ignore this datagram */
+    AssertIntEQ(send(fd, (MESSAGE_TYPE_CAST)ch, sizeof(ch), 0), sizeof(ch));
+    generateDTLSMsg(ch, sizeof(ch), 20, client_hello, 10000);
+    /* Server should ignore this datagram */
+    AssertIntEQ(send(fd, (MESSAGE_TYPE_CAST)ch, sizeof(ch), 0), sizeof(ch));
+
+    AssertIntEQ(wolfSSL_write(ssl, msg, sizeof(msg)), sizeof(msg));
+    AssertIntGT(wolfSSL_read(ssl, reply, sizeof(reply)),0);
+    reply[sizeof(reply) - 1] = '\0';
+    fprintf(stderr, "Server response: %s\n", reply);
+}
+
+int test_wolfSSL_dtls_plaintext(void)
+{
+    callback_functions func_cb_client;
+    callback_functions func_cb_server;
+    size_t i;
+    struct test_params {
+        method_provider client_meth;
+        method_provider server_meth;
+        ssl_callback on_result_server;
+        ssl_callback on_result_client;
+    } params[] = {
+        {wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method,
+                test_wolfSSL_dtls_plaintext_server,
+                test_wolfSSL_dtls_plaintext_client},
+    };
+
+    for (i = 0; i < sizeof(params)/sizeof(*params); i++) {
+        XMEMSET(&func_cb_client, 0, sizeof(callback_functions));
+        XMEMSET(&func_cb_server, 0, sizeof(callback_functions));
+
+        func_cb_client.doUdp = func_cb_server.doUdp = 1;
+        func_cb_server.method = params[i].server_meth;
+        func_cb_client.method = params[i].client_meth;
+        func_cb_client.on_result = params[i].on_result_client;
+        func_cb_server.on_result = params[i].on_result_server;
+
+        test_wolfSSL_client_server_nofail(&func_cb_client, &func_cb_server);
+
+        if (!func_cb_client.return_code)
+            return TEST_FAIL;
+        if (!func_cb_server.return_code)
+            return TEST_FAIL;
+    }
+
+    return TEST_RES_CHECK(1);
+}
+#else
+int test_wolfSSL_dtls_plaintext(void)
+{
+    return TEST_SKIPPED;
+}
+#endif
+
+#if defined(HAVE_IO_TESTS_DEPENDENCIES) && !defined(SINGLE_THREADED) && \
+    defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12)
+
+static void test_wolfSSL_dtls12_fragments_spammer(WOLFSSL* ssl)
+{
+    byte b[1100]; /* buffer for the messages to send */
+    size_t idx = 0;
+    size_t seq_offset = 0;
+    size_t msg_offset = 0;
+    int i;
+    int fd = wolfSSL_get_wfd(ssl);
+    int ret = wolfSSL_connect_cert(ssl); /* This gets us past the cookie */
+    word32 seq_number = 100; /* start high so server definitely reads this */
+    word16 msg_number = 50; /* start high so server has to buffer this */
+    AssertIntEQ(ret, 1);
+    /* Now let's start spamming the peer with fragments it needs to store */
+    XMEMSET(b, -1, sizeof(b));
+
+    /* record layer */
+    /* handshake type */
+    b[idx++] = 22;
+    /* protocol version */
+    b[idx++] = 0xfe;
+    b[idx++] = 0xfd; /* DTLS 1.2 */
+    /* epoch 0 */
+    XMEMSET(b + idx, 0, 2);
+    idx += 2;
+    /* sequence number */
+    XMEMSET(b + idx, 0, 6);
+    seq_offset = idx + 2; /* increment only the low 32 bits */
+    idx += 6;
+    /* static length in BE */
+    c16toa(42, b + idx);
+    idx += 2;
+
+    /* handshake layer */
+    /* cert type */
+    b[idx++] = 11;
+    /* length */
+    c32to24(1000, b + idx);
+    idx += 3;
+    /* message seq */
+    c16toa(0, b + idx);
+    msg_offset = idx;
+    idx += 2;
+    /* frag offset */
+    c32to24(500, b + idx);
+    idx += 3;
+    /* frag length */
+    c32to24(30, b + idx);
+    idx += 3;
+    (void)idx; /* inhibit clang-analyzer-deadcode.DeadStores */
+
+    for (i = 0; i < DTLS_POOL_SZ * 2 && ret > 0;
+            seq_number++, msg_number++, i++) {
+        struct timespec delay;
+        XMEMSET(&delay, 0, sizeof(delay));
+        delay.tv_nsec = 10000000; /* wait 0.01 seconds */
+        c32toa(seq_number, b + seq_offset);
+        c16toa(msg_number, b + msg_offset);
+        ret = (int)send(fd, (MESSAGE_TYPE_CAST)b, 55, 0);
+        nanosleep(&delay, NULL);
+    }
+}
+
+#ifdef WOLFSSL_DTLS13
+static void test_wolfSSL_dtls13_fragments_spammer(WOLFSSL* ssl)
+{
+    const word16 sendCountMax = 100;
+    byte b[150]; /* buffer for the messages to send */
+    size_t idx = 0;
+    size_t msg_offset = 0;
+    int fd = wolfSSL_get_wfd(ssl);
+    word16 msg_number = 10; /* start high so server has to buffer this */
+    int ret = wolfSSL_connect_cert(ssl); /* This gets us past the cookie */
+    AssertIntEQ(ret, 1);
+    /* Now let's start spamming the peer with fragments it needs to store */
+    XMEMSET(b, -1, sizeof(b));
+
+    /* handshake type */
+    b[idx++] = 11;
+    /* length */
+    c32to24(10000, b + idx);
+    idx += 3;
+    /* message_seq */
+    msg_offset = idx;
+    idx += 2;
+    /* fragment_offset */
+    c32to24(5000, b + idx);
+    idx += 3;
+    /* fragment_length */
+    c32to24(100, b + idx);
+    idx += 3;
+    /* fragment contents */
+    idx += 100;
+
+    for (; ret > 0 && msg_number < sendCountMax; msg_number++) {
+        byte sendBuf[150];
+        int sendSz = sizeof(sendBuf);
+        struct timespec delay;
+        XMEMSET(&delay, 0, sizeof(delay));
+        delay.tv_nsec = 10000000; /* wait 0.01 seconds */
+        c16toa(msg_number, b + msg_offset);
+        ret = sendSz = BuildTls13Message(ssl, sendBuf, sendSz, b,
+            (int)idx, handshake, 0, 0, 0);
+        if (sendSz > 0)
+            ret = (int)send(fd, (MESSAGE_TYPE_CAST)sendBuf, (size_t)sendSz, 0);
+        nanosleep(&delay, NULL);
+    }
+}
+#endif
+
+int test_wolfSSL_dtls_fragments(void)
+{
+    EXPECT_DECLS;
+    callback_functions func_cb_client;
+    callback_functions func_cb_server;
+    size_t i;
+    struct test_params {
+        method_provider client_meth;
+        method_provider server_meth;
+        ssl_callback spammer;
+    } params[] = {
+#if !defined(WOLFSSL_NO_TLS12)
+        {wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method,
+                test_wolfSSL_dtls12_fragments_spammer},
+#endif
+#ifdef WOLFSSL_DTLS13
+        {wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method,
+                test_wolfSSL_dtls13_fragments_spammer},
+#endif
+    };
+
+    for (i = 0; i < sizeof(params)/sizeof(*params); i++) {
+        XMEMSET(&func_cb_client, 0, sizeof(callback_functions));
+        XMEMSET(&func_cb_server, 0, sizeof(callback_functions));
+
+        func_cb_client.doUdp = func_cb_server.doUdp = 1;
+        func_cb_server.method = params[i].server_meth;
+        func_cb_client.method = params[i].client_meth;
+        func_cb_client.ssl_ready = params[i].spammer;
+
+        test_wolfSSL_client_server_nofail(&func_cb_client, &func_cb_server);
+
+        /* If the client failed, check that the error it encountered was from
+         * the server aborting, resulting in a socket error, fatal error or
+         * reading a close notify alert.
+         *
+         * Under slow execution (e.g. valgrind + noasm), the server may
+         * still be processing fragments when the client completes its
+         * handshake and write, so the client may succeed -- in that
+         * case return_code is TEST_SUCCESS and these checks don't apply.
+         */
+        if (func_cb_client.return_code == TEST_FAIL) {
+            if (func_cb_client.last_err != WC_NO_ERR_TRACE(SOCKET_ERROR_E) &&
+                    func_cb_client.last_err != WOLFSSL_ERROR_ZERO_RETURN &&
+                    func_cb_client.last_err != WC_NO_ERR_TRACE(FATAL_ERROR)) {
+                ExpectIntEQ(func_cb_client.last_err, WC_NO_ERR_TRACE(SOCKET_ERROR_E));
+            }
+        }
+        /* Check the server returned an error indicating the msg buffer
+         * was full.
+         *
+         * Under slow execution (e.g. valgrind + noasm), the real handshake
+         * from wolfSSL_negotiate() may complete before enough spam fragments
+         * accumulate to trigger DTLS_TOO_MANY_FRAGMENTS_E. Accept both
+         * outcomes: server hit the fragment limit, or completed normally.
+         */
+        if (func_cb_server.return_code == TEST_FAIL) {
+            ExpectIntEQ(func_cb_server.last_err, WC_NO_ERR_TRACE(DTLS_TOO_MANY_FRAGMENTS_E));
+        }
+
+        if (EXPECT_FAIL())
+            break;
+    }
+
+    return EXPECT_RESULT();
+}
+
+static void test_wolfSSL_dtls_send_alert(WOLFSSL* ssl)
+{
+    int fd, ret;
+    byte alert_msg[] = {
+        0x15, /* alert type */
+        0xfe, 0xfd, /* version */
+        0x00, 0x00, /* epoch */
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x01, /* seq number */
+        0x00, 0x02, /* length */
+        0x02, /* level: fatal */
+        0x46 /* protocol version */
+    };
+
+    fd = wolfSSL_get_wfd(ssl);
+    AssertIntGE(fd, 0);
+    ret = (int)send(fd, (MESSAGE_TYPE_CAST)alert_msg, sizeof(alert_msg), 0);
+    AssertIntGT(ret, 0);
+}
+
+static int _test_wolfSSL_ignore_alert_before_cookie(byte version12)
+{
+    callback_functions client_cbs, server_cbs;
+
+    XMEMSET(&client_cbs, 0, sizeof(client_cbs));
+    XMEMSET(&server_cbs, 0, sizeof(server_cbs));
+    client_cbs.doUdp = server_cbs.doUdp = 1;
+    if (version12) {
+#if !defined(WOLFSSL_NO_TLS12)
+        client_cbs.method = wolfDTLSv1_2_client_method;
+        server_cbs.method = wolfDTLSv1_2_server_method;
+#else
+        return TEST_SKIPPED;
+#endif
+    }
+    else
+    {
+#ifdef WOLFSSL_DTLS13
+        client_cbs.method = wolfDTLSv1_3_client_method;
+        server_cbs.method = wolfDTLSv1_3_server_method;
+#else
+        return TEST_SKIPPED;
+#endif /* WOLFSSL_DTLS13 */
+    }
+
+    client_cbs.ssl_ready = test_wolfSSL_dtls_send_alert;
+    test_wolfSSL_client_server_nofail(&client_cbs, &server_cbs);
+
+    if (!client_cbs.return_code)
+        return TEST_FAIL;
+    if (!server_cbs.return_code)
+        return TEST_FAIL;
+
+    return TEST_SUCCESS;
+}
+
+int test_wolfSSL_ignore_alert_before_cookie(void)
+{
+    int ret;
+    ret =_test_wolfSSL_ignore_alert_before_cookie(0);
+    if (ret != 0)
+        return ret;
+    ret =_test_wolfSSL_ignore_alert_before_cookie(1);
+    if (ret != 0)
+        return ret;
+    return 0;
+}
+
+static void test_wolfSSL_send_bad_record(WOLFSSL* ssl)
+{
+    int ret;
+    int fd;
+
+    byte bad_msg[] = {
+        0x17, /* app data  */
+        0xaa, 0xfd, /* bad version */
+        0x00, 0x01, /* epoch 1 */
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x55, /* not seen seq number */
+        0x00, 0x26, /* length: 38 bytes */
+        0xae, 0x30, 0x31, 0xb1, 0xf1, 0xb9, 0x6f, 0xda, 0x17, 0x19, 0xd9, 0x57,
+        0xa9, 0x9d, 0x5c, 0x51, 0x9b, 0x53, 0x63, 0xa5, 0x24, 0x70, 0xa1,
+        0xae, 0xdf, 0x1c, 0xb9, 0xfc, 0xe3, 0xd7, 0x77, 0x6d, 0xb6, 0x89, 0x0f,
+        0x03, 0x18, 0x72
+    };
+
+    fd = wolfSSL_get_wfd(ssl);
+    AssertIntGE(fd, 0);
+    ret = (int)send(fd, (MESSAGE_TYPE_CAST)bad_msg, sizeof(bad_msg), 0);
+    AssertIntEQ(ret, sizeof(bad_msg));
+    ret = wolfSSL_write(ssl, "badrecordtest", sizeof("badrecordtest"));
+    AssertIntEQ(ret, sizeof("badrecordtest"));
+}
+
+static void test_wolfSSL_read_string(WOLFSSL* ssl)
+{
+    byte buf[100];
+    int ret;
+
+    ret = wolfSSL_read(ssl, buf, sizeof(buf));
+    AssertIntGT(ret, 0);
+    AssertIntEQ(strcmp((char*)buf, "badrecordtest"), 0);
+}
+
+static int _test_wolfSSL_dtls_bad_record(
+    method_provider client_method, method_provider server_method)
+{
+    callback_functions client_cbs, server_cbs;
+
+    XMEMSET(&client_cbs, 0, sizeof(client_cbs));
+    XMEMSET(&server_cbs, 0, sizeof(server_cbs));
+    client_cbs.doUdp = server_cbs.doUdp = 1;
+    client_cbs.method = client_method;
+    server_cbs.method = server_method;
+
+    client_cbs.on_result = test_wolfSSL_send_bad_record;
+    server_cbs.on_result = test_wolfSSL_read_string;
+
+    test_wolfSSL_client_server_nofail(&client_cbs, &server_cbs);
+
+    if (!client_cbs.return_code)
+        return TEST_FAIL;
+    if (!server_cbs.return_code)
+        return TEST_FAIL;
+
+    return TEST_SUCCESS;
+}
+
+int test_wolfSSL_dtls_bad_record(void)
+{
+    int ret = TEST_SUCCESS;
+#if !defined(WOLFSSL_NO_TLS12)
+    ret = _test_wolfSSL_dtls_bad_record(wolfDTLSv1_2_client_method,
+        wolfDTLSv1_2_server_method);
+#endif
+#ifdef WOLFSSL_DTLS13
+    if (ret == TEST_SUCCESS) {
+        ret = _test_wolfSSL_dtls_bad_record(wolfDTLSv1_3_client_method,
+        wolfDTLSv1_3_server_method);
+    }
+#endif /* WOLFSSL_DTLS13 */
+    return ret;
+
+}
+
+#else
+int test_wolfSSL_dtls_fragments(void)
+{
+    return TEST_SKIPPED;
+}
+int test_wolfSSL_ignore_alert_before_cookie(void)
+{
+    return TEST_SKIPPED;
+}
+int test_wolfSSL_dtls_bad_record(void)
+{
+    return TEST_SKIPPED;
+}
+#endif
+
+#if defined(WOLFSSL_DTLS13) && !defined(WOLFSSL_TLS13_IGNORE_AEAD_LIMITS) && \
+    !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER) && \
+    defined(HAVE_IO_TESTS_DEPENDENCIES)
+static volatile int test_AEAD_seq_num = 0;
+#ifdef WOLFSSL_NO_ATOMICS
+static volatile int test_AEAD_done = 0;
+#else
+wolfSSL_Atomic_Int test_AEAD_done = WOLFSSL_ATOMIC_INITIALIZER(0);
+#endif
+#ifdef WOLFSSL_MUTEX_INITIALIZER
+static wolfSSL_Mutex test_AEAD_mutex = WOLFSSL_MUTEX_INITIALIZER(test_AEAD_mutex);
+#endif
+
+static int test_AEAD_fail_decryption = 0;
+static int test_AEAD_cbiorecv(WOLFSSL *ssl, char *buf, int sz, void *ctx)
+{
+    int fd = wolfSSL_get_fd(ssl);
+    int ret = -1;
+    if (fd >= 0 && (ret = (int)recv(fd, buf, sz, 0)) > 0) {
+        if (test_AEAD_fail_decryption) {
+            /* Modify the packet to trigger a decryption failure */
+            buf[ret/2] ^= 0xFF;
+            if (test_AEAD_fail_decryption == 1)
+                test_AEAD_fail_decryption = 0;
+        }
+    }
+    (void)ctx;
+    return ret;
+}
+
+static void test_AEAD_get_limits(WOLFSSL* ssl, w64wrapper* hardLimit,
+        w64wrapper* keyUpdateLimit, w64wrapper* sendLimit)
+{
+    if (sendLimit)
+        w64Zero(sendLimit);
+    switch (ssl->specs.bulk_cipher_algorithm) {
+        case wolfssl_aes_gcm:
+            if (sendLimit)
+                *sendLimit = AEAD_AES_LIMIT;
+            FALL_THROUGH;
+        case wolfssl_chacha:
+            if (hardLimit)
+                *hardLimit = DTLS_AEAD_AES_GCM_CHACHA_FAIL_LIMIT;
+            if (keyUpdateLimit)
+                *keyUpdateLimit = DTLS_AEAD_AES_GCM_CHACHA_FAIL_KU_LIMIT;
+            break;
+        case wolfssl_aes_ccm:
+            if (sendLimit)
+                *sendLimit = DTLS_AEAD_AES_CCM_LIMIT;
+            if (ssl->specs.aead_mac_size == AES_CCM_8_AUTH_SZ) {
+                if (hardLimit)
+                    *hardLimit = DTLS_AEAD_AES_CCM_8_FAIL_LIMIT;
+                if (keyUpdateLimit)
+                    *keyUpdateLimit = DTLS_AEAD_AES_CCM_8_FAIL_KU_LIMIT;
+            }
+            else {
+                if (hardLimit)
+                    *hardLimit = DTLS_AEAD_AES_CCM_FAIL_LIMIT;
+                if (keyUpdateLimit)
+                    *keyUpdateLimit = DTLS_AEAD_AES_CCM_FAIL_KU_LIMIT;
+            }
+            break;
+        default:
+            fprintf(stderr, "Unrecognized bulk cipher");
+            AssertFalse(1);
+            break;
+    }
+}
+
+static void test_AEAD_limit_client(WOLFSSL* ssl)
+{
+    int ret;
+    int i;
+    int didReKey = 0;
+    char msgBuf[20];
+    w64wrapper hardLimit;
+    w64wrapper keyUpdateLimit;
+    w64wrapper counter;
+    w64wrapper sendLimit;
+
+    test_AEAD_get_limits(ssl, &hardLimit, &keyUpdateLimit, &sendLimit);
+
+    w64Zero(&counter);
+    AssertTrue(w64Equal(Dtls13GetEpoch(ssl, ssl->dtls13Epoch)->dropCount, counter));
+
+    wolfSSL_SSLSetIORecv(ssl, test_AEAD_cbiorecv);
+
+    for (i = 0; i < 10; i++) {
+        /* Test some failed decryptions */
+        test_AEAD_fail_decryption = 1;
+        w64Increment(&counter);
+        ret = wolfSSL_read(ssl, msgBuf, sizeof(msgBuf));
+        /* Should succeed since decryption failures are dropped */
+        AssertIntGT(ret, 0);
+        AssertTrue(w64Equal(Dtls13GetEpoch(ssl, ssl->dtls13PeerEpoch)->dropCount, counter));
+    }
+
+    test_AEAD_fail_decryption = 1;
+    Dtls13GetEpoch(ssl, ssl->dtls13PeerEpoch)->dropCount = keyUpdateLimit;
+    w64Increment(&Dtls13GetEpoch(ssl, ssl->dtls13PeerEpoch)->dropCount);
+    /* 100 read calls should be enough to complete the key update */
+    w64Zero(&counter);
+    for (i = 0; i < 100; i++) {
+        /* Key update should be sent and negotiated */
+        ret = wolfSSL_read(ssl, msgBuf, sizeof(msgBuf));
+        AssertIntGT(ret, 0);
+        /* Epoch after one key update is 4 */
+        if (w64Equal(ssl->dtls13PeerEpoch, w64From32(0, 4)) &&
+                w64Equal(Dtls13GetEpoch(ssl, ssl->dtls13PeerEpoch)->dropCount, counter)) {
+            didReKey = 1;
+            break;
+        }
+    }
+    AssertTrue(didReKey);
+
+    if (!w64IsZero(sendLimit)) {
+        /* Test the sending limit for AEAD ciphers */
+#ifdef WOLFSSL_MUTEX_INITIALIZER
+        (void)wc_LockMutex(&test_AEAD_mutex);
+#endif
+        Dtls13GetEpoch(ssl, ssl->dtls13Epoch)->nextSeqNumber = sendLimit;
+        test_AEAD_seq_num = 1;
+        XMEMSET(msgBuf, 0, sizeof(msgBuf));
+        ret = wolfSSL_write(ssl, msgBuf, sizeof(msgBuf));
+        AssertIntGT(ret, 0);
+        didReKey = 0;
+        w64Zero(&counter);
+#ifdef WOLFSSL_MUTEX_INITIALIZER
+        wc_UnLockMutex(&test_AEAD_mutex);
+#endif
+        /* 100 read calls should be enough to complete the key update */
+        for (i = 0; i < 100; i++) {
+            /* Key update should be sent and negotiated */
+            ret = wolfSSL_read(ssl, msgBuf, sizeof(msgBuf));
+            AssertIntGT(ret, 0);
+            /* Epoch after another key update is 5 */
+            if (w64Equal(ssl->dtls13Epoch, w64From32(0, 5)) &&
+                    w64Equal(Dtls13GetEpoch(ssl, ssl->dtls13Epoch)->dropCount, counter)) {
+                didReKey = 1;
+                break;
+            }
+        }
+        AssertTrue(didReKey);
+    }
+
+    test_AEAD_fail_decryption = 2;
+    Dtls13GetEpoch(ssl, ssl->dtls13PeerEpoch)->dropCount = hardLimit;
+    w64Decrement(&Dtls13GetEpoch(ssl, ssl->dtls13PeerEpoch)->dropCount);
+    /* Connection should fail with a DECRYPT_ERROR */
+    ret = wolfSSL_read(ssl, msgBuf, sizeof(msgBuf));
+    AssertIntEQ(ret, WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR));
+    AssertIntEQ(wolfSSL_get_error(ssl, ret), WC_NO_ERR_TRACE(DECRYPT_ERROR));
+
+#ifdef WOLFSSL_ATOMIC_INITIALIZER
+    WOLFSSL_ATOMIC_STORE(test_AEAD_done, 1);
+#else
+    test_AEAD_done = 1;
+#endif
+}
+
+int counter = 0;
+static void test_AEAD_limit_server(WOLFSSL* ssl)
+{
+    char msgBuf[] = "Sending data";
+    int ret = WOLFSSL_SUCCESS;
+    w64wrapper sendLimit;
+    SOCKET_T fd = wolfSSL_get_fd(ssl);
+    struct timespec delay;
+    XMEMSET(&delay, 0, sizeof(delay));
+    delay.tv_nsec = 100000000; /* wait 0.1 seconds */
+    tcp_set_nonblocking(&fd); /* So that read doesn't block */
+    wolfSSL_dtls_set_using_nonblock(ssl, 1);
+    test_AEAD_get_limits(ssl, NULL, NULL, &sendLimit);
+    while (!
+    #ifdef WOLFSSL_ATOMIC_INITIALIZER
+           WOLFSSL_ATOMIC_LOAD(test_AEAD_done)
+    #else
+           test_AEAD_done
+    #endif
+           && ret > 0)
+    {
+        counter++;
+#ifdef WOLFSSL_MUTEX_INITIALIZER
+        (void)wc_LockMutex(&test_AEAD_mutex);
+#endif
+        if (test_AEAD_seq_num) {
+            /* We need to update the seq number so that we can understand the
+             * peer. Otherwise we will incorrectly interpret the seq number. */
+            Dtls13Epoch* e = Dtls13GetEpoch(ssl, ssl->dtls13PeerEpoch);
+            AssertNotNull(e);
+            e->nextPeerSeqNumber = sendLimit;
+            test_AEAD_seq_num = 0;
+        }
+#ifdef WOLFSSL_MUTEX_INITIALIZER
+        wc_UnLockMutex(&test_AEAD_mutex);
+#endif
+        (void)wolfSSL_read(ssl, msgBuf, sizeof(msgBuf));
+        ret = wolfSSL_write(ssl, msgBuf, sizeof(msgBuf));
+        nanosleep(&delay, NULL);
+    }
+}
+
+int test_wolfSSL_dtls_AEAD_limit(void)
+{
+    callback_functions func_cb_client;
+    callback_functions func_cb_server;
+    XMEMSET(&func_cb_client, 0, sizeof(callback_functions));
+    XMEMSET(&func_cb_server, 0, sizeof(callback_functions));
+
+    func_cb_client.doUdp = func_cb_server.doUdp = 1;
+    func_cb_server.method = wolfDTLSv1_3_server_method;
+    func_cb_client.method = wolfDTLSv1_3_client_method;
+    func_cb_server.on_result = test_AEAD_limit_server;
+    func_cb_client.on_result = test_AEAD_limit_client;
+
+    test_wolfSSL_client_server_nofail(&func_cb_client, &func_cb_server);
+
+    if (!func_cb_client.return_code)
+        return TEST_FAIL;
+    if (!func_cb_server.return_code)
+        return TEST_FAIL;
+
+    return TEST_SUCCESS;
+}
+#else
+int test_wolfSSL_dtls_AEAD_limit(void)
+{
+    return TEST_SKIPPED;
+}
+#endif
+
+#if defined(WOLFSSL_DTLS) && \
+    defined(HAVE_IO_TESTS_DEPENDENCIES) && !defined(SINGLE_THREADED) && \
+    !defined(DEBUG_VECTOR_REGISTER_ACCESS_FUZZING)
+static void test_wolfSSL_dtls_send_ch(WOLFSSL* ssl)
+{
+    int fd, ret;
+    byte ch_msg[] = {
+        0x16, 0xfe, 0xfd, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01,
+        0xfa, 0x01, 0x00, 0x01, 0xee, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01,
+        0xee, 0xfe, 0xfd, 0xc0, 0xca, 0xb5, 0x6f, 0x3d, 0x23, 0xcc, 0x53, 0x9a,
+        0x67, 0x17, 0x70, 0xd3, 0xfb, 0x23, 0x16, 0x9e, 0x4e, 0xd6, 0x7e, 0x29,
+        0xab, 0xfa, 0x4c, 0xa5, 0x84, 0x95, 0xc3, 0xdb, 0x21, 0x9a, 0x52, 0x00,
+        0x00, 0x00, 0x36, 0x13, 0x01, 0x13, 0x02, 0x13, 0x03, 0xc0, 0x2c, 0xc0,
+        0x2b, 0xc0, 0x30, 0xc0, 0x2f, 0x00, 0x9f, 0x00, 0x9e, 0xcc, 0xa9, 0xcc,
+        0xa8, 0xcc, 0xaa, 0xc0, 0x27, 0xc0, 0x23, 0xc0, 0x28, 0xc0, 0x24, 0xc0,
+        0x0a, 0xc0, 0x09, 0xc0, 0x14, 0xc0, 0x13, 0x00, 0x6b, 0x00, 0x67, 0x00,
+        0x39, 0x00, 0x33, 0xcc, 0x14, 0xcc, 0x13, 0xcc, 0x15, 0x01, 0x00, 0x01,
+        0x8e, 0x00, 0x2b, 0x00, 0x03, 0x02, 0xfe, 0xfc, 0x00, 0x0d, 0x00, 0x20,
+        0x00, 0x1e, 0x06, 0x03, 0x05, 0x03, 0x04, 0x03, 0x02, 0x03, 0x08, 0x06,
+        0x08, 0x0b, 0x08, 0x05, 0x08, 0x0a, 0x08, 0x04, 0x08, 0x09, 0x06, 0x01,
+        0x05, 0x01, 0x04, 0x01, 0x03, 0x01, 0x02, 0x01, 0x00, 0x0a, 0x00, 0x0c,
+        0x00, 0x0a, 0x00, 0x19, 0x00, 0x18, 0x00, 0x17, 0x00, 0x15, 0x01, 0x00,
+        0x00, 0x16, 0x00, 0x00, 0x00, 0x33, 0x01, 0x4b, 0x01, 0x49, 0x00, 0x17,
+        0x00, 0x41, 0x04, 0x96, 0xcb, 0x2e, 0x4e, 0xd9, 0x88, 0x71, 0xc7, 0xf3,
+        0x1a, 0x16, 0xdd, 0x7a, 0x7c, 0xf7, 0x67, 0x8a, 0x5d, 0x9a, 0x55, 0xa6,
+        0x4a, 0x90, 0xd9, 0xfb, 0xc7, 0xfb, 0xbe, 0x09, 0xa9, 0x8a, 0xb5, 0x7a,
+        0xd1, 0xde, 0x83, 0x74, 0x27, 0x31, 0x1c, 0xaa, 0xae, 0xef, 0x58, 0x43,
+        0x13, 0x7d, 0x15, 0x4d, 0x7f, 0x68, 0xf6, 0x8a, 0x38, 0xef, 0x0e, 0xb3,
+        0xcf, 0xb8, 0x4a, 0xa9, 0xb4, 0xd7, 0xcb, 0x01, 0x00, 0x01, 0x00, 0x1d,
+        0x0a, 0x22, 0x8a, 0xd1, 0x78, 0x85, 0x1e, 0x5a, 0xe1, 0x1d, 0x1e, 0xb7,
+        0x2d, 0xbc, 0x5f, 0x52, 0xbc, 0x97, 0x5d, 0x8b, 0x6a, 0x8b, 0x9d, 0x1e,
+        0xb1, 0xfc, 0x8a, 0xb2, 0x56, 0xcd, 0xed, 0x4b, 0xfb, 0x66, 0x3f, 0x59,
+        0x3f, 0x15, 0x5d, 0x09, 0x9e, 0x2f, 0x60, 0x5b, 0x31, 0x81, 0x27, 0xf0,
+        0x1c, 0xda, 0xcd, 0x48, 0x66, 0xc6, 0xbb, 0x25, 0xf0, 0x5f, 0xda, 0x4c,
+        0xcf, 0x1d, 0x88, 0xc8, 0xda, 0x1b, 0x53, 0xea, 0xbd, 0xce, 0x6d, 0xf6,
+        0x4a, 0x76, 0xdb, 0x75, 0x99, 0xaf, 0xcf, 0x76, 0x4a, 0xfb, 0xe3, 0xef,
+        0xb2, 0xcb, 0xae, 0x4a, 0xc0, 0xe8, 0x63, 0x1f, 0xd6, 0xe8, 0xe6, 0x45,
+        0xf9, 0xea, 0x0d, 0x06, 0x19, 0xfc, 0xb1, 0xfd, 0x5d, 0x92, 0x89, 0x7b,
+        0xc7, 0x9f, 0x1a, 0xb3, 0x2b, 0xc7, 0xad, 0x0e, 0xfb, 0x13, 0x41, 0x83,
+        0x84, 0x58, 0x3a, 0x25, 0xb9, 0x49, 0x35, 0x1c, 0x23, 0xcb, 0xd6, 0xe7,
+        0xc2, 0x8c, 0x4b, 0x2a, 0x73, 0xa1, 0xdf, 0x4f, 0x73, 0x9b, 0xb3, 0xd2,
+        0xb2, 0x95, 0x00, 0x3c, 0x26, 0x09, 0x89, 0x71, 0x05, 0x39, 0xc8, 0x98,
+        0x8f, 0xed, 0x32, 0x15, 0x78, 0xcd, 0xd3, 0x7e, 0xfb, 0x5a, 0x78, 0x2a,
+        0xdc, 0xca, 0x20, 0x09, 0xb5, 0x14, 0xf9, 0xd4, 0x58, 0xf6, 0x69, 0xf8,
+        0x65, 0x9f, 0xb7, 0xe4, 0x93, 0xf1, 0xa3, 0x84, 0x7e, 0x1b, 0x23, 0x5d,
+        0xea, 0x59, 0x3e, 0x4d, 0xca, 0xfd, 0xa5, 0x55, 0xdd, 0x99, 0xb5, 0x02,
+        0xf8, 0x0d, 0xe5, 0xf4, 0x06, 0xb0, 0x43, 0x9e, 0x2e, 0xbf, 0x05, 0x33,
+        0x65, 0x7b, 0x13, 0x8c, 0xf9, 0x16, 0x4d, 0xc5, 0x15, 0x0b, 0x40, 0x2f,
+        0x66, 0x94, 0xf2, 0x43, 0x95, 0xe7, 0xa9, 0xb6, 0x39, 0x99, 0x73, 0xb3,
+        0xb0, 0x06, 0xfe, 0x52, 0x9e, 0x57, 0xba, 0x75, 0xfd, 0x76, 0x7b, 0x20,
+        0x31, 0x68, 0x4c
+    };
+
+    fd = wolfSSL_get_wfd(ssl);
+    AssertIntGE(fd, 0);
+    ret = (int)send(fd, (MESSAGE_TYPE_CAST)ch_msg, sizeof(ch_msg), 0);
+    AssertIntGT(ret, 0);
+    /* consume the HRR otherwise handshake will fail */
+    ret = (int)recv(fd, (MESSAGE_TYPE_CAST)ch_msg, sizeof(ch_msg), 0);
+    AssertIntGT(ret, 0);
+}
+
+#if defined(WOLFSSL_DTLS13) && defined(WOLFSSL_SEND_HRR_COOKIE)
+static void test_wolfSSL_dtls_send_ch_with_invalid_cookie(WOLFSSL* ssl)
+{
+    int fd, ret;
+    byte ch_msh_invalid_cookie[] = {
+      0x16, 0xfe, 0xfd, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x02,
+      0x4e, 0x01, 0x00, 0x02, 0x42, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x02,
+      0x42, 0xfe, 0xfd, 0x69, 0xca, 0x77, 0x60, 0x6f, 0xfc, 0xd1, 0x5b, 0x60,
+      0x5d, 0xf1, 0xa6, 0x5c, 0x44, 0x71, 0xae, 0xca, 0x62, 0x19, 0x0c, 0xb6,
+      0xf7, 0x2c, 0xa6, 0xd5, 0xd2, 0x99, 0x9d, 0x18, 0xae, 0xac, 0x11, 0x00,
+      0x00, 0x00, 0x36, 0x13, 0x01, 0x13, 0x02, 0x13, 0x03, 0xc0, 0x2c, 0xc0,
+      0x2b, 0xc0, 0x30, 0xc0, 0x2f, 0x00, 0x9f, 0x00, 0x9e, 0xcc, 0xa9, 0xcc,
+      0xa8, 0xcc, 0xaa, 0xc0, 0x27, 0xc0, 0x23, 0xc0, 0x28, 0xc0, 0x24, 0xc0,
+      0x0a, 0xc0, 0x09, 0xc0, 0x14, 0xc0, 0x13, 0x00, 0x6b, 0x00, 0x67, 0x00,
+      0x39, 0x00, 0x33, 0xcc, 0x14, 0xcc, 0x13, 0xcc, 0x15, 0x01, 0x00, 0x01,
+      0xe2, 0x00, 0x2b, 0x00, 0x03, 0x02, 0xfe, 0xfc, 0x00, 0x0d, 0x00, 0x20,
+      0x00, 0x1e, 0x06, 0x03, 0x05, 0x03, 0x04, 0x03, 0x02, 0x03, 0x08, 0x06,
+      0x08, 0x0b, 0x08, 0x05, 0x08, 0x0a, 0x08, 0x04, 0x08, 0x09, 0x06, 0x01,
+      0x05, 0x01, 0x04, 0x01, 0x03, 0x01, 0x02, 0x01, 0x00, 0x2c, 0x00, 0x45,
+      0x00, 0x43, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+      0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+      0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+      0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+      0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
+      0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00, 0x2d, 0x00,
+      0x03, 0x02, 0x00, 0x01, 0x00, 0x0a, 0x00, 0x0c, 0x00, 0x0a, 0x00, 0x19,
+      0x00, 0x18, 0x00, 0x17, 0x00, 0x15, 0x01, 0x00, 0x00, 0x16, 0x00, 0x00,
+      0x00, 0x33, 0x01, 0x4b, 0x01, 0x49, 0x00, 0x17, 0x00, 0x41, 0x04, 0x7c,
+      0x5a, 0xc2, 0x5a, 0xfd, 0xcd, 0x2b, 0x08, 0xb2, 0xeb, 0x8e, 0xc0, 0x02,
+      0x03, 0x9d, 0xb1, 0xc1, 0x0d, 0x7b, 0x7f, 0x46, 0x43, 0xdf, 0xf3, 0xee,
+      0x2b, 0x78, 0x0e, 0x29, 0x8c, 0x42, 0x11, 0x2c, 0xde, 0xd7, 0x41, 0x0f,
+      0x28, 0x94, 0x80, 0x41, 0x70, 0xc4, 0x17, 0xfd, 0x6d, 0xfa, 0xee, 0x9a,
+      0xf2, 0xc4, 0x15, 0x4c, 0x5f, 0x54, 0xb6, 0x78, 0x6e, 0xf9, 0x63, 0x27,
+      0x33, 0xb8, 0x7b, 0x01, 0x00, 0x01, 0x00, 0xd4, 0x46, 0x62, 0x9c, 0xbf,
+      0x8f, 0x1b, 0x65, 0x9b, 0xf0, 0x29, 0x64, 0xd8, 0x50, 0x0e, 0x74, 0xf1,
+      0x58, 0x10, 0xc9, 0xd9, 0x82, 0x5b, 0xd9, 0xbe, 0x14, 0xdf, 0xde, 0x86,
+      0xb4, 0x2e, 0x15, 0xee, 0x4f, 0xf6, 0x74, 0x9e, 0x59, 0x11, 0x36, 0x2d,
+      0xb9, 0x67, 0xaa, 0x5a, 0x09, 0x9b, 0x45, 0xf1, 0x01, 0x4c, 0x4e, 0xf6,
+      0xda, 0x6a, 0xae, 0xa7, 0x73, 0x7b, 0x2e, 0xb6, 0x24, 0x89, 0x99, 0xb7,
+      0x52, 0x16, 0x62, 0x0a, 0xab, 0x58, 0xf8, 0x3f, 0x10, 0x5b, 0x83, 0xfd,
+      0x7b, 0x81, 0x77, 0x81, 0x8d, 0xef, 0x24, 0x56, 0x6d, 0xba, 0x49, 0xd4,
+      0x8b, 0xb5, 0xa0, 0xb1, 0xc9, 0x8c, 0x32, 0x95, 0x1c, 0x5e, 0x0a, 0x4b,
+      0xf6, 0x00, 0x50, 0x0a, 0x87, 0x99, 0x59, 0xcf, 0x6f, 0x9d, 0x02, 0xd0,
+      0x1b, 0xa1, 0x96, 0x45, 0x28, 0x76, 0x40, 0x33, 0x28, 0xc9, 0xa1, 0xfd,
+      0x46, 0xab, 0x2c, 0x9e, 0x5e, 0xc6, 0x74, 0x19, 0x9a, 0xf5, 0x9b, 0x51,
+      0x11, 0x4f, 0xc8, 0xb9, 0x99, 0x6b, 0x4e, 0x3e, 0x31, 0x64, 0xb4, 0x92,
+      0xf4, 0x0d, 0x41, 0x4b, 0x2c, 0x65, 0x23, 0xf7, 0x47, 0xe3, 0xa5, 0x2e,
+      0xe4, 0x9c, 0x2b, 0xc9, 0x41, 0x22, 0x83, 0x8a, 0x23, 0xef, 0x29, 0x7e,
+      0x4f, 0x3f, 0xa3, 0xbf, 0x73, 0x2b, 0xd7, 0xcc, 0xc8, 0xc6, 0xe9, 0xbc,
+      0x01, 0xb7, 0x32, 0x63, 0xd4, 0x7e, 0x7f, 0x9a, 0xaf, 0x5f, 0x05, 0x31,
+      0x53, 0xd6, 0x1f, 0xa2, 0xd0, 0xdf, 0x67, 0x56, 0xf1, 0x9c, 0x4a, 0x9d,
+      0x83, 0xb4, 0xef, 0xb3, 0xf2, 0xcc, 0xf1, 0x91, 0x6c, 0x47, 0xc3, 0x8b,
+      0xd0, 0x92, 0x79, 0x3d, 0xa0, 0xc0, 0x3a, 0x57, 0x26, 0x6d, 0x0a, 0xad,
+      0x5f, 0xad, 0xb4, 0x74, 0x48, 0x4a, 0x51, 0xe1, 0xb5, 0x82, 0x0a, 0x4c,
+      0x4f, 0x9d, 0xaf, 0xee, 0x5a, 0xa2, 0x4d, 0x4d, 0x5f, 0xe0, 0x17, 0x00,
+      0x23, 0x00, 0x00
+    };
+    byte alert_reply[50];
+    byte expected_alert_reply[] = {
+        0x15, 0xfe, 0xfd, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00,
+        0x02, 0x02, 0x2f
+    };
+
+    fd = wolfSSL_get_wfd(ssl);
+    if (fd >= 0) {
+        ret = (int)send(fd, (MESSAGE_TYPE_CAST)ch_msh_invalid_cookie,
+                sizeof(ch_msh_invalid_cookie), 0);
+        AssertIntGT(ret, 0);
+        /* should reply with an illegal_parameter reply */
+        ret = (int)recv(fd, (MESSAGE_TYPE_CAST)alert_reply, sizeof(alert_reply), 0);
+        AssertIntEQ(ret, sizeof(expected_alert_reply));
+        AssertIntEQ(XMEMCMP(alert_reply, expected_alert_reply,
+                sizeof(expected_alert_reply)), 0);
+    }
+}
+#endif
+
+static word32 test_wolfSSL_dtls_stateless_HashWOLFSSL(const WOLFSSL* ssl)
+{
+#ifndef NO_MD5
+    enum wc_HashType hashType = WC_HASH_TYPE_MD5;
+#elif !defined(NO_SHA)
+    enum wc_HashType hashType = WC_HASH_TYPE_SHA;
+#elif !defined(NO_SHA256)
+    enum wc_HashType hashType = WC_HASH_TYPE_SHA256;
+#else
+    #error "We need a digest to hash the WOLFSSL object"
+#endif
+    byte hashBuf[WC_MAX_DIGEST_SIZE];
+    wc_HashAlg hash;
+    const TLSX* exts = ssl->extensions;
+    WOLFSSL sslCopy; /* Use a copy to omit certain fields */
+#ifndef WOLFSSL_SMALL_STACK_CACHE
+    HS_Hashes* hsHashes = ssl->hsHashes; /* Is re-allocated in
+                                          * InitHandshakeHashes */
+#endif
+
+    XMEMCPY(&sslCopy, ssl, sizeof(*ssl));
+    XMEMSET(hashBuf, 0, sizeof(hashBuf));
+
+    /* Following fields are not important to compare */
+    XMEMSET(sslCopy.buffers.inputBuffer.staticBuffer, 0, STATIC_BUFFER_LEN);
+    sslCopy.buffers.inputBuffer.buffer = NULL;
+    sslCopy.buffers.inputBuffer.bufferSize = 0;
+    sslCopy.buffers.inputBuffer.dynamicFlag = 0;
+    sslCopy.buffers.inputBuffer.offset = 0;
+    XMEMSET(sslCopy.buffers.outputBuffer.staticBuffer, 0, STATIC_BUFFER_LEN);
+    sslCopy.buffers.outputBuffer.buffer = NULL;
+    sslCopy.buffers.outputBuffer.bufferSize = 0;
+    sslCopy.buffers.outputBuffer.dynamicFlag = 0;
+    sslCopy.buffers.outputBuffer.offset = 0;
+    sslCopy.error = 0;
+    sslCopy.curSize = 0;
+    sslCopy.curStartIdx = 0;
+    sslCopy.keys.curSeq_lo = 0;
+    XMEMSET(&sslCopy.curRL, 0, sizeof(sslCopy.curRL));
+#ifdef WOLFSSL_DTLS13
+    XMEMSET(&sslCopy.keys.curSeq, 0, sizeof(sslCopy.keys.curSeq));
+    sslCopy.dtls13FastTimeout = 0;
+#endif
+    sslCopy.keys.dtls_peer_handshake_number = 0;
+    XMEMSET(&sslCopy.alert_history, 0, sizeof(sslCopy.alert_history));
+    sslCopy.hsHashes = NULL;
+#if !defined(WOLFSSL_NO_CLIENT_AUTH) && \
+    ((defined(WOLFSSL_SM2) && defined(WOLFSSL_SM3)) || \
+     (defined(HAVE_ED25519) && !defined(NO_ED25519_CLIENT_AUTH)) || \
+     (defined(HAVE_ED448) && !defined(NO_ED448_CLIENT_AUTH)))
+    sslCopy.options.cacheMessages = 0;
+#endif
+#ifdef WOLFSSL_ASYNC_IO
+#ifdef WOLFSSL_ASYNC_CRYPT
+    sslCopy.asyncDev = NULL;
+#endif
+    sslCopy.async = NULL;
+#endif
+
+    AssertIntEQ(wc_HashInit(&hash, hashType), 0);
+    AssertIntEQ(wc_HashUpdate(&hash, hashType, (byte*)&sslCopy, sizeof(sslCopy)), 0);
+    /* hash extension list */
+    while (exts != NULL) {
+        AssertIntEQ(wc_HashUpdate(&hash, hashType, (byte*)exts, sizeof(*exts)), 0);
+        exts = exts->next;
+    }
+    /* Hash suites */
+    if (sslCopy.suites != NULL) {
+        AssertIntEQ(wc_HashUpdate(&hash, hashType, (byte*)sslCopy.suites,
+                sizeof(struct Suites)), 0);
+    }
+
+#ifdef WOLFSSL_SMALL_STACK_CACHE
+    /* with WOLFSSL_SMALL_STACK_CACHE, the SHA-2 objects always differ after
+     * initialization because of cached W and (for SHA512) X buffers.
+     */
+#else
+    /* Hash hsHashes */
+    AssertIntEQ(wc_HashUpdate(&hash, hashType, (byte*)hsHashes,
+            sizeof(*hsHashes)), 0);
+#endif
+
+    AssertIntEQ(wc_HashFinal(&hash, hashType, hashBuf), 0);
+    AssertIntEQ(wc_HashFree(&hash, hashType), 0);
+
+    return MakeWordFromHash(hashBuf);
+}
+
+static CallbackIORecv test_wolfSSL_dtls_compare_stateless_cb;
+static int test_wolfSSL_dtls_compare_stateless_cb_call_once;
+static int test_wolfSSL_dtls_compare_stateless_read_cb_once(WOLFSSL *ssl,
+        char *buf, int sz, void *ctx)
+{
+    if (test_wolfSSL_dtls_compare_stateless_cb_call_once) {
+        test_wolfSSL_dtls_compare_stateless_cb_call_once = 0;
+        return test_wolfSSL_dtls_compare_stateless_cb(ssl, buf, sz, ctx);
+    }
+    else {
+        return WOLFSSL_CBIO_ERR_WANT_READ;
+    }
+}
+
+static void test_wolfSSL_dtls_compare_stateless(WOLFSSL* ssl)
+{
+    /* Compare the ssl object before and after one ClientHello msg */
+    SOCKET_T fd = wolfSSL_get_fd(ssl);
+    int res;
+    int err;
+    word32 initHash;
+
+    test_wolfSSL_dtls_compare_stateless_cb = ssl->CBIORecv;
+    test_wolfSSL_dtls_compare_stateless_cb_call_once = 1;
+    wolfSSL_dtls_set_using_nonblock(ssl, 1);
+    ssl->CBIORecv = test_wolfSSL_dtls_compare_stateless_read_cb_once;
+
+    initHash = test_wolfSSL_dtls_stateless_HashWOLFSSL(ssl);
+    (void)initHash;
+
+    res = tcp_select(fd, 5);
+    /* We are expecting a msg. A timeout indicates failure. */
+    AssertIntEQ(res, TEST_RECV_READY);
+
+    res = wolfSSL_accept(ssl);
+    err = wolfSSL_get_error(ssl, res);
+    AssertIntEQ(res, WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR));
+    AssertIntEQ(err, WOLFSSL_ERROR_WANT_READ);
+
+    AssertIntEQ(initHash, test_wolfSSL_dtls_stateless_HashWOLFSSL(ssl));
+
+    wolfSSL_dtls_set_using_nonblock(ssl, 0);
+    ssl->CBIORecv = test_wolfSSL_dtls_compare_stateless_cb;
+
+}
+
+#if defined(WOLFSSL_DTLS13) && defined(WOLFSSL_SEND_HRR_COOKIE)
+static void test_wolfSSL_dtls_enable_hrrcookie(WOLFSSL* ssl)
+{
+    int ret;
+    ret = wolfSSL_send_hrr_cookie(ssl, NULL, 0);
+    AssertIntEQ(ret, WOLFSSL_SUCCESS);
+    test_wolfSSL_dtls_compare_stateless(ssl);
+}
+#endif
+
+int test_wolfSSL_dtls_stateless(void)
+{
+    callback_functions client_cbs, server_cbs;
+    size_t i;
+    struct {
+        method_provider client_meth;
+        method_provider server_meth;
+        ssl_callback client_ssl_ready;
+        ssl_callback server_ssl_ready;
+    } test_params[] = {
+#if !defined(WOLFSSL_NO_TLS12)
+        {wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method,
+                test_wolfSSL_dtls_send_ch, test_wolfSSL_dtls_compare_stateless},
+#endif
+#if defined(WOLFSSL_DTLS13) && defined(WOLFSSL_SEND_HRR_COOKIE)
+        {wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method,
+                test_wolfSSL_dtls_send_ch, test_wolfSSL_dtls_enable_hrrcookie},
+        {wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method,
+                test_wolfSSL_dtls_send_ch_with_invalid_cookie, test_wolfSSL_dtls_enable_hrrcookie},
+#endif
+    };
+
+    if (0 == sizeof(test_params)){
+        return TEST_SKIPPED;
+    }
+
+    for (i = 0; i < sizeof(test_params)/sizeof(*test_params); i++) {
+        XMEMSET(&client_cbs, 0, sizeof(client_cbs));
+        XMEMSET(&server_cbs, 0, sizeof(server_cbs));
+        client_cbs.doUdp = server_cbs.doUdp = 1;
+        client_cbs.method = test_params[i].client_meth;
+        server_cbs.method = test_params[i].server_meth;
+
+        client_cbs.ssl_ready = test_params[i].client_ssl_ready;
+        server_cbs.ssl_ready = test_params[i].server_ssl_ready;
+        test_wolfSSL_client_server_nofail(&client_cbs, &server_cbs);
+
+        if (!client_cbs.return_code)
+            return TEST_FAIL;
+        if (!server_cbs.return_code)
+            return TEST_FAIL;
+    }
+
+    return TEST_SUCCESS;
+}
+
+/* DTLS stateless API handling multiple CHs with different HRR groups */
+int test_wolfSSL_dtls_stateless_hrr_group(void)
+{
+    EXPECT_DECLS;
+#if defined(WOLFSSL_SEND_HRR_COOKIE)
+    size_t i;
+    word32 initHash;
+    struct {
+        method_provider client_meth;
+        method_provider server_meth;
+    } params[] = {
+#if defined(WOLFSSL_TLS13) && defined(WOLFSSL_DTLS13)
+        { wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method },
+#endif
+#if !defined(WOLFSSL_NO_TLS12) && defined(WOLFSSL_DTLS)
+        { wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method },
+#endif
+    };
+    for (i = 0; i < XELEM_CNT(params) && !EXPECT_FAIL(); i++) {
+        WOLFSSL_CTX *ctx_s = NULL, *ctx_c = NULL;
+        WOLFSSL *ssl_s = NULL, *ssl_c = NULL, *ssl_c2 = NULL;
+        struct test_memio_ctx test_ctx;
+        int groups_1[] = {
+            WOLFSSL_ECC_SECP256R1,
+            WOLFSSL_ECC_SECP384R1,
+            WOLFSSL_ECC_SECP521R1
+        };
+        int groups_2[] = {
+            WOLFSSL_ECC_SECP384R1,
+            WOLFSSL_ECC_SECP521R1
+        };
+        char hrrBuf[1000];
+        int hrrSz = sizeof(hrrBuf);
+
+        XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+
+        ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+                params[i].client_meth, params[i].server_meth), 0);
+
+        ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, NULL, &ssl_c2, NULL,
+                params[i].client_meth, params[i].server_meth), 0);
+
+
+        wolfSSL_SetLoggingPrefix("server");
+        wolfSSL_dtls_set_using_nonblock(ssl_s, 1);
+
+        initHash = test_wolfSSL_dtls_stateless_HashWOLFSSL(ssl_s);
+
+        /* Set groups and disable key shares. This ensures that only the given
+         * groups are in the SupportedGroups extension and that an empty key
+         * share extension is sent in the initial ClientHello of each session.
+         * This triggers the server to send a HelloRetryRequest with the first
+         * group in the SupportedGroups extension selected. */
+        wolfSSL_SetLoggingPrefix("client1");
+        ExpectIntEQ(wolfSSL_set_groups(ssl_c, groups_1, 3), WOLFSSL_SUCCESS);
+        ExpectIntEQ(wolfSSL_NoKeyShares(ssl_c), WOLFSSL_SUCCESS);
+
+        wolfSSL_SetLoggingPrefix("client2");
+        ExpectIntEQ(wolfSSL_set_groups(ssl_c2, groups_2, 2), WOLFSSL_SUCCESS);
+        ExpectIntEQ(wolfSSL_NoKeyShares(ssl_c2), WOLFSSL_SUCCESS);
+
+        /* Start handshake, send first ClientHello */
+        wolfSSL_SetLoggingPrefix("client1");
+        ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
+        ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+
+        /* Read first ClientHello, send HRR with WOLFSSL_ECC_SECP256R1 */
+        wolfSSL_SetLoggingPrefix("server");
+        ExpectIntEQ(wolfDTLS_accept_stateless(ssl_s), 0);
+        ExpectIntEQ(test_memio_copy_message(&test_ctx, 1, hrrBuf, &hrrSz, 0), 0);
+        ExpectIntGT(hrrSz, 0);
+        ExpectIntEQ(initHash, test_wolfSSL_dtls_stateless_HashWOLFSSL(ssl_s));
+        test_memio_clear_buffer(&test_ctx, 1);
+
+        /* Send second ClientHello */
+        wolfSSL_SetLoggingPrefix("client2");
+        ExpectIntEQ(wolfSSL_connect(ssl_c2), -1);
+        ExpectIntEQ(wolfSSL_get_error(ssl_c2, -1), WOLFSSL_ERROR_WANT_READ);
+
+        /* Read second ClientHello, send HRR now with WOLFSSL_ECC_SECP384R1 */
+        wolfSSL_SetLoggingPrefix("server");
+        ExpectIntEQ(wolfDTLS_accept_stateless(ssl_s), 0);
+        ExpectIntEQ(initHash, test_wolfSSL_dtls_stateless_HashWOLFSSL(ssl_s));
+        test_memio_clear_buffer(&test_ctx, 1);
+
+        /* Complete first handshake with WOLFSSL_ECC_SECP256R1 */
+        wolfSSL_SetLoggingPrefix("client1");
+        ExpectIntEQ(test_memio_inject_message(&test_ctx, 1, hrrBuf, hrrSz), 0);
+        ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
+        ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+
+        wolfSSL_SetLoggingPrefix("server");
+        ExpectIntEQ(wolfDTLS_accept_stateless(ssl_s), WOLFSSL_SUCCESS);
+
+        ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+
+        wolfSSL_free(ssl_s);
+        wolfSSL_free(ssl_c);
+        wolfSSL_free(ssl_c2);
+        wolfSSL_CTX_free(ctx_s);
+        wolfSSL_CTX_free(ctx_c);
+    }
+#endif /* WOLFSSL_SEND_HRR_COOKIE */
+    return EXPECT_RESULT();
+}
+#else
+int test_wolfSSL_dtls_stateless(void)
+{
+    return TEST_SKIPPED;
+}
+
+int test_wolfSSL_dtls_stateless_hrr_group(void)
+{
+    return TEST_SKIPPED;
+}
+#endif /* WOLFSSL_DTLS13 && WOLFSSL_SEND_HRR_COOKIE &&
+        * HAVE_IO_TESTS_DEPENDENCIES && !SINGLE_THREADED */
+
+/*----------------------------------------------------------------------------*/
+/* DTLS window updates, fragment buckets, stateless resume/downgrade, alerts  */
+/*----------------------------------------------------------------------------*/
+
+#ifdef WOLFSSL_DTLS
+
+/* Prints out the current window */
+static void DUW_TEST_print_window_binary(word32 h, word32 l, word32* w) {
+#ifdef WOLFSSL_DEBUG_DTLS_WINDOW
+    int i;
+    for (i = WOLFSSL_DTLS_WINDOW_WORDS - 1; i >= 0; i--) {
+        word32 b = w[i];
+        int j;
+        /* Prints out a 32 bit binary number in big endian order */
+        for (j = 0; j < 32; j++, b <<= 1) {
+            if (b & (((word32)1) << 31))
+                fprintf(stderr, "1");
+            else
+                fprintf(stderr, "0");
+        }
+        fprintf(stderr, " ");
+    }
+    fprintf(stderr, "cur_hi %u cur_lo %u\n", h, l);
+#else
+    (void)h;
+    (void)l;
+    (void)w;
+#endif
+}
+
+/* a - cur_hi
+ * b - cur_lo
+ * c - next_hi
+ * d - next_lo
+ * e - window
+ * f - expected next_hi
+ * g - expected next_lo
+ * h - expected window[1]
+ * i - expected window[0]
+ */
+#define DUW_TEST(a,b,c,d,e,f,g,h,i) do { \
+    ExpectIntEQ(wolfSSL_DtlsUpdateWindow((a), (b), &(c), &(d), (e)), 1); \
+    DUW_TEST_print_window_binary((a), (b), (e)); \
+    ExpectIntEQ((c), (f)); \
+    ExpectIntEQ((d), (g)); \
+    ExpectIntEQ((e)[1], (h)); \
+    ExpectIntEQ((e)[0], (i)); \
+} while (0)
+
+int test_wolfSSL_DtlsUpdateWindow(void)
+{
+    EXPECT_DECLS;
+    word32 window[WOLFSSL_DTLS_WINDOW_WORDS];
+    word32 next_lo = 0;
+    word16 next_hi = 0;
+
+#ifdef WOLFSSL_DEBUG_DTLS_WINDOW
+    fprintf(stderr, "\n");
+#endif
+
+    XMEMSET(window, 0, sizeof window);
+    DUW_TEST(0, 0, next_hi, next_lo, window, 0, 1, 0, 0x01);
+    DUW_TEST(0, 1, next_hi, next_lo, window, 0, 2, 0, 0x03);
+    DUW_TEST(0, 5, next_hi, next_lo, window, 0, 6, 0, 0x31);
+    DUW_TEST(0, 4, next_hi, next_lo, window, 0, 6, 0, 0x33);
+    DUW_TEST(0, 100, next_hi, next_lo, window, 0, 101, 0, 0x01);
+    DUW_TEST(0, 101, next_hi, next_lo, window, 0, 102, 0, 0x03);
+    DUW_TEST(0, 133, next_hi, next_lo, window, 0, 134, 0x03, 0x01);
+    DUW_TEST(0, 200, next_hi, next_lo, window, 0, 201, 0, 0x01);
+    DUW_TEST(0, 264, next_hi, next_lo, window, 0, 265, 0, 0x01);
+    DUW_TEST(0, 0xFFFFFFFF, next_hi, next_lo, window, 1, 0, 0, 0x01);
+    DUW_TEST(0, 0xFFFFFFFD, next_hi, next_lo, window, 1, 0, 0, 0x05);
+    DUW_TEST(0, 0xFFFFFFFE, next_hi, next_lo, window, 1, 0, 0, 0x07);
+    DUW_TEST(1, 3, next_hi, next_lo, window, 1, 4, 0, 0x71);
+    DUW_TEST(1, 0, next_hi, next_lo, window, 1, 4, 0, 0x79);
+    DUW_TEST(1, 0xFFFFFFFF, next_hi, next_lo, window, 2, 0, 0, 0x01);
+    DUW_TEST(2, 3, next_hi, next_lo, window, 2, 4, 0, 0x11);
+    DUW_TEST(2, 0, next_hi, next_lo, window, 2, 4, 0, 0x19);
+    DUW_TEST(2, 25, next_hi, next_lo, window, 2, 26, 0, 0x6400001);
+    DUW_TEST(2, 27, next_hi, next_lo, window, 2, 28, 0, 0x19000005);
+    DUW_TEST(2, 29, next_hi, next_lo, window, 2, 30, 0, 0x64000015);
+    DUW_TEST(2, 33, next_hi, next_lo, window, 2, 34, 6, 0x40000151);
+    DUW_TEST(2, 60, next_hi, next_lo, window, 2, 61, 0x3200000A, 0x88000001);
+    DUW_TEST(1, 0xFFFFFFF0, next_hi, next_lo, window, 2, 61, 0x3200000A, 0x88000001);
+    DUW_TEST(2, 0xFFFFFFFD, next_hi, next_lo, window, 2, 0xFFFFFFFE, 0, 0x01);
+    DUW_TEST(3, 1, next_hi, next_lo, window, 3, 2, 0, 0x11);
+    DUW_TEST(99, 66, next_hi, next_lo, window, 99, 67, 0, 0x01);
+    DUW_TEST(50, 66, next_hi, next_lo, window, 99, 67, 0, 0x01);
+    DUW_TEST(100, 68, next_hi, next_lo, window, 100, 69, 0, 0x01);
+    DUW_TEST(99, 50, next_hi, next_lo, window, 100, 69, 0, 0x01);
+    DUW_TEST(99, 0xFFFFFFFF, next_hi, next_lo, window, 100, 69, 0, 0x01);
+    DUW_TEST(150, 0xFFFFFFFF, next_hi, next_lo, window, 151, 0, 0, 0x01);
+    DUW_TEST(152, 0xFFFFFFFF, next_hi, next_lo, window, 153, 0, 0, 0x01);
+
+    return EXPECT_RESULT();
+}
+#else
+int test_wolfSSL_DtlsUpdateWindow(void)
+{
+    return TEST_SKIPPED;
+}
+#endif /* WOLFSSL_DTLS */
+
+#ifdef WOLFSSL_DTLS
+static int DFB_TEST(WOLFSSL* ssl, word32 seq, word32 len, word32 f_offset,
+        word32 f_len, word32 f_count, byte ready, word32 bytesReceived)
+{
+    EXPECT_DECLS;
+    DtlsMsg* cur = NULL;
+    static byte msg[100];
+    static byte msgInit = 0;
+
+    if (!msgInit) {
+        int i;
+        for (i = 0; i < 100; i++)
+            msg[i] = i + 1;
+        msgInit = 1;
+    }
+
+    /* Sanitize test parameters */
+    ExpectIntLE(len, sizeof(msg));
+    ExpectIntLE(f_offset + f_len, sizeof(msg));
+
+    if (EXPECT_SUCCESS())
+        DtlsMsgStore(ssl, 0, seq, msg + f_offset, len, certificate, f_offset, f_len, NULL);
+
+    ExpectNotNull(ssl->dtls_rx_msg_list);
+
+    ExpectNotNull(cur = DtlsMsgFind(ssl->dtls_rx_msg_list, 0, seq));
+    ExpectIntEQ(cur->fragBucketListCount, f_count);
+    ExpectIntEQ(cur->ready, ready);
+    ExpectIntEQ(cur->bytesReceived, bytesReceived);
+    if (ready) {
+        ExpectNull(cur->fragBucketList);
+        ExpectBufEQ(cur->fullMsg, msg, cur->sz);
+    }
+    else {
+        DtlsFragBucket* fb;
+        ExpectNotNull(cur->fragBucketList);
+        for (fb = cur != NULL ? cur->fragBucketList : NULL;
+                EXPECT_SUCCESS() && fb != NULL; fb = fb->m.m.next)
+            ExpectBufEQ(fb->buf, msg + fb->m.m.offset, fb->m.m.sz);
+    }
+    if (EXPECT_FAIL()) {
+        printf("Test parameters: seq %u len %u f_offset %u f_len %u f_count %u ready %u bytesReceived %u\n",
+                                 seq, len, f_offset, f_len, f_count, ready, bytesReceived);
+    }
+    return EXPECT_RESULT();
+}
+
+int test_wolfSSL_DTLS_fragment_buckets(void)
+{
+    EXPECT_DECLS;
+    WOLFSSL ssl[1];
+
+    XMEMSET(ssl, 0, sizeof(*ssl));
+
+    EXPECT_TEST(DFB_TEST(ssl, 0, 100, 0, 100, 0, 1, 100)); /*  0-100 */
+
+    EXPECT_TEST(DFB_TEST(ssl, 1, 100,  0, 20, 1, 0,  20)); /*  0-20  */
+    EXPECT_TEST(DFB_TEST(ssl, 1, 100, 20, 20, 1, 0,  40)); /* 20-40  */
+    EXPECT_TEST(DFB_TEST(ssl, 1, 100, 40, 20, 1, 0,  60)); /* 40-60  */
+    EXPECT_TEST(DFB_TEST(ssl, 1, 100, 60, 20, 1, 0,  80)); /* 60-80  */
+    EXPECT_TEST(DFB_TEST(ssl, 1, 100, 80, 20, 0, 1, 100)); /* 80-100 */
+
+    /* Test all permutations of 3 regions */
+    /* 1 2 3 */
+    EXPECT_TEST(DFB_TEST(ssl, 2, 100,  0, 30, 1, 0,  30)); /*  0-30  */
+    EXPECT_TEST(DFB_TEST(ssl, 2, 100, 30, 30, 1, 0,  60)); /* 30-60  */
+    EXPECT_TEST(DFB_TEST(ssl, 2, 100, 60, 40, 0, 1, 100)); /* 60-100 */
+    /* 1 3 2 */
+    EXPECT_TEST(DFB_TEST(ssl, 3, 100,  0, 30, 1, 0,  30)); /*  0-30  */
+    EXPECT_TEST(DFB_TEST(ssl, 3, 100, 60, 40, 2, 0,  70)); /* 60-100 */
+    EXPECT_TEST(DFB_TEST(ssl, 3, 100, 30, 30, 0, 1, 100)); /* 30-60  */
+    /* 2 1 3 */
+    EXPECT_TEST(DFB_TEST(ssl, 4, 100, 30, 30, 1, 0,  30)); /* 30-60  */
+    EXPECT_TEST(DFB_TEST(ssl, 4, 100,  0, 30, 1, 0,  60)); /*  0-30  */
+    EXPECT_TEST(DFB_TEST(ssl, 4, 100, 60, 40, 0, 1, 100)); /* 60-100 */
+    /* 2 3 1 */
+    EXPECT_TEST(DFB_TEST(ssl, 5, 100, 30, 30, 1, 0,  30)); /* 30-60  */
+    EXPECT_TEST(DFB_TEST(ssl, 5, 100, 60, 40, 1, 0,  70)); /* 60-100 */
+    EXPECT_TEST(DFB_TEST(ssl, 5, 100,  0, 30, 0, 1, 100)); /*  0-30  */
+    /* 3 1 2 */
+    EXPECT_TEST(DFB_TEST(ssl, 6, 100, 60, 40, 1, 0,  40)); /* 60-100 */
+    EXPECT_TEST(DFB_TEST(ssl, 6, 100,  0, 30, 2, 0,  70)); /*  0-30  */
+    EXPECT_TEST(DFB_TEST(ssl, 6, 100, 30, 30, 0, 1, 100)); /* 30-60  */
+    /* 3 2 1 */
+    EXPECT_TEST(DFB_TEST(ssl, 7, 100, 60, 40, 1, 0,  40)); /* 60-100 */
+    EXPECT_TEST(DFB_TEST(ssl, 7, 100, 30, 30, 1, 0,  70)); /* 30-60  */
+    EXPECT_TEST(DFB_TEST(ssl, 7, 100,  0, 30, 0, 1, 100)); /*  0-30  */
+
+    /* Test overlapping regions */
+    EXPECT_TEST(DFB_TEST(ssl, 8, 100,  0, 30, 1, 0,  30)); /*  0-30  */
+    EXPECT_TEST(DFB_TEST(ssl, 8, 100, 20, 10, 1, 0,  30)); /* 20-30  */
+    EXPECT_TEST(DFB_TEST(ssl, 8, 100, 70, 10, 2, 0,  40)); /* 70-80  */
+    EXPECT_TEST(DFB_TEST(ssl, 8, 100, 20, 30, 2, 0,  60)); /* 20-50  */
+    EXPECT_TEST(DFB_TEST(ssl, 8, 100, 40, 60, 0, 1, 100)); /* 40-100 */
+
+    /* Test overlapping multiple regions */
+    EXPECT_TEST(DFB_TEST(ssl, 9, 100,  0, 20, 1, 0,  20)); /*  0-20  */
+    EXPECT_TEST(DFB_TEST(ssl, 9, 100, 30,  5, 2, 0,  25)); /* 30-35  */
+    EXPECT_TEST(DFB_TEST(ssl, 9, 100, 40,  5, 3, 0,  30)); /* 40-45  */
+    EXPECT_TEST(DFB_TEST(ssl, 9, 100, 50,  5, 4, 0,  35)); /* 50-55  */
+    EXPECT_TEST(DFB_TEST(ssl, 9, 100, 60,  5, 5, 0,  40)); /* 60-65  */
+    EXPECT_TEST(DFB_TEST(ssl, 9, 100, 70,  5, 6, 0,  45)); /* 70-75  */
+    EXPECT_TEST(DFB_TEST(ssl, 9, 100, 30, 25, 4, 0,  55)); /* 30-55  */
+    EXPECT_TEST(DFB_TEST(ssl, 9, 100, 55, 15, 2, 0,  65)); /* 55-70  */
+    EXPECT_TEST(DFB_TEST(ssl, 9, 100, 75, 25, 2, 0,  90)); /* 75-100 */
+    EXPECT_TEST(DFB_TEST(ssl, 9, 100, 10, 25, 0, 1, 100)); /* 10-35 */
+
+    EXPECT_TEST(DFB_TEST(ssl,10, 100,  0, 20, 1, 0,  20)); /*  0-20  */
+    EXPECT_TEST(DFB_TEST(ssl,10, 100, 30, 20, 2, 0,  40)); /* 30-50  */
+    EXPECT_TEST(DFB_TEST(ssl,10, 100,  0, 40, 1, 0,  50)); /*  0-40  */
+    EXPECT_TEST(DFB_TEST(ssl,10, 100, 50, 50, 0, 1, 100)); /* 50-100 */
+
+    /* Test region between other regions */
+    EXPECT_TEST(DFB_TEST(ssl,11, 100,  0, 20, 1, 0,  20)); /*  0-20  */
+    EXPECT_TEST(DFB_TEST(ssl,11, 100, 80, 20, 2, 0,  40)); /* 80-100 */
+    EXPECT_TEST(DFB_TEST(ssl,11, 100, 40, 20, 3, 0,  60)); /* 40-60  */
+    EXPECT_TEST(DFB_TEST(ssl,11, 100, 20, 20, 2, 0,  80)); /* 20-40  */
+    EXPECT_TEST(DFB_TEST(ssl,11, 100, 60, 20, 0, 1, 100)); /* 60-80  */
+
+    /* Test gap before first bucket (prev==NULL in gap-before branch) */
+    EXPECT_TEST(DFB_TEST(ssl,12, 100, 50, 20, 1, 0,  20)); /* 50-70  */
+    EXPECT_TEST(DFB_TEST(ssl,12, 100,  0, 20, 2, 0,  40)); /*  0-20  gap before first */
+    EXPECT_TEST(DFB_TEST(ssl,12, 100, 20, 30, 1, 0,  70)); /* 20-50  bridges gap */
+    EXPECT_TEST(DFB_TEST(ssl,12, 100, 70, 30, 0, 1, 100)); /* 70-100 */
+
+    /* Test fragment after message is already complete (ready early return) */
+    EXPECT_TEST(DFB_TEST(ssl,13, 100,  0,100, 0, 1, 100)); /*  0-100 complete */
+    EXPECT_TEST(DFB_TEST(ssl,13, 100,  0, 50, 0, 1, 100)); /*  0-50  dup on ready */
+
+    /* Test combine where next bucket is larger than cur (chosenBucket=&next) */
+    EXPECT_TEST(DFB_TEST(ssl,14, 100,  0, 10, 1, 0,  10)); /*  0-10  */
+    EXPECT_TEST(DFB_TEST(ssl,14, 100, 30, 50, 2, 0,  60)); /* 30-80  */
+    EXPECT_TEST(DFB_TEST(ssl,14, 100,  5, 30, 1, 0,  80)); /*  5-35  next>cur */
+    EXPECT_TEST(DFB_TEST(ssl,14, 100, 80, 20, 0, 1, 100)); /* 80-100 */
+
+    /* Test super fragment covering all existing buckets */
+    EXPECT_TEST(DFB_TEST(ssl,15, 100, 10, 10, 1, 0,  10)); /* 10-20  */
+    EXPECT_TEST(DFB_TEST(ssl,15, 100, 30, 10, 2, 0,  20)); /* 30-40  */
+    EXPECT_TEST(DFB_TEST(ssl,15, 100, 60, 10, 3, 0,  30)); /* 60-70  */
+    EXPECT_TEST(DFB_TEST(ssl,15, 100,  0,100, 0, 1, 100)); /*  0-100 super frag */
+
+    /* Test exact duplicate fragment */
+    EXPECT_TEST(DFB_TEST(ssl,16, 100, 20, 40, 1, 0,  40)); /* 20-60  */
+    EXPECT_TEST(DFB_TEST(ssl,16, 100, 20, 40, 1, 0,  40)); /* 20-60  exact dup */
+    EXPECT_TEST(DFB_TEST(ssl,16, 100,  0, 20, 1, 0,  60)); /*  0-20  */
+    EXPECT_TEST(DFB_TEST(ssl,16, 100, 60, 40, 0, 1, 100)); /* 60-100 */
+
+    /* Test combine bridging two buckets (combineNext, cur->data) */
+    EXPECT_TEST(DFB_TEST(ssl,17, 100,  0, 30, 1, 0,  30)); /*  0-30  */
+    EXPECT_TEST(DFB_TEST(ssl,17, 100, 60, 20, 2, 0,  50)); /* 60-80  */
+    EXPECT_TEST(DFB_TEST(ssl,17, 100, 20, 45, 1, 0,  80)); /* 20-65  bridge */
+    EXPECT_TEST(DFB_TEST(ssl,17, 100, 80, 20, 0, 1, 100)); /* 80-100 */
+
+    /* Test progressive left-extension with partial overlaps */
+    EXPECT_TEST(DFB_TEST(ssl,18, 100, 70, 30, 1, 0,  30)); /* 70-100 */
+    EXPECT_TEST(DFB_TEST(ssl,18, 100, 50, 30, 1, 0,  50)); /* 50-80  extend left */
+    EXPECT_TEST(DFB_TEST(ssl,18, 100, 30, 30, 1, 0,  70)); /* 30-60  extend left */
+    EXPECT_TEST(DFB_TEST(ssl,18, 100,  0, 40, 0, 1, 100)); /*  0-40  complete left */
+
+    DtlsMsgListDelete(ssl->dtls_rx_msg_list, ssl->heap);
+    ssl->dtls_rx_msg_list = NULL;
+    ssl->dtls_rx_msg_list_sz = 0;
+    return EXPECT_RESULT();
+}
+
+#else
+int test_wolfSSL_DTLS_fragment_buckets(void)
+{
+    return TEST_SKIPPED;
+}
+#endif
+
+
+#if !defined(NO_FILESYSTEM) && \
+     defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) && \
+     defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && !defined(NO_RSA)
+
+int test_wolfSSL_dtls_stateless2(void)
+{
+    EXPECT_DECLS;
+    WOLFSSL *ssl_c = NULL;
+    WOLFSSL *ssl_c2 = NULL;
+    WOLFSSL *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+    WOLFSSL_CTX *ctx_c = NULL;
+    WOLFSSL_CTX *ctx_s = NULL;
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method), 0);
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, NULL, &ssl_c2, NULL,
+        wolfDTLSv1_2_client_method, NULL), 0);
+    ExpectFalse(wolfSSL_is_stateful(ssl_s));
+    /* send CH */
+    ExpectTrue((wolfSSL_connect(ssl_c2) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
+        (ssl_c2->error == WC_NO_ERR_TRACE(WANT_READ)));
+    ExpectIntEQ(wolfDTLS_accept_stateless(ssl_s), WOLFSSL_FAILURE);
+    ExpectFalse(wolfSSL_is_stateful(ssl_s));
+    ExpectIntNE(test_ctx.c_len, 0);
+    /* consume HRR */
+    test_memio_clear_buffer(&test_ctx, 1);
+    /* send CH1 */
+    ExpectIntEQ(wolfSSL_connect(ssl_c), WOLFSSL_FATAL_ERROR);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, WOLFSSL_FATAL_ERROR),
+            WOLFSSL_ERROR_WANT_READ);
+    /* send HRR */
+    ExpectIntEQ(wolfDTLS_accept_stateless(ssl_s), WOLFSSL_FAILURE);
+    /* send CH2 */
+    ExpectIntEQ(wolfSSL_connect(ssl_c), WOLFSSL_FATAL_ERROR);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, WOLFSSL_FATAL_ERROR),
+            WOLFSSL_ERROR_WANT_READ);
+    /* send HRR */
+    ExpectIntEQ(wolfDTLS_accept_stateless(ssl_s), WOLFSSL_SUCCESS);
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+    ExpectTrue(wolfSSL_is_stateful(ssl_s));
+
+    wolfSSL_free(ssl_c2);
+    wolfSSL_free(ssl_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_CTX_free(ctx_s);
+    return EXPECT_RESULT();
+}
+
+int test_wolfSSL_dtls_stateless_maxfrag(void)
+{
+    EXPECT_DECLS;
+#ifdef HAVE_MAX_FRAGMENT
+    WOLFSSL *ssl_c = NULL;
+    WOLFSSL *ssl_c2 = NULL;
+    WOLFSSL *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+    WOLFSSL_CTX *ctx_c = NULL;
+    WOLFSSL_CTX *ctx_s = NULL;
+    word16 max_fragment = 0;
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method), 0);
+    ExpectNotNull(ssl_s);
+    ExpectNotNull(ssl_c2 = wolfSSL_new(ctx_c));
+    ExpectIntEQ(wolfSSL_UseMaxFragment(ssl_c2, WOLFSSL_MFL_2_8),
+        WOLFSSL_SUCCESS);
+    wolfSSL_SetIOWriteCtx(ssl_c2, &test_ctx);
+    wolfSSL_SetIOReadCtx(ssl_c2, &test_ctx);
+    if (EXPECT_SUCCESS()) {
+        max_fragment = ssl_s->max_fragment;
+    }
+    /* send CH */
+    ExpectTrue((wolfSSL_connect(ssl_c2) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
+        (ssl_c2->error == WC_NO_ERR_TRACE(WANT_READ)));
+    ExpectTrue((wolfSSL_accept(ssl_s) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
+        (ssl_s->error == WC_NO_ERR_TRACE(WANT_READ)));
+    /* CH without cookie shouldn't change state */
+    ExpectIntEQ(ssl_s->max_fragment, max_fragment);
+    ExpectIntNE(test_ctx.c_len, 0);
+
+    /* consume HRR from buffer */
+    test_memio_clear_buffer(&test_ctx, 1);
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+
+    wolfSSL_free(ssl_c2);
+    wolfSSL_free(ssl_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_CTX_free(ctx_s);
+#endif /* HAVE_MAX_FRAGMENT */
+    return EXPECT_RESULT();
+}
+
+#if defined(WOLFSSL_DTLS_NO_HVR_ON_RESUME)
+#define ROUNDS_WITH_HVR 4
+#define ROUNDS_WITHOUT_HVR 2
+#define HANDSHAKE_TYPE_OFFSET DTLS_RECORD_HEADER_SZ
+static int buf_is_hvr(const byte *data, int len)
+{
+    if (len < DTLS_RECORD_HEADER_SZ + DTLS_HANDSHAKE_HEADER_SZ)
+        return 0;
+    return data[HANDSHAKE_TYPE_OFFSET] == hello_verify_request;
+}
+
+static int _test_wolfSSL_dtls_stateless_resume(byte useticket, byte bad)
+{
+    EXPECT_DECLS;
+    struct test_memio_ctx test_ctx;
+    WOLFSSL_CTX *ctx_c = NULL;
+    WOLFSSL_CTX *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL;
+    WOLFSSL *ssl_s = NULL;
+    WOLFSSL_SESSION *sess = NULL;
+    int round_trips;
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c,
+        &ssl_s, wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method), 0);
+#ifdef HAVE_SESSION_TICKET
+    if (useticket) {
+        ExpectIntEQ(wolfSSL_UseSessionTicket(ssl_c), WOLFSSL_SUCCESS);
+    }
+#endif
+    round_trips = ROUNDS_WITH_HVR;
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, round_trips,
+        &round_trips), 0);
+    ExpectIntEQ(round_trips, ROUNDS_WITH_HVR);
+    ExpectNotNull(sess = wolfSSL_get1_session(ssl_c));
+    wolfSSL_shutdown(ssl_c);
+    wolfSSL_shutdown(ssl_s);
+    wolfSSL_free(ssl_c);
+    ssl_c = NULL;
+    wolfSSL_free(ssl_s);
+    ssl_s = NULL;
+
+    test_memio_clear_buffer(&test_ctx, 1);
+    test_memio_clear_buffer(&test_ctx, 0);
+    /* make resumption invalid */
+    if (bad && (sess != NULL)) {
+        if (useticket) {
+#ifdef HAVE_SESSION_TICKET
+            if (sess->ticket != NULL) {
+                sess->ticket[0] = !sess->ticket[0];
+            }
+#endif /* HAVE_SESSION_TICKET */
+        }
+        else {
+            sess->sessionID[0] = !sess->sessionID[0];
+        }
+    }
+    ExpectNotNull(ssl_c = wolfSSL_new(ctx_c));
+    ExpectNotNull(ssl_s = wolfSSL_new(ctx_s));
+    wolfSSL_SetIOWriteCtx(ssl_c, &test_ctx);
+    wolfSSL_SetIOReadCtx(ssl_c, &test_ctx);
+    wolfSSL_SetIOWriteCtx(ssl_s, &test_ctx);
+    wolfSSL_SetIOReadCtx(ssl_s, &test_ctx);
+    ExpectIntEQ(wolfSSL_set_session(ssl_c, sess), WOLFSSL_SUCCESS);
+    ExpectTrue((wolfSSL_connect(ssl_c) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
+        (ssl_c->error == WC_NO_ERR_TRACE(WANT_READ)));
+    ExpectTrue((wolfSSL_accept(ssl_s) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
+        (ssl_s->error == WC_NO_ERR_TRACE(WANT_READ)));
+    ExpectFalse(bad && !buf_is_hvr(test_ctx.c_buff, test_ctx.c_len));
+    ExpectFalse(!bad && buf_is_hvr(test_ctx.c_buff, test_ctx.c_len));
+    if (!useticket) {
+        ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, &round_trips), 0);
+        ExpectFalse(bad && round_trips != ROUNDS_WITH_HVR - 1);
+        ExpectFalse(!bad && round_trips != ROUNDS_WITHOUT_HVR - 1);
+    }
+    wolfSSL_SESSION_free(sess);
+    wolfSSL_free(ssl_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_CTX_free(ctx_s);
+    return EXPECT_RESULT();
+}
+
+int test_wolfSSL_dtls_stateless_resume(void)
+{
+    EXPECT_DECLS;
+#ifdef HAVE_SESSION_TICKET
+    ExpectIntEQ(_test_wolfSSL_dtls_stateless_resume(1, 0), TEST_SUCCESS);
+    ExpectIntEQ(_test_wolfSSL_dtls_stateless_resume(1, 1), TEST_SUCCESS);
+#endif /* HAVE_SESION_TICKET */
+    ExpectIntEQ(_test_wolfSSL_dtls_stateless_resume(0, 0), TEST_SUCCESS);
+    ExpectIntEQ(_test_wolfSSL_dtls_stateless_resume(0, 1), TEST_SUCCESS);
+    return EXPECT_RESULT();
+}
+#else
+int test_wolfSSL_dtls_stateless_resume(void)
+{
+    return TEST_SKIPPED;
+}
+#endif /* WOLFSSL_DTLS_NO_HVR_ON_RESUME */
+
+int test_wolfSSL_dtls_stateless_downgrade(void)
+{
+    EXPECT_DECLS;
+#if !defined(NO_OLD_TLS)
+    WOLFSSL_CTX *ctx_c = NULL;
+    WOLFSSL_CTX *ctx_c2 = NULL;
+    WOLFSSL_CTX *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL;
+    WOLFSSL *ssl_c2 = NULL;
+    WOLFSSL *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method), 0);
+    ExpectIntEQ(wolfSSL_CTX_SetMinVersion(ctx_s, WOLFSSL_DTLSV1),
+        WOLFSSL_SUCCESS);
+    ExpectNotNull(ctx_c2 = wolfSSL_CTX_new(wolfDTLSv1_client_method()));
+    wolfSSL_SetIORecv(ctx_c2, test_memio_read_cb);
+    wolfSSL_SetIOSend(ctx_c2, test_memio_write_cb);
+    ExpectNotNull(ssl_c2 = wolfSSL_new(ctx_c2));
+    wolfSSL_SetIOWriteCtx(ssl_c2, &test_ctx);
+    wolfSSL_SetIOReadCtx(ssl_c2, &test_ctx);
+    /* send CH */
+    ExpectTrue((wolfSSL_connect(ssl_c2) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
+        (ssl_c2->error == WC_NO_ERR_TRACE(WANT_READ)));
+    ExpectTrue((wolfSSL_accept(ssl_s) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
+        (ssl_s->error == WC_NO_ERR_TRACE(WANT_READ)));
+    ExpectIntNE(test_ctx.c_len, 0);
+    /* consume HRR */
+    test_memio_clear_buffer(&test_ctx, 1);
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+
+    wolfSSL_free(ssl_c2);
+    wolfSSL_free(ssl_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_CTX_free(ctx_c2);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
+
+#else /* outer wrap: WOLFSSL_DTLS && !WOLFSSL_NO_TLS12 && client/server && !NO_RSA */
+int test_wolfSSL_dtls_stateless2(void)
+{
+    return TEST_SKIPPED;
+}
+int test_wolfSSL_dtls_stateless_maxfrag(void)
+{
+    return TEST_SKIPPED;
+}
+int test_wolfSSL_dtls_stateless_resume(void)
+{
+    return TEST_SKIPPED;
+}
+int test_wolfSSL_dtls_stateless_downgrade(void)
+{
+    return TEST_SKIPPED;
+}
+#endif /* defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) && \
+    !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER)*/
+
+#if defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) && \
+    !defined(NO_OLD_TLS) && defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES)
+int test_WOLFSSL_dtls_version_alert(void)
+{
+    EXPECT_DECLS;
+    struct test_memio_ctx test_ctx;
+    WOLFSSL_CTX *ctx_c = NULL;
+    WOLFSSL_CTX *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL;
+    WOLFSSL *ssl_s = NULL;
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_2_client_method, wolfDTLSv1_server_method), 0);
+
+    /* client hello */
+    ExpectTrue((wolfSSL_connect(ssl_c) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
+        (ssl_c->error == WC_NO_ERR_TRACE(WANT_READ)));
+    /* hrr */
+    ExpectTrue((wolfSSL_accept(ssl_s) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
+        (ssl_s->error == WC_NO_ERR_TRACE(WANT_READ)));
+    /* client hello 1 */
+    ExpectTrue((wolfSSL_connect(ssl_c) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
+        (ssl_c->error == WC_NO_ERR_TRACE(WANT_READ)));
+    /* server hello */
+    ExpectTrue((wolfSSL_accept(ssl_s) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
+        (ssl_s->error == WC_NO_ERR_TRACE(WANT_READ)));
+    /* should fail */
+    ExpectTrue((wolfSSL_connect(ssl_c) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
+        (ssl_c->error == WC_NO_ERR_TRACE(VERSION_ERROR)));
+    /* shuould fail */
+    ExpectTrue((wolfSSL_accept(ssl_s) == WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)) &&
+        (ssl_s->error == WC_NO_ERR_TRACE(VERSION_ERROR) || ssl_s->error == WC_NO_ERR_TRACE(FATAL_ERROR)));
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_CTX_free(ctx_s);
+
+    return EXPECT_RESULT();
+}
+#else
+int test_WOLFSSL_dtls_version_alert(void)
+{
+    return TEST_SKIPPED;
+}
+#endif /* defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) &&
+        * !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER) &&
+        * !defined(NO_OLD_TLS) && !defined(NO_RSA)
+        */
+
+/*----------------------------------------------------------------------------*/
+/* Remaining DTLS tests moved out of api.c (msg, ipv6, downgrade, ccs, etc.)  */
+/*----------------------------------------------------------------------------*/
+
+/*-- msg_helpers_and_from_other_peer (api.c lines 30572,30671) ---*/
+#if defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) &&          \
+    !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER) &&   \
+    !defined(SINGLE_THREADED) && !defined(NO_RSA)
+
+static int test_dtls_msg_get_connected_port(int fd, word16 *port)
+{
+    SOCKADDR_S peer;
+    XSOCKLENT len;
+    int ret;
+
+    XMEMSET((byte*)&peer, 0, sizeof(peer));
+    len = sizeof(peer);
+    ret = getpeername(fd,  (SOCKADDR*)&peer, &len);
+    if (ret != 0 || len > (XSOCKLENT)sizeof(peer))
+        return -1;
+    switch (peer.ss_family) {
+#ifdef WOLFSSL_IPV6
+    case WOLFSSL_IP6: {
+        *port = ntohs(((SOCKADDR_IN6*)&peer)->sin6_port);
+        break;
+    }
+#endif /* WOLFSSL_IPV6 */
+    case WOLFSSL_IP4:
+        *port = ntohs(((SOCKADDR_IN*)&peer)->sin_port);
+        break;
+    default:
+        return -1;
+    }
+    return 0;
+}
+
+static int test_dtls_msg_from_other_peer_cb(WOLFSSL_CTX *ctx, WOLFSSL *ssl)
+{
+    char buf[1] = {'t'};
+    SOCKADDR_IN_T addr;
+    int sock_fd;
+    word16 port;
+    int err;
+
+    (void)ssl;
+    (void)ctx;
+
+    if (ssl == NULL)
+        return -1;
+
+    err = test_dtls_msg_get_connected_port(wolfSSL_get_fd(ssl), &port);
+    if (err != 0)
+        return -1;
+
+    sock_fd = socket(AF_INET_V, SOCK_DGRAM, 0);
+    if (sock_fd == -1)
+        return -1;
+    build_addr(&addr, wolfSSLIP, port, 1, 0);
+
+    /* send a packet to the server. Being another socket, the kernel will ensure
+     * the source port will be different. */
+    err = (int)sendto(sock_fd, buf, sizeof(buf), 0, (SOCKADDR*)&addr,
+        sizeof(addr));
+
+    close(sock_fd);
+    if (err == -1)
+        return -1;
+
+    return 0;
+}
+
+/* setup a SSL session but just after the handshake send a packet to the server
+ * with a source address different than the one of the connected client. The I/O
+ * callback EmbedRecvFrom should just ignore the packet. Sending of the packet
+ * is done in test_dtls_msg_from_other_peer_cb */
+int test_dtls_msg_from_other_peer(void)
+{
+    EXPECT_DECLS;
+    callback_functions client_cbs;
+    callback_functions server_cbs;
+
+    XMEMSET((byte*)&client_cbs, 0, sizeof(client_cbs));
+    XMEMSET((byte*)&server_cbs, 0, sizeof(server_cbs));
+
+    client_cbs.method = wolfDTLSv1_2_client_method;
+    server_cbs.method = wolfDTLSv1_2_server_method;
+    client_cbs.doUdp = 1;
+    server_cbs.doUdp = 1;
+
+    test_wolfSSL_client_server_nofail_ex(&client_cbs, &server_cbs,
+        test_dtls_msg_from_other_peer_cb);
+
+    ExpectIntEQ(client_cbs.return_code, WOLFSSL_SUCCESS);
+    ExpectIntEQ(server_cbs.return_code, WOLFSSL_SUCCESS);
+
+    return EXPECT_RESULT();
+}
+#else
+int test_dtls_msg_from_other_peer(void)
+{
+    return TEST_SKIPPED;
+}
+#endif /* defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) &&          \
+        *  !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER) &&  \
+        *  !defined(SINGLE_THREADED) && !defined(NO_RSA) */
+
+/*-- ipv6_check (api.c lines 30672,30730) ---*/
+#if defined(WOLFSSL_DTLS) && !defined(WOLFSSL_IPV6) &&               \
+    !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER) &&   \
+    defined(HAVE_IO_TESTS_DEPENDENCIES) && !defined(WOLFSSL_NO_TLS12) \
+    && !defined(USE_WINDOWS_API)
+int test_dtls_ipv6_check(void)
+{
+    EXPECT_DECLS;
+    WOLFSSL_CTX *ctx_c = NULL;
+    WOLFSSL_CTX *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL;
+    WOLFSSL *ssl_s = NULL;
+    SOCKADDR_IN fake_addr6;
+    int sockfd = -1;
+
+    ExpectNotNull(ctx_c = wolfSSL_CTX_new(wolfDTLSv1_2_client_method()));
+    ExpectNotNull(ssl_c  = wolfSSL_new(ctx_c));
+    ExpectNotNull(ctx_s = wolfSSL_CTX_new(wolfDTLSv1_2_server_method()));
+    ExpectIntEQ(wolfSSL_CTX_use_PrivateKey_file(ctx_s, svrKeyFile,
+        WOLFSSL_FILETYPE_PEM), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_CTX_use_certificate_file(ctx_s, svrCertFile,
+        WOLFSSL_FILETYPE_PEM), WOLFSSL_SUCCESS);
+    ExpectNotNull(ssl_s  = wolfSSL_new(ctx_s));
+    XMEMSET((byte*)&fake_addr6, 0, sizeof(fake_addr6));
+    /* mimic a sockaddr_in6 struct, this way we can't test without
+     *  WOLFSSL_IPV6 */
+    fake_addr6.sin_family = WOLFSSL_IP6;
+    ExpectIntNE(sockfd = socket(AF_INET, SOCK_DGRAM, 0), -1);
+    ExpectIntEQ(wolfSSL_set_fd(ssl_c, sockfd), WOLFSSL_SUCCESS);
+    /* can't return error here, as the peer is opaque for wolfssl library at
+     * this point */
+    ExpectIntEQ(wolfSSL_dtls_set_peer(ssl_c, &fake_addr6, sizeof(fake_addr6)),
+        WOLFSSL_SUCCESS);
+    ExpectIntNE(fcntl(sockfd, F_SETFL, O_NONBLOCK), -1);
+    wolfSSL_dtls_set_using_nonblock(ssl_c, 1);
+    ExpectIntNE(wolfSSL_connect(ssl_c), WOLFSSL_SUCCESS);
+    ExpectIntEQ(ssl_c->error, WC_NO_ERR_TRACE(SOCKET_ERROR_E));
+
+    ExpectIntEQ(wolfSSL_dtls_set_peer(ssl_s, &fake_addr6, sizeof(fake_addr6)),
+        WOLFSSL_SUCCESS);
+    /* reuse the socket */
+    ExpectIntEQ(wolfSSL_set_fd(ssl_c, sockfd), WOLFSSL_SUCCESS);
+    wolfSSL_dtls_set_using_nonblock(ssl_s, 1);
+    ExpectIntNE(wolfSSL_accept(ssl_s), WOLFSSL_SUCCESS);
+    ExpectIntEQ(ssl_s->error, WC_NO_ERR_TRACE(SOCKET_ERROR_E));
+    if (sockfd != -1)
+        close(sockfd);
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_s);
+    return EXPECT_RESULT();
+}
+#else
+int test_dtls_ipv6_check(void)
+{
+    return TEST_SKIPPED;
+}
+#endif
+
+/*-- no_extensions (api.c lines 30824,30913) ---*/
+int test_dtls_no_extensions(void)
+{
+    EXPECT_DECLS;
+#if defined(WOLFSSL_DTLS) && defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && \
+    !defined(WOLFSSL_NO_TLS12)
+    WOLFSSL *ssl_s = NULL;
+    WOLFSSL_CTX *ctx_s = NULL;
+    struct test_memio_ctx test_ctx;
+    const byte chNoExtensions[] = {
+        /* Handshake type */
+        0x16,
+        /* Version */
+        0xfe, 0xff,
+        /* Epoch */
+        0x00, 0x00,
+        /* Seq number */
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+        /* Length */
+        0x00, 0x40,
+        /* CH type */
+        0x01,
+        /* Length */
+        0x00, 0x00, 0x34,
+        /* Msg Seq */
+        0x00, 0x00,
+        /* Frag offset */
+        0x00, 0x00, 0x00,
+        /* Frag length */
+        0x00, 0x00, 0x34,
+        /* Version */
+        0xfe, 0xff,
+        /* Random */
+        0x62, 0xfe, 0xbc, 0xfe, 0x2b, 0xfe, 0x3f, 0xeb, 0x03, 0xc4, 0xea, 0x37,
+        0xe7, 0x47, 0x7e, 0x8a, 0xd9, 0xbf, 0x77, 0x0f, 0x6c, 0xb6, 0x77, 0x0b,
+        0x03, 0x3f, 0x82, 0x2b, 0x21, 0x64, 0x57, 0x1d,
+        /* Session Length */
+        0x00,
+        /* Cookie Length */
+        0x00,
+        /* CS Length */
+        0x00, 0x0c,
+        /* CS */
+        0xc0, 0x0a, 0xc0, 0x09, 0xc0, 0x14, 0xc0, 0x13, 0x00, 0x39, 0x00, 0x33,
+        /* Comp Meths Length */
+        0x01,
+        /* Comp Meths */
+        0x00
+        /* And finally... no extensions */
+    };
+    int i;
+#ifdef OPENSSL_EXTRA
+    int repeats = 2;
+#else
+    int repeats = 1;
+#endif
+
+    for (i = 0; i < repeats; i++) {
+        XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+        ssl_s = NULL;
+        ctx_s = NULL;
+
+        ExpectIntEQ(test_memio_setup(&test_ctx, NULL, &ctx_s, NULL, &ssl_s,
+            NULL, wolfDTLS_server_method), 0);
+
+        test_memio_clear_buffer(&test_ctx, 0);
+        ExpectIntEQ(
+            test_memio_inject_message(&test_ctx, 1,
+                (const char *)chNoExtensions, sizeof(chNoExtensions)), 0);
+
+
+#ifdef OPENSSL_EXTRA
+        if (i > 0) {
+            ExpectIntEQ(wolfSSL_set_max_proto_version(ssl_s, DTLS1_2_VERSION),
+                        WOLFSSL_SUCCESS);
+        }
+#endif
+
+        ExpectIntEQ(wolfSSL_accept(ssl_s), -1);
+        ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+
+        /* Expecting a handshake msg. Either HVR or SH. */
+        ExpectIntGT(test_ctx.c_len, 0);
+        ExpectIntEQ(test_ctx.c_buff[0], 0x16);
+
+        wolfSSL_free(ssl_s);
+        wolfSSL_CTX_free(ctx_s);
+    }
+#endif
+    return EXPECT_RESULT();
+}
+
+/*-- dtls_1_0_hvr_downgrade (api.c lines 31038,31073) ---*/
+#if defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) && \
+    defined(HAVE_IO_TESTS_DEPENDENCIES)
+static void test_dtls_1_0_hvr_downgrade_ctx_ready(WOLFSSL_CTX* ctx)
+{
+    AssertIntEQ(wolfSSL_CTX_SetMinVersion(ctx, WOLFSSL_DTLSV1_2),
+                WOLFSSL_SUCCESS);
+}
+
+int test_dtls_1_0_hvr_downgrade(void)
+{
+    EXPECT_DECLS;
+    callback_functions func_cb_client;
+    callback_functions func_cb_server;
+
+    XMEMSET(&func_cb_client, 0, sizeof(callback_functions));
+    XMEMSET(&func_cb_server, 0, sizeof(callback_functions));
+
+    func_cb_client.doUdp = func_cb_server.doUdp = 1;
+    func_cb_client.method = wolfDTLS_client_method;
+    func_cb_server.method = wolfDTLSv1_2_server_method;
+    func_cb_client.ctx_ready = test_dtls_1_0_hvr_downgrade_ctx_ready;
+
+    test_wolfSSL_client_server_nofail(&func_cb_client, &func_cb_server);
+
+    ExpectIntEQ(func_cb_client.return_code, TEST_SUCCESS);
+    ExpectIntEQ(func_cb_server.return_code, TEST_SUCCESS);
+
+    return EXPECT_RESULT();
+}
+#else
+int test_dtls_1_0_hvr_downgrade(void)
+{
+    EXPECT_DECLS;
+    return EXPECT_RESULT();
+}
+#endif
+
+/*-- downgrade_scr_server (api.c lines 31230,31293) ---*/
+#if defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) && \
+    defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_SECURE_RENEGOTIATION)
+static void test_dtls_downgrade_scr_server_ctx_ready_server(WOLFSSL_CTX* ctx)
+{
+    AssertIntEQ(wolfSSL_CTX_SetMinVersion(ctx, WOLFSSL_DTLSV1_2),
+                WOLFSSL_SUCCESS);
+    AssertIntEQ(wolfSSL_CTX_UseSecureRenegotiation(ctx), WOLFSSL_SUCCESS);
+}
+
+static void test_dtls_downgrade_scr_server_ctx_ready(WOLFSSL_CTX* ctx)
+{
+    AssertIntEQ(wolfSSL_CTX_UseSecureRenegotiation(ctx), WOLFSSL_SUCCESS);
+}
+
+static void test_dtls_downgrade_scr_server_on_result(WOLFSSL* ssl)
+{
+    char testMsg[] = "Message after SCR";
+    char msgBuf[sizeof(testMsg)];
+    if (wolfSSL_is_server(ssl)) {
+        AssertIntEQ(wolfSSL_Rehandshake(ssl), WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR));
+        AssertIntEQ(wolfSSL_get_error(ssl, -1), WC_NO_ERR_TRACE(APP_DATA_READY));
+        AssertIntEQ(wolfSSL_read(ssl, msgBuf, sizeof(msgBuf)), sizeof(msgBuf));
+        AssertIntEQ(wolfSSL_Rehandshake(ssl), WOLFSSL_SUCCESS);
+        AssertIntEQ(wolfSSL_write(ssl, testMsg, sizeof(testMsg)),
+                    sizeof(testMsg));
+    }
+    else {
+        AssertIntEQ(wolfSSL_write(ssl, testMsg, sizeof(testMsg)),
+                    sizeof(testMsg));
+        AssertIntEQ(wolfSSL_read(ssl, msgBuf, sizeof(msgBuf)), sizeof(msgBuf));
+    }
+}
+
+int test_dtls_downgrade_scr_server(void)
+{
+    EXPECT_DECLS;
+    callback_functions func_cb_client;
+    callback_functions func_cb_server;
+
+    XMEMSET(&func_cb_client, 0, sizeof(callback_functions));
+    XMEMSET(&func_cb_server, 0, sizeof(callback_functions));
+
+    func_cb_client.doUdp = func_cb_server.doUdp = 1;
+    func_cb_client.method = wolfDTLSv1_2_client_method;
+    func_cb_server.method = wolfDTLS_server_method;
+    func_cb_client.ctx_ready = test_dtls_downgrade_scr_server_ctx_ready;
+    func_cb_server.ctx_ready = test_dtls_downgrade_scr_server_ctx_ready_server;
+    func_cb_client.on_result = test_dtls_downgrade_scr_server_on_result;
+    func_cb_server.on_result = test_dtls_downgrade_scr_server_on_result;
+
+    test_wolfSSL_client_server_nofail(&func_cb_client, &func_cb_server);
+
+    ExpectIntEQ(func_cb_client.return_code, TEST_SUCCESS);
+    ExpectIntEQ(func_cb_server.return_code, TEST_SUCCESS);
+
+    return EXPECT_RESULT();
+}
+#else
+int test_dtls_downgrade_scr_server(void)
+{
+    EXPECT_DECLS;
+    return EXPECT_RESULT();
+}
+#endif
+
+/*-- downgrade_scr (api.c lines 31295,31352) ---*/
+#if defined(WOLFSSL_DTLS) && !defined(WOLFSSL_NO_TLS12) && \
+    defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_SECURE_RENEGOTIATION)
+static void test_dtls_downgrade_scr_ctx_ready(WOLFSSL_CTX* ctx)
+{
+    AssertIntEQ(wolfSSL_CTX_SetMinVersion(ctx, WOLFSSL_DTLSV1_2),
+                WOLFSSL_SUCCESS);
+    AssertIntEQ(wolfSSL_CTX_UseSecureRenegotiation(ctx), WOLFSSL_SUCCESS);
+}
+
+static void test_dtls_downgrade_scr_on_result(WOLFSSL* ssl)
+{
+    char testMsg[] = "Message after SCR";
+    char msgBuf[sizeof(testMsg)];
+    if (wolfSSL_is_server(ssl)) {
+        AssertIntEQ(wolfSSL_Rehandshake(ssl), WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR));
+        AssertIntEQ(wolfSSL_get_error(ssl, -1), WC_NO_ERR_TRACE(APP_DATA_READY));
+        AssertIntEQ(wolfSSL_read(ssl, msgBuf, sizeof(msgBuf)), sizeof(msgBuf));
+        AssertIntEQ(wolfSSL_Rehandshake(ssl), WOLFSSL_SUCCESS);
+        AssertIntEQ(wolfSSL_write(ssl, testMsg, sizeof(testMsg)),
+                    sizeof(testMsg));
+    }
+    else {
+        AssertIntEQ(wolfSSL_write(ssl, testMsg, sizeof(testMsg)),
+                    sizeof(testMsg));
+        AssertIntEQ(wolfSSL_read(ssl, msgBuf, sizeof(msgBuf)), sizeof(msgBuf));
+    }
+}
+
+int test_dtls_downgrade_scr(void)
+{
+    EXPECT_DECLS;
+    callback_functions func_cb_client;
+    callback_functions func_cb_server;
+
+    XMEMSET(&func_cb_client, 0, sizeof(callback_functions));
+    XMEMSET(&func_cb_server, 0, sizeof(callback_functions));
+
+    func_cb_client.doUdp = func_cb_server.doUdp = 1;
+    func_cb_client.method = wolfDTLS_client_method;
+    func_cb_server.method = wolfDTLSv1_2_server_method;
+    func_cb_client.ctx_ready = test_dtls_downgrade_scr_ctx_ready;
+    func_cb_client.on_result = test_dtls_downgrade_scr_on_result;
+    func_cb_server.on_result = test_dtls_downgrade_scr_on_result;
+
+    test_wolfSSL_client_server_nofail(&func_cb_client, &func_cb_server);
+
+    ExpectIntEQ(func_cb_client.return_code, TEST_SUCCESS);
+    ExpectIntEQ(func_cb_server.return_code, TEST_SUCCESS);
+
+    return EXPECT_RESULT();
+}
+#else
+int test_dtls_downgrade_scr(void)
+{
+    EXPECT_DECLS;
+    return EXPECT_RESULT();
+}
+#endif
+
+/*-- client_hello_timeout_downgrade_with_helper (api.c lines 31354,31490) ---*/
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13) \
+    && !defined(WOLFSSL_NO_TLS12)
+
+static int test_dtls_client_hello_timeout_downgrade_read_cb(WOLFSSL *ssl,
+        char *data, int sz, void *ctx)
+{
+    static int call_counter = 0;
+    call_counter++;
+    (void)ssl;
+    (void)data;
+    (void)sz;
+    (void)ctx;
+    switch (call_counter) {
+        case 1:
+        case 2:
+            return WOLFSSL_CBIO_ERR_TIMEOUT;
+        case 3:
+            return WOLFSSL_CBIO_ERR_WANT_READ;
+        default:
+            AssertIntLE(call_counter, 3);
+            return -1;
+    }
+}
+#endif
+
+/* Make sure we don't send acks before getting a server hello */
+int test_dtls_client_hello_timeout_downgrade(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13) \
+    && !defined(WOLFSSL_NO_TLS12)
+
+    WOLFSSL_CTX *ctx_c = NULL;
+    WOLFSSL_CTX *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL;
+    WOLFSSL *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+    DtlsRecordLayerHeader* dtlsRH;
+    size_t len;
+    byte sequence_number[8];
+    int i;
+
+    for (i = 0; i < 2; i++) {
+        XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+
+        ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+            wolfDTLS_client_method, wolfDTLSv1_2_server_method), 0);
+
+        if (i == 0) {
+            /* First time simulate timeout in IO layer */
+            /* CH1 */
+            ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+            ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+            /* HVR */
+            ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+            ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+            /* CH2 */
+            ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+            ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+            /* SH flight */
+            ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+            ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+            /* Drop the SH */
+            if (EXPECT_SUCCESS()) {
+                ExpectIntEQ(test_memio_drop_message(&test_ctx, 1, 0), 0);
+            }
+            /* Read the remainder of the flight */
+            ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+            ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+            wolfSSL_SSLSetIORecv(ssl_c,
+                    test_dtls_client_hello_timeout_downgrade_read_cb);
+            /* CH3 */
+            ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+            ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+            wolfSSL_SSLSetIORecv(ssl_c, test_memio_read_cb);
+        }
+        else {
+            /* Second time call wolfSSL_dtls_got_timeout */
+            /* CH1 */
+            ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+            ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+            /* HVR */
+            ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+            ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+            /* CH2 */
+            ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+            ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+            /* SH flight */
+            ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+            ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+            /* Drop the SH */
+            if (EXPECT_SUCCESS()) {
+                ExpectIntEQ(test_memio_drop_message(&test_ctx, 1, 0), 0);
+            }
+            /* Read the remainder of the flight */
+            ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+            ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+            /* Quick timeout should be set as we received at least one msg */
+            ExpectIntEQ(wolfSSL_dtls13_use_quick_timeout(ssl_c), 1);
+            ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
+            /* Quick timeout should be cleared after a quick timeout */
+            /* CH3 */
+            ExpectIntEQ(wolfSSL_dtls13_use_quick_timeout(ssl_c), 0);
+            ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
+        }
+
+        /* Parse out to make sure we got exactly one ClientHello message */
+        XMEMSET(&sequence_number, 0, sizeof(sequence_number));
+        /* Second ClientHello after HVR */
+        sequence_number[7] = 2;
+        dtlsRH = (DtlsRecordLayerHeader*)test_ctx.s_buff;
+        ExpectIntEQ(dtlsRH->type, handshake);
+        ExpectIntEQ(dtlsRH->pvMajor, DTLS_MAJOR);
+        ExpectIntEQ(dtlsRH->pvMinor, DTLSv1_2_MINOR);
+        ExpectIntEQ(XMEMCMP(sequence_number, dtlsRH->sequence_number,
+                sizeof(sequence_number)), 0);
+        len = (size_t)((dtlsRH->length[0] << 8) | dtlsRH->length[1]);
+        ExpectIntEQ(sizeof(DtlsRecordLayerHeader) + len, test_ctx.s_len);
+
+        /* Connection should be able to continue */
+        ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+
+        wolfSSL_free(ssl_c);
+        wolfSSL_free(ssl_s);
+        wolfSSL_CTX_free(ctx_c);
+        wolfSSL_CTX_free(ctx_s);
+        ssl_c = NULL;
+        ssl_s = NULL;
+        ctx_c = NULL;
+        ctx_s = NULL;
+        if (!EXPECT_SUCCESS())
+            break;
+    }
+
+#endif
+    return EXPECT_RESULT();
+}
+
+/*-- client_hello_timeout_with_helper (api.c lines 31492,31581) ---*/
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
+static int test_dtls_client_hello_timeout_read_cb(WOLFSSL *ssl, char *data,
+        int sz, void *ctx)
+{
+    static int call_counter = 0;
+    call_counter++;
+    (void)ssl;
+    (void)data;
+    (void)sz;
+    (void)ctx;
+    switch (call_counter) {
+        case 1:
+            return WOLFSSL_CBIO_ERR_TIMEOUT;
+        case 2:
+            return WOLFSSL_CBIO_ERR_WANT_READ;
+        default:
+            AssertIntLE(call_counter, 2);
+            return -1;
+    }
+}
+#endif
+
+/* Make sure we don't send acks before getting a server hello */
+int test_dtls_client_hello_timeout(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
+    WOLFSSL *ssl_c = NULL;
+    WOLFSSL_CTX *ctx_c = NULL;
+    struct test_memio_ctx test_ctx;
+    DtlsRecordLayerHeader* dtlsRH;
+    size_t idx;
+    size_t len;
+    byte sequence_number[8];
+    int i;
+
+    for (i = 0; i < 2; i++) {
+        XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+
+        ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, NULL, &ssl_c, NULL,
+            wolfDTLSv1_3_client_method, NULL), 0);
+
+        if (i == 0) {
+            /* First time simulate timeout in IO layer */
+            wolfSSL_SSLSetIORecv(ssl_c, test_dtls_client_hello_timeout_read_cb);
+            ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
+            ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+        }
+        else {
+            /* Second time call wolfSSL_dtls_got_timeout */
+            ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
+            ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+            ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
+        }
+
+        /* Parse out to make sure we got exactly two ClientHello messages */
+        idx = 0;
+        XMEMSET(&sequence_number, 0, sizeof(sequence_number));
+        /* First ClientHello */
+        dtlsRH = (DtlsRecordLayerHeader*)(test_ctx.s_buff + idx);
+        ExpectIntEQ(dtlsRH->type, handshake);
+        ExpectIntEQ(dtlsRH->pvMajor, DTLS_MAJOR);
+        ExpectIntEQ(dtlsRH->pvMinor, DTLSv1_2_MINOR);
+        ExpectIntEQ(XMEMCMP(sequence_number, dtlsRH->sequence_number,
+                sizeof(sequence_number)), 0);
+        len = (size_t)((dtlsRH->length[0] << 8) | dtlsRH->length[1]);
+        ExpectIntLT(idx + sizeof(DtlsRecordLayerHeader) + len, test_ctx.s_len);
+        idx += sizeof(DtlsRecordLayerHeader) + len;
+        /* Second ClientHello */
+        sequence_number[7] = 1;
+        dtlsRH = (DtlsRecordLayerHeader*)(test_ctx.s_buff + idx);
+        ExpectIntEQ(dtlsRH->type, handshake);
+        ExpectIntEQ(dtlsRH->pvMajor, DTLS_MAJOR);
+        ExpectIntEQ(dtlsRH->pvMinor, DTLSv1_2_MINOR);
+        ExpectIntEQ(XMEMCMP(sequence_number, dtlsRH->sequence_number,
+                sizeof(sequence_number)), 0);
+        len = (size_t)((dtlsRH->length[0] << 8) | dtlsRH->length[1]);
+        ExpectIntEQ(idx + sizeof(DtlsRecordLayerHeader) + len, test_ctx.s_len);
+
+        wolfSSL_free(ssl_c);
+        wolfSSL_CTX_free(ctx_c);
+        ssl_c = NULL;
+        ctx_c = NULL;
+        if (!EXPECT_SUCCESS())
+            break;
+    }
+
+#endif
+    return EXPECT_RESULT();
+}
+
+/*-- dropped_ccs (api.c lines 31584,31648) ---*/
+int test_dtls_dropped_ccs(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS) \
+    && !defined(WOLFSSL_NO_TLS12)
+
+    WOLFSSL_CTX *ctx_c = NULL;
+    WOLFSSL_CTX *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL;
+    WOLFSSL *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+    DtlsRecordLayerHeader* dtlsRH;
+    size_t len;
+    byte data[1];
+
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method), 0);
+
+    /* CH1 */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* HVR */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    /* CH2 */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Server first flight */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Client flight */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Server ccs + finished */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), 1);
+
+    /* Drop the ccs */
+    dtlsRH = (DtlsRecordLayerHeader*)test_ctx.c_buff;
+    len = (size_t)((dtlsRH->length[0] << 8) | dtlsRH->length[1]);
+    ExpectIntEQ(len, 1);
+    ExpectIntEQ(dtlsRH->type, change_cipher_spec);
+    if (EXPECT_SUCCESS()) {
+        ExpectIntEQ(test_memio_drop_message(&test_ctx, 1, 0), 0);
+    }
+
+    /* Client rtx flight */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
+    /* Server ccs + finished rtx */
+    ExpectIntEQ(wolfSSL_read(ssl_s, data, sizeof(data)), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Client processes finished */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), 1);
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
+
+/*-- seq_num_downgrade_with_helper (api.c lines 31650,31722) ---*/
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS) \
+    && !defined(WOLFSSL_NO_TLS12)
+static int test_dtls_seq_num_downgrade_check_num(byte* ioBuf, int ioBufLen,
+        byte seq_num)
+{
+    EXPECT_DECLS;
+    DtlsRecordLayerHeader* dtlsRH;
+    byte sequence_number[8];
+
+    XMEMSET(&sequence_number, 0, sizeof(sequence_number));
+
+    ExpectIntGE(ioBufLen, sizeof(*dtlsRH));
+    dtlsRH = (DtlsRecordLayerHeader*)ioBuf;
+    ExpectIntEQ(dtlsRH->type, handshake);
+    ExpectIntEQ(dtlsRH->pvMajor, DTLS_MAJOR);
+    ExpectIntEQ(dtlsRH->pvMinor, DTLSv1_2_MINOR);
+    sequence_number[7] = seq_num;
+    ExpectIntEQ(XMEMCMP(sequence_number, dtlsRH->sequence_number,
+            sizeof(sequence_number)), 0);
+
+    return EXPECT_RESULT();
+}
+#endif
+
+/*
+ * Make sure that we send the correct sequence number after a HelloVerifyRequest
+ * and after a HelloRetryRequest. This is testing the server side as it is
+ * operating statelessly and should copy the sequence number of the ClientHello.
+ */
+int test_dtls_seq_num_downgrade(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS) \
+    && !defined(WOLFSSL_NO_TLS12)
+    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_2_client_method, wolfDTLS_server_method), 0);
+
+    /* CH1 */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    ExpectIntEQ(test_dtls_seq_num_downgrade_check_num(test_ctx.s_buff,
+            test_ctx.s_len, 0), TEST_SUCCESS);
+    /* HVR */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    ExpectIntEQ(test_dtls_seq_num_downgrade_check_num(test_ctx.c_buff,
+            test_ctx.c_len, 0), TEST_SUCCESS);
+    /* CH2 */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    ExpectIntEQ(test_dtls_seq_num_downgrade_check_num(test_ctx.s_buff,
+            test_ctx.s_len, 1), TEST_SUCCESS);
+    /* Server first flight */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    ExpectIntEQ(test_dtls_seq_num_downgrade_check_num(test_ctx.c_buff,
+            test_ctx.c_len, 1), TEST_SUCCESS);
+
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
+
+/*-- old_seq_number (api.c lines 31953,32005) ---*/
+int test_dtls_old_seq_number(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS) && \
+    !defined(WOLFSSL_NO_TLS12)
+    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method), 0);
+
+    /* CH1 */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* HVR */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    /* CH2 */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Server first flight */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Client second flight */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Modify the sequence number */
+    {
+        DtlsRecordLayerHeader* dtlsRH = (DtlsRecordLayerHeader*)test_ctx.s_buff;
+        XMEMSET(dtlsRH->sequence_number, 0, sizeof(dtlsRH->sequence_number));
+    }
+    /* Server second flight */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Server should not do anything as a pkt was dropped */
+    ExpectIntEQ(test_ctx.c_len, 0);
+    ExpectIntEQ(test_ctx.s_len, 0);
+    /* Trigger rtx */
+    ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
+
+    /* Complete connection */
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
+
+/*-- dtls12_missing_finished (api.c lines 32007,32068) ---*/
+int test_dtls12_missing_finished(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS) && \
+    !defined(WOLFSSL_NO_TLS12)
+    WOLFSSL_CTX *ctx_c = NULL;
+    WOLFSSL_CTX *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL;
+    WOLFSSL *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+    const char test_str[] = "test string";
+    char test_buf[sizeof(test_str)];
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method), 0);
+
+    /* CH1 */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* HVR */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    /* CH2 */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Server first flight */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Client second flight with finished */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Server second flight with finished */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), 1);
+    /* Let's clear the output */
+    test_memio_clear_buffer(&test_ctx, 1);
+    /* Let's send some app data */
+    ExpectIntEQ(wolfSSL_write(ssl_s, test_str, sizeof(test_str)),
+                sizeof(test_str));
+    /* Client should not error out on a missing finished */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Server rtx second flight with finished */
+    ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_s), 1);
+    /* Client process rest of handshake */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), 1);
+
+    /* Let's send some app data */
+    ExpectIntEQ(wolfSSL_write(ssl_s, test_str, sizeof(test_str)),
+                sizeof(test_str));
+    ExpectIntEQ(wolfSSL_read(ssl_c, test_buf, sizeof(test_buf)),
+                sizeof(test_str));
+    ExpectBufEQ(test_buf, test_str, sizeof(test_str));
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
+
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS) && \
+    defined(WOLFSSL_SESSION_EXPORT) && defined(HAVE_ENCRYPT_THEN_MAC) && \
+    !defined(WOLFSSL_AEAD_ONLY) && !defined(WOLFSSL_NO_TLS12) && \
+    !defined(NO_RSA) && !defined(NO_AES) && defined(HAVE_AES_CBC) && \
+    !defined(NO_SHA256) && defined(HAVE_ECC)
+/* Dummy peer callbacks so the DTLS exporter/importer has peer information to
+ * work with (the library requires these unless built with
+ * WOLFSSL_SESSION_EXPORT_NOPEER). */
+static int test_dtls_export_etm_get_peer(WOLFSSL* ssl, char* ip, int* ipSz,
+        unsigned short* port, int* fam)
+{
+    (void)ssl;
+    ip[0] = -1;
+    *ipSz = 1;
+    *port = 1;
+    *fam = 2;
+    return 1;
+}
+
+static int test_dtls_export_etm_set_peer(WOLFSSL* ssl, char* ip, int ipSz,
+        unsigned short port, int fam)
+{
+    (void)ssl;
+    if (ip[0] != -1 || ipSz != 1 || port != 1 || fam != 2)
+        return 0;
+    return 1;
+}
+#endif
+
+/* Regression test for DTLS session export/import dropping the Encrypt-Then-MAC
+ * options. Historically the ETM option fields were only serialized for TLS, so
+ * a re-imported DTLS session lost the negotiated ETM state and broke the record
+ * layer. Establish a DTLS 1.2 connection with a CBC cipher suite (where ETM
+ * applies), export the session, re-import it into a fresh WOLFSSL, and confirm
+ * the ETM option fields survive the round trip. */
+int test_dtls12_export_import_etm(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS) && \
+    defined(WOLFSSL_SESSION_EXPORT) && defined(HAVE_ENCRYPT_THEN_MAC) && \
+    !defined(WOLFSSL_AEAD_ONLY) && !defined(WOLFSSL_NO_TLS12) && \
+    !defined(NO_RSA) && !defined(NO_AES) && defined(HAVE_AES_CBC) && \
+    !defined(NO_SHA256) && defined(HAVE_ECC)
+    /* TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256 - a CBC suite, where ETM applies. */
+    const char* cbcSuite = "ECDHE-RSA-AES128-SHA256";
+    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
+    WOLFSSL *ssl_imp = NULL;
+    struct test_memio_ctx test_ctx;
+    unsigned char* session = NULL;
+    unsigned int sessionSz = 0;
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+                    wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method), 0);
+    ExpectIntEQ(wolfSSL_set_cipher_list(ssl_c, cbcSuite), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_set_cipher_list(ssl_s, cbcSuite), WOLFSSL_SUCCESS);
+
+    /* The exporter/importer needs peer info callbacks. */
+    wolfSSL_CTX_SetIOGetPeer(ctx_s, test_dtls_export_etm_get_peer);
+    wolfSSL_CTX_SetIOSetPeer(ctx_s, test_dtls_export_etm_set_peer);
+
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+
+    /* Sanity: the handshake itself negotiated ETM on both sides. */
+    if (ssl_c != NULL)
+        ExpectIntEQ(ssl_c->options.encThenMac, 1);
+    if (ssl_s != NULL)
+        ExpectIntEQ(ssl_s->options.encThenMac, 1);
+
+    /* Export the server's DTLS session. */
+    ExpectIntGE(wolfSSL_dtls_export(ssl_s, NULL, &sessionSz), 0);
+    ExpectIntGT(sessionSz, 0);
+    ExpectNotNull(session = (unsigned char*)XMALLOC(sessionSz, NULL,
+                    DYNAMIC_TYPE_TMP_BUFFER));
+    ExpectIntGE(wolfSSL_dtls_export(ssl_s, session, &sessionSz), 0);
+
+    /* Import into a fresh WOLFSSL and confirm the ETM state survived. */
+    ExpectNotNull(ssl_imp = wolfSSL_new(ctx_s));
+    ExpectIntGE(wolfSSL_dtls_import(ssl_imp, session, sessionSz), 0);
+    if (ssl_imp != NULL) {
+        /* Regression check: pre-fix these were all reset to 0 for DTLS. */
+        ExpectIntEQ(ssl_imp->options.encThenMac, 1);
+        ExpectIntEQ(ssl_imp->options.startedETMRead, 1);
+        ExpectIntEQ(ssl_imp->options.startedETMWrite, 1);
+        ExpectIntEQ(ssl_imp->options.disallowEncThenMac, 0);
+    }
+
+    XFREE(session, NULL, DYNAMIC_TYPE_TMP_BUFFER);
+    wolfSSL_free(ssl_imp);
+    wolfSSL_free(ssl_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
diff --git a/tests/api/test_dtls.h b/tests/api/test_dtls.h
index 823b9b9f45..fd22d125bf 100644
--- a/tests/api/test_dtls.h
+++ b/tests/api/test_dtls.h
@@ -23,15 +23,8 @@
 #define TESTS_API_DTLS_H
 
 int test_dtls12_basic_connection_id(void);
-int test_dtls13_basic_connection_id(void);
-int test_dtls13_hrr_want_write(void);
-int test_dtls13_every_write_want_write(void);
 int test_wolfSSL_dtls_cid_parse(void);
 int test_wolfSSL_dtls_set_pending_peer(void);
-int test_dtls13_epochs(void);
-int test_dtls13_ack_order(void);
-int test_dtls13_ack_overflow(void);
-int test_dtls13_ack_dup_write_counter(void);
 int test_dtls_version_checking(void);
 int test_dtls_short_ciphertext(void);
 int test_dtls12_record_length_mismatch(void);
@@ -41,34 +34,56 @@ int test_dtls13_short_read(void);
 int test_records_span_network_boundaries(void);
 int test_dtls_record_cross_boundaries(void);
 int test_dtls_rtx_across_epoch_change(void);
-int test_dtls13_ch2_rtx_no_ch1(void);
-int test_dtls13_frag_ch2_with_ch1_rtx(void);
 int test_dtls_drop_client_ack(void);
 int test_dtls_bogus_finished_epoch_zero(void);
 int test_dtls_replay(void);
-int test_dtls_srtp(void);
 int test_dtls_timeout(void);
 int test_dtls_certreq_order(void);
 int test_dtls_memio_wolfio(void);
 int test_dtls_memio_wolfio_stateless(void);
 int test_dtls_mtu_fragment_headroom(void);
 int test_dtls_mtu_split_messages(void);
-int test_dtls13_min_rtx_interval(void);
-int test_dtls13_no_session_id_echo(void);
-int test_dtls13_5_9_0_compat(void);
 int test_dtls_set_session_min_downgrade(void);
+int test_dtls12_export_import_etm(void);
+
+/* DTLS tests moved out of tests/api.c. */
+int test_dtls_msg_from_other_peer(void);
+int test_dtls_ipv6_check(void);
+int test_dtls_no_extensions(void);
+int test_dtls_1_0_hvr_downgrade(void);
+int test_dtls_downgrade_scr_server(void);
+int test_dtls_downgrade_scr(void);
+int test_dtls_client_hello_timeout_downgrade(void);
+int test_dtls_client_hello_timeout(void);
+int test_dtls_dropped_ccs(void);
+int test_dtls_seq_num_downgrade(void);
+int test_dtls_old_seq_number(void);
+int test_dtls12_missing_finished(void);
+int test_wolfSSL_dtls_export(void);
+int test_wolfSSL_dtls_export_peers(void);
+int test_wolfSSL_dtls_import_state_extra_window_words(void);
+int test_wolfSSL_DTLS_either_side(void);
+int test_generate_cookie(void);
+int test_wolfSSL_dtls_set_mtu(void);
+int test_wolfSSL_dtls_plaintext(void);
+int test_wolfSSL_dtls_fragments(void);
+int test_wolfSSL_ignore_alert_before_cookie(void);
+int test_wolfSSL_dtls_bad_record(void);
+int test_wolfSSL_dtls_AEAD_limit(void);
+int test_wolfSSL_dtls_stateless(void);
+int test_wolfSSL_dtls_stateless_hrr_group(void);
+int test_wolfSSL_DtlsUpdateWindow(void);
+int test_wolfSSL_DTLS_fragment_buckets(void);
+int test_wolfSSL_dtls_stateless2(void);
+int test_wolfSSL_dtls_stateless_maxfrag(void);
+int test_wolfSSL_dtls_stateless_resume(void);
+int test_wolfSSL_dtls_stateless_downgrade(void);
+int test_WOLFSSL_dtls_version_alert(void);
 
 #define TEST_DTLS_DECLS                                                        \
         TEST_DECL_GROUP("dtls", test_dtls12_basic_connection_id),              \
-        TEST_DECL_GROUP("dtls", test_dtls13_basic_connection_id),              \
-        TEST_DECL_GROUP("dtls", test_dtls13_hrr_want_write),                   \
-        TEST_DECL_GROUP("dtls", test_dtls13_every_write_want_write),           \
         TEST_DECL_GROUP("dtls", test_wolfSSL_dtls_cid_parse),                  \
         TEST_DECL_GROUP("dtls", test_wolfSSL_dtls_set_pending_peer),           \
-        TEST_DECL_GROUP("dtls", test_dtls13_epochs),                           \
-        TEST_DECL_GROUP("dtls", test_dtls13_ack_order),                        \
-        TEST_DECL_GROUP("dtls", test_dtls13_ack_overflow),                     \
-        TEST_DECL_GROUP("dtls", test_dtls13_ack_dup_write_counter),            \
         TEST_DECL_GROUP("dtls", test_dtls_version_checking),                   \
         TEST_DECL_GROUP("dtls", test_dtls_short_ciphertext),                   \
         TEST_DECL_GROUP("dtls", test_dtls12_record_length_mismatch),           \
@@ -78,20 +93,47 @@ int test_dtls_set_session_min_downgrade(void);
         TEST_DECL_GROUP("dtls", test_records_span_network_boundaries),         \
         TEST_DECL_GROUP("dtls", test_dtls_record_cross_boundaries),            \
         TEST_DECL_GROUP("dtls", test_dtls_rtx_across_epoch_change),            \
-        TEST_DECL_GROUP("dtls", test_dtls13_ch2_rtx_no_ch1),                   \
-        TEST_DECL_GROUP("dtls", test_dtls13_frag_ch2_with_ch1_rtx),            \
         TEST_DECL_GROUP("dtls", test_dtls_drop_client_ack),                    \
         TEST_DECL_GROUP("dtls", test_dtls_bogus_finished_epoch_zero),          \
         TEST_DECL_GROUP("dtls", test_dtls_replay),                             \
-        TEST_DECL_GROUP("dtls", test_dtls_srtp),                               \
         TEST_DECL_GROUP("dtls", test_dtls_certreq_order),                      \
         TEST_DECL_GROUP("dtls", test_dtls_timeout),                            \
         TEST_DECL_GROUP("dtls", test_dtls_memio_wolfio),                       \
         TEST_DECL_GROUP("dtls", test_dtls_mtu_fragment_headroom),              \
         TEST_DECL_GROUP("dtls", test_dtls_mtu_split_messages),                 \
         TEST_DECL_GROUP("dtls", test_dtls_memio_wolfio_stateless),             \
-        TEST_DECL_GROUP("dtls", test_dtls13_min_rtx_interval),                 \
-        TEST_DECL_GROUP("dtls", test_dtls13_no_session_id_echo),               \
         TEST_DECL_GROUP("dtls", test_dtls_set_session_min_downgrade),          \
-        TEST_DECL_GROUP("dtls", test_dtls13_5_9_0_compat)
+        TEST_DECL_GROUP("dtls", test_wolfSSL_dtls_export),                     \
+        TEST_DECL_GROUP("dtls", test_wolfSSL_dtls_export_peers),               \
+        TEST_DECL_GROUP("dtls", test_wolfSSL_dtls_import_state_extra_window_words), \
+        TEST_DECL_GROUP("dtls", test_wolfSSL_DTLS_either_side),                \
+        TEST_DECL_GROUP("dtls", test_generate_cookie),                         \
+        TEST_DECL_GROUP("dtls", test_wolfSSL_dtls_set_mtu),                    \
+        TEST_DECL_GROUP("dtls", test_wolfSSL_dtls_plaintext),                  \
+        TEST_DECL_GROUP("dtls", test_wolfSSL_dtls_fragments),                  \
+        TEST_DECL_GROUP("dtls", test_wolfSSL_ignore_alert_before_cookie),      \
+        TEST_DECL_GROUP("dtls", test_wolfSSL_dtls_bad_record),                 \
+        TEST_DECL_GROUP("dtls", test_wolfSSL_dtls_AEAD_limit),                 \
+        TEST_DECL_GROUP("dtls", test_wolfSSL_dtls_stateless),                  \
+        TEST_DECL_GROUP("dtls", test_wolfSSL_dtls_stateless_hrr_group),        \
+        TEST_DECL_GROUP("dtls", test_wolfSSL_DtlsUpdateWindow),                \
+        TEST_DECL_GROUP("dtls", test_wolfSSL_DTLS_fragment_buckets),           \
+        TEST_DECL_GROUP("dtls", test_wolfSSL_dtls_stateless2),                 \
+        TEST_DECL_GROUP("dtls", test_wolfSSL_dtls_stateless_maxfrag),          \
+        TEST_DECL_GROUP("dtls", test_wolfSSL_dtls_stateless_resume),           \
+        TEST_DECL_GROUP("dtls", test_wolfSSL_dtls_stateless_downgrade),        \
+        TEST_DECL_GROUP("dtls", test_WOLFSSL_dtls_version_alert),              \
+        TEST_DECL_GROUP("dtls", test_dtls_msg_from_other_peer),                \
+        TEST_DECL_GROUP("dtls", test_dtls_ipv6_check),                         \
+        TEST_DECL_GROUP("dtls", test_dtls_no_extensions),                      \
+        TEST_DECL_GROUP("dtls", test_dtls_1_0_hvr_downgrade),                  \
+        TEST_DECL_GROUP("dtls", test_dtls_downgrade_scr_server),               \
+        TEST_DECL_GROUP("dtls", test_dtls_downgrade_scr),                      \
+        TEST_DECL_GROUP("dtls", test_dtls_client_hello_timeout_downgrade),     \
+        TEST_DECL_GROUP("dtls", test_dtls_client_hello_timeout),               \
+        TEST_DECL_GROUP("dtls", test_dtls_dropped_ccs),                        \
+        TEST_DECL_GROUP("dtls", test_dtls_seq_num_downgrade),                  \
+        TEST_DECL_GROUP("dtls", test_dtls_old_seq_number),                     \
+        TEST_DECL_GROUP("dtls", test_dtls12_missing_finished),                 \
+        TEST_DECL_GROUP("dtls", test_dtls12_export_import_etm)
 #endif /* TESTS_API_DTLS_H */
diff --git a/tests/api/test_dtls13.c b/tests/api/test_dtls13.c
new file mode 100644
index 0000000000..1716fd439d
--- /dev/null
+++ b/tests/api/test_dtls13.c
@@ -0,0 +1,1781 @@
+/* test_dtls13.c
+ *
+ * Copyright (C) 2006-2026 wolfSSL Inc.
+ *
+ * This file is part of wolfSSL.
+ *
+ * wolfSSL is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * wolfSSL is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1335, USA
+ */
+
+#include <tests/unit.h>
+
+#ifdef NO_INLINE
+    #include <wolfssl/wolfcrypt/misc.h>
+#else
+    #define WOLFSSL_MISC_INCLUDED
+    #include <wolfcrypt/src/misc.c>
+#endif
+
+#include <wolfssl/ssl.h>
+#include <wolfssl/internal.h>
+#include <tests/api/api.h>
+#include <tests/utils.h>
+#include <tests/api/test_dtls13.h>
+
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
+
+
+int test_dtls13_bad_epoch_ch(void)
+{
+    EXPECT_DECLS;
+    WOLFSSL_CTX *ctx_c = NULL;
+    WOLFSSL_CTX *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL;
+    WOLFSSL *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+    const int EPOCH_OFF = 3;
+    int groups[] = {
+        WOLFSSL_ECC_SECP256R1,
+        WOLFSSL_ECC_SECP384R1,
+    };
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
+
+    /* disable hrr cookie so we can later check msgsReceived.got_client_hello
+     *  with just one message */
+    ExpectIntEQ(wolfSSL_disable_hrr_cookie(ssl_s), WOLFSSL_SUCCESS);
+
+    /* Set client groups to traditional only to avoid CH fragmentation */
+    ExpectIntEQ(wolfSSL_set_groups(ssl_c, groups, 2), WOLFSSL_SUCCESS);
+
+    ExpectIntNE(wolfSSL_connect(ssl_c), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)),
+        WOLFSSL_ERROR_WANT_READ);
+
+    ExpectIntGE(test_ctx.s_len, EPOCH_OFF + 2);
+
+    /* first CH should use epoch 0x0 */
+    ExpectTrue((test_ctx.s_buff[EPOCH_OFF] == 0x0) &&
+        (test_ctx.s_buff[EPOCH_OFF + 1] == 0x0));
+
+    /* change epoch to 2 */
+    test_ctx.s_buff[EPOCH_OFF + 1] = 0x2;
+
+    ExpectIntNE(wolfSSL_accept(ssl_s), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, WC_NO_ERR_TRACE(WOLFSSL_FATAL_ERROR)),
+        WOLFSSL_ERROR_WANT_READ);
+
+    ExpectIntNE(ssl_s->msgsReceived.got_client_hello, 1);
+
+    /* resend the CH */
+    ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
+
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_s);
+
+    return EXPECT_RESULT();
+}
+#else
+int test_dtls13_bad_epoch_ch(void)
+{
+    return TEST_SKIPPED;
+}
+#endif
+
+#if defined(HAVE_NULL_CIPHER) && defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) \
+    && defined(WOLFSSL_DTLS13)
+static byte* test_find_string(const char *string,
+    byte *buf, int buf_size)
+{
+    int string_size, i;
+
+    string_size = (int)XSTRLEN(string);
+    for (i = 0; i < buf_size - string_size - 1; i++) {
+        if (XSTRCMP((char*)&buf[i], string) == 0)
+            return &buf[i];
+    }
+    return NULL;
+}
+
+int test_wolfSSL_dtls13_null_cipher(void)
+{
+    EXPECT_DECLS;
+    WOLFSSL_CTX *ctx_c = NULL;
+    WOLFSSL_CTX *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL;
+    WOLFSSL *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+    const char *test_str = "test";
+    int test_str_size;
+    byte buf[255], *ptr = NULL;
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+    test_ctx.c_ciphers = test_ctx.s_ciphers = "TLS13-SHA256-SHA256";
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+    test_str_size = XSTRLEN("test") + 1;
+    ExpectIntEQ(wolfSSL_write(ssl_c, test_str, test_str_size), test_str_size);
+    ExpectIntEQ(wolfSSL_read(ssl_s, buf, sizeof(buf)), test_str_size);
+    ExpectIntEQ(XSTRCMP((char*)buf, test_str), 0);
+
+    ExpectIntEQ(wolfSSL_write(ssl_c, test_str, test_str_size), test_str_size);
+
+    /* check that the packet was sent cleartext */
+    ExpectNotNull(ptr = test_find_string(test_str, test_ctx.s_buff,
+        test_ctx.s_len));
+    if (ptr != NULL) {
+        /* modify the message */
+        *ptr = 'H';
+        /* bad messages should be ignored in DTLS */
+        ExpectIntEQ(wolfSSL_read(ssl_s, buf, sizeof(buf)), -1);
+        ExpectIntEQ(ssl_s->error, WC_NO_ERR_TRACE(WANT_READ));
+    }
+
+    ExpectIntEQ(wolfSSL_shutdown(ssl_c), WOLFSSL_SHUTDOWN_NOT_DONE);
+    ExpectIntEQ(wolfSSL_shutdown(ssl_s), WOLFSSL_SHUTDOWN_NOT_DONE);
+    ExpectIntEQ(wolfSSL_shutdown(ssl_c), 1);
+    ExpectIntEQ(wolfSSL_shutdown(ssl_s), 1);
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_CTX_free(ctx_s);
+    return TEST_SUCCESS;
+}
+#else
+int test_wolfSSL_dtls13_null_cipher(void)
+{
+    return TEST_SKIPPED;
+}
+#endif
+
+int test_dtls13_frag_ch_pq(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13) \
+    && defined(WOLFSSL_DTLS_CH_FRAG) && defined(WOLFSSL_HAVE_MLKEM)
+    WOLFSSL_CTX *ctx_c = NULL;
+    WOLFSSL_CTX *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL;
+    WOLFSSL *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+    const char *test_str = "test";
+    int test_str_size;
+    byte buf[255];
+#if defined(WOLFSSL_MLKEM_KYBER)
+    #if !defined(WOLFSSL_NO_KYBER1024)
+    int group = WOLFSSL_KYBER_LEVEL5;
+    const char *group_name = "KYBER_LEVEL5";
+    #elif !defined(WOLFSSL_NO_KYBER768)
+    int group = WOLFSSL_KYBER_LEVEL3;
+    const char *group_name = "KYBER_LEVEL3";
+    #else
+    int group = WOLFSSL_KYBER_LEVEL1;
+    const char *group_name = "KYBER_LEVEL1";
+    #endif
+#elif !defined(WOLFSSL_NO_ML_KEM) && !defined(WOLFSSL_TLS_NO_MLKEM_STANDALONE)
+    #if !defined(WOLFSSL_NO_ML_KEM_1024)
+    int group = WOLFSSL_ML_KEM_1024;
+    const char *group_name = "ML_KEM_1024";
+    #elif !defined(WOLFSSL_NO_ML_KEM_768)
+    int group = WOLFSSL_ML_KEM_768;
+    const char *group_name = "ML_KEM_768";
+    #else
+    int group = WOLFSSL_ML_KEM_512;
+    const char *group_name = "ML_KEM_512";
+    #endif
+#elif defined(WOLFSSL_PQC_HYBRIDS)
+    #if defined(HAVE_CURVE25519) && !defined(WOLFSSL_NO_ML_KEM_768)
+    int group = WOLFSSL_X25519MLKEM768;
+    const char *group_name = "X25519MLKEM768";
+    #elif !defined(WOLFSSL_NO_ML_KEM_768)
+    int group = WOLFSSL_SECP256R1MLKEM768;
+    const char *group_name = "SecP256r1MLKEM768";
+    #else
+    int group = WOLFSSL_SECP384R1MLKEM1024;
+    const char *group_name = "SecP384r1MLKEM1024";
+    #endif
+#endif
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
+    /* Add in a large post-quantum key share to make the CH long. */
+    ExpectIntEQ(wolfSSL_set_groups(ssl_c, &group, 1), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_UseKeyShare(ssl_c, group), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_dtls13_allow_ch_frag(ssl_s, 1), WOLFSSL_SUCCESS);
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+    ExpectStrEQ(wolfSSL_get_curve_name(ssl_c), group_name);
+    ExpectStrEQ(wolfSSL_get_curve_name(ssl_s), group_name);
+    test_str_size = XSTRLEN("test") + 1;
+    ExpectIntEQ(wolfSSL_write(ssl_c, test_str, test_str_size), test_str_size);
+    ExpectIntEQ(wolfSSL_read(ssl_s, buf, sizeof(buf)), test_str_size);
+    ExpectIntEQ(XSTRCMP((char*)buf, test_str), 0);
+    ExpectIntEQ(wolfSSL_write(ssl_c, test_str, test_str_size), test_str_size);
+    wolfSSL_free(ssl_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
+
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS) \
+    && defined(WOLFSSL_DTLS_MTU) && defined(WOLFSSL_DTLS_CH_FRAG) && \
+    defined(WOLFSSL_AES_256)
+static int test_dtls_frag_ch_count_records(byte* b, int len)
+{
+    DtlsRecordLayerHeader* dtlsRH;
+    int records = 0;
+    size_t recordLen;
+    while (len > 0) {
+        records++;
+        dtlsRH = (DtlsRecordLayerHeader*)b;
+        recordLen = (dtlsRH->length[0] << 8) | dtlsRH->length[1];
+        if (recordLen > (size_t)len)
+            break;
+        b += sizeof(DtlsRecordLayerHeader) + recordLen;
+        len -= sizeof(DtlsRecordLayerHeader) + recordLen;
+    }
+    return records;
+}
+#endif
+
+int test_dtls_frag_ch(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13) \
+    && defined(WOLFSSL_DTLS_MTU) && defined(WOLFSSL_DTLS_CH_FRAG) && \
+    defined(WOLFSSL_AES_256)
+    WOLFSSL_CTX *ctx_c = NULL;
+    WOLFSSL_CTX *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL;
+    WOLFSSL *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+    static unsigned int DUMMY_MTU = 256;
+    unsigned int len;
+    unsigned char four_frag_CH[] = {
+      0x16, 0xfe, 0xfd, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+      0xda, 0x01, 0x00, 0x02, 0xdc, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+      0xce, 0xfe, 0xfd, 0xf3, 0x94, 0x01, 0x33, 0x2c, 0xcf, 0x2c, 0x47, 0xb1,
+      0xe5, 0xa1, 0x7b, 0x19, 0x3e, 0xac, 0x68, 0xdd, 0xe6, 0x17, 0x6b, 0x85,
+      0xad, 0x5f, 0xfc, 0x7f, 0x6e, 0xf0, 0xb9, 0xe0, 0x2e, 0xca, 0x47, 0x00,
+      0x00, 0x00, 0x36, 0x13, 0x01, 0x13, 0x02, 0x13, 0x03, 0xc0, 0x2c, 0xc0,
+      0x2b, 0xc0, 0x30, 0xc0, 0x2f, 0x00, 0x9f, 0x00, 0x9e, 0xcc, 0xa9, 0xcc,
+      0xa8, 0xcc, 0xaa, 0xc0, 0x27, 0xc0, 0x23, 0xc0, 0x28, 0xc0, 0x24, 0xc0,
+      0x0a, 0xc0, 0x09, 0xc0, 0x14, 0xc0, 0x13, 0x00, 0x6b, 0x00, 0x67, 0x00,
+      0x39, 0x00, 0x33, 0xcc, 0x14, 0xcc, 0x13, 0xcc, 0x15, 0x01, 0x00, 0x02,
+      0x7c, 0x00, 0x2b, 0x00, 0x03, 0x02, 0xfe, 0xfc, 0x00, 0x0d, 0x00, 0x20,
+      0x00, 0x1e, 0x06, 0x03, 0x05, 0x03, 0x04, 0x03, 0x02, 0x03, 0x08, 0x06,
+      0x08, 0x0b, 0x08, 0x05, 0x08, 0x0a, 0x08, 0x04, 0x08, 0x09, 0x06, 0x01,
+      0x05, 0x01, 0x04, 0x01, 0x03, 0x01, 0x02, 0x01, 0x00, 0x0a, 0x00, 0x0c,
+      0x00, 0x0a, 0x00, 0x19, 0x00, 0x18, 0x00, 0x17, 0x00, 0x15, 0x01, 0x00,
+      0x00, 0x16, 0x00, 0x00, 0x00, 0x33, 0x02, 0x39, 0x02, 0x37, 0x00, 0x17,
+      0x00, 0x41, 0x04, 0x94, 0xdf, 0x36, 0xd7, 0xb3, 0x90, 0x6d, 0x01, 0xa1,
+      0xe6, 0xed, 0x67, 0xf4, 0xd9, 0x9d, 0x2c, 0xac, 0x57, 0x74, 0xff, 0x19,
+      0xbe, 0x5a, 0xc9, 0x30, 0x11, 0xb7, 0x2b, 0x59, 0x47, 0x80, 0x7c, 0xa9,
+      0xb7, 0x31, 0x8c, 0x16, 0xfe, 0xfd, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+      0x00, 0x01, 0x00, 0xda, 0x01, 0x00, 0x02, 0xdc, 0x00, 0x00, 0x00, 0x00,
+      0xce, 0x00, 0x00, 0xce, 0x9e, 0x13, 0x74, 0x3b, 0x86, 0xba, 0x69, 0x1f,
+      0x12, 0xf7, 0xcd, 0x78, 0x53, 0xe8, 0x50, 0x4d, 0x71, 0x3f, 0x4b, 0x4e,
+      0xeb, 0x3e, 0xe5, 0x43, 0x54, 0x78, 0x17, 0x6d, 0x00, 0x18, 0x00, 0x61,
+      0x04, 0xd1, 0x99, 0x66, 0x4f, 0xda, 0xc7, 0x12, 0x3b, 0xff, 0xb2, 0xd6,
+      0x2f, 0x35, 0xb6, 0x17, 0x1f, 0xb3, 0xd0, 0xb6, 0x52, 0xff, 0x97, 0x8b,
+      0x01, 0xe8, 0xd9, 0x68, 0x71, 0x40, 0x02, 0xd5, 0x68, 0x3a, 0x58, 0xb2,
+      0x5d, 0xee, 0xa4, 0xe9, 0x5f, 0xf4, 0xaf, 0x3e, 0x30, 0x9c, 0x3e, 0x2b,
+      0xda, 0x61, 0x43, 0x99, 0x02, 0x35, 0x33, 0x9f, 0xcf, 0xb5, 0xd3, 0x28,
+      0x19, 0x9d, 0x1c, 0xbe, 0x69, 0x07, 0x9e, 0xfc, 0xe4, 0x8e, 0xcd, 0x86,
+      0x4a, 0x1b, 0xf0, 0xfc, 0x17, 0x94, 0x66, 0x53, 0xda, 0x24, 0x5e, 0xaf,
+      0xce, 0xec, 0x62, 0x4c, 0x06, 0xb4, 0x52, 0x94, 0xb1, 0x4a, 0x7a, 0x8c,
+      0x4f, 0x00, 0x19, 0x00, 0x85, 0x04, 0x00, 0x27, 0xeb, 0x99, 0x49, 0x7f,
+      0xcb, 0x2c, 0x46, 0x54, 0x2d, 0x93, 0x5d, 0x25, 0x92, 0x58, 0x5e, 0x06,
+      0xc3, 0x7c, 0xfb, 0x9a, 0xa7, 0xec, 0xcd, 0x9f, 0xe1, 0x6b, 0x2d, 0x78,
+      0xf5, 0x16, 0xa9, 0x20, 0x52, 0x48, 0x19, 0x0f, 0x1a, 0xd0, 0xce, 0xd8,
+      0x68, 0xb1, 0x4e, 0x7f, 0x33, 0x03, 0x7d, 0x0c, 0x39, 0xdb, 0x9c, 0x4b,
+      0xf4, 0xe7, 0xc2, 0xf5, 0xdd, 0x51, 0x9b, 0x03, 0xa8, 0x53, 0x2b, 0xe6,
+      0x00, 0x15, 0x4b, 0xff, 0xd2, 0xa0, 0x16, 0xfe, 0xfd, 0x00, 0x00, 0x00,
+      0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0xda, 0x01, 0x00, 0x02, 0xdc, 0x00,
+      0x00, 0x00, 0x01, 0x9c, 0x00, 0x00, 0xce, 0x58, 0x30, 0x10, 0x3d, 0x46,
+      0xcc, 0xca, 0x1a, 0x44, 0xc8, 0x58, 0x9b, 0x27, 0x17, 0x67, 0x31, 0x96,
+      0x8a, 0x66, 0x39, 0xf4, 0xcc, 0xc1, 0x9f, 0x12, 0x1f, 0x01, 0x30, 0x50,
+      0x16, 0xd6, 0x89, 0x97, 0xa3, 0x66, 0xd7, 0x99, 0x50, 0x09, 0x6e, 0x80,
+      0x87, 0xe4, 0xa2, 0x88, 0xae, 0xb4, 0x23, 0x57, 0x2f, 0x12, 0x60, 0xe7,
+      0x7d, 0x44, 0x2d, 0xad, 0xbe, 0xe9, 0x0d, 0x01, 0x00, 0x01, 0x00, 0xd5,
+      0xdd, 0x62, 0xee, 0xf3, 0x0e, 0xd9, 0x30, 0x0e, 0x38, 0xf3, 0x48, 0xf4,
+      0xc9, 0x8f, 0x8c, 0x20, 0xf7, 0xd3, 0xa8, 0xb3, 0x87, 0x3c, 0x98, 0x5d,
+      0x70, 0xc5, 0x03, 0x76, 0xb7, 0xd5, 0x0b, 0x7b, 0x23, 0x97, 0x6b, 0xe3,
+      0xb5, 0x18, 0xeb, 0x64, 0x55, 0x18, 0xb2, 0x8a, 0x90, 0x1a, 0x8f, 0x0e,
+      0x15, 0xda, 0xb1, 0x8e, 0x7f, 0xee, 0x1f, 0xe0, 0x3b, 0xb9, 0xed, 0xfc,
+      0x4e, 0x3f, 0x78, 0x16, 0x39, 0x95, 0x5f, 0xb7, 0xcb, 0x65, 0x55, 0x72,
+      0x7b, 0x7d, 0x86, 0x2f, 0x8a, 0xe5, 0xee, 0xf7, 0x57, 0x40, 0xf3, 0xc4,
+      0x96, 0x4f, 0x11, 0x4d, 0x85, 0xf9, 0x56, 0xfa, 0x3d, 0xf0, 0xc9, 0xa4,
+      0xec, 0x1e, 0xaa, 0x47, 0x90, 0x53, 0xdf, 0xe1, 0xb7, 0x78, 0x18, 0xeb,
+      0xdd, 0x0d, 0x89, 0xb7, 0xf6, 0x15, 0x0e, 0x55, 0x12, 0xb3, 0x23, 0x17,
+      0x0b, 0x59, 0x6f, 0x83, 0x05, 0x6b, 0xa6, 0xf8, 0x6c, 0x3a, 0x9b, 0x1b,
+      0x50, 0x93, 0x51, 0xea, 0x95, 0x2d, 0x99, 0x96, 0x38, 0x16, 0xfe, 0xfd,
+      0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03, 0x00, 0x7e, 0x01, 0x00,
+      0x02, 0xdc, 0x00, 0x00, 0x00, 0x02, 0x6a, 0x00, 0x00, 0x72, 0x2d, 0x66,
+      0x3e, 0xf2, 0x36, 0x5a, 0xf2, 0x23, 0x8f, 0x28, 0x09, 0xa9, 0x55, 0x8c,
+      0x8f, 0xc0, 0x0d, 0x61, 0x98, 0x33, 0x56, 0x87, 0x7a, 0xfd, 0xa7, 0x50,
+      0x71, 0x84, 0x2e, 0x41, 0x58, 0x00, 0x87, 0xd9, 0x27, 0xe5, 0x7b, 0xf4,
+      0x6d, 0x84, 0x4e, 0x2e, 0x0c, 0x80, 0x0c, 0xf3, 0x8a, 0x02, 0x4b, 0x99,
+      0x3a, 0x1f, 0x9f, 0x18, 0x7d, 0x1c, 0xec, 0xad, 0x60, 0x54, 0xa6, 0xa3,
+      0x2c, 0x82, 0x5e, 0xf8, 0x8f, 0xae, 0xe1, 0xc4, 0x82, 0x7e, 0x43, 0x43,
+      0xc5, 0x99, 0x49, 0x05, 0xd3, 0xf6, 0xdf, 0xa1, 0xb5, 0x2d, 0x0c, 0x13,
+      0x2f, 0x1e, 0xb6, 0x28, 0x7c, 0x5c, 0xa1, 0x02, 0x6b, 0x8d, 0xa3, 0xeb,
+      0xd4, 0x58, 0xe6, 0xa0, 0x7e, 0x6b, 0xaa, 0x09, 0x43, 0x67, 0x71, 0x87,
+      0xa5, 0xcb, 0x68, 0xf3
+    };
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
+
+    /* Fragment msgs */
+    ExpectIntEQ(wolfSSL_dtls_set_mtu(ssl_c, DUMMY_MTU), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_dtls_set_mtu(ssl_s, DUMMY_MTU), WOLFSSL_SUCCESS);
+
+    /* Add in some key shares to make the CH long */
+    ExpectIntEQ(wolfSSL_UseKeyShare(ssl_c, WOLFSSL_ECC_SECP256R1),
+            WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_UseKeyShare(ssl_c, WOLFSSL_ECC_SECP384R1),
+            WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_UseKeyShare(ssl_c, WOLFSSL_ECC_SECP521R1),
+            WOLFSSL_SUCCESS);
+#ifdef HAVE_FFDHE_2048
+    ExpectIntEQ(wolfSSL_UseKeyShare(ssl_c, WOLFSSL_FFDHE_2048),
+            WOLFSSL_SUCCESS);
+#endif
+#ifdef HAVE_FFDHE_3072
+    ExpectIntEQ(wolfSSL_UseKeyShare(ssl_c, WOLFSSL_FFDHE_3072),
+            WOLFSSL_SUCCESS);
+#endif
+#ifdef HAVE_FFDHE_4096
+    ExpectIntEQ(wolfSSL_UseKeyShare(ssl_c, WOLFSSL_FFDHE_4096),
+            WOLFSSL_SUCCESS);
+#endif
+
+    ExpectIntEQ(wolfSSL_dtls13_allow_ch_frag(ssl_s, 1), WOLFSSL_SUCCESS);
+
+    /* Reject fragmented first CH */
+    ExpectIntEQ(test_dtls_frag_ch_count_records(four_frag_CH,
+            sizeof(four_frag_CH)), 4);
+    len = sizeof(four_frag_CH);
+    test_memio_clear_buffer(&test_ctx, 0);
+    while (len > 0 && EXPECT_SUCCESS()) {
+        unsigned int inj_len = len > DUMMY_MTU ? DUMMY_MTU : len;
+        unsigned char *idx = four_frag_CH + sizeof(four_frag_CH) - len;
+        ExpectIntEQ(test_memio_inject_message(&test_ctx, 0, (const char *)idx,
+            inj_len), 0);
+        len -= inj_len;
+    }
+    ExpectIntEQ(test_ctx.s_len, sizeof(four_frag_CH));
+    while (test_ctx.s_len > 0 && EXPECT_SUCCESS()) {
+        int s_len = test_ctx.s_len;
+        ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+        ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+        /* Fail if we didn't advance the buffer to avoid infinite loops */
+        ExpectIntLT(test_ctx.s_len, s_len);
+    }
+    /* Expect all fragments read */
+    ExpectIntEQ(test_ctx.s_len, 0);
+    /* Expect quietly dropping fragmented first CH */
+    ExpectIntEQ(test_ctx.c_len, 0);
+
+#if defined(WOLFSSL_TLS13) && defined(HAVE_ECH)
+    /* Disable ECH as it pushes it over our MTU */
+    wolfSSL_SetEchEnable(ssl_c, 0);
+#endif
+
+    /* Limit options to make the CH a fixed length */
+    /* See wolfSSL_parse_cipher_list for reason why we provide 1.3 AND 1.2
+     * ciphersuite. This is only necessary when building with OPENSSL_EXTRA. */
+#ifdef OPENSSL_EXTRA
+    ExpectTrue(wolfSSL_set_cipher_list(ssl_c, "TLS13-AES256-GCM-SHA384"
+                                       ":DHE-RSA-AES256-GCM-SHA384"));
+#else
+    ExpectTrue(wolfSSL_set_cipher_list(ssl_c, "TLS13-AES256-GCM-SHA384"));
+#endif
+
+    /* CH1 */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Count records. Expect 1 unfragmented CH */
+    ExpectIntEQ(test_dtls_frag_ch_count_records(test_ctx.s_buff,
+            test_ctx.s_len), 1);
+    /* HRR */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    /* CH2 */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Count records. Expect fragmented CH */
+    ExpectIntGT(test_dtls_frag_ch_count_records(test_ctx.s_buff,
+            test_ctx.s_len), 1);
+
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_CTX_free(ctx_s);
+    ssl_c = ssl_s = NULL;
+    ctx_c = ctx_s = NULL;
+#endif
+    return EXPECT_RESULT();
+}
+
+int test_dtls_empty_keyshare_with_cookie(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
+    WOLFSSL_CTX *ctx_s = NULL;
+    WOLFSSL *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+    unsigned char ch_empty_keyshare_with_cookie[] = {
+        0x16, 0xfe, 0xfd, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01,
+        0x12, 0x01, 0x00, 0x01, 0x06, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x01,
+        0x06, 0xfe, 0xfd, 0xfb, 0x8c, 0x9b, 0x28, 0xae, 0x50, 0x1c, 0x4d, 0xf3,
+        0xb8, 0xcf, 0x4d, 0xd8, 0x7e, 0x93, 0x13, 0x7b, 0x9e, 0xd9, 0xeb, 0xe9,
+        0x13, 0x4b, 0x0d, 0x7f, 0x2e, 0x43, 0x62, 0x8c, 0xe4, 0x57, 0x79, 0x00,
+        0x00, 0x00, 0x36, 0x13, 0x01, 0x13, 0x02, 0x13, 0x03, 0xc0, 0x2c, 0xc0,
+        0x2b, 0xc0, 0x30, 0xc0, 0x2f, 0x00, 0x9f, 0x00, 0x9e, 0xcc, 0xa9, 0xcc,
+        0xa8, 0xcc, 0xaa, 0xc0, 0x27, 0xc0, 0x23, 0xc0, 0x28, 0xc0, 0x24, 0xc0,
+        0x0a, 0xc0, 0x09, 0xc0, 0x14, 0xc0, 0x13, 0x00, 0x6b, 0x00, 0x67, 0x00,
+        0x39, 0x00, 0x33, 0xcc, 0x14, 0xcc, 0x13, 0xcc, 0x15, 0x01, 0x00, 0x00,
+        0xa6, 0x00, 0x2b, 0x00, 0x03, 0x02, 0xfe, 0xfc, 0x00, 0x2c, 0x00, 0x47,
+        0x00, 0x45, 0x20, 0xee, 0x4b, 0x17, 0x70, 0x63, 0xa0, 0x4c, 0x82, 0xbf,
+        0x43, 0x01, 0x7d, 0x8d, 0xc1, 0x1b, 0x4e, 0x9b, 0xa0, 0x3c, 0x53, 0x1f,
+        0xb7, 0xd1, 0x10, 0x81, 0xa8, 0xdf, 0xdf, 0x8c, 0x7f, 0xf3, 0x11, 0x13,
+        0x01, 0x02, 0x3d, 0x3b, 0x7d, 0x14, 0x2c, 0x31, 0xb3, 0x60, 0x72, 0x4d,
+        0xe5, 0x1a, 0xb2, 0xa3, 0x61, 0x77, 0x73, 0x03, 0x40, 0x0e, 0x5f, 0xc5,
+        0x61, 0x38, 0x43, 0x56, 0x21, 0x4a, 0x95, 0xd5, 0x35, 0xa8, 0x0d, 0x00,
+        0x0d, 0x00, 0x2a, 0x00, 0x28, 0x06, 0x03, 0x05, 0x03, 0x04, 0x03, 0x02,
+        0x03, 0xfe, 0x0b, 0xfe, 0x0e, 0xfe, 0xa0, 0xfe, 0xa3, 0xfe, 0xa5, 0x08,
+        0x06, 0x08, 0x0b, 0x08, 0x05, 0x08, 0x0a, 0x08, 0x04, 0x08, 0x09, 0x06,
+        0x01, 0x05, 0x01, 0x04, 0x01, 0x03, 0x01, 0x02, 0x01, 0x00, 0x0a, 0x00,
+        0x18, 0x00, 0x16, 0x00, 0x19, 0x00, 0x18, 0x00, 0x17, 0x00, 0x15, 0x01,
+        0x00, 0x02, 0x3a, 0x02, 0x3c, 0x02, 0x3d, 0x2f, 0x3a, 0x2f, 0x3c, 0x2f,
+        0x3d, 0x00, 0x16, 0x00, 0x00, 0x00, 0x33, 0x00, 0x02, 0x00, 0x00
+    };
+    DtlsRecordLayerHeader* dtlsRH;
+    byte sequence_number[8];
+
+    XMEMSET(&sequence_number, 0, sizeof(sequence_number));
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+    ExpectIntEQ(test_memio_setup(&test_ctx, NULL, &ctx_s, NULL, &ssl_s,
+        NULL, wolfDTLSv1_3_server_method), 0);
+    ExpectIntEQ(test_memio_inject_message(&test_ctx, 0,
+            (const char *)ch_empty_keyshare_with_cookie,
+            sizeof(ch_empty_keyshare_with_cookie)), 0);
+
+    /* CH1 */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Expect an alert. A plaintext alert should be exactly 15 bytes. */
+    ExpectIntEQ(test_ctx.c_len, 15);
+    dtlsRH = (DtlsRecordLayerHeader*)test_ctx.c_buff;
+    ExpectIntEQ(dtlsRH->type, alert);
+    ExpectIntEQ(dtlsRH->pvMajor, DTLS_MAJOR);
+    ExpectIntEQ(dtlsRH->pvMinor, DTLSv1_2_MINOR);
+    sequence_number[7] = 1;
+    ExpectIntEQ(XMEMCMP(sequence_number, dtlsRH->sequence_number,
+            sizeof(sequence_number)), 0);
+    ExpectIntEQ(dtlsRH->length[0], 0);
+    ExpectIntEQ(dtlsRH->length[1], 2);
+    ExpectIntEQ(test_ctx.c_buff[13], alert_fatal);
+    ExpectIntEQ(test_ctx.c_buff[14], illegal_parameter);
+
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
+
+int test_dtls13_missing_finished_client(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
+    WOLFSSL_CTX *ctx_c = NULL;
+    WOLFSSL_CTX *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL;
+    WOLFSSL *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+    const char test_str[] = "test string";
+    char test_buf[sizeof(test_str)];
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
+
+    /* CH1 */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* HRR */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    /* CH2 */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Server first flight with finished */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Let's clear the output */
+    test_memio_clear_buffer(&test_ctx, 1);
+    /* Let's send some app data */
+    ExpectIntEQ(wolfSSL_write(ssl_s, test_str, sizeof(test_str)),
+                sizeof(test_str));
+    /* Client second flight with finished */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Server should not error out on a missing finished */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Client rtx second flight with finished */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Server */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), 1);
+    /* Client */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), 1);
+    /* Let's send some app data */
+    ExpectIntEQ(wolfSSL_write(ssl_s, test_str, sizeof(test_str)),
+                sizeof(test_str));
+    ExpectIntEQ(wolfSSL_read(ssl_c, test_buf, sizeof(test_buf)),
+                sizeof(test_str));
+    ExpectBufEQ(test_buf, test_str, sizeof(test_str));
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
+
+int test_dtls13_missing_finished_server(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
+    WOLFSSL_CTX *ctx_c = NULL;
+    WOLFSSL_CTX *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL;
+    WOLFSSL *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+    const char test_str[] = "test string";
+    char test_buf[sizeof(test_str)];
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
+
+    /* CH1 */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* HRR */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    /* CH2 */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Server first flight with finished */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Client second flight with finished */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Let's clear the output */
+    test_memio_clear_buffer(&test_ctx, 0);
+    ExpectFalse(wolfSSL_is_init_finished(ssl_c));
+    /* Let's send some app data */
+    ExpectIntEQ(wolfSSL_write(ssl_c, test_str, sizeof(test_str)),
+                sizeof(test_str));
+    /* Server should not error out on a missing finished */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Client rtx second flight with finished */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Server first flight with finished */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), 1);
+    /* Let's send some app data */
+    ExpectIntEQ(wolfSSL_write(ssl_c, test_str, sizeof(test_str)),
+                sizeof(test_str));
+    ExpectIntEQ(wolfSSL_read(ssl_s, test_buf, sizeof(test_buf)),
+                sizeof(test_str));
+    ExpectBufEQ(test_buf, test_str, sizeof(test_str));
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
+
+int test_dtls13_finished_send_error_propagation(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
+    WOLFSSL_CTX *ctx_c = NULL;
+    WOLFSSL_CTX *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL;
+    WOLFSSL *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
+
+    /* CH1 */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* HRR */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    /* CH2 */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Server first flight with finished */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    /* Client second flight with finished - block sends to force error */
+    test_ctx.s_len = TEST_MEMIO_BUF_SZ;
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    /* Verify the error is propagated, not silently swallowed as success */
+    ExpectIntNE(wolfSSL_get_error(ssl_c, -1), 0);
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
+
+
+/*----------------------------------------------------------------------------*/
+/* DTLSv1.3-only tests moved from test_dtls.c (isolated from DTLS<=1.2 code)  */
+/*----------------------------------------------------------------------------*/
+
+/*-- basic_connection_id (test_dtls.c lines 422,580) ---*/
+int test_dtls13_basic_connection_id(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13) \
+    && defined(WOLFSSL_DTLS_CID)
+    unsigned char client_cid[] = { 9, 8, 7, 6, 5, 4, 3, 2, 1, 0 };
+    unsigned char server_cid[] = { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9 };
+    unsigned char readBuf[50];
+    void *        cid = NULL;
+    const char* params[] = {
+#ifndef NO_SHA256
+#ifdef WOLFSSL_AES_128
+#ifdef HAVE_AESGCM
+        "TLS13-AES128-GCM-SHA256",
+#endif
+#if defined(HAVE_CHACHA) && defined(HAVE_POLY1305)
+        "TLS13-CHACHA20-POLY1305-SHA256",
+#endif
+#ifdef HAVE_AESCCM
+        "TLS13-AES128-CCM-8-SHA256",
+        "TLS13-AES128-CCM-SHA256",
+#endif
+#endif
+#ifdef HAVE_NULL_CIPHER
+        "TLS13-SHA256-SHA256",
+#endif
+#endif
+    };
+    size_t i;
+
+    /* We check if the side included the CID in their output */
+#define CLIENT_CID() mymemmem(test_ctx.s_buff, test_ctx.s_len, \
+                              client_cid, sizeof(client_cid))
+#define SERVER_CID() mymemmem(test_ctx.c_buff, test_ctx.c_len, \
+                              server_cid, sizeof(server_cid))
+#define RESET_CID(cid) if ((cid) != NULL) { \
+                           ((char*)(cid))[0] = -1; \
+                       }
+
+
+    printf("\n");
+    for (i = 0; i < XELEM_CNT(params) && EXPECT_SUCCESS(); i++) {
+        WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
+        WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
+        struct test_memio_ctx test_ctx;
+
+        printf("Testing %s ... ", params[i]);
+
+        XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+
+        ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+            wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
+
+        ExpectIntEQ(wolfSSL_set_cipher_list(ssl_c, params[i]), WOLFSSL_SUCCESS);
+        ExpectIntEQ(wolfSSL_set_cipher_list(ssl_s, params[i]), WOLFSSL_SUCCESS);
+
+        ExpectIntEQ(wolfSSL_dtls_cid_use(ssl_c), 1);
+        ExpectIntEQ(wolfSSL_dtls_cid_set(ssl_c, server_cid, sizeof(server_cid)),
+                1);
+        ExpectIntEQ(wolfSSL_dtls_cid_use(ssl_s), 1);
+        ExpectIntEQ(wolfSSL_dtls_cid_set(ssl_s, client_cid, sizeof(client_cid)),
+                1);
+
+        /* CH1 */
+        ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+        ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+        ExpectNull(CLIENT_CID());
+        /* HRR */
+        ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+        ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+        ExpectNull(SERVER_CID());
+        /* CH2 */
+        ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+        ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+        ExpectNull(CLIENT_CID());
+        /* Server first flight */
+        ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+        ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+        ExpectNotNull(SERVER_CID());
+        /* Client second flight */
+        ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+        ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+        ExpectNotNull(CLIENT_CID());
+        /* Server process flight */
+        ExpectIntEQ(wolfSSL_negotiate(ssl_s), 1);
+        /* Client process flight */
+        ExpectIntEQ(wolfSSL_negotiate(ssl_c), 1);
+
+        /* Write some data */
+        ExpectIntEQ(wolfSSL_write(ssl_c, params[i], (int)XSTRLEN(params[i])),
+                XSTRLEN(params[i]));
+        ExpectNotNull(CLIENT_CID());
+        ExpectIntEQ(wolfSSL_write(ssl_s, params[i], (int)XSTRLEN(params[i])),
+                XSTRLEN(params[i]));
+        ExpectNotNull(SERVER_CID());
+        /* Read the data */
+        XMEMSET(readBuf, 0, sizeof(readBuf));
+        ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)),
+                XSTRLEN(params[i]));
+        ExpectStrEQ(readBuf, params[i]);
+        XMEMSET(readBuf, 0, sizeof(readBuf));
+        ExpectIntEQ(wolfSSL_read(ssl_s, readBuf, sizeof(readBuf)),
+                XSTRLEN(params[i]));
+        ExpectStrEQ(readBuf, params[i]);
+        /* Write short data */
+        ExpectIntEQ(wolfSSL_write(ssl_c, params[i], 1), 1);
+        ExpectNotNull(CLIENT_CID());
+        ExpectIntEQ(wolfSSL_write(ssl_s, params[i], 1), 1);
+        ExpectNotNull(SERVER_CID());
+        /* Read the short data */
+        XMEMSET(readBuf, 0, sizeof(readBuf));
+        ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), 1);
+        ExpectIntEQ(readBuf[0], params[i][0]);
+        XMEMSET(readBuf, 0, sizeof(readBuf));
+        ExpectIntEQ(wolfSSL_read(ssl_s, readBuf, sizeof(readBuf)), 1);
+        ExpectIntEQ(readBuf[0], params[i][0]);
+        /* Write some data but with wrong CID */
+        ExpectIntEQ(wolfSSL_write(ssl_c, params[i], (int)XSTRLEN(params[i])),
+                XSTRLEN(params[i]));
+        /* Reset client cid. */
+        ExpectNotNull(cid = CLIENT_CID());
+        RESET_CID(cid);
+        ExpectIntEQ(wolfSSL_write(ssl_s, params[i], (int)XSTRLEN(params[i])),
+                XSTRLEN(params[i]));
+        /* Reset server cid. */
+        ExpectNotNull(cid = SERVER_CID());
+        RESET_CID(cid);
+        /* Try to read the data but it shouldn't be there */
+        ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), -1);
+        ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+        ExpectIntEQ(wolfSSL_read(ssl_s, readBuf, sizeof(readBuf)), -1);
+        ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+
+        /* Close connection */
+        ExpectIntEQ(wolfSSL_shutdown(ssl_c), WOLFSSL_SHUTDOWN_NOT_DONE);
+        ExpectNotNull(CLIENT_CID());
+        ExpectIntEQ(wolfSSL_shutdown(ssl_s), WOLFSSL_SHUTDOWN_NOT_DONE);
+        ExpectNotNull(SERVER_CID());
+        ExpectIntEQ(wolfSSL_shutdown(ssl_c), 1);
+        ExpectIntEQ(wolfSSL_shutdown(ssl_s), 1);
+
+        if (EXPECT_SUCCESS())
+            printf("ok\n");
+        else
+            printf("failed\n");
+
+        wolfSSL_free(ssl_c);
+        wolfSSL_CTX_free(ctx_c);
+        wolfSSL_free(ssl_s);
+        wolfSSL_CTX_free(ctx_s);
+    }
+
+#undef CLIENT_CID
+#undef SERVER_CID
+#undef RESET_CID
+
+#endif
+    return EXPECT_RESULT();
+}
+
+/*-- hrr_want_write (test_dtls.c lines 588,639) ---*/
+int test_dtls13_hrr_want_write(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
+    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
+    const char msg[] = "hello";
+    const int msgLen = sizeof(msg);
+    struct test_memio_ctx test_ctx;
+    char readBuf[sizeof(msg)];
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+                    wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method),
+        0);
+
+    /* Client sends first ClientHello */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+
+    /* Force server to hit WANT_WRITE when producing the HRR */
+    test_memio_simulate_want_write(&test_ctx, 0, 1);
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_WRITE);
+
+    /* Allow the server to flush the HRR and proceed */
+    test_memio_simulate_want_write(&test_ctx, 0, 0);
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+
+    /* Resume the DTLS 1.3 handshake */
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+
+    /* Verify post-handshake application data in both directions */
+    XMEMSET(readBuf, 0, sizeof(readBuf));
+    ExpectIntEQ(wolfSSL_write(ssl_c, msg, msgLen), msgLen);
+    ExpectIntEQ(wolfSSL_read(ssl_s, readBuf, sizeof(readBuf)), msgLen);
+    ExpectStrEQ(readBuf, msg);
+
+    XMEMSET(readBuf, 0, sizeof(readBuf));
+    ExpectIntEQ(wolfSSL_write(ssl_s, msg, msgLen), msgLen);
+    ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), msgLen);
+    ExpectStrEQ(readBuf, msg);
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
+
+/*-- want_write_send_cb_helper (test_dtls.c lines 641,655) ---*/
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
+struct test_dtls13_wwrite_ctx {
+    int want_write;
+    struct test_memio_ctx *text_ctx;
+};
+static int test_dtls13_want_write_send_cb(WOLFSSL *ssl, char *data, int sz, void *ctx)
+{
+    struct test_dtls13_wwrite_ctx *wwctx = (struct test_dtls13_wwrite_ctx *)ctx;
+    wwctx->want_write = !wwctx->want_write;
+    if (wwctx->want_write) {
+        return WOLFSSL_CBIO_ERR_WANT_WRITE;
+    }
+    return test_memio_write_cb(ssl, data, sz, wwctx->text_ctx);
+}
+#endif
+
+/*-- every_write_want_write (test_dtls.c lines 665,740) ---*/
+int test_dtls13_every_write_want_write(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
+    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+    const char msg[] = "want-write";
+    const int msgLen = sizeof(msg);
+    char readBuf[sizeof(msg)];
+    struct test_dtls13_wwrite_ctx wwctx_c;
+    struct test_dtls13_wwrite_ctx wwctx_s;
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+                    wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method),
+        0);
+
+    wwctx_c.want_write = 0;
+    wwctx_c.text_ctx = &test_ctx;
+    wolfSSL_SetIOWriteCtx(ssl_c, &wwctx_c);
+    wolfSSL_SSLSetIOSend(ssl_c, test_dtls13_want_write_send_cb);
+    wwctx_s.want_write = 0;
+    wwctx_s.text_ctx = &test_ctx;
+    wolfSSL_SetIOWriteCtx(ssl_s, &wwctx_s);
+    wolfSSL_SSLSetIOSend(ssl_s, test_dtls13_want_write_send_cb);
+
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 20, NULL), 0);
+
+    ExpectTrue(wolfSSL_is_init_finished(ssl_c));
+    ExpectTrue(wolfSSL_is_init_finished(ssl_s));
+
+    test_memio_simulate_want_write(&test_ctx, 0, 0);
+    test_memio_simulate_want_write(&test_ctx, 1, 0);
+
+    wolfSSL_SetIOWriteCtx(ssl_c, &test_ctx);
+    wolfSSL_SSLSetIOSend(ssl_c, test_memio_write_cb);
+    wolfSSL_SetIOWriteCtx(ssl_s, &test_ctx);
+    wolfSSL_SSLSetIOSend(ssl_s, test_memio_write_cb);
+
+    XMEMSET(readBuf, 0, sizeof(readBuf));
+    ExpectIntEQ(wolfSSL_write(ssl_c, msg, msgLen), msgLen);
+    ExpectIntEQ(wolfSSL_read(ssl_s, readBuf, sizeof(readBuf)), msgLen);
+    ExpectStrEQ(readBuf, msg);
+
+    XMEMSET(readBuf, 0, sizeof(readBuf));
+    ExpectIntEQ(wolfSSL_write(ssl_s, msg, msgLen), msgLen);
+    ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), msgLen);
+    ExpectStrEQ(readBuf, msg);
+
+    test_memio_simulate_want_write(&test_ctx, 0, 1);
+    XMEMSET(readBuf, 0, sizeof(readBuf));
+    ExpectIntEQ(wolfSSL_write(ssl_s, msg, msgLen), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_WRITE);
+    test_memio_simulate_want_write(&test_ctx, 0, 0);
+    ExpectIntEQ(wolfSSL_write(ssl_s, msg, msgLen), msgLen);
+    ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), msgLen);
+    ExpectStrEQ(readBuf, msg);
+
+    XMEMSET(readBuf, 0, sizeof(readBuf));
+    test_memio_simulate_want_write(&test_ctx, 1, 1);
+    ExpectIntEQ(wolfSSL_write(ssl_c, msg, msgLen), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_WRITE);
+    test_memio_simulate_want_write(&test_ctx, 1, 0);
+    ExpectIntEQ(wolfSSL_write(ssl_c, msg, msgLen), msgLen);
+    ExpectIntEQ(wolfSSL_read(ssl_s, readBuf, sizeof(readBuf)), msgLen);
+    ExpectStrEQ(readBuf, msg);
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
+
+/*-- epochs (test_dtls.c lines 823,871) ---*/
+int test_dtls13_epochs(void) {
+    EXPECT_DECLS;
+#if defined(WOLFSSL_DTLS13) && !defined(NO_WOLFSSL_CLIENT)
+    WOLFSSL_CTX* ctx = NULL;
+    WOLFSSL* ssl = NULL;
+    byte input[20];
+    word32 inOutIdx = 0;
+
+    XMEMSET(input, 0, sizeof(input));
+
+    ExpectNotNull(ctx = wolfSSL_CTX_new(wolfDTLSv1_3_client_method()));
+    ExpectNotNull(ssl = wolfSSL_new(ctx));
+    /* Some manual setup to enter the epoch check */
+    ExpectTrue(ssl->options.tls1_3 = 1);
+
+    inOutIdx = 0;
+    if (ssl != NULL) ssl->keys.curEpoch64 = w64From32(0x0, 0x0);
+    ExpectIntEQ(DoApplicationData(ssl, input, &inOutIdx, 0), SANITY_MSG_E);
+    inOutIdx = 0;
+    if (ssl != NULL) ssl->keys.curEpoch64 = w64From32(0x0, 0x2);
+    ExpectIntEQ(DoApplicationData(ssl, input, &inOutIdx, 0), SANITY_MSG_E);
+
+    if (ssl != NULL) ssl->keys.curEpoch64 = w64From32(0x0, 0x1);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, client_hello), SANITY_MSG_E);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, server_hello), SANITY_MSG_E);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, hello_verify_request), SANITY_MSG_E);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, hello_retry_request), SANITY_MSG_E);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, hello_request), SANITY_MSG_E);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, encrypted_extensions), SANITY_MSG_E);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, server_key_exchange), SANITY_MSG_E);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, server_hello_done), SANITY_MSG_E);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, client_key_exchange), SANITY_MSG_E);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, certificate_request), SANITY_MSG_E);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, certificate), SANITY_MSG_E);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, certificate_verify), SANITY_MSG_E);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, finished), SANITY_MSG_E);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, certificate_status), SANITY_MSG_E);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, change_cipher_hs), SANITY_MSG_E);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, key_update), SANITY_MSG_E);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, session_ticket), SANITY_MSG_E);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, end_of_early_data), SANITY_MSG_E);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, message_hash), SANITY_MSG_E);
+    ExpectIntEQ(Dtls13CheckEpoch(ssl, no_shake), SANITY_MSG_E);
+
+    wolfSSL_CTX_free(ctx);
+    wolfSSL_free(ssl);
+#endif
+    return EXPECT_RESULT();
+}
+
+/*-- ack_order (test_dtls.c lines 873,951) ---*/
+int test_dtls13_ack_order(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
+    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+    unsigned char readBuf[50];
+    word32 length = 0;
+    /* struct {
+     *     uint64 epoch;
+     *     uint64 sequence_number;
+     * } RecordNumber;
+     * Big endian */
+    static const unsigned char expected_output[] = {
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03,
+        0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x06,
+    };
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+
+    /* Get a populated DTLS object */
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+    ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), -1);
+    /* Clear the buffer of any extra messages */
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    ExpectIntEQ(wolfSSL_read(ssl_s, readBuf, sizeof(readBuf)), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    ExpectIntEQ(test_ctx.c_len, 0);
+    ExpectIntEQ(test_ctx.s_len, 0);
+
+    /* Add seen records */
+    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 3), w64From32(0, 2)), 0);
+    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 3), w64From32(0, 0)), 0);
+    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 3), w64From32(0, 1)), 0);
+    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 3), w64From32(0, 4)), 0);
+    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 2), w64From32(0, 0)), 0);
+    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 3), w64From32(0, 6)), 0);
+    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 3), w64From32(0, 6)), 0);
+    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 2), w64From32(0, 1)), 0);
+    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 2), w64From32(0, 2)), 0);
+    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 2), w64From32(0, 2)), 0);
+    ExpectIntEQ(Dtls13WriteAckMessage(ssl_c, ssl_c->dtls13Rtx.seenRecords,
+            ssl_c->dtls13Rtx.seenRecordsCount, &length), 0);
+
+    /* must zero the span reserved for the header to avoid read of uninited
+     * data.
+     */
+    XMEMSET(ssl_c->buffers.outputBuffer.buffer, 0,
+            5 /* DTLS13_UNIFIED_HEADER_SIZE */);
+    /* N * RecordNumber + 2 extra bytes for length */
+    ExpectIntEQ(length, sizeof(expected_output) + 2);
+    ExpectNotNull(mymemmem(ssl_c->buffers.outputBuffer.buffer,
+            ssl_c->buffers.outputBuffer.bufferSize, expected_output,
+            sizeof(expected_output)));
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
+
+/*-- ack_overflow (test_dtls.c lines 953,1010) ---*/
+int test_dtls13_ack_overflow(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13)
+    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+    unsigned char readBuf[50];
+    word32 length = 0;
+    int i;
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+    ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    ExpectIntEQ(wolfSSL_read(ssl_s, readBuf, sizeof(readBuf)), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+
+    /* Edge case 1: one below limit - all inserts must succeed */
+    for (i = 0; i < DTLS13_ACK_MAX_RECORDS - 1; i++) {
+        ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 0),
+                                    w64From32(0, (word32)i)), 0);
+    }
+    ExpectIntEQ(ssl_c->dtls13Rtx.seenRecordsCount, DTLS13_ACK_MAX_RECORDS - 1);
+
+    /* Edge case 2: insert the last allowed record - must succeed */
+    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 0),
+                    w64From32(0, (word32)(DTLS13_ACK_MAX_RECORDS - 1))), 0);
+    ExpectIntEQ(ssl_c->dtls13Rtx.seenRecordsCount, DTLS13_ACK_MAX_RECORDS);
+
+    /* Writing a full-but-valid list must succeed */
+    ExpectIntEQ(Dtls13WriteAckMessage(ssl_c, ssl_c->dtls13Rtx.seenRecords,
+                    ssl_c->dtls13Rtx.seenRecordsCount, &length), 0);
+
+    /* Edge case 3: one over limit - must be silently dropped */
+    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 0),
+                    w64From32(0, (word32)DTLS13_ACK_MAX_RECORDS)), 0);
+    ExpectIntEQ(ssl_c->dtls13Rtx.seenRecordsCount, DTLS13_ACK_MAX_RECORDS);
+
+    /* Bypass the insert guard to force the list one element over the limit,
+     * then verify Dtls13WriteAckMessage errors out instead of overflowing */
+    ssl_c->dtls13Rtx.seenRecordsCount = 0;
+    ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 1),
+                    w64From32(0, (word32)DTLS13_ACK_MAX_RECORDS)), 0);
+    ssl_c->dtls13Rtx.seenRecordsCount = (word16)(DTLS13_ACK_MAX_RECORDS + 1);
+    ExpectIntEQ(Dtls13WriteAckMessage(ssl_c, ssl_c->dtls13Rtx.seenRecords,
+                    ssl_c->dtls13Rtx.seenRecordsCount, &length), BUFFER_E);
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
+
+/*-- ack_dup_write_counter (test_dtls.c lines 1012,1069) ---*/
+int test_dtls13_ack_dup_write_counter(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13) \
+    && defined(HAVE_WRITE_DUP)
+    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
+    WOLFSSL *ssl_c2 = NULL;
+    struct test_memio_ctx test_ctx;
+    unsigned char readBuf[50];
+    int i;
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+    /* Drain any post-handshake messages */
+    ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    ExpectIntEQ(wolfSSL_read(ssl_s, readBuf, sizeof(readBuf)), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+
+    /* Split ssl_c: ssl_c becomes READ_DUP_SIDE, ssl_c2 becomes WRITE_DUP_SIDE */
+    ExpectNotNull(ssl_c2 = wolfSSL_write_dup(ssl_c));
+
+    /* Cycle 1: add records, trigger handoff, verify counter is reset to 0 */
+    for (i = 0; i < 5; i++)
+        ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 3),
+                                    w64From32(0, (word32)i)), 0);
+    ExpectIntEQ(ssl_c->dtls13Rtx.seenRecordsCount, 5);
+    ssl_c->dtls13Rtx.sendAcks = 1;
+    ExpectIntEQ(Dtls13DoScheduledWork(ssl_c), 0);
+    /* seenRecords ownership was transferred to dupWrite->sendAckList;
+     * seenRecordsCount must be reset to 0,  not left at 5. */
+    ExpectNull(ssl_c->dtls13Rtx.seenRecords);
+    ExpectIntEQ(ssl_c->dtls13Rtx.seenRecordsCount, 0);
+
+    /* Cycle 2 (different epoch to avoid the dup-filter): verify the counter
+     * did not accumulate across the previous transfer.  Without the fix,
+     * seenRecordsCount would now be 10 after this second batch. */
+    for (i = 0; i < 5; i++)
+        ExpectIntEQ(Dtls13RtxAddAck(ssl_c, w64From32(0, 4),
+                                    w64From32(0, (word32)i)), 0);
+    ExpectIntEQ(ssl_c->dtls13Rtx.seenRecordsCount, 5);
+    ssl_c->dtls13Rtx.sendAcks = 1;
+    ExpectIntEQ(Dtls13DoScheduledWork(ssl_c), 0);
+    ExpectNull(ssl_c->dtls13Rtx.seenRecords);
+    ExpectIntEQ(ssl_c->dtls13Rtx.seenRecordsCount, 0);
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_free(ssl_c2);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
+
+/*-- get_message_seq_helper (test_dtls.c lines 1850,1867) ---*/
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) &&                           \
+    defined(WOLFSSL_DTLS13) && defined(WOLFSSL_DTLS)
+static int test_dtls13_get_message_seq(const char* msg, int msgSz,
+    word16* msgSeq)
+{
+    int hsOff = DTLS_RECORD_HEADER_SZ;
+
+    if (msg == NULL || msgSeq == NULL ||
+            msgSz < DTLS_RECORD_HEADER_SZ + DTLS_HANDSHAKE_HEADER_SZ) {
+        return BAD_FUNC_ARG;
+    }
+
+    *msgSeq = ((word16)(byte)msg[hsOff + 4] << 8) |
+              (word16)(byte)msg[hsOff + 5];
+
+    return WOLFSSL_SUCCESS;
+}
+#endif
+
+/*-- ch2_rtx_no_ch1 (test_dtls.c lines 1869,1940) ---*/
+int test_dtls13_ch2_rtx_no_ch1(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) &&                           \
+    defined(WOLFSSL_DTLS13) && defined(WOLFSSL_DTLS)
+    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+    const char* msg = NULL;
+    int msgSz = 0;
+    word16 ch1Seq = 0;
+    int i;
+    int foundCh1Seq = 0;
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+                    wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method),
+        0);
+
+    /* To force HRR */
+    ExpectIntEQ(wolfSSL_NoKeyShares(ssl_c), WOLFSSL_SUCCESS);
+
+    /* CH1 */
+    ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    ExpectIntEQ(test_memio_get_message(&test_ctx, 0, &msg, &msgSz, 0), 0);
+    ExpectIntGE(msgSz, DTLS_RECORD_HEADER_SZ + DTLS_HANDSHAKE_HEADER_SZ);
+    ExpectIntEQ(test_dtls13_get_message_seq(msg, msgSz, &ch1Seq),
+        WOLFSSL_SUCCESS);
+
+    /* HRR */
+    ExpectIntEQ(wolfSSL_accept(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    ExpectIntGT(test_ctx.c_msg_count, 0);
+
+    /* CH2 */
+    ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+    ExpectIntGT(test_ctx.s_msg_count, 0);
+
+    /* Drop CH2 and trigger the client retransmission timeout. */
+    test_memio_clear_buffer(&test_ctx, 0);
+    if (wolfSSL_dtls13_use_quick_timeout(ssl_c))
+        ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
+    ExpectIntGT(test_ctx.s_msg_count, 0);
+
+    for (i = 0; i < test_ctx.s_msg_count && EXPECT_SUCCESS(); i++) {
+        int hsOff = DTLS_RECORD_HEADER_SZ;
+        word16 msgSeq = 0;
+
+        ExpectIntEQ(test_memio_get_message(&test_ctx, 0, &msg, &msgSz, i), 0);
+        /* memio stores one DTLS record per message in this handshake path. */
+        if (msgSz >= DTLS_RECORD_HEADER_SZ + DTLS_HANDSHAKE_HEADER_SZ &&
+                (byte)msg[0] == handshake && msg[hsOff] == client_hello) {
+            ExpectIntEQ(test_dtls13_get_message_seq(msg, msgSz, &msgSeq),
+                WOLFSSL_SUCCESS);
+            if (msgSeq == ch1Seq)
+                foundCh1Seq = 1;
+        }
+    }
+
+    ExpectIntEQ(foundCh1Seq, 0);
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
+
+/*-- frag_ch2_with_ch1_rtx (test_dtls.c lines 1942,2068) ---*/
+int test_dtls13_frag_ch2_with_ch1_rtx(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) &&                           \
+    defined(WOLFSSL_DTLS13) && defined(WOLFSSL_DTLS) &&                        \
+    defined(WOLFSSL_DTLS_MTU) && defined(WOLFSSL_DTLS_CH_FRAG) &&              \
+    (!defined(HAVE_FIPS) || FIPS_VERSION3_GE(7,0,0))
+    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+    char hrr[TEST_MEMIO_BUF_SZ];
+    int hrrSz = (int)sizeof(hrr);
+    char ch1Rtx[TEST_MEMIO_BUF_SZ];
+    int ch1RtxSz = (int)sizeof(ch1Rtx);
+    char ch2[TEST_MEMIO_BUF_SZ];
+    int ch2Sz = 0;
+    int ch2MsgCount = 0;
+    int ch2MsgSizes[TEST_MEMIO_MAX_MSGS] = {0};
+    /* The DTLS record sequence number occupies the last 8 bytes of the
+     * record header. */
+    int recordSeqOff = DTLS_RECORD_HEADER_SZ - 8;
+    int ch2Seq = 0;
+    int ch1RtxSeq = 0;
+    int off;
+    int i;
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+                    wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method),
+        0);
+
+    /* To force HRR */
+    ExpectIntEQ(wolfSSL_NoKeyShares(ssl_c), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_dtls13_allow_ch_frag(ssl_s, 1), WOLFSSL_SUCCESS);
+
+    /* CH1 */
+    ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+
+    /* HRR */
+    ExpectIntEQ(wolfSSL_accept(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    ExpectIntEQ(test_memio_copy_message(&test_ctx, 1, hrr, &hrrSz, 0), 0);
+
+    /* Drop HRR, trigger CH1 retransmission, copy and drop it */
+    test_memio_clear_buffer(&test_ctx, 1);
+    if (wolfSSL_dtls13_use_quick_timeout(ssl_c))
+        ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_c), WOLFSSL_SUCCESS);
+    ExpectIntEQ(test_memio_copy_message(&test_ctx, 0, ch1Rtx, &ch1RtxSz, 0), 0);
+    test_memio_clear_buffer(&test_ctx, 0);
+
+    /* Force CH2 fragmentation. MTU must be small enough to fragment but large
+     * enough that the cookie extension lands in the first fragment, otherwise
+     * the server can't validate it statelessly and the test scenario (server
+     * stateful after frag 1) does not hold. With --enable-all (PQ groups in
+     * supported_groups), the cookie extension can sit ~400 bytes into CH2; 600
+     * gives margin while still producing multiple fragments (CH2 is ~2KB). */
+    ExpectIntEQ(wolfSSL_dtls_set_mtu(ssl_c, 600), WOLFSSL_SUCCESS);
+
+    /* Forward HRR and let the client create fragmented CH2 */
+    ExpectIntEQ(test_memio_inject_message(&test_ctx, 1, hrr, hrrSz), 0);
+    ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+
+    ExpectIntGT(test_ctx.s_msg_count, 1);
+    ExpectIntLE(test_ctx.s_msg_count, TEST_MEMIO_MAX_MSGS);
+    ExpectIntLE(test_ctx.s_len, (int)sizeof(ch2));
+    if (EXPECT_SUCCESS()) {
+        ch2Sz = test_ctx.s_len;
+        ch2MsgCount = test_ctx.s_msg_count;
+        XMEMCPY(ch2, test_ctx.s_buff, ch2Sz);
+        XMEMCPY(ch2MsgSizes, test_ctx.s_msg_sizes,
+            sizeof(ch2MsgSizes[0]) * (size_t)ch2MsgCount);
+
+        ch2Seq = ((byte)ch2[recordSeqOff + 4] << 8) |
+                 (byte)ch2[recordSeqOff + 5];
+        ch1RtxSeq = ch2Seq + ch2MsgCount;
+
+        /* Synthesize a CH1 retransmission that can pass the replay window after
+         * the first CH2 fragment makes the server stateful. The handshake
+         * message_seq remains the original CH1 value; only the DTLS record
+         * sequence is moved past the fragmented CH2 flight */
+        ch1Rtx[recordSeqOff + 0] = 0;
+        ch1Rtx[recordSeqOff + 1] = 0;
+        ch1Rtx[recordSeqOff + 2] = 0;
+        ch1Rtx[recordSeqOff + 3] = 0;
+        ch1Rtx[recordSeqOff + 4] = (byte)(ch1RtxSeq >> 8);
+        ch1Rtx[recordSeqOff + 5] = (byte)ch1RtxSeq;
+    }
+
+    test_memio_clear_buffer(&test_ctx, 0);
+
+    /* Deliver CH2 first fragment only. Now server is stateful */
+    ExpectIntEQ(test_memio_inject_message(&test_ctx, 0, ch2, ch2MsgSizes[0]), 0);
+    ExpectIntEQ(wolfSSL_accept(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+
+    /* Deliver the retransmitted CH1 between CH2 fragments, it should be
+     * discarded as rtx */
+    ExpectIntEQ(test_memio_inject_message(&test_ctx, 0, ch1Rtx, ch1RtxSz), 0);
+    ExpectIntEQ(wolfSSL_accept(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+    test_memio_clear_buffer(&test_ctx, 1);
+
+    /* Deliver the rest of CH2 */
+    off = ch2MsgSizes[0];
+    for (i = 1; i < ch2MsgCount && EXPECT_SUCCESS(); i++) {
+        ExpectIntEQ(test_memio_inject_message(&test_ctx, 0, ch2 + off,
+            ch2MsgSizes[i]), 0);
+        off += ch2MsgSizes[i];
+    }
+
+    /* Restore MTU so the client's input buffer can hold the full server
+     * flight (e.g. an SH carrying a hybrid PQC key share). */
+    ExpectIntEQ(wolfSSL_dtls_set_mtu(ssl_c, 1500), WOLFSSL_SUCCESS);
+
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
+
+/*-- srtp_with_helper_and_stub (test_dtls.c lines 2276,2312) ---*/
+#if defined(WOLFSSL_DTLS13) && defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES) && \
+    defined(WOLFSSL_SRTP)
+static int test_dtls_srtp_ctx_ready(WOLFSSL_CTX* ctx)
+{
+    EXPECT_DECLS;
+    ExpectIntEQ(wolfSSL_CTX_set_tlsext_use_srtp(ctx, "SRTP_AEAD_AES_256_GCM:"
+         "SRTP_AEAD_AES_128_GCM:SRTP_AES128_CM_SHA1_80:SRTP_AES128_CM_SHA1_32"),
+          0);
+    return EXPECT_RESULT();
+}
+
+int test_dtls_srtp(void)
+{
+    EXPECT_DECLS;
+    test_ssl_cbf client_cbf;
+    test_ssl_cbf server_cbf;
+
+    XMEMSET(&client_cbf, 0, sizeof(client_cbf));
+    XMEMSET(&server_cbf, 0, sizeof(server_cbf));
+
+    client_cbf.method = wolfDTLSv1_3_client_method;
+    client_cbf.ctx_ready = test_dtls_srtp_ctx_ready;
+    server_cbf.method = wolfDTLSv1_3_server_method;
+    server_cbf.ctx_ready = test_dtls_srtp_ctx_ready;
+
+    ExpectIntEQ(test_wolfSSL_client_server_nofail_memio(&client_cbf,
+        &server_cbf, NULL), TEST_SUCCESS);
+
+    return EXPECT_RESULT();
+}
+#else
+int test_dtls_srtp(void)
+{
+    EXPECT_DECLS;
+    return EXPECT_RESULT();
+}
+#endif
+
+/*-- min_rtx_interval (test_dtls.c lines 2890,2960) ---*/
+int test_dtls13_min_rtx_interval(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && \
+    defined(WOLFSSL_DTLS13) && !defined(DTLS13_MIN_RTX_INTERVAL) && \
+    !defined(NO_ASN_TIME)
+    /* We don't want to test when DTLS13_MIN_RTX_INTERVAL is defined because
+     * it may be too low to trigger reliably in a test. The default value is
+     * 1 second which is sufficient for testing here. */
+    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
+    struct test_memio_ctx test_ctx;
+    int c_msg_count = 0;
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+
+    /* Setup DTLS 1.3 contexts */
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+                    wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
+
+    /* CH0 */
+    ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), SSL_ERROR_WANT_READ);
+
+    /* HRR */
+    ExpectIntEQ(wolfSSL_accept(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), SSL_ERROR_WANT_READ);
+
+    /* CH1 */
+    ExpectIntEQ(wolfSSL_connect(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), SSL_ERROR_WANT_READ);
+
+    /* SH ... FINISHED */
+    ExpectIntEQ(wolfSSL_accept(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), SSL_ERROR_WANT_READ);
+
+    /* We should have SH ... FINISHED messages in the buffer */
+    ExpectIntGE(test_ctx.c_msg_count, 2);
+
+    /* Drop everything */
+    test_memio_clear_buffer(&test_ctx, 1);
+
+    /* First timeout. This one should trigger a retransmission */
+    if (wolfSSL_dtls13_use_quick_timeout(ssl_s))
+        ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_s), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_s), WOLFSSL_SUCCESS);
+    /* Save the message count to make sure no new messages are sent */
+    ExpectIntGE(test_ctx.c_msg_count, 2);
+    c_msg_count = test_ctx.c_msg_count;
+
+    /* Second timeout. This one should not trigger a retransmission */
+    if (wolfSSL_dtls13_use_quick_timeout(ssl_s))
+        ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_s), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_dtls_got_timeout(ssl_s), WOLFSSL_SUCCESS);
+    /* This is the critical check. The message count should not increase
+     * after the second timeout. DTLS13_MIN_RTX_INTERVAL should have blocked
+     * retransmission here. */
+    ExpectIntEQ(c_msg_count, test_ctx.c_msg_count);
+
+    /* Now complete the handshake. We didn't clear the first retransmission
+     * so the handshake should proceed without issues. */
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+
+    /* Cleanup */
+    wolfSSL_free(ssl_c);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
+
+/*-- no_session_id_echo (test_dtls.c lines 2965,3044) ---*/
+int test_dtls13_no_session_id_echo(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13) && \
+    defined(HAVE_SESSION_TICKET) && defined(HAVE_ECC) && \
+    !defined(WOLFSSL_DTLS13_ECHO_LEGACY_SESSION_ID)
+    struct test_memio_ctx test_ctx;
+    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
+    WOLFSSL_SESSION *sess = NULL;
+    char readBuf[1];
+    /* Pin to SECP256R1 to avoid a PQ-induced key-share HRR */
+    int groups[] = { WOLFSSL_ECC_SECP256R1 };
+
+    /* First connection: complete a DTLS 1.3 handshake to get a session */
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
+    ExpectIntEQ(wolfSSL_set_groups(ssl_c, groups, 1), WOLFSSL_SUCCESS);
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+
+    /* Read to process any NewSessionTicket */
+    ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+
+    ExpectNotNull(sess = wolfSSL_get1_session(ssl_c));
+
+    /* Ensure the session has a non-empty session ID so the ClientHello
+     * will have a populated legacy_session_id field (which is legal per
+     * RFC 9147). */
+    if (sess != NULL && sess->sessionIDSz == 0) {
+        sess->sessionIDSz = ID_LEN;
+        XMEMSET(sess->sessionID, 0x42, ID_LEN);
+    }
+
+    wolfSSL_free(ssl_c); ssl_c = NULL;
+    wolfSSL_free(ssl_s); ssl_s = NULL;
+    wolfSSL_CTX_free(ctx_c); ctx_c = NULL;
+    wolfSSL_CTX_free(ctx_s); ctx_s = NULL;
+
+    /* Second connection: set the session on the client so the ClientHello
+     * contains a non-empty legacy_session_id. Verify the server does NOT
+     * echo it in the ServerHello. */
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
+    ExpectIntEQ(wolfSSL_set_session(ssl_c, sess), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_set_groups(ssl_c, groups, 1), WOLFSSL_SUCCESS);
+    /* Disable HRR cookie so the server directly sends a ServerHello */
+    ExpectIntEQ(wolfSSL_disable_hrr_cookie(ssl_s), WOLFSSL_SUCCESS);
+
+    /* Client sends ClientHello (with non-empty legacy_session_id) */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+
+    /* Server processes ClientHello and sends ServerHello + flight */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+
+    /* Verify the ServerHello on the wire.
+     * Layout: DTLS Record Header (13) + DTLS Handshake Header (12) +
+     *         ProtocolVersion (2) + Random (32) = offset 59 for
+     *         legacy_session_id_echo length byte. */
+    ExpectIntGE(test_ctx.c_len, 60);
+    ExpectIntEQ(test_ctx.c_buff[0], handshake);
+    ExpectIntEQ(test_ctx.c_buff[DTLS_RECORD_HEADER_SZ], server_hello);
+    ExpectIntEQ(test_ctx.c_buff[DTLS_RECORD_HEADER_SZ +
+        DTLS_HANDSHAKE_HEADER_SZ + OPAQUE16_LEN + RAN_LEN], 0);
+
+    /* Complete the handshake */
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+
+    wolfSSL_SESSION_free(sess);
+    wolfSSL_free(ssl_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
+
+/*-- 5_9_0_compat (test_dtls.c lines 3049,3170) ---*/
+int test_dtls13_5_9_0_compat(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && defined(WOLFSSL_DTLS13) && \
+    defined(HAVE_SESSION_TICKET) && defined(WOLFSSL_DTLS13_ECHO_LEGACY_SESSION_ID) && \
+    defined(HAVE_ECC)
+    struct test_memio_ctx test_ctx;
+    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
+    WOLFSSL_SESSION *sess = NULL;
+    char readBuf[1];
+    /* Pin to SECP256R1 to avoid a PQ-induced key-share HRR */
+    int groups[] = { WOLFSSL_ECC_SECP256R1 };
+    /* RFC 8446 Section 4.1.3: an HRR is a ServerHello carrying this magic
+     * random. Used to assert sub-test 1 is a real ServerHello, not an HRR. */
+    static const byte hrrRandom[RAN_LEN] = {
+        0xCF, 0x21, 0xAD, 0x74, 0xE5, 0x9A, 0x61, 0x11,
+        0xBE, 0x1D, 0x8C, 0x02, 0x1E, 0x65, 0xB8, 0x91,
+        0xC2, 0xA2, 0x11, 0x16, 0x7A, 0xBB, 0x8C, 0x5E,
+        0x07, 0x9E, 0x09, 0xE2, 0xC8, 0xA8, 0x33, 0x9C
+    };
+
+    /* --- initial connection: get a real session to carry the session ID --- */
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
+    ExpectIntEQ(wolfSSL_set_groups(ssl_c, groups, 1), WOLFSSL_SUCCESS);
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+
+    /* drain any NewSessionTicket before calling get1_session */
+    ExpectIntEQ(wolfSSL_read(ssl_c, readBuf, sizeof(readBuf)), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+
+    ExpectNotNull(sess = wolfSSL_get1_session(ssl_c));
+
+    /* Force a non-zero session ID - simulates a wolfSSL <=v5.9.0 client that
+     * mistakenly sends 32 bytes as legacy_session_id in DTLS 1.3. */
+    if (sess != NULL && sess->sessionIDSz == 0) {
+        sess->sessionIDSz = ID_LEN;
+        XMEMSET(sess->sessionID, 0x42, ID_LEN);
+    }
+
+    wolfSSL_free(ssl_c); ssl_c = NULL;
+    wolfSSL_free(ssl_s); ssl_s = NULL;
+    wolfSSL_CTX_free(ctx_c); ctx_c = NULL;
+    wolfSSL_CTX_free(ctx_s); ctx_s = NULL;
+
+    /* --- sub-test 1: direct ServerHello (HRR cookie disabled) ---
+     * Exercises DoTls13ClientHello (change 1) and
+     * SendTls13ServerHello (change 2). */
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
+    ExpectIntEQ(wolfSSL_set_session(ssl_c, sess), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_set_groups(ssl_c, groups, 1), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_disable_hrr_cookie(ssl_s), WOLFSSL_SUCCESS);
+
+    /* Client sends CH1 with non-empty legacy_session_id */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+
+    /* Server processes CH1 and sends ServerHello */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+
+    /* Verify that the ServerHello on the wire echoes the session ID.
+     * Layout: DTLS Record Header (13) + DTLS Handshake Header (12) +
+     *         ProtocolVersion (2) + Random (32) = byte 59 for
+     *         legacy_session_id_echo length. */
+    ExpectIntGE(test_ctx.c_len, 60);
+    ExpectIntEQ(test_ctx.c_buff[0], handshake);
+    ExpectIntEQ(test_ctx.c_buff[DTLS_RECORD_HEADER_SZ], server_hello);
+    /* Confirm it is a real ServerHello, not an HRR (also encoded as a
+     * ServerHello but bearing the HelloRetryRequest magic random). */
+    ExpectIntNE(XMEMCMP(&test_ctx.c_buff[DTLS_RECORD_HEADER_SZ +
+        DTLS_HANDSHAKE_HEADER_SZ + OPAQUE16_LEN], hrrRandom, RAN_LEN), 0);
+    ExpectIntEQ(test_ctx.c_buff[DTLS_RECORD_HEADER_SZ +
+        DTLS_HANDSHAKE_HEADER_SZ + OPAQUE16_LEN + RAN_LEN], ID_LEN);
+
+    /* Complete the handshake - Finished MAC validates the transcript */
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+
+    wolfSSL_free(ssl_c); ssl_c = NULL;
+    wolfSSL_free(ssl_s); ssl_s = NULL;
+    wolfSSL_CTX_free(ctx_c); ctx_c = NULL;
+    wolfSSL_CTX_free(ctx_s); ctx_s = NULL;
+
+    /* --- sub-test 2: stateless HRR (HRR cookie enabled by default) ---
+     * Exercises SendStatelessReplyDtls13 (change 4) and
+     * RestartHandshakeHashWithCookie (change 3). */
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+        wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method), 0);
+    ExpectIntEQ(wolfSSL_set_session(ssl_c, sess), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_set_groups(ssl_c, groups, 1), WOLFSSL_SUCCESS);
+
+    /* Client sends CH1 */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_c), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_c, -1), WOLFSSL_ERROR_WANT_READ);
+
+    /* Server sends stateless HRR (SendStatelessReplyDtls13) */
+    ExpectIntEQ(wolfSSL_negotiate(ssl_s), -1);
+    ExpectIntEQ(wolfSSL_get_error(ssl_s, -1), WOLFSSL_ERROR_WANT_READ);
+
+    /* Verify the HRR echoes the session ID at the same wire offset */
+    ExpectIntGE(test_ctx.c_len, 60);
+    ExpectIntEQ(test_ctx.c_buff[0], handshake);
+    ExpectIntEQ(test_ctx.c_buff[DTLS_RECORD_HEADER_SZ], server_hello);
+    ExpectIntEQ(test_ctx.c_buff[DTLS_RECORD_HEADER_SZ +
+        DTLS_HANDSHAKE_HEADER_SZ + OPAQUE16_LEN + RAN_LEN], ID_LEN);
+
+    /* Complete the handshake - Finished MAC validates RestartHandshakeHashWithCookie */
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+
+    wolfSSL_SESSION_free(sess);
+    wolfSSL_free(ssl_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_CTX_free(ctx_s);
+#endif
+    return EXPECT_RESULT();
+}
diff --git a/tests/api/test_dtls13.h b/tests/api/test_dtls13.h
new file mode 100644
index 0000000000..6b82a50483
--- /dev/null
+++ b/tests/api/test_dtls13.h
@@ -0,0 +1,79 @@
+/* test_dtls13.h
+ *
+ * Copyright (C) 2006-2026 wolfSSL Inc.
+ *
+ * This file is part of wolfSSL.
+ *
+ * wolfSSL is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * wolfSSL is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1335, USA
+ */
+
+#ifndef TESTS_API_DTLS13_H
+#define TESTS_API_DTLS13_H
+
+#include <tests/api/api_decl.h>
+
+/* DTLSv1.3-only tests. None share helpers with DTLS<=1.2 tests, so they
+ * live in their own translation unit and register under the "dtls13"
+ * group. Each function is defined unconditionally with the body (or the
+ * whole function via #else stub) guarded by WOLFSSL_DTLS13. */
+int test_dtls13_bad_epoch_ch(void);
+int test_wolfSSL_dtls13_null_cipher(void);
+int test_dtls13_frag_ch_pq(void);
+int test_dtls_frag_ch(void);
+int test_dtls_empty_keyshare_with_cookie(void);
+int test_dtls13_missing_finished_client(void);
+int test_dtls13_missing_finished_server(void);
+int test_dtls13_finished_send_error_propagation(void);
+
+/* DTLSv1.3-only tests moved from test_dtls.c (isolated from DTLS<=1.2 code,
+ * none share helpers with non-DTLS13 tests). */
+int test_dtls13_basic_connection_id(void);
+int test_dtls13_hrr_want_write(void);
+int test_dtls13_every_write_want_write(void);
+int test_dtls13_epochs(void);
+int test_dtls13_ack_order(void);
+int test_dtls13_ack_overflow(void);
+int test_dtls13_ack_dup_write_counter(void);
+int test_dtls13_ch2_rtx_no_ch1(void);
+int test_dtls13_frag_ch2_with_ch1_rtx(void);
+int test_dtls_srtp(void);
+int test_dtls13_min_rtx_interval(void);
+int test_dtls13_no_session_id_echo(void);
+int test_dtls13_5_9_0_compat(void);
+
+#define TEST_DTLS13_DECLS                                                      \
+    TEST_DECL_GROUP("dtls13", test_dtls13_bad_epoch_ch),                       \
+    TEST_DECL_GROUP("dtls13", test_wolfSSL_dtls13_null_cipher),                \
+    TEST_DECL_GROUP("dtls13", test_dtls13_frag_ch_pq),                         \
+    TEST_DECL_GROUP("dtls13", test_dtls_frag_ch),                              \
+    TEST_DECL_GROUP("dtls13", test_dtls_empty_keyshare_with_cookie),           \
+    TEST_DECL_GROUP("dtls13", test_dtls13_missing_finished_client),            \
+    TEST_DECL_GROUP("dtls13", test_dtls13_missing_finished_server),            \
+    TEST_DECL_GROUP("dtls13", test_dtls13_finished_send_error_propagation),    \
+    TEST_DECL_GROUP("dtls13", test_dtls13_basic_connection_id),                \
+    TEST_DECL_GROUP("dtls13", test_dtls13_hrr_want_write),                     \
+    TEST_DECL_GROUP("dtls13", test_dtls13_every_write_want_write),             \
+    TEST_DECL_GROUP("dtls13", test_dtls13_epochs),                             \
+    TEST_DECL_GROUP("dtls13", test_dtls13_ack_order),                          \
+    TEST_DECL_GROUP("dtls13", test_dtls13_ack_overflow),                       \
+    TEST_DECL_GROUP("dtls13", test_dtls13_ack_dup_write_counter),              \
+    TEST_DECL_GROUP("dtls13", test_dtls13_ch2_rtx_no_ch1),                     \
+    TEST_DECL_GROUP("dtls13", test_dtls13_frag_ch2_with_ch1_rtx),              \
+    TEST_DECL_GROUP("dtls13", test_dtls_srtp),                                 \
+    TEST_DECL_GROUP("dtls13", test_dtls13_min_rtx_interval),                   \
+    TEST_DECL_GROUP("dtls13", test_dtls13_no_session_id_echo),                 \
+    TEST_DECL_GROUP("dtls13", test_dtls13_5_9_0_compat)
+
+#endif /* TESTS_API_DTLS13_H */
diff --git a/tests/api/test_ed25519.c b/tests/api/test_ed25519.c
index 2df8d41270..e68505b862 100644
--- a/tests/api/test_ed25519.c
+++ b/tests/api/test_ed25519.c
@@ -725,3 +725,184 @@ int test_wc_Ed25519KeyToDer_oneasymkey_version(void)
     return EXPECT_RESULT();
 }
 
+/* Ed25519 identity and small-order public keys must be rejected. When
+ * the public key is the identity point (or any small-order point), any
+ * signature of the form (R = [S]B, S) verifies for arbitrary messages
+ * because h*A is the neutral element. Gated on FIPS_VERSION3_GE(7,0,0)
+ * because older FIPS-certified modules do not have this check in their
+ * frozen copy of ed25519.c and would fail this test. */
+int test_wc_ed25519_reject_small_order_keys(void)
+{
+    EXPECT_DECLS;
+#if (!defined(HAVE_FIPS) || FIPS_VERSION3_GE(7,0,0)) && \
+    defined(HAVE_ED25519) && defined(HAVE_ED25519_KEY_IMPORT)
+    /* Each entry holds an encoded small-order Ed25519 public key. The
+     * sign-bit variants of each y-coordinate are listed explicitly so
+     * the test catches both possible encodings of each y. */
+    static const byte small_order_keys[][ED25519_PUB_KEY_SIZE] = {
+        /* identity (y = 1) */
+        {0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00},
+        /* identity with x-sign bit set */
+        {0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x80},
+        /* order 2: y = p - 1 */
+        {0xec,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f},
+        /* order 2: y = p - 1 with x-sign bit set */
+        {0xec,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff},
+        /* non-canonical y = p (decodes to y = 0) */
+        {0xed,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f},
+        /* non-canonical y = p with x-sign bit set */
+        {0xed,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff},
+        /* non-canonical y = p + 1 (decodes to y = 1) */
+        {0xee,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f},
+        /* non-canonical y = p + 1 with x-sign bit set */
+        {0xee,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff},
+        /* order 4: y = 0 */
+        {0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00},
+        /* order 4 with x-sign bit set */
+        {0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x80},
+        /* order 8 */
+        {0x26,0xe8,0x95,0x8f,0xc2,0xb2,0x27,0xb0,
+         0x45,0xc3,0xf4,0x89,0xf2,0xef,0x98,0xf0,
+         0xd5,0xdf,0xac,0x05,0xd3,0xc6,0x33,0x39,
+         0xb1,0x38,0x02,0x88,0x6d,0x53,0xfc,0x05},
+        /* order 8 with x-sign bit set */
+        {0x26,0xe8,0x95,0x8f,0xc2,0xb2,0x27,0xb0,
+         0x45,0xc3,0xf4,0x89,0xf2,0xef,0x98,0xf0,
+         0xd5,0xdf,0xac,0x05,0xd3,0xc6,0x33,0x39,
+         0xb1,0x38,0x02,0x88,0x6d,0x53,0xfc,0x85},
+        /* order 8 (other y) */
+        {0xc7,0x17,0x6a,0x70,0x3d,0x4d,0xd8,0x4f,
+         0xba,0x3c,0x0b,0x76,0x0d,0x10,0x67,0x0f,
+         0x2a,0x20,0x53,0xfa,0x2c,0x39,0xcc,0xc6,
+         0x4e,0xc7,0xfd,0x77,0x92,0xac,0x03,0x7a},
+        /* order 8 (other y) with x-sign bit set */
+        {0xc7,0x17,0x6a,0x70,0x3d,0x4d,0xd8,0x4f,
+         0xba,0x3c,0x0b,0x76,0x0d,0x10,0x67,0x0f,
+         0x2a,0x20,0x53,0xfa,0x2c,0x39,0xcc,0xc6,
+         0x4e,0xc7,0xfd,0x77,0x92,0xac,0x03,0xfa},
+    };
+    /* Forged signature: R = B (base point), S = 1.
+     * With public key A = identity, S*B - h*A = B = R for any message. */
+    static const byte forged_sig[ED25519_SIG_SIZE] = {
+        0x58,0x66,0x66,0x66,0x66,0x66,0x66,0x66,
+        0x66,0x66,0x66,0x66,0x66,0x66,0x66,0x66,
+        0x66,0x66,0x66,0x66,0x66,0x66,0x66,0x66,
+        0x66,0x66,0x66,0x66,0x66,0x66,0x66,0x66,
+        0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+        0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+        0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+        0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00
+    };
+    ed25519_key key;
+    word32 i;
+    word32 num_keys = (word32)(sizeof(small_order_keys) / ED25519_PUB_KEY_SIZE);
+
+    /* (1) Untrusted wc_ed25519_import_public must reject every small-order
+     * encoding (it runs wc_ed25519_check_key as part of the import). */
+    for (i = 0; i < num_keys; i++) {
+        int rc;
+        XMEMSET(&key, 0, sizeof(key));
+        ExpectIntEQ(wc_ed25519_init(&key), 0);
+        rc = wc_ed25519_import_public(small_order_keys[i],
+            ED25519_PUB_KEY_SIZE, &key);
+        if (rc != WC_NO_ERR_TRACE(PUBLIC_KEY_E)) {
+            fprintf(stderr, "small_order_keys[%u]: import_public returned %d, "
+                "expected PUBLIC_KEY_E\n", (unsigned)i, rc);
+        }
+        ExpectIntEQ(rc, WC_NO_ERR_TRACE(PUBLIC_KEY_E));
+        wc_ed25519_free(&key);
+    }
+
+    /* (2) wc_ed25519_check_key called directly must also reject. Guards
+     * against a refactor that moves the small-order check out of
+     * check_key and into the import path: (1) would still pass, but the
+     * documented check_key contract would silently regress. */
+    for (i = 0; i < num_keys; i++) {
+        int rc;
+        XMEMSET(&key, 0, sizeof(key));
+        ExpectIntEQ(wc_ed25519_init(&key), 0);
+        /* trusted = 1 bypasses the import-time check_key call so the
+         * direct check_key below is what's under test. */
+        ExpectIntEQ(wc_ed25519_import_public_ex(small_order_keys[i],
+            ED25519_PUB_KEY_SIZE, &key, 1), 0);
+        rc = wc_ed25519_check_key(&key);
+        if (rc != WC_NO_ERR_TRACE(PUBLIC_KEY_E)) {
+            fprintf(stderr, "small_order_keys[%u]: check_key returned %d, "
+                "expected PUBLIC_KEY_E\n", (unsigned)i, rc);
+        }
+        ExpectIntEQ(rc, WC_NO_ERR_TRACE(PUBLIC_KEY_E));
+        wc_ed25519_free(&key);
+    }
+
+    /* (3) Even a "trusted" import (which bypasses wc_ed25519_check_key)
+     * must not let wc_ed25519_verify_msg accept a forged signature against
+     * an identity public key. Test both the canonical encoding (y = 1,
+     * small_order_keys[0]) and the non-canonical encoding (y = p + 1,
+     * small_order_keys[6]) so the verify-side check is exercised against
+     * the canonical-form bypass route, not just the byte-for-byte
+     * identity. The forged sig (R = B, S = 1) verifies for an identity
+     * public key only - other small-order points would reject it on the
+     * math alone, so they aren't useful here. */
+    {
+        static const word32 identity_indices[] = { 0, 6 };
+        const char* msg = "forged message";
+        word32 j;
+
+        for (j = 0;
+             j < sizeof(identity_indices)/sizeof(identity_indices[0]);
+             j++) {
+            word32 idx = identity_indices[j];
+            int verify_result = 1;
+            int rc;
+
+            XMEMSET(&key, 0, sizeof(key));
+            ExpectIntEQ(wc_ed25519_init(&key), 0);
+            ExpectIntEQ(wc_ed25519_import_public_ex(small_order_keys[idx],
+                ED25519_PUB_KEY_SIZE, &key, 1), 0);
+            rc = wc_ed25519_verify_msg(forged_sig, sizeof(forged_sig),
+                (const byte*)msg, (word32)XSTRLEN(msg), &verify_result, &key);
+            if (rc != WC_NO_ERR_TRACE(BAD_FUNC_ARG) || verify_result != 0) {
+                fprintf(stderr, "verify_msg with identity-equiv "
+                    "small_order_keys[%u]: rc=%d verify_result=%d "
+                    "(expected BAD_FUNC_ARG and 0)\n",
+                    (unsigned)idx, rc, verify_result);
+            }
+            ExpectIntEQ(rc, WC_NO_ERR_TRACE(BAD_FUNC_ARG));
+            ExpectIntEQ(verify_result, 0);
+            wc_ed25519_free(&key);
+        }
+    }
+#endif
+    return EXPECT_RESULT();
+}
+
diff --git a/tests/api/test_ed25519.h b/tests/api/test_ed25519.h
index 4ac3693f66..a14b2a4496 100644
--- a/tests/api/test_ed25519.h
+++ b/tests/api/test_ed25519.h
@@ -37,6 +37,7 @@ int test_wc_Ed25519PublicKeyToDer(void);
 int test_wc_Ed25519KeyToDer(void);
 int test_wc_Ed25519PrivateKeyToDer(void);
 int test_wc_Ed25519KeyToDer_oneasymkey_version(void);
+int test_wc_ed25519_reject_small_order_keys(void);
 
 #define TEST_ED25519_DECLS                                          \
     TEST_DECL_GROUP("ed25519", test_wc_ed25519_make_key),           \
@@ -51,6 +52,7 @@ int test_wc_Ed25519KeyToDer_oneasymkey_version(void);
     TEST_DECL_GROUP("ed25519", test_wc_Ed25519PublicKeyToDer),      \
     TEST_DECL_GROUP("ed25519", test_wc_Ed25519KeyToDer),            \
     TEST_DECL_GROUP("ed25519", test_wc_Ed25519PrivateKeyToDer),     \
-    TEST_DECL_GROUP("ed25519", test_wc_Ed25519KeyToDer_oneasymkey_version)
+    TEST_DECL_GROUP("ed25519", test_wc_Ed25519KeyToDer_oneasymkey_version), \
+    TEST_DECL_GROUP("ed25519", test_wc_ed25519_reject_small_order_keys)
 
 #endif /* WOLFCRYPT_TEST_ED25519_H */
diff --git a/tests/api/test_ed448.c b/tests/api/test_ed448.c
index f6b7552302..6bb7934615 100644
--- a/tests/api/test_ed448.c
+++ b/tests/api/test_ed448.c
@@ -649,3 +649,224 @@ int test_wc_Ed448KeyToDer_oneasymkey_version(void)
     return EXPECT_RESULT();
 }
 
+/* Ed448 identity and small-order public keys must be rejected.
+ * Edwards448 has cofactor 4, so the small-order subgroup contains the
+ * identity, an order-2 point, and two order-4 points. With any of these
+ * as the public key, h*A is the neutral element and forged signatures
+ * verify for arbitrary messages. Gated on FIPS_VERSION3_GE(7,0,0)
+ * because older FIPS-certified modules do not have this check in their
+ * frozen copy of ed448.c. */
+int test_wc_ed448_reject_small_order_keys(void)
+{
+    EXPECT_DECLS;
+#if (!defined(HAVE_FIPS) || FIPS_VERSION3_GE(7,0,0)) && \
+    defined(HAVE_ED448) && defined(HAVE_ED448_KEY_IMPORT)
+    /* Two regressions are guarded here. Both sign-bit variants of each
+     * y are listed so weakening the "clear all of byte 56" mask in
+     * ed448_is_small_order() would be caught. The non-canonical rows
+     * (y = p, y = p + 1) guard against dropping the canonical-form
+     * coverage: fe448_from_bytes reads bytes 0-55 modulo p with no
+     * canonical-form check, so y = p decodes to 0 and y = p + 1
+     * decodes to 1, both of which are small order. */
+    static const byte small_order_keys[][ED448_PUB_KEY_SIZE] = {
+        /* identity (y = 1), sign 0 */
+        {0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00},
+        /* identity (y = 1), sign bit set */
+        {0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x80},
+        /* order 4: y = 0, x-sign 0 */
+        {0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00},
+        /* order 4: y = 0, x-sign 1 */
+        {0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x80},
+        /* order 2: y = p - 1, x = 0, sign 0 */
+        {0xfe,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xfe,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0x00},
+        /* order 2: y = p - 1, sign bit set */
+        {0xfe,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xfe,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0x80},
+        /* non-canonical y = p (decodes to y = 0), sign 0 */
+        {0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xfe,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0x00},
+        /* non-canonical y = p, sign bit set */
+        {0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xfe,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0x80},
+        /* non-canonical y = p + 1 (decodes to y = 1), sign 0 */
+        {0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0x00},
+        /* non-canonical y = p + 1, sign bit set */
+        {0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0x80},
+    };
+    /* Arbitrary signature bytes: S = 1 (must be below the Ed448 group
+     * order or wc_ed448_verify_msg() returns BAD_FUNC_ARG before the
+     * small-order check has a chance to fire). The R bytes do not need
+     * to encode a valid curve point for this test - the small-order
+     * defence in ed448_verify_msg_final_with_sha() rejects the public
+     * key before the R/S verification math runs. */
+    static const byte forged_sig[ED448_SIG_SIZE] = {
+        /* R: 57 bytes of arbitrary data (last byte 0 to satisfy the
+         * spec-mandated zero of byte 56 bits 0-6; sign bit doesn't
+         * matter here). */
+        0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+        0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+        0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+        0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+        0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+        0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+        0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+        0x00,
+        /* S = 1 */
+        0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+        0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+        0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+        0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+        0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+        0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+        0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+        0x00
+    };
+    ed448_key key;
+    word32 i;
+    word32 num_keys = (word32)(sizeof(small_order_keys) / ED448_PUB_KEY_SIZE);
+
+    /* (1) Untrusted wc_ed448_import_public must reject every small-order
+     * encoding (it runs wc_ed448_check_key as part of the import). */
+    for (i = 0; i < num_keys; i++) {
+        int rc;
+        XMEMSET(&key, 0, sizeof(key));
+        ExpectIntEQ(wc_ed448_init(&key), 0);
+        rc = wc_ed448_import_public(small_order_keys[i],
+            ED448_PUB_KEY_SIZE, &key);
+        if (rc != WC_NO_ERR_TRACE(PUBLIC_KEY_E)) {
+            fprintf(stderr, "small_order_keys[%u]: import_public returned %d, "
+                "expected PUBLIC_KEY_E\n", (unsigned)i, rc);
+        }
+        ExpectIntEQ(rc, WC_NO_ERR_TRACE(PUBLIC_KEY_E));
+        wc_ed448_free(&key);
+    }
+
+    /* (2) wc_ed448_check_key called directly must also reject. Guards
+     * against a refactor that moves the small-order check out of
+     * check_key and into the import path: (1) would still pass, but the
+     * documented check_key contract would silently regress. */
+    for (i = 0; i < num_keys; i++) {
+        int rc;
+        XMEMSET(&key, 0, sizeof(key));
+        ExpectIntEQ(wc_ed448_init(&key), 0);
+        /* trusted = 1 bypasses the import-time check_key call so the
+         * direct check_key below is what's under test. */
+        ExpectIntEQ(wc_ed448_import_public_ex(small_order_keys[i],
+            ED448_PUB_KEY_SIZE, &key, 1), 0);
+        rc = wc_ed448_check_key(&key);
+        if (rc != WC_NO_ERR_TRACE(PUBLIC_KEY_E)) {
+            fprintf(stderr, "small_order_keys[%u]: check_key returned %d, "
+                "expected PUBLIC_KEY_E\n", (unsigned)i, rc);
+        }
+        ExpectIntEQ(rc, WC_NO_ERR_TRACE(PUBLIC_KEY_E));
+        wc_ed448_free(&key);
+    }
+
+    /* (3) Even a "trusted" import (which bypasses wc_ed448_check_key)
+     * must not let wc_ed448_verify_msg accept a forged signature against
+     * an identity public key. Test both the canonical encoding (y = 1,
+     * small_order_keys[0]) and the non-canonical encoding (y = p + 1,
+     * small_order_keys[8]) so the verify-side check is exercised against
+     * the canonical-form bypass route, not just the byte-for-byte
+     * identity. */
+    {
+        static const word32 identity_indices[] = { 0, 8 };
+        const char* msg = "forged message";
+        word32 j;
+
+        for (j = 0;
+             j < sizeof(identity_indices)/sizeof(identity_indices[0]);
+             j++) {
+            word32 idx = identity_indices[j];
+            int verify_result = 1;
+            int rc;
+
+            XMEMSET(&key, 0, sizeof(key));
+            ExpectIntEQ(wc_ed448_init(&key), 0);
+            ExpectIntEQ(wc_ed448_import_public_ex(small_order_keys[idx],
+                ED448_PUB_KEY_SIZE, &key, 1), 0);
+            rc = wc_ed448_verify_msg(forged_sig, sizeof(forged_sig),
+                (const byte*)msg, (word32)XSTRLEN(msg), &verify_result,
+                &key, NULL, 0);
+            if (rc != WC_NO_ERR_TRACE(BAD_FUNC_ARG) || verify_result != 0) {
+                fprintf(stderr, "verify_msg with identity-equiv "
+                    "small_order_keys[%u]: rc=%d verify_result=%d "
+                    "(expected BAD_FUNC_ARG and 0)\n",
+                    (unsigned)idx, rc, verify_result);
+            }
+            ExpectIntEQ(rc, WC_NO_ERR_TRACE(BAD_FUNC_ARG));
+            ExpectIntEQ(verify_result, 0);
+            wc_ed448_free(&key);
+        }
+    }
+#endif
+    return EXPECT_RESULT();
+}
+
diff --git a/tests/api/test_ed448.h b/tests/api/test_ed448.h
index 68d85a6338..a40edd17bb 100644
--- a/tests/api/test_ed448.h
+++ b/tests/api/test_ed448.h
@@ -37,6 +37,7 @@ int test_wc_Ed448PublicKeyToDer(void);
 int test_wc_Ed448KeyToDer(void);
 int test_wc_Ed448PrivateKeyToDer(void);
 int test_wc_Ed448KeyToDer_oneasymkey_version(void);
+int test_wc_ed448_reject_small_order_keys(void);
 
 #define TEST_ED448_DECLS                                          \
     TEST_DECL_GROUP("ed448", test_wc_ed448_make_key),             \
@@ -51,6 +52,7 @@ int test_wc_Ed448KeyToDer_oneasymkey_version(void);
     TEST_DECL_GROUP("ed448", test_wc_Ed448PublicKeyToDer),        \
     TEST_DECL_GROUP("ed448", test_wc_Ed448KeyToDer),              \
     TEST_DECL_GROUP("ed448", test_wc_Ed448PrivateKeyToDer),       \
-    TEST_DECL_GROUP("ed448", test_wc_Ed448KeyToDer_oneasymkey_version)
+    TEST_DECL_GROUP("ed448", test_wc_Ed448KeyToDer_oneasymkey_version), \
+    TEST_DECL_GROUP("ed448", test_wc_ed448_reject_small_order_keys)
 
 #endif /* WOLFCRYPT_TEST_ED448_H */
diff --git a/tests/api/test_evp_pkey.c b/tests/api/test_evp_pkey.c
index 13de8ebe25..6b15f38d96 100644
--- a/tests/api/test_evp_pkey.c
+++ b/tests/api/test_evp_pkey.c
@@ -2702,3 +2702,162 @@ int test_wolfSSL_EVP_PKEY_ed448(void)
     return EXPECT_RESULT();
 }
 
+int test_wolfSSL_EVP_PKEY_x25519(void)
+{
+    EXPECT_DECLS;
+#if defined(OPENSSL_EXTRA) && defined(HAVE_CURVE25519)
+    EVP_PKEY* pkey = NULL;
+    EVP_PKEY* peer = NULL;
+    EVP_PKEY_CTX* genCtx = NULL;
+    EVP_PKEY_CTX* ctx = NULL;
+    unsigned char rawPriv[32];
+    unsigned char rawPub[32];
+    unsigned char secretA[32];
+    unsigned char secretB[32];
+    size_t secretLen;
+    int i;
+
+    for (i = 0; i < 32; i++) {
+        rawPriv[i] = (unsigned char)i;
+        rawPub[i]  = (unsigned char)(0x40 + i);
+    }
+
+    /* Raw import with the correct length reports the X25519 type. */
+    ExpectNotNull(pkey = EVP_PKEY_new_raw_public_key(
+        EVP_PKEY_X25519, NULL, rawPub, sizeof(rawPub)));
+    ExpectIntEQ(EVP_PKEY_id(pkey), EVP_PKEY_X25519);
+    EVP_PKEY_free(pkey);
+    pkey = NULL;
+
+    ExpectNotNull(pkey = EVP_PKEY_new_raw_private_key(
+        EVP_PKEY_X25519, NULL, rawPriv, sizeof(rawPriv)));
+    ExpectIntEQ(EVP_PKEY_id(pkey), EVP_PKEY_X25519);
+
+    /* X25519 is key-agreement only: signing must be rejected. */
+    ExpectNotNull(ctx = EVP_PKEY_CTX_new(pkey, NULL));
+    ExpectIntNE(EVP_PKEY_sign_init(ctx), WOLFSSL_SUCCESS);
+    EVP_PKEY_CTX_free(ctx);
+    ctx = NULL;
+    EVP_PKEY_free(pkey);
+    pkey = NULL;
+
+    /* Wrong raw lengths are rejected. */
+    ExpectNull(EVP_PKEY_new_raw_public_key(
+        EVP_PKEY_X25519, NULL, rawPub, 16));
+    ExpectNull(EVP_PKEY_new_raw_private_key(
+        EVP_PKEY_X25519, NULL, rawPriv, 16));
+
+    /* Generate two key pairs and confirm ECDH agreement is symmetric. This
+     * also exercises the little-endian convention used on import/derive. */
+    ExpectNotNull(genCtx = EVP_PKEY_CTX_new_id(EVP_PKEY_X25519, NULL));
+    ExpectIntEQ(EVP_PKEY_keygen_init(genCtx), WOLFSSL_SUCCESS);
+    ExpectIntEQ(EVP_PKEY_keygen(genCtx, &pkey), WOLFSSL_SUCCESS);
+    ExpectIntEQ(EVP_PKEY_keygen(genCtx, &peer), WOLFSSL_SUCCESS);
+    EVP_PKEY_CTX_free(genCtx);
+    genCtx = NULL;
+
+    ExpectNotNull(ctx = EVP_PKEY_CTX_new(pkey, NULL));
+    ExpectIntEQ(EVP_PKEY_derive_init(ctx), WOLFSSL_SUCCESS);
+    ExpectIntEQ(EVP_PKEY_derive_set_peer(ctx, peer), WOLFSSL_SUCCESS);
+    secretLen = sizeof(secretA);
+    ExpectIntEQ(EVP_PKEY_derive(ctx, secretA, &secretLen), WOLFSSL_SUCCESS);
+    ExpectIntEQ((int)secretLen, 32);
+    EVP_PKEY_CTX_free(ctx);
+    ctx = NULL;
+
+    ExpectNotNull(ctx = EVP_PKEY_CTX_new(peer, NULL));
+    ExpectIntEQ(EVP_PKEY_derive_init(ctx), WOLFSSL_SUCCESS);
+    ExpectIntEQ(EVP_PKEY_derive_set_peer(ctx, pkey), WOLFSSL_SUCCESS);
+    secretLen = sizeof(secretB);
+    ExpectIntEQ(EVP_PKEY_derive(ctx, secretB, &secretLen), WOLFSSL_SUCCESS);
+    ExpectIntEQ((int)secretLen, 32);
+    EVP_PKEY_CTX_free(ctx);
+    ctx = NULL;
+
+    ExpectIntEQ(XMEMCMP(secretA, secretB, 32), 0);
+
+    EVP_PKEY_free(peer);
+    EVP_PKEY_free(pkey);
+#endif
+    return EXPECT_RESULT();
+}
+
+int test_wolfSSL_EVP_PKEY_x448(void)
+{
+    EXPECT_DECLS;
+#if defined(OPENSSL_EXTRA) && defined(HAVE_CURVE448)
+    EVP_PKEY* pkey = NULL;
+    EVP_PKEY* peer = NULL;
+    EVP_PKEY_CTX* genCtx = NULL;
+    EVP_PKEY_CTX* ctx = NULL;
+    unsigned char rawPriv[56];
+    unsigned char rawPub[56];
+    unsigned char secretA[56];
+    unsigned char secretB[56];
+    size_t secretLen;
+    int i;
+
+    for (i = 0; i < 56; i++) {
+        rawPriv[i] = (unsigned char)i;
+        rawPub[i]  = (unsigned char)(0x40 + i);
+    }
+
+    /* Raw import with the correct length reports the X448 type. */
+    ExpectNotNull(pkey = EVP_PKEY_new_raw_public_key(
+        EVP_PKEY_X448, NULL, rawPub, sizeof(rawPub)));
+    ExpectIntEQ(EVP_PKEY_id(pkey), EVP_PKEY_X448);
+    EVP_PKEY_free(pkey);
+    pkey = NULL;
+
+    ExpectNotNull(pkey = EVP_PKEY_new_raw_private_key(
+        EVP_PKEY_X448, NULL, rawPriv, sizeof(rawPriv)));
+    ExpectIntEQ(EVP_PKEY_id(pkey), EVP_PKEY_X448);
+
+    /* X448 is key-agreement only: signing must be rejected. */
+    ExpectNotNull(ctx = EVP_PKEY_CTX_new(pkey, NULL));
+    ExpectIntNE(EVP_PKEY_sign_init(ctx), WOLFSSL_SUCCESS);
+    EVP_PKEY_CTX_free(ctx);
+    ctx = NULL;
+    EVP_PKEY_free(pkey);
+    pkey = NULL;
+
+    /* Wrong raw lengths are rejected. */
+    ExpectNull(EVP_PKEY_new_raw_public_key(
+        EVP_PKEY_X448, NULL, rawPub, 16));
+    ExpectNull(EVP_PKEY_new_raw_private_key(
+        EVP_PKEY_X448, NULL, rawPriv, 16));
+
+    /* Generate two key pairs and confirm ECDH agreement is symmetric. */
+    ExpectNotNull(genCtx = EVP_PKEY_CTX_new_id(EVP_PKEY_X448, NULL));
+    ExpectIntEQ(EVP_PKEY_keygen_init(genCtx), WOLFSSL_SUCCESS);
+    ExpectIntEQ(EVP_PKEY_keygen(genCtx, &pkey), WOLFSSL_SUCCESS);
+    ExpectIntEQ(EVP_PKEY_keygen(genCtx, &peer), WOLFSSL_SUCCESS);
+    EVP_PKEY_CTX_free(genCtx);
+    genCtx = NULL;
+
+    ExpectNotNull(ctx = EVP_PKEY_CTX_new(pkey, NULL));
+    ExpectIntEQ(EVP_PKEY_derive_init(ctx), WOLFSSL_SUCCESS);
+    ExpectIntEQ(EVP_PKEY_derive_set_peer(ctx, peer), WOLFSSL_SUCCESS);
+    secretLen = sizeof(secretA);
+    ExpectIntEQ(EVP_PKEY_derive(ctx, secretA, &secretLen), WOLFSSL_SUCCESS);
+    ExpectIntEQ((int)secretLen, 56);
+    EVP_PKEY_CTX_free(ctx);
+    ctx = NULL;
+
+    ExpectNotNull(ctx = EVP_PKEY_CTX_new(peer, NULL));
+    ExpectIntEQ(EVP_PKEY_derive_init(ctx), WOLFSSL_SUCCESS);
+    ExpectIntEQ(EVP_PKEY_derive_set_peer(ctx, pkey), WOLFSSL_SUCCESS);
+    secretLen = sizeof(secretB);
+    ExpectIntEQ(EVP_PKEY_derive(ctx, secretB, &secretLen), WOLFSSL_SUCCESS);
+    ExpectIntEQ((int)secretLen, 56);
+    EVP_PKEY_CTX_free(ctx);
+    ctx = NULL;
+
+    ExpectIntEQ(XMEMCMP(secretA, secretB, 56), 0);
+
+    EVP_PKEY_free(peer);
+    EVP_PKEY_free(pkey);
+#endif
+    return EXPECT_RESULT();
+}
+
diff --git a/tests/api/test_evp_pkey.h b/tests/api/test_evp_pkey.h
index 2b77782a59..97388dfd7a 100644
--- a/tests/api/test_evp_pkey.h
+++ b/tests/api/test_evp_pkey.h
@@ -66,6 +66,8 @@ int test_wolfSSL_EVP_PKEY_print_public(void);
 int test_wolfSSL_EVP_PKEY_ed25519(void);
 int test_wolfSSL_CTX_use_PrivateKey_ed25519(void);
 int test_wolfSSL_EVP_PKEY_ed448(void);
+int test_wolfSSL_EVP_PKEY_x25519(void);
+int test_wolfSSL_EVP_PKEY_x448(void);
 
 #define TEST_EVP_PKEY_DECLS                                                    \
     TEST_DECL_GROUP("evp_pkey", test_wolfSSL_EVP_PKEY_CTX_new_id),             \
@@ -110,6 +112,8 @@ int test_wolfSSL_EVP_PKEY_ed448(void);
     TEST_DECL_GROUP("evp_pkey", test_wolfSSL_EVP_PKEY_print_public),           \
     TEST_DECL_GROUP("evp_pkey", test_wolfSSL_EVP_PKEY_ed25519),                \
     TEST_DECL_GROUP("evp_pkey", test_wolfSSL_CTX_use_PrivateKey_ed25519),      \
-    TEST_DECL_GROUP("evp_pkey", test_wolfSSL_EVP_PKEY_ed448)
+    TEST_DECL_GROUP("evp_pkey", test_wolfSSL_EVP_PKEY_ed448),                  \
+    TEST_DECL_GROUP("evp_pkey", test_wolfSSL_EVP_PKEY_x25519),                 \
+    TEST_DECL_GROUP("evp_pkey", test_wolfSSL_EVP_PKEY_x448)
 
 #endif /* WOLFCRYPT_TEST_EVP_PKEY_H */
diff --git a/tests/api/test_lms_xmss.c b/tests/api/test_lms_xmss.c
new file mode 100644
index 0000000000..2dc3ff2158
--- /dev/null
+++ b/tests/api/test_lms_xmss.c
@@ -0,0 +1,934 @@
+/* test_lms_xmss.c
+ *
+ * Copyright (C) 2006-2026 wolfSSL Inc.
+ *
+ * This file is part of wolfSSL.
+ *
+ * wolfSSL is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * wolfSSL is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1335, USA
+ */
+
+#include <tests/unit.h>
+
+#ifdef NO_INLINE
+    #include <wolfssl/wolfcrypt/misc.h>
+#else
+    #define WOLFSSL_MISC_INCLUDED
+    #include <wolfcrypt/src/misc.c>
+#endif
+
+#include <wolfssl/ssl.h>
+#include <wolfssl/wolfcrypt/asn.h>
+#include <tests/api/api.h>
+#include <tests/utils.h>
+#include <tests/api/test_lms_xmss.h>
+
+/*----------------------------------------------------------------------------*/
+/* LMS tests                                                                  */
+/*----------------------------------------------------------------------------*/
+
+#if defined(WOLFSSL_HAVE_LMS) && !defined(WOLFSSL_LMS_VERIFY_ONLY)
+
+#include <wolfssl/wolfcrypt/wc_lms.h>
+
+#define LMS_TEST_PRIV_KEY_FILE "/tmp/wolfssl_test_lms.key"
+
+static int test_lms_write_key(const byte* priv, word32 privSz, void* context)
+{
+    FILE* f = fopen((const char*)context, "wb");
+    int ret = WC_LMS_RC_SAVED_TO_NV_MEMORY;
+    if (f == NULL)
+        return -1;
+    if (fwrite(priv, 1, privSz, f) != privSz)
+        ret = -1;
+    fclose(f);
+    return ret;
+}
+
+static int test_lms_read_key(byte* priv, word32 privSz, void* context)
+{
+    FILE* f = fopen((const char*)context, "rb");
+    if (f == NULL)
+        return -1;
+    if (fread(priv, 1, privSz, f) == 0) {
+        fclose(f);
+        return -1;
+    }
+    fclose(f);
+    return WC_LMS_RC_READ_TO_MEMORY;
+}
+
+/* Helper: init an LMS key with callbacks and L1-H10-W8 params */
+static int test_lms_init_key(LmsKey* key, WC_RNG* rng)
+{
+    int ret;
+
+    ret = wc_LmsKey_Init(key, NULL, INVALID_DEVID);
+    if (ret != 0) return ret;
+
+#if !defined(WOLFSSL_LMS_MAX_HEIGHT) || (WOLFSSL_LMS_MAX_HEIGHT >= 10)
+    ret = wc_LmsKey_SetParameters(key, 1, 10, 8);
+#else
+    ret = wc_LmsKey_SetParameters(key, 1, 5, 8);
+#endif
+    if (ret != 0) return ret;
+
+    ret = wc_LmsKey_SetWriteCb(key, test_lms_write_key);
+    if (ret != 0) return ret;
+
+    ret = wc_LmsKey_SetReadCb(key, test_lms_read_key);
+    if (ret != 0) return ret;
+
+    ret = wc_LmsKey_SetContext(key, (void*)LMS_TEST_PRIV_KEY_FILE);
+    if (ret != 0) return ret;
+
+    (void)rng;
+    return 0;
+}
+
+#endif /* WOLFSSL_HAVE_LMS && !WOLFSSL_LMS_VERIFY_ONLY */
+
+/*
+ * Test basic LMS sign/verify with multiple signings.
+ * Uses L1-H10-W8 (1024 total signatures, 32-entry leaf cache).
+ */
+int test_wc_LmsKey_sign_verify(void)
+{
+    EXPECT_DECLS;
+#if defined(WOLFSSL_HAVE_LMS) && !defined(WOLFSSL_LMS_VERIFY_ONLY)
+    LmsKey  key;
+    WC_RNG  rng;
+    byte    msg[] = "test message for LMS signing";
+    byte    sig[2048];
+    word32  sigSz;
+    int     i;
+    int     numSigs = 5;
+
+    ExpectIntEQ(wc_InitRng(&rng), 0);
+
+    remove(LMS_TEST_PRIV_KEY_FILE);
+    ExpectIntEQ(test_lms_init_key(&key, &rng), 0);
+    ExpectIntEQ(wc_LmsKey_MakeKey(&key, &rng), 0);
+
+    for (i = 0; i < numSigs; i++) {
+        sigSz = sizeof(sig);
+        ExpectIntEQ(wc_LmsKey_Sign(&key, sig, &sigSz, msg, sizeof(msg)), 0);
+        ExpectIntEQ(wc_LmsKey_Verify(&key, sig, sigSz, msg, sizeof(msg)), 0);
+    }
+
+    wc_LmsKey_Free(&key);
+    wc_FreeRng(&rng);
+    remove(LMS_TEST_PRIV_KEY_FILE);
+#endif
+    return EXPECT_RESULT();
+}
+
+/*
+ * Test LMS key reload after advancing past the leaf cache window.
+ *
+ * Reproduces a heap-buffer-overflow bug in wc_lms_treehash_init() where the
+ * leaf cache write uses (i * hash_len) instead of ((i - leaf->idx) * hash_len).
+ * When q > max_cb (default 32), wc_LmsKey_Reload calls wc_hss_init_auth_path
+ * which calls wc_lms_treehash_init with q > 0, causing writes past the end of
+ * the leaf cache buffer.
+ *
+ * Reproduction steps:
+ *   1. Generate L1-H10-W8 key (cacheBits=5, max_cb=32)
+ *   2. Sign 33 times to advance q past the cache window
+ *   3. Free the key and reload from persisted state
+ *   4. Sign and verify after reload
+ *
+ * Without the fix: heap-buffer-overflow at wc_lms_impl.c:1965
+ * With the fix:    all operations succeed, signatures verify
+ */
+int test_wc_LmsKey_reload_cache(void)
+{
+    EXPECT_DECLS;
+#if defined(WOLFSSL_HAVE_LMS) && !defined(WOLFSSL_LMS_VERIFY_ONLY) && \
+    (!defined(WOLFSSL_LMS_MAX_HEIGHT) || (WOLFSSL_LMS_MAX_HEIGHT >= 10))
+    LmsKey  key;
+    LmsKey  vkey;
+    WC_RNG  rng;
+    byte    msg[] = "test message for LMS signing";
+    byte    sig[2048];
+    word32  sigSz;
+    byte    pub[64];
+    word32  pubSz = sizeof(pub);
+    int     i;
+    /* Sign 33 times to advance q past the 32-entry cache window. */
+    int     preSigs = 33;
+
+    ExpectIntEQ(wc_InitRng(&rng), 0);
+
+    /* Phase 1: Generate key and sign past cache window */
+    remove(LMS_TEST_PRIV_KEY_FILE);
+    ExpectIntEQ(test_lms_init_key(&key, &rng), 0);
+    ExpectIntEQ(wc_LmsKey_MakeKey(&key, &rng), 0);
+
+    for (i = 0; i < preSigs; i++) {
+        sigSz = sizeof(sig);
+        ExpectIntEQ(wc_LmsKey_Sign(&key, sig, &sigSz, msg, sizeof(msg)), 0);
+    }
+
+    /* Save public key for verification after reload */
+    ExpectIntEQ(wc_LmsKey_ExportPubRaw(&key, pub, &pubSz), 0);
+
+    wc_LmsKey_Free(&key);
+
+    /* Phase 2: Reload key. Triggers wc_lms_treehash_init with q=33 */
+    ExpectIntEQ(test_lms_init_key(&key, &rng), 0);
+    ExpectIntEQ(wc_LmsKey_Reload(&key), 0);
+
+    /* Phase 3: Sign after reload and verify with separate verify-only key */
+    sigSz = sizeof(sig);
+    ExpectIntEQ(wc_LmsKey_Sign(&key, sig, &sigSz, msg, sizeof(msg)), 0);
+
+    ExpectIntEQ(wc_LmsKey_Init(&vkey, NULL, INVALID_DEVID), 0);
+#if !defined(WOLFSSL_LMS_MAX_HEIGHT) || (WOLFSSL_LMS_MAX_HEIGHT >= 10)
+    ExpectIntEQ(wc_LmsKey_SetParameters(&vkey, 1, 10, 8), 0);
+#else
+    ExpectIntEQ(wc_LmsKey_SetParameters(&vkey, 1, 5, 8), 0);
+#endif
+    ExpectIntEQ(wc_LmsKey_ImportPubRaw(&vkey, pub, pubSz), 0);
+    ExpectIntEQ(wc_LmsKey_Verify(&vkey, sig, sigSz, msg, sizeof(msg)), 0);
+
+    wc_LmsKey_Free(&vkey);
+    wc_LmsKey_Free(&key);
+    wc_FreeRng(&rng);
+    remove(LMS_TEST_PRIV_KEY_FILE);
+#endif
+    return EXPECT_RESULT();
+}
+
+/*----------------------------------------------------------------------------*/
+/* RFC 9802 (HSS/LMS and XMSS/XMSS^MT in X.509) tests                         */
+/*----------------------------------------------------------------------------*/
+
+/* For every committed self-signed test certificate confirm:
+ *   - wc_ParseCert succeeds on the RFC 9802 AlgorithmIdentifier encoding
+ *     (OID-only SEQUENCE, no NULL parameters)
+ *   - keyOID and signatureOID are set to the expected values
+ *   - loading as a trust anchor and verifying the same bytes through
+ *     wolfSSL_CertManagerVerifyBuffer exercises the ConfirmSignature
+ *     path and succeeds on a valid cert
+ *   - flipping a byte in the signature AND flipping a byte in the
+ *     TBSCertificate both cause verification to fail.
+ *
+ * Test vectors are in certs/lms/ and certs/xmss/, generated with Bouncy
+ * Castle 1.81. BC's default XMSS / XMSS^MT X.509 encoding uses pre-
+ * standard ISARA OIDs and wraps the raw RFC 8391 pub key in an OCTET
+ * STRING, so the fixtures were produced with a small generator that
+ * overrides the AlgorithmIdentifier and SPKI to match RFC 9802. */
+#if (defined(WOLFSSL_HAVE_LMS) || defined(WOLFSSL_HAVE_XMSS)) && \
+    !defined(NO_FILESYSTEM) && !defined(NO_CERTS)
+/* Sanity bound on a test fixture cert. The largest BC-generated
+ * fixture we ship (XMSS^MT 40/8) is ~19 KiB; 1 MiB is well above
+ * any realistic RFC 9802 cert and catches a wild XFTELL. Typed as
+ * long to match XFTELL's return so the size comparison below isn't
+ * a mixed long-vs-int compare. */
+#define RFC9802_TEST_MAX_CERT_SIZE ((long)(1L << 20))
+
+/* Load a whole file into a freshly-allocated buffer. Caller frees. */
+static int rfc9802_load_file(const char* path, byte** out, int* outLen)
+{
+    EXPECT_DECLS;
+    XFILE  f = XBADFILE;
+    long   sz = 0;
+    size_t got = 0;
+    byte*  buf = NULL;
+
+    *out = NULL;
+    *outLen = 0;
+    ExpectTrue((f = XFOPEN(path, "rb")) != XBADFILE);
+    if (f == XBADFILE)
+        return TEST_FAIL;
+    if (XFSEEK(f, 0, XSEEK_END) == 0)
+        sz = XFTELL(f);
+    (void)XFSEEK(f, 0, XSEEK_SET);
+    ExpectIntGT(sz, 0);
+    ExpectIntLT(sz, RFC9802_TEST_MAX_CERT_SIZE);
+    /* Hard-fail before XMALLOC if XFSEEK / XFTELL produced an unusable
+     * size: ExpectInt* records the failure but doesn't short-circuit,
+     * so without this guard a -1 from XFTELL would cast to a multi-GiB
+     * (size_t) allocation, and a 0 would request a zero-byte malloc. */
+    if (sz <= 0 || sz >= RFC9802_TEST_MAX_CERT_SIZE) {
+        XFCLOSE(f);
+        return TEST_FAIL;
+    }
+    ExpectNotNull(buf = (byte*)XMALLOC((size_t)sz, NULL,
+        DYNAMIC_TYPE_TMP_BUFFER));
+    if (buf != NULL) {
+        got = XFREAD(buf, 1, (size_t)sz, f);
+        ExpectIntEQ(got, (size_t)sz);
+        /* On a short read the caller would otherwise proceed with a
+         * partially-initialized buffer and produce cascading parse
+         * failures driven by the uninitialized tail. Free here so the
+         * caller's `if (buf == NULL) return TEST_FAIL;` short-circuits
+         * cleanly with a single recorded failure. */
+        if (got != (size_t)sz) {
+            XFREE(buf, NULL, DYNAMIC_TYPE_TMP_BUFFER);
+            buf = NULL;
+            sz = 0;
+        }
+    }
+    XFCLOSE(f);
+    *out = buf;
+    *outLen = (int)sz;
+    return EXPECT_RESULT();
+}
+
+static int rfc9802_verify_one_cert(const char* path, word32 expectedKeyOID,
+    word32 expectedSigOID)
+{
+    EXPECT_DECLS;
+    byte*                 buf = NULL;
+    byte*                 tampered = NULL;
+    int                   bytes = 0;
+    DecodedCert           cert;
+    WOLFSSL_CERT_MANAGER* cm = NULL;
+    word32                certBegin = 0;
+    word32                sigIndex = 0;
+
+    ExpectIntEQ(rfc9802_load_file(path, &buf, &bytes), TEST_SUCCESS);
+    if (buf == NULL)
+        return TEST_FAIL;
+
+    /* Parse + check OIDs, capture certBegin and sigIndex for later tamper. */
+    wc_InitDecodedCert(&cert, buf, (word32)bytes, NULL);
+    ExpectIntEQ(wc_ParseCert(&cert, CERT_TYPE, NO_VERIFY, NULL), 0);
+    ExpectIntEQ((int)cert.keyOID, (int)expectedKeyOID);
+    ExpectIntEQ((int)cert.signatureOID, (int)expectedSigOID);
+    certBegin = cert.certBegin;
+    sigIndex  = cert.sigIndex;
+    wc_FreeDecodedCert(&cert);
+
+    /* Full verify against a self-installed trust anchor. */
+    ExpectNotNull(cm = wolfSSL_CertManagerNew());
+    ExpectIntEQ(wolfSSL_CertManagerLoadCABuffer(cm, buf, (long)bytes,
+        WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_CertManagerVerifyBuffer(cm, buf, (long)bytes,
+        WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
+    if (cm != NULL) {
+        wolfSSL_CertManagerFree(cm);
+        cm = NULL;
+    }
+
+    ExpectNotNull(tampered = (byte*)XMALLOC((size_t)bytes, NULL,
+        DYNAMIC_TYPE_TMP_BUFFER));
+
+    /* Negative 1: flip a byte inside the signatureValue BIT STRING.
+     * Everything after sigIndex is the signatureAlgorithm + the BIT
+     * STRING payload, so flipping the last byte is always inside the
+     * signature content. */
+    if (tampered != NULL) {
+        XMEMCPY(tampered, buf, (size_t)bytes);
+        tampered[bytes - 1] ^= 0x01;
+        ExpectNotNull(cm = wolfSSL_CertManagerNew());
+        ExpectIntEQ(wolfSSL_CertManagerLoadCABuffer(cm, buf, (long)bytes,
+            WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
+        ExpectIntNE(wolfSSL_CertManagerVerifyBuffer(cm, tampered,
+            (long)bytes, WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
+        if (cm != NULL) {
+            wolfSSL_CertManagerFree(cm);
+            cm = NULL;
+        }
+    }
+
+    /* Negative 2: flip a byte at the midpoint of the TBSCertificate. The
+     * TBS is the first element of the outer Certificate SEQUENCE and
+     * its bytes lie between (certBegin + outerSeqHeader) and sigIndex.
+     * Picking the midpoint ensures we're inside TBS regardless of the
+     * fixture's DN / extensions layout. */
+    if (tampered != NULL && sigIndex > certBegin + 8U) {
+        word32 midTbs = certBegin + 8 + ((sigIndex - (certBegin + 8)) / 2);
+        XMEMCPY(tampered, buf, (size_t)bytes);
+        tampered[midTbs] ^= 0x01;
+        ExpectNotNull(cm = wolfSSL_CertManagerNew());
+        ExpectIntEQ(wolfSSL_CertManagerLoadCABuffer(cm, buf, (long)bytes,
+            WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
+        ExpectIntNE(wolfSSL_CertManagerVerifyBuffer(cm, tampered,
+            (long)bytes, WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
+        if (cm != NULL) {
+            wolfSSL_CertManagerFree(cm);
+            cm = NULL;
+        }
+    }
+
+    /* The fixtures MUST carry a KeyUsage extension with at least one of
+     * digitalSignature / nonRepudiation / keyCertSign / cRLSign set per
+     * RFC 9802 sec 3. Re-parse and assert that wolfSSL recorded a non-
+     * empty set of KeyUsage bits from one of those values. */
+    wc_InitDecodedCert(&cert, buf, (word32)bytes, NULL);
+    ExpectIntEQ(wc_ParseCert(&cert, CERT_TYPE, NO_VERIFY, NULL), 0);
+    ExpectIntEQ(cert.extKeyUsageSet, 1);
+    ExpectIntNE(cert.extKeyUsage & (KEYUSE_DIGITAL_SIG | KEYUSE_CONTENT_COMMIT |
+        KEYUSE_KEY_CERT_SIGN | KEYUSE_CRL_SIGN), 0);
+    wc_FreeDecodedCert(&cert);
+
+    XFREE(tampered, NULL, DYNAMIC_TYPE_TMP_BUFFER);
+    XFREE(buf, NULL, DYNAMIC_TYPE_TMP_BUFFER);
+    return EXPECT_RESULT();
+}
+#endif
+
+/* Direct wolfCrypt-level negative tests for the parameter-derivation
+ * helpers used by the RFC 9802 parse path. These exercise failure modes
+ * (unknown algorithm bytes, truncated inputs, mismatches) that a real
+ * cert body wouldn't easily reach. */
+#if defined(WOLFSSL_HAVE_LMS)
+static int rfc9802_lms_import_negative(void)
+{
+    EXPECT_DECLS;
+    LmsKey key;
+    /* 60-byte buffer matches HSS_PUBLIC_KEY_LEN(32), just like a valid
+     * SHA-256/M32/H5 key; the algorithm-type bytes are junk so param
+     * derivation must fail cleanly. */
+    byte   junk[60];
+
+    XMEMSET(junk, 0, sizeof(junk));
+    /* levels=1, lmsType=0xFFFFFFFF, lmOtsType=0xFFFFFFFF. */
+    junk[3] = 1;
+    XMEMSET(junk + 4, 0xFF, 4);
+    XMEMSET(junk + 8, 0xFF, 4);
+
+    /* Unknown algorithm types must be rejected. */
+    ExpectIntEQ(wc_LmsKey_Init(&key, NULL, INVALID_DEVID), 0);
+    ExpectIntEQ(wc_LmsKey_ImportPubRaw(&key, junk, sizeof(junk)),
+        WC_NO_ERR_TRACE(NOT_COMPILED_IN));
+    wc_LmsKey_Free(&key);
+
+    /* Too-short buffer: only L + lmsType, no lmOtsType. */
+    ExpectIntEQ(wc_LmsKey_Init(&key, NULL, INVALID_DEVID), 0);
+    ExpectIntEQ(wc_LmsKey_ImportPubRaw(&key, junk, 8),
+        WC_NO_ERR_TRACE(BUFFER_E));
+    wc_LmsKey_Free(&key);
+
+#if !defined(WOLFSSL_NO_LMS_SHA256_256)
+    /* The two cases below pin specific SHA-256/M32 parameter codes
+     * (L1_H5_W8, L1_H5_W4, L1_H10_W2). Skip them in builds where the
+     * SHA-256/M32 family is disabled -- the family-agnostic checks
+     * above (junk algorithm types, too-short buffer, GetSigLen on
+     * unconfigured key) still cover the universal invariants. */
+
+    /* Pre-set params that disagree with the raw key's algorithm bytes:
+     * configure H=5/W=8 but feed buffer that claims H=10 / W=2. */
+    XMEMSET(junk, 0, sizeof(junk));
+    junk[3] = 1;       /* levels=1     */
+    junk[7] = 6;       /* lmsType = LMS_SHA256_M32_H10 = 6 */
+    junk[11] = 2;      /* lmOtsType = LMOTS_SHA256_N32_W2 = 2 */
+    ExpectIntEQ(wc_LmsKey_Init(&key, NULL, INVALID_DEVID), 0);
+    ExpectIntEQ(wc_LmsKey_SetParameters(&key, 1, 5, 8), 0);
+    ExpectIntEQ(wc_LmsKey_ImportPubRaw(&key, junk, sizeof(junk)),
+        WC_NO_ERR_TRACE(BAD_FUNC_ARG));
+    wc_LmsKey_Free(&key);
+#endif /* !WOLFSSL_NO_LMS_SHA256_256 */
+
+    /* GetSigLen on a key with no params set must not NULL-deref the
+     * params pointer; it must return BAD_FUNC_ARG instead. */
+    {
+        word32 sigLen = 0;
+        ExpectIntEQ(wc_LmsKey_Init(&key, NULL, INVALID_DEVID), 0);
+        ExpectIntEQ(wc_LmsKey_GetSigLen(&key, &sigLen),
+            WC_NO_ERR_TRACE(BAD_FUNC_ARG));
+        wc_LmsKey_Free(&key);
+    }
+
+#if !defined(WOLFSSL_NO_LMS_SHA256_256)
+    /* Partial-write invariant: a length mismatch after a successful
+     * auto-derive must leave key->params NULL. Build a buffer whose
+     * leading u32str(L) || lmsType || lmOtsType identifies a known
+     * parameter set, but truncate to one byte less than the real pub
+     * key length so the post-derive length check fails. */
+    {
+        byte truncated[59];   /* HSS_PUBLIC_KEY_LEN(32) is 60 */
+        XMEMSET(truncated, 0, sizeof(truncated));
+        truncated[3] = 1;     /* L = 1 */
+        truncated[7] = 5;     /* lmsType = LMS_SHA256_M32_H5 */
+        truncated[11] = 4;    /* lmOtsType = LMOTS_SHA256_N32_W4 */
+        ExpectIntEQ(wc_LmsKey_Init(&key, NULL, INVALID_DEVID), 0);
+        ExpectNull(key.params);
+        ExpectIntEQ(wc_LmsKey_ImportPubRaw(&key, truncated,
+            sizeof(truncated)), WC_NO_ERR_TRACE(BUFFER_E));
+        ExpectNull(key.params);
+        wc_LmsKey_Free(&key);
+    }
+#endif /* !WOLFSSL_NO_LMS_SHA256_256 */
+
+    return EXPECT_RESULT();
+}
+#endif
+
+#if defined(WOLFSSL_HAVE_XMSS)
+static int rfc9802_xmss_import_negative(void)
+{
+    EXPECT_DECLS;
+    XmssKey key;
+    byte    junk[8];
+
+    XMEMSET(junk, 0, sizeof(junk));
+
+    /* Too-short buffer. */
+    ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
+    ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, junk, 2, 0),
+        WC_NO_ERR_TRACE(BUFFER_E));
+    wc_XmssKey_Free(&key);
+
+    /* Unknown OID (all-zero) for both XMSS and XMSS^MT. */
+    ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
+    ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, junk, sizeof(junk), 0),
+        WC_NO_ERR_TRACE(NOT_COMPILED_IN));
+    wc_XmssKey_Free(&key);
+    ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
+    ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, junk, sizeof(junk), 1),
+        WC_NO_ERR_TRACE(NOT_COMPILED_IN));
+    wc_XmssKey_Free(&key);
+
+    /* NULL key / input. */
+    ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(NULL, junk, sizeof(junk), 0),
+        WC_NO_ERR_TRACE(BAD_FUNC_ARG));
+    ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
+    ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, NULL, 8, 0),
+        WC_NO_ERR_TRACE(BAD_FUNC_ARG));
+    wc_XmssKey_Free(&key);
+
+    /* GetSigLen on a key with no params set must not NULL-deref the
+     * params pointer; it must return BAD_FUNC_ARG instead. */
+    {
+        word32 sigLen = 0;
+        ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
+        ExpectIntEQ(wc_XmssKey_GetSigLen(&key, &sigLen),
+            WC_NO_ERR_TRACE(BAD_FUNC_ARG));
+        wc_XmssKey_Free(&key);
+    }
+
+    /* Once params have been configured (state != INITED), the OID
+     * prefix in the raw key MUST match key->oid and is_xmssmt MUST
+     * match key->is_xmssmt. Set XMSS-SHA2_10_256 and feed a valid-
+     * sized buffer whose 4-byte OID prefix is bogus -> BAD_FUNC_ARG. */
+    {
+        byte mismatch[XMSS_SHA256_PUBLEN];
+        ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
+        ExpectIntEQ(wc_XmssKey_SetParamStr(&key, "XMSS-SHA2_10_256"), 0);
+        XMEMSET(mismatch, 0, sizeof(mismatch));
+        mismatch[3] = 0x77; /* nonsense OID */
+        ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, mismatch,
+            sizeof(mismatch), 0), WC_NO_ERR_TRACE(BAD_FUNC_ARG));
+        /* Same buffer with the correct OID, but is_xmssmt hint
+         * contradicts the configured family -> BAD_FUNC_ARG. */
+        mismatch[3] = 0x01; /* WC_XMSS_OID_SHA2_10_256 */
+        ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, mismatch,
+            sizeof(mismatch), 1), WC_NO_ERR_TRACE(BAD_FUNC_ARG));
+        wc_XmssKey_Free(&key);
+    }
+
+    /* Partial-write invariant: a length mismatch after a successful
+     * auto-derive must leave the key in its INITED state, with
+     * key->params NULL. */
+    {
+        byte truncated[XMSS_SHA256_PUBLEN - 1];
+        XMEMSET(truncated, 0, sizeof(truncated));
+        truncated[3] = 0x01;
+        ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
+        ExpectNull(key.params);
+        ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, truncated,
+            sizeof(truncated), 0), WC_NO_ERR_TRACE(BUFFER_E));
+        ExpectNull(key.params);
+        wc_XmssKey_Free(&key);
+    }
+
+    /* is_xmssmt disambiguation: XMSS oid=1 and XMSS^MT oid=1 share
+     * the wire-numeric value but resolve to different parameter sets.
+     * Importing the same 68-byte buffer with hint=0 vs hint=1 must
+     * land in different tables and produce distinct is_xmssmt. */
+    {
+        byte buf[XMSS_SHA256_PUBLEN];
+        XMEMSET(buf, 0, sizeof(buf));
+        buf[3] = 0x01;
+
+        ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
+        ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, buf, sizeof(buf), 0), 0);
+        ExpectIntEQ((int)key.is_xmssmt, 0);
+        wc_XmssKey_Free(&key);
+
+        ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
+        ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, buf, sizeof(buf), 1), 0);
+        ExpectIntEQ((int)key.is_xmssmt, 1);
+        wc_XmssKey_Free(&key);
+    }
+
+    /* Lenient state: re-importing the same pub key into a VERIFYONLY
+     * key (params set, no private material) succeeds. The second
+     * call exercises the lenient-state branch. */
+    {
+        byte buf[XMSS_SHA256_PUBLEN];
+        XMEMSET(buf, 0, sizeof(buf));
+        buf[3] = 0x01;
+
+        ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
+        ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, buf, sizeof(buf), 0), 0);
+        ExpectIntEQ((int)key.state, (int)WC_XMSS_STATE_VERIFYONLY);
+        ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, buf, sizeof(buf), 0), 0);
+        ExpectIntEQ((int)key.state, (int)WC_XMSS_STATE_VERIFYONLY);
+        wc_XmssKey_Free(&key);
+    }
+
+    /* Strict signature-length check: wc_XmssKey_Verify rejects any
+     * sigLen != key->params->sig_len. This guards every consumer
+     * (RFC 9802 X.509, PKCS#7, CMS, ...) against a longer wrapper that
+     * happens to start with a valid signature. Construct a key in
+     * VERIFYONLY state, then verify with sig_len + 1 and sig_len - 1
+     * byte buffers; both must fail with BUFFER_E before any crypto
+     * runs. The buffer contents are irrelevant since the length check
+     * fires first. */
+    {
+        byte    pub[XMSS_SHA256_PUBLEN];
+        byte*   sigBuf = NULL;
+        word32  sigLen = 0;
+        const byte msg[1] = { 0 };
+
+        XMEMSET(pub, 0, sizeof(pub));
+        pub[3] = 0x01;
+        ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
+        ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, pub, sizeof(pub), 0), 0);
+        ExpectIntEQ((int)key.state, (int)WC_XMSS_STATE_VERIFYONLY);
+        ExpectIntEQ(wc_XmssKey_GetSigLen(&key, &sigLen), 0);
+        ExpectIntGT(sigLen, 0);
+        ExpectNotNull(sigBuf = (byte*)XMALLOC((size_t)sigLen + 1, NULL,
+            DYNAMIC_TYPE_TMP_BUFFER));
+        if (sigBuf != NULL) {
+            XMEMSET(sigBuf, 0, (size_t)sigLen + 1);
+            ExpectIntEQ(wc_XmssKey_Verify(&key, sigBuf, sigLen + 1,
+                msg, (int)sizeof(msg)), WC_NO_ERR_TRACE(BUFFER_E));
+            ExpectIntEQ(wc_XmssKey_Verify(&key, sigBuf, sigLen - 1,
+                msg, (int)sizeof(msg)), WC_NO_ERR_TRACE(BUFFER_E));
+            XFREE(sigBuf, NULL, DYNAMIC_TYPE_TMP_BUFFER);
+        }
+        wc_XmssKey_Free(&key);
+    }
+
+    /* BAD_STATE_E branch: WC_XMSS_STATE_OK must be rejected. Reaching
+     * OK normally requires a successful private-key Reload / sign,
+     * which is unavailable in WOLFSSL_XMSS_VERIFY_ONLY builds. Force
+     * the state directly to exercise the rejection without coupling
+     * this helper to the signing test fixture; sk stays NULL so Free
+     * is still safe. */
+    {
+        byte pub[XMSS_SHA256_PUBLEN];
+
+        XMEMSET(pub, 0, sizeof(pub));
+        pub[3] = 0x01;
+        ExpectIntEQ(wc_XmssKey_Init(&key, NULL, INVALID_DEVID), 0);
+        ExpectIntEQ(wc_XmssKey_SetParamStr(&key, "XMSS-SHA2_10_256"), 0);
+        key.state = WC_XMSS_STATE_OK;
+        ExpectIntEQ(wc_XmssKey_ImportPubRaw_ex(&key, pub, sizeof(pub), 0),
+            WC_NO_ERR_TRACE(BAD_STATE_E));
+        wc_XmssKey_Free(&key);
+    }
+
+    return EXPECT_RESULT();
+}
+#endif
+
+/* Walk the AlgorithmIdentifier SEQUENCE that begins at sigIndex and
+ * locate the byte offset of the last byte of its OID content. Handles
+ * both short-form (length < 128) and long-form DER length encodings,
+ * so a future fixture-regenerator that emits longer OIDs / SEQUENCEs
+ * still drives this test rather than tripping the loud-fail branch.
+ *
+ * Returns 0 on success with *oidLastByte set; returns -1 on any DER
+ * shape mismatch. */
+#if defined(WOLFSSL_HAVE_XMSS) && !defined(NO_FILESYSTEM) && !defined(NO_CERTS)
+static int rfc9802_find_sig_alg_oid_last_byte(const byte* buf, word32 bufLen,
+    word32 sigIndex, word32* oidLastByte)
+{
+    word32 idx = sigIndex;
+    word32 oidContentLen = 0;
+
+    /* AlgorithmIdentifier ::= SEQUENCE { algorithm OID, ... } */
+    if (idx >= bufLen || buf[idx] != 0x30)
+        return -1;
+    idx++;
+    /* Skip SEQUENCE length (short or long form). */
+    if (idx >= bufLen)
+        return -1;
+    if (buf[idx] < 0x80) {
+        idx++;
+    }
+    else {
+        word32 nbytes = (word32)(buf[idx] & 0x7F);
+        if (nbytes == 0 || nbytes > 4 || idx + 1 + nbytes > bufLen)
+            return -1;
+        idx += 1 + nbytes;
+    }
+    /* algorithm OID tag. */
+    if (idx >= bufLen || buf[idx] != 0x06)
+        return -1;
+    idx++;
+    /* OID length (short or long form). */
+    if (idx >= bufLen)
+        return -1;
+    if (buf[idx] < 0x80) {
+        oidContentLen = buf[idx];
+        idx++;
+    }
+    else {
+        word32 nbytes = (word32)(buf[idx] & 0x7F);
+        word32 i;
+        if (nbytes == 0 || nbytes > 4 || idx + 1 + nbytes > bufLen)
+            return -1;
+        for (i = 0; i < nbytes; i++)
+            oidContentLen = (oidContentLen << 8) | buf[idx + 1 + i];
+        idx += 1 + nbytes;
+    }
+    if (oidContentLen == 0 || idx + oidContentLen > bufLen)
+        return -1;
+    *oidLastByte = idx + oidContentLen - 1;
+    return 0;
+}
+
+/* Helper: load fixture, locate last byte of outer signatureAlgorithm
+ * OID, patch it from `expected` to `swap`, and assert that verifying
+ * the patched cert against itself as a trust anchor fails. */
+static int rfc9802_assert_oid_patch_breaks_verify(const char* path,
+    byte expectedLastByte, byte patchedLastByte)
+{
+    EXPECT_DECLS;
+    byte*                 buf = NULL;
+    int                   bytes = 0;
+    DecodedCert           cert;
+    WOLFSSL_CERT_MANAGER* cm = NULL;
+    word32                sigIndex = 0;
+    word32                lastOidByte = 0;
+
+    ExpectIntEQ(rfc9802_load_file(path, &buf, &bytes), TEST_SUCCESS);
+    if (buf == NULL)
+        return TEST_FAIL;
+
+    wc_InitDecodedCert(&cert, buf, (word32)bytes, NULL);
+    ExpectIntEQ(wc_ParseCert(&cert, CERT_TYPE, NO_VERIFY, NULL), 0);
+    sigIndex = cert.sigIndex;
+    wc_FreeDecodedCert(&cert);
+
+    ExpectIntEQ(rfc9802_find_sig_alg_oid_last_byte(buf, (word32)bytes,
+        sigIndex, &lastOidByte), 0);
+    /* Sanity-check the fixture matches the family the caller asserted,
+     * so a future regenerator swapping fixtures fails loudly here
+     * rather than silently testing the wrong direction. */
+    ExpectIntEQ((int)buf[lastOidByte], (int)expectedLastByte);
+
+    if (lastOidByte < (word32)bytes &&
+            buf[lastOidByte] == expectedLastByte) {
+        buf[lastOidByte] = patchedLastByte;
+        ExpectNotNull(cm = wolfSSL_CertManagerNew());
+        /* After the patch the cert's outer signatureAlgorithm and SPKI
+         * disagree. Verification must fail somewhere (at parse, at
+         * load, or at ConfirmSignature). The load is best-effort -
+         * some shape changes get caught there, others only at verify. */
+        (void)wolfSSL_CertManagerLoadCABuffer(cm, buf, (long)bytes,
+            WOLFSSL_FILETYPE_ASN1);
+        ExpectIntNE(wolfSSL_CertManagerVerifyBuffer(cm, buf,
+            (long)bytes, WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
+        if (cm != NULL) {
+            wolfSSL_CertManagerFree(cm);
+            cm = NULL;
+        }
+    }
+
+    XFREE(buf, NULL, DYNAMIC_TYPE_TMP_BUFFER);
+    return EXPECT_RESULT();
+}
+
+/* X.509-level negative: swap the outer signatureAlgorithm OID byte so
+ * the cert declares XMSS where the SPKI is XMSS^MT, and vice versa.
+ * SigOidMatchesKeyOid must reject both directions before any crypto. */
+static int rfc9802_xmss_sig_oid_mismatch(void)
+{
+    EXPECT_DECLS;
+    /* XMSS sigOID ends 0x22; XMSS^MT sigOID ends 0x23. Patch each
+     * direction so the asymmetric-key path is exercised both ways -
+     * a regression that only stripped the check from one branch of
+     * SigOidMatchesKeyOid would otherwise be missed. */
+    ExpectIntEQ(rfc9802_assert_oid_patch_breaks_verify(
+        "./certs/xmss/bc_xmss_sha2_10_256_root.der",
+        /* expected XMSS */ 0x22, /* patched to XMSS^MT */ 0x23),
+        TEST_SUCCESS);
+    ExpectIntEQ(rfc9802_assert_oid_patch_breaks_verify(
+        "./certs/xmss/bc_xmssmt_sha2_20_2_256_root.der",
+        /* expected XMSS^MT */ 0x23, /* patched to XMSS */ 0x22),
+        TEST_SUCCESS);
+    return EXPECT_RESULT();
+}
+#endif
+
+/* Exercise a real CA -> leaf certificate chain, not just self-signed.
+ * Loads the CA as a trust anchor and verifies the leaf against it. */
+#if defined(WOLFSSL_HAVE_LMS) && !defined(NO_FILESYSTEM) && !defined(NO_CERTS)
+static int rfc9802_lms_chain_verify(void)
+{
+    EXPECT_DECLS;
+    byte*                 caBuf   = NULL;
+    byte*                 leafBuf = NULL;
+    int                   caLen   = 0;
+    int                   leafLen = 0;
+    WOLFSSL_CERT_MANAGER* cm      = NULL;
+
+    ExpectIntEQ(rfc9802_load_file("./certs/lms/bc_lms_chain_ca.der",
+        &caBuf, &caLen), TEST_SUCCESS);
+    ExpectIntEQ(rfc9802_load_file("./certs/lms/bc_lms_chain_leaf.der",
+        &leafBuf, &leafLen), TEST_SUCCESS);
+
+    ExpectNotNull(cm = wolfSSL_CertManagerNew());
+    /* Only the CA is a trust anchor; the leaf is verified against it. */
+    ExpectIntEQ(wolfSSL_CertManagerLoadCABuffer(cm, caBuf, (long)caLen,
+        WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_CertManagerVerifyBuffer(cm, leafBuf, (long)leafLen,
+        WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
+
+    /* Without loading the CA the leaf must NOT verify. */
+    if (cm != NULL) {
+        wolfSSL_CertManagerFree(cm);
+        cm = NULL;
+    }
+    ExpectNotNull(cm = wolfSSL_CertManagerNew());
+    ExpectIntNE(wolfSSL_CertManagerVerifyBuffer(cm, leafBuf, (long)leafLen,
+        WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
+    if (cm != NULL) {
+        wolfSSL_CertManagerFree(cm);
+        cm = NULL;
+    }
+
+    XFREE(leafBuf, NULL, DYNAMIC_TYPE_TMP_BUFFER);
+    XFREE(caBuf, NULL, DYNAMIC_TYPE_TMP_BUFFER);
+    return EXPECT_RESULT();
+}
+#endif
+
+/* Mirror of rfc9802_lms_chain_verify but for an XMSS CA -> leaf pair. */
+#if defined(WOLFSSL_HAVE_XMSS) && !defined(NO_FILESYSTEM) && !defined(NO_CERTS)
+static int rfc9802_xmss_chain_verify(void)
+{
+    EXPECT_DECLS;
+    byte*                 caBuf   = NULL;
+    byte*                 leafBuf = NULL;
+    int                   caLen   = 0;
+    int                   leafLen = 0;
+    WOLFSSL_CERT_MANAGER* cm      = NULL;
+
+    ExpectIntEQ(rfc9802_load_file("./certs/xmss/bc_xmss_chain_ca.der",
+        &caBuf, &caLen), TEST_SUCCESS);
+    ExpectIntEQ(rfc9802_load_file("./certs/xmss/bc_xmss_chain_leaf.der",
+        &leafBuf, &leafLen), TEST_SUCCESS);
+
+    ExpectNotNull(cm = wolfSSL_CertManagerNew());
+    ExpectIntEQ(wolfSSL_CertManagerLoadCABuffer(cm, caBuf, (long)caLen,
+        WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_CertManagerVerifyBuffer(cm, leafBuf, (long)leafLen,
+        WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
+
+    if (cm != NULL) {
+        wolfSSL_CertManagerFree(cm);
+        cm = NULL;
+    }
+    ExpectNotNull(cm = wolfSSL_CertManagerNew());
+    ExpectIntNE(wolfSSL_CertManagerVerifyBuffer(cm, leafBuf, (long)leafLen,
+        WOLFSSL_FILETYPE_ASN1), WOLFSSL_SUCCESS);
+    if (cm != NULL) {
+        wolfSSL_CertManagerFree(cm);
+        cm = NULL;
+    }
+
+    XFREE(leafBuf, NULL, DYNAMIC_TYPE_TMP_BUFFER);
+    XFREE(caBuf, NULL, DYNAMIC_TYPE_TMP_BUFFER);
+    return EXPECT_RESULT();
+}
+#endif
+
+int test_rfc9802_lms_x509_verify(void)
+{
+    EXPECT_DECLS;
+#if defined(WOLFSSL_HAVE_LMS)
+#if !defined(NO_FILESYSTEM) && !defined(NO_CERTS) && \
+    !defined(WOLFSSL_NO_LMS_SHA256_256)
+    /* Mixed single-level LMS and multi-level HSS fixtures. The HSS
+     * public key carries only the top-level LMS/LM-OTS types, so
+     * wc_LmsKey_ImportPubRaw's auto-derive path searches the map
+     * by (levels, lmsType, lmOtsType). The bc_lms_native_bc_root
+     * fixture is generated through Bouncy Castle's stock
+     * JcaContentSignerBuilder("LMS") + JcaX509v3CertificateBuilder
+     * with no overrides; including it here is the cross-impl interop
+     * gate (BC's native LMS X.509 path is RFC 9802-compliant for HSS/
+     * LMS, so wolfSSL must accept it end-to-end).
+     *
+     * All fixtures use the SHA-256/M32 family, so the whole block
+     * is gated on that family being compiled in. Truncated SHA-256/192
+     * or SHAKE-only builds skip this block. */
+    static const char* const lmsFiles[] = {
+        "./certs/lms/bc_lms_sha256_h5_w4_root.der",
+#if !defined(WOLFSSL_LMS_MAX_HEIGHT) || (WOLFSSL_LMS_MAX_HEIGHT >= 10)
+        "./certs/lms/bc_lms_sha256_h10_w8_root.der",
+#endif
+#if !defined(WOLFSSL_LMS_MAX_LEVELS) || (WOLFSSL_LMS_MAX_LEVELS >= 2)
+        "./certs/lms/bc_hss_L2_H5_W8_root.der",
+#endif
+#if !defined(WOLFSSL_LMS_MAX_LEVELS) || (WOLFSSL_LMS_MAX_LEVELS >= 3)
+        "./certs/lms/bc_hss_L3_H5_W4_root.der",
+#endif
+        "./certs/lms/bc_lms_native_bc_root.der",
+    };
+    size_t i;
+    for (i = 0; i < sizeof(lmsFiles) / sizeof(lmsFiles[0]); i++) {
+        ExpectIntEQ(rfc9802_verify_one_cert(lmsFiles[i],
+            HSS_LMSk, CTC_HSS_LMS), TEST_SUCCESS);
+    }
+    ExpectIntEQ(rfc9802_lms_chain_verify(), TEST_SUCCESS);
+#endif /* !NO_FILESYSTEM && !NO_CERTS && !WOLFSSL_NO_LMS_SHA256_256 */
+    /* Pure wolfCrypt-level negative tests don't need filesystem or cert
+     * support, so they run for any LMS-enabled build. */
+    ExpectIntEQ(rfc9802_lms_import_negative(), TEST_SUCCESS);
+#endif
+    return EXPECT_RESULT();
+}
+
+int test_rfc9802_xmss_x509_verify(void)
+{
+    EXPECT_DECLS;
+#if defined(WOLFSSL_HAVE_XMSS)
+#if !defined(NO_FILESYSTEM) && !defined(NO_CERTS)
+    static const char* const xmssFiles[] = {
+        "./certs/xmss/bc_xmss_sha2_10_256_root.der",
+        "./certs/xmss/bc_xmss_sha2_16_256_root.der",
+    };
+    static const char* const xmssmtFiles[] = {
+        "./certs/xmss/bc_xmssmt_sha2_20_2_256_root.der",
+        "./certs/xmss/bc_xmssmt_sha2_20_4_256_root.der",
+        "./certs/xmss/bc_xmssmt_sha2_40_8_256_root.der",
+    };
+    size_t i;
+    for (i = 0; i < sizeof(xmssFiles) / sizeof(xmssFiles[0]); i++) {
+        ExpectIntEQ(rfc9802_verify_one_cert(xmssFiles[i],
+            XMSSk, CTC_XMSS), TEST_SUCCESS);
+    }
+    for (i = 0; i < sizeof(xmssmtFiles) / sizeof(xmssmtFiles[0]); i++) {
+        ExpectIntEQ(rfc9802_verify_one_cert(xmssmtFiles[i],
+            XMSSMTk, CTC_XMSSMT), TEST_SUCCESS);
+    }
+    ExpectIntEQ(rfc9802_xmss_sig_oid_mismatch(), TEST_SUCCESS);
+    ExpectIntEQ(rfc9802_xmss_chain_verify(), TEST_SUCCESS);
+#endif /* !NO_FILESYSTEM && !NO_CERTS */
+    /* Pure wolfCrypt-level negative tests don't need filesystem or cert
+     * support, so they run for any XMSS-enabled build. */
+    ExpectIntEQ(rfc9802_xmss_import_negative(), TEST_SUCCESS);
+#endif
+    return EXPECT_RESULT();
+}
diff --git a/tests/api/test_lms_xmss.h b/tests/api/test_lms_xmss.h
new file mode 100644
index 0000000000..b2ff579987
--- /dev/null
+++ b/tests/api/test_lms_xmss.h
@@ -0,0 +1,39 @@
+/* test_lms_xmss.h
+ *
+ * Copyright (C) 2006-2026 wolfSSL Inc.
+ *
+ * This file is part of wolfSSL.
+ *
+ * wolfSSL is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * wolfSSL is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1335, USA
+ */
+
+#ifndef WOLFCRYPT_TEST_LMS_XMSS_H
+#define WOLFCRYPT_TEST_LMS_XMSS_H
+
+#include <tests/api/api_decl.h>
+
+int test_wc_LmsKey_sign_verify(void);
+int test_wc_LmsKey_reload_cache(void);
+int test_rfc9802_lms_x509_verify(void);
+int test_rfc9802_xmss_x509_verify(void);
+
+/* LMS, and RFC 9802 (HSS/LMS and XMSS/XMSS^MT in X.509). */
+#define TEST_LMS_XMSS_DECLS                                                  \
+    TEST_DECL_GROUP("lms", test_wc_LmsKey_sign_verify),                 \
+    TEST_DECL_GROUP("lms", test_wc_LmsKey_reload_cache),                \
+    TEST_DECL_GROUP("lms", test_rfc9802_lms_x509_verify),               \
+    TEST_DECL_GROUP("xmss", test_rfc9802_xmss_x509_verify)
+
+#endif /* WOLFCRYPT_TEST_LMS_XMSS_H */
diff --git a/tests/api/test_mldsa.c b/tests/api/test_mldsa.c
index dc52a54453..5a91f220c3 100644
--- a/tests/api/test_mldsa.c
+++ b/tests/api/test_mldsa.c
@@ -30508,3 +30508,85 @@ WOLFSSL_MLDSA_API_CHECK_INLINE void wc_mldsa_canonical_api_check(void)
 PRAGMA_CLANG_DIAG_POP
 
 #endif /* WOLFSSL_HAVE_MLDSA */
+
+/*----------------------------------------------------------------------------*/
+/* ML-DSA / Dilithium negative length-validation tests                        */
+/*----------------------------------------------------------------------------*/
+
+/* ML-DSA HashML-DSA verify must reject hashLen > WC_MAX_DIGEST_SIZE */
+int test_mldsa_verify_hash(void)
+{
+    EXPECT_DECLS;
+#if defined(WOLFSSL_HAVE_MLDSA) && \
+    !defined(WOLFSSL_MLDSA_NO_MAKE_KEY) && \
+    !defined(WOLFSSL_MLDSA_NO_VERIFY)
+    wc_MlDsaKey key;
+    WC_RNG rng;
+    int res = 0;
+    byte sig[4000];
+    byte hash[4096]; /* larger than WC_MAX_DIGEST_SIZE */
+
+    XMEMSET(&key, 0, sizeof(key));
+    XMEMSET(&rng, 0, sizeof(rng));
+    XMEMSET(sig, 0x41, sizeof(sig));
+    XMEMSET(hash, 'A', sizeof(hash));
+
+    ExpectIntEQ(wc_InitRng(&rng), 0);
+    ExpectIntEQ(wc_MlDsaKey_Init(&key, NULL, INVALID_DEVID), 0);
+#ifndef WOLFSSL_NO_ML_DSA_65
+    ExpectIntEQ(wc_MlDsaKey_SetParams(&key, WC_ML_DSA_65), 0);
+#elif !defined(WOLFSSL_NO_ML_DSA_44)
+    ExpectIntEQ(wc_MlDsaKey_SetParams(&key, WC_ML_DSA_44), 0);
+#else
+    ExpectIntEQ(wc_MlDsaKey_SetParams(&key, WC_ML_DSA_87), 0);
+#endif
+    ExpectIntEQ(wc_MlDsaKey_MakeKey(&key, &rng), 0);
+
+    /* hashLen=4096 must be rejected, not overflow the stack */
+    ExpectIntEQ(wc_MlDsaKey_VerifyCtxHash(&key, sig, sizeof(sig), NULL, 0,
+        hash, sizeof(hash), WC_HASH_TYPE_SHA256, &res),
+        WC_NO_ERR_TRACE(BAD_LENGTH_E));
+
+    wc_MlDsaKey_Free(&key);
+    DoExpectIntEQ(wc_FreeRng(&rng), 0);
+#endif
+    return EXPECT_RESULT();
+}
+
+/* Dilithium verify_ctx_msg must reject absurdly large msgLen */
+int test_dilithium_hash(void)
+{
+    EXPECT_DECLS;
+#if defined(WOLFSSL_HAVE_MLDSA) && \
+    !defined(WOLFSSL_MLDSA_NO_MAKE_KEY) && \
+    !defined(WOLFSSL_MLDSA_NO_VERIFY)
+    wc_MlDsaKey key;
+    WC_RNG rng;
+    int res = 0;
+    byte sig[4000];
+    byte msg[64];
+
+    XMEMSET(&key, 0, sizeof(key));
+    XMEMSET(&rng, 0, sizeof(rng));
+    XMEMSET(sig, 0, sizeof(sig));
+    XMEMSET(msg, 'A', sizeof(msg));
+
+    ExpectIntEQ(wc_InitRng(&rng), 0);
+    ExpectIntEQ(wc_MlDsaKey_Init(&key, NULL, INVALID_DEVID), 0);
+#ifndef WOLFSSL_NO_ML_DSA_65
+    ExpectIntEQ(wc_MlDsaKey_SetParams(&key, WC_ML_DSA_65), 0);
+#elif !defined(WOLFSSL_NO_ML_DSA_44)
+    ExpectIntEQ(wc_MlDsaKey_SetParams(&key, WC_ML_DSA_44), 0);
+#else
+    ExpectIntEQ(wc_MlDsaKey_SetParams(&key, WC_ML_DSA_87), 0);
+#endif
+    ExpectIntEQ(wc_MlDsaKey_MakeKey(&key, &rng), 0);
+
+    ExpectIntEQ(wc_MlDsaKey_VerifyCtx(&key, sig, sizeof(sig), NULL, 0,
+        msg, 0xFFFFFFC0, &res), WC_NO_ERR_TRACE(BAD_FUNC_ARG));
+
+    wc_MlDsaKey_Free(&key);
+    DoExpectIntEQ(wc_FreeRng(&rng), 0);
+#endif
+    return EXPECT_RESULT();
+}
diff --git a/tests/api/test_mldsa.h b/tests/api/test_mldsa.h
index 82c844b54e..631ebabdc1 100644
--- a/tests/api/test_mldsa.h
+++ b/tests/api/test_mldsa.h
@@ -51,6 +51,11 @@ int test_mldsa_encode_w1_large_values(void);
 int test_mldsa_pkcs12(void);
 int test_mldsa_x509_pubkey_sigtype(void);
 
+/* Negative length-validation regression tests, also defined in
+ * tests/api/test_mldsa.c. */
+int test_mldsa_verify_hash(void);
+int test_dilithium_hash(void);
+
 /* Legacy-name shim coverage defined in tests/api/test_mldsa_legacy.c.
  * Single function -- compile-time wc_static_assert checks for every alias
  * + one runtime smoke test that drives each arg-reordering macro family.
@@ -82,6 +87,8 @@ int test_mldsa_legacy_shim(void);
     TEST_DECL_GROUP("mldsa", test_mldsa_encode_w1_large_values),               \
     TEST_DECL_GROUP("mldsa", test_mldsa_pkcs12),                               \
     TEST_DECL_GROUP("mldsa", test_mldsa_x509_pubkey_sigtype),                  \
+    TEST_DECL_GROUP("mldsa", test_mldsa_verify_hash),                          \
+    TEST_DECL_GROUP("mldsa", test_dilithium_hash),                             \
     TEST_DECL_GROUP("mldsa", test_mldsa_legacy_shim)
 
 #endif /* WOLFCRYPT_TEST_MLDSA_H */
diff --git a/tests/api/test_ocsp.c b/tests/api/test_ocsp.c
index bd4a0633bb..e9df5a6068 100644
--- a/tests/api/test_ocsp.c
+++ b/tests/api/test_ocsp.c
@@ -247,6 +247,15 @@ int test_ocsp_basic_verify(void)
     ExpectNull(
         response = wolfSSL_d2i_OCSP_RESPONSE(NULL, &ptr, sizeof(resp_bad)));
 
+    /* reuse failure must clear caller pointer */
+    ptr = (const unsigned char*)resp;
+    ExpectNotNull(
+        response = wolfSSL_d2i_OCSP_RESPONSE(&response, &ptr, sizeof(resp)));
+    ptr = (const unsigned char*)resp_bad;
+    ExpectNull(
+        wolfSSL_d2i_OCSP_RESPONSE(&response, &ptr, sizeof(resp_bad)));
+    ExpectNull(response);
+
     ptr = (const unsigned char*)resp;
     ExpectNotNull(
         response = wolfSSL_d2i_OCSP_RESPONSE(NULL, &ptr, sizeof(resp)));
@@ -765,6 +774,63 @@ int test_ocsp_certid_dup(void)
 }
 #endif
 
+int test_ocsp_resp_find_status_serial_prefix(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_OCSP) && defined(OPENSSL_EXTRA) && !defined(NO_SHA)
+    OcspResponse response;
+    OcspEntry    single;
+    CertStatus   responseStatus;
+    OcspEntry    requestedId;
+    CertStatus   requestedStatus;
+    int          status;
+
+    XMEMSET(&response, 0, sizeof(response));
+    XMEMSET(&single, 0, sizeof(single));
+    XMEMSET(&responseStatus, 0, sizeof(responseStatus));
+    XMEMSET(&requestedId, 0, sizeof(requestedId));
+    XMEMSET(&requestedStatus, 0, sizeof(requestedStatus));
+
+    single.status   = &responseStatus;
+    requestedId.status = &requestedStatus;
+    response.single  = &single;
+
+    /* Matching issuer name and key hashes on both sides. */
+    XMEMSET(single.issuerHash, 0x41, OCSP_DIGEST_SIZE);
+    XMEMSET(single.issuerKeyHash, 0x42, OCSP_DIGEST_SIZE);
+    XMEMCPY(requestedId.issuerHash, single.issuerHash, OCSP_DIGEST_SIZE);
+    XMEMCPY(requestedId.issuerKeyHash, single.issuerKeyHash, OCSP_DIGEST_SIZE);
+
+    /* Response carries a CERT_GOOD status for serial 01:02 (2 bytes). */
+    responseStatus.serial[0] = 0x01;
+    responseStatus.serial[1] = 0x02;
+    responseStatus.serialSz  = 2;
+    responseStatus.status    = CERT_GOOD;
+
+    /* Sanity check: an exact serial match must be found and report CERT_GOOD. */
+    requestedStatus.serial[0] = 0x01;
+    requestedStatus.serial[1] = 0x02;
+    requestedStatus.serialSz  = 2;
+    status = -1;
+    ExpectIntEQ(wolfSSL_OCSP_resp_find_status(&response, &requestedId, &status,
+                    NULL, NULL, NULL, NULL), WOLFSSL_SUCCESS);
+    ExpectIntEQ(status, CERT_GOOD);
+
+    /* Request serial 01:02:03 (3 bytes) shares the 01:02 prefix of the
+     * response serial.
+     * The lookup must not bind the good response to this longer and
+     * differing serial. */
+    requestedStatus.serial[0] = 0x01;
+    requestedStatus.serial[1] = 0x02;
+    requestedStatus.serial[2] = 0x03;
+    requestedStatus.serialSz  = 3;
+    status = -1;
+    ExpectIntEQ(wolfSSL_OCSP_resp_find_status(&response, &requestedId, &status,
+                    NULL, NULL, NULL, NULL), WOLFSSL_FAILURE);
+#endif
+    return EXPECT_RESULT();
+}
+
 #if defined(HAVE_OCSP) && defined(WOLFSSL_CERT_SETUP_CB) && \
     defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES) && !defined(NO_RSA) && \
     (defined(HAVE_CERTIFICATE_STATUS_REQUEST) || \
diff --git a/tests/api/test_ocsp.h b/tests/api/test_ocsp.h
index a139f56824..41dd21329d 100644
--- a/tests/api/test_ocsp.h
+++ b/tests/api/test_ocsp.h
@@ -24,6 +24,7 @@
 
 int test_ocsp_certid_enc_dec(void);
 int test_ocsp_certid_dup(void);
+int test_ocsp_resp_find_status_serial_prefix(void);
 int test_ocsp_status_callback(void);
 int test_ocsp_basic_verify(void);
 int test_ocsp_responder_keyhash_binding(void);
diff --git a/tests/api/test_ossl_pem.c b/tests/api/test_ossl_pem.c
index 547768854e..9715319ae6 100644
--- a/tests/api/test_ossl_pem.c
+++ b/tests/api/test_ossl_pem.c
@@ -755,6 +755,126 @@ int test_wolfSSL_PEM_PrivateKey(void)
     return EXPECT_RESULT();
 }
 
+int test_wolfSSL_PEM_write_PrivateKey(void)
+{
+    EXPECT_DECLS;
+#if defined(OPENSSL_EXTRA) && !defined(NO_CERTS) && !defined(NO_RSA) && \
+    !defined(NO_FILESYSTEM) && defined(USE_CERT_BUFFERS_2048) && \
+    !defined(NO_ASN) && !defined(NO_PWDBASED)
+    const char* privFile = "./test-pem-write-private-key.pem";
+    const unsigned char* serverKey =
+        (const unsigned char*)server_key_der_2048;
+    EVP_PKEY* pkey = NULL;
+    EVP_PKEY* readPriv = NULL;
+    XFILE fp = XBADFILE;
+
+    remove(privFile);
+
+    ExpectNotNull(wolfSSL_d2i_PrivateKey(EVP_PKEY_RSA, &pkey, &serverKey,
+        (long)sizeof_server_key_der_2048));
+
+    /* Bad-argument checks. */
+    ExpectIntEQ(PEM_write_PrivateKey(XBADFILE, pkey, NULL, NULL, 0, NULL,
+        NULL), 0);
+    ExpectIntEQ(PEM_write_PrivateKey(stderr, NULL, NULL, NULL, 0, NULL,
+        NULL), 0);
+
+    /* Write private key to file. */
+    ExpectTrue((fp = XFOPEN(privFile, "wb")) != XBADFILE);
+    if (fp != XBADFILE) {
+        ExpectIntEQ(PEM_write_PrivateKey(fp, pkey, NULL, NULL, 0, NULL, NULL),
+            1);
+        XFCLOSE(fp);
+        fp = XBADFILE;
+    }
+
+    /* Read it back and verify the DER content matches. */
+    ExpectTrue((fp = XFOPEN(privFile, "rb")) != XBADFILE);
+    if (fp != XBADFILE) {
+        ExpectNotNull(readPriv = PEM_read_PrivateKey(fp, NULL, NULL, NULL));
+        XFCLOSE(fp);
+        fp = XBADFILE;
+    }
+    if ((pkey != NULL) && (readPriv != NULL) && (pkey->pkey.ptr != NULL) &&
+            (readPriv->pkey.ptr != NULL)) {
+        ExpectIntEQ(pkey->pkey_sz, readPriv->pkey_sz);
+        ExpectIntEQ(XMEMCMP(pkey->pkey.ptr, readPriv->pkey.ptr,
+            pkey->pkey_sz), 0);
+    }
+
+    EVP_PKEY_free(readPriv);
+    EVP_PKEY_free(pkey);
+    if (fp != XBADFILE) {
+        XFCLOSE(fp);
+    }
+    remove(privFile);
+#endif
+    return EXPECT_RESULT();
+}
+
+int test_wolfSSL_PEM_write_PUBKEY(void)
+{
+    EXPECT_DECLS;
+#if defined(OPENSSL_EXTRA) && !defined(NO_CERTS) && !defined(NO_RSA) && \
+    !defined(NO_FILESYSTEM) && defined(USE_CERT_BUFFERS_2048) && \
+    !defined(NO_ASN) && !defined(NO_PWDBASED)
+    const char* pubFile = "./test-pem-write-pubkey.pem";
+    const unsigned char* serverKey =
+        (const unsigned char*)server_key_der_2048;
+    EVP_PKEY* pkey = NULL;
+    EVP_PKEY* readPub = NULL;
+    unsigned char* pubDer = NULL;
+    unsigned char* readPubDer = NULL;
+    XFILE fp = XBADFILE;
+    int pubDerSz = 0;
+    int readPubDerSz = 0;
+
+    remove(pubFile);
+
+    ExpectNotNull(wolfSSL_d2i_PrivateKey(EVP_PKEY_RSA, &pkey, &serverKey,
+        (long)sizeof_server_key_der_2048));
+
+    /* Bad-argument checks. */
+    ExpectIntEQ(PEM_write_PUBKEY(XBADFILE, pkey), 0);
+    ExpectIntEQ(PEM_write_PUBKEY(stderr, NULL), 0);
+
+    /* Capture the expected public-key DER for later comparison. */
+    ExpectIntGT(pubDerSz = wolfSSL_i2d_PUBKEY(pkey, &pubDer), 0);
+
+    /* Write public key to file. */
+    ExpectTrue((fp = XFOPEN(pubFile, "wb")) != XBADFILE);
+    if (fp != XBADFILE) {
+        ExpectIntEQ(PEM_write_PUBKEY(fp, pkey), 1);
+        XFCLOSE(fp);
+        fp = XBADFILE;
+    }
+
+    /* Read it back and verify the DER content matches. */
+    ExpectTrue((fp = XFOPEN(pubFile, "rb")) != XBADFILE);
+    if (fp != XBADFILE) {
+        ExpectNotNull(readPub = PEM_read_PUBKEY(fp, NULL, NULL, NULL));
+        XFCLOSE(fp);
+        fp = XBADFILE;
+    }
+    ExpectIntGT(readPubDerSz = wolfSSL_i2d_PUBKEY(readPub, &readPubDer), 0);
+    ExpectIntEQ(pubDerSz, readPubDerSz);
+    if ((pubDer != NULL) && (readPubDer != NULL) && (pubDerSz > 0) &&
+            (pubDerSz == readPubDerSz)) {
+        ExpectIntEQ(XMEMCMP(pubDer, readPubDer, pubDerSz), 0);
+    }
+
+    XFREE(readPubDer, NULL, DYNAMIC_TYPE_PUBLIC_KEY);
+    XFREE(pubDer, NULL, DYNAMIC_TYPE_PUBLIC_KEY);
+    EVP_PKEY_free(readPub);
+    EVP_PKEY_free(pkey);
+    if (fp != XBADFILE) {
+        XFCLOSE(fp);
+    }
+    remove(pubFile);
+#endif
+    return EXPECT_RESULT();
+}
+
 int test_wolfSSL_PEM_file_RSAKey(void)
 {
     EXPECT_DECLS;
diff --git a/tests/api/test_ossl_pem.h b/tests/api/test_ossl_pem.h
index 55480d218e..3de900a6fa 100644
--- a/tests/api/test_ossl_pem.h
+++ b/tests/api/test_ossl_pem.h
@@ -32,6 +32,8 @@ int test_wolfSSL_PEM_PrivateKey_ecc(void);
 int test_wolfSSL_PEM_PrivateKey_dsa(void);
 int test_wolfSSL_PEM_PrivateKey_dh(void);
 int test_wolfSSL_PEM_PrivateKey(void);
+int test_wolfSSL_PEM_write_PrivateKey(void);
+int test_wolfSSL_PEM_write_PUBKEY(void);
 int test_wolfSSL_PEM_file_RSAKey(void);
 int test_wolfSSL_PEM_file_RSAPrivateKey(void);
 int test_wolfSSL_PEM_read_RSA_PUBKEY(void);
@@ -52,6 +54,8 @@ int test_wolfSSL_PEM_PUBKEY(void);
     TEST_DECL_GROUP("ossl_pem", test_wolfSSL_PEM_PrivateKey_dsa),       \
     TEST_DECL_GROUP("ossl_pem", test_wolfSSL_PEM_PrivateKey_dh),        \
     TEST_DECL_GROUP("ossl_pem", test_wolfSSL_PEM_PrivateKey),           \
+    TEST_DECL_GROUP("ossl_pem", test_wolfSSL_PEM_write_PrivateKey),     \
+    TEST_DECL_GROUP("ossl_pem", test_wolfSSL_PEM_write_PUBKEY),         \
     TEST_DECL_GROUP("ossl_pem", test_wolfSSL_PEM_file_RSAKey),          \
     TEST_DECL_GROUP("ossl_pem", test_wolfSSL_PEM_file_RSAPrivateKey),   \
     TEST_DECL_GROUP("ossl_pem", test_wolfSSL_PEM_read_RSA_PUBKEY),      \
diff --git a/tests/api/test_ossl_x509_crypto.c b/tests/api/test_ossl_x509_crypto.c
index 01f5735480..4c84d40617 100644
--- a/tests/api/test_ossl_x509_crypto.c
+++ b/tests/api/test_ossl_x509_crypto.c
@@ -71,6 +71,114 @@ int test_wolfSSL_X509_check_private_key(void)
     return EXPECT_RESULT();
 }
 
+/* EVP_PKCS82PKEY() must populate pkey.ptr/pkey_sz for ML-DSA so
+ * X509_check_private_key() (wc_CheckPrivateKey) can redecode the DER, and
+ * d2i_PKCS8_PKEY() must keep the full PKCS#8 wrapper for ML-DSA level recovery
+ * from the AlgorithmIdentifier. */
+int test_wolfSSL_X509_check_private_key_mldsa(void)
+{
+    EXPECT_DECLS;
+#if defined(OPENSSL_EXTRA) && !defined(NO_FILESYSTEM) && \
+    !defined(NO_BIO) && !defined(NO_CHECK_PRIVATE_KEY) && \
+    defined(HAVE_DILITHIUM) && !defined(WOLFSSL_DILITHIUM_NO_SIGN) && \
+    !defined(WOLFSSL_DILITHIUM_NO_VERIFY) && \
+    (defined(OPENSSL_ALL) || defined(WOLFSSL_WPAS_SMALL)) && \
+    (!defined(WOLFSSL_NO_ML_DSA_44) || !defined(WOLFSSL_NO_ML_DSA_65) || \
+     !defined(WOLFSSL_NO_ML_DSA_87))
+    static const struct {
+        const char* keyPath;
+        const char* certPath;
+        const char* mismatchCertPath;  /* NULL if no other level available */
+    } cases[] = {
+    #if !defined(WOLFSSL_NO_ML_DSA_44)
+        { "./certs/mldsa/mldsa44-key.pem",
+          "./certs/mldsa/mldsa44-cert.der",
+        #if !defined(WOLFSSL_NO_ML_DSA_65)
+          "./certs/mldsa/mldsa65-cert.der"
+        #elif !defined(WOLFSSL_NO_ML_DSA_87)
+          "./certs/mldsa/mldsa87-cert.der"
+        #else
+          NULL
+        #endif
+        },
+    #endif
+    #if !defined(WOLFSSL_NO_ML_DSA_65)
+        { "./certs/mldsa/mldsa65-key.pem",
+          "./certs/mldsa/mldsa65-cert.der",
+        #if !defined(WOLFSSL_NO_ML_DSA_87)
+          "./certs/mldsa/mldsa87-cert.der"
+        #elif !defined(WOLFSSL_NO_ML_DSA_44)
+          "./certs/mldsa/mldsa44-cert.der"
+        #else
+          NULL
+        #endif
+        },
+    #endif
+    #if !defined(WOLFSSL_NO_ML_DSA_87)
+        { "./certs/mldsa/mldsa87-key.pem",
+          "./certs/mldsa/mldsa87-cert.der",
+        #if !defined(WOLFSSL_NO_ML_DSA_44)
+          "./certs/mldsa/mldsa44-cert.der"
+        #elif !defined(WOLFSSL_NO_ML_DSA_65)
+          "./certs/mldsa/mldsa65-cert.der"
+        #else
+          NULL
+        #endif
+        },
+    #endif
+    };
+    size_t i;
+
+    for (i = 0; i < sizeof(cases) / sizeof(cases[0]); i++) {
+        PKCS8_PRIV_KEY_INFO* pt = NULL;
+        EVP_PKEY* pkey = NULL;
+        X509* x509 = NULL;
+        X509* mismatchX509 = NULL;
+        BIO* bio = NULL;
+        byte* buf = NULL;
+        size_t sz = 0;
+
+        ExpectIntEQ(load_file(cases[i].keyPath, &buf, &sz), 0);
+
+        ExpectNotNull(bio = BIO_new_mem_buf((void*)buf, (int)sz));
+        ExpectNotNull(pt = d2i_PKCS8_PRIV_KEY_INFO_bio(bio, NULL));
+
+        ExpectNotNull(pkey = EVP_PKCS82PKEY(pt));
+        if (pkey != NULL) {
+            ExpectIntEQ(EVP_PKEY_id(pkey), EVP_PKEY_DILITHIUM);
+            /* pkey.ptr must hold the DER so that X509_check_private_key() to
+             * wc_CheckPrivateKey() can re-decode it. */
+            ExpectNotNull(pkey->pkey.ptr);
+            ExpectIntGT(pkey->pkey_sz, 0);
+        }
+
+        ExpectNotNull(x509 = X509_load_certificate_file(
+            cases[i].certPath, SSL_FILETYPE_ASN1));
+        ExpectIntEQ(X509_check_private_key(x509, pkey), 1);
+
+        if (cases[i].mismatchCertPath != NULL) {
+            ExpectNotNull(mismatchX509 = X509_load_certificate_file(
+                cases[i].mismatchCertPath, SSL_FILETYPE_ASN1));
+            ExpectIntEQ(X509_check_private_key(mismatchX509, pkey), 0);
+        }
+
+        /* Negative check, corrupt the outer SEQ tag so the key DER fails */
+        if (EXPECT_SUCCESS() && (pkey != NULL) && (pkey->pkey.ptr != NULL)) {
+            pkey->pkey.ptr[0] ^= 0xFF;
+            ExpectIntEQ(X509_check_private_key(x509, pkey), 0);
+        }
+
+        X509_free(mismatchX509);
+        X509_free(x509);
+        EVP_PKEY_free(pkey);
+        PKCS8_PRIV_KEY_INFO_free(pt);
+        BIO_free(bio);
+        XFREE(buf, NULL, DYNAMIC_TYPE_TMP_BUFFER);
+    }
+#endif
+    return EXPECT_RESULT();
+}
+
 int test_wolfSSL_X509_verify(void)
 {
     EXPECT_DECLS;
diff --git a/tests/api/test_ossl_x509_crypto.h b/tests/api/test_ossl_x509_crypto.h
index fe09acf6d1..cf40bca9a3 100644
--- a/tests/api/test_ossl_x509_crypto.h
+++ b/tests/api/test_ossl_x509_crypto.h
@@ -25,6 +25,7 @@
 #include <tests/api/api_decl.h>
 
 int test_wolfSSL_X509_check_private_key(void);
+int test_wolfSSL_X509_check_private_key_mldsa(void);
 int test_wolfSSL_X509_verify(void);
 int test_wolfSSL_X509_sign(void);
 int test_wolfSSL_X509_sign2(void);
@@ -32,6 +33,8 @@ int test_wolfSSL_make_cert(void);
 
 #define TEST_OSSL_X509_CRYPTO_DECLS                                            \
     TEST_DECL_GROUP("ossl_x509_crypto", test_wolfSSL_X509_check_private_key),  \
+    TEST_DECL_GROUP("ossl_x509_crypto",                                        \
+        test_wolfSSL_X509_check_private_key_mldsa),                            \
     TEST_DECL_GROUP("ossl_x509_crypto", test_wolfSSL_X509_verify),             \
     TEST_DECL_GROUP("ossl_x509_crypto", test_wolfSSL_X509_sign),               \
     TEST_DECL_GROUP("ossl_x509_crypto", test_wolfSSL_X509_sign2),              \
diff --git a/tests/api/test_ossl_x509_ext.c b/tests/api/test_ossl_x509_ext.c
index dfc2451529..fb0967be5e 100644
--- a/tests/api/test_ossl_x509_ext.c
+++ b/tests/api/test_ossl_x509_ext.c
@@ -593,6 +593,70 @@ int test_wolfSSL_X509_add_ext(void)
     return EXPECT_RESULT();
 }
 
+int test_wolfSSL_X509_add_ext_dirname_san_rejected(void)
+{
+    EXPECT_DECLS;
+#if defined(OPENSSL_ALL) && !defined(NO_RSA)
+    WOLFSSL_X509* x509 = NULL;
+    WOLFSSL_X509_EXTENSION* ext = NULL;
+    WOLFSSL_ASN1_OBJECT* obj = NULL;
+    WOLFSSL_GENERAL_NAME* gn = NULL;
+    WOLFSSL_X509_NAME* dirName = NULL;
+    WOLFSSL_STACK* sk = NULL;
+
+    ExpectNotNull(x509 = wolfSSL_X509_new());
+    ExpectNotNull(ext = wolfSSL_X509_EXTENSION_new());
+
+    /* Build a GEN_DIRNAME GENERAL_NAME with a real directoryName so that
+     * gn->d.directoryName aliases an X509_NAME object via the union. */
+    ExpectNotNull(gn = wolfSSL_GENERAL_NAME_new());
+    ExpectNotNull(dirName = wolfSSL_X509_NAME_new());
+    ExpectIntEQ(wolfSSL_X509_NAME_add_entry_by_NID(dirName, NID_commonName,
+        MBSTRING_UTF8, (unsigned char*)"dirname-san-test", -1, -1, 0), 1);
+    if (gn != NULL) {
+        /* Drop the default IA5 string and install the X509_NAME. */
+        wolfSSL_ASN1_STRING_free(gn->d.ia5);
+        gn->type = GEN_DIRNAME;
+        gn->d.directoryName = dirName;
+        dirName = NULL; /* gn owns the X509_NAME now */
+    }
+
+    /* Build the ext: SAN OID + ext_sk containing the DirName GENERAL_NAME. */
+    ExpectNotNull(sk = wolfSSL_sk_new_null());
+    if (sk != NULL) {
+        sk->type = STACK_TYPE_GEN_NAME;
+    }
+    ExpectIntGT(wolfSSL_sk_GENERAL_NAME_push(sk, gn), 0);
+    gn = NULL; /* sk owns gn now */
+
+    ExpectNotNull(obj = wolfSSL_OBJ_nid2obj(NID_subject_alt_name));
+    if (obj != NULL) {
+        obj->type = NID_subject_alt_name;
+        obj->nid  = NID_subject_alt_name;
+    }
+    if ((ext != NULL) && (obj != NULL) && (sk != NULL)) {
+        ext->obj = obj;
+        obj = NULL;       /* ext owns obj now */
+        ext->ext_sk = sk;
+        sk = NULL;        /* ext owns sk now */
+    }
+
+    /* The unsupported GeneralName type must be rejected safely, NOT crash
+     * or read OOB via a type-confused d.ia5 dereference. */
+    ExpectIntEQ(wolfSSL_X509_add_ext(x509, ext, -1),
+        WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+
+    /* Cleanup. The success-path owners (set to NULL above) are no-ops. */
+    wolfSSL_ASN1_OBJECT_free(obj);
+    wolfSSL_sk_GENERAL_NAME_pop_free(sk, wolfSSL_GENERAL_NAME_free);
+    wolfSSL_GENERAL_NAME_free(gn);
+    wolfSSL_X509_NAME_free(dirName);
+    wolfSSL_X509_EXTENSION_free(ext);
+    wolfSSL_X509_free(x509);
+#endif
+    return EXPECT_RESULT();
+}
+
 int test_wolfSSL_X509_get_ext_count(void)
 {
     EXPECT_DECLS;
@@ -1098,10 +1162,6 @@ int test_wolfSSL_X509V3_EXT_bc(void)
 
     ExpectNotNull(ext = wolfSSL_X509_EXTENSION_new());
     ExpectNotNull(obj = wolfSSL_ASN1_OBJECT_new());
-    ExpectNotNull(pathLen = wolfSSL_ASN1_INTEGER_new());
-    if (pathLen != NULL) {
-        pathLen->length = 2;
-    }
 
     if (obj != NULL) {
         obj->type = NID_basic_constraints;
@@ -1109,17 +1169,47 @@ int test_wolfSSL_X509V3_EXT_bc(void)
     }
     ExpectIntEQ(wolfSSL_X509_EXTENSION_set_object(ext, obj), WOLFSSL_SUCCESS);
     ExpectNotNull(wolfSSL_X509V3_EXT_get(ext));
-    /* No pathlen set. */
+
+    /* No pathLenConstraint present. Per RFC 5280 4.2.1.9 no limit is imposed,
+     * so pathlen must be NULL (and distinguishable from a value of 0). */
     ExpectNotNull(bc = (WOLFSSL_BASIC_CONSTRAINTS*)wolfSSL_X509V3_EXT_d2i(ext));
+    ExpectNull(bc->pathlen);
     wolfSSL_BASIC_CONSTRAINTS_free(bc);
     bc = NULL;
 
+    /* pathLenConstraint of 0 is valid and meaningful (the CA may only issue
+     * end-entity certificates). It must be preserved, not conflated with an
+     * absent constraint. */
+    ExpectNotNull(pathLen = wolfSSL_ASN1_INTEGER_new());
+    if (pathLen != NULL) {
+        pathLen->length = 0;
+    }
     if ((ext != NULL) && (ext->obj != NULL)) {
         ext->obj->pathlen = pathLen;
         pathLen = NULL;
     }
-    /* pathlen set. */
     ExpectNotNull(bc = (WOLFSSL_BASIC_CONSTRAINTS*)wolfSSL_X509V3_EXT_d2i(ext));
+    ExpectNotNull(bc->pathlen);
+    ExpectIntEQ(bc->pathlen->length, 0);
+    wolfSSL_BASIC_CONSTRAINTS_free(bc);
+    bc = NULL;
+
+    /* A non-zero pathLenConstraint is preserved as-is. */
+    if ((ext != NULL) && (ext->obj != NULL)) {
+        wolfSSL_ASN1_INTEGER_free(ext->obj->pathlen);
+        ext->obj->pathlen = NULL;
+    }
+    ExpectNotNull(pathLen = wolfSSL_ASN1_INTEGER_new());
+    if (pathLen != NULL) {
+        pathLen->length = 2;
+    }
+    if ((ext != NULL) && (ext->obj != NULL)) {
+        ext->obj->pathlen = pathLen;
+        pathLen = NULL;
+    }
+    ExpectNotNull(bc = (WOLFSSL_BASIC_CONSTRAINTS*)wolfSSL_X509V3_EXT_d2i(ext));
+    ExpectNotNull(bc->pathlen);
+    ExpectIntEQ(bc->pathlen->length, 2);
 
     wolfSSL_ASN1_INTEGER_free(pathLen);
     wolfSSL_BASIC_CONSTRAINTS_free(bc);
@@ -1129,6 +1219,72 @@ int test_wolfSSL_X509V3_EXT_bc(void)
     return EXPECT_RESULT();
 }
 
+int test_wolfSSL_X509_get_ext_d2i_basic_constraints(void)
+{
+    EXPECT_DECLS;
+#if !defined(NO_FILESYSTEM) && defined(OPENSSL_ALL) && !defined(NO_RSA)
+    XFILE f = XBADFILE;
+    WOLFSSL_X509* x509 = NULL;
+    WOLFSSL_BASIC_CONSTRAINTS* bc = NULL;
+    int crit = 0;
+
+    /* CA certificate with basicConstraints CA:TRUE and *no* pathLenConstraint.
+     * Per RFC 5280 4.2.1.9 no path length limit is imposed, so the returned
+     * pathlen must be NULL - it must not be reported as a value of 0. */
+    ExpectTrue((f = XFOPEN("./certs/ca-cert.pem", "rb")) != XBADFILE);
+    ExpectNotNull(x509 = wolfSSL_PEM_read_X509(f, NULL, NULL, NULL));
+    if (f != XBADFILE) {
+        XFCLOSE(f);
+        f = XBADFILE;
+    }
+    ExpectNotNull(bc = (WOLFSSL_BASIC_CONSTRAINTS*)wolfSSL_X509_get_ext_d2i(
+        x509, NID_basic_constraints, &crit, NULL));
+    ExpectNull(bc->pathlen);
+    wolfSSL_BASIC_CONSTRAINTS_free(bc);
+    bc = NULL;
+    wolfSSL_X509_free(x509);
+    x509 = NULL;
+
+    /* Intermediate CA with basicConstraints CA:TRUE, pathlen:1. */
+    ExpectTrue((f = XFOPEN("./certs/intermediate/ca-int-cert.pem", "rb")) !=
+        XBADFILE);
+    ExpectNotNull(x509 = wolfSSL_PEM_read_X509(f, NULL, NULL, NULL));
+    if (f != XBADFILE) {
+        XFCLOSE(f);
+        f = XBADFILE;
+    }
+    ExpectNotNull(bc = (WOLFSSL_BASIC_CONSTRAINTS*)wolfSSL_X509_get_ext_d2i(
+        x509, NID_basic_constraints, &crit, NULL));
+    ExpectNotNull(bc->pathlen);
+    ExpectIntEQ(bc->pathlen->length, 1);
+    wolfSSL_BASIC_CONSTRAINTS_free(bc);
+    bc = NULL;
+    wolfSSL_X509_free(x509);
+    x509 = NULL;
+
+    /* CA with basicConstraints CA:TRUE, pathlen:0. A pathLenConstraint of 0 is
+     * valid and meaningful (the CA may only issue end-entity certificates) and
+     * must be reported (non-NULL pathlen, value 0) - it must not be conflated
+     * with an absent constraint. */
+    ExpectTrue((f = XFOPEN("./certs/test-pathlen/chainG-ICA1-pathlen0.pem",
+        "rb")) != XBADFILE);
+    ExpectNotNull(x509 = wolfSSL_PEM_read_X509(f, NULL, NULL, NULL));
+    if (f != XBADFILE) {
+        XFCLOSE(f);
+        f = XBADFILE;
+    }
+    ExpectNotNull(bc = (WOLFSSL_BASIC_CONSTRAINTS*)wolfSSL_X509_get_ext_d2i(
+        x509, NID_basic_constraints, &crit, NULL));
+    ExpectNotNull(bc->pathlen);
+    ExpectIntEQ(bc->pathlen->length, 0);
+    wolfSSL_BASIC_CONSTRAINTS_free(bc);
+    bc = NULL;
+    wolfSSL_X509_free(x509);
+    x509 = NULL;
+#endif
+    return EXPECT_RESULT();
+}
+
 int test_wolfSSL_X509V3_EXT_san(void)
 {
     EXPECT_DECLS;
diff --git a/tests/api/test_ossl_x509_ext.h b/tests/api/test_ossl_x509_ext.h
index ee749facd9..68be4ebf5e 100644
--- a/tests/api/test_ossl_x509_ext.h
+++ b/tests/api/test_ossl_x509_ext.h
@@ -30,6 +30,7 @@ int test_wolfSSL_X509_get_ext_by_NID(void);
 int test_wolfSSL_X509_get_ext_subj_alt_name(void);
 int test_wolfSSL_X509_set_ext(void);
 int test_wolfSSL_X509_add_ext(void);
+int test_wolfSSL_X509_add_ext_dirname_san_rejected(void);
 int test_wolfSSL_X509_get_ext_count(void);
 int test_wolfSSL_X509_stack_extensions(void);
 int test_wolfSSL_X509_EXTENSION_new(void);
@@ -42,6 +43,7 @@ int test_wolfSSL_X509V3_set_ctx(void);
 int test_wolfSSL_X509V3_EXT_get(void);
 int test_wolfSSL_X509V3_EXT_nconf(void);
 int test_wolfSSL_X509V3_EXT_bc(void);
+int test_wolfSSL_X509_get_ext_d2i_basic_constraints(void);
 int test_wolfSSL_X509V3_EXT_san(void);
 int test_wolfSSL_X509V3_EXT_aia(void);
 int test_wolfSSL_X509V3_EXT(void);
@@ -62,6 +64,8 @@ int test_wolfSSL_NAME_CONSTRAINTS_excluded(void);
     TEST_DECL_GROUP("ossl_x509_ext", test_wolfSSL_X509_get_ext_subj_alt_name), \
     TEST_DECL_GROUP("ossl_x509_ext", test_wolfSSL_X509_set_ext),               \
     TEST_DECL_GROUP("ossl_x509_ext", test_wolfSSL_X509_add_ext),               \
+    TEST_DECL_GROUP("ossl_x509_ext",                                           \
+                      test_wolfSSL_X509_add_ext_dirname_san_rejected),         \
     TEST_DECL_GROUP("ossl_x509_ext", test_wolfSSL_X509_get_ext_count),         \
     TEST_DECL_GROUP("ossl_x509_ext", test_wolfSSL_X509_stack_extensions),      \
     TEST_DECL_GROUP("ossl_x509_ext", test_wolfSSL_X509_EXTENSION_new),         \
@@ -76,6 +80,8 @@ int test_wolfSSL_NAME_CONSTRAINTS_excluded(void);
     TEST_DECL_GROUP("ossl_x509_ext", test_wolfSSL_X509V3_EXT_get),             \
     TEST_DECL_GROUP("ossl_x509_ext", test_wolfSSL_X509V3_EXT_nconf),           \
     TEST_DECL_GROUP("ossl_x509_ext", test_wolfSSL_X509V3_EXT_bc),              \
+    TEST_DECL_GROUP("ossl_x509_ext",                                           \
+                          test_wolfSSL_X509_get_ext_d2i_basic_constraints),    \
     TEST_DECL_GROUP("ossl_x509_ext", test_wolfSSL_X509V3_EXT_san),             \
     TEST_DECL_GROUP("ossl_x509_ext", test_wolfSSL_X509V3_EXT_aia),             \
     TEST_DECL_GROUP("ossl_x509_ext", test_wolfSSL_X509V3_EXT),                 \
diff --git a/tests/api/test_pkcs12.c b/tests/api/test_pkcs12.c
index 8b9e4863e1..cbe2186e8e 100644
--- a/tests/api/test_pkcs12.c
+++ b/tests/api/test_pkcs12.c
@@ -648,10 +648,12 @@ int test_wc_PKCS12_PBKDF(void)
     ExpectIntEQ(XMEMCMP(derived, verify2, 24), 0);
 
     /* iterations <= 0 must be rejected */
+    #if !defined(HAVE_FIPS) || FIPS_VERSION3_GE(7,0,0)
     ExpectIntEQ(wc_PKCS12_PBKDF(derived, passwd, (int)sizeof(passwd),
                     salt, (int)sizeof(salt), 0, 24, WC_SHA256, 1), BAD_FUNC_ARG);
     ExpectIntEQ(wc_PKCS12_PBKDF(derived, passwd, (int)sizeof(passwd),
                     salt, (int)sizeof(salt), -1, 24, WC_SHA256, 1), BAD_FUNC_ARG);
+    #endif /* !HAVE_FIPS || FIPS_VERSION3_GE(7,0,0) */
 #endif
     return EXPECT_RESULT();
 }
@@ -717,12 +719,14 @@ int test_wc_PKCS12_PBKDF_ex(void)
                     salt, (int)sizeof(salt), 1, 24, WC_SHA256, 3, NULL), 0);
 
     /* iterations <= 0 must be rejected */
+    #if !defined(HAVE_FIPS) || FIPS_VERSION3_GE(7,0,0)
     ExpectIntEQ(wc_PKCS12_PBKDF_ex(derived, passwd, (int)sizeof(passwd),
                     salt, (int)sizeof(salt), 0, 24, WC_SHA256, 1, NULL),
                 BAD_FUNC_ARG);
     ExpectIntEQ(wc_PKCS12_PBKDF_ex(derived, passwd, (int)sizeof(passwd),
                     salt, (int)sizeof(salt), -1, 24, WC_SHA256, 1, NULL),
                 BAD_FUNC_ARG);
+    #endif /* !HAVE_FIPS || FIPS_VERSION3_GE(7,0,0) */
 #endif
     return EXPECT_RESULT();
 }
diff --git a/tests/api/test_pwdbased.c b/tests/api/test_pwdbased.c
index 7facfd2005..92a144c42c 100644
--- a/tests/api/test_pwdbased.c
+++ b/tests/api/test_pwdbased.c
@@ -29,7 +29,8 @@ int test_wc_PBKDF1_ex_iterations(void)
 {
     EXPECT_DECLS;
 #if defined(HAVE_PBKDF1) && !defined(NO_PWDBASED) && !defined(NO_SHA) && \
-    !defined(HAVE_SELFTEST)
+    !defined(HAVE_SELFTEST) && (!defined(HAVE_FIPS) || FIPS_VERSION3_GE(7,0,0))
+
     static const byte passwd[] = { 'p', 'a', 's', 's' };
     static const byte salt[]   = { 0x78, 0x57, 0x8E, 0x5a,
                                    0x5d, 0x63, 0xcb, 0x06 };
diff --git a/tests/api/test_session.c b/tests/api/test_session.c
new file mode 100644
index 0000000000..e7557ff007
--- /dev/null
+++ b/tests/api/test_session.c
@@ -0,0 +1,1616 @@
+/* test_session.c
+ *
+ * Copyright (C) 2006-2026 wolfSSL Inc.
+ *
+ * This file is part of wolfSSL.
+ *
+ * wolfSSL is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * wolfSSL is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1335, USA
+ */
+
+#include <tests/unit.h>
+
+#ifdef NO_INLINE
+    #include <wolfssl/wolfcrypt/misc.h>
+#else
+    #define WOLFSSL_MISC_INCLUDED
+    #include <wolfcrypt/src/misc.c>
+#endif
+
+#include <wolfssl/ssl.h>
+#include <wolfssl/internal.h>
+#include <wolfssl/error-ssl.h>
+#include <tests/api/api.h>
+#include <tests/utils.h>
+#include <tests/api/test_session.h>
+
+/*----------------------------------------------------------------------------*/
+/* WOLFSSL_CTX_add_session / session resumption                               */
+/*----------------------------------------------------------------------------*/
+
+#if defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_EXT_CACHE) && \
+    !defined(SINGLE_THREADED) && defined(WOLFSSL_TLS13) && \
+    !defined(NO_SESSION_CACHE)
+
+/* Sessions to restore/store */
+static WOLFSSL_SESSION* test_wolfSSL_CTX_add_session_client_sess;
+static WOLFSSL_SESSION* test_wolfSSL_CTX_add_session_server_sess;
+static WOLFSSL_CTX*     test_wolfSSL_CTX_add_session_server_ctx;
+
+static void test_wolfSSL_CTX_add_session_ctx_ready(WOLFSSL_CTX* ctx)
+{
+    /* Don't store sessions. Lookup is still enabled. */
+    AssertIntEQ(wolfSSL_CTX_set_session_cache_mode(ctx,
+            WOLFSSL_SESS_CACHE_NO_INTERNAL_STORE), WOLFSSL_SUCCESS);
+#ifdef OPENSSL_EXTRA
+    AssertIntEQ(wolfSSL_CTX_get_session_cache_mode(ctx) &
+            WOLFSSL_SESS_CACHE_NO_INTERNAL_STORE,
+            WOLFSSL_SESS_CACHE_NO_INTERNAL_STORE);
+#endif
+    /* Require both peers to provide certs */
+    wolfSSL_CTX_set_verify(ctx, WOLFSSL_VERIFY_PEER, NULL);
+}
+
+static void test_wolfSSL_CTX_add_session_on_result(WOLFSSL* ssl)
+{
+    WOLFSSL_SESSION** sess;
+#ifdef WOLFSSL_MUTEX_INITIALIZER
+    static wolfSSL_Mutex m = WOLFSSL_MUTEX_INITIALIZER(m);
+
+    (void)wc_LockMutex(&m);
+#endif
+    if (wolfSSL_is_server(ssl))
+        sess = &test_wolfSSL_CTX_add_session_server_sess;
+    else
+        sess = &test_wolfSSL_CTX_add_session_client_sess;
+    if (*sess == NULL) {
+#ifdef NO_SESSION_CACHE_REF
+        *sess = wolfSSL_get1_session(ssl);
+        AssertNotNull(*sess);
+#else
+        /* Test for backwards compatibility */
+        if (wolfSSL_is_server(ssl)) {
+            *sess = wolfSSL_get1_session(ssl);
+            AssertNotNull(*sess);
+        }
+        else {
+            *sess = wolfSSL_get_session(ssl);
+            AssertNotNull(*sess);
+        }
+#endif
+        /* Now save the session in the internal store to make it available
+         * for lookup. For TLS 1.3, we can't save the session without
+         * WOLFSSL_TICKET_HAVE_ID because there is no way to retrieve the
+         * session from cache. */
+        if (wolfSSL_is_server(ssl)
+#ifndef WOLFSSL_TICKET_HAVE_ID
+                && wolfSSL_version(ssl) != TLS1_3_VERSION
+#endif
+                )
+            AssertIntEQ(wolfSSL_CTX_add_session(wolfSSL_get_SSL_CTX(ssl),
+                    *sess), WOLFSSL_SUCCESS);
+    }
+    else {
+        /* If we have a session retrieved then remaining connections should be
+         * resuming on that session */
+        AssertIntEQ(wolfSSL_session_reused(ssl), 1);
+    }
+#ifdef WOLFSSL_MUTEX_INITIALIZER
+    wc_UnLockMutex(&m);
+#endif
+
+    /* Save CTX to be able to decrypt tickets */
+    if (wolfSSL_is_server(ssl) &&
+            test_wolfSSL_CTX_add_session_server_ctx == NULL) {
+        test_wolfSSL_CTX_add_session_server_ctx = wolfSSL_get_SSL_CTX(ssl);
+        AssertNotNull(test_wolfSSL_CTX_add_session_server_ctx);
+        AssertIntEQ(wolfSSL_CTX_up_ref(wolfSSL_get_SSL_CTX(ssl)),
+                WOLFSSL_SUCCESS);
+    }
+#if defined(SESSION_CERTS) && !defined(WOLFSSL_NO_CLIENT_AUTH)
+#ifndef WOLFSSL_TICKET_HAVE_ID
+    if (wolfSSL_version(ssl) != TLS1_3_VERSION &&
+            wolfSSL_session_reused(ssl))
+#endif
+    {
+        /* With WOLFSSL_TICKET_HAVE_ID the peer certs should be available
+         * for all connections. TLS 1.3 only has tickets so if we don't
+         * include the session id in the ticket then the certificates
+         * will not be available on resumption. */
+    #ifdef KEEP_PEER_CERT
+        WOLFSSL_X509* peer = wolfSSL_get_peer_certificate(ssl);
+        AssertNotNull(peer);
+        wolfSSL_X509_free(peer);
+    #endif
+        AssertNotNull(wolfSSL_SESSION_get_peer_chain(*sess));
+    #ifdef OPENSSL_EXTRA
+        AssertNotNull(SSL_SESSION_get0_peer(*sess));
+    #endif
+    }
+#endif /* SESSION_CERTS && !WOLFSSL_NO_CLIENT_AUTH */
+}
+
+static void test_wolfSSL_CTX_add_session_ssl_ready(WOLFSSL* ssl)
+{
+    /* Set the session to reuse for the client */
+    AssertIntEQ(wolfSSL_set_session(ssl,
+            test_wolfSSL_CTX_add_session_client_sess), WOLFSSL_SUCCESS);
+}
+#endif
+
+int test_wolfSSL_CTX_add_session(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_EXT_CACHE) && \
+    !defined(SINGLE_THREADED) && defined(WOLFSSL_TLS13) && \
+    !defined(NO_SESSION_CACHE)
+    tcp_ready ready;
+    func_args client_args;
+    func_args server_args;
+    THREAD_TYPE serverThread;
+    callback_functions client_cb;
+    callback_functions server_cb;
+    method_provider methods[][2] = {
+#if !defined(NO_OLD_TLS) && ((!defined(NO_AES) && !defined(NO_AES_CBC)) || \
+        !defined(NO_DES3))
+        /* Without AES there are almost no ciphersuites available. This leads
+         * to no ciphersuites being available and an error. */
+        { wolfTLSv1_1_client_method, wolfTLSv1_1_server_method },
+#endif
+#ifndef WOLFSSL_NO_TLS12
+        { wolfTLSv1_2_client_method, wolfTLSv1_2_server_method },
+#endif
+        /* Needs the default ticket callback since it is tied to the
+         * connection context and this makes it easy to carry over the ticket
+         * crypto context between connections */
+#if defined(WOLFSSL_TLS13) && !defined(WOLFSSL_NO_DEF_TICKET_ENC_CB) && \
+    defined(HAVE_SESSION_TICKET)
+        { wolfTLSv1_3_client_method, wolfTLSv1_3_server_method },
+#endif
+    };
+    const size_t methodsLen = sizeof(methods)/sizeof(*methods);
+    size_t i, j;
+
+    for (i = 0; i < methodsLen; i++) {
+        /* First run creates a connection while the second+ run will attempt
+         * to resume the connection. The trick is that the internal cache
+         * is turned off. wolfSSL_CTX_add_session should put the session in
+         * the cache anyway. */
+        test_wolfSSL_CTX_add_session_client_sess = NULL;
+        test_wolfSSL_CTX_add_session_server_sess = NULL;
+        test_wolfSSL_CTX_add_session_server_ctx = NULL;
+
+#ifdef NO_SESSION_CACHE_REF
+        for (j = 0; j < 4; j++) {
+#else
+        /* The session may be overwritten in this case. Do only one resumption
+         * to stop this test from failing intermittently. */
+        for (j = 0; j < 2; j++) {
+#endif
+#ifdef WOLFSSL_TIRTOS
+            fdOpenSession(Task_self());
+#endif
+
+            StartTCP();
+            InitTcpReady(&ready);
+
+            XMEMSET(&client_args, 0, sizeof(func_args));
+            XMEMSET(&server_args, 0, sizeof(func_args));
+
+            XMEMSET(&client_cb, 0, sizeof(callback_functions));
+            XMEMSET(&server_cb, 0, sizeof(callback_functions));
+            client_cb.method  = methods[i][0];
+            server_cb.method  = methods[i][1];
+
+            server_args.signal    = &ready;
+            server_args.callbacks = &server_cb;
+            client_args.signal    = &ready;
+            client_args.callbacks = &client_cb;
+
+            if (test_wolfSSL_CTX_add_session_server_ctx != NULL) {
+                server_cb.ctx = test_wolfSSL_CTX_add_session_server_ctx;
+                server_cb.isSharedCtx = 1;
+            }
+            server_cb.ctx_ready = test_wolfSSL_CTX_add_session_ctx_ready;
+            client_cb.ctx_ready = test_wolfSSL_CTX_add_session_ctx_ready;
+            if (j != 0)
+                client_cb.ssl_ready = test_wolfSSL_CTX_add_session_ssl_ready;
+            server_cb.on_result = test_wolfSSL_CTX_add_session_on_result;
+            client_cb.on_result = test_wolfSSL_CTX_add_session_on_result;
+            server_cb.ticNoInit = 1; /* Use default builtin */
+
+            start_thread(test_server_nofail, &server_args, &serverThread);
+            wait_tcp_ready(&server_args);
+            test_client_nofail(&client_args, NULL);
+            join_thread(serverThread);
+
+            ExpectTrue(client_args.return_code);
+            ExpectTrue(server_args.return_code);
+
+            FreeTcpReady(&ready);
+
+            if (EXPECT_FAIL())
+                break;
+        }
+        wolfSSL_SESSION_free(test_wolfSSL_CTX_add_session_client_sess);
+        wolfSSL_SESSION_free(test_wolfSSL_CTX_add_session_server_sess);
+        wolfSSL_CTX_free(test_wolfSSL_CTX_add_session_server_ctx);
+
+        if (EXPECT_FAIL())
+            break;
+    }
+#endif
+
+    return EXPECT_RESULT();
+}
+#if defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES) && defined(HAVE_EXT_CACHE) && \
+     defined(WOLFSSL_TLS13) && !defined(NO_SESSION_CACHE) && \
+     defined(OPENSSL_EXTRA) && defined(SESSION_CERTS) && \
+     defined(HAVE_SESSION_TICKET) && \
+    !defined(TITAN_SESSION_CACHE) && \
+    !defined(HUGE_SESSION_CACHE) && \
+    !defined(BIG_SESSION_CACHE) && \
+    !defined(MEDIUM_SESSION_CACHE)
+
+/* twcase - prefix for test_wolfSSL_CTX_add_session_ext */
+/* Sessions to restore/store */
+static WOLFSSL_SESSION* twcase_server_first_session_ptr;
+static WOLFSSL_SESSION* twcase_client_first_session_ptr;
+static WOLFSSL_CTX*     twcase_server_current_ctx_ptr;
+static int twcase_new_session_called    = 0;
+static int twcase_remove_session_called = 0;
+static int twcase_get_session_called    = 0;
+
+/* Test default, SESSIONS_PER_ROW*SESSION_ROWS = 3*11, see ssl.c */
+#define SESSION_CACHE_SIZE 33
+
+typedef struct {
+    const byte* key;  /* key, altSessionID, session ID, NULL if empty */
+    WOLFSSL_SESSION* value;
+} hashTable_entry;
+
+typedef struct {
+    hashTable_entry entries[SESSION_CACHE_SIZE];  /* hash slots */
+    size_t capacity;                     /* size of entries */
+    size_t length;                       /* number of items in the hash table */
+    wolfSSL_Mutex htLock;                /* lock */
+}hashTable;
+
+static hashTable server_sessionCache;
+
+static int twcase_new_sessionCb(WOLFSSL *ssl, WOLFSSL_SESSION *sess)
+{
+    int i;
+    unsigned int len;
+    (void)ssl;
+
+    /*
+     * This example uses a hash table.
+     * Steps you should take for a non-demo code:
+     * - acquire a lock for the file named according to the session id
+     * - open the file
+     * - encrypt and write the SSL_SESSION object to the file
+     * - release the lock
+     *
+     * Return:
+     *  0: The callback does not wish to hold a reference of the sess
+     *  1: The callback wants to hold a reference of the sess. The callback is
+     *     now also responsible for calling wolfSSL_SESSION_free() on sess.
+     */
+    if (sess == NULL)
+        return 0;
+
+    if (wc_LockMutex(&server_sessionCache.htLock) != 0) {
+        return 0;
+    }
+    for (i = 0; i < SESSION_CACHE_SIZE; i++) {
+        if (server_sessionCache.entries[i].value == NULL) {
+            server_sessionCache.entries[i].key = SSL_SESSION_get_id(sess, &len);
+            server_sessionCache.entries[i].value = sess;
+            server_sessionCache.length++;
+            break;
+        }
+    }
+    ++twcase_new_session_called;
+    wc_UnLockMutex(&server_sessionCache.htLock);
+    fprintf(stderr, "\t\ttwcase_new_session_called %d\n",
+            twcase_new_session_called);
+    return 1;
+}
+
+static void twcase_remove_sessionCb(WOLFSSL_CTX *ctx, WOLFSSL_SESSION *sess)
+{
+    int i;
+    (void)ctx;
+    (void)sess;
+
+    if (sess == NULL)
+        return;
+    /*
+     * This example uses a hash table.
+     * Steps you should take for a non-demo code:
+     * - acquire a lock for the file named according to the session id
+     * - remove the file
+     * - release the lock
+     */
+    if (wc_LockMutex(&server_sessionCache.htLock) != 0) {
+        return;
+    }
+    for (i = 0; i < SESSION_CACHE_SIZE; i++) {
+        if (server_sessionCache.entries[i].key != NULL &&
+           XMEMCMP(server_sessionCache.entries[i].key,
+                   sess->sessionID, SSL_MAX_SSL_SESSION_ID_LENGTH) == 0) {
+            wolfSSL_SESSION_free(server_sessionCache.entries[i].value);
+            server_sessionCache.entries[i].value = NULL;
+            server_sessionCache.entries[i].key = NULL;
+            server_sessionCache.length--;
+            break;
+        }
+    }
+    ++twcase_remove_session_called;
+    wc_UnLockMutex(&server_sessionCache.htLock);
+    fprintf(stderr, "\t\ttwcase_remove_session_called %d\n",
+            twcase_remove_session_called);
+}
+
+static WOLFSSL_SESSION *twcase_get_sessionCb(WOLFSSL *ssl,
+                                  const unsigned char *id, int len, int *ref)
+{
+    int i;
+    (void)ssl;
+    (void)id;
+    (void)len;
+
+    /*
+     * This example uses a hash table.
+     * Steps you should take for a non-demo code:
+     * - acquire a lock for the file named according to the session id in the
+     *   2nd arg
+     * - read and decrypt contents of file and create a new SSL_SESSION
+     * - object release the lock
+     * - return the new session object
+     */
+    fprintf(stderr, "\t\ttwcase_get_session_called %d\n",
+            ++twcase_get_session_called);
+    /* This callback want to retain a copy of the object. If we want wolfSSL to
+     * be responsible for the pointer then set to 0. */
+    *ref = 1;
+
+    for (i = 0; i < SESSION_CACHE_SIZE; i++) {
+        if (server_sessionCache.entries[i].key != NULL &&
+           XMEMCMP(server_sessionCache.entries[i].key, id,
+                   SSL_MAX_SSL_SESSION_ID_LENGTH) == 0) {
+           return server_sessionCache.entries[i].value;
+        }
+    }
+    return NULL;
+}
+static int twcase_get_sessionCb_cleanup(void)
+{
+    int i;
+    int cnt = 0;
+
+    /* If  twcase_get_sessionCb sets *ref = 1, the application is responsible
+     * for freeing sessions */
+
+    for (i = 0; i < SESSION_CACHE_SIZE; i++) {
+        if (server_sessionCache.entries[i].value != NULL) {
+            wolfSSL_SESSION_free(server_sessionCache.entries[i].value);
+            cnt++;
+        }
+    }
+
+    fprintf(stderr, "\t\ttwcase_get_sessionCb_cleanup freed %d sessions\n",
+            cnt);
+
+    return TEST_SUCCESS;
+}
+
+static int twcase_cache_intOff_extOff(WOLFSSL_CTX* ctx)
+{
+    EXPECT_DECLS;
+    /* off - Disable internal cache */
+    ExpectIntEQ(wolfSSL_CTX_set_session_cache_mode(ctx,
+            WOLFSSL_SESS_CACHE_NO_INTERNAL_STORE), WOLFSSL_SUCCESS);
+#ifdef OPENSSL_EXTRA
+    ExpectIntEQ(wolfSSL_CTX_get_session_cache_mode(ctx) &
+            WOLFSSL_SESS_CACHE_NO_INTERNAL_STORE,
+            WOLFSSL_SESS_CACHE_NO_INTERNAL_STORE);
+#endif
+    /* off - Do not setup external cache */
+
+    /* Require both peers to provide certs */
+    wolfSSL_CTX_set_verify(ctx, WOLFSSL_VERIFY_PEER, NULL);
+    return EXPECT_RESULT();
+}
+
+static int twcase_cache_intOn_extOff(WOLFSSL_CTX* ctx)
+{
+    /* on - internal cache is on by default */
+    /* off - Do not setup external cache */
+    /* Require both peers to provide certs */
+    wolfSSL_CTX_set_verify(ctx, WOLFSSL_VERIFY_PEER, NULL);
+    return TEST_SUCCESS;
+}
+
+static int twcase_cache_intOff_extOn(WOLFSSL_CTX* ctx)
+{
+    EXPECT_DECLS;
+    /* off - Disable internal cache */
+    ExpectIntEQ(wolfSSL_CTX_set_session_cache_mode(ctx,
+            WOLFSSL_SESS_CACHE_NO_INTERNAL_STORE), WOLFSSL_SUCCESS);
+#ifdef OPENSSL_EXTRA
+    ExpectIntEQ(wolfSSL_CTX_get_session_cache_mode(ctx) &
+            WOLFSSL_SESS_CACHE_NO_INTERNAL_STORE,
+            WOLFSSL_SESS_CACHE_NO_INTERNAL_STORE);
+#endif
+    /* on - Enable external cache */
+    wolfSSL_CTX_sess_set_new_cb(ctx, twcase_new_sessionCb);
+    wolfSSL_CTX_sess_set_remove_cb(ctx, twcase_remove_sessionCb);
+    wolfSSL_CTX_sess_set_get_cb(ctx, twcase_get_sessionCb);
+
+    /* Require both peers to provide certs */
+    wolfSSL_CTX_set_verify(ctx, WOLFSSL_VERIFY_PEER, NULL);
+    return EXPECT_RESULT();
+}
+
+static int twcase_cache_intOn_extOn(WOLFSSL_CTX* ctx)
+{
+    /* on - internal cache is on by default */
+    /* on - Enable external cache */
+    wolfSSL_CTX_sess_set_new_cb(ctx, twcase_new_sessionCb);
+    wolfSSL_CTX_sess_set_remove_cb(ctx, twcase_remove_sessionCb);
+    wolfSSL_CTX_sess_set_get_cb(ctx, twcase_get_sessionCb);
+
+    /* Require both peers to provide certs */
+    wolfSSL_CTX_set_verify(ctx, WOLFSSL_VERIFY_PEER, NULL);
+    return TEST_SUCCESS;
+}
+static int twcase_cache_intOn_extOn_noTicket(WOLFSSL_CTX* ctx)
+{
+    /* on - internal cache is on by default */
+    /* on - Enable external cache */
+    wolfSSL_CTX_sess_set_new_cb(ctx, twcase_new_sessionCb);
+    wolfSSL_CTX_sess_set_remove_cb(ctx, twcase_remove_sessionCb);
+    wolfSSL_CTX_sess_set_get_cb(ctx, twcase_get_sessionCb);
+
+    wolfSSL_CTX_set_options(ctx, WOLFSSL_OP_NO_TICKET);
+    /* Require both peers to provide certs */
+    wolfSSL_CTX_set_verify(ctx, WOLFSSL_VERIFY_PEER, NULL);
+    return TEST_SUCCESS;
+}
+static int twcase_server_sess_ctx_pre_shutdown(WOLFSSL* ssl)
+{
+    EXPECT_DECLS;
+    WOLFSSL_SESSION** sess;
+    if (wolfSSL_is_server(ssl))
+        sess = &twcase_server_first_session_ptr;
+    else
+        return TEST_SUCCESS;
+
+    if (*sess == NULL) {
+        ExpectNotNull(*sess = wolfSSL_get1_session(ssl));
+        /* Now save the session in the internal store to make it available
+         * for lookup. For TLS 1.3, we can't save the session without
+         * WOLFSSL_TICKET_HAVE_ID because there is no way to retrieve the
+         * session from cache. */
+        if (wolfSSL_is_server(ssl)
+#ifndef WOLFSSL_TICKET_HAVE_ID
+                && wolfSSL_version(ssl) != TLS1_3_VERSION
+                && wolfSSL_version(ssl) != DTLS1_3_VERSION
+#endif
+                ) {
+            ExpectIntEQ(wolfSSL_CTX_add_session(wolfSSL_get_SSL_CTX(ssl),
+                    *sess), WOLFSSL_SUCCESS);
+        }
+    }
+    /* Save CTX to be able to decrypt tickets */
+    if (twcase_server_current_ctx_ptr == NULL) {
+        ExpectNotNull(twcase_server_current_ctx_ptr = wolfSSL_get_SSL_CTX(ssl));
+        ExpectIntEQ(wolfSSL_CTX_up_ref(wolfSSL_get_SSL_CTX(ssl)),
+                    WOLFSSL_SUCCESS);
+    }
+#if defined(SESSION_CERTS) && !defined(WOLFSSL_NO_CLIENT_AUTH)
+#ifndef WOLFSSL_TICKET_HAVE_ID
+    if (wolfSSL_version(ssl) != TLS1_3_VERSION &&
+            wolfSSL_session_reused(ssl))
+#endif
+    {
+        /* With WOLFSSL_TICKET_HAVE_ID the peer certs should be available
+         * for all connections. TLS 1.3 only has tickets so if we don't
+         * include the session id in the ticket then the certificates
+         * will not be available on resumption. */
+    #ifdef KEEP_PEER_CERT
+        WOLFSSL_X509* peer = NULL;
+        ExpectNotNull(peer = wolfSSL_get_peer_certificate(ssl));
+        wolfSSL_X509_free(peer);
+    #endif
+        ExpectNotNull(wolfSSL_SESSION_get_peer_chain(*sess));
+    }
+#endif
+    return EXPECT_RESULT();
+}
+
+static int twcase_client_sess_ctx_pre_shutdown(WOLFSSL* ssl)
+{
+    EXPECT_DECLS;
+    WOLFSSL_SESSION** sess;
+    sess = &twcase_client_first_session_ptr;
+    if (*sess == NULL) {
+        ExpectNotNull(*sess = wolfSSL_get1_session(ssl));
+    }
+    else {
+        /* If we have a session retrieved then remaining connections should be
+         * resuming on that session */
+        ExpectIntEQ(wolfSSL_session_reused(ssl), 1);
+    }
+
+#if defined(SESSION_CERTS) && !defined(WOLFSSL_NO_CLIENT_AUTH)
+#ifndef WOLFSSL_TICKET_HAVE_ID
+    if (wolfSSL_version(ssl) != TLS1_3_VERSION &&
+            wolfSSL_session_reused(ssl))
+#endif
+    {
+    #ifdef KEEP_PEER_CERT
+        WOLFSSL_X509* peer = wolfSSL_get_peer_certificate(ssl);
+        ExpectNotNull(peer);
+        wolfSSL_X509_free(peer);
+    #endif
+        ExpectNotNull(wolfSSL_SESSION_get_peer_chain(*sess));
+#ifdef OPENSSL_EXTRA
+        ExpectNotNull(wolfSSL_SESSION_get0_peer(*sess));
+#endif
+    }
+#endif
+    return EXPECT_RESULT();
+}
+static int twcase_client_set_sess_ssl_ready(WOLFSSL* ssl)
+{
+    EXPECT_DECLS;
+    /* Set the session to reuse for the client */
+    ExpectNotNull(ssl);
+    ExpectNotNull(twcase_client_first_session_ptr);
+    ExpectIntEQ(wolfSSL_set_session(ssl,twcase_client_first_session_ptr),
+                WOLFSSL_SUCCESS);
+    return EXPECT_RESULT();
+}
+
+struct test_add_session_ext_params {
+    method_provider client_meth;
+    method_provider server_meth;
+    const char* tls_version;
+};
+
+/* Marked WC_MAYBE_UNUSED: each registered test_wolfSSL_CTX_add_session_ext_*
+ * variant below calls this helper, but each variant is gated on a specific
+ * TLS/DTLS version combination. In builds where no version combination is
+ * enabled, the helper is defined but unused. */
+static WC_MAYBE_UNUSED int test_wolfSSL_CTX_add_session_ext(
+    struct test_add_session_ext_params* param)
+{
+    EXPECT_DECLS;
+    /* Test the default 33 sessions */
+    int j;
+
+    /* Clear cache before starting */
+    wolfSSL_CTX_flush_sessions(NULL, -1);
+
+    XMEMSET(&server_sessionCache, 0, sizeof(hashTable));
+    if (wc_InitMutex(&server_sessionCache.htLock) != 0)
+        return BAD_MUTEX_E;
+    server_sessionCache.capacity = SESSION_CACHE_SIZE;
+
+    fprintf(stderr, "\tBegin %s\n", param->tls_version);
+    for (j = 0; j < 5; j++) {
+        int tls13 = XSTRSTR(param->tls_version, "TLSv1_3") != NULL;
+        int dtls = XSTRSTR(param->tls_version, "DTLS") != NULL;
+        test_ssl_cbf client_cb;
+        test_ssl_cbf server_cb;
+
+        (void)dtls;
+
+        /* Test five cache configurations */
+        twcase_client_first_session_ptr = NULL;
+        twcase_server_first_session_ptr = NULL;
+        twcase_server_current_ctx_ptr = NULL;
+        twcase_new_session_called    = 0;
+        twcase_remove_session_called = 0;
+        twcase_get_session_called    = 0;
+
+        /* connection 1 - first connection */
+        fprintf(stderr, "\tconnect: %s: j=%d\n", param->tls_version, j);
+
+        XMEMSET(&client_cb, 0, sizeof(client_cb));
+        XMEMSET(&server_cb, 0, sizeof(server_cb));
+        client_cb.method  = param->client_meth;
+        server_cb.method  = param->server_meth;
+
+        if (dtls)
+            client_cb.doUdp = server_cb.doUdp = 1;
+
+        /* Setup internal and external cache */
+        switch (j) {
+            case 0:
+                /* SSL_OP_NO_TICKET stateful ticket case */
+                server_cb.ctx_ready = twcase_cache_intOn_extOn_noTicket;
+                break;
+            case 1:
+                server_cb.ctx_ready = twcase_cache_intOn_extOn;
+                break;
+            case 2:
+                server_cb.ctx_ready = twcase_cache_intOff_extOn;
+                break;
+            case 3:
+                server_cb.ctx_ready = twcase_cache_intOn_extOff;
+                break;
+            case 4:
+                server_cb.ctx_ready = twcase_cache_intOff_extOff;
+                break;
+        }
+        client_cb.ctx_ready = twcase_cache_intOff_extOff;
+
+        /* Add session to internal cache and save SSL session for testing */
+        server_cb.on_result = twcase_server_sess_ctx_pre_shutdown;
+        /* Save client SSL session for testing */
+        client_cb.on_result = twcase_client_sess_ctx_pre_shutdown;
+        server_cb.ticNoInit = 1; /* Use default builtin */
+        /* Don't free/release ctx */
+        server_cb.ctx = twcase_server_current_ctx_ptr;
+        server_cb.isSharedCtx = 1;
+
+        ExpectIntEQ(test_wolfSSL_client_server_nofail_memio(&client_cb,
+            &server_cb, NULL), TEST_SUCCESS);
+
+        ExpectIntEQ(twcase_get_session_called, 0);
+        if (EXPECT_FAIL()) {
+            wolfSSL_SESSION_free(twcase_client_first_session_ptr);
+            wolfSSL_SESSION_free(twcase_server_first_session_ptr);
+            wolfSSL_CTX_free(twcase_server_current_ctx_ptr);
+            break;
+        }
+
+        switch (j) {
+            case 0:
+            case 1:
+            case 2:
+                /* cache cannot be searched with out a connection */
+                /* Add a new session */
+                ExpectIntEQ(twcase_new_session_called, 1);
+                /* In twcase_server_sess_ctx_pre_shutdown
+                 * wolfSSL_CTX_add_session which evicts the existing session
+                 * in cache and adds it back in */
+                ExpectIntLE(twcase_remove_session_called, 1);
+                break;
+            case 3:
+            case 4:
+                /* no external cache  */
+                ExpectIntEQ(twcase_new_session_called, 0);
+                ExpectIntEQ(twcase_remove_session_called, 0);
+                break;
+        }
+
+        /* connection 2 - session resume */
+        fprintf(stderr, "\tresume: %s: j=%d\n", param->tls_version, j);
+        twcase_new_session_called    = 0;
+        twcase_remove_session_called = 0;
+        twcase_get_session_called    = 0;
+        server_cb.on_result = 0;
+        client_cb.on_result = 0;
+        server_cb.ticNoInit = 1; /* Use default builtin */
+
+        server_cb.ctx = twcase_server_current_ctx_ptr;
+
+        /* try session resumption */
+        client_cb.ssl_ready = twcase_client_set_sess_ssl_ready;
+
+        ExpectIntEQ(test_wolfSSL_client_server_nofail_memio(&client_cb,
+            &server_cb, NULL), TEST_SUCCESS);
+
+        /* Clear cache before checking */
+        wolfSSL_CTX_flush_sessions(NULL, -1);
+
+        switch (j) {
+            case 0:
+                if (tls13) {
+                    /* (D)TLSv1.3 stateful case */
+                    /* cache hit */
+                    /* DTLS accesses cache once for stateless parsing and
+                     * once for stateful parsing */
+                    ExpectIntEQ(twcase_get_session_called, !dtls ? 1 : 2);
+
+                    /* (D)TLSv1.3 creates a new ticket,
+                     * updates both internal and external cache */
+                    ExpectIntEQ(twcase_new_session_called, 1);
+                    /* A new session ID is created for a new ticket */
+                    ExpectIntEQ(twcase_remove_session_called, 2);
+
+                }
+                else {
+                    /* non (D)TLSv1.3 case, no update */
+                    /* DTLS accesses cache once for stateless parsing and
+                     * once for stateful parsing */
+#ifdef WOLFSSL_DTLS_NO_HVR_ON_RESUME
+                    ExpectIntEQ(twcase_get_session_called, !dtls ? 1 : 2);
+#else
+                    ExpectIntEQ(twcase_get_session_called, 1);
+#endif
+                    ExpectIntEQ(twcase_new_session_called, 0);
+                    /* Called on session added in
+                     * twcase_server_sess_ctx_pre_shutdown */
+                    ExpectIntEQ(twcase_remove_session_called, 1);
+                }
+                break;
+            case 1:
+                if (tls13) {
+                    /* (D)TLSv1.3 case */
+                    /* cache hit */
+                    ExpectIntEQ(twcase_get_session_called, 1);
+                    /* (D)TLSv1.3 creates a new ticket,
+                     * updates both internal and external cache */
+                    ExpectIntEQ(twcase_new_session_called, 1);
+                    /* Called on session added in
+                     * twcase_server_sess_ctx_pre_shutdown and by wolfSSL */
+                    ExpectIntEQ(twcase_remove_session_called, 1);
+                }
+                else {
+                    /* non (D)TLSv1.3 case */
+                    /* cache hit */
+                    /* DTLS accesses cache once for stateless parsing and
+                     * once for stateful parsing */
+#ifdef WOLFSSL_DTLS_NO_HVR_ON_RESUME
+                    ExpectIntEQ(twcase_get_session_called, !dtls ? 1 : 2);
+#else
+                    ExpectIntEQ(twcase_get_session_called, 1);
+#endif
+                    ExpectIntEQ(twcase_new_session_called, 0);
+                    /* Called on session added in
+                     * twcase_server_sess_ctx_pre_shutdown */
+                    ExpectIntEQ(twcase_remove_session_called, 1);
+                }
+                break;
+            case 2:
+                if (tls13) {
+                    /* (D)TLSv1.3 case */
+                    /* cache hit */
+                    ExpectIntEQ(twcase_get_session_called, 1);
+                    /* (D)TLSv1.3 creates a new ticket,
+                     * updates both internal and external cache */
+                    ExpectIntEQ(twcase_new_session_called, 1);
+                    /* Called on session added in
+                     * twcase_server_sess_ctx_pre_shutdown and by wolfSSL */
+                    ExpectIntEQ(twcase_remove_session_called, 1);
+                }
+                else {
+                    /* non (D)TLSv1.3 case */
+                    /* cache hit */
+                    /* DTLS accesses cache once for stateless parsing and
+                     * once for stateful parsing */
+#ifdef WOLFSSL_DTLS_NO_HVR_ON_RESUME
+                    ExpectIntEQ(twcase_get_session_called, !dtls ? 1 : 2);
+#else
+                    ExpectIntEQ(twcase_get_session_called, 1);
+#endif
+                    ExpectIntEQ(twcase_new_session_called, 0);
+                    /* Called on session added in
+                     * twcase_server_sess_ctx_pre_shutdown */
+                    ExpectIntEQ(twcase_remove_session_called, 1);
+                }
+                break;
+            case 3:
+            case 4:
+                /* no external cache */
+                ExpectIntEQ(twcase_get_session_called, 0);
+                ExpectIntEQ(twcase_new_session_called, 0);
+                ExpectIntEQ(twcase_remove_session_called, 0);
+                break;
+        }
+        wolfSSL_SESSION_free(twcase_client_first_session_ptr);
+        wolfSSL_SESSION_free(twcase_server_first_session_ptr);
+        wolfSSL_CTX_free(twcase_server_current_ctx_ptr);
+
+        if (EXPECT_FAIL())
+            break;
+    }
+    twcase_get_sessionCb_cleanup();
+    XMEMSET(&server_sessionCache.entries, 0,
+            sizeof(server_sessionCache.entries));
+    fprintf(stderr, "\tEnd %s\n", param->tls_version);
+
+    wc_FreeMutex(&server_sessionCache.htLock);
+
+    return EXPECT_RESULT();
+}
+#endif
+
+int test_wolfSSL_CTX_add_session_ext_tls13(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_EXT_CACHE) && \
+     defined(WOLFSSL_TLS13) && !defined(NO_SESSION_CACHE) && \
+     defined(OPENSSL_EXTRA) && defined(SESSION_CERTS) && \
+     defined(HAVE_SESSION_TICKET) && \
+    !defined(TITAN_SESSION_CACHE) && \
+    !defined(HUGE_SESSION_CACHE) && \
+    !defined(BIG_SESSION_CACHE) && \
+    !defined(MEDIUM_SESSION_CACHE)
+#if defined(WOLFSSL_TLS13) && !defined(WOLFSSL_NO_DEF_TICKET_ENC_CB) && \
+    defined(HAVE_SESSION_TICKET) && defined(WOLFSSL_TICKET_HAVE_ID)
+    struct test_add_session_ext_params param[1] =  {
+        { wolfTLSv1_3_client_method, wolfTLSv1_3_server_method, "TLSv1_3" }
+    };
+    ExpectIntEQ(test_wolfSSL_CTX_add_session_ext(param), TEST_SUCCESS);
+#endif
+#endif
+    return EXPECT_RESULT();
+}
+int test_wolfSSL_CTX_add_session_ext_dtls13(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_EXT_CACHE) && \
+     defined(WOLFSSL_TLS13) && !defined(NO_SESSION_CACHE) && \
+     defined(OPENSSL_EXTRA) && defined(SESSION_CERTS) && \
+     defined(HAVE_SESSION_TICKET) && \
+    !defined(TITAN_SESSION_CACHE) && \
+    !defined(HUGE_SESSION_CACHE) && \
+    !defined(BIG_SESSION_CACHE) && \
+    !defined(MEDIUM_SESSION_CACHE)
+#if defined(WOLFSSL_TLS13) && !defined(WOLFSSL_NO_DEF_TICKET_ENC_CB) && \
+    defined(HAVE_SESSION_TICKET) && defined(WOLFSSL_TICKET_HAVE_ID)
+#ifdef WOLFSSL_DTLS13
+    struct test_add_session_ext_params param[1] =  {
+        { wolfDTLSv1_3_client_method, wolfDTLSv1_3_server_method, "DTLSv1_3" }
+    };
+    ExpectIntEQ(test_wolfSSL_CTX_add_session_ext(param), TEST_SUCCESS);
+#endif
+#endif
+#endif
+    return EXPECT_RESULT();
+}
+int test_wolfSSL_CTX_add_session_ext_tls12(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_EXT_CACHE) && \
+     defined(WOLFSSL_TLS13) && !defined(NO_SESSION_CACHE) && \
+     defined(OPENSSL_EXTRA) && defined(SESSION_CERTS) && \
+     defined(HAVE_SESSION_TICKET) && \
+    !defined(TITAN_SESSION_CACHE) && \
+    !defined(HUGE_SESSION_CACHE) && \
+    !defined(BIG_SESSION_CACHE) && \
+    !defined(MEDIUM_SESSION_CACHE)
+#ifndef WOLFSSL_NO_TLS12
+    struct test_add_session_ext_params param[1] =  {
+        { wolfTLSv1_2_client_method, wolfTLSv1_2_server_method, "TLSv1_2" }
+    };
+    ExpectIntEQ(test_wolfSSL_CTX_add_session_ext(param), TEST_SUCCESS);
+#endif
+#endif
+    return EXPECT_RESULT();
+}
+int test_wolfSSL_CTX_add_session_ext_dtls12(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_EXT_CACHE) && \
+     defined(WOLFSSL_TLS13) && !defined(NO_SESSION_CACHE) && \
+     defined(OPENSSL_EXTRA) && defined(SESSION_CERTS) && \
+     defined(HAVE_SESSION_TICKET) && \
+    !defined(TITAN_SESSION_CACHE) && \
+    !defined(HUGE_SESSION_CACHE) && \
+    !defined(BIG_SESSION_CACHE) && \
+    !defined(MEDIUM_SESSION_CACHE)
+#ifndef WOLFSSL_NO_TLS12
+#ifdef WOLFSSL_DTLS
+    struct test_add_session_ext_params param[1] =  {
+        { wolfDTLSv1_2_client_method, wolfDTLSv1_2_server_method, "DTLSv1_2" }
+    };
+    ExpectIntEQ(test_wolfSSL_CTX_add_session_ext(param), TEST_SUCCESS);
+#endif
+#endif
+#endif
+    return EXPECT_RESULT();
+}
+int test_wolfSSL_CTX_add_session_ext_tls11(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_EXT_CACHE) && \
+     defined(WOLFSSL_TLS13) && !defined(NO_SESSION_CACHE) && \
+     defined(OPENSSL_EXTRA) && defined(SESSION_CERTS) && \
+     defined(HAVE_SESSION_TICKET) && \
+    !defined(TITAN_SESSION_CACHE) && \
+    !defined(HUGE_SESSION_CACHE) && \
+    !defined(BIG_SESSION_CACHE) && \
+    !defined(MEDIUM_SESSION_CACHE)
+#if !defined(NO_OLD_TLS) && ((!defined(NO_AES) && !defined(NO_AES_CBC)) || \
+        !defined(NO_DES3))
+    struct test_add_session_ext_params param[1] =  {
+        { wolfTLSv1_1_client_method, wolfTLSv1_1_server_method, "TLSv1_1" }
+    };
+    ExpectIntEQ(test_wolfSSL_CTX_add_session_ext(param), TEST_SUCCESS);
+#endif
+#endif
+    return EXPECT_RESULT();
+}
+int test_wolfSSL_CTX_add_session_ext_dtls1(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_IO_TESTS_DEPENDENCIES) && defined(HAVE_EXT_CACHE) && \
+     defined(WOLFSSL_TLS13) && !defined(NO_SESSION_CACHE) && \
+     defined(OPENSSL_EXTRA) && defined(SESSION_CERTS) && \
+     defined(HAVE_SESSION_TICKET) && \
+    !defined(TITAN_SESSION_CACHE) && \
+    !defined(HUGE_SESSION_CACHE) && \
+    !defined(BIG_SESSION_CACHE) && \
+    !defined(MEDIUM_SESSION_CACHE)
+#if !defined(NO_OLD_TLS) && ((!defined(NO_AES) && !defined(NO_AES_CBC)) || \
+        !defined(NO_DES3))
+#ifdef WOLFSSL_DTLS
+    struct test_add_session_ext_params param[1] =  {
+        { wolfDTLSv1_client_method, wolfDTLSv1_server_method, "DTLSv1_0" }
+    };
+    ExpectIntEQ(test_wolfSSL_CTX_add_session_ext(param), TEST_SUCCESS);
+#endif
+#endif
+#endif
+    return EXPECT_RESULT();
+}
+
+/*----------------------------------------------------------------------------*/
+/* WOLFSSL_SESSION, ticket keys and session removal callbacks                 */
+/*----------------------------------------------------------------------------*/
+
+int test_wolfSSL_SESSION(void)
+{
+    EXPECT_DECLS;
+#if !defined(NO_FILESYSTEM) && !defined(NO_CERTS) && \
+    !defined(NO_RSA) && !defined(NO_SHA256) && \
+    defined(HAVE_IO_TESTS_DEPENDENCIES) && !defined(NO_SESSION_CACHE)
+    WOLFSSL*     ssl = NULL;
+    WOLFSSL_CTX* ctx = NULL;
+    WOLFSSL_SESSION* sess = NULL;
+    WOLFSSL_SESSION* sess_copy = NULL;
+#ifdef OPENSSL_EXTRA
+#ifdef HAVE_EXT_CACHE
+    unsigned char* sessDer = NULL;
+    unsigned char* ptr     = NULL;
+    int sz = 0;
+#endif
+    const unsigned char context[] = "user app context";
+    unsigned int contextSz = (unsigned int)sizeof(context);
+#endif
+    int ret = 0, err = 0;
+    SOCKET_T sockfd;
+    tcp_ready ready;
+    func_args server_args;
+    THREAD_TYPE serverThread;
+    char msg[80];
+    const char* sendGET = "GET";
+
+    /* TLS v1.3 requires session tickets */
+    /* CHACHA and POLY1305 required for myTicketEncCb */
+#if !defined(WOLFSSL_NO_TLS12) && (!defined(WOLFSSL_TLS13) || \
+    !(defined(HAVE_SESSION_TICKET) && ((defined(HAVE_CHACHA) && \
+            defined(HAVE_POLY1305)) || defined(HAVE_AESGCM))))
+    ExpectNotNull(ctx = wolfSSL_CTX_new(wolfTLSv1_2_client_method()));
+#else
+    ExpectNotNull(ctx = wolfSSL_CTX_new(wolfSSLv23_client_method()));
+#endif
+
+    ExpectTrue(wolfSSL_CTX_use_certificate_file(ctx, cliCertFile,
+        CERT_FILETYPE));
+    ExpectTrue(wolfSSL_CTX_use_PrivateKey_file(ctx, cliKeyFile,
+        CERT_FILETYPE));
+    ExpectIntEQ(wolfSSL_CTX_load_verify_locations(ctx, caCertFile, 0),
+        WOLFSSL_SUCCESS);
+#ifdef WOLFSSL_ENCRYPTED_KEYS
+    wolfSSL_CTX_set_default_passwd_cb(ctx, PasswordCallBack);
+#endif
+#ifdef HAVE_SESSION_TICKET
+    /* Use session tickets, for ticket tests below */
+    ExpectIntEQ(wolfSSL_CTX_UseSessionTicket(ctx), WOLFSSL_SUCCESS);
+#endif
+
+    XMEMSET(&server_args, 0, sizeof(func_args));
+#ifdef WOLFSSL_TIRTOS
+    fdOpenSession(Task_self());
+#endif
+
+    StartTCP();
+    InitTcpReady(&ready);
+
+    server_args.signal = &ready;
+    start_thread(test_server_nofail, &server_args, &serverThread);
+    wait_tcp_ready(&server_args);
+
+    /* client connection */
+    ExpectNotNull(ssl = wolfSSL_new(ctx));
+    tcp_connect(&sockfd, wolfSSLIP, ready.port, 0, 0, ssl);
+    ExpectIntEQ(wolfSSL_set_fd(ssl, sockfd), WOLFSSL_SUCCESS);
+
+    WOLFSSL_ASYNC_WHILE_PENDING(ret = wolfSSL_connect(ssl),
+                                ret != WOLFSSL_SUCCESS);
+    ExpectIntEQ(ret, WOLFSSL_SUCCESS);
+
+    WOLFSSL_ASYNC_WHILE_PENDING(
+        ret = wolfSSL_write(ssl, sendGET, (int)XSTRLEN(sendGET)),
+        ret <= 0);
+    ExpectIntEQ(ret, (int)XSTRLEN(sendGET));
+
+    WOLFSSL_ASYNC_WHILE_PENDING(ret = wolfSSL_read(ssl, msg, sizeof(msg)),
+                                ret != 23);
+    ExpectIntEQ(ret, 23);
+
+    ExpectPtrNE((sess = wolfSSL_get1_session(ssl)), NULL); /* ref count 1 */
+    ExpectPtrNE((sess_copy = wolfSSL_get1_session(ssl)), NULL); /* ref count 2 */
+    ExpectIntEQ(wolfSSL_SessionIsSetup(sess), 1);
+#ifdef HAVE_EXT_CACHE
+    ExpectPtrEq(sess, sess_copy); /* they should be the same pointer but without
+                                   * HAVE_EXT_CACHE we get new objects each time */
+#endif
+    wolfSSL_SESSION_free(sess_copy); sess_copy = NULL;
+    wolfSSL_SESSION_free(sess);      sess = NULL; /* free session ref */
+
+    sess = wolfSSL_get_session(ssl);
+
+#ifdef OPENSSL_EXTRA
+    ExpectIntEQ(SSL_SESSION_is_resumable(NULL), 0);
+    ExpectIntEQ(SSL_SESSION_is_resumable(sess), 1);
+
+    ExpectIntEQ(wolfSSL_SESSION_has_ticket(NULL), 0);
+    ExpectIntEQ(wolfSSL_SESSION_get_ticket_lifetime_hint(NULL), 0);
+    #ifdef HAVE_SESSION_TICKET
+    ExpectIntEQ(wolfSSL_SESSION_has_ticket(sess), 1);
+    ExpectIntEQ(wolfSSL_SESSION_get_ticket_lifetime_hint(sess),
+                SESSION_TICKET_HINT_DEFAULT);
+    #else
+    ExpectIntEQ(wolfSSL_SESSION_has_ticket(sess), 0);
+    #endif
+#else
+    (void)sess;
+#endif /* OPENSSL_EXTRA */
+
+    /* Retain copy of the session for later testing */
+    ExpectNotNull(sess = wolfSSL_get1_session(ssl));
+
+    wolfSSL_shutdown(ssl);
+    wolfSSL_free(ssl); ssl = NULL;
+
+    CloseSocket(sockfd);
+
+    join_thread(serverThread);
+
+    FreeTcpReady(&ready);
+
+#ifdef WOLFSSL_TIRTOS
+    fdOpenSession(Task_self());
+#endif
+
+#if defined(SESSION_CERTS) && defined(OPENSSL_EXTRA)
+    {
+        X509 *x509 = NULL;
+        char buf[30];
+        int  bufSz = 0;
+
+        ExpectNotNull(x509 = SSL_SESSION_get0_peer(sess));
+        ExpectIntGT((bufSz = X509_NAME_get_text_by_NID(
+            X509_get_subject_name(x509), NID_organizationalUnitName, buf,
+            sizeof(buf))), 0);
+        ExpectIntNE((bufSz == 7 || bufSz == 16), 0); /* should be one of these*/
+        if (bufSz == 7) {
+            ExpectIntEQ(XMEMCMP(buf, "Support", bufSz), 0);
+        }
+        if (bufSz == 16) {
+            ExpectIntEQ(XMEMCMP(buf, "Programming-2048", bufSz), 0);
+        }
+    }
+#endif
+
+#ifdef HAVE_EXT_CACHE
+    ExpectNotNull(sess_copy = wolfSSL_SESSION_dup(sess));
+    wolfSSL_SESSION_free(sess_copy); sess_copy = NULL;
+    sess_copy = NULL;
+#endif
+
+#if defined(OPENSSL_EXTRA) && defined(HAVE_EXT_CACHE)
+    /* get session from DER and update the timeout */
+    ExpectIntEQ(wolfSSL_i2d_SSL_SESSION(NULL, &sessDer), WC_NO_ERR_TRACE(BAD_FUNC_ARG));
+    ExpectIntGT((sz = wolfSSL_i2d_SSL_SESSION(sess, &sessDer)), 0);
+    wolfSSL_SESSION_free(sess); sess = NULL;
+    sess = NULL;
+    ptr = sessDer;
+    ExpectNull(sess = wolfSSL_d2i_SSL_SESSION(NULL, NULL, sz));
+    ExpectNotNull(sess = wolfSSL_d2i_SSL_SESSION(NULL,
+                (const unsigned char**)&ptr, sz));
+    XFREE(sessDer, NULL, DYNAMIC_TYPE_OPENSSL);
+    sessDer = NULL;
+
+    ExpectIntGT(wolfSSL_SESSION_get_time(sess), 0);
+    ExpectIntEQ(wolfSSL_SSL_SESSION_set_timeout(sess, 500), SSL_SUCCESS);
+#endif
+
+    /* successful set session test */
+    ExpectNotNull(ssl = wolfSSL_new(ctx));
+    ExpectIntEQ(wolfSSL_set_session(ssl, sess), WOLFSSL_SUCCESS);
+
+#ifdef HAVE_SESSION_TICKET
+    /* Test set/get session ticket */
+    {
+        const char* ticket = "This is a session ticket";
+        char buf[64] = {0};
+        word32 bufSz = (word32)sizeof(buf);
+        word32 retSz = 0;
+
+        ExpectIntEQ(WOLFSSL_SUCCESS,
+            wolfSSL_set_SessionTicket(ssl, (byte *)ticket,
+                (word32)XSTRLEN(ticket)));
+        ExpectIntEQ(WOLFSSL_SUCCESS,
+            wolfSSL_get_SessionTicket(ssl, (byte *)buf, &bufSz));
+        ExpectStrEQ(ticket, buf);
+
+        /* return ticket length if buffer parameter is null */
+        wolfSSL_get_SessionTicket(ssl, NULL, &retSz);
+        ExpectIntEQ(bufSz, retSz);
+    }
+#endif
+
+#ifdef OPENSSL_EXTRA
+    /* session timeout case */
+    /* make the session to be expired */
+    ExpectIntEQ(SSL_SESSION_set_timeout(sess,1), SSL_SUCCESS);
+    XSLEEP_MS(1200);
+
+    /* SSL_set_session should reject specified session but return success
+     * if WOLFSSL_ERROR_CODE_OPENSSL macro is defined for OpenSSL compatibility.
+     */
+#if defined(WOLFSSL_ERROR_CODE_OPENSSL)
+    ExpectIntEQ(wolfSSL_set_session(ssl,sess), SSL_SUCCESS);
+#else
+    ExpectIntEQ(wolfSSL_set_session(ssl,sess), WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+#endif
+    ExpectIntEQ(wolfSSL_SSL_SESSION_set_timeout(sess, 500), SSL_SUCCESS);
+
+#ifdef WOLFSSL_SESSION_ID_CTX
+    /* fail case with miss match session context IDs (use compatibility API) */
+    ExpectIntEQ(SSL_set_session_id_context(ssl, context, contextSz),
+            SSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_set_session(ssl, sess), WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+    wolfSSL_free(ssl); ssl = NULL;
+
+    ExpectIntEQ(SSL_CTX_set_session_id_context(NULL, context, contextSz),
+            WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+    ExpectIntEQ(SSL_CTX_set_session_id_context(ctx, context, contextSz),
+            SSL_SUCCESS);
+    ExpectNotNull(ssl = wolfSSL_new(ctx));
+    ExpectIntEQ(wolfSSL_set_session(ssl, sess), WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+#endif
+#endif /* OPENSSL_EXTRA */
+
+    wolfSSL_free(ssl);
+    wolfSSL_SESSION_free(sess);
+    wolfSSL_CTX_free(ctx);
+#endif
+    return EXPECT_RESULT();
+}
+
+#if !defined(NO_FILESYSTEM) && !defined(NO_CERTS) && \
+    !defined(NO_RSA) && defined(HAVE_IO_TESTS_DEPENDENCIES) && \
+    !defined(NO_SESSION_CACHE) && defined(OPENSSL_EXTRA) && \
+    !defined(WOLFSSL_NO_TLS12)
+static WOLFSSL_SESSION* test_wolfSSL_SESSION_expire_sess = NULL;
+
+static void test_wolfSSL_SESSION_expire_downgrade_ctx_ready(WOLFSSL_CTX* ctx)
+{
+    #ifdef WOLFSSL_ERROR_CODE_OPENSSL
+    /* returns previous timeout value */
+    AssertIntEQ(wolfSSL_CTX_set_timeout(ctx, 1), 500);
+    #else
+    AssertIntEQ(wolfSSL_CTX_set_timeout(ctx, 1), WOLFSSL_SUCCESS);
+    #endif
+}
+
+
+/* set the session to timeout in a second */
+static void test_wolfSSL_SESSION_expire_downgrade_ssl_ready(WOLFSSL* ssl)
+{
+    AssertIntEQ(wolfSSL_set_timeout(ssl, 2), 1);
+}
+
+
+/* store the client side session from the first successful connection */
+static void test_wolfSSL_SESSION_expire_downgrade_ssl_result(WOLFSSL* ssl)
+{
+    AssertPtrNE((test_wolfSSL_SESSION_expire_sess = wolfSSL_get1_session(ssl)),
+        NULL); /* ref count 1 */
+}
+
+
+/* wait till session is expired then set it in the WOLFSSL struct for use */
+static void test_wolfSSL_SESSION_expire_downgrade_ssl_ready_wait(WOLFSSL* ssl)
+{
+    AssertIntEQ(wolfSSL_set_timeout(ssl, 1), 1);
+    AssertIntEQ(wolfSSL_set_session(ssl, test_wolfSSL_SESSION_expire_sess),
+        WOLFSSL_SUCCESS);
+    XSLEEP_MS(2000); /* wait 2 seconds for session to expire */
+}
+
+
+/* set expired session in the WOLFSSL struct for use */
+static void test_wolfSSL_SESSION_expire_downgrade_ssl_ready_set(WOLFSSL* ssl)
+{
+    XSLEEP_MS(1200); /* wait a second for session to expire */
+
+    /* set the expired session, call to set session fails but continuing on
+       after failure should be handled here */
+#if defined(OPENSSL_EXTRA) && defined(WOLFSSL_ERROR_CODE_OPENSSL)
+    AssertIntEQ(wolfSSL_set_session(ssl, test_wolfSSL_SESSION_expire_sess),
+        WOLFSSL_SUCCESS);
+#else
+    AssertIntNE(wolfSSL_set_session(ssl, test_wolfSSL_SESSION_expire_sess),
+        WOLFSSL_SUCCESS);
+#endif
+}
+
+
+/* check that the expired session was not reused */
+static void test_wolfSSL_SESSION_expire_downgrade_ssl_result_reuse(WOLFSSL* ssl)
+{
+    /* since the session has expired it should not have been reused */
+    AssertIntEQ(wolfSSL_session_reused(ssl), 0);
+}
+#endif
+
+int test_wolfSSL_SESSION_expire_downgrade(void)
+{
+    EXPECT_DECLS;
+#if !defined(NO_FILESYSTEM) && !defined(NO_CERTS) && \
+    !defined(NO_RSA) && defined(HAVE_IO_TESTS_DEPENDENCIES) && \
+    !defined(NO_SESSION_CACHE) && defined(OPENSSL_EXTRA) && \
+    !defined(WOLFSSL_NO_TLS12)
+    callback_functions server_cbf, client_cbf;
+
+    XMEMSET(&server_cbf, 0, sizeof(callback_functions));
+    XMEMSET(&client_cbf, 0, sizeof(callback_functions));
+
+    /* force server side to use TLS 1.2 */
+    server_cbf.method = wolfTLSv1_2_server_method;
+
+    client_cbf.method = wolfSSLv23_client_method;
+    server_cbf.ctx_ready = test_wolfSSL_SESSION_expire_downgrade_ctx_ready;
+    client_cbf.ssl_ready = test_wolfSSL_SESSION_expire_downgrade_ssl_ready;
+    client_cbf.on_result = test_wolfSSL_SESSION_expire_downgrade_ssl_result;
+
+    test_wolfSSL_client_server_nofail(&client_cbf, &server_cbf);
+    ExpectIntEQ(client_cbf.return_code, TEST_SUCCESS);
+    ExpectIntEQ(server_cbf.return_code, TEST_SUCCESS);
+
+    client_cbf.method = wolfSSLv23_client_method;
+    server_cbf.ctx_ready = test_wolfSSL_SESSION_expire_downgrade_ctx_ready;
+    client_cbf.ssl_ready = test_wolfSSL_SESSION_expire_downgrade_ssl_ready_wait;
+    client_cbf.on_result =
+        test_wolfSSL_SESSION_expire_downgrade_ssl_result_reuse;
+
+    test_wolfSSL_client_server_nofail(&client_cbf, &server_cbf);
+    ExpectIntEQ(client_cbf.return_code, TEST_SUCCESS);
+    ExpectIntEQ(server_cbf.return_code, TEST_SUCCESS);
+
+    client_cbf.method = wolfSSLv23_client_method;
+    server_cbf.ctx_ready = test_wolfSSL_SESSION_expire_downgrade_ctx_ready;
+    client_cbf.ssl_ready = test_wolfSSL_SESSION_expire_downgrade_ssl_ready_set;
+    client_cbf.on_result =
+        test_wolfSSL_SESSION_expire_downgrade_ssl_result_reuse;
+
+    test_wolfSSL_client_server_nofail(&client_cbf, &server_cbf);
+    ExpectIntEQ(client_cbf.return_code, TEST_SUCCESS);
+    ExpectIntEQ(server_cbf.return_code, TEST_SUCCESS);
+
+    wolfSSL_SESSION_free(test_wolfSSL_SESSION_expire_sess);
+#endif
+    return EXPECT_RESULT();
+}
+
+#if defined(OPENSSL_EXTRA) && defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES) && \
+    defined(HAVE_EX_DATA) && !defined(NO_SESSION_CACHE)
+#ifdef WOLFSSL_ATOMIC_OPS
+    typedef wolfSSL_Atomic_Int SessRemCounter_t;
+#else
+    typedef int SessRemCounter_t;
+#endif
+static SessRemCounter_t clientSessRemCountMalloc;
+static SessRemCounter_t serverSessRemCountMalloc;
+static SessRemCounter_t clientSessRemCountFree;
+static SessRemCounter_t serverSessRemCountFree;
+
+static WOLFSSL_CTX* serverSessCtx = NULL;
+static WOLFSSL_SESSION* serverSess = NULL;
+#if (defined(WOLFSSL_TLS13) && defined(HAVE_SESSION_TICKET)) || \
+        !defined(NO_SESSION_CACHE_REF)
+static WOLFSSL_CTX* clientSessCtx = NULL;
+static WOLFSSL_SESSION* clientSess = NULL;
+#endif
+static int serverSessRemIdx = 3;
+static int sessRemCtx_Server = WOLFSSL_SERVER_END;
+static int sessRemCtx_Client = WOLFSSL_CLIENT_END;
+
+static void SessRemCtxCb(WOLFSSL_CTX *ctx, WOLFSSL_SESSION *sess)
+{
+    int* side;
+
+    (void)ctx;
+
+    side = (int*)SSL_SESSION_get_ex_data(sess, serverSessRemIdx);
+    if (side != NULL) {
+        if (*side == WOLFSSL_CLIENT_END)
+            (void)wolfSSL_Atomic_Int_FetchAdd(&clientSessRemCountFree, 1);
+        else
+            (void)wolfSSL_Atomic_Int_FetchAdd(&serverSessRemCountFree, 1);
+
+        SSL_SESSION_set_ex_data(sess, serverSessRemIdx, NULL);
+    }
+}
+
+static int SessRemCtxSetupCb(WOLFSSL_CTX* ctx)
+{
+    SSL_CTX_sess_set_remove_cb(ctx, SessRemCtxCb);
+#if defined(WOLFSSL_TLS13) && !defined(HAVE_SESSION_TICKET) && \
+        !defined(NO_SESSION_CACHE_REF)
+    {
+        EXPECT_DECLS;
+        /* Allow downgrade, set min version, and disable TLS 1.3.
+         * Do this because without NO_SESSION_CACHE_REF we will want to return a
+         * reference to the session cache. But with WOLFSSL_TLS13 and without
+         * HAVE_SESSION_TICKET we won't have a session ID to be able to place
+         * the session in the cache. In this case we need to downgrade to
+         * previous versions to just use the legacy session ID field. */
+        ExpectIntEQ(SSL_CTX_set_min_proto_version(ctx, SSL3_VERSION),
+            SSL_SUCCESS);
+        ExpectIntEQ(SSL_CTX_set_max_proto_version(ctx, TLS1_2_VERSION),
+            SSL_SUCCESS);
+        return EXPECT_RESULT();
+    }
+#else
+    return TEST_SUCCESS;
+#endif
+}
+
+static int SessRemSslSetupCb(WOLFSSL* ssl)
+{
+    EXPECT_DECLS;
+    int* side;
+
+    if (SSL_is_server(ssl)) {
+        side = &sessRemCtx_Server;
+        (void)wolfSSL_Atomic_Int_FetchAdd(&serverSessRemCountMalloc, 1);
+        ExpectNotNull(serverSess = SSL_get1_session(ssl));
+        ExpectIntEQ(SSL_CTX_up_ref(serverSessCtx = SSL_get_SSL_CTX(ssl)),
+                SSL_SUCCESS);
+    }
+    else {
+        side = &sessRemCtx_Client;
+        (void)wolfSSL_Atomic_Int_FetchAdd(&clientSessRemCountMalloc, 1);
+#if (defined(WOLFSSL_TLS13) && defined(HAVE_SESSION_TICKET)) || \
+    !defined(NO_SESSION_CACHE_REF)
+        ExpectNotNull(clientSess = SSL_get1_session(ssl));
+        ExpectIntEQ(SSL_CTX_up_ref(clientSessCtx = SSL_get_SSL_CTX(ssl)),
+                SSL_SUCCESS);
+#endif
+    }
+    ExpectIntEQ(SSL_SESSION_set_ex_data(SSL_get_session(ssl),
+        serverSessRemIdx, side), SSL_SUCCESS);
+
+    return EXPECT_RESULT();
+}
+#endif
+
+int test_wolfSSL_CTX_sess_set_remove_cb(void)
+{
+    EXPECT_DECLS;
+#if defined(OPENSSL_EXTRA) && defined(HAVE_SSL_MEMIO_TESTS_DEPENDENCIES) && \
+    defined(HAVE_EX_DATA) && !defined(NO_SESSION_CACHE)
+    /* Check that the remove callback gets called for external data in a
+     * session object */
+    test_ssl_cbf func_cb;
+
+    wolfSSL_Atomic_Int_Init(&clientSessRemCountMalloc, 0);
+    wolfSSL_Atomic_Int_Init(&serverSessRemCountMalloc, 0);
+    wolfSSL_Atomic_Int_Init(&clientSessRemCountFree, 0);
+    wolfSSL_Atomic_Int_Init(&serverSessRemCountFree, 0);
+
+    XMEMSET(&func_cb, 0, sizeof(func_cb));
+    func_cb.ctx_ready = SessRemCtxSetupCb;
+    func_cb.on_result = SessRemSslSetupCb;
+
+    ExpectIntEQ(test_wolfSSL_client_server_nofail_memio(&func_cb, &func_cb,
+        NULL), TEST_SUCCESS);
+
+    /* Both should have been allocated */
+    ExpectIntEQ(clientSessRemCountMalloc, 1);
+    ExpectIntEQ(serverSessRemCountMalloc, 1);
+
+    /* This should not be called yet. Session wasn't evicted from cache yet. */
+    ExpectIntEQ(clientSessRemCountFree, 0);
+#if (defined(WOLFSSL_TLS13) && defined(HAVE_SESSION_TICKET)) || \
+        !defined(NO_SESSION_CACHE_REF)
+    /* Force a cache lookup */
+    ExpectNotNull(SSL_SESSION_get_ex_data(clientSess, serverSessRemIdx));
+    /* Force a cache update */
+    ExpectNotNull(SSL_SESSION_set_ex_data(clientSess, serverSessRemIdx - 1, 0));
+    /* This should set the timeout to 0 and call the remove callback from within
+     * the session cache. Returns 1 per OpenSSL semantics (session was
+     * present in the cache and removed). */
+    ExpectIntEQ(SSL_CTX_remove_session(clientSessCtx, clientSess), 1);
+    ExpectNull(SSL_SESSION_get_ex_data(clientSess, serverSessRemIdx));
+    ExpectIntEQ(clientSessRemCountFree, 1);
+#endif
+    /* Server session is in the cache so ex_data isn't free'd with the SSL
+     * object */
+    ExpectIntEQ(serverSessRemCountFree, 0);
+    /* Force a cache lookup */
+    ExpectNotNull(SSL_SESSION_get_ex_data(serverSess, serverSessRemIdx));
+    /* Force a cache update */
+    ExpectNotNull(SSL_SESSION_set_ex_data(serverSess, serverSessRemIdx - 1, 0));
+    /* This should set the timeout to 0 and call the remove callback from within
+     * the session cache. Returns 1 per OpenSSL semantics (session was
+     * present in the cache and removed). */
+    ExpectIntEQ(SSL_CTX_remove_session(serverSessCtx, serverSess), 1);
+    ExpectNull(SSL_SESSION_get_ex_data(serverSess, serverSessRemIdx));
+    ExpectIntEQ(serverSessRemCountFree, 1);
+    /* Need to free the references that we kept */
+    SSL_CTX_free(serverSessCtx);
+    SSL_SESSION_free(serverSess);
+#if (defined(WOLFSSL_TLS13) && defined(HAVE_SESSION_TICKET)) || \
+        !defined(NO_SESSION_CACHE_REF)
+    SSL_CTX_free(clientSessCtx);
+    SSL_SESSION_free(clientSess);
+#endif
+#endif
+    return EXPECT_RESULT();
+}
+
+int test_wolfSSL_ticket_keys(void)
+{
+    EXPECT_DECLS;
+#if defined(HAVE_SESSION_TICKET) && !defined(WOLFSSL_NO_DEF_TICKET_ENC_CB) && \
+    !defined(NO_WOLFSSL_SERVER) && !defined(NO_TLS)
+    WOLFSSL_CTX* ctx = NULL;
+    byte keys[WOLFSSL_TICKET_KEYS_SZ];
+
+    ExpectNotNull(ctx = wolfSSL_CTX_new(wolfSSLv23_server_method()));
+
+    ExpectIntEQ(wolfSSL_CTX_get_tlsext_ticket_keys(NULL, NULL, 0),
+                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+    ExpectIntEQ(wolfSSL_CTX_get_tlsext_ticket_keys(ctx, NULL, 0),
+                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+    ExpectIntEQ(wolfSSL_CTX_get_tlsext_ticket_keys(ctx, keys, 0),
+                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+    ExpectIntEQ(wolfSSL_CTX_get_tlsext_ticket_keys(NULL, keys, 0),
+                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+    ExpectIntEQ(wolfSSL_CTX_get_tlsext_ticket_keys(NULL, NULL, sizeof(keys)),
+                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+    ExpectIntEQ(wolfSSL_CTX_get_tlsext_ticket_keys(ctx, NULL, sizeof(keys)),
+                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+    ExpectIntEQ(wolfSSL_CTX_get_tlsext_ticket_keys(NULL, keys, sizeof(keys)),
+                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+
+    ExpectIntEQ(wolfSSL_CTX_set_tlsext_ticket_keys(NULL, NULL, 0),
+                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+    ExpectIntEQ(wolfSSL_CTX_set_tlsext_ticket_keys(ctx, NULL, 0),
+                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+    ExpectIntEQ(wolfSSL_CTX_set_tlsext_ticket_keys(ctx, keys, 0),
+                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+    ExpectIntEQ(wolfSSL_CTX_set_tlsext_ticket_keys(NULL, keys, 0),
+                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+    ExpectIntEQ(wolfSSL_CTX_set_tlsext_ticket_keys(NULL, NULL, sizeof(keys)),
+                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+    ExpectIntEQ(wolfSSL_CTX_set_tlsext_ticket_keys(ctx, NULL, sizeof(keys)),
+                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+    ExpectIntEQ(wolfSSL_CTX_set_tlsext_ticket_keys(NULL, keys, sizeof(keys)),
+                WC_NO_ERR_TRACE(WOLFSSL_FAILURE));
+
+    ExpectIntEQ(wolfSSL_CTX_get_tlsext_ticket_keys(ctx, keys, sizeof(keys)),
+                WOLFSSL_SUCCESS);
+    ExpectIntEQ(wolfSSL_CTX_set_tlsext_ticket_keys(ctx, keys, sizeof(keys)),
+                WOLFSSL_SUCCESS);
+
+    wolfSSL_CTX_free(ctx);
+#endif
+    return EXPECT_RESULT();
+}
+
+/*----------------------------------------------------------------------------*/
+/* SESSION ex_data new index                                                  */
+/*----------------------------------------------------------------------------*/
+
+#if defined(HAVE_EX_DATA_CRYPTO) && defined(OPENSSL_EXTRA)
+
+#define SESSION_NEW_IDX_LONG 0xDEADBEEF
+#define SESSION_NEW_IDX_VAL  ((void*)0xAEADAEAD)
+#define SESSION_DUP_IDX_VAL  ((void*)0xDEDEDEDE)
+#define SESSION_NEW_IDX_PTR  "Testing"
+
+static void test_wolfSSL_SESSION_get_ex_new_index_new_cb(void* p, void* ptr,
+        CRYPTO_EX_DATA* a, int idx, long argValue, void* arg)
+{
+    AssertNotNull(p);
+    AssertNull(ptr);
+    AssertIntEQ(CRYPTO_set_ex_data(a, idx, SESSION_NEW_IDX_VAL), SSL_SUCCESS);
+    AssertIntEQ(argValue, SESSION_NEW_IDX_LONG);
+    AssertStrEQ(arg, SESSION_NEW_IDX_PTR);
+}
+
+static int test_wolfSSL_SESSION_get_ex_new_index_dup_cb(CRYPTO_EX_DATA* out,
+        const CRYPTO_EX_DATA* in, void* inPtr, int idx, long argV,
+        void* arg)
+{
+    EXPECT_DECLS;
+
+    ExpectNotNull(out);
+    ExpectNotNull(in);
+    ExpectPtrEq(*(void**)inPtr, SESSION_NEW_IDX_VAL);
+    ExpectPtrEq(CRYPTO_get_ex_data(in, idx), SESSION_NEW_IDX_VAL);
+    ExpectPtrEq(CRYPTO_get_ex_data(out, idx), SESSION_NEW_IDX_VAL);
+    ExpectIntEQ(argV, SESSION_NEW_IDX_LONG);
+    ExpectStrEQ(arg, SESSION_NEW_IDX_PTR);
+    *(void**)inPtr = SESSION_DUP_IDX_VAL;
+    if (EXPECT_SUCCESS()) {
+        return SSL_SUCCESS;
+    }
+    else {
+        return SSL_FAILURE;
+    }
+}
+
+static int test_wolfSSL_SESSION_get_ex_new_index_free_cb_called = 0;
+static void test_wolfSSL_SESSION_get_ex_new_index_free_cb(void* p, void* ptr,
+        CRYPTO_EX_DATA* a, int idx, long argValue, void* arg)
+{
+    EXPECT_DECLS;
+
+    ExpectNotNull(p);
+    ExpectNull(ptr);
+    ExpectPtrNE(CRYPTO_get_ex_data(a, idx), 0);
+    ExpectIntEQ(argValue, SESSION_NEW_IDX_LONG);
+    ExpectStrEQ(arg, SESSION_NEW_IDX_PTR);
+    if (EXPECT_SUCCESS()) {
+        test_wolfSSL_SESSION_get_ex_new_index_free_cb_called++;
+    }
+}
+
+int test_wolfSSL_SESSION_get_ex_new_index(void)
+{
+    EXPECT_DECLS;
+    int idx = SSL_SESSION_get_ex_new_index(SESSION_NEW_IDX_LONG,
+                (void*)SESSION_NEW_IDX_PTR,
+                test_wolfSSL_SESSION_get_ex_new_index_new_cb,
+                test_wolfSSL_SESSION_get_ex_new_index_dup_cb,
+                test_wolfSSL_SESSION_get_ex_new_index_free_cb);
+    SSL_SESSION* s = SSL_SESSION_new();
+    SSL_SESSION* d = NULL;
+
+    ExpectNotNull(s);
+    ExpectPtrEq(SSL_SESSION_get_ex_data(s, idx), SESSION_NEW_IDX_VAL);
+    ExpectNotNull(d = SSL_SESSION_dup(s));
+    ExpectPtrEq(SSL_SESSION_get_ex_data(d, idx), SESSION_DUP_IDX_VAL);
+    SSL_SESSION_free(s);
+    ExpectIntEQ(test_wolfSSL_SESSION_get_ex_new_index_free_cb_called, 1);
+    SSL_SESSION_free(d);
+    ExpectIntEQ(test_wolfSSL_SESSION_get_ex_new_index_free_cb_called, 2);
+
+    crypto_ex_cb_free(crypto_ex_cb_ctx_session);
+    crypto_ex_cb_ctx_session = NULL;
+    return EXPECT_RESULT();
+}
+#else
+int test_wolfSSL_SESSION_get_ex_new_index(void)
+{
+    return TEST_SKIPPED;
+}
+#endif
diff --git a/tests/api/test_session.h b/tests/api/test_session.h
new file mode 100644
index 0000000000..9dc2e258f1
--- /dev/null
+++ b/tests/api/test_session.h
@@ -0,0 +1,54 @@
+/* test_session.h
+ *
+ * Copyright (C) 2006-2026 wolfSSL Inc.
+ *
+ * This file is part of wolfSSL.
+ *
+ * wolfSSL is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * wolfSSL is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1335, USA
+ */
+
+#ifndef WOLFCRYPT_TEST_SESSION_H
+#define WOLFCRYPT_TEST_SESSION_H
+
+#include <tests/api/api_decl.h>
+
+int test_wolfSSL_CTX_add_session(void);
+int test_wolfSSL_CTX_add_session_ext_tls13(void);
+int test_wolfSSL_CTX_add_session_ext_dtls13(void);
+int test_wolfSSL_CTX_add_session_ext_tls12(void);
+int test_wolfSSL_CTX_add_session_ext_dtls12(void);
+int test_wolfSSL_CTX_add_session_ext_tls11(void);
+int test_wolfSSL_CTX_add_session_ext_dtls1(void);
+int test_wolfSSL_SESSION(void);
+int test_wolfSSL_SESSION_expire_downgrade(void);
+int test_wolfSSL_CTX_sess_set_remove_cb(void);
+int test_wolfSSL_ticket_keys(void);
+int test_wolfSSL_SESSION_get_ex_new_index(void);
+
+#define TEST_SESSION_DECLS                                                     \
+    TEST_DECL_GROUP("session", test_wolfSSL_CTX_add_session),                  \
+    TEST_DECL_GROUP("session", test_wolfSSL_CTX_add_session_ext_tls13),        \
+    TEST_DECL_GROUP("session", test_wolfSSL_CTX_add_session_ext_dtls13),       \
+    TEST_DECL_GROUP("session", test_wolfSSL_CTX_add_session_ext_tls12),        \
+    TEST_DECL_GROUP("session", test_wolfSSL_CTX_add_session_ext_dtls12),       \
+    TEST_DECL_GROUP("session", test_wolfSSL_CTX_add_session_ext_tls11),        \
+    TEST_DECL_GROUP("session", test_wolfSSL_CTX_add_session_ext_dtls1),        \
+    TEST_DECL_GROUP("session", test_wolfSSL_SESSION),                          \
+    TEST_DECL_GROUP("session", test_wolfSSL_SESSION_expire_downgrade),         \
+    TEST_DECL_GROUP("session", test_wolfSSL_CTX_sess_set_remove_cb),           \
+    TEST_DECL_GROUP("session", test_wolfSSL_ticket_keys),                      \
+    TEST_DECL_GROUP("session", test_wolfSSL_SESSION_get_ex_new_index)
+
+#endif /* WOLFCRYPT_TEST_SESSION_H */
diff --git a/tests/api/test_tls13.c b/tests/api/test_tls13.c
index 8684d1bd96..ca37fb48cc 100644
--- a/tests/api/test_tls13.c
+++ b/tests/api/test_tls13.c
@@ -6310,3 +6310,216 @@ int test_tls13_cipher_fuzz_aes128_ccm_8_sha256(void)
 #endif
     return EXPECT_RESULT();
 }
+
+/* Regression test for the AEAD record-protection limit constants in
+ * internal.h. The macros expand to w64From32(hi, lo). A prior version split
+ * the intended 32-bit constants into 16-bit halves and passed each half as
+ * a separate 32-bit argument, producing a 64-bit value many orders of
+ * magnitude larger than RFC 8446 / RFC 9147 require. That made
+ * CheckTLS13AEADSendLimit's key-update trigger effectively unreachable.
+ * Compare against the hard-coded spec values so a recurrence is caught even
+ * if the macro is reused on both sides of the comparison. */
+int test_tls13_AEAD_limit_macros(void)
+{
+    EXPECT_DECLS;
+#if defined(WOLFSSL_TLS13) && !defined(WOLFSSL_TLS13_IGNORE_AEAD_LIMITS)
+    w64wrapper limit;
+
+    /* RFC 8446 5.5: 2^24.5 ~= 23726566 (0x016A09E6). */
+    limit = AEAD_AES_LIMIT;
+    ExpectIntEQ(w64GetHigh32(limit), 0);
+    ExpectIntEQ(w64GetLow32(limit),  0x016A09E6);
+
+#ifdef WOLFSSL_DTLS13
+    /* RFC 9147 (AES-CCM integrity): 2^23.5 ~= 11863283 (0x00B504F3). */
+    limit = DTLS_AEAD_AES_CCM_FAIL_LIMIT;
+    ExpectIntEQ(w64GetHigh32(limit), 0);
+    ExpectIntEQ(w64GetLow32(limit),  0x00B504F3);
+
+    /* Key-update threshold is half the fail limit: 5931641 (0x005A8279). */
+    limit = DTLS_AEAD_AES_CCM_FAIL_KU_LIMIT;
+    ExpectIntEQ(w64GetHigh32(limit), 0);
+    ExpectIntEQ(w64GetLow32(limit),  0x005A8279);
+#endif
+#endif
+    return EXPECT_RESULT();
+}
+
+#if defined(WOLFSSL_TLS13) && !defined(WOLFSSL_TLS13_IGNORE_AEAD_LIMITS) && \
+    !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER) && \
+    defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && \
+    (defined(BUILD_TLS_AES_128_GCM_SHA256) || \
+     defined(BUILD_TLS_AES_256_GCM_SHA384) || \
+     defined(BUILD_TLS_AES_128_CCM_SHA256) || \
+     defined(BUILD_TLS_AES_128_CCM_8_SHA256))
+/* Drive the client's encrypt sequence number towards the spec limit for
+ * `suite` and verify CheckTLS13AEADSendLimit's KeyUpdate trigger fires at
+ * exactly the right boundary.
+ *
+ * Two writes are exercised:
+ *   1. Counter set to limit - 2. After the write the counter must read
+ *      limit - 1 (record incremented it by 1) and no KeyUpdate must have
+ *      been emitted. CheckTLS13AEADSendLimit uses `seq >= limit`, so neither
+ *      the pre-send check nor the trailing loop check (which runs once more
+ *      after the last record before wolfSSL_write exits) is allowed to fire.
+ *   2. A second write follows with the counter already sitting at limit - 1
+ *      from the previous record. The user record goes out at seq = limit-1,
+ *      which bumps the counter to limit; the trailing limit check then
+ *      fires SendTls13KeyUpdate. SetKeysSide zeroes the encrypt counter, so
+ *      the post-write counter is 0.
+ *
+ * With the previous broken AEAD-limit macros the limit was unreachable, no
+ * KeyUpdate would ever fire, and the counter would simply advance to
+ * limit_lo + 1 in the second case instead of being reset.
+ *
+ * The AEAD nonce mixes in the record sequence number on both sides, so the
+ * server's decrypt counter has to be advanced in lockstep with the client's
+ * encrypt counter or the record fails the integrity check. */
+static int test_tls13_AEAD_limit_triggers_KeyUpdate_cs(const char* suite,
+        word32 limit_hi, word32 limit_lo, int expected_bulk_cipher)
+{
+    EXPECT_DECLS;
+    struct test_memio_ctx test_ctx;
+    WOLFSSL_CTX *ctx_c = NULL, *ctx_s = NULL;
+    WOLFSSL *ssl_c = NULL, *ssl_s = NULL;
+    const char msg[] = "post-limit-record";
+    char buf[sizeof(msg)];
+    int written;
+
+    XMEMSET(&test_ctx, 0, sizeof(test_ctx));
+    test_ctx.c_ciphers = suite;
+    test_ctx.s_ciphers = suite;
+
+    ExpectIntEQ(test_memio_setup(&test_ctx, &ctx_c, &ctx_s, &ssl_c, &ssl_s,
+            wolfTLSv1_3_client_method, wolfTLSv1_3_server_method), 0);
+    ExpectIntEQ(test_memio_do_handshake(ssl_c, ssl_s, 10, NULL), 0);
+
+    if (EXPECT_SUCCESS() && ssl_c != NULL && ssl_s != NULL) {
+        /* Sanity check: the negotiated bulk cipher matches what the caller
+         * intends to exercise. If a build flag combination falls through to
+         * a different suite, the limit constant would be wrong. */
+        ExpectIntEQ(ssl_c->specs.bulk_cipher_algorithm, expected_bulk_cipher);
+
+        /* Stage the counters two below the limit so the first write stays
+         * comfortably below the trigger threshold. */
+        ssl_c->keys.sequence_number_hi      = limit_hi;
+        ssl_c->keys.sequence_number_lo      = limit_lo - 2;
+        ssl_s->keys.peer_sequence_number_hi = limit_hi;
+        ssl_s->keys.peer_sequence_number_lo = limit_lo - 2;
+    }
+
+    /* First write: below the limit, no KeyUpdate expected. */
+    written = wolfSSL_write(ssl_c, msg, (int)sizeof(msg));
+    ExpectIntEQ(written, (int)sizeof(msg));
+
+    if (EXPECT_SUCCESS() && ssl_c != NULL) {
+        /* The record bumped the counter from limit-2 to limit-1. A
+         * KeyUpdate would have zeroed it via SetKeysSide and bumped to 1. */
+        ExpectIntEQ((int)ssl_c->keys.sequence_number_hi, (int)limit_hi);
+        ExpectIntEQ(ssl_c->keys.sequence_number_lo,      limit_lo - 1);
+    }
+
+    /* Server consumes the below-limit record with its existing keys. */
+    XMEMSET(buf, 0, sizeof(buf));
+    ExpectIntEQ(wolfSSL_read(ssl_s, buf, (int)sizeof(buf)), (int)sizeof(msg));
+    ExpectIntEQ(XMEMCMP(buf, msg, sizeof(msg)), 0);
+
+    /* Second write: the client's counter is now at limit-1. Sending this
+     * record will push it to limit, at which point the trailing check
+     * inside SendData's loop fires SendTls13KeyUpdate. No manual counter
+     * adjustment is needed -- the counter is allowed to "naturally" reach
+     * the limit through the previous send. */
+    written = wolfSSL_write(ssl_c, msg, (int)sizeof(msg));
+    ExpectIntEQ(written, (int)sizeof(msg));
+
+    if (EXPECT_SUCCESS() && ssl_c != NULL) {
+        /* SendTls13KeyUpdate -> DeriveTls13Keys -> SetKeysSide zeroes the
+         * encrypt sequence number. The user record went out before the
+         * trigger fired, so no record was sent on the new keys. */
+        ExpectIntEQ((int)ssl_c->keys.sequence_number_hi, 0);
+        ExpectIntEQ((int)ssl_c->keys.sequence_number_lo, 0);
+    }
+
+    /* The server reads the user record (sent under the pre-update keys at
+     * seq = limit - 1) before it sees the KeyUpdate record. The KeyUpdate
+     * is consumed transparently on a subsequent read; for the test we just
+     * need to confirm the user data round-trips. */
+    XMEMSET(buf, 0, sizeof(buf));
+    {
+        int r = -1, attempts;
+        for (attempts = 0; attempts < 5; attempts++) {
+            r = wolfSSL_read(ssl_s, buf, (int)sizeof(buf));
+            if (r > 0)
+                break;
+            if (wolfSSL_get_error(ssl_s, r) != WOLFSSL_ERROR_WANT_READ)
+                break;
+        }
+        ExpectIntEQ(r, (int)sizeof(msg));
+    }
+    ExpectIntEQ(XMEMCMP(buf, msg, sizeof(msg)), 0);
+
+    wolfSSL_free(ssl_c);
+    wolfSSL_free(ssl_s);
+    wolfSSL_CTX_free(ctx_c);
+    wolfSSL_CTX_free(ctx_s);
+
+    return EXPECT_RESULT();
+}
+#endif
+
+int test_tls13_AEAD_limit_KU_aes128_gcm_sha256(void)
+{
+    EXPECT_DECLS;
+#if defined(WOLFSSL_TLS13) && !defined(WOLFSSL_TLS13_IGNORE_AEAD_LIMITS) && \
+    !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER) && \
+    defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && \
+    defined(BUILD_TLS_AES_128_GCM_SHA256)
+    ExpectIntEQ(test_tls13_AEAD_limit_triggers_KeyUpdate_cs(
+            "TLS13-AES128-GCM-SHA256", 0, 0x016A09E6, wolfssl_aes_gcm),
+            TEST_SUCCESS);
+#endif
+    return EXPECT_RESULT();
+}
+
+int test_tls13_AEAD_limit_KU_aes256_gcm_sha384(void)
+{
+    EXPECT_DECLS;
+#if defined(WOLFSSL_TLS13) && !defined(WOLFSSL_TLS13_IGNORE_AEAD_LIMITS) && \
+    !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER) && \
+    defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && \
+    defined(BUILD_TLS_AES_256_GCM_SHA384)
+    ExpectIntEQ(test_tls13_AEAD_limit_triggers_KeyUpdate_cs(
+            "TLS13-AES256-GCM-SHA384", 0, 0x016A09E6, wolfssl_aes_gcm),
+            TEST_SUCCESS);
+#endif
+    return EXPECT_RESULT();
+}
+
+int test_tls13_AEAD_limit_KU_aes128_ccm_sha256(void)
+{
+    EXPECT_DECLS;
+#if defined(WOLFSSL_TLS13) && !defined(WOLFSSL_TLS13_IGNORE_AEAD_LIMITS) && \
+    !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER) && \
+    defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && \
+    defined(BUILD_TLS_AES_128_CCM_SHA256)
+    ExpectIntEQ(test_tls13_AEAD_limit_triggers_KeyUpdate_cs(
+            "TLS13-AES128-CCM-SHA256", 0, 0x016A09E6, wolfssl_aes_ccm),
+            TEST_SUCCESS);
+#endif
+    return EXPECT_RESULT();
+}
+
+int test_tls13_AEAD_limit_KU_aes128_ccm_8_sha256(void)
+{
+    EXPECT_DECLS;
+#if defined(WOLFSSL_TLS13) && !defined(WOLFSSL_TLS13_IGNORE_AEAD_LIMITS) && \
+    !defined(NO_WOLFSSL_CLIENT) && !defined(NO_WOLFSSL_SERVER) && \
+    defined(HAVE_MANUAL_MEMIO_TESTS_DEPENDENCIES) && \
+    defined(BUILD_TLS_AES_128_CCM_8_SHA256)
+    ExpectIntEQ(test_tls13_AEAD_limit_triggers_KeyUpdate_cs(
+            "TLS13-AES128-CCM-8-SHA256", 0, 0x016A09E6, wolfssl_aes_ccm),
+            TEST_SUCCESS);
+#endif
+    return EXPECT_RESULT();
+}
+
diff --git a/tests/api/test_tls13.h b/tests/api/test_tls13.h
index 2a04b9069a..e01527f0e0 100644
--- a/tests/api/test_tls13.h
+++ b/tests/api/test_tls13.h
@@ -76,6 +76,11 @@ int test_tls13_cipher_fuzz_aes256_gcm_sha384(void);
 int test_tls13_cipher_fuzz_chacha20_poly1305_sha256(void);
 int test_tls13_cipher_fuzz_aes128_ccm_sha256(void);
 int test_tls13_cipher_fuzz_aes128_ccm_8_sha256(void);
+int test_tls13_AEAD_limit_macros(void);
+int test_tls13_AEAD_limit_KU_aes128_gcm_sha256(void);
+int test_tls13_AEAD_limit_KU_aes256_gcm_sha384(void);
+int test_tls13_AEAD_limit_KU_aes128_ccm_sha256(void);
+int test_tls13_AEAD_limit_KU_aes128_ccm_8_sha256(void);
 
 #define TEST_TLS13_DECLS                                        \
     TEST_DECL_GROUP("tls13", test_tls13_apis),                  \
@@ -129,6 +134,11 @@ int test_tls13_cipher_fuzz_aes128_ccm_8_sha256(void);
     TEST_DECL_GROUP("tls13", test_tls13_cipher_fuzz_aes256_gcm_sha384), \
     TEST_DECL_GROUP("tls13", test_tls13_cipher_fuzz_chacha20_poly1305_sha256), \
     TEST_DECL_GROUP("tls13", test_tls13_cipher_fuzz_aes128_ccm_sha256), \
-    TEST_DECL_GROUP("tls13", test_tls13_cipher_fuzz_aes128_ccm_8_sha256)
+    TEST_DECL_GROUP("tls13", test_tls13_cipher_fuzz_aes128_ccm_8_sha256), \
+    TEST_DECL_GROUP("tls13", test_tls13_AEAD_limit_macros), \
+    TEST_DECL_GROUP("tls13", test_tls13_AEAD_limit_KU_aes128_gcm_sha256), \
+    TEST_DECL_GROUP("tls13", test_tls13_AEAD_limit_KU_aes256_gcm_sha384), \
+    TEST_DECL_GROUP("tls13", test_tls13_AEAD_limit_KU_aes128_ccm_sha256), \
+    TEST_DECL_GROUP("tls13", test_tls13_AEAD_limit_KU_aes128_ccm_8_sha256)
 
 #endif /* WOLFCRYPT_TEST_TLS13_H */
diff --git a/tests/utils.h b/tests/utils.h
index 8a7942086d..bd9a150e01 100644
--- a/tests/utils.h
+++ b/tests/utils.h
@@ -87,6 +87,15 @@ int test_memio_modify_message_len(struct test_memio_ctx *ctx, int client, int ms
 int test_memio_remove_from_buffer(struct test_memio_ctx *ctx, int client, int off, int sz);
 #endif
 
+/* Shared TLS server/client thread bodies, defined in tests/api.c. The
+ * definitions are gated on ENABLE_TLS_CALLBACK_TEST (a composite condition
+ * locally #defined inside api.c) or (WOLFSSL_DTLS && WOLFSSL_SESSION_EXPORT).
+ * Declared unconditionally here so api.c itself sees the prototype regardless
+ * of which side of the local #define triggers; absent the definition the
+ * prototypes are harmless and any caller would get a link error. */
+THREAD_RETURN WOLFSSL_THREAD run_wolfssl_server(void* args);
+void run_wolfssl_client(void* args);
+
 #if !defined(NO_FILESYSTEM) && defined(OPENSSL_EXTRA) && \
     defined(DEBUG_UNIT_TEST_CERTS)
 void DEBUG_WRITE_CERT_X509(WOLFSSL_X509* x509, const char* fileName);
diff --git a/wolfcrypt/benchmark/benchmark.c b/wolfcrypt/benchmark/benchmark.c
index c99fc71822..1f05e157de 100644
--- a/wolfcrypt/benchmark/benchmark.c
+++ b/wolfcrypt/benchmark/benchmark.c
@@ -1298,9 +1298,6 @@ static const bench_alg bench_other_opt[] = {
 #endif
     { NULL, 0}
 };
-#endif /* MAIN_NO_ARGS */
-
-#endif /* !WOLFSSL_BENCHMARK_ALL && !NO_MAIN_DRIVER */
 
 #if defined(BENCH_PQ_STATEFUL_HBS)
 typedef struct bench_pq_hash_sig_alg {
@@ -1353,6 +1350,10 @@ static const bench_pq_hash_sig_alg bench_pq_hash_sig_opt[] = {
 };
 #endif /* BENCH_PQ_STATEFUL_HBS */
 
+#endif /* MAIN_NO_ARGS */
+
+#endif /* !WOLFSSL_BENCHMARK_ALL && !NO_MAIN_DRIVER */
+
 #if !defined(WOLFSSL_BENCHMARK_ALL) && !defined(MAIN_NO_ARGS)
 #if defined(WOLFSSL_HAVE_MLKEM) || defined(HAVE_FALCON) || \
     defined(WOLFSSL_HAVE_MLDSA)
diff --git a/wolfcrypt/src/aes.c b/wolfcrypt/src/aes.c
index b9ef36f472..dccab8ff78 100644
--- a/wolfcrypt/src/aes.c
+++ b/wolfcrypt/src/aes.c
@@ -977,6 +977,30 @@ static WARN_UNUSED_RESULT int wc_AesDecrypt(Aes* aes, const byte* inBlock,
 }
 #endif /* HAVE_AES_DECRYPT && WOLFSSL_AES_DIRECT */
 
+#elif defined(WOLFSSL_PPC64_ASM)
+
+#if defined(WOLFSSL_AES_DIRECT) || defined(HAVE_AESCCM) || \
+    defined(WOLFSSL_AESGCM_STREAM) || defined(HAVE_AESGCM)
+static WARN_UNUSED_RESULT int wc_AesEncrypt(Aes* aes, const byte* inBlock,
+    byte* outBlock)
+{
+    AES_ECB_encrypt(inBlock, outBlock, WC_AES_BLOCK_SIZE, (byte*)aes->key,
+        (int)aes->rounds);
+
+    return 0;
+}
+#endif
+
+#if defined(HAVE_AES_DECRYPT) && defined(WOLFSSL_AES_DIRECT)
+static WARN_UNUSED_RESULT int wc_AesDecrypt(Aes* aes, const byte* inBlock,
+    byte* outBlock)
+{
+    AES_ECB_decrypt(inBlock, outBlock, WC_AES_BLOCK_SIZE, (byte*)aes->key,
+        (int)aes->rounds);
+    return 0;
+}
+#endif /* HAVE_AES_DECRYPT && WOLFSSL_AES_DIRECT */
+
 #elif defined(FREESCALE_MMCAU)
     /* Freescale mmCAU hardware AES support for Direct, CBC, CCM, GCM modes
      * through the CAU/mmCAU library. Documentation located in
@@ -1285,12 +1309,14 @@ static WARN_UNUSED_RESULT int wc_AesDecrypt(Aes* aes, const byte* inBlock,
 #if !defined(WOLFSSL_ESP32_CRYPT) || \
     (defined(NO_ESP32_CRYPT) || defined(NO_WOLFSSL_ESP32_CRYPT_AES) || \
      defined(NEED_AES_HW_FALLBACK))
+#ifndef WOLFSSL_PPC64_ASM
 static const FLASH_QUALIFIER word32 rcon[] = {
     0x01000000, 0x02000000, 0x04000000, 0x08000000,
     0x10000000, 0x20000000, 0x40000000, 0x80000000,
     0x1B000000, 0x36000000,
     /* for 128-bit blocks, Rijndael never uses more than 10 rcon values */
 };
+#endif
 #endif /* ESP32 */
 #endif /* __aarch64__ || !WOLFSSL_ARMASM */
 
@@ -4637,6 +4663,102 @@ static WARN_UNUSED_RESULT int wc_AesDecrypt(Aes* aes, const byte* inBlock,
         return AesSetKey(aes, userKey, keylen, iv, dir);
     }
 
+    #if defined(WOLFSSL_AES_DIRECT) || defined(WOLFSSL_AES_COUNTER)
+        /* AES-CTR and AES-DIRECT need to use this for key setup */
+        /* This function allows key sizes that are not 128/192/256 bits */
+    int wc_AesSetKeyDirect(Aes* aes, const byte* userKey, word32 keylen,
+                           const byte* iv, int dir)
+    {
+        if (aes == NULL) {
+            return BAD_FUNC_ARG;
+        }
+        if (keylen > sizeof(aes->key)) {
+            return BAD_FUNC_ARG;
+        }
+
+        return AesSetKey(aes, userKey, keylen, iv, dir);
+    }
+    #endif /* WOLFSSL_AES_DIRECT || WOLFSSL_AES_COUNTER */
+#elif defined(WOLFSSL_PPC64_ASM)
+    static int AesSetKey(Aes* aes, const byte* userKey, word32 keylen,
+            const byte* iv, int dir)
+    {
+    #if defined(WOLFSSL_AES_COUNTER) || defined(WOLFSSL_AES_CFB) || \
+        defined(WOLFSSL_AES_OFB) || defined(WOLFSSL_AES_XTS) || \
+        defined(WOLFSSL_AES_CTS)
+        aes->left = 0;
+    #endif
+
+        aes->keylen = (int)keylen;
+        aes->rounds = (keylen/4) + 6;
+
+        AES_set_encrypt_key(userKey, keylen * 8, (byte*)aes->key);
+
+    #ifdef HAVE_AES_DECRYPT
+        if (dir == AES_DECRYPTION) {
+            AES_invert_key((byte*)aes->key, aes->rounds);
+        }
+    #else
+        (void)dir;
+    #endif
+        return wc_AesSetIV(aes, iv);
+    }
+
+    int wc_AesSetKey(Aes* aes, const byte* userKey, word32 keylen,
+            const byte* iv, int dir)
+    {
+        if ((aes == NULL) || (userKey == NULL)) {
+            return BAD_FUNC_ARG;
+        }
+
+        switch (keylen) {
+    #if defined(AES_MAX_KEY_SIZE) && AES_MAX_KEY_SIZE >= 128 &&     \
+        defined(WOLFSSL_AES_128)
+        case 16:
+    #endif
+    #if defined(AES_MAX_KEY_SIZE) && AES_MAX_KEY_SIZE >= 192 &&     \
+        defined(WOLFSSL_AES_192)
+        case 24:
+    #endif
+    #if defined(AES_MAX_KEY_SIZE) && AES_MAX_KEY_SIZE >= 256 &&     \
+        defined(WOLFSSL_AES_256)
+        case 32:
+    #endif
+            break;
+        default:
+            return BAD_FUNC_ARG;
+        }
+
+    #ifdef WOLF_CRYPTO_CB
+        if (aes->devId != INVALID_DEVID) {
+        #ifdef WOLF_CRYPTO_CB_AES_SETKEY
+            int ret = wc_CryptoCb_AesSetKey(aes, userKey, keylen);
+            if (ret == 0) {
+                /* Callback succeeded - SE owns the key */
+                aes->keylen = (int)keylen;
+                if (iv != NULL)
+                    XMEMCPY(aes->reg, iv, WC_AES_BLOCK_SIZE);
+                else
+                    XMEMSET(aes->reg, 0, WC_AES_BLOCK_SIZE);
+                return 0;
+            }
+            else if (ret != WC_NO_ERR_TRACE(CRYPTOCB_UNAVAILABLE)) {
+                aes->devCtx = NULL;
+                return ret;
+            }
+            /* CRYPTOCB_UNAVAILABLE: continue to software setup */
+        #endif
+            /* Standard CryptoCB path - copy key to devKey for encrypt/decrypt offload */
+            if (keylen > sizeof(aes->devKey)) {
+                return BAD_FUNC_ARG;
+            }
+            XMEMCPY(aes->devKey, userKey, keylen);
+        }
+    #endif
+
+        return AesSetKey(aes, userKey, keylen, iv, dir);
+    }
+
     #if defined(WOLFSSL_AES_DIRECT) || defined(WOLFSSL_AES_COUNTER)
         /* AES-CTR and AES-DIRECT need to use this for key setup */
         /* This function allows key sizes that are not 128/192/256 bits */
@@ -6598,7 +6720,7 @@ int wc_AesSetIV(Aes* aes, const byte* iv)
 
 int wc_AesCbcEncrypt(Aes* aes, byte* out, const byte* in, word32 sz)
     {
-#if !defined(WOLFSSL_ARMASM)
+#if !defined(WOLFSSL_ARMASM) && !defined(WOLFSSL_PPC64_ASM)
         word32 blocks;
         int ret;
 #endif
@@ -6611,7 +6733,7 @@ int wc_AesCbcEncrypt(Aes* aes, byte* out, const byte* in, word32 sz)
             return 0;
         }
 
-#if !defined(WOLFSSL_ARMASM)
+#if !defined(WOLFSSL_ARMASM) && !defined(WOLFSSL_PPC64_ASM)
         blocks = sz / WC_AES_BLOCK_SIZE;
 #endif
 #ifdef WOLFSSL_AES_CBC_LENGTH_CHECKS
@@ -6687,6 +6809,10 @@ int wc_AesCbcEncrypt(Aes* aes, byte* out, const byte* in, word32 sz)
         }
     #endif
         return 0;
+#elif defined(WOLFSSL_PPC64_ASM)
+        AES_CBC_encrypt(in, out, sz, (const unsigned char*)aes->key,
+            aes->rounds, (unsigned char*)aes->reg);
+        return 0;
 #else
     #if defined(WOLFSSL_SE050) && defined(WOLFSSL_SE050_CRYPT)
         /* Implemented in wolfcrypt/src/port/nxp/se050_port.c */
@@ -6788,7 +6914,7 @@ int wc_AesCbcEncrypt(Aes* aes, byte* out, const byte* in, word32 sz)
     /* Software AES - CBC Decrypt */
     int wc_AesCbcDecrypt(Aes* aes, byte* out, const byte* in, word32 sz)
     {
-#if !defined(WOLFSSL_ARMASM)
+#if !defined(WOLFSSL_ARMASM) && !defined(WOLFSSL_PPC64_ASM)
         word32 blocks;
         int ret;
 #endif
@@ -6816,7 +6942,7 @@ int wc_AesCbcEncrypt(Aes* aes, byte* out, const byte* in, word32 sz)
         }
     #endif
 
-#if !defined(WOLFSSL_ARMASM)
+#if !defined(WOLFSSL_ARMASM) && !defined(WOLFSSL_PPC64_ASM)
         blocks = sz / WC_AES_BLOCK_SIZE;
 #endif
         if (sz % WC_AES_BLOCK_SIZE) {
@@ -6909,6 +7035,10 @@ int wc_AesCbcEncrypt(Aes* aes, byte* out, const byte* in, word32 sz)
     #endif
     #endif /* __aarch64__ || WOLFSSL_ARMASM_NO_HW_CRYPTO */
         return 0;
+#elif defined(WOLFSSL_PPC64_ASM)
+        AES_CBC_decrypt(in, out, sz, (const unsigned char*)aes->key,
+            aes->rounds, (unsigned char*)aes->reg);
+        return 0;
 #else
         VECTOR_REGISTERS_PUSH;
 
@@ -7266,7 +7396,7 @@ int wc_AesCbcEncrypt(Aes* aes, byte* out, const byte* in, word32 sz)
           !defined(WOLFSSL_ARMASM_NO_HW_CRYPTO))
             byte scratch[WC_AES_BLOCK_SIZE];
     #endif
-    #if !defined(WOLFSSL_ARMASM)
+    #if !defined(WOLFSSL_ARMASM) && !defined(WOLFSSL_PPC64_ASM)
             int ret = 0;
     #endif
             word32 processed;
@@ -7389,6 +7519,47 @@ int wc_AesCbcEncrypt(Aes* aes, byte* out, const byte* in, word32 sz)
             }
         #endif /* __aarch64__ || WOLFSSL_ARMASM_NO_HW_CRYPTO */
             return 0;
+    #elif defined(WOLFSSL_PPC64_ASM)
+            {
+                word32 numBlocks;
+                byte* tmp = (byte*)aes->tmp + WC_AES_BLOCK_SIZE - aes->left;
+                /* consume any unused bytes left in aes->tmp */
+                while ((aes->left != 0) && (sz != 0)) {
+                   *(out++) = *(in++) ^ *(tmp++);
+                   aes->left--;
+                   sz--;
+                }
+
+                /* do as many block size ops as possible */
+                numBlocks = sz / WC_AES_BLOCK_SIZE;
+                if (numBlocks > 0) {
+                    AES_CTR_encrypt(in, out, numBlocks * WC_AES_BLOCK_SIZE,
+                        (byte*)aes->key, aes->rounds, (byte*)aes->reg);
+
+                    sz  -= numBlocks * WC_AES_BLOCK_SIZE;
+                    out += numBlocks * WC_AES_BLOCK_SIZE;
+                    in  += numBlocks * WC_AES_BLOCK_SIZE;
+                }
+
+                /* handle non block size remaining */
+                if (sz) {
+                    byte zeros[WC_AES_BLOCK_SIZE] = { 0, 0, 0, 0, 0, 0, 0, 0,
+                                                      0, 0, 0, 0, 0, 0, 0, 0 };
+
+                    AES_CTR_encrypt(zeros, (byte*)aes->tmp,
+                        WC_AES_BLOCK_SIZE, (byte*)aes->key, aes->rounds,
+                        (byte*)aes->reg);
+
+                    aes->left = WC_AES_BLOCK_SIZE;
+                    tmp = (byte*)aes->tmp;
+
+                    while (sz--) {
+                        *(out++) = *(in++) ^ *(tmp++);
+                        aes->left--;
+                    }
+                }
+            }
+            return 0;
     #else
             VECTOR_REGISTERS_PUSH;
 
@@ -7640,6 +7811,13 @@ void GenerateM0(Gcm* gcm)
         }
     }
 
+#if defined(WOLFSSL_PPC64_ASM)
+    for (i = 1; i < 256; i++) {
+        word64* m64 = (word64*)gcm->M0[i];
+        m64[0] = ByteReverseWord64(m64[0]);
+        m64[1] = ByteReverseWord64(m64[1]);
+    }
+#endif
     XMEMSET(m[0], 0, WC_AES_BLOCK_SIZE);
 }
 
@@ -7725,7 +7903,6 @@ void GenerateM0(Gcm* gcm)
     #endif
     }
 #endif
-
 }
 
 #endif /* GCM_TABLE */
@@ -8125,6 +8302,37 @@ static void GCM_gmult_len_armasm_C(
 #elif defined(WOLFSSL_ARMASM)
 #define GCM_GMULT_LEN(gcm, x, a, len) \
     GCM_gmult_len_NEON(x, (const byte*)((gcm)->H), a, len)
+#elif defined(WOLFSSL_PPC64_ASM)
+static void GCM_gmult_len_armasm_C(
+    byte* x, const byte* h, const unsigned char* a, unsigned long len)
+{
+    byte Z[AES_BLOCK_SIZE];
+    byte V[AES_BLOCK_SIZE];
+    int i;
+    int j;
+
+    while (len >= AES_BLOCK_SIZE) {
+        xorbuf(x, a, AES_BLOCK_SIZE);
+        XMEMSET(Z, 0, AES_BLOCK_SIZE);
+        XMEMCPY(V, x, AES_BLOCK_SIZE);
+        for (i = 0; i < AES_BLOCK_SIZE; i++) {
+            byte y = h[i];
+            for (j = 0; j < 8; j++) {
+                if (y & 0x80) {
+                    xorbuf(Z, V, AES_BLOCK_SIZE);
+                }
+                RIGHTSHIFTX(V);
+                y = y << 1;
+            }
+        }
+        XMEMCPY(x, Z, AES_BLOCK_SIZE);
+        len -= AES_BLOCK_SIZE;
+        a += AES_BLOCK_SIZE;
+    }
+}
+
+#define GCM_GMULT_LEN(gcm, x, a, len) \
+    GCM_gmult_len_armasm_C(x, (gcm)->H, a, len)
 #endif
 
 #elif defined(GCM_TABLE)
@@ -8138,6 +8346,9 @@ static void GCM_gmult_len_armasm_C(
 #define GCM_GMULT_LEN(gcm, x, a, len) \
     GCM_gmult_len(x, (const byte**)((gcm)->M0), a, len)
 #endif
+#elif defined(WOLFSSL_PPC64_ASM)
+#define GCM_GMULT_LEN(gcm, x, a, len) \
+    GCM_gmult_len(x, (const byte**)((gcm)->M0), a, len)
 #else
 ALIGN16 static const byte R[256][2] = {
     {0x00, 0x00}, {0x01, 0xc2}, {0x03, 0x84}, {0x02, 0x46},
@@ -8400,7 +8611,7 @@ void GHASH(Gcm* gcm, const byte* a, word32 aSz, const byte* c,
 #endif /* WOLFSSL_AESGCM_STREAM */
 /* end GCM_TABLE */
 #elif defined(GCM_TABLE_4BIT)
-
+/* ARM assembly */
 #if defined(WOLFSSL_ARMASM) && (defined(__aarch64__) || \
     defined(WOLFSSL_ARMASM_NO_HW_CRYPTO))
 #if !defined(WOLFSSL_ARMASM_NO_NEON) && defined(__aarch64__)
@@ -8414,6 +8625,14 @@ void GHASH(Gcm* gcm, const byte* a, word32 aSz, const byte* c,
 #define GMULT(x, m)                                                      \
     GCM_gmult(x, (const byte**)m)
 #endif
+
+/* PPC64 assembly */
+#elif defined(WOLFSSL_PPC64_ASM)
+#define GCM_GMULT_LEN(gcm, x, a, len)                                    \
+    GCM_gmult_len(x, (const byte**)((gcm)->M0), a, len)
+#define GMULT(x, m)                                                      \
+    GCM_gmult(x, (const byte**)m)
+
 #else
 /* remainder = x^7 + x^2 + x^1 + 1 => 0xe1
  *  R shifts right a reverse bit pair of bytes such that:
@@ -9928,7 +10147,7 @@ static WARN_UNUSED_RESULT int wc_AesGcmEncrypt_STM32(
 
 #endif /* STM32_CRYPTO_AES_GCM */
 
-#if !defined(WOLFSSL_ARMASM)
+#if !defined(WOLFSSL_ARMASM) && !defined(WOLFSSL_PPC64_ASM)
 #ifdef WOLFSSL_AESNI
 /* For performance reasons, this code needs to be not inlined. */
 WARN_UNUSED_RESULT int AES_GCM_encrypt_C(
@@ -10048,8 +10267,9 @@ WARN_UNUSED_RESULT int AES_GCM_encrypt_C(
 
     return ret;
 }
-#elif defined(__aarch64__) || defined(WOLFSSL_ARMASM_NO_HW_CRYPTO)
-static int AES_GCM_encrypt_ARM(Aes* aes, byte* out, const byte* in,
+#elif (defined(__aarch64__) || defined(WOLFSSL_ARMASM_NO_HW_CRYPTO)) || \
+      defined(WOLFSSL_PPC64_ASM)
+static int AES_GCM_encrypt_ASM(Aes* aes, byte* out, const byte* in,
     word32 sz, const byte* iv, word32 ivSz, byte* authTag, word32 authTagSz,
     const byte* authIn, word32 authInSz)
 {
@@ -10305,10 +10525,13 @@ int wc_AesGcmEncrypt(Aes* aes, byte* out, const byte* in, word32 sz,
 #endif /* !WOLFSSL_ARMASM_NO_HW_CRYPTO */
 #if defined(__aarch64__) || defined(WOLFSSL_ARMASM_NO_HW_CRYPTO)
     {
-        ret = AES_GCM_encrypt_ARM(aes, out, in, sz, iv, ivSz, authTag,
+        ret = AES_GCM_encrypt_ASM(aes, out, in, sz, iv, ivSz, authTag,
             authTagSz, authIn, authInSz);
     }
 #endif /* __aarch64__ || WOLFSSL_ARMASM_NO_HW_CRYPTO */
+#elif defined(WOLFSSL_PPC64_ASM)
+    ret = AES_GCM_encrypt_ASM(aes, out, in, sz, iv, ivSz, authTag, authTagSz,
+        authIn, authInSz);
 #else
 #ifdef WOLFSSL_AESNI
     if (aes->use_aesni) {
@@ -10653,7 +10876,7 @@ static WARN_UNUSED_RESULT int wc_AesGcmDecrypt_STM32(
 
 #endif /* STM32_CRYPTO_AES_GCM */
 
-#if !defined(WOLFSSL_ARMASM)
+#if !defined(WOLFSSL_ARMASM) && !defined(WOLFSSL_PPC64_ASM)
 #ifdef WOLFSSL_AESNI
 /* For performance reasons, this code needs to be not inlined. */
 int WARN_UNUSED_RESULT AES_GCM_decrypt_C(
@@ -10802,8 +11025,9 @@ int WARN_UNUSED_RESULT AES_GCM_decrypt_C(
 #endif
     return ret;
 }
-#elif defined(__aarch64__) || defined(WOLFSSL_ARMASM_NO_HW_CRYPTO)
-static int AES_GCM_decrypt_ARM(Aes* aes, byte* out, const byte* in,
+#elif (defined(__aarch64__) || defined(WOLFSSL_ARMASM_NO_HW_CRYPTO)) || \
+      defined(WOLFSSL_PPC64_ASM)
+static int AES_GCM_decrypt_ASM(Aes* aes, byte* out, const byte* in,
     word32 sz, const byte* iv, word32 ivSz, const byte* authTag,
     word32 authTagSz, const byte* authIn, word32 authInSz)
 {
@@ -11057,10 +11281,15 @@ int wc_AesGcmDecrypt(Aes* aes, byte* out, const byte* in, word32 sz,
 #endif /* !WOLFSSL_ARMASM_NO_HW_CRYPTO */
 #if defined(__aarch64__) || defined(WOLFSSL_ARMASM_NO_HW_CRYPTO)
     {
-        ret = AES_GCM_decrypt_ARM(aes, out, in, sz, iv, ivSz, authTag,
+        ret = AES_GCM_decrypt_ASM(aes, out, in, sz, iv, ivSz, authTag,
             authTagSz, authIn, authInSz);
     }
 #endif /* __aarch64__ || WOLFSSL_ARMASM_NO_HW_CRYPTO */
+#elif defined(WOLFSSL_PPC64_ASM)
+    {
+        ret = AES_GCM_decrypt_ASM(aes, out, in, sz, iv, ivSz, authTag,
+            authTagSz, authIn, authInSz);
+    }
 #else
 #ifdef WOLFSSL_AESNI
     if (aes->use_aesni) {
@@ -14233,6 +14462,9 @@ static WARN_UNUSED_RESULT int _AesEcbEncrypt(
             aes->rounds);
     }
 #endif
+#elif defined(WOLFSSL_PPC64_ASM)
+    AES_ECB_encrypt(in, out, sz, (const unsigned char*)aes->key, aes->rounds);
+    ret = 0;
 #else
 #ifdef WOLFSSL_AESNI
     if (aes->use_aesni) {
@@ -14325,6 +14557,9 @@ static WARN_UNUSED_RESULT int _AesEcbDecrypt(
             aes->rounds);
     }
 #endif
+#elif defined(WOLFSSL_PPC64_ASM)
+    AES_ECB_decrypt(in, out, sz, (const unsigned char*)aes->key, aes->rounds);
+    ret = 0;
 #else
 #ifdef WOLFSSL_AESNI
     if (aes->use_aesni) {
@@ -15584,8 +15819,8 @@ static WARN_UNUSED_RESULT int _AesXtsHelper(
  */
 /* Software AES - XTS Encrypt  */
 
-#if !defined(WOLFSSL_ARMASM) || (!defined(__aarch64__) && \
-    defined(WOLFSSL_ARMASM_NO_HW_CRYPTO))
+#if (!defined(WOLFSSL_ARMASM) || (!defined(__aarch64__) && \
+     defined(WOLFSSL_ARMASM_NO_HW_CRYPTO))) && !defined(WOLFSSL_PPC64_ASM)
 static int AesXtsEncryptUpdate_sw(XtsAes* xaes, byte* out, const byte* in,
                                   word32 sz,
                                   byte *i);
@@ -15825,6 +16060,10 @@ int wc_AesXtsEncrypt(XtsAes* xaes, byte* out, const byte* in, word32 sz,
         ret = 0;
     }
 #endif
+#elif defined(WOLFSSL_PPC64_ASM)
+    AES_XTS_encrypt(in, out, sz, i, (byte*)xaes->aes.key,
+        (byte*)xaes->tweak.key, (byte*)xaes->aes.tmp, xaes->aes.rounds);
+    ret = 0;
 #else
     ret = AesXtsEncrypt_sw(xaes, out, in, sz, i);
 #endif
@@ -16043,8 +16282,8 @@ int wc_AesXtsEncryptFinal(XtsAes* xaes, byte* out, const byte* in, word32 sz,
  */
 /* Software AES - XTS Decrypt */
 
-#if !defined(WOLFSSL_ARMASM) || (!defined(__aarch64__) && \
-    defined(WOLFSSL_ARMASM_NO_HW_CRYPTO))
+#if (!defined(WOLFSSL_ARMASM) || (!defined(__aarch64__) && \
+     defined(WOLFSSL_ARMASM_NO_HW_CRYPTO))) && !defined(WOLFSSL_PPC64_ASM)
 static int AesXtsDecryptUpdate_sw(XtsAes* xaes, byte* out, const byte* in,
                                   word32 sz, byte *i);
 
@@ -16302,6 +16541,10 @@ int wc_AesXtsDecrypt(XtsAes* xaes, byte* out, const byte* in, word32 sz,
         ret = 0;
     }
 #endif
+#elif defined(WOLFSSL_PPC64_ASM)
+    AES_XTS_decrypt(in, out, sz, i, (byte*)xaes->aes.key,
+        (byte*)xaes->tweak.key, (byte*)xaes->aes.tmp, xaes->aes.rounds);
+    ret = 0;
 #else
     ret = AesXtsDecrypt_sw(xaes, out, in, sz, i);
 #endif
diff --git a/wolfcrypt/src/asn.c b/wolfcrypt/src/asn.c
index a4bd0d7ae2..9c76e4d5ee 100644
--- a/wolfcrypt/src/asn.c
+++ b/wolfcrypt/src/asn.c
@@ -8743,9 +8743,8 @@ int wc_RsaPrivateKeyValidate(const byte* input, word32* inOutIdx, int* keySz,
 #endif /* NO_RSA */
 
 #ifdef WOLFSSL_ASN_TEMPLATE
-/* ASN.1 template for a PKCS #8 key.
- * Ignoring optional attributes and public key.
- * PKCS #8: RFC 5958, 2 - PrivateKeyInfo
+/* ASN.1 template for a PKCS #8 PrivateKeyInfo / RFC 5958 OneAsymmetricKey.
+ * Includes the optional [0] attributes and [1] publicKey trailing fields.
  */
 static const ASNItem pkcs8KeyASN[] = {
 /*  SEQ                 */    { 0, ASN_SEQUENCE, 1, 1, 0 },
@@ -8758,9 +8757,10 @@ static const ASNItem pkcs8KeyASN[] = {
 /*  PKEY_ALGO_PARAM_SEQ */            { 2, ASN_SEQUENCE, 1, 0, 1 },
 #endif
 /*  PKEY_DATA           */        { 1, ASN_OCTET_STRING, 0, 0, 0 },
-/*  OPTIONAL Attributes IMPLICIT [0] */
+/*  Attributes [0] OPTIONAL */
                                   { 1, ASN_CONTEXT_SPECIFIC | 0, 1, 0, 1 },
-/* [[2: publicKey        [1] PublicKey OPTIONAL ]] */
+/*  publicKey [1] OPTIONAL */
+                                  { 1, ASN_CONTEXT_SPECIFIC | 1, 0, 0, 1 },
 };
 enum {
     PKCS8KEYASN_IDX_SEQ = 0,
@@ -8774,6 +8774,7 @@ enum {
 #endif
     PKCS8KEYASN_IDX_PKEY_DATA,
     PKCS8KEYASN_IDX_PKEY_ATTRIBUTES,
+    PKCS8KEYASN_IDX_PKEY_PUBKEY,
     WOLF_ENUM_DUMMY_LAST_ELEMENT(PKCS8KEYASN_IDX)
 };
 
@@ -8833,13 +8834,15 @@ int ToTraditionalInline_ex2(const byte* input, word32* inOutIdx, word32 sz,
         /* Key type OID. */
         oid = dataASN[PKCS8KEYASN_IDX_PKEY_ALGO_OID_KEY].data.oid.sum;
 
-        /* Version 1 includes an optional public key.
-         * If public key is included then the parsing will fail as it did not
-         * use all the data.
-         */
+        /* Only v1(0) and v2(1) are supported (RFC 5958). The [1] publicKey
+         * trailer is permitted only when version == v1. */
         if (version > PKCS8v1) {
             ret = ASN_PARSE_E;
         }
+        else if ((version < PKCS8v1) &&
+                 (dataASN[PKCS8KEYASN_IDX_PKEY_PUBKEY].tag != 0)) {
+            ret = ASN_PARSE_E;
+        }
     }
     if (ret == 0) {
         switch (oid) {
@@ -8935,9 +8938,71 @@ int ToTraditionalInline_ex2(const byte* input, word32* inOutIdx, word32 sz,
                 }
                 break;
         #endif
-            /* DSAk not supported. */
-            /* Falcon, Dilithium and SLH-DSA not supported. */
-            /* Ignore OID lookup failures. */
+        #ifdef HAVE_FALCON
+            case FALCON_LEVEL1k:
+            case FALCON_LEVEL5k:
+                /* Neither NULL item nor OBJECT_ID item allowed. */
+                if ((dataASN[PKCS8KEYASN_IDX_PKEY_ALGO_NULL].tag != 0) ||
+                    (dataASN[PKCS8KEYASN_IDX_PKEY_ALGO_OID_CURVE].tag != 0)) {
+                    ret = ASN_PARSE_E;
+                }
+                break;
+        #endif
+        #ifdef HAVE_DILITHIUM
+            #ifdef WOLFSSL_DILITHIUM_FIPS204_DRAFT
+            case DILITHIUM_LEVEL2k:
+            case DILITHIUM_LEVEL3k:
+            case DILITHIUM_LEVEL5k:
+            #endif
+            case ML_DSA_LEVEL2k:
+            case ML_DSA_LEVEL3k:
+            case ML_DSA_LEVEL5k:
+                /* Neither NULL item nor OBJECT_ID item allowed. */
+                if ((dataASN[PKCS8KEYASN_IDX_PKEY_ALGO_NULL].tag != 0) ||
+                    (dataASN[PKCS8KEYASN_IDX_PKEY_ALGO_OID_CURVE].tag != 0)) {
+                    ret = ASN_PARSE_E;
+                }
+                break;
+        #endif
+        #ifdef WOLFSSL_HAVE_SLHDSA
+            case SLH_DSA_SHA2_128Sk:
+            case SLH_DSA_SHA2_128Fk:
+            case SLH_DSA_SHA2_192Sk:
+            case SLH_DSA_SHA2_192Fk:
+            case SLH_DSA_SHA2_256Sk:
+            case SLH_DSA_SHA2_256Fk:
+            case SLH_DSA_SHAKE_128Sk:
+            case SLH_DSA_SHAKE_128Fk:
+            case SLH_DSA_SHAKE_192Sk:
+            case SLH_DSA_SHAKE_192Fk:
+            case SLH_DSA_SHAKE_256Sk:
+            case SLH_DSA_SHAKE_256Fk:
+                /* Neither NULL item nor OBJECT_ID item allowed. */
+                if ((dataASN[PKCS8KEYASN_IDX_PKEY_ALGO_NULL].tag != 0) ||
+                    (dataASN[PKCS8KEYASN_IDX_PKEY_ALGO_OID_CURVE].tag != 0)) {
+                    ret = ASN_PARSE_E;
+                }
+                break;
+        #endif
+        #ifdef WOLFSSL_HAVE_LMS
+            case HSS_LMSk:
+                /* Neither NULL item nor OBJECT_ID item allowed. */
+                if ((dataASN[PKCS8KEYASN_IDX_PKEY_ALGO_NULL].tag != 0) ||
+                    (dataASN[PKCS8KEYASN_IDX_PKEY_ALGO_OID_CURVE].tag != 0)) {
+                    ret = ASN_PARSE_E;
+                }
+                break;
+        #endif
+        #ifdef WOLFSSL_HAVE_XMSS
+            case XMSSk:
+                /* Neither NULL item nor OBJECT_ID item allowed. */
+                if ((dataASN[PKCS8KEYASN_IDX_PKEY_ALGO_NULL].tag != 0) ||
+                    (dataASN[PKCS8KEYASN_IDX_PKEY_ALGO_OID_CURVE].tag != 0)) {
+                    ret = ASN_PARSE_E;
+                }
+                break;
+        #endif
+            /* Other OIDs (DSAk), no parameter validation. */
             default:
                 break;
         }
@@ -9037,9 +9102,9 @@ int wc_GetPkcs8TraditionalOffset(byte* input, word32* inOutIdx, word32 sz)
 int wc_CreatePKCS8Key(byte* out, word32* outSz, byte* key, word32 keySz,
         int algoID, const byte* curveOID, word32 oidSz)
 {
-    /* pkcs8KeyASN_Length-1, the -1 is because we are not adding the optional
-     * set of attributes */
-    DECL_ASNSETDATA(dataASN, pkcs8KeyASN_Length-1);
+    /* pkcs8KeyASN_Length-2, the -2 is because we are not adding the optional
+     * set of attributes or publicKey */
+    DECL_ASNSETDATA(dataASN, pkcs8KeyASN_Length-2);
     word32 sz = 0;
     int ret = 0;
     word32 keyIdx = 0;
@@ -9066,7 +9131,7 @@ int wc_CreatePKCS8Key(byte* out, word32* outSz, byte* key, word32 keySz,
 #endif
 
     if (ret == 0)
-        CALLOC_ASNSETDATA(dataASN, pkcs8KeyASN_Length-1, ret, NULL);
+        CALLOC_ASNSETDATA(dataASN, pkcs8KeyASN_Length-2, ret, NULL);
 
     if (ret == 0) {
         /* Only support default PKCS #8 format - v0. */
@@ -9092,7 +9157,7 @@ int wc_CreatePKCS8Key(byte* out, word32* outSz, byte* key, word32 keySz,
         SetASN_Buffer(&dataASN[PKCS8KEYASN_IDX_PKEY_DATA], key, keySz);
 
         /* Get the size of the DER encoding. */
-        ret = SizeASN_Items(pkcs8KeyASN, dataASN, pkcs8KeyASN_Length-1, &sz);
+        ret = SizeASN_Items(pkcs8KeyASN, dataASN, pkcs8KeyASN_Length-2, &sz);
     }
     if ((ret == 0) || (ret == WC_NO_ERR_TRACE(LENGTH_ONLY_E))) {
         /* Always return the calculated size. */
@@ -9105,7 +9170,7 @@ int wc_CreatePKCS8Key(byte* out, word32* outSz, byte* key, word32 keySz,
     }
     if (ret == 0) {
         /*  Encode PKCS #8 key into buffer. */
-        SetASN_Items(pkcs8KeyASN, dataASN, pkcs8KeyASN_Length-1, out);
+        SetASN_Items(pkcs8KeyASN, dataASN, pkcs8KeyASN_Length-2, out);
         ret = (int)sz;
     }
 
@@ -17833,12 +17898,31 @@ int wolfssl_local_MatchBaseName(int type, const char* name, int nameSz,
     const char* base, int baseSz)
 {
     if (base == NULL || baseSz <= 0 || name == NULL || nameSz <= 0 ||
-            name[0] == '.' || nameSz < baseSz ||
+            name[0] == '.' ||
             (type != ASN_RFC822_TYPE && type != ASN_DNS_TYPE &&
              type != ASN_DIR_TYPE)) {
         return 0;
     }
 
+    if (type == ASN_DNS_TYPE) {
+        /* MatchDomainName treats one trailing dot as the absolute-FQDN marker.
+         * Apply the same normalization before enforcing DNS name constraints.
+         */
+        if (name[nameSz - 1] == '.') {
+            nameSz--;
+        }
+        if (base[baseSz - 1] == '.') {
+            baseSz--;
+        }
+        if (nameSz <= 0 || baseSz <= 0) {
+            return 0;
+        }
+    }
+
+    if (nameSz < baseSz) {
+        return 0;
+    }
+
     if (type == ASN_DIR_TYPE)
         return XMEMCMP(name, base, (size_t)baseSz) == 0;
 
@@ -17927,8 +18011,8 @@ int wolfssl_local_MatchBaseName(int type, const char* name, int nameSz,
     return 1;
 }
 
-static int MatchUriNameConstraint(const char* uri, int uriSz, const char* base,
-    int baseSz)
+int wolfssl_local_MatchUriNameConstraint(const char* uri, int uriSz,
+    const char* base, int baseSz)
 {
     const char* hostStart;
     const char* hostEnd;
@@ -17936,7 +18020,10 @@ static int MatchUriNameConstraint(const char* uri, int uriSz, const char* base,
     const char* uriEnd;
     int hostSz;
 
-    if (uri == NULL || uriSz <= 0 || base == NULL || baseSz <= 0) {
+    /* Need at least 3 bytes for the "://" scheme separator; rejecting short
+     * inputs early also keeps the loop bound (uriEnd - 2) from forming a
+     * pointer before `uri`. */
+    if (uri == NULL || uriSz < 3 || base == NULL || baseSz <= 0) {
         return 0;
     }
 
@@ -17992,8 +18079,168 @@ static int MatchUriNameConstraint(const char* uri, int uriSz, const char* base,
         return 0;
     }
 
-    return wolfssl_local_MatchBaseName(ASN_DNS_TYPE, hostStart, hostSz, base,
-        baseSz);
+    /* RFC 5280 4.2.1.10: for URIs the constraint applies to the host part.
+     * A constraint that begins with a '.' matches any host with one or more
+     * additional leading labels (the bare host is excluded) - this is the
+     * DNS subtree behaviour. A constraint without a leading '.' specifies a
+     * single host and must match it exactly. */
+    if (base[0] == '.') {
+        return wolfssl_local_MatchBaseName(ASN_DNS_TYPE, hostStart, hostSz,
+            base, baseSz);
+    }
+    else {
+        int i;
+        if (hostSz != baseSz) {
+            return 0;
+        }
+        for (i = 0; i < baseSz; i++) {
+            if (XTOLOWER((unsigned char)hostStart[i]) !=
+                XTOLOWER((unsigned char)base[i])) {
+                return 0;
+            }
+        }
+        return 1;
+    }
+}
+
+/* Locate the right-most label of `s` that ends (exclusive) at index `end`.
+ * Sets *outStart to its starting index (the index of its first byte). If
+ * outHasWild is non-NULL, sets *outHasWild to 1 iff the label contains '*'.
+ * The '.' immediately before *outStart (if any) is the label separator and is
+ * not part of either the current or the preceding label. */
+static void PrevDnsLabel(const char* s, int end, int* outStart,
+    int* outHasWild)
+{
+    int start = end;
+    int hasWild = 0;
+    while (start > 0 && s[start - 1] != '.') {
+        if (s[start - 1] == '*') {
+            hasWild = 1;
+        }
+        start--;
+    }
+    *outStart = start;
+    if (outHasWild != NULL) {
+        *outHasWild = hasWild;
+    }
+}
+
+/* Match a wildcard DNS SAN against a DNS name-constraint subtree.
+ *
+ * A wildcard SAN denotes the set of names its '*'(s) can expand to. Because a
+ * '*' never crosses a label boundary (see MatchDomainName), every expansion
+ * has the same number of labels and only the content of '*'-bearing labels
+ * varies. Matching is therefore done label-by-label from the right against the
+ * constraint base.
+ *
+ * permitted != 0: containment. Accept only if EVERY expansion stays inside the
+ *   subtree, i.e. each of the right-most base-length labels of the name is
+ *   wildcard-free and equal to the corresponding base label. Extra labels to
+ *   the left may be anything (adding labels on the left stays in-subtree).
+ *
+ * permitted == 0: intersection (for excluded subtrees). Reject if SOME
+ *   expansion falls inside the subtree. A label containing a '*' is
+ *   conservatively treated as able to match any single base label; a literal
+ *   label must equal the base label.
+ *
+ * A leading '.' on the base denotes a proper subtree (the apex is excluded),
+ * which requires at least one extra label on the left of the name.
+ *
+ * Returns 1 on match (contained / intersecting), 0 otherwise.
+ */
+int wolfssl_local_MatchDnsConstraintWildcard(const char* name, int nameSz,
+    const char* base, int baseSz, int permitted)
+{
+    int baseLead;
+    int ni, bi;
+
+    if (name == NULL || base == NULL || nameSz <= 0 || baseSz <= 0) {
+        return 0;
+    }
+
+    /* MatchDomainName treats one trailing dot as the absolute-FQDN marker.
+     * Apply the same normalization before label-wise constraint matching.
+     */
+    if (name[nameSz - 1] == '.') {
+        nameSz--;
+    }
+    if (base[baseSz - 1] == '.') {
+        baseSz--;
+    }
+    if (nameSz <= 0 || baseSz <= 0 || name[0] == '.') {
+        return 0;
+    }
+
+    baseLead = (base[0] == '.');
+    if (baseLead) {
+        base++;
+        baseSz--;
+    }
+    /* A base of only dots (".", "..") has no labels to match. */
+    if (baseSz <= 0 || base[0] == '.') {
+        return 0;
+    }
+
+    ni = nameSz; /* exclusive end of the unconsumed name */
+    bi = baseSz; /* exclusive end of the unconsumed base */
+
+    /* Compare each base label (right to left) with the aligned name label. */
+    while (bi > 0) {
+        int nStart, bStart, nLen, bLen, k;
+        int hasWild = 0;
+
+        /* Base labels remain but the name has none left -> name is shorter in
+         * labels and cannot contain the base. */
+        if (ni <= 0) {
+            return 0;
+        }
+
+        PrevDnsLabel(name, ni, &nStart, &hasWild);
+        PrevDnsLabel(base, bi, &bStart, NULL);
+        nLen = ni - nStart;
+        bLen = bi - bStart;
+
+        /* Empty label (e.g. "a..b" or an extra trailing dot) is invalid. */
+        if (nLen == 0 || bLen == 0) {
+            return 0;
+        }
+
+        if (hasWild) {
+            /* permitted: a wildcard label cannot prove containment.
+             * excluded: a wildcard label is conservatively treated as
+             * compatible, so nothing more to check. */
+            if (permitted) {
+                return 0;
+            }
+        }
+        else {
+            /* Literal label: both modes require exact case-insensitive
+             * equality with the base label. */
+            if (nLen != bLen) {
+                return 0;
+            }
+            for (k = 0; k < bLen; k++) {
+                if (XTOLOWER((unsigned char)name[nStart + k]) !=
+                    XTOLOWER((unsigned char)base[bStart + k])) {
+                    return 0;
+                }
+            }
+        }
+
+        /* Consume both labels and the '.' that precedes them (if any). */
+        ni = nStart - 1;
+        bi = bStart - 1;
+    }
+
+    /* All base labels matched. ni >= 0 means name[ni] == '.' and at least one
+     * extra label remains on the left; ni < 0 means the name had exactly the
+     * base's label count (an apex match). A leading-dot base is a proper
+     * subtree and requires at least one extra left label. */
+    if (baseLead && ni < 0) {
+        return 0;
+    }
+
+    return 1;
 }
 
 /* Check if IP address matches a name constraint.
@@ -18046,6 +18293,18 @@ static int MatchOtherNameConstraint(DNS_entry* name, Base_entry* current)
     return XMEMCMP(name->name, current->name, (size_t)current->nameSz) == 0;
 }
 
+/* Return 1 if the name contains a wildcard '*' character. */
+static int DnsNameHasWildcard(const char* name, int nameSz)
+{
+    int i;
+    for (i = 0; i < nameSz; i++) {
+        if (name[i] == '*') {
+            return 1;
+        }
+    }
+    return 0;
+}
+
 /* Search through the list to find if the name is permitted.
  * name     The DNS name to search for
  * dnsList  The list to search through
@@ -18073,7 +18332,7 @@ static int PermittedListOk(DNS_entry* name, Base_entry* dnsList, byte nameType)
                 }
             }
             else if (nameType == ASN_URI_TYPE) {
-                if (MatchUriNameConstraint(name->name, name->len,
+                if (wolfssl_local_MatchUriNameConstraint(name->name, name->len,
                         current->name, current->nameSz)) {
                     match = 1;
                     break;
@@ -18096,6 +18355,17 @@ static int PermittedListOk(DNS_entry* name, Base_entry* dnsList, byte nameType)
                     break;
                 }
             }
+            else if (nameType == ASN_DNS_TYPE &&
+                     DnsNameHasWildcard(name->name, name->len)) {
+                /* Wildcard DNS SAN: a '*' can expand to a longer label, so the
+                 * byte-length guard used for literal names below is invalid.
+                 * Permit only if every expansion stays inside the subtree. */
+                if (wolfssl_local_MatchDnsConstraintWildcard(name->name,
+                        name->len, current->name, current->nameSz, 1)) {
+                    match = 1;
+                    break;
+                }
+            }
             else if (name->len >= current->nameSz &&
                 wolfssl_local_MatchBaseName(nameType, name->name, name->len,
                                             current->name, current->nameSz)) {
@@ -18137,7 +18407,7 @@ static int IsInExcludedList(DNS_entry* name, Base_entry* dnsList, byte nameType)
                 }
             }
             else if (nameType == ASN_URI_TYPE) {
-                if (MatchUriNameConstraint(name->name, name->len,
+                if (wolfssl_local_MatchUriNameConstraint(name->name, name->len,
                         current->name, current->nameSz)) {
                     ret = 1;
                     break;
@@ -18159,6 +18429,17 @@ static int IsInExcludedList(DNS_entry* name, Base_entry* dnsList, byte nameType)
                     break;
                 }
             }
+            else if (nameType == ASN_DNS_TYPE &&
+                     DnsNameHasWildcard(name->name, name->len)) {
+                /* Wildcard DNS SAN: a '*' can expand to a longer label, so the
+                 * byte-length guard used for literal names below is invalid.
+                 * Exclude if any expansion can fall inside the subtree. */
+                if (wolfssl_local_MatchDnsConstraintWildcard(name->name,
+                        name->len, current->name, current->nameSz, 0)) {
+                    ret = 1;
+                    break;
+                }
+            }
             else if (name->len >= current->nameSz &&
                 wolfssl_local_MatchBaseName(nameType, name->name, name->len,
                                             current->name, current->nameSz)) {
@@ -18579,7 +18860,7 @@ static int DecodeGeneralName(const byte* input, word32* inOutIdx, byte tag,
         }
         WOLFSSL_MSG("\tPutting URI into list but not using");
 
-    #if !defined(WOLFSSL_NO_ASN_STRICT) && !defined(WOLFSSL_FPKI)
+    #ifndef WOLFSSL_NO_ASN_STRICT
         /* Verify RFC 5280 Sec 4.2.1.6 rule:
             "The name MUST NOT be a relative URI"
             As per RFC 3986 Sec 4.3, an absolute URI is only required to contain
@@ -18595,9 +18876,8 @@ static int DecodeGeneralName(const byte* input, word32* inOutIdx, byte tag,
                     break;
                 }
                 if (input[idx + (word32)i] == '/') {
-                    i = len; /* error, found relative path since '/' was
-                              * encountered before ':'. Returning error
-                              * value in next if statement. */
+                    /* path is relative since '/' was encountered before ':'. */
+                    return ASN_ALT_NAME_E;
                 }
             }
 
@@ -18669,6 +18949,10 @@ static int DecodeGeneralName(const byte* input, word32* inOutIdx, byte tag,
      *   - CheckForAltNames (TLS hostname matching): skips ASN_RID_TYPE
      *     unconditionally and excludes them from *checkCN, so a cert
      *     with only registeredID SANs still falls back to CN.
+     *   - CheckForAltNames (TLS hostname matching): skips ASN_URI_TYPE
+     *     for DNS hostname checks (RFC 9525 Sec. 6.3) but URI SAN presence
+     *     still suppresses CN fallback because URI-ID is a distinct presented
+     *     identifier.
      *   - DNS_to_GENERAL_NAME (used by wolfSSL_X509_get_ext) and the
      *     ALT_NAMES_OID arm of wolfSSL_X509_get_ext_d2i: build a proper
      *     ASN1_OBJECT in d.registeredID from raw OID bytes regardless
@@ -37657,7 +37941,7 @@ static int DecodeAcertGeneralName(const byte* input, word32* inOutIdx,
     else if (tag == (ASN_CONTEXT_SPECIFIC | ASN_URI_TYPE)) {
         WOLFSSL_MSG("\tPutting URI into list but not using");
 
-    #if !defined(WOLFSSL_NO_ASN_STRICT) && !defined(WOLFSSL_FPKI)
+    #ifndef WOLFSSL_NO_ASN_STRICT
         /* Verify RFC 5280 Sec 4.2.1.6 rule:
            "The name MUST NOT be a relative URI"
            As per RFC 3986 Sec 4.3, an absolute URI is only required to contain
@@ -37673,9 +37957,8 @@ static int DecodeAcertGeneralName(const byte* input, word32* inOutIdx,
                     break;
                 }
                 if (input[idx + (word32)i] == '/') {
-                    i = len; /* error, found relative path since '/' was
-                              * encountered before ':'. Returning error
-                              * value in next if statement. */
+                    /* path is relative since '/' was encountered before ':'. */
+                    return ASN_ALT_NAME_E;
                 }
             }
 
diff --git a/wolfcrypt/src/asn_orig.c b/wolfcrypt/src/asn_orig.c
index 7db2a4eed1..f1e67ff4f5 100644
--- a/wolfcrypt/src/asn_orig.c
+++ b/wolfcrypt/src/asn_orig.c
@@ -750,81 +750,6 @@ int EncryptContent(byte* input, word32 inputSz, byte* out, word32* outSz,
 #endif
 #endif
 #ifndef NO_RSA
-#if defined(WOLFSSL_RENESAS_TSIP_TLS) || defined(WOLFSSL_RENESAS_FSPSM_TLS)
-static int RsaPublicKeyDecodeRawIndex(const byte* input, word32* inOutIdx,
-                                      word32 inSz, word32* key_n,
-                                      word32* key_n_len, word32* key_e,
-                                      word32* key_e_len)
-{
-    int ret = 0;
-    int length = 0;
-
-#if defined(OPENSSL_EXTRA) || defined(RSA_DECODE_EXTRA)
-    byte b;
-#endif
-
-    if (input == NULL || inOutIdx == NULL)
-        return BAD_FUNC_ARG;
-
-    if (GetSequence(input, inOutIdx, &length, inSz) < 0)
-        return ASN_PARSE_E;
-
-#if defined(OPENSSL_EXTRA) || defined(RSA_DECODE_EXTRA)
-    if ((*inOutIdx + 1) > inSz)
-        return BUFFER_E;
-
-    b = input[*inOutIdx];
-    if (b != ASN_INTEGER) {
-        /* not from decoded cert, will have algo id, skip past */
-        if (GetSequence(input, inOutIdx, &length, inSz) < 0)
-            return ASN_PARSE_E;
-
-        if (SkipObjectId(input, inOutIdx, inSz) < 0)
-            return ASN_PARSE_E;
-
-        /* Option NULL ASN.1 tag */
-        if (*inOutIdx  >= inSz) {
-            return BUFFER_E;
-        }
-        if (input[*inOutIdx] == ASN_TAG_NULL) {
-            ret = GetASNNull(input, inOutIdx, inSz);
-            if (ret != 0)
-                return ret;
-        }
-        /* TODO: support RSA PSS */
-
-        /* should have bit tag length and seq next */
-        ret = CheckBitString(input, inOutIdx, NULL, inSz, 1, NULL);
-        if (ret != 0)
-            return ret;
-
-        if (GetSequence(input, inOutIdx, &length, inSz) < 0)
-            return ASN_PARSE_E;
-    }
-#endif /* OPENSSL_EXTRA */
-
-    /* Get modulus */
-    ret = GetASNInt(input, inOutIdx, &length, inSz);
-    *key_n += *inOutIdx;
-    if (ret < 0) {
-        return ASN_RSA_KEY_E;
-    }
-    if (key_n_len)
-        *key_n_len = length;
-    *inOutIdx += length;
-
-    /* Get exponent */
-    ret = GetASNInt(input, inOutIdx, &length, inSz);
-    *key_e += *inOutIdx;
-    if (ret < 0) {
-        return ASN_RSA_KEY_E;
-    }
-    if (key_e_len)
-        *key_e_len = length;
-    return ret;
-}
-
-#endif
 int wc_RsaPublicKeyDecode_ex(const byte* input, word32* inOutIdx, word32 inSz,
     const byte** n, word32* nSz, const byte** e, word32* eSz)
 {
@@ -3420,7 +3345,7 @@ static int DecodeAltNames(const byte* input, word32 sz, DecodedCert* cert)
                 return ASN_PARSE_E;
             }
 
-        #if !defined(WOLFSSL_NO_ASN_STRICT) && !defined(WOLFSSL_FPKI)
+        #ifndef WOLFSSL_NO_ASN_STRICT
             /* Verify RFC 5280 Sec 4.2.1.6 rule:
                 "The name MUST NOT be a relative URI"
                 As per RFC 3986 Sec 4.3, an absolute URI is only required to contain
diff --git a/wolfcrypt/src/cryptocb.c b/wolfcrypt/src/cryptocb.c
index d647fda7bf..c3af7ed608 100644
--- a/wolfcrypt/src/cryptocb.c
+++ b/wolfcrypt/src/cryptocb.c
@@ -2033,10 +2033,20 @@ int wc_CryptoCb_Sha384Hash(wc_Sha384* sha384, const byte* in,
 
 #ifdef WOLFSSL_SHA512
 int wc_CryptoCb_Sha512Hash(wc_Sha512* sha512, const byte* in,
-    word32 inSz, byte* digest, size_t digestSz)
+    word32 inSz, byte* digest
+#if !(defined(HAVE_FIPS) && FIPS_VERSION_LT(7,0))
+    , size_t digestSz
+#endif
+    )
 {
     int ret = WC_NO_ERR_TRACE(CRYPTOCB_UNAVAILABLE);
     CryptoCb* dev;
+#if defined(HAVE_FIPS) && FIPS_VERSION_LT(7,0)
+    /* Older FIPS sha512.c snapshots call the 4-arg API (no digestSz). Treat as
+     * full-size SHA-512 with no variant dispatch, matching pre-digestSz
+     * behavior. */
+    size_t digestSz = WC_SHA512_DIGEST_SIZE;
+#endif
 
     /* locate registered callback */
     #ifndef NO_SHA2_CRYPTO_CB
diff --git a/wolfcrypt/src/dh.c b/wolfcrypt/src/dh.c
index 26166c7310..c7560d185c 100644
--- a/wolfcrypt/src/dh.c
+++ b/wolfcrypt/src/dh.c
@@ -57,14 +57,6 @@
     }
 #endif
 
-#if defined(WOLFSSL_USE_SAVE_VECTOR_REGISTERS) && !defined(WOLFSSL_SP_ASM)
-    /* force off unneeded vector register save/restore. */
-    #undef SAVE_VECTOR_REGISTERS
-    #define SAVE_VECTOR_REGISTERS(fail_clause) SAVE_NO_VECTOR_REGISTERS(fail_clause)
-    #undef RESTORE_VECTOR_REGISTERS
-    #define RESTORE_VECTOR_REGISTERS() RESTORE_NO_VECTOR_REGISTERS()
-#endif
-
 /*
 Possible DH enable options:
  * NO_RSA:              Overall control of DH                 default: on (not defined)
@@ -1425,8 +1417,6 @@ int wc_DhGeneratePublic(DhKey* key, byte* priv, word32 privSz,
         return BAD_FUNC_ARG;
     }
 
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
-
     ret = GeneratePublicDh(key, priv, privSz, pub, pubSz);
 
     #if FIPS_VERSION_GE(5,0) || defined(WOLFSSL_VALIDATE_DH_KEYGEN)
@@ -1436,8 +1426,6 @@ int wc_DhGeneratePublic(DhKey* key, byte* priv, word32 privSz,
         ret = _ffc_pairwise_consistency_test(key, pub, *pubSz, priv, privSz);
     #endif /* FIPS V5 or later || WOLFSSL_VALIDATE_DH_KEYGEN */
 
-    RESTORE_VECTOR_REGISTERS();
-
     return ret;
 }
 
@@ -1451,8 +1439,6 @@ static int wc_DhGenerateKeyPair_Sync(DhKey* key, WC_RNG* rng,
         return BAD_FUNC_ARG;
     }
 
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
-
     ret = GeneratePrivateDh(key, rng, priv, privSz);
 
     if (ret == 0)
@@ -1464,9 +1450,6 @@ static int wc_DhGenerateKeyPair_Sync(DhKey* key, WC_RNG* rng,
         ret = _ffc_pairwise_consistency_test(key, pub, *pubSz, priv, *privSz);
 #endif /* FIPS V5 or later || WOLFSSL_VALIDATE_DH_KEYGEN */
 
-
-    RESTORE_VECTOR_REGISTERS();
-
     return ret;
 }
 #endif /* !WOLFSSL_KCAPI_DH */
@@ -1589,8 +1572,6 @@ static int _ffc_validate_public_key(DhKey* key, const byte* pub, word32 pubSz,
         return MP_INIT_E;
     }
 
-    SAVE_VECTOR_REGISTERS(ret = _svr_ret;);
-
     if (mp_read_unsigned_bin(y, pub, pubSz) != MP_OKAY) {
         ret = MP_READ_E;
     }
@@ -1679,8 +1660,6 @@ static int _ffc_validate_public_key(DhKey* key, const byte* pub, word32 pubSz,
     mp_clear(p);
     mp_clear(q);
 
-    RESTORE_VECTOR_REGISTERS();
-
 #if defined(WOLFSSL_SMALL_STACK) && !defined(WOLFSSL_NO_MALLOC)
     XFREE(q, key->heap, DYNAMIC_TYPE_DH);
     XFREE(p, key->heap, DYNAMIC_TYPE_DH);
@@ -1919,8 +1898,6 @@ static int _ffc_pairwise_consistency_test(DhKey* key,
         return MP_INIT_E;
     }
 
-    SAVE_VECTOR_REGISTERS(ret = _svr_ret;);
-
     /* Load the private and public keys into big integers. */
     if (mp_read_unsigned_bin(publicKey, pub, pubSz) != MP_OKAY ||
         mp_read_unsigned_bin(privateKey, priv, privSz) != MP_OKAY) {
@@ -1979,8 +1956,6 @@ static int _ffc_pairwise_consistency_test(DhKey* key,
     mp_clear(publicKey);
     mp_clear(checkKey);
 
-    RESTORE_VECTOR_REGISTERS();
-
 #if defined(WOLFSSL_SMALL_STACK) && !defined(WOLFSSL_NO_MALLOC)
     XFREE(checkKey, key->heap, DYNAMIC_TYPE_DH);
     XFREE(privateKey, key->heap, DYNAMIC_TYPE_DH);
@@ -2174,8 +2149,6 @@ static int wc_DhAgree_Sync(DhKey* key, byte* agree, word32* agreeSz,
             ret = MP_INIT_E;
 
         if (ret == 0) {
-            SAVE_VECTOR_REGISTERS(ret = _svr_ret;);
-
             if (ret == 0 && mp_read_unsigned_bin(y, otherPub, pubSz) != MP_OKAY)
                 ret = MP_READ_E;
 
@@ -2201,8 +2174,6 @@ static int wc_DhAgree_Sync(DhKey* key, byte* agree, word32* agreeSz,
             }
 
             mp_clear(y);
-
-            RESTORE_VECTOR_REGISTERS();
         }
 
         /* make sure agree is > 1 (SP800-56A, 5.7.1.1) */
@@ -2253,8 +2224,6 @@ static int wc_DhAgree_Sync(DhKey* key, byte* agree, word32* agreeSz,
     }
 #endif
 
-    SAVE_VECTOR_REGISTERS(ret = _svr_ret;);
-
     if (mp_read_unsigned_bin(x, priv, privSz) != MP_OKAY)
         ret = MP_READ_E;
 #ifdef WOLFSSL_CHECK_MEM_ZERO
@@ -2313,8 +2282,6 @@ static int wc_DhAgree_Sync(DhKey* key, byte* agree, word32* agreeSz,
     mp_clear(y);
     mp_forcezero(x);
 
-    RESTORE_VECTOR_REGISTERS();
-
 #else
     (void)ct;
     ret = WC_KEY_SIZE_E;
@@ -2601,8 +2568,6 @@ static int _DhSetKey(DhKey* key, const byte* p, word32 pSz, const byte* g,
         ret = BAD_FUNC_ARG;
     }
 
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
-
     if (ret == 0) {
         /* may have leading 0 */
         if (p[0] == 0) {
@@ -2714,8 +2679,6 @@ static int _DhSetKey(DhKey* key, const byte* p, word32 pSz, const byte* g,
             mp_clear(keyP);
     }
 
-    RESTORE_VECTOR_REGISTERS();
-
     return ret;
 }
 
@@ -3204,8 +3167,6 @@ int wc_DhGenerateParams(WC_RNG *rng, int modSz, DhKey *dh)
     }
 #endif
 
-    SAVE_VECTOR_REGISTERS(ret = _svr_ret;);
-
     if (ret == 0) {
         /* force magnitude */
         buf[0] |= 0xC0;
@@ -3264,9 +3225,10 @@ int wc_DhGenerateParams(WC_RNG *rng, int modSz, DhKey *dh)
             if (ret != 0 || primeCheck == MP_YES)
                 break;
 
-            /* linuxkm: release the kernel for a moment before iterating. */
-            RESTORE_VECTOR_REGISTERS();
-            SAVE_VECTOR_REGISTERS(ret = _svr_ret; break;);
+            ret = WC_CHECK_FOR_INTR_SIGNALS();
+            if (ret != 0)
+                break;
+            WC_RELAX_LONG_LOOP();
         };
     }
 
@@ -3308,8 +3270,6 @@ int wc_DhGenerateParams(WC_RNG *rng, int modSz, DhKey *dh)
         mp_clear(&dh->g);
     }
 
-    RESTORE_VECTOR_REGISTERS();
-
 #ifndef WOLFSSL_NO_MALLOC
     if (buf != NULL)
 #endif
diff --git a/wolfcrypt/src/dsa.c b/wolfcrypt/src/dsa.c
index 0e849c2349..5df809c2da 100644
--- a/wolfcrypt/src/dsa.c
+++ b/wolfcrypt/src/dsa.c
@@ -36,14 +36,6 @@
     #include <wolfcrypt/src/misc.c>
 #endif
 
-#if defined(WOLFSSL_USE_SAVE_VECTOR_REGISTERS) && !defined(WOLFSSL_SP_ASM)
-    /* force off unneeded vector register save/restore. */
-    #undef SAVE_VECTOR_REGISTERS
-    #define SAVE_VECTOR_REGISTERS(fail_clause) SAVE_NO_VECTOR_REGISTERS(fail_clause)
-    #undef RESTORE_VECTOR_REGISTERS
-    #define RESTORE_VECTOR_REGISTERS() RESTORE_NO_VECTOR_REGISTERS()
-#endif
-
 #ifdef _MSC_VER
     /* disable for while(0) cases (MSVC bug) */
     #pragma warning(disable:4127)
@@ -269,8 +261,6 @@ int wc_MakeDsaKey(WC_RNG *rng, DsaKey *dsa)
     }
 #endif
 
-    SAVE_VECTOR_REGISTERS(;);
-
 #if defined(WOLFSSL_SMALL_STACK) && !defined(WOLFSSL_NO_MALLOC)
     if ((tmpQ = (mp_int *)XMALLOC(sizeof(*tmpQ), dsa->heap,
             DYNAMIC_TYPE_TMP_BUFFER)) == NULL)
@@ -338,8 +328,6 @@ int wc_MakeDsaKey(WC_RNG *rng, DsaKey *dsa)
     mp_clear(tmpQ);
 #endif
 
-    RESTORE_VECTOR_REGISTERS();
-
     return err;
 }
 
@@ -454,6 +442,11 @@ int wc_MakeDsaParameters(WC_RNG *rng, int modulus_size, DsaKey *dsa)
                     break;
                 loop_check_prime++;
             }
+
+            err = WC_CHECK_FOR_INTR_SIGNALS();
+            if (err != 0)
+                break;
+            WC_RELAX_LONG_LOOP();
         }
     }
 
@@ -794,8 +787,6 @@ int wc_DsaSign_ex(const byte* digest, word32 digestSz, byte* out, DsaKey* key,
         return BAD_LENGTH_E;
     }
 
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
-
     do {
 #ifdef WOLFSSL_SMALL_STACK
         k = (mp_int *)XMALLOC(sizeof *k, key->heap, DYNAMIC_TYPE_TMP_BUFFER);
@@ -1040,8 +1031,6 @@ int wc_DsaSign_ex(const byte* digest, word32 digestSz, byte* out, DsaKey* key,
         }
     } while (0);
 
-    RESTORE_VECTOR_REGISTERS();
-
 #ifdef WOLFSSL_SMALL_STACK
     if (k) {
         if ((ret != WC_NO_ERR_TRACE(MP_INIT_E)) &&
diff --git a/wolfcrypt/src/ecc.c b/wolfcrypt/src/ecc.c
index 57b29197d6..11276ce6e9 100644
--- a/wolfcrypt/src/ecc.c
+++ b/wolfcrypt/src/ecc.c
@@ -275,14 +275,6 @@ ECC Curve Sizes:
     #include <wolfssl/wolfcrypt/hmac.h>
 #endif
 
-#if defined(WOLFSSL_USE_SAVE_VECTOR_REGISTERS) && !defined(WOLFSSL_SP_ASM)
-    /* force off unneeded vector register save/restore. */
-    #undef SAVE_VECTOR_REGISTERS
-    #define SAVE_VECTOR_REGISTERS(fail_clause) SAVE_NO_VECTOR_REGISTERS(fail_clause)
-    #undef RESTORE_VECTOR_REGISTERS
-    #define RESTORE_VECTOR_REGISTERS() RESTORE_NO_VECTOR_REGISTERS()
-#endif
-
 #if !defined(WOLFSSL_ATECC508A) && !defined(WOLFSSL_ATECC608A) && \
     !defined(WOLFSSL_MICROCHIP_TA100) && \
     !defined(WOLFSSL_CRYPTOCELL) && !defined(WOLFSSL_SILABS_SE_ACCEL) && \
@@ -5150,8 +5142,6 @@ int wc_ecc_shared_secret_ex(ecc_key* private_key, ecc_point* point,
         return ECC_BAD_ARG_E;
     }
 
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
-
     switch (private_key->state) {
         case ECC_STATE_NONE:
         case ECC_STATE_SHARED_SEC_GEN:
@@ -5194,8 +5184,6 @@ int wc_ecc_shared_secret_ex(ecc_key* private_key, ecc_point* point,
             err = BAD_STATE_E;
     } /* switch */
 
-    RESTORE_VECTOR_REGISTERS();
-
     /* if async pending then return and skip done cleanup below */
     if (err == WC_NO_ERR_TRACE(WC_PENDING_E)) {
         return err;
@@ -5280,8 +5268,6 @@ int wc_ecc_point_is_on_curve(ecc_point *p, int curve_idx)
        return ECC_BAD_ARG_E;
     }
 
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
-
     ALLOC_CURVE_SPECS(3, err);
     if (err == MP_OKAY) {
         err = wc_ecc_curve_load(wc_ecc_get_curve_params(curve_idx), &curve,
@@ -5296,8 +5282,6 @@ int wc_ecc_point_is_on_curve(ecc_point *p, int curve_idx)
     wc_ecc_curve_free(curve);
     FREE_CURVE_SPECS();
 
-    RESTORE_VECTOR_REGISTERS();
-
     return err;
 }
 
@@ -5456,8 +5440,6 @@ static int ecc_make_pub_ex(ecc_key* key, ecc_curve_spec* curve,
         return BAD_FUNC_ARG;
     }
 
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
-
 #ifdef HAVE_ECC_MAKE_PUB
     /* if ecc_point passed in then use it as output for public key point */
     if (pubOut != NULL) {
@@ -5603,8 +5585,6 @@ static int ecc_make_pub_ex(ecc_key* key, ecc_curve_spec* curve,
         key->type = ECC_PRIVATEKEY;
     }
 
-    RESTORE_VECTOR_REGISTERS();
-
     return err;
 }
 
@@ -6107,8 +6087,6 @@ int wc_ecc_make_key_ex2(WC_RNG* rng, int keysize, ecc_key* key, int curve_id,
 {
     int err;
 
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
-
     err = _ecc_make_key_ex(rng, keysize, key, curve_id, flags);
 
 #if (FIPS_VERSION_GE(5,0) || defined(WOLFSSL_VALIDATE_ECC_KEYGEN)) && \
@@ -6132,8 +6110,6 @@ int wc_ecc_make_key_ex2(WC_RNG* rng, int keysize, ecc_key* key, int curve_id,
     }
 #endif
 
-    RESTORE_VECTOR_REGISTERS();
-
     return err;
 }
 
@@ -7195,6 +7171,11 @@ static int ecc_sign_hash_sw(ecc_key* key, ecc_key* pubkey, WC_RNG* rng,
          mp_clear(pubkey->pubkey.z);
      #endif
          mp_forcezero(pubkey->k);
+
+         err = WC_CHECK_FOR_INTR_SIGNALS();
+         if (err != 0)
+             break;
+         WC_RELAX_LONG_LOOP();
     }
     mp_forcezero(b);
     FREE_MP_INT_SIZE(b, key->heap, DYNAMIC_TYPE_ECC);
@@ -7244,10 +7225,8 @@ static int ecc_sign_hash_sp(const byte* in, word32 inlen, WC_RNG* rng,
         #if !defined(WC_ECC_NONBLOCK) || (defined(WC_ECC_NONBLOCK) && !defined(WC_ECC_NONBLOCK_ONLY))
             {
                 int ret;
-                SAVE_VECTOR_REGISTERS(return _svr_ret;);
                 ret = sp_ecc_sign_256(in, inlen, rng, ecc_get_k(key), r, s,
                                       sign_k, key->heap);
-                RESTORE_VECTOR_REGISTERS();
                 return ret;
             }
         #endif
@@ -7256,10 +7235,8 @@ static int ecc_sign_hash_sp(const byte* in, word32 inlen, WC_RNG* rng,
     #if defined(WOLFSSL_SM2) && defined(WOLFSSL_SP_SM2)
         if (ecc_sets[key->idx].id == ECC_SM2P256V1) {
             int ret;
-            SAVE_VECTOR_REGISTERS(return _svr_ret;);
             ret = sp_ecc_sign_sm2_256(in, inlen, rng, ecc_get_k(key), r, s,
                                       sign_k, key->heap);
-            RESTORE_VECTOR_REGISTERS();
             return ret;
         }
     #endif
@@ -7284,10 +7261,8 @@ static int ecc_sign_hash_sp(const byte* in, word32 inlen, WC_RNG* rng,
         #if !defined(WC_ECC_NONBLOCK) || (defined(WC_ECC_NONBLOCK) && !defined(WC_ECC_NONBLOCK_ONLY))
             {
                 int ret;
-                SAVE_VECTOR_REGISTERS(return _svr_ret;);
                 ret = sp_ecc_sign_384(in, inlen, rng, ecc_get_k(key), r, s,
                                       sign_k, key->heap);
-                RESTORE_VECTOR_REGISTERS();
                 return ret;
             }
         #endif
@@ -7314,10 +7289,8 @@ static int ecc_sign_hash_sp(const byte* in, word32 inlen, WC_RNG* rng,
         #if !defined(WC_ECC_NONBLOCK) || (defined(WC_ECC_NONBLOCK) && !defined(WC_ECC_NONBLOCK_ONLY))
             {
                 int ret;
-                SAVE_VECTOR_REGISTERS(return _svr_ret;);
                 ret = sp_ecc_sign_521(in, inlen, rng, ecc_get_k(key), r, s,
                                       sign_k, key->heap);
-                RESTORE_VECTOR_REGISTERS();
                 return ret;
             }
         #endif
@@ -8938,10 +8911,8 @@ static int ecc_verify_hash_sp(mp_int *r, mp_int *s, const byte* hash,
         #if !defined(WC_ECC_NONBLOCK) || (defined(WC_ECC_NONBLOCK) && !defined(WC_ECC_NONBLOCK_ONLY))
             {
                 int ret;
-                SAVE_VECTOR_REGISTERS(return _svr_ret;);
                 ret = sp_ecc_verify_256(hash, hashlen, key->pubkey.x,
                     key->pubkey.y, key->pubkey.z, r, s, res, key->heap);
-                RESTORE_VECTOR_REGISTERS();
                 return ret;
             }
         #endif
@@ -8982,10 +8953,8 @@ static int ecc_verify_hash_sp(mp_int *r, mp_int *s, const byte* hash,
         #if !defined(WC_ECC_NONBLOCK) || (defined(WC_ECC_NONBLOCK) && !defined(WC_ECC_NONBLOCK_ONLY))
             {
                 int ret;
-                SAVE_VECTOR_REGISTERS(return _svr_ret;);
                 ret = sp_ecc_verify_384(hash, hashlen, key->pubkey.x,
                     key->pubkey.y, key->pubkey.z, r, s, res, key->heap);
-                RESTORE_VECTOR_REGISTERS();
                 return ret;
             }
         #endif
@@ -9011,10 +8980,8 @@ static int ecc_verify_hash_sp(mp_int *r, mp_int *s, const byte* hash,
         #if !defined(WC_ECC_NONBLOCK) || (defined(WC_ECC_NONBLOCK) && !defined(WC_ECC_NONBLOCK_ONLY))
             {
                 int ret;
-                SAVE_VECTOR_REGISTERS(return _svr_ret;);
                 ret = sp_ecc_verify_521(hash, hashlen, key->pubkey.x,
                     key->pubkey.y, key->pubkey.z, r, s, res, key->heap);
-                RESTORE_VECTOR_REGISTERS();
                 return ret;
             }
         #endif
@@ -9652,8 +9619,6 @@ int wc_ecc_import_point_der_ex(const byte* in, word32 inLen,
     if (err != MP_OKAY)
         return MEMORY_E;
 
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
-
     /* check for point type (4, 2, or 3) */
     pointType = in[0];
     if (pointType != ECC_POINT_UNCOMP && pointType != ECC_POINT_COMP_EVEN &&
@@ -9845,8 +9810,6 @@ int wc_ecc_import_point_der_ex(const byte* in, word32 inLen,
         mp_clear(point->z);
     }
 
-    RESTORE_VECTOR_REGISTERS();
-
     return err;
 }
 
@@ -10197,8 +10160,6 @@ static int _ecc_is_point(ecc_point* ecp, mp_int* a, mp_int* b, mp_int* prime)
       return err;
    }
 
-   SAVE_VECTOR_REGISTERS(err = _svr_ret;);
-
    /* compute y^2 */
    if (err == MP_OKAY)
        err = mp_sqr(ecp->y, t1);
@@ -10267,8 +10228,6 @@ static int _ecc_is_point(ecc_point* ecp, mp_int* a, mp_int* b, mp_int* prime)
    mp_clear(t1);
    mp_clear(t2);
 
-   RESTORE_VECTOR_REGISTERS();
-
    WC_FREE_VAR_EX(t2, NULL, DYNAMIC_TYPE_ECC);
    WC_FREE_VAR_EX(t1, NULL, DYNAMIC_TYPE_ECC);
 
@@ -10718,8 +10677,6 @@ static int _ecc_validate_public_key(ecc_key* key, int partial, int priv)
     DECLARE_CURVE_SPECS(4);
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (key == NULL)
         return BAD_FUNC_ARG;
 
@@ -10871,9 +10828,7 @@ WOLFSSL_ABI
 int wc_ecc_check_key(ecc_key* key)
 {
     int ret;
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
     ret = _ecc_validate_public_key(key, 0, 1);
-    RESTORE_VECTOR_REGISTERS();
     return ret;
 }
 
@@ -10938,8 +10893,6 @@ static int _ecc_import_x963_ex2(const byte* in, word32 inLen, ecc_key* key,
     mp_forcezero(key->kb);
 #endif
 
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
-
     /* check for point type (4, 2, or 3) */
     pointType = in[0];
     if (pointType != ECC_POINT_UNCOMP && pointType != ECC_POINT_COMP_EVEN &&
@@ -11271,8 +11224,6 @@ static int _ecc_import_x963_ex2(const byte* in, word32 inLen, ecc_key* key,
         mp_forcezero(key->k);
     }
 
-    RESTORE_VECTOR_REGISTERS();
-
     return err;
 }
 
@@ -11691,10 +11642,6 @@ static int _ecc_import_private_key_ex(const byte* priv, word32 privSz,
     }
 #else
 
-#ifdef WOLFSSL_VALIDATE_ECC_IMPORT
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
-#endif
-
     ret = mp_read_unsigned_bin(key->k, priv, privSz);
 #ifdef HAVE_WOLF_BIGINT
     if (ret == 0 && wc_bigint_from_unsigned_bin(&key->k->raw, priv,
@@ -11745,10 +11692,6 @@ static int _ecc_import_private_key_ex(const byte* priv, word32 privSz,
 
 #endif
 
-#ifdef WOLFSSL_VALIDATE_ECC_IMPORT
-    RESTORE_VECTOR_REGISTERS();
-#endif
-
 #ifdef WOLFSSL_MAXQ10XX_CRYPTO
     if ((ret == 0) && (key->devId != INVALID_DEVID)) {
         ret = wc_MAXQ10XX_EccSetKey(key, key->dp->size);
@@ -12137,10 +12080,6 @@ static int _ecc_import_raw_private(ecc_key* key, const char* qx,
     }
 #endif
 
-#ifdef WOLFSSL_VALIDATE_ECC_IMPORT
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
-#endif
-
     /* import private key */
     if (err == MP_OKAY) {
         if (d != NULL) {
@@ -12234,10 +12173,6 @@ static int _ecc_import_raw_private(ecc_key* key, const char* qx,
     }
 #endif
 
-#ifdef WOLFSSL_VALIDATE_ECC_IMPORT
-    RESTORE_VECTOR_REGISTERS();
-#endif
-
 #ifdef WOLFSSL_MAXQ10XX_CRYPTO
     if (err == MP_OKAY) {
         err = wc_MAXQ10XX_EccSetKey(key, key->dp->size);
@@ -13858,8 +13793,6 @@ int ecc_mul2add(ecc_point* A, mp_int* kA,
    }
 #endif /* HAVE_THREAD_LS */
 
-      SAVE_VECTOR_REGISTERS(err = _svr_ret;);
-
       /* find point */
       idx1 = find_base(A);
 
@@ -13942,8 +13875,6 @@ int ecc_mul2add(ecc_point* A, mp_int* kA,
         }
       }
 
-      RESTORE_VECTOR_REGISTERS();
-
 #ifndef HAVE_THREAD_LS
     wc_UnLockMutex(&ecc_fp_lock);
 #endif /* HAVE_THREAD_LS */
@@ -14011,8 +13942,6 @@ int wc_ecc_mulmod_ex(const mp_int* k, ecc_point *G, ecc_point *R, mp_int* a,
    got_ecc_fp_lock = 1;
 #endif /* HAVE_THREAD_LS */
 
-      SAVE_VECTOR_REGISTERS(err = _svr_ret; goto out;);
-
       /* find point */
       idx = find_base(G);
 
@@ -14061,8 +13990,6 @@ int wc_ecc_mulmod_ex(const mp_int* k, ecc_point *G, ecc_point *R, mp_int* a,
         }
       }
 
-      RESTORE_VECTOR_REGISTERS();
-
   out:
 
 #ifndef HAVE_THREAD_LS
@@ -14088,36 +14015,28 @@ int wc_ecc_mulmod_ex(const mp_int* k, ecc_point *G, ecc_point *R, mp_int* a,
 #if defined(WOLFSSL_SM2) && defined(WOLFSSL_SP_SM2)
     if ((mp_count_bits(modulus) == 256) && (!mp_is_bit_set(modulus, 224))) {
         int ret;
-        SAVE_VECTOR_REGISTERS(return _svr_ret);
         ret = sp_ecc_mulmod_sm2_256(k, G, R, map, heap);
-        RESTORE_VECTOR_REGISTERS();
         return ret;
     }
 #endif
 #ifndef WOLFSSL_SP_NO_256
     if (mp_count_bits(modulus) == 256) {
         int ret;
-        SAVE_VECTOR_REGISTERS(return _svr_ret;);
         ret = sp_ecc_mulmod_256(k, G, R, map, heap);
-        RESTORE_VECTOR_REGISTERS();
         return ret;
     }
 #endif
 #ifdef WOLFSSL_SP_384
     if (mp_count_bits(modulus) == 384) {
         int ret;
-        SAVE_VECTOR_REGISTERS(return _svr_ret;);
         ret = sp_ecc_mulmod_384(k, G, R, map, heap);
-        RESTORE_VECTOR_REGISTERS();
         return ret;
     }
 #endif
 #ifdef WOLFSSL_SP_521
     if (mp_count_bits(modulus) == 521) {
         int ret;
-        SAVE_VECTOR_REGISTERS(return _svr_ret;);
         ret = sp_ecc_mulmod_521(k, G, R, map, heap);
-        RESTORE_VECTOR_REGISTERS();
         return ret;
     }
 #endif
@@ -14182,8 +14101,6 @@ int wc_ecc_mulmod_ex2(const mp_int* k, ecc_point *G, ecc_point *R, mp_int* a,
    got_ecc_fp_lock = 1;
 #endif /* HAVE_THREAD_LS */
 
-      SAVE_VECTOR_REGISTERS(err = _svr_ret; goto out;);
-
       /* find point */
       idx = find_base(G);
 
@@ -14232,8 +14149,6 @@ int wc_ecc_mulmod_ex2(const mp_int* k, ecc_point *G, ecc_point *R, mp_int* a,
         }
       }
 
-      RESTORE_VECTOR_REGISTERS();
-
   out:
 
 #ifndef HAVE_THREAD_LS
@@ -14262,36 +14177,28 @@ int wc_ecc_mulmod_ex2(const mp_int* k, ecc_point *G, ecc_point *R, mp_int* a,
 #if defined(WOLFSSL_SM2) && defined(WOLFSSL_SP_SM2)
     if ((mp_count_bits(modulus) == 256) && (!mp_is_bit_set(modulus, 224))) {
         int ret;
-        SAVE_VECTOR_REGISTERS(return _svr_ret;);
         ret = sp_ecc_mulmod_sm2_256(k, G, R, map, heap);
-        RESTORE_VECTOR_REGISTERS();
         return ret;
     }
 #endif
 #ifndef WOLFSSL_SP_NO_256
     if (mp_count_bits(modulus) == 256) {
         int ret;
-        SAVE_VECTOR_REGISTERS(return _svr_ret;);
         ret = sp_ecc_mulmod_256(k, G, R, map, heap);
-        RESTORE_VECTOR_REGISTERS();
         return ret;
     }
 #endif
 #ifdef WOLFSSL_SP_384
     if (mp_count_bits(modulus) == 384) {
         int ret;
-        SAVE_VECTOR_REGISTERS(return _svr_ret;);
         ret = sp_ecc_mulmod_384(k, G, R, map, heap);
-        RESTORE_VECTOR_REGISTERS();
         return ret;
     }
 #endif
 #ifdef WOLFSSL_SP_521
     if (mp_count_bits(modulus) == 521) {
         int ret;
-        SAVE_VECTOR_REGISTERS(return _svr_ret;);
         ret = sp_ecc_mulmod_521(k, G, R, map, heap);
-        RESTORE_VECTOR_REGISTERS();
         return ret;
     }
 #endif
@@ -14887,8 +14794,6 @@ int wc_ecc_encrypt_ex(ecc_key* privKey, ecc_key* pubKey, const byte* msg,
     }
 #endif
 
-    SAVE_VECTOR_REGISTERS(ret = _svr_ret;);
-
 #ifdef WOLFSSL_ECIES_ISO18033
     XMEMCPY(sharedSecret, out - pubKeySz, pubKeySz);
     sharedSz -= pubKeySz;
@@ -15105,8 +15010,6 @@ int wc_ecc_encrypt_ex(ecc_key* privKey, ecc_key* pubKey, const byte* msg,
 #endif
     }
 
-    RESTORE_VECTOR_REGISTERS();
-
     ForceZero(sharedSecret, sharedSz);
     ForceZero(keys, (word32)keysLen);
     WC_FREE_VAR_EX(sharedSecret, ctx->heap, DYNAMIC_TYPE_ECC_BUFFER);
@@ -15266,8 +15169,6 @@ int wc_ecc_decrypt(ecc_key* privKey, ecc_key* pubKey, const byte* msg,
     }
 #endif
 
-    SAVE_VECTOR_REGISTERS(ret = _svr_ret;);
-
 #ifndef WOLFSSL_ECIES_OLD
     if (pubKey == NULL) {
         WC_ALLOC_VAR_EX(peerKey, ecc_key, 1, ctx->heap,
@@ -15501,8 +15402,6 @@ int wc_ecc_decrypt(ecc_key* privKey, ecc_key* pubKey, const byte* msg,
     if (ret == 0)
        *outSz = msgSz - digestSz;
 
-    RESTORE_VECTOR_REGISTERS();
-
 #ifndef WOLFSSL_ECIES_OLD
     if (pubKey == peerKey)
         wc_ecc_free(peerKey);
@@ -15576,8 +15475,6 @@ static int mp_jacobi(mp_int* a, mp_int* n, int* c)
         return res;
     }
 
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
-
     if ((res = mp_mod(a, n, a1)) != MP_OKAY) {
         goto done;
     }
@@ -15635,8 +15532,6 @@ static int mp_jacobi(mp_int* a, mp_int* n, int* c)
 
 done:
 
-    RESTORE_VECTOR_REGISTERS();
-
     /* cleanup */
     mp_clear(n1);
     mp_clear(a1);
@@ -15674,8 +15569,6 @@ static int mp_sqrtmod_prime(mp_int* n, mp_int* prime, mp_int* ret)
       return MP_VAL;
   }
 
-  SAVE_VECTOR_REGISTERS(return _svr_ret;);
-
   res = mp_init(&e);
   if (res == MP_OKAY)
       res = mp_mod_d(prime, 8, &i);
@@ -15699,8 +15592,6 @@ static int mp_sqrtmod_prime(mp_int* n, mp_int* prime, mp_int* ret)
 
   mp_clear(&e);
 
-  RESTORE_VECTOR_REGISTERS();
-
   return res;
 #else
   int res, legendre, done = 0;
@@ -15720,8 +15611,6 @@ static int mp_sqrtmod_prime(mp_int* n, mp_int* prime, mp_int* ret)
   mp_int t1[1], C[1], Q[1], S[1], Z[1], M[1], T[1], R[1], N[1], two[1];
 #endif
 
-  SAVE_VECTOR_REGISTERS(res = _svr_ret; goto out;);
-
   if ((mp_init_multi(t1, C, Q, S, Z, M) != MP_OKAY) ||
       (mp_init_multi(T, R, N, two, NULL, NULL) != MP_OKAY)) {
     res = MP_INIT_E;
@@ -15933,8 +15822,6 @@ static int mp_sqrtmod_prime(mp_int* n, mp_int* prime, mp_int* ret)
 
   out:
 
-  RESTORE_VECTOR_REGISTERS();
-
 #ifdef WOLFSSL_SMALL_STACK
   if (t1) {
     if (res != WC_NO_ERR_TRACE(MP_INIT_E))
diff --git a/wolfcrypt/src/eccsi.c b/wolfcrypt/src/eccsi.c
index d919dd8a34..d0417ec879 100644
--- a/wolfcrypt/src/eccsi.c
+++ b/wolfcrypt/src/eccsi.c
@@ -36,14 +36,6 @@
     #include <wolfssl/wolfcrypt/sp.h>
 #endif
 
-#if defined(WOLFSSL_USE_SAVE_VECTOR_REGISTERS) && !defined(WOLFSSL_SP_ASM)
-    /* force off unneeded vector register save/restore. */
-    #undef SAVE_VECTOR_REGISTERS
-    #define SAVE_VECTOR_REGISTERS(fail_clause) SAVE_NO_VECTOR_REGISTERS(fail_clause)
-    #undef RESTORE_VECTOR_REGISTERS
-    #define RESTORE_VECTOR_REGISTERS() RESTORE_NO_VECTOR_REGISTERS()
-#endif
-
 #ifndef WOLFSSL_HAVE_ECC_KEY_GET_PRIV
     /* FIPS build has replaced ecc.h. */
     #define wc_ecc_key_get_priv(key) (&((key)->k))
@@ -1507,8 +1499,6 @@ int wc_ValidateEccsiPair(EccsiKey* key, enum wc_HashType hashType,
     if (err != 0)
         return err;
 
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
-
     params = &key->params;
     hs = &key->tmp;
     res = &key->pubkey.pubkey;
@@ -1563,8 +1553,6 @@ int wc_ValidateEccsiPair(EccsiKey* key, enum wc_HashType hashType,
         }
     }
 
-    RESTORE_VECTOR_REGISTERS();
-
     return err;
 }
 
@@ -2231,8 +2219,6 @@ int wc_VerifyEccsiHash(EccsiKey* key, enum wc_HashType hashType,
     if (err != 0)
         return err;
 
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
-
     /* Decode the signature into components. */
     r = wc_ecc_key_get_priv(&key->pubkey);
     pvt = &key->pubkey.pubkey;
@@ -2318,8 +2304,6 @@ int wc_VerifyEccsiHash(EccsiKey* key, enum wc_HashType hashType,
         *verified = ((err == 0) && (mp_cmp(jx, r) == MP_EQ));
     }
 
-    RESTORE_VECTOR_REGISTERS();
-
     return err;
 }
 #endif /* WOLFCRYPT_ECCSI_CLIENT */
diff --git a/wolfcrypt/src/ed25519.c b/wolfcrypt/src/ed25519.c
index de12e87b83..b84d7e9b15 100644
--- a/wolfcrypt/src/ed25519.c
+++ b/wolfcrypt/src/ed25519.c
@@ -205,6 +205,64 @@ static int ed25519_hash(ed25519_key* key, const byte* in, word32 inLen,
     return ret;
 }
 
+/* Reject small-order Ed25519 public keys: h*A vanishes during verification
+ * so any (R = [S]B, S) verifies for an arbitrary message. */
+static int ed25519_is_small_order(const byte p[ED25519_PUB_KEY_SIZE])
+{
+    /* y-coordinates of every order-1/2/4/8 point plus the two non-canonical
+     * encodings y = p / y = p+1. Sign bit masked before compare. Only
+     * {y, y + p} fits in 32 bytes (2p overflows the 255-bit y field), so
+     * listing y and y + p exhausts the reachable encodings for each
+     * small-order y. */
+    static const byte small_order_y[][ED25519_PUB_KEY_SIZE] = {
+        /* order 4: y = 0 */
+        {0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00},
+        /* order 1: y = 1 (identity) */
+        {0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00},
+        /* order 8 */
+        {0x26,0xe8,0x95,0x8f,0xc2,0xb2,0x27,0xb0,
+         0x45,0xc3,0xf4,0x89,0xf2,0xef,0x98,0xf0,
+         0xd5,0xdf,0xac,0x05,0xd3,0xc6,0x33,0x39,
+         0xb1,0x38,0x02,0x88,0x6d,0x53,0xfc,0x05},
+        /* order 8 */
+        {0xc7,0x17,0x6a,0x70,0x3d,0x4d,0xd8,0x4f,
+         0xba,0x3c,0x0b,0x76,0x0d,0x10,0x67,0x0f,
+         0x2a,0x20,0x53,0xfa,0x2c,0x39,0xcc,0xc6,
+         0x4e,0xc7,0xfd,0x77,0x92,0xac,0x03,0x7a},
+        /* order 2: y = p - 1 */
+        {0xec,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f},
+        /* non-canonical y = p (decodes to y = 0) */
+        {0xed,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f},
+        /* non-canonical y = p + 1 (decodes to y = 1) */
+        {0xee,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f},
+    };
+    byte y[ED25519_PUB_KEY_SIZE];
+    word32 i;
+
+    XMEMCPY(y, p, ED25519_PUB_KEY_SIZE);
+    y[ED25519_PUB_KEY_SIZE - 1] &= 0x7f;
+    for (i = 0; i < sizeof(small_order_y) / ED25519_PUB_KEY_SIZE; i++) {
+        if (XMEMCMP(y, small_order_y[i], ED25519_PUB_KEY_SIZE) == 0)
+            return 1;
+    }
+    return 0;
+}
+
 #ifdef HAVE_ED25519_MAKE_KEY
 #if FIPS_VERSION3_GE(6,0,0)
 /* Performs a Pairwise Consistency Test on an Ed25519 key pair.
@@ -808,6 +866,13 @@ static int ed25519_verify_msg_final_with_sha(const byte* sig, word32 sigLen,
     if (i == -1)
         return BAD_FUNC_ARG;
 
+    /* Defence in depth: also catch small-order keys imported with trusted=1. */
+    if (ed25519_is_small_order(key->p)) {
+        WOLFSSL_MSG("Ed25519 small-order public key rejected during "
+                    "signature verification");
+        return BAD_FUNC_ARG;
+    }
+
     /* uncompress A (public key), test if valid, and negate it */
 #ifndef FREESCALE_LTC_ECC
     if (ge_frombytes_negate_vartime(&A, key->p) != 0)
@@ -1502,6 +1567,13 @@ int wc_ed25519_check_key(ed25519_key* key)
         ret = PUBLIC_KEY_E;
     }
 
+    /* Reject small-order pub key before the priv-vs-pub compare so the
+     * diagnostic isn't masked by a "mismatch" error. */
+    if ((ret == 0) && ed25519_is_small_order(key->p)) {
+        WOLFSSL_MSG("Ed25519 small-order public key rejected during key check");
+        ret = PUBLIC_KEY_E;
+    }
+
 #ifdef HAVE_ED25519_MAKE_KEY
     /* If we have a private key just make the public key and compare. */
     if ((ret == 0) && (key->privKeySet)) {
@@ -1521,9 +1593,6 @@ int wc_ed25519_check_key(ed25519_key* key)
         && (!key->privKeySet)
 #endif
         ) {
-        /* Verify that Q is not identity element 0.
-         * 0 has no representation for Ed25519. */
-
         /* Verify that xQ and yQ are integers in the interval [0, p - 1].
          * Only have yQ so check that ordinate. p = 2^255 - 19 */
         if ((key->p[ED25519_PUB_KEY_SIZE - 1] & 0x7f) == 0x7f) {
diff --git a/wolfcrypt/src/ed448.c b/wolfcrypt/src/ed448.c
index 72f1724856..d6c35c6245 100644
--- a/wolfcrypt/src/ed448.c
+++ b/wolfcrypt/src/ed448.c
@@ -238,6 +238,77 @@ static int ed448_pairwise_consistency_test(ed448_key* key, WC_RNG* rng)
 }
 #endif
 
+/* Reject small-order Ed448 public keys: h*A vanishes during verification
+ * so any (R = [S]B, S) verifies for an arbitrary message. Cofactor is 4. */
+static int ed448_is_small_order(const byte p[ED448_PUB_KEY_SIZE])
+{
+    /* y-coordinates of every order-1/2/4 point plus the non-canonical
+     * encodings y = p / y = p+1. Byte 56 is cleared in both table and
+     * input before compare, masking the x-sign bit and the
+     * spec-mandated-zero (but decoder-ignored) bits 0-6. The decoder
+     * (fe448_from_bytes) reads bytes 0-55 modulo p with no canonical-form
+     * check, so y = p decodes to 0 and y = p+1 decodes to 1; both must
+     * be rejected here. Only {y, y + p} fits in 56 bytes (2p overflows),
+     * so listing y and y + p exhausts the reachable encodings. */
+    static const byte small_order_y[][ED448_PUB_KEY_SIZE] = {
+        /* order 1: identity y = 1, x = 0 */
+        {0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00},
+        /* order 4: y = 0 (x = +/-1; sign bit covered by mask) */
+        {0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00},
+        /* order 2: y = p - 1, x = 0; p = 2^448 - 2^224 - 1 */
+        {0xfe,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xfe,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0x00},
+        /* non-canonical y = p (decodes to y = 0) */
+        {0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xfe,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0x00},
+        /* non-canonical y = p + 1 (decodes to y = 1) */
+        {0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
+         0x00,0x00,0x00,0x00,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
+         0x00},
+    };
+    byte y[ED448_PUB_KEY_SIZE];
+    word32 i;
+
+    XMEMCPY(y, p, ED448_PUB_KEY_SIZE);
+    y[ED448_PUB_KEY_SIZE - 1] = 0;
+    for (i = 0; i < sizeof(small_order_y) / ED448_PUB_KEY_SIZE; i++) {
+        if (XMEMCMP(y, small_order_y[i], ED448_PUB_KEY_SIZE) == 0)
+            return 1;
+    }
+    return 0;
+}
+
 /* Derive the public key for the private key.
  *
  * key       [in]  Ed448 key object.
@@ -731,16 +802,11 @@ static int ed448_verify_msg_final_with_sha(const byte* sig, word32 sigLen,
     if (i == -1)
         return BAD_FUNC_ARG;
 
-    /* Reject identity public key (0,1): 0x01 followed by 56 zero bytes. */
-    {
-        int isIdentity = (key->p[0] == 0x01);
-        int j;
-        for (j = 1; j < ED448_PUB_KEY_SIZE && isIdentity; j++) {
-            if (key->p[j] != 0x00)
-                isIdentity = 0;
-        }
-        if (isIdentity)
-            return BAD_FUNC_ARG;
+    /* Defence in depth: also catch small-order keys imported with trusted=1. */
+    if (ed448_is_small_order(key->p)) {
+        WOLFSSL_MSG("Ed448 small-order public key rejected during "
+                    "signature verification");
+        return BAD_FUNC_ARG;
     }
 
     /* uncompress A (public key), test if valid, and negate it */
@@ -1412,6 +1478,13 @@ int wc_ed448_check_key(ed448_key* key)
         ret = PUBLIC_KEY_E;
     }
 
+    /* Reject small-order pub key before the priv-vs-pub compare so the
+     * diagnostic isn't masked by a "mismatch" error. */
+    if ((ret == 0) && ed448_is_small_order(key->p)) {
+        WOLFSSL_MSG("Ed448 small-order public key rejected during key check");
+        ret = PUBLIC_KEY_E;
+    }
+
     /* If we have a private key just make the public key and compare. */
     if ((ret == 0) && key->privKeySet) {
         ret = wc_ed448_make_public(key, pubKey, sizeof(pubKey));
@@ -1421,23 +1494,6 @@ int wc_ed448_check_key(ed448_key* key)
     }
     /* No private key, check Y is valid. */
     else if ((ret == 0) && (!key->privKeySet)) {
-        /* Reject the identity element (0, 1).
-         * Encoding: 0x01 followed by 56 zero bytes. */
-        {
-            int isIdentity = 1;
-            int i;
-            if (key->p[0] != 0x01)
-                isIdentity = 0;
-            for (i = 1; i < ED448_PUB_KEY_SIZE && isIdentity; i++) {
-                if (key->p[i] != 0x00)
-                    isIdentity = 0;
-            }
-            if (isIdentity) {
-                WOLFSSL_MSG("Ed448 public key is the identity element");
-                ret = PUBLIC_KEY_E;
-            }
-        }
-
         /* Verify that xQ and yQ are integers in the interval [0, p - 1].
          * Only have yQ so check that ordinate.
          * p = 2^448-2^224-1 = 0xff..fe..ff
diff --git a/wolfcrypt/src/evp.c b/wolfcrypt/src/evp.c
index 20f921ba37..c7e1e4c885 100644
--- a/wolfcrypt/src/evp.c
+++ b/wolfcrypt/src/evp.c
@@ -47,6 +47,12 @@
 #ifdef HAVE_ED448
 #include <wolfssl/wolfcrypt/ed448.h>
 #endif
+#ifdef HAVE_CURVE25519
+#include <wolfssl/wolfcrypt/curve25519.h>
+#endif
+#ifdef HAVE_CURVE448
+#include <wolfssl/wolfcrypt/curve448.h>
+#endif
 
 static const struct s_ent {
     const enum wc_HashType macType;
@@ -2767,7 +2773,7 @@ int wolfSSL_EVP_PKEY_CTX_ctrl_str(WOLFSSL_EVP_PKEY_CTX *ctx,
 #endif /* NO_WOLFSSL_STUB */
 
 #if (!defined(NO_DH) && defined(WOLFSSL_DH_EXTRA)) || defined(HAVE_ECC) || \
-    defined(HAVE_HKDF)
+    defined(HAVE_HKDF) || defined(HAVE_CURVE25519) || defined(HAVE_CURVE448)
 int wolfSSL_EVP_PKEY_derive(WOLFSSL_EVP_PKEY_CTX *ctx, unsigned char *key, size_t *keylen)
 {
     int len;
@@ -2884,6 +2890,54 @@ int wolfSSL_EVP_PKEY_derive(WOLFSSL_EVP_PKEY_CTX *ctx, unsigned char *key, size_
         *keylen = (size_t)len;
         break;
 #endif
+#ifdef HAVE_CURVE25519
+    case WC_EVP_PKEY_X25519:
+        if (!ctx->pkey->curve25519 || !ctx->peerKey->curve25519) {
+            return WOLFSSL_FAILURE;
+        }
+        len = CURVE25519_KEYSIZE;
+        if (key) {
+            word32 len32 = (word32)*keylen;
+            if (*keylen < (size_t)len) {
+                WOLFSSL_MSG("buffer too short");
+                return WOLFSSL_FAILURE;
+            }
+            /* X25519 shared secret is little-endian (RFC 7748). */
+            if (wc_curve25519_shared_secret_ex(ctx->pkey->curve25519,
+                    ctx->peerKey->curve25519, key, &len32,
+                    EC25519_LITTLE_ENDIAN) != 0) {
+                WOLFSSL_MSG("wc_curve25519_shared_secret_ex failed");
+                return WOLFSSL_FAILURE;
+            }
+            len = (int)len32;
+        }
+        *keylen = (size_t)len;
+        break;
+#endif
+#ifdef HAVE_CURVE448
+    case WC_EVP_PKEY_X448:
+        if (!ctx->pkey->curve448 || !ctx->peerKey->curve448) {
+            return WOLFSSL_FAILURE;
+        }
+        len = CURVE448_KEY_SIZE;
+        if (key) {
+            word32 len32 = (word32)*keylen;
+            if (*keylen < (size_t)len) {
+                WOLFSSL_MSG("buffer too short");
+                return WOLFSSL_FAILURE;
+            }
+            /* X448 shared secret is little-endian (RFC 7748). */
+            if (wc_curve448_shared_secret_ex(ctx->pkey->curve448,
+                    ctx->peerKey->curve448, key, &len32,
+                    EC448_LITTLE_ENDIAN) != 0) {
+                WOLFSSL_MSG("wc_curve448_shared_secret_ex failed");
+                return WOLFSSL_FAILURE;
+            }
+            len = (int)len32;
+        }
+        *keylen = (size_t)len;
+        break;
+#endif
 #ifdef HAVE_HKDF
     case WC_EVP_PKEY_HKDF:
         (void)len;
@@ -3761,6 +3815,12 @@ int wolfSSL_EVP_PKEY_keygen(WOLFSSL_EVP_PKEY_CTX *ctx,
         if (ctx->pkey == NULL ||
                 (ctx->pkey->type != WC_EVP_PKEY_EC &&
                  ctx->pkey->type != WC_EVP_PKEY_RSA &&
+            #ifdef HAVE_CURVE25519
+                 ctx->pkey->type != WC_EVP_PKEY_X25519 &&
+            #endif
+            #ifdef HAVE_CURVE448
+                 ctx->pkey->type != WC_EVP_PKEY_X448 &&
+            #endif
                  ctx->pkey->type != WC_EVP_PKEY_DH)) {
             WOLFSSL_MSG("Key not set or key type not supported");
             return WOLFSSL_FAILURE;
@@ -3821,6 +3881,57 @@ int wolfSSL_EVP_PKEY_keygen(WOLFSSL_EVP_PKEY_CTX *ctx,
                 }
             }
             break;
+#endif
+#ifdef HAVE_CURVE25519
+        case WC_EVP_PKEY_X25519:
+            if (pkey->curve25519 == NULL) {
+                pkey->curve25519 = (curve25519_key*)XMALLOC(
+                    sizeof(curve25519_key), pkey->heap, DYNAMIC_TYPE_CURVE25519);
+                if (pkey->curve25519 == NULL) {
+                    ret = MEMORY_E;
+                    break;
+                }
+                if (wc_curve25519_init_ex(pkey->curve25519, pkey->heap,
+                        INVALID_DEVID) != 0) {
+                    XFREE(pkey->curve25519, pkey->heap, DYNAMIC_TYPE_CURVE25519);
+                    pkey->curve25519 = NULL;
+                    break;
+                }
+            #ifdef WOLFSSL_CURVE25519_BLINDING
+                /* Use the EVP_PKEY's RNG for scalar blinding on derive. */
+                (void)wc_curve25519_set_rng(pkey->curve25519, &pkey->rng);
+            #endif
+                pkey->ownCurve25519 = 1;
+            }
+            /* Reuse the RNG already initialized on the EVP_PKEY. */
+            if (wc_curve25519_make_key(&pkey->rng, CURVE25519_KEYSIZE,
+                    pkey->curve25519) == 0) {
+                ret = WOLFSSL_SUCCESS;
+            }
+            break;
+#endif
+#ifdef HAVE_CURVE448
+        case WC_EVP_PKEY_X448:
+            if (pkey->curve448 == NULL) {
+                pkey->curve448 = (curve448_key*)XMALLOC(sizeof(curve448_key),
+                    pkey->heap, DYNAMIC_TYPE_CURVE448);
+                if (pkey->curve448 == NULL) {
+                    ret = MEMORY_E;
+                    break;
+                }
+                if (wc_curve448_init(pkey->curve448) != 0) {
+                    XFREE(pkey->curve448, pkey->heap, DYNAMIC_TYPE_CURVE448);
+                    pkey->curve448 = NULL;
+                    break;
+                }
+                pkey->ownCurve448 = 1;
+            }
+            /* Reuse the RNG already initialized on the EVP_PKEY. */
+            if (wc_curve448_make_key(&pkey->rng, CURVE448_KEY_SIZE,
+                    pkey->curve448) == 0) {
+                ret = WOLFSSL_SUCCESS;
+            }
+            break;
 #endif
         default:
             break;
@@ -3871,6 +3982,16 @@ int wolfSSL_EVP_PKEY_size(WOLFSSL_EVP_PKEY *pkey)
         return wc_ecc_sig_size((ecc_key*)(pkey->ecc->internal));
 #endif /* HAVE_ECC */
 
+#ifdef HAVE_CURVE25519
+    case WC_EVP_PKEY_X25519:
+        return CURVE25519_KEYSIZE;
+#endif
+
+#ifdef HAVE_CURVE448
+    case WC_EVP_PKEY_X448:
+        return CURVE448_KEY_SIZE;
+#endif
+
     default:
         break;
     }
@@ -11822,6 +11943,27 @@ void wolfSSL_EVP_PKEY_free(WOLFSSL_EVP_PKEY* key)
                     break;
                 #endif /* HAVE_ED448 */
 
+                #ifdef HAVE_CURVE25519
+                case WC_EVP_PKEY_X25519:
+                    if (key->curve25519 != NULL && key->ownCurve25519 == 1) {
+                        wc_curve25519_free(key->curve25519);
+                        XFREE(key->curve25519, key->heap,
+                            DYNAMIC_TYPE_CURVE25519);
+                        key->curve25519 = NULL;
+                    }
+                    break;
+                #endif /* HAVE_CURVE25519 */
+
+                #ifdef HAVE_CURVE448
+                case WC_EVP_PKEY_X448:
+                    if (key->curve448 != NULL && key->ownCurve448 == 1) {
+                        wc_curve448_free(key->curve448);
+                        XFREE(key->curve448, key->heap, DYNAMIC_TYPE_CURVE448);
+                        key->curve448 = NULL;
+                    }
+                    break;
+                #endif /* HAVE_CURVE448 */
+
                 #ifdef HAVE_HKDF
                 case WC_EVP_PKEY_HKDF:
                     XFREE(key->hkdfSalt, NULL, DYNAMIC_TYPE_SALT);
diff --git a/wolfcrypt/src/evp_pk.c b/wolfcrypt/src/evp_pk.c
index 03eb059b8c..9ef4d459bd 100644
--- a/wolfcrypt/src/evp_pk.c
+++ b/wolfcrypt/src/evp_pk.c
@@ -439,6 +439,64 @@ WOLFSSL_EVP_PKEY* wolfSSL_EVP_PKEY_new_raw_public_key(int type,
             ok = 1;
             break;
         }
+    #endif
+    #ifdef HAVE_CURVE25519
+        case WC_EVP_PKEY_X25519: {
+            curve25519_key* cKey;
+            if (len != CURVE25519_PUB_KEY_SIZE) {
+                break;
+            }
+            cKey = (curve25519_key*)XMALLOC(sizeof(curve25519_key), pkey->heap,
+                DYNAMIC_TYPE_CURVE25519);
+            if (cKey == NULL) {
+                break;
+            }
+            if (wc_curve25519_init_ex(cKey, pkey->heap, INVALID_DEVID) != 0) {
+                XFREE(cKey, pkey->heap, DYNAMIC_TYPE_CURVE25519);
+                break;
+            }
+            /* Raw X25519 keys are little-endian (RFC 7748). */
+            if (wc_curve25519_import_public_ex(pub, (word32)len, cKey,
+                    EC25519_LITTLE_ENDIAN) != 0) {
+                wc_curve25519_free(cKey);
+                XFREE(cKey, pkey->heap, DYNAMIC_TYPE_CURVE25519);
+                break;
+            }
+            pkey->type          = WC_EVP_PKEY_X25519;
+            pkey->curve25519    = cKey;
+            pkey->ownCurve25519 = 1;
+            ok = 1;
+            break;
+        }
+    #endif
+    #ifdef HAVE_CURVE448
+        case WC_EVP_PKEY_X448: {
+            curve448_key* cKey;
+            if (len != CURVE448_PUB_KEY_SIZE) {
+                break;
+            }
+            cKey = (curve448_key*)XMALLOC(sizeof(curve448_key), pkey->heap,
+                DYNAMIC_TYPE_CURVE448);
+            if (cKey == NULL) {
+                break;
+            }
+            if (wc_curve448_init(cKey) != 0) {
+                XFREE(cKey, pkey->heap, DYNAMIC_TYPE_CURVE448);
+                break;
+            }
+            /* Raw X448 keys are little-endian (RFC 7748). */
+            if (wc_curve448_import_public_ex(pub, (word32)len, cKey,
+                    EC448_LITTLE_ENDIAN) != 0) {
+                wc_curve448_free(cKey);
+                XFREE(cKey, pkey->heap, DYNAMIC_TYPE_CURVE448);
+                break;
+            }
+            pkey->type        = WC_EVP_PKEY_X448;
+            pkey->curve448    = cKey;
+            pkey->ownCurve448 = 1;
+            ok = 1;
+            break;
+        }
     #endif
         default:
             break;
@@ -526,6 +584,68 @@ WOLFSSL_EVP_PKEY* wolfSSL_EVP_PKEY_new_raw_private_key(int type,
             ok = 1;
             break;
         }
+    #endif
+    #ifdef HAVE_CURVE25519
+        case WC_EVP_PKEY_X25519: {
+            curve25519_key* cKey;
+            if (len != CURVE25519_KEYSIZE) {
+                break;
+            }
+            cKey = (curve25519_key*)XMALLOC(sizeof(curve25519_key), pkey->heap,
+                DYNAMIC_TYPE_CURVE25519);
+            if (cKey == NULL) {
+                break;
+            }
+            if (wc_curve25519_init_ex(cKey, pkey->heap, INVALID_DEVID) != 0) {
+                XFREE(cKey, pkey->heap, DYNAMIC_TYPE_CURVE25519);
+                break;
+            }
+        #ifdef WOLFSSL_CURVE25519_BLINDING
+            /* Use the EVP_PKEY's RNG for scalar blinding on shared-secret. */
+            (void)wc_curve25519_set_rng(cKey, &pkey->rng);
+        #endif
+            /* Raw X25519 keys are little-endian (RFC 7748). */
+            if (wc_curve25519_import_private_ex(priv, (word32)len, cKey,
+                    EC25519_LITTLE_ENDIAN) != 0) {
+                wc_curve25519_free(cKey);
+                XFREE(cKey, pkey->heap, DYNAMIC_TYPE_CURVE25519);
+                break;
+            }
+            pkey->type          = WC_EVP_PKEY_X25519;
+            pkey->curve25519    = cKey;
+            pkey->ownCurve25519 = 1;
+            ok = 1;
+            break;
+        }
+    #endif
+    #ifdef HAVE_CURVE448
+        case WC_EVP_PKEY_X448: {
+            curve448_key* cKey;
+            if (len != CURVE448_KEY_SIZE) {
+                break;
+            }
+            cKey = (curve448_key*)XMALLOC(sizeof(curve448_key), pkey->heap,
+                DYNAMIC_TYPE_CURVE448);
+            if (cKey == NULL) {
+                break;
+            }
+            if (wc_curve448_init(cKey) != 0) {
+                XFREE(cKey, pkey->heap, DYNAMIC_TYPE_CURVE448);
+                break;
+            }
+            /* Raw X448 keys are little-endian (RFC 7748). */
+            if (wc_curve448_import_private_ex(priv, (word32)len, cKey,
+                    EC448_LITTLE_ENDIAN) != 0) {
+                wc_curve448_free(cKey);
+                XFREE(cKey, pkey->heap, DYNAMIC_TYPE_CURVE448);
+                break;
+            }
+            pkey->type        = WC_EVP_PKEY_X448;
+            pkey->curve448    = cKey;
+            pkey->ownCurve448 = 1;
+            ok = 1;
+            break;
+        }
     #endif
         default:
             break;
@@ -942,7 +1062,8 @@ static int d2iTryMlDsaKey(WOLFSSL_EVP_PKEY** out, const unsigned char* mem,
         return WOLFSSL_FATAL_ERROR;
     }
 
-    return d2i_make_pkey(out, NULL, 0, priv, WC_EVP_PKEY_DILITHIUM);
+    /* Copy the consumed DER into pkey->pkey.ptr when the input was DER */
+    return d2i_make_pkey(out, mem, keyIdx, priv, WC_EVP_PKEY_DILITHIUM);
 }
 #endif /* WOLFSSL_HAVE_MLDSA */
 
@@ -1696,6 +1817,10 @@ WOLFSSL_PKCS8_PRIV_KEY_INFO* wolfSSL_d2i_PKCS8_PKEY(
     DerBuffer rawDer;
     EncryptedInfo info;
     int advanceLen = 0;
+#ifdef HAVE_DILITHIUM
+    word32 outerIdx = 0;
+    int    outerLen = 0;
+#endif
 
     /* Clear the encryption information and DER buffer. */
     XMEMSET(&info, 0, sizeof(info));
@@ -1737,15 +1862,40 @@ WOLFSSL_PKCS8_PRIV_KEY_INFO* wolfSSL_d2i_PKCS8_PKEY(
             }
             if (algId == DHk) {
                 /* Special case for DH as we expect the DER buffer to be always
-                 * be in PKCS8 format */
+                 * in PKCS8 format */
                 rawDer.buffer = pkcs8Der->buffer;
                 rawDer.length = inOutIdx + (word32)ret;
             }
+        #ifdef HAVE_DILITHIUM
+            else if (
+            #ifdef WOLFSSL_DILITHIUM_FIPS204_DRAFT
+                (algId == DILITHIUM_LEVEL2k) ||
+                (algId == DILITHIUM_LEVEL3k) ||
+                (algId == DILITHIUM_LEVEL5k) ||
+            #endif
+                (algId == ML_DSA_LEVEL2k) ||
+                (algId == ML_DSA_LEVEL3k) ||
+                (algId == ML_DSA_LEVEL5k)) {
+
+                /* Keep full PKCS#8 wrapper for level recovery from
+                 * AlgorithmIdentifier parameters */
+                rawDer.buffer = pkcs8Der->buffer;
+                if (GetSequence(pkcs8Der->buffer, &outerIdx, &outerLen,
+                    pkcs8Der->length) < 0) {
+                    ret = ASN_PARSE_E;
+                }
+                else {
+                    rawDer.length = outerIdx + (word32)outerLen;
+                }
+            }
+        #endif
             else {
                 rawDer.buffer = pkcs8Der->buffer + inOutIdx;
                 rawDer.length = (word32)ret;
             }
-            ret = 0; /* good DER */
+            if (ret > 0) {
+                ret = 0; /* good DER */
+            }
         }
     }
 
diff --git a/wolfcrypt/src/fe_x25519_asm.S b/wolfcrypt/src/fe_x25519_asm.S
index ed53856bca..3f0e0dd6a8 100644
--- a/wolfcrypt/src/fe_x25519_asm.S
+++ b/wolfcrypt/src/fe_x25519_asm.S
@@ -4639,6 +4639,7 @@ L_curve25519_base_x64_3:
         adcq	$0x00, %r8
         adcq	$0x00, %r9
         adcq	$0x00, %r10
+        andq	%rax, %r10
         # Store
         movq	%rcx, (%rdi)
         movq	%r8, 8(%rdi)
@@ -7054,6 +7055,7 @@ L_curve25519_x64_3:
         adcq	$0x00, %r9
         adcq	$0x00, %r10
         adcq	$0x00, %r11
+        andq	%rax, %r11
         # Store
         movq	%rcx, (%rdi)
         movq	%r9, 8(%rdi)
@@ -15107,6 +15109,7 @@ L_curve25519_base_avx2_last_3:
         adcq	$0x00, %r9
         adcq	$0x00, %r10
         adcq	$0x00, %r11
+        andq	%rcx, %r11
         # Store
         movq	%r8, (%rdi)
         movq	%r9, 8(%rdi)
@@ -17116,6 +17119,7 @@ L_curve25519_avx2_last_3:
         adcq	$0x00, %r10
         adcq	$0x00, %r11
         adcq	$0x00, %r12
+        andq	%rcx, %r12
         # Store
         movq	%r9, (%rdi)
         movq	%r10, 8(%rdi)
diff --git a/wolfcrypt/src/port/pic32/pic32mz-crypt.c b/wolfcrypt/src/port/pic32/pic32mz-crypt.c
index 75be92b306..2a86445957 100644
--- a/wolfcrypt/src/port/pic32/pic32mz-crypt.c
+++ b/wolfcrypt/src/port/pic32/pic32mz-crypt.c
@@ -657,9 +657,9 @@ static int wc_Pic32HashFinal(hashUpdCache* cache, byte* stdBuf,
     return ret;
 }
 
-static void wc_Pic32HashFree(hashUpdCache* cache, void* heap)
+static void wc_Pic32HashFree(hashUpdCache* cache, void* stdBuf, void* heap)
 {
-    if (cache && cache->buf && !cache->isCopy) {
+    if (cache && cache->buf && cache->buf != stdBuf && !cache->isCopy) {
         XFREE(cache->buf, heap, DYNAMIC_TYPE_HASH_TMP);
         cache->buf = NULL;
     }
@@ -713,7 +713,7 @@ static void wc_Pic32HashFree(hashUpdCache* cache, void* heap)
     void wc_Md5Pic32Free(wc_Md5* md5)
     {
         if (md5) {
-            wc_Pic32HashFree(&md5->cache, md5->heap);
+            wc_Pic32HashFree(&md5->cache, (byte*)md5->buffer, md5->heap);
         }
     }
 #endif /* !NO_MD5 */
@@ -764,7 +764,7 @@ static void wc_Pic32HashFree(hashUpdCache* cache, void* heap)
     void wc_ShaPic32Free(wc_Sha* sha)
     {
         if (sha) {
-            wc_Pic32HashFree(&sha->cache, sha->heap);
+            wc_Pic32HashFree(&sha->cache, (byte*)sha->buffer, sha->heap);
         }
     }
 #endif /* !NO_SHA */
@@ -815,7 +815,7 @@ static void wc_Pic32HashFree(hashUpdCache* cache, void* heap)
     void wc_Sha256Pic32Free(wc_Sha256* sha256)
     {
         if (sha256) {
-            wc_Pic32HashFree(&sha256->cache, sha256->heap);
+            wc_Pic32HashFree(&sha256->cache, (byte*)sha256->buffer, sha256->heap);
         }
     }
 #endif /* !NO_SHA256 */
diff --git a/wolfcrypt/src/port/ppc64/ppc64-aes-asm.S b/wolfcrypt/src/port/ppc64/ppc64-aes-asm.S
new file mode 100644
index 0000000000..4e3c681efc
--- /dev/null
+++ b/wolfcrypt/src/port/ppc64/ppc64-aes-asm.S
@@ -0,0 +1,8627 @@
+/* ppc64-aes-asm
+ *
+ * Copyright (C) 2006-2025 wolfSSL Inc.
+ *
+ * This file is part of wolfSSL.
+ *
+ * wolfSSL is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * wolfSSL is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1335, USA
+ */
+
+#include <wolfssl/wolfcrypt/libwolfssl_sources_asm.h>
+
+/* Generated using (from wolfssl):
+ *   cd ../scripts
+ *   ruby ./aes/aes.rb ppc64 \
+ *       ../wolfssl/wolfcrypt/src/port/ppc64/ppc64-aes-asm.S
+ */
+#ifdef WOLFSSL_PPC64_ASM
+#ifndef WOLFSSL_PPC64_ASM_INLINE
+#if !defined(NO_AES) && defined(WOLFSSL_PPC64_ASM)
+#ifdef HAVE_AES_DECRYPT
+#ifndef __APPLE__
+	.text
+	.section	.rodata
+	.type	L_AES_PPC64_td, @object
+	.size	L_AES_PPC64_td, 1024
+#else
+	.section	__DATA,__data
+#endif /* __APPLE__ */
+#ifndef __APPLE__
+	.align	4
+#else
+	.p2align	4
+#endif /* __APPLE__ */
+L_AES_PPC64_td:
+	.long	0x5051f4a7
+	.long	0x537e4165
+	.long	0xc31a17a4
+	.long	0x963a275e
+	.long	0xcb3bab6b
+	.long	0xf11f9d45
+	.long	0xabacfa58
+	.long	0x934be303
+	.long	0x552030fa
+	.long	0xf6ad766d
+	.long	0x9188cc76
+	.long	0x25f5024c
+	.long	0xfc4fe5d7
+	.long	0xd7c52acb
+	.long	0x80263544
+	.long	0x8fb562a3
+	.long	0x49deb15a
+	.long	0x6725ba1b
+	.long	0x9845ea0e
+	.long	0xe15dfec0
+	.long	0x2c32f75
+	.long	0x12814cf0
+	.long	0xa38d4697
+	.long	0xc66bd3f9
+	.long	0xe7038f5f
+	.long	0x9515929c
+	.long	0xebbf6d7a
+	.long	0xda955259
+	.long	0x2dd4be83
+	.long	0xd3587421
+	.long	0x2949e069
+	.long	0x448ec9c8
+	.long	0x6a75c289
+	.long	0x78f48e79
+	.long	0x6b99583e
+	.long	0xdd27b971
+	.long	0xb6bee14f
+	.long	0x17f088ad
+	.long	0x66c920ac
+	.long	0xb47dce3a
+	.long	0x1863df4a
+	.long	0x82e51a31
+	.long	0x60975133
+	.long	0x4562537f
+	.long	0xe0b16477
+	.long	0x84bb6bae
+	.long	0x1cfe81a0
+	.long	0x94f9082b
+	.long	0x58704868
+	.long	0x198f45fd
+	.long	0x8794de6c
+	.long	0xb7527bf8
+	.long	0x23ab73d3
+	.long	0xe2724b02
+	.long	0x57e31f8f
+	.long	0x2a6655ab
+	.long	0x7b2eb28
+	.long	0x32fb5c2
+	.long	0x9a86c57b
+	.long	0xa5d33708
+	.long	0xf2302887
+	.long	0xb223bfa5
+	.long	0xba02036a
+	.long	0x5ced1682
+	.long	0x2b8acf1c
+	.long	0x92a779b4
+	.long	0xf0f307f2
+	.long	0xa14e69e2
+	.long	0xcd65daf4
+	.long	0xd50605be
+	.long	0x1fd13462
+	.long	0x8ac4a6fe
+	.long	0x9d342e53
+	.long	0xa0a2f355
+	.long	0x32058ae1
+	.long	0x75a4f6eb
+	.long	0x390b83ec
+	.long	0xaa4060ef
+	.long	0x65e719f
+	.long	0x51bd6e10
+	.long	0xf93e218a
+	.long	0x3d96dd06
+	.long	0xaedd3e05
+	.long	0x464de6bd
+	.long	0xb591548d
+	.long	0x571c45d
+	.long	0x6f0406d4
+	.long	0xff605015
+	.long	0x241998fb
+	.long	0x97d6bde9
+	.long	0xcc894043
+	.long	0x7767d99e
+	.long	0xbdb0e842
+	.long	0x8807898b
+	.long	0x38e7195b
+	.long	0xdb79c8ee
+	.long	0x47a17c0a
+	.long	0xe97c420f
+	.long	0xc9f8841e
+	.long	0x0
+	.long	0x83098086
+	.long	0x48322bed
+	.long	0xac1e1170
+	.long	0x4e6c5a72
+	.long	0xfbfd0eff
+	.long	0x560f8538
+	.long	0x1e3daed5
+	.long	0x27362d39
+	.long	0x640a0fd9
+	.long	0x21685ca6
+	.long	0xd19b5b54
+	.long	0x3a24362e
+	.long	0xb10c0a67
+	.long	0xf9357e7
+	.long	0xd2b4ee96
+	.long	0x9e1b9b91
+	.long	0x4f80c0c5
+	.long	0xa261dc20
+	.long	0x695a774b
+	.long	0x161c121a
+	.long	0xae293ba
+	.long	0xe5c0a02a
+	.long	0x433c22e0
+	.long	0x1d121b17
+	.long	0xb0e090d
+	.long	0xadf28bc7
+	.long	0xb92db6a8
+	.long	0xc8141ea9
+	.long	0x8557f119
+	.long	0x4caf7507
+	.long	0xbbee99dd
+	.long	0xfda37f60
+	.long	0x9ff70126
+	.long	0xbc5c72f5
+	.long	0xc544663b
+	.long	0x345bfb7e
+	.long	0x768b4329
+	.long	0xdccb23c6
+	.long	0x68b6edfc
+	.long	0x63b8e4f1
+	.long	0xcad731dc
+	.long	0x10426385
+	.long	0x40139722
+	.long	0x2084c611
+	.long	0x7d854a24
+	.long	0xf8d2bb3d
+	.long	0x11aef932
+	.long	0x6dc729a1
+	.long	0x4b1d9e2f
+	.long	0xf3dcb230
+	.long	0xec0d8652
+	.long	0xd077c1e3
+	.long	0x6c2bb316
+	.long	0x99a970b9
+	.long	0xfa119448
+	.long	0x2247e964
+	.long	0xc4a8fc8c
+	.long	0x1aa0f03f
+	.long	0xd8567d2c
+	.long	0xef223390
+	.long	0xc787494e
+	.long	0xc1d938d1
+	.long	0xfe8ccaa2
+	.long	0x3698d40b
+	.long	0xcfa6f581
+	.long	0x28a57ade
+	.long	0x26dab78e
+	.long	0xa43fadbf
+	.long	0xe42c3a9d
+	.long	0xd507892
+	.long	0x9b6a5fcc
+	.long	0x62547e46
+	.long	0xc2f68d13
+	.long	0xe890d8b8
+	.long	0x5e2e39f7
+	.long	0xf582c3af
+	.long	0xbe9f5d80
+	.long	0x7c69d093
+	.long	0xa96fd52d
+	.long	0xb3cf2512
+	.long	0x3bc8ac99
+	.long	0xa710187d
+	.long	0x6ee89c63
+	.long	0x7bdb3bbb
+	.long	0x9cd2678
+	.long	0xf46e5918
+	.long	0x1ec9ab7
+	.long	0xa8834f9a
+	.long	0x65e6956e
+	.long	0x7eaaffe6
+	.long	0x821bccf
+	.long	0xe6ef15e8
+	.long	0xd9bae79b
+	.long	0xce4a6f36
+	.long	0xd4ea9f09
+	.long	0xd629b07c
+	.long	0xaf31a4b2
+	.long	0x312a3f23
+	.long	0x30c6a594
+	.long	0xc035a266
+	.long	0x37744ebc
+	.long	0xa6fc82ca
+	.long	0xb0e090d0
+	.long	0x1533a7d8
+	.long	0x4af10498
+	.long	0xf741ecda
+	.long	0xe7fcd50
+	.long	0x2f1791f6
+	.long	0x8d764dd6
+	.long	0x4d43efb0
+	.long	0x54ccaa4d
+	.long	0xdfe49604
+	.long	0xe39ed1b5
+	.long	0x1b4c6a88
+	.long	0xb8c12c1f
+	.long	0x7f466551
+	.long	0x49d5eea
+	.long	0x5d018c35
+	.long	0x73fa8774
+	.long	0x2efb0b41
+	.long	0x5ab3671d
+	.long	0x5292dbd2
+	.long	0x33e91056
+	.long	0x136dd647
+	.long	0x8c9ad761
+	.long	0x7a37a10c
+	.long	0x8e59f814
+	.long	0x89eb133c
+	.long	0xeecea927
+	.long	0x35b761c9
+	.long	0xede11ce5
+	.long	0x3c7a47b1
+	.long	0x599cd2df
+	.long	0x3f55f273
+	.long	0x791814ce
+	.long	0xbf73c737
+	.long	0xea53f7cd
+	.long	0x5b5ffdaa
+	.long	0x14df3d6f
+	.long	0x867844db
+	.long	0x81caaff3
+	.long	0x3eb968c4
+	.long	0x2c382434
+	.long	0x5fc2a340
+	.long	0x72161dc3
+	.long	0xcbce225
+	.long	0x8b283c49
+	.long	0x41ff0d95
+	.long	0x7139a801
+	.long	0xde080cb3
+	.long	0x9cd8b4e4
+	.long	0x906456c1
+	.long	0x617bcb84
+	.long	0x70d532b6
+	.long	0x74486c5c
+	.long	0x42d0b857
+#endif /* HAVE_AES_DECRYPT */
+#if defined(HAVE_AES_DECRYPT) || defined(HAVE_AES_CBC) || \
+    defined(HAVE_AESCCM) || defined(HAVE_AESGCM) || \
+    defined(WOLFSSL_AES_DIRECT) || defined(WOLFSSL_AES_COUNTER)
+#ifndef __APPLE__
+	.text
+	.section	.rodata
+	.type	L_AES_PPC64_te, @object
+	.size	L_AES_PPC64_te, 1024
+#else
+	.section	__DATA,__data
+#endif /* __APPLE__ */
+#ifndef __APPLE__
+	.align	4
+#else
+	.p2align	4
+#endif /* __APPLE__ */
+L_AES_PPC64_te:
+	.long	0xa5c66363
+	.long	0x84f87c7c
+	.long	0x99ee7777
+	.long	0x8df67b7b
+	.long	0xdfff2f2
+	.long	0xbdd66b6b
+	.long	0xb1de6f6f
+	.long	0x5491c5c5
+	.long	0x50603030
+	.long	0x3020101
+	.long	0xa9ce6767
+	.long	0x7d562b2b
+	.long	0x19e7fefe
+	.long	0x62b5d7d7
+	.long	0xe64dabab
+	.long	0x9aec7676
+	.long	0x458fcaca
+	.long	0x9d1f8282
+	.long	0x4089c9c9
+	.long	0x87fa7d7d
+	.long	0x15effafa
+	.long	0xebb25959
+	.long	0xc98e4747
+	.long	0xbfbf0f0
+	.long	0xec41adad
+	.long	0x67b3d4d4
+	.long	0xfd5fa2a2
+	.long	0xea45afaf
+	.long	0xbf239c9c
+	.long	0xf753a4a4
+	.long	0x96e47272
+	.long	0x5b9bc0c0
+	.long	0xc275b7b7
+	.long	0x1ce1fdfd
+	.long	0xae3d9393
+	.long	0x6a4c2626
+	.long	0x5a6c3636
+	.long	0x417e3f3f
+	.long	0x2f5f7f7
+	.long	0x4f83cccc
+	.long	0x5c683434
+	.long	0xf451a5a5
+	.long	0x34d1e5e5
+	.long	0x8f9f1f1
+	.long	0x93e27171
+	.long	0x73abd8d8
+	.long	0x53623131
+	.long	0x3f2a1515
+	.long	0xc080404
+	.long	0x5295c7c7
+	.long	0x65462323
+	.long	0x5e9dc3c3
+	.long	0x28301818
+	.long	0xa1379696
+	.long	0xf0a0505
+	.long	0xb52f9a9a
+	.long	0x90e0707
+	.long	0x36241212
+	.long	0x9b1b8080
+	.long	0x3ddfe2e2
+	.long	0x26cdebeb
+	.long	0x694e2727
+	.long	0xcd7fb2b2
+	.long	0x9fea7575
+	.long	0x1b120909
+	.long	0x9e1d8383
+	.long	0x74582c2c
+	.long	0x2e341a1a
+	.long	0x2d361b1b
+	.long	0xb2dc6e6e
+	.long	0xeeb45a5a
+	.long	0xfb5ba0a0
+	.long	0xf6a45252
+	.long	0x4d763b3b
+	.long	0x61b7d6d6
+	.long	0xce7db3b3
+	.long	0x7b522929
+	.long	0x3edde3e3
+	.long	0x715e2f2f
+	.long	0x97138484
+	.long	0xf5a65353
+	.long	0x68b9d1d1
+	.long	0x0
+	.long	0x2cc1eded
+	.long	0x60402020
+	.long	0x1fe3fcfc
+	.long	0xc879b1b1
+	.long	0xedb65b5b
+	.long	0xbed46a6a
+	.long	0x468dcbcb
+	.long	0xd967bebe
+	.long	0x4b723939
+	.long	0xde944a4a
+	.long	0xd4984c4c
+	.long	0xe8b05858
+	.long	0x4a85cfcf
+	.long	0x6bbbd0d0
+	.long	0x2ac5efef
+	.long	0xe54faaaa
+	.long	0x16edfbfb
+	.long	0xc5864343
+	.long	0xd79a4d4d
+	.long	0x55663333
+	.long	0x94118585
+	.long	0xcf8a4545
+	.long	0x10e9f9f9
+	.long	0x6040202
+	.long	0x81fe7f7f
+	.long	0xf0a05050
+	.long	0x44783c3c
+	.long	0xba259f9f
+	.long	0xe34ba8a8
+	.long	0xf3a25151
+	.long	0xfe5da3a3
+	.long	0xc0804040
+	.long	0x8a058f8f
+	.long	0xad3f9292
+	.long	0xbc219d9d
+	.long	0x48703838
+	.long	0x4f1f5f5
+	.long	0xdf63bcbc
+	.long	0xc177b6b6
+	.long	0x75afdada
+	.long	0x63422121
+	.long	0x30201010
+	.long	0x1ae5ffff
+	.long	0xefdf3f3
+	.long	0x6dbfd2d2
+	.long	0x4c81cdcd
+	.long	0x14180c0c
+	.long	0x35261313
+	.long	0x2fc3ecec
+	.long	0xe1be5f5f
+	.long	0xa2359797
+	.long	0xcc884444
+	.long	0x392e1717
+	.long	0x5793c4c4
+	.long	0xf255a7a7
+	.long	0x82fc7e7e
+	.long	0x477a3d3d
+	.long	0xacc86464
+	.long	0xe7ba5d5d
+	.long	0x2b321919
+	.long	0x95e67373
+	.long	0xa0c06060
+	.long	0x98198181
+	.long	0xd19e4f4f
+	.long	0x7fa3dcdc
+	.long	0x66442222
+	.long	0x7e542a2a
+	.long	0xab3b9090
+	.long	0x830b8888
+	.long	0xca8c4646
+	.long	0x29c7eeee
+	.long	0xd36bb8b8
+	.long	0x3c281414
+	.long	0x79a7dede
+	.long	0xe2bc5e5e
+	.long	0x1d160b0b
+	.long	0x76addbdb
+	.long	0x3bdbe0e0
+	.long	0x56643232
+	.long	0x4e743a3a
+	.long	0x1e140a0a
+	.long	0xdb924949
+	.long	0xa0c0606
+	.long	0x6c482424
+	.long	0xe4b85c5c
+	.long	0x5d9fc2c2
+	.long	0x6ebdd3d3
+	.long	0xef43acac
+	.long	0xa6c46262
+	.long	0xa8399191
+	.long	0xa4319595
+	.long	0x37d3e4e4
+	.long	0x8bf27979
+	.long	0x32d5e7e7
+	.long	0x438bc8c8
+	.long	0x596e3737
+	.long	0xb7da6d6d
+	.long	0x8c018d8d
+	.long	0x64b1d5d5
+	.long	0xd29c4e4e
+	.long	0xe049a9a9
+	.long	0xb4d86c6c
+	.long	0xfaac5656
+	.long	0x7f3f4f4
+	.long	0x25cfeaea
+	.long	0xafca6565
+	.long	0x8ef47a7a
+	.long	0xe947aeae
+	.long	0x18100808
+	.long	0xd56fbaba
+	.long	0x88f07878
+	.long	0x6f4a2525
+	.long	0x725c2e2e
+	.long	0x24381c1c
+	.long	0xf157a6a6
+	.long	0xc773b4b4
+	.long	0x5197c6c6
+	.long	0x23cbe8e8
+	.long	0x7ca1dddd
+	.long	0x9ce87474
+	.long	0x213e1f1f
+	.long	0xdd964b4b
+	.long	0xdc61bdbd
+	.long	0x860d8b8b
+	.long	0x850f8a8a
+	.long	0x90e07070
+	.long	0x427c3e3e
+	.long	0xc471b5b5
+	.long	0xaacc6666
+	.long	0xd8904848
+	.long	0x5060303
+	.long	0x1f7f6f6
+	.long	0x121c0e0e
+	.long	0xa3c26161
+	.long	0x5f6a3535
+	.long	0xf9ae5757
+	.long	0xd069b9b9
+	.long	0x91178686
+	.long	0x5899c1c1
+	.long	0x273a1d1d
+	.long	0xb9279e9e
+	.long	0x38d9e1e1
+	.long	0x13ebf8f8
+	.long	0xb32b9898
+	.long	0x33221111
+	.long	0xbbd26969
+	.long	0x70a9d9d9
+	.long	0x89078e8e
+	.long	0xa7339494
+	.long	0xb62d9b9b
+	.long	0x223c1e1e
+	.long	0x92158787
+	.long	0x20c9e9e9
+	.long	0x4987cece
+	.long	0xffaa5555
+	.long	0x78502828
+	.long	0x7aa5dfdf
+	.long	0x8f038c8c
+	.long	0xf859a1a1
+	.long	0x80098989
+	.long	0x171a0d0d
+	.long	0xda65bfbf
+	.long	0x31d7e6e6
+	.long	0xc6844242
+	.long	0xb8d06868
+	.long	0xc3824141
+	.long	0xb0299999
+	.long	0x775a2d2d
+	.long	0x111e0f0f
+	.long	0xcb7bb0b0
+	.long	0xfca85454
+	.long	0xd66dbbbb
+	.long	0x3a2c1616
+#endif /* HAVE_AES_DECRYPT || HAVE_AES_CBC || HAVE_AESCCM || HAVE_AESGCM ||
+        * WOLFSSL_AES_DIRECT || WOLFSSL_AES_COUNTER */
+#ifdef HAVE_AES_DECRYPT
+#ifndef __APPLE__
+.text
+.globl	AES_invert_key
+.type	AES_invert_key,@function
+.align	16
+AES_invert_key:
+#else
+.section	__TEXT,__text
+.globl	_AES_invert_key
+.p2align	4
+_AES_invert_key:
+#endif /* __APPLE__ */
+	stdu	1, -32(1)
+	mflr	0
+	std	0, 24(1)
+	std	14, 0(1)
+	std	15, 8(1)
+	std	16, 16(1)
+#if defined(__ARCH_PWR8) || (defined(_ARCH_PWR8))
+	addis	5, 2, L_AES_PPC64_te@toc@ha
+	addi	5, 5, L_AES_PPC64_te@toc@l
+#else
+	lis	5, L_AES_PPC64_te@ha
+	la	5, L_AES_PPC64_te@l(5)
+#endif /* defined(__ARCH_PWR8) || (defined(_ARCH_PWR8)) */
+#if defined(__ARCH_PWR8) || (defined(_ARCH_PWR8))
+	addis	6, 2, L_AES_PPC64_td@toc@ha
+	addi	6, 6, L_AES_PPC64_td@toc@l
+#else
+	lis	6, L_AES_PPC64_td@ha
+	la	6, L_AES_PPC64_td@l(6)
+#endif /* defined(__ARCH_PWR8) || (defined(_ARCH_PWR8)) */
+	addi	5, 5, 3
+	sldi	16, 4, 4
+	add	16, 16, 3
+	ld	7, 0(3)
+	ld	8, 8(3)
+	ld	11, 0(16)
+	ld	12, 8(16)
+	rldicr	7, 7, 32, 63
+	rldicr	8, 8, 32, 63
+	rldicr	11, 11, 32, 63
+	rldicr	12, 12, 32, 63
+	std	7, 0(16)
+	std	8, 8(16)
+	std	11, 0(3)
+	std	12, 8(3)
+	addi	3, 3, 16
+	addi	16, 16, -16
+	addi	0, 4, -2
+	srdi	0, 0, 1
+	mtctr	0
+L_AES_invert_key_loop:
+	ld	7, 0(3)
+	ld	8, 8(3)
+	ld	11, 0(16)
+	ld	12, 8(16)
+	std	7, 0(16)
+	std	8, 8(16)
+	std	11, 0(3)
+	std	12, 8(3)
+	addi	3, 3, 16
+	addi	16, 16, -16
+	bdnz	L_AES_invert_key_loop
+	sldi	16, 4, 3
+	subf	3, 16, 3
+	addi	3, 3, 16
+	addi	0, 4, -1
+	mtctr	0
+L_AES_invert_key_mix_loop:
+	lwz	8, 0(3)
+	lwz	7, 4(3)
+	lwz	10, 8(3)
+	lwz	9, 12(3)
+	rlwinm	11, 7, 2, 22, 29
+	rlwinm	12, 7, 26, 22, 29
+	rlwinm	14, 7, 18, 22, 29
+	rlwinm	15, 7, 10, 22, 29
+	lbzx	11, 5, 11
+	lbzx	12, 5, 12
+	lbzx	14, 5, 14
+	lbzx	15, 5, 15
+	slwi	11, 11, 2
+	slwi	12, 12, 2
+	slwi	14, 14, 2
+	slwi	15, 15, 2
+	lwzx	11, 6, 11
+	lwzx	12, 6, 12
+	lwzx	14, 6, 14
+	lwzx	15, 6, 15
+	rlwimi	11, 11, 16, 0, 31
+	rlwimi	12, 12, 24, 0, 31
+	rlwimi	15, 15, 8, 0, 31
+	xor	14, 14, 11
+	xor	14, 14, 12
+	xor	14, 14, 15
+	stw	14, 4(3)
+	rlwinm	11, 8, 2, 22, 29
+	rlwinm	12, 8, 26, 22, 29
+	rlwinm	14, 8, 18, 22, 29
+	rlwinm	15, 8, 10, 22, 29
+	lbzx	11, 5, 11
+	lbzx	12, 5, 12
+	lbzx	14, 5, 14
+	lbzx	15, 5, 15
+	slwi	11, 11, 2
+	slwi	12, 12, 2
+	slwi	14, 14, 2
+	slwi	15, 15, 2
+	lwzx	11, 6, 11
+	lwzx	12, 6, 12
+	lwzx	14, 6, 14
+	lwzx	15, 6, 15
+	rlwimi	11, 11, 16, 0, 31
+	rlwimi	12, 12, 24, 0, 31
+	rlwimi	15, 15, 8, 0, 31
+	xor	14, 14, 11
+	xor	14, 14, 12
+	xor	14, 14, 15
+	stw	14, 0(3)
+	addi	3, 3, 8
+	rlwinm	11, 9, 2, 22, 29
+	rlwinm	12, 9, 26, 22, 29
+	rlwinm	14, 9, 18, 22, 29
+	rlwinm	15, 9, 10, 22, 29
+	lbzx	11, 5, 11
+	lbzx	12, 5, 12
+	lbzx	14, 5, 14
+	lbzx	15, 5, 15
+	slwi	11, 11, 2
+	slwi	12, 12, 2
+	slwi	14, 14, 2
+	slwi	15, 15, 2
+	lwzx	11, 6, 11
+	lwzx	12, 6, 12
+	lwzx	14, 6, 14
+	lwzx	15, 6, 15
+	rlwimi	11, 11, 16, 0, 31
+	rlwimi	12, 12, 24, 0, 31
+	rlwimi	15, 15, 8, 0, 31
+	xor	14, 14, 11
+	xor	14, 14, 12
+	xor	14, 14, 15
+	stw	14, 4(3)
+	rlwinm	11, 10, 2, 22, 29
+	rlwinm	12, 10, 26, 22, 29
+	rlwinm	14, 10, 18, 22, 29
+	rlwinm	15, 10, 10, 22, 29
+	lbzx	11, 5, 11
+	lbzx	12, 5, 12
+	lbzx	14, 5, 14
+	lbzx	15, 5, 15
+	slwi	11, 11, 2
+	slwi	12, 12, 2
+	slwi	14, 14, 2
+	slwi	15, 15, 2
+	lwzx	11, 6, 11
+	lwzx	12, 6, 12
+	lwzx	14, 6, 14
+	lwzx	15, 6, 15
+	rlwimi	11, 11, 16, 0, 31
+	rlwimi	12, 12, 24, 0, 31
+	rlwimi	15, 15, 8, 0, 31
+	xor	14, 14, 11
+	xor	14, 14, 12
+	xor	14, 14, 15
+	stw	14, 0(3)
+	addi	3, 3, 8
+	bdnz	L_AES_invert_key_mix_loop
+	ld	0, 24(1)
+	mtlr	0
+	ld	14, 0(1)
+	ld	15, 8(1)
+	ld	16, 16(1)
+	addi	1, 1, 32
+	blr
+#ifndef __APPLE__
+	.size	AES_invert_key,.-AES_invert_key
+#endif /* __APPLE__ */
+#endif /* HAVE_AES_DECRYPT */
+#ifndef __APPLE__
+	.text
+	.section	.rodata
+	.type	L_AES_PPC64_rcon, @object
+	.size	L_AES_PPC64_rcon, 40
+#else
+	.section	__DATA,__data
+#endif /* __APPLE__ */
+#ifndef __APPLE__
+	.align	4
+#else
+	.p2align	4
+#endif /* __APPLE__ */
+L_AES_PPC64_rcon:
+	.long	0x1000000
+	.long	0x2000000
+	.long	0x4000000
+	.long	0x8000000
+	.long	0x10000000
+	.long	0x20000000
+	.long	0x40000000
+	.long	0x80000000
+	.long	0x1b000000
+	.long	0x36000000
+#ifndef __APPLE__
+.text
+.globl	AES_set_encrypt_key
+.type	AES_set_encrypt_key,@function
+.align	16
+AES_set_encrypt_key:
+#else
+.section	__TEXT,__text
+.globl	_AES_set_encrypt_key
+.p2align	4
+_AES_set_encrypt_key:
+#endif /* __APPLE__ */
+	stdu	1, -24(1)
+	mflr	0
+	std	0, 16(1)
+	std	14, 0(1)
+	std	15, 8(1)
+#if defined(__ARCH_PWR8) || (defined(_ARCH_PWR8))
+	addis	6, 2, L_AES_PPC64_te@toc@ha
+	addi	6, 6, L_AES_PPC64_te@toc@l
+#else
+	lis	6, L_AES_PPC64_te@ha
+	la	6, L_AES_PPC64_te@l(6)
+#endif /* defined(__ARCH_PWR8) || (defined(_ARCH_PWR8)) */
+#if defined(__ARCH_PWR8) || (defined(_ARCH_PWR8))
+	addis	7, 2, L_AES_PPC64_rcon@toc@ha
+	addi	7, 7, L_AES_PPC64_rcon@toc@l
+#else
+	lis	7, L_AES_PPC64_rcon@ha
+	la	7, L_AES_PPC64_rcon@l(7)
+#endif /* defined(__ARCH_PWR8) || (defined(_ARCH_PWR8)) */
+	addi	6, 6, 3
+	cmplwi	4, 0x80
+	beq	L_AES_set_encrypt_key_start_128
+	cmplwi	4, 0xc0
+	beq	L_AES_set_encrypt_key_start_192
+	ld	9, 0(3)
+	ld	10, 8(3)
+	ld	11, 16(3)
+	ld	12, 24(3)
+	rldicr	9, 9, 32, 63
+	rldicr	10, 10, 32, 63
+	rldicr	11, 11, 32, 63
+	rldicr	12, 12, 32, 63
+	std	9, 0(5)
+	std	10, 8(5)
+	std	11, 16(5)
+	std	12, 24(5)
+	rldicr	12, 12, 32, 63
+	li	8, 6
+	mtctr	8
+L_AES_set_encrypt_key_loop_256:
+	rlwinm	9, 12, 2, 22, 29
+	rlwinm	10, 12, 26, 22, 29
+	rlwinm	11, 12, 18, 22, 29
+	rlwinm	12, 12, 10, 22, 29
+	lbzx	9, 6, 9
+	lbzx	10, 6, 10
+	lbzx	11, 6, 11
+	lbzx	0, 6, 12
+	rlwimi	0, 9, 8, 16, 23
+	rlwimi	0, 10, 16, 8, 15
+	rlwimi	0, 11, 24, 0, 7
+	lwz	10, 0(5)
+	lwz	9, 4(5)
+	lwz	12, 8(5)
+	lwz	11, 12(5)
+	addi	5, 5, 16
+	xor	9, 9, 0
+	lwz	0, 0(7)
+	addi	7, 7, 4
+	xor	9, 9, 0
+	xor	10, 10, 9
+	xor	11, 11, 10
+	xor	12, 12, 11
+	stw	10, 16(5)
+	stw	9, 20(5)
+	stw	12, 24(5)
+	stw	11, 28(5)
+	rlwinm	9, 12, 26, 22, 29
+	rlwinm	10, 12, 18, 22, 29
+	rlwinm	11, 12, 10, 22, 29
+	rlwinm	12, 12, 2, 22, 29
+	lbzx	9, 6, 9
+	lbzx	10, 6, 10
+	lbzx	11, 6, 11
+	lbzx	0, 6, 12
+	rlwimi	0, 9, 8, 16, 23
+	rlwimi	0, 10, 16, 8, 15
+	rlwimi	0, 11, 24, 0, 7
+	lwz	10, 0(5)
+	lwz	9, 4(5)
+	lwz	12, 8(5)
+	lwz	11, 12(5)
+	addi	5, 5, 16
+	xor	9, 9, 0
+	xor	10, 10, 9
+	xor	11, 11, 10
+	xor	12, 12, 11
+	stw	10, 16(5)
+	stw	9, 20(5)
+	stw	12, 24(5)
+	stw	11, 28(5)
+	bdnz	L_AES_set_encrypt_key_loop_256
+	rlwinm	9, 12, 2, 22, 29
+	rlwinm	10, 12, 26, 22, 29
+	rlwinm	11, 12, 18, 22, 29
+	rlwinm	12, 12, 10, 22, 29
+	lbzx	9, 6, 9
+	lbzx	10, 6, 10
+	lbzx	11, 6, 11
+	lbzx	0, 6, 12
+	rlwimi	0, 9, 8, 16, 23
+	rlwimi	0, 10, 16, 8, 15
+	rlwimi	0, 11, 24, 0, 7
+	lwz	10, 0(5)
+	lwz	9, 4(5)
+	lwz	12, 8(5)
+	lwz	11, 12(5)
+	addi	5, 5, 16
+	xor	9, 9, 0
+	lwz	0, 0(7)
+	addi	7, 7, 4
+	xor	9, 9, 0
+	xor	10, 10, 9
+	xor	11, 11, 10
+	xor	12, 12, 11
+	stw	9, 16(5)
+	stw	10, 20(5)
+	stw	11, 24(5)
+	stw	12, 28(5)
+	b	L_AES_set_encrypt_key_end
+L_AES_set_encrypt_key_start_192:
+	ld	12, 0(3)
+	ld	14, 8(3)
+	ld	15, 16(3)
+	rldicr	12, 12, 32, 63
+	rldicr	14, 14, 32, 63
+	rldicr	15, 15, 32, 63
+	std	12, 0(5)
+	std	14, 8(5)
+	std	15, 16(5)
+	rldicr	15, 15, 32, 63
+	li	8, 7
+	mtctr	8
+L_AES_set_encrypt_key_loop_192:
+	rlwinm	9, 15, 2, 22, 29
+	rlwinm	10, 15, 26, 22, 29
+	rlwinm	11, 15, 18, 22, 29
+	rlwinm	15, 15, 10, 22, 29
+	lbzx	9, 6, 9
+	lbzx	10, 6, 10
+	lbzx	11, 6, 11
+	lbzx	0, 6, 15
+	rlwimi	0, 9, 8, 16, 23
+	rlwimi	0, 10, 16, 8, 15
+	rlwimi	0, 11, 24, 0, 7
+	lwz	10, 0(5)
+	lwz	9, 4(5)
+	lwz	12, 8(5)
+	lwz	11, 12(5)
+	lwz	15, 16(5)
+	lwz	14, 20(5)
+	addi	5, 5, 24
+	xor	9, 9, 0
+	lwz	0, 0(7)
+	addi	7, 7, 4
+	xor	9, 9, 0
+	xor	10, 10, 9
+	xor	11, 11, 10
+	xor	12, 12, 11
+	xor	14, 14, 12
+	xor	15, 15, 14
+	stw	10, 0(5)
+	stw	9, 4(5)
+	stw	12, 8(5)
+	stw	11, 12(5)
+	stw	15, 16(5)
+	stw	14, 20(5)
+	bdnz	L_AES_set_encrypt_key_loop_192
+	rlwinm	9, 15, 2, 22, 29
+	rlwinm	10, 15, 26, 22, 29
+	rlwinm	11, 15, 18, 22, 29
+	rlwinm	15, 15, 10, 22, 29
+	lbzx	9, 6, 9
+	lbzx	10, 6, 10
+	lbzx	11, 6, 11
+	lbzx	0, 6, 15
+	rlwimi	0, 9, 8, 16, 23
+	rlwimi	0, 10, 16, 8, 15
+	rlwimi	0, 11, 24, 0, 7
+	lwz	10, 0(5)
+	lwz	9, 4(5)
+	lwz	12, 8(5)
+	lwz	11, 12(5)
+	lwz	15, 16(5)
+	lwz	14, 20(5)
+	addi	5, 5, 24
+	xor	9, 9, 0
+	lwz	0, 0(7)
+	addi	7, 7, 4
+	xor	9, 9, 0
+	xor	10, 10, 9
+	xor	11, 11, 10
+	xor	12, 12, 11
+	stw	9, 0(5)
+	stw	10, 4(5)
+	stw	11, 8(5)
+	stw	12, 12(5)
+	b	L_AES_set_encrypt_key_end
+L_AES_set_encrypt_key_start_128:
+	li	8, 0
+	ld	11, 0(3)
+	ld	12, 8(3)
+	rldicr	11, 11, 32, 63
+	rldicr	12, 12, 32, 63
+	std	11, 0(5)
+	std	12, 8(5)
+	rldicr	12, 12, 32, 63
+	li	8, 9
+	mtctr	8
+L_AES_set_encrypt_key_loop_128:
+	rlwinm	9, 12, 2, 22, 29
+	rlwinm	10, 12, 26, 22, 29
+	rlwinm	11, 12, 18, 22, 29
+	rlwinm	12, 12, 10, 22, 29
+	lbzx	9, 6, 9
+	lbzx	10, 6, 10
+	lbzx	11, 6, 11
+	lbzx	0, 6, 12
+	rlwimi	0, 9, 8, 16, 23
+	rlwimi	0, 10, 16, 8, 15
+	rlwimi	0, 11, 24, 0, 7
+	lwz	10, 0(5)
+	lwz	9, 4(5)
+	lwz	12, 8(5)
+	lwz	11, 12(5)
+	addi	5, 5, 16
+	xor	9, 9, 0
+	lwz	0, 0(7)
+	addi	7, 7, 4
+	xor	9, 9, 0
+	xor	10, 10, 9
+	xor	11, 11, 10
+	xor	12, 12, 11
+	stw	10, 0(5)
+	stw	9, 4(5)
+	stw	12, 8(5)
+	stw	11, 12(5)
+	bdnz	L_AES_set_encrypt_key_loop_128
+	rlwinm	9, 12, 2, 22, 29
+	rlwinm	10, 12, 26, 22, 29
+	rlwinm	11, 12, 18, 22, 29
+	rlwinm	12, 12, 10, 22, 29
+	lbzx	9, 6, 9
+	lbzx	10, 6, 10
+	lbzx	11, 6, 11
+	lbzx	0, 6, 12
+	rlwimi	0, 9, 8, 16, 23
+	rlwimi	0, 10, 16, 8, 15
+	rlwimi	0, 11, 24, 0, 7
+	lwz	10, 0(5)
+	lwz	9, 4(5)
+	lwz	12, 8(5)
+	lwz	11, 12(5)
+	addi	5, 5, 16
+	xor	9, 9, 0
+	lwz	0, 0(7)
+	addi	7, 7, 4
+	xor	9, 9, 0
+	xor	10, 10, 9
+	xor	11, 11, 10
+	xor	12, 12, 11
+	stw	9, 0(5)
+	stw	10, 4(5)
+	stw	11, 8(5)
+	stw	12, 12(5)
+L_AES_set_encrypt_key_end:
+	ld	0, 16(1)
+	mtlr	0
+	ld	14, 0(1)
+	ld	15, 8(1)
+	addi	1, 1, 24
+	blr
+#ifndef __APPLE__
+	.size	AES_set_encrypt_key,.-AES_set_encrypt_key
+#endif /* __APPLE__ */
+#ifndef __APPLE__
+	.text
+	.section	.rodata
+	.type	L_AES_PPC64_te4_0, @object
+	.size	L_AES_PPC64_te4_0, 4096
+#else
+	.section	__DATA,__data
+#endif /* __APPLE__ */
+#ifndef __APPLE__
+	.align	4
+#else
+	.p2align	4
+#endif /* __APPLE__ */
+L_AES_PPC64_te4_0:
+	.long	0xc66363a5
+	.long	0xf87c7c84
+	.long	0xee777799
+	.long	0xf67b7b8d
+	.long	0xfff2f20d
+	.long	0xd66b6bbd
+	.long	0xde6f6fb1
+	.long	0x91c5c554
+	.long	0x60303050
+	.long	0x2010103
+	.long	0xce6767a9
+	.long	0x562b2b7d
+	.long	0xe7fefe19
+	.long	0xb5d7d762
+	.long	0x4dababe6
+	.long	0xec76769a
+	.long	0x8fcaca45
+	.long	0x1f82829d
+	.long	0x89c9c940
+	.long	0xfa7d7d87
+	.long	0xeffafa15
+	.long	0xb25959eb
+	.long	0x8e4747c9
+	.long	0xfbf0f00b
+	.long	0x41adadec
+	.long	0xb3d4d467
+	.long	0x5fa2a2fd
+	.long	0x45afafea
+	.long	0x239c9cbf
+	.long	0x53a4a4f7
+	.long	0xe4727296
+	.long	0x9bc0c05b
+	.long	0x75b7b7c2
+	.long	0xe1fdfd1c
+	.long	0x3d9393ae
+	.long	0x4c26266a
+	.long	0x6c36365a
+	.long	0x7e3f3f41
+	.long	0xf5f7f702
+	.long	0x83cccc4f
+	.long	0x6834345c
+	.long	0x51a5a5f4
+	.long	0xd1e5e534
+	.long	0xf9f1f108
+	.long	0xe2717193
+	.long	0xabd8d873
+	.long	0x62313153
+	.long	0x2a15153f
+	.long	0x804040c
+	.long	0x95c7c752
+	.long	0x46232365
+	.long	0x9dc3c35e
+	.long	0x30181828
+	.long	0x379696a1
+	.long	0xa05050f
+	.long	0x2f9a9ab5
+	.long	0xe070709
+	.long	0x24121236
+	.long	0x1b80809b
+	.long	0xdfe2e23d
+	.long	0xcdebeb26
+	.long	0x4e272769
+	.long	0x7fb2b2cd
+	.long	0xea75759f
+	.long	0x1209091b
+	.long	0x1d83839e
+	.long	0x582c2c74
+	.long	0x341a1a2e
+	.long	0x361b1b2d
+	.long	0xdc6e6eb2
+	.long	0xb45a5aee
+	.long	0x5ba0a0fb
+	.long	0xa45252f6
+	.long	0x763b3b4d
+	.long	0xb7d6d661
+	.long	0x7db3b3ce
+	.long	0x5229297b
+	.long	0xdde3e33e
+	.long	0x5e2f2f71
+	.long	0x13848497
+	.long	0xa65353f5
+	.long	0xb9d1d168
+	.long	0x0
+	.long	0xc1eded2c
+	.long	0x40202060
+	.long	0xe3fcfc1f
+	.long	0x79b1b1c8
+	.long	0xb65b5bed
+	.long	0xd46a6abe
+	.long	0x8dcbcb46
+	.long	0x67bebed9
+	.long	0x7239394b
+	.long	0x944a4ade
+	.long	0x984c4cd4
+	.long	0xb05858e8
+	.long	0x85cfcf4a
+	.long	0xbbd0d06b
+	.long	0xc5efef2a
+	.long	0x4faaaae5
+	.long	0xedfbfb16
+	.long	0x864343c5
+	.long	0x9a4d4dd7
+	.long	0x66333355
+	.long	0x11858594
+	.long	0x8a4545cf
+	.long	0xe9f9f910
+	.long	0x4020206
+	.long	0xfe7f7f81
+	.long	0xa05050f0
+	.long	0x783c3c44
+	.long	0x259f9fba
+	.long	0x4ba8a8e3
+	.long	0xa25151f3
+	.long	0x5da3a3fe
+	.long	0x804040c0
+	.long	0x58f8f8a
+	.long	0x3f9292ad
+	.long	0x219d9dbc
+	.long	0x70383848
+	.long	0xf1f5f504
+	.long	0x63bcbcdf
+	.long	0x77b6b6c1
+	.long	0xafdada75
+	.long	0x42212163
+	.long	0x20101030
+	.long	0xe5ffff1a
+	.long	0xfdf3f30e
+	.long	0xbfd2d26d
+	.long	0x81cdcd4c
+	.long	0x180c0c14
+	.long	0x26131335
+	.long	0xc3ecec2f
+	.long	0xbe5f5fe1
+	.long	0x359797a2
+	.long	0x884444cc
+	.long	0x2e171739
+	.long	0x93c4c457
+	.long	0x55a7a7f2
+	.long	0xfc7e7e82
+	.long	0x7a3d3d47
+	.long	0xc86464ac
+	.long	0xba5d5de7
+	.long	0x3219192b
+	.long	0xe6737395
+	.long	0xc06060a0
+	.long	0x19818198
+	.long	0x9e4f4fd1
+	.long	0xa3dcdc7f
+	.long	0x44222266
+	.long	0x542a2a7e
+	.long	0x3b9090ab
+	.long	0xb888883
+	.long	0x8c4646ca
+	.long	0xc7eeee29
+	.long	0x6bb8b8d3
+	.long	0x2814143c
+	.long	0xa7dede79
+	.long	0xbc5e5ee2
+	.long	0x160b0b1d
+	.long	0xaddbdb76
+	.long	0xdbe0e03b
+	.long	0x64323256
+	.long	0x743a3a4e
+	.long	0x140a0a1e
+	.long	0x924949db
+	.long	0xc06060a
+	.long	0x4824246c
+	.long	0xb85c5ce4
+	.long	0x9fc2c25d
+	.long	0xbdd3d36e
+	.long	0x43acacef
+	.long	0xc46262a6
+	.long	0x399191a8
+	.long	0x319595a4
+	.long	0xd3e4e437
+	.long	0xf279798b
+	.long	0xd5e7e732
+	.long	0x8bc8c843
+	.long	0x6e373759
+	.long	0xda6d6db7
+	.long	0x18d8d8c
+	.long	0xb1d5d564
+	.long	0x9c4e4ed2
+	.long	0x49a9a9e0
+	.long	0xd86c6cb4
+	.long	0xac5656fa
+	.long	0xf3f4f407
+	.long	0xcfeaea25
+	.long	0xca6565af
+	.long	0xf47a7a8e
+	.long	0x47aeaee9
+	.long	0x10080818
+	.long	0x6fbabad5
+	.long	0xf0787888
+	.long	0x4a25256f
+	.long	0x5c2e2e72
+	.long	0x381c1c24
+	.long	0x57a6a6f1
+	.long	0x73b4b4c7
+	.long	0x97c6c651
+	.long	0xcbe8e823
+	.long	0xa1dddd7c
+	.long	0xe874749c
+	.long	0x3e1f1f21
+	.long	0x964b4bdd
+	.long	0x61bdbddc
+	.long	0xd8b8b86
+	.long	0xf8a8a85
+	.long	0xe0707090
+	.long	0x7c3e3e42
+	.long	0x71b5b5c4
+	.long	0xcc6666aa
+	.long	0x904848d8
+	.long	0x6030305
+	.long	0xf7f6f601
+	.long	0x1c0e0e12
+	.long	0xc26161a3
+	.long	0x6a35355f
+	.long	0xae5757f9
+	.long	0x69b9b9d0
+	.long	0x17868691
+	.long	0x99c1c158
+	.long	0x3a1d1d27
+	.long	0x279e9eb9
+	.long	0xd9e1e138
+	.long	0xebf8f813
+	.long	0x2b9898b3
+	.long	0x22111133
+	.long	0xd26969bb
+	.long	0xa9d9d970
+	.long	0x78e8e89
+	.long	0x339494a7
+	.long	0x2d9b9bb6
+	.long	0x3c1e1e22
+	.long	0x15878792
+	.long	0xc9e9e920
+	.long	0x87cece49
+	.long	0xaa5555ff
+	.long	0x50282878
+	.long	0xa5dfdf7a
+	.long	0x38c8c8f
+	.long	0x59a1a1f8
+	.long	0x9898980
+	.long	0x1a0d0d17
+	.long	0x65bfbfda
+	.long	0xd7e6e631
+	.long	0x844242c6
+	.long	0xd06868b8
+	.long	0x824141c3
+	.long	0x299999b0
+	.long	0x5a2d2d77
+	.long	0x1e0f0f11
+	.long	0x7bb0b0cb
+	.long	0xa85454fc
+	.long	0x6dbbbbd6
+	.long	0x2c16163a
+	.long	0xa5c66363
+	.long	0x84f87c7c
+	.long	0x99ee7777
+	.long	0x8df67b7b
+	.long	0xdfff2f2
+	.long	0xbdd66b6b
+	.long	0xb1de6f6f
+	.long	0x5491c5c5
+	.long	0x50603030
+	.long	0x3020101
+	.long	0xa9ce6767
+	.long	0x7d562b2b
+	.long	0x19e7fefe
+	.long	0x62b5d7d7
+	.long	0xe64dabab
+	.long	0x9aec7676
+	.long	0x458fcaca
+	.long	0x9d1f8282
+	.long	0x4089c9c9
+	.long	0x87fa7d7d
+	.long	0x15effafa
+	.long	0xebb25959
+	.long	0xc98e4747
+	.long	0xbfbf0f0
+	.long	0xec41adad
+	.long	0x67b3d4d4
+	.long	0xfd5fa2a2
+	.long	0xea45afaf
+	.long	0xbf239c9c
+	.long	0xf753a4a4
+	.long	0x96e47272
+	.long	0x5b9bc0c0
+	.long	0xc275b7b7
+	.long	0x1ce1fdfd
+	.long	0xae3d9393
+	.long	0x6a4c2626
+	.long	0x5a6c3636
+	.long	0x417e3f3f
+	.long	0x2f5f7f7
+	.long	0x4f83cccc
+	.long	0x5c683434
+	.long	0xf451a5a5
+	.long	0x34d1e5e5
+	.long	0x8f9f1f1
+	.long	0x93e27171
+	.long	0x73abd8d8
+	.long	0x53623131
+	.long	0x3f2a1515
+	.long	0xc080404
+	.long	0x5295c7c7
+	.long	0x65462323
+	.long	0x5e9dc3c3
+	.long	0x28301818
+	.long	0xa1379696
+	.long	0xf0a0505
+	.long	0xb52f9a9a
+	.long	0x90e0707
+	.long	0x36241212
+	.long	0x9b1b8080
+	.long	0x3ddfe2e2
+	.long	0x26cdebeb
+	.long	0x694e2727
+	.long	0xcd7fb2b2
+	.long	0x9fea7575
+	.long	0x1b120909
+	.long	0x9e1d8383
+	.long	0x74582c2c
+	.long	0x2e341a1a
+	.long	0x2d361b1b
+	.long	0xb2dc6e6e
+	.long	0xeeb45a5a
+	.long	0xfb5ba0a0
+	.long	0xf6a45252
+	.long	0x4d763b3b
+	.long	0x61b7d6d6
+	.long	0xce7db3b3
+	.long	0x7b522929
+	.long	0x3edde3e3
+	.long	0x715e2f2f
+	.long	0x97138484
+	.long	0xf5a65353
+	.long	0x68b9d1d1
+	.long	0x0
+	.long	0x2cc1eded
+	.long	0x60402020
+	.long	0x1fe3fcfc
+	.long	0xc879b1b1
+	.long	0xedb65b5b
+	.long	0xbed46a6a
+	.long	0x468dcbcb
+	.long	0xd967bebe
+	.long	0x4b723939
+	.long	0xde944a4a
+	.long	0xd4984c4c
+	.long	0xe8b05858
+	.long	0x4a85cfcf
+	.long	0x6bbbd0d0
+	.long	0x2ac5efef
+	.long	0xe54faaaa
+	.long	0x16edfbfb
+	.long	0xc5864343
+	.long	0xd79a4d4d
+	.long	0x55663333
+	.long	0x94118585
+	.long	0xcf8a4545
+	.long	0x10e9f9f9
+	.long	0x6040202
+	.long	0x81fe7f7f
+	.long	0xf0a05050
+	.long	0x44783c3c
+	.long	0xba259f9f
+	.long	0xe34ba8a8
+	.long	0xf3a25151
+	.long	0xfe5da3a3
+	.long	0xc0804040
+	.long	0x8a058f8f
+	.long	0xad3f9292
+	.long	0xbc219d9d
+	.long	0x48703838
+	.long	0x4f1f5f5
+	.long	0xdf63bcbc
+	.long	0xc177b6b6
+	.long	0x75afdada
+	.long	0x63422121
+	.long	0x30201010
+	.long	0x1ae5ffff
+	.long	0xefdf3f3
+	.long	0x6dbfd2d2
+	.long	0x4c81cdcd
+	.long	0x14180c0c
+	.long	0x35261313
+	.long	0x2fc3ecec
+	.long	0xe1be5f5f
+	.long	0xa2359797
+	.long	0xcc884444
+	.long	0x392e1717
+	.long	0x5793c4c4
+	.long	0xf255a7a7
+	.long	0x82fc7e7e
+	.long	0x477a3d3d
+	.long	0xacc86464
+	.long	0xe7ba5d5d
+	.long	0x2b321919
+	.long	0x95e67373
+	.long	0xa0c06060
+	.long	0x98198181
+	.long	0xd19e4f4f
+	.long	0x7fa3dcdc
+	.long	0x66442222
+	.long	0x7e542a2a
+	.long	0xab3b9090
+	.long	0x830b8888
+	.long	0xca8c4646
+	.long	0x29c7eeee
+	.long	0xd36bb8b8
+	.long	0x3c281414
+	.long	0x79a7dede
+	.long	0xe2bc5e5e
+	.long	0x1d160b0b
+	.long	0x76addbdb
+	.long	0x3bdbe0e0
+	.long	0x56643232
+	.long	0x4e743a3a
+	.long	0x1e140a0a
+	.long	0xdb924949
+	.long	0xa0c0606
+	.long	0x6c482424
+	.long	0xe4b85c5c
+	.long	0x5d9fc2c2
+	.long	0x6ebdd3d3
+	.long	0xef43acac
+	.long	0xa6c46262
+	.long	0xa8399191
+	.long	0xa4319595
+	.long	0x37d3e4e4
+	.long	0x8bf27979
+	.long	0x32d5e7e7
+	.long	0x438bc8c8
+	.long	0x596e3737
+	.long	0xb7da6d6d
+	.long	0x8c018d8d
+	.long	0x64b1d5d5
+	.long	0xd29c4e4e
+	.long	0xe049a9a9
+	.long	0xb4d86c6c
+	.long	0xfaac5656
+	.long	0x7f3f4f4
+	.long	0x25cfeaea
+	.long	0xafca6565
+	.long	0x8ef47a7a
+	.long	0xe947aeae
+	.long	0x18100808
+	.long	0xd56fbaba
+	.long	0x88f07878
+	.long	0x6f4a2525
+	.long	0x725c2e2e
+	.long	0x24381c1c
+	.long	0xf157a6a6
+	.long	0xc773b4b4
+	.long	0x5197c6c6
+	.long	0x23cbe8e8
+	.long	0x7ca1dddd
+	.long	0x9ce87474
+	.long	0x213e1f1f
+	.long	0xdd964b4b
+	.long	0xdc61bdbd
+	.long	0x860d8b8b
+	.long	0x850f8a8a
+	.long	0x90e07070
+	.long	0x427c3e3e
+	.long	0xc471b5b5
+	.long	0xaacc6666
+	.long	0xd8904848
+	.long	0x5060303
+	.long	0x1f7f6f6
+	.long	0x121c0e0e
+	.long	0xa3c26161
+	.long	0x5f6a3535
+	.long	0xf9ae5757
+	.long	0xd069b9b9
+	.long	0x91178686
+	.long	0x5899c1c1
+	.long	0x273a1d1d
+	.long	0xb9279e9e
+	.long	0x38d9e1e1
+	.long	0x13ebf8f8
+	.long	0xb32b9898
+	.long	0x33221111
+	.long	0xbbd26969
+	.long	0x70a9d9d9
+	.long	0x89078e8e
+	.long	0xa7339494
+	.long	0xb62d9b9b
+	.long	0x223c1e1e
+	.long	0x92158787
+	.long	0x20c9e9e9
+	.long	0x4987cece
+	.long	0xffaa5555
+	.long	0x78502828
+	.long	0x7aa5dfdf
+	.long	0x8f038c8c
+	.long	0xf859a1a1
+	.long	0x80098989
+	.long	0x171a0d0d
+	.long	0xda65bfbf
+	.long	0x31d7e6e6
+	.long	0xc6844242
+	.long	0xb8d06868
+	.long	0xc3824141
+	.long	0xb0299999
+	.long	0x775a2d2d
+	.long	0x111e0f0f
+	.long	0xcb7bb0b0
+	.long	0xfca85454
+	.long	0xd66dbbbb
+	.long	0x3a2c1616
+	.long	0x63a5c663
+	.long	0x7c84f87c
+	.long	0x7799ee77
+	.long	0x7b8df67b
+	.long	0xf20dfff2
+	.long	0x6bbdd66b
+	.long	0x6fb1de6f
+	.long	0xc55491c5
+	.long	0x30506030
+	.long	0x1030201
+	.long	0x67a9ce67
+	.long	0x2b7d562b
+	.long	0xfe19e7fe
+	.long	0xd762b5d7
+	.long	0xabe64dab
+	.long	0x769aec76
+	.long	0xca458fca
+	.long	0x829d1f82
+	.long	0xc94089c9
+	.long	0x7d87fa7d
+	.long	0xfa15effa
+	.long	0x59ebb259
+	.long	0x47c98e47
+	.long	0xf00bfbf0
+	.long	0xadec41ad
+	.long	0xd467b3d4
+	.long	0xa2fd5fa2
+	.long	0xafea45af
+	.long	0x9cbf239c
+	.long	0xa4f753a4
+	.long	0x7296e472
+	.long	0xc05b9bc0
+	.long	0xb7c275b7
+	.long	0xfd1ce1fd
+	.long	0x93ae3d93
+	.long	0x266a4c26
+	.long	0x365a6c36
+	.long	0x3f417e3f
+	.long	0xf702f5f7
+	.long	0xcc4f83cc
+	.long	0x345c6834
+	.long	0xa5f451a5
+	.long	0xe534d1e5
+	.long	0xf108f9f1
+	.long	0x7193e271
+	.long	0xd873abd8
+	.long	0x31536231
+	.long	0x153f2a15
+	.long	0x40c0804
+	.long	0xc75295c7
+	.long	0x23654623
+	.long	0xc35e9dc3
+	.long	0x18283018
+	.long	0x96a13796
+	.long	0x50f0a05
+	.long	0x9ab52f9a
+	.long	0x7090e07
+	.long	0x12362412
+	.long	0x809b1b80
+	.long	0xe23ddfe2
+	.long	0xeb26cdeb
+	.long	0x27694e27
+	.long	0xb2cd7fb2
+	.long	0x759fea75
+	.long	0x91b1209
+	.long	0x839e1d83
+	.long	0x2c74582c
+	.long	0x1a2e341a
+	.long	0x1b2d361b
+	.long	0x6eb2dc6e
+	.long	0x5aeeb45a
+	.long	0xa0fb5ba0
+	.long	0x52f6a452
+	.long	0x3b4d763b
+	.long	0xd661b7d6
+	.long	0xb3ce7db3
+	.long	0x297b5229
+	.long	0xe33edde3
+	.long	0x2f715e2f
+	.long	0x84971384
+	.long	0x53f5a653
+	.long	0xd168b9d1
+	.long	0x0
+	.long	0xed2cc1ed
+	.long	0x20604020
+	.long	0xfc1fe3fc
+	.long	0xb1c879b1
+	.long	0x5bedb65b
+	.long	0x6abed46a
+	.long	0xcb468dcb
+	.long	0xbed967be
+	.long	0x394b7239
+	.long	0x4ade944a
+	.long	0x4cd4984c
+	.long	0x58e8b058
+	.long	0xcf4a85cf
+	.long	0xd06bbbd0
+	.long	0xef2ac5ef
+	.long	0xaae54faa
+	.long	0xfb16edfb
+	.long	0x43c58643
+	.long	0x4dd79a4d
+	.long	0x33556633
+	.long	0x85941185
+	.long	0x45cf8a45
+	.long	0xf910e9f9
+	.long	0x2060402
+	.long	0x7f81fe7f
+	.long	0x50f0a050
+	.long	0x3c44783c
+	.long	0x9fba259f
+	.long	0xa8e34ba8
+	.long	0x51f3a251
+	.long	0xa3fe5da3
+	.long	0x40c08040
+	.long	0x8f8a058f
+	.long	0x92ad3f92
+	.long	0x9dbc219d
+	.long	0x38487038
+	.long	0xf504f1f5
+	.long	0xbcdf63bc
+	.long	0xb6c177b6
+	.long	0xda75afda
+	.long	0x21634221
+	.long	0x10302010
+	.long	0xff1ae5ff
+	.long	0xf30efdf3
+	.long	0xd26dbfd2
+	.long	0xcd4c81cd
+	.long	0xc14180c
+	.long	0x13352613
+	.long	0xec2fc3ec
+	.long	0x5fe1be5f
+	.long	0x97a23597
+	.long	0x44cc8844
+	.long	0x17392e17
+	.long	0xc45793c4
+	.long	0xa7f255a7
+	.long	0x7e82fc7e
+	.long	0x3d477a3d
+	.long	0x64acc864
+	.long	0x5de7ba5d
+	.long	0x192b3219
+	.long	0x7395e673
+	.long	0x60a0c060
+	.long	0x81981981
+	.long	0x4fd19e4f
+	.long	0xdc7fa3dc
+	.long	0x22664422
+	.long	0x2a7e542a
+	.long	0x90ab3b90
+	.long	0x88830b88
+	.long	0x46ca8c46
+	.long	0xee29c7ee
+	.long	0xb8d36bb8
+	.long	0x143c2814
+	.long	0xde79a7de
+	.long	0x5ee2bc5e
+	.long	0xb1d160b
+	.long	0xdb76addb
+	.long	0xe03bdbe0
+	.long	0x32566432
+	.long	0x3a4e743a
+	.long	0xa1e140a
+	.long	0x49db9249
+	.long	0x60a0c06
+	.long	0x246c4824
+	.long	0x5ce4b85c
+	.long	0xc25d9fc2
+	.long	0xd36ebdd3
+	.long	0xacef43ac
+	.long	0x62a6c462
+	.long	0x91a83991
+	.long	0x95a43195
+	.long	0xe437d3e4
+	.long	0x798bf279
+	.long	0xe732d5e7
+	.long	0xc8438bc8
+	.long	0x37596e37
+	.long	0x6db7da6d
+	.long	0x8d8c018d
+	.long	0xd564b1d5
+	.long	0x4ed29c4e
+	.long	0xa9e049a9
+	.long	0x6cb4d86c
+	.long	0x56faac56
+	.long	0xf407f3f4
+	.long	0xea25cfea
+	.long	0x65afca65
+	.long	0x7a8ef47a
+	.long	0xaee947ae
+	.long	0x8181008
+	.long	0xbad56fba
+	.long	0x7888f078
+	.long	0x256f4a25
+	.long	0x2e725c2e
+	.long	0x1c24381c
+	.long	0xa6f157a6
+	.long	0xb4c773b4
+	.long	0xc65197c6
+	.long	0xe823cbe8
+	.long	0xdd7ca1dd
+	.long	0x749ce874
+	.long	0x1f213e1f
+	.long	0x4bdd964b
+	.long	0xbddc61bd
+	.long	0x8b860d8b
+	.long	0x8a850f8a
+	.long	0x7090e070
+	.long	0x3e427c3e
+	.long	0xb5c471b5
+	.long	0x66aacc66
+	.long	0x48d89048
+	.long	0x3050603
+	.long	0xf601f7f6
+	.long	0xe121c0e
+	.long	0x61a3c261
+	.long	0x355f6a35
+	.long	0x57f9ae57
+	.long	0xb9d069b9
+	.long	0x86911786
+	.long	0xc15899c1
+	.long	0x1d273a1d
+	.long	0x9eb9279e
+	.long	0xe138d9e1
+	.long	0xf813ebf8
+	.long	0x98b32b98
+	.long	0x11332211
+	.long	0x69bbd269
+	.long	0xd970a9d9
+	.long	0x8e89078e
+	.long	0x94a73394
+	.long	0x9bb62d9b
+	.long	0x1e223c1e
+	.long	0x87921587
+	.long	0xe920c9e9
+	.long	0xce4987ce
+	.long	0x55ffaa55
+	.long	0x28785028
+	.long	0xdf7aa5df
+	.long	0x8c8f038c
+	.long	0xa1f859a1
+	.long	0x89800989
+	.long	0xd171a0d
+	.long	0xbfda65bf
+	.long	0xe631d7e6
+	.long	0x42c68442
+	.long	0x68b8d068
+	.long	0x41c38241
+	.long	0x99b02999
+	.long	0x2d775a2d
+	.long	0xf111e0f
+	.long	0xb0cb7bb0
+	.long	0x54fca854
+	.long	0xbbd66dbb
+	.long	0x163a2c16
+	.long	0x6363a5c6
+	.long	0x7c7c84f8
+	.long	0x777799ee
+	.long	0x7b7b8df6
+	.long	0xf2f20dff
+	.long	0x6b6bbdd6
+	.long	0x6f6fb1de
+	.long	0xc5c55491
+	.long	0x30305060
+	.long	0x1010302
+	.long	0x6767a9ce
+	.long	0x2b2b7d56
+	.long	0xfefe19e7
+	.long	0xd7d762b5
+	.long	0xababe64d
+	.long	0x76769aec
+	.long	0xcaca458f
+	.long	0x82829d1f
+	.long	0xc9c94089
+	.long	0x7d7d87fa
+	.long	0xfafa15ef
+	.long	0x5959ebb2
+	.long	0x4747c98e
+	.long	0xf0f00bfb
+	.long	0xadadec41
+	.long	0xd4d467b3
+	.long	0xa2a2fd5f
+	.long	0xafafea45
+	.long	0x9c9cbf23
+	.long	0xa4a4f753
+	.long	0x727296e4
+	.long	0xc0c05b9b
+	.long	0xb7b7c275
+	.long	0xfdfd1ce1
+	.long	0x9393ae3d
+	.long	0x26266a4c
+	.long	0x36365a6c
+	.long	0x3f3f417e
+	.long	0xf7f702f5
+	.long	0xcccc4f83
+	.long	0x34345c68
+	.long	0xa5a5f451
+	.long	0xe5e534d1
+	.long	0xf1f108f9
+	.long	0x717193e2
+	.long	0xd8d873ab
+	.long	0x31315362
+	.long	0x15153f2a
+	.long	0x4040c08
+	.long	0xc7c75295
+	.long	0x23236546
+	.long	0xc3c35e9d
+	.long	0x18182830
+	.long	0x9696a137
+	.long	0x5050f0a
+	.long	0x9a9ab52f
+	.long	0x707090e
+	.long	0x12123624
+	.long	0x80809b1b
+	.long	0xe2e23ddf
+	.long	0xebeb26cd
+	.long	0x2727694e
+	.long	0xb2b2cd7f
+	.long	0x75759fea
+	.long	0x9091b12
+	.long	0x83839e1d
+	.long	0x2c2c7458
+	.long	0x1a1a2e34
+	.long	0x1b1b2d36
+	.long	0x6e6eb2dc
+	.long	0x5a5aeeb4
+	.long	0xa0a0fb5b
+	.long	0x5252f6a4
+	.long	0x3b3b4d76
+	.long	0xd6d661b7
+	.long	0xb3b3ce7d
+	.long	0x29297b52
+	.long	0xe3e33edd
+	.long	0x2f2f715e
+	.long	0x84849713
+	.long	0x5353f5a6
+	.long	0xd1d168b9
+	.long	0x0
+	.long	0xeded2cc1
+	.long	0x20206040
+	.long	0xfcfc1fe3
+	.long	0xb1b1c879
+	.long	0x5b5bedb6
+	.long	0x6a6abed4
+	.long	0xcbcb468d
+	.long	0xbebed967
+	.long	0x39394b72
+	.long	0x4a4ade94
+	.long	0x4c4cd498
+	.long	0x5858e8b0
+	.long	0xcfcf4a85
+	.long	0xd0d06bbb
+	.long	0xefef2ac5
+	.long	0xaaaae54f
+	.long	0xfbfb16ed
+	.long	0x4343c586
+	.long	0x4d4dd79a
+	.long	0x33335566
+	.long	0x85859411
+	.long	0x4545cf8a
+	.long	0xf9f910e9
+	.long	0x2020604
+	.long	0x7f7f81fe
+	.long	0x5050f0a0
+	.long	0x3c3c4478
+	.long	0x9f9fba25
+	.long	0xa8a8e34b
+	.long	0x5151f3a2
+	.long	0xa3a3fe5d
+	.long	0x4040c080
+	.long	0x8f8f8a05
+	.long	0x9292ad3f
+	.long	0x9d9dbc21
+	.long	0x38384870
+	.long	0xf5f504f1
+	.long	0xbcbcdf63
+	.long	0xb6b6c177
+	.long	0xdada75af
+	.long	0x21216342
+	.long	0x10103020
+	.long	0xffff1ae5
+	.long	0xf3f30efd
+	.long	0xd2d26dbf
+	.long	0xcdcd4c81
+	.long	0xc0c1418
+	.long	0x13133526
+	.long	0xecec2fc3
+	.long	0x5f5fe1be
+	.long	0x9797a235
+	.long	0x4444cc88
+	.long	0x1717392e
+	.long	0xc4c45793
+	.long	0xa7a7f255
+	.long	0x7e7e82fc
+	.long	0x3d3d477a
+	.long	0x6464acc8
+	.long	0x5d5de7ba
+	.long	0x19192b32
+	.long	0x737395e6
+	.long	0x6060a0c0
+	.long	0x81819819
+	.long	0x4f4fd19e
+	.long	0xdcdc7fa3
+	.long	0x22226644
+	.long	0x2a2a7e54
+	.long	0x9090ab3b
+	.long	0x8888830b
+	.long	0x4646ca8c
+	.long	0xeeee29c7
+	.long	0xb8b8d36b
+	.long	0x14143c28
+	.long	0xdede79a7
+	.long	0x5e5ee2bc
+	.long	0xb0b1d16
+	.long	0xdbdb76ad
+	.long	0xe0e03bdb
+	.long	0x32325664
+	.long	0x3a3a4e74
+	.long	0xa0a1e14
+	.long	0x4949db92
+	.long	0x6060a0c
+	.long	0x24246c48
+	.long	0x5c5ce4b8
+	.long	0xc2c25d9f
+	.long	0xd3d36ebd
+	.long	0xacacef43
+	.long	0x6262a6c4
+	.long	0x9191a839
+	.long	0x9595a431
+	.long	0xe4e437d3
+	.long	0x79798bf2
+	.long	0xe7e732d5
+	.long	0xc8c8438b
+	.long	0x3737596e
+	.long	0x6d6db7da
+	.long	0x8d8d8c01
+	.long	0xd5d564b1
+	.long	0x4e4ed29c
+	.long	0xa9a9e049
+	.long	0x6c6cb4d8
+	.long	0x5656faac
+	.long	0xf4f407f3
+	.long	0xeaea25cf
+	.long	0x6565afca
+	.long	0x7a7a8ef4
+	.long	0xaeaee947
+	.long	0x8081810
+	.long	0xbabad56f
+	.long	0x787888f0
+	.long	0x25256f4a
+	.long	0x2e2e725c
+	.long	0x1c1c2438
+	.long	0xa6a6f157
+	.long	0xb4b4c773
+	.long	0xc6c65197
+	.long	0xe8e823cb
+	.long	0xdddd7ca1
+	.long	0x74749ce8
+	.long	0x1f1f213e
+	.long	0x4b4bdd96
+	.long	0xbdbddc61
+	.long	0x8b8b860d
+	.long	0x8a8a850f
+	.long	0x707090e0
+	.long	0x3e3e427c
+	.long	0xb5b5c471
+	.long	0x6666aacc
+	.long	0x4848d890
+	.long	0x3030506
+	.long	0xf6f601f7
+	.long	0xe0e121c
+	.long	0x6161a3c2
+	.long	0x35355f6a
+	.long	0x5757f9ae
+	.long	0xb9b9d069
+	.long	0x86869117
+	.long	0xc1c15899
+	.long	0x1d1d273a
+	.long	0x9e9eb927
+	.long	0xe1e138d9
+	.long	0xf8f813eb
+	.long	0x9898b32b
+	.long	0x11113322
+	.long	0x6969bbd2
+	.long	0xd9d970a9
+	.long	0x8e8e8907
+	.long	0x9494a733
+	.long	0x9b9bb62d
+	.long	0x1e1e223c
+	.long	0x87879215
+	.long	0xe9e920c9
+	.long	0xcece4987
+	.long	0x5555ffaa
+	.long	0x28287850
+	.long	0xdfdf7aa5
+	.long	0x8c8c8f03
+	.long	0xa1a1f859
+	.long	0x89898009
+	.long	0xd0d171a
+	.long	0xbfbfda65
+	.long	0xe6e631d7
+	.long	0x4242c684
+	.long	0x6868b8d0
+	.long	0x4141c382
+	.long	0x9999b029
+	.long	0x2d2d775a
+	.long	0xf0f111e
+	.long	0xb0b0cb7b
+	.long	0x5454fca8
+	.long	0xbbbbd66d
+	.long	0x16163a2c
+#if defined(HAVE_AESCCM) || defined(HAVE_AESGCM) || \
+    defined(WOLFSSL_AES_DIRECT) || defined(WOLFSSL_AES_COUNTER) || \
+    defined(HAVE_AES_ECB)
+#ifndef __APPLE__
+.text
+.globl	AES_ECB_encrypt
+.type	AES_ECB_encrypt,@function
+.align	16
+AES_ECB_encrypt:
+#else
+.section	__TEXT,__text
+.globl	_AES_ECB_encrypt
+.p2align	4
+_AES_ECB_encrypt:
+#endif /* __APPLE__ */
+	stdu	1, -104(1)
+	mflr	0
+	std	0, 96(1)
+	std	14, 0(1)
+	std	15, 8(1)
+	std	16, 16(1)
+	std	17, 24(1)
+	std	18, 32(1)
+	std	19, 40(1)
+	std	20, 48(1)
+	std	21, 56(1)
+	std	22, 64(1)
+	std	23, 72(1)
+	std	24, 80(1)
+	std	25, 88(1)
+#if defined(__ARCH_PWR8) || (defined(_ARCH_PWR8))
+	addis	8, 2, L_AES_PPC64_te4_0@toc@ha
+	addi	8, 8, L_AES_PPC64_te4_0@toc@l
+#else
+	lis	8, L_AES_PPC64_te4_0@ha
+	la	8, L_AES_PPC64_te4_0@l(8)
+#endif /* defined(__ARCH_PWR8) || (defined(_ARCH_PWR8)) */
+	addi	9, 8, 0x400
+	addi	10, 8, 0x800
+	addi	11, 8, 0xc00
+L_AES_ECB_encrypt_loop_block_128:
+	addi	25, 6, 0
+	ld	12, 0(3)
+	ld	14, 8(3)
+	ld	17, 0(25)
+	ld	18, 8(25)
+	addi	25, 25, 16
+	rldicr	12, 12, 32, 63
+	rldicr	14, 14, 32, 63
+	# Round: 0 - XOR in key schedule
+	xor	12, 12, 17
+	xor	14, 14, 18
+	addi	0, 7, -2
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+	ld	20, 0(8)
+	ld	20, 64(8)
+	ld	20, 128(8)
+	ld	20, 192(8)
+	ld	20, 256(8)
+	ld	20, 320(8)
+	ld	20, 384(8)
+	ld	20, 448(8)
+	ld	20, 512(8)
+	ld	20, 576(8)
+	ld	20, 640(8)
+	ld	20, 704(8)
+	ld	20, 768(8)
+	ld	20, 832(8)
+	ld	20, 896(8)
+	ld	20, 960(8)
+	ld	20, 1024(8)
+	ld	20, 1088(8)
+	ld	20, 1152(8)
+	ld	20, 1216(8)
+	ld	20, 1280(8)
+	ld	20, 1344(8)
+	ld	20, 1408(8)
+	ld	20, 1472(8)
+	ld	20, 1536(8)
+	ld	20, 1600(8)
+	ld	20, 1664(8)
+	ld	20, 1728(8)
+	ld	20, 1792(8)
+	ld	20, 1856(8)
+	ld	20, 1920(8)
+	ld	20, 1984(8)
+	ld	20, 2048(8)
+	ld	20, 2112(8)
+	ld	20, 2176(8)
+	ld	20, 2240(8)
+	ld	20, 2304(8)
+	ld	20, 2368(8)
+	ld	20, 2432(8)
+	ld	20, 2496(8)
+	ld	20, 2560(8)
+	ld	20, 2624(8)
+	ld	20, 2688(8)
+	ld	20, 2752(8)
+	ld	20, 2816(8)
+	ld	20, 2880(8)
+	ld	20, 2944(8)
+	ld	20, 3008(8)
+	ld	20, 3072(8)
+	ld	20, 3136(8)
+	ld	20, 3200(8)
+	ld	20, 3264(8)
+	ld	20, 3328(8)
+	ld	20, 3392(8)
+	ld	20, 3456(8)
+	ld	20, 3520(8)
+	ld	20, 3584(8)
+	ld	20, 3648(8)
+	ld	20, 3712(8)
+	ld	20, 3776(8)
+	ld	20, 3840(8)
+	ld	20, 3904(8)
+	ld	20, 3968(8)
+	ld	20, 4032(8)
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+	srwi	0, 0, 1
+	mtctr	0
+L_AES_ECB_encrypt_loop_nr:
+	srdi	15, 12, 32
+	srdi	16, 14, 32
+	rlwinm	17, 12, 10, 22, 29
+	rlwinm	18, 14, 10, 22, 29
+	rlwinm	19, 15, 10, 22, 29
+	rlwinm	20, 16, 10, 22, 29
+	lwzx	17, 8, 17
+	lwzx	18, 8, 18
+	lwzx	19, 8, 19
+	lwzx	20, 8, 20
+	rlwinm	21, 15, 18, 22, 29
+	rlwinm	22, 14, 18, 22, 29
+	rlwinm	23, 16, 18, 22, 29
+	rlwinm	24, 12, 18, 22, 29
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	lwzx	24, 9, 24
+	xor	17, 17, 21
+	xor	19, 19, 22
+	xor	18, 18, 23
+	xor	20, 20, 24
+	rlwinm	21, 14, 26, 22, 29
+	rlwinm	22, 16, 26, 22, 29
+	rlwinm	23, 12, 26, 22, 29
+	rlwinm	24, 15, 26, 22, 29
+	lwzx	21, 10, 21
+	lwzx	22, 10, 22
+	lwzx	23, 10, 23
+	lwzx	24, 10, 24
+	xor	17, 17, 21
+	xor	19, 19, 22
+	xor	18, 18, 23
+	xor	20, 20, 24
+	rlwinm	21, 16, 2, 22, 29
+	rlwinm	22, 12, 2, 22, 29
+	rlwinm	23, 15, 2, 22, 29
+	rlwinm	24, 14, 2, 22, 29
+	lwzx	21, 11, 21
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	xor	17, 17, 21
+	xor	19, 19, 22
+	xor	18, 18, 23
+	xor	20, 20, 24
+	ld	12, 0(25)
+	ld	14, 8(25)
+	rldimi	17, 19, 32, 0
+	rldimi	18, 20, 32, 0
+	addi	25, 25, 16
+	#   XOR in Key Schedule
+	xor	17, 17, 12
+	xor	18, 18, 14
+	srdi	19, 17, 32
+	srdi	20, 18, 32
+	rlwinm	12, 17, 10, 22, 29
+	rlwinm	14, 18, 10, 22, 29
+	rlwinm	15, 19, 10, 22, 29
+	rlwinm	16, 20, 10, 22, 29
+	lwzx	12, 8, 12
+	lwzx	14, 8, 14
+	lwzx	15, 8, 15
+	lwzx	16, 8, 16
+	rlwinm	21, 19, 18, 22, 29
+	rlwinm	22, 18, 18, 22, 29
+	rlwinm	23, 20, 18, 22, 29
+	rlwinm	24, 17, 18, 22, 29
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	lwzx	24, 9, 24
+	xor	12, 12, 21
+	xor	15, 15, 22
+	xor	14, 14, 23
+	xor	16, 16, 24
+	rlwinm	21, 18, 26, 22, 29
+	rlwinm	22, 20, 26, 22, 29
+	rlwinm	23, 17, 26, 22, 29
+	rlwinm	24, 19, 26, 22, 29
+	lwzx	21, 10, 21
+	lwzx	22, 10, 22
+	lwzx	23, 10, 23
+	lwzx	24, 10, 24
+	xor	12, 12, 21
+	xor	15, 15, 22
+	xor	14, 14, 23
+	xor	16, 16, 24
+	rlwinm	21, 20, 2, 22, 29
+	rlwinm	22, 17, 2, 22, 29
+	rlwinm	23, 19, 2, 22, 29
+	rlwinm	24, 18, 2, 22, 29
+	lwzx	21, 11, 21
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	xor	12, 12, 21
+	xor	15, 15, 22
+	xor	14, 14, 23
+	xor	16, 16, 24
+	ld	17, 0(25)
+	ld	18, 8(25)
+	rldimi	12, 15, 32, 0
+	rldimi	14, 16, 32, 0
+	addi	25, 25, 16
+	#   XOR in Key Schedule
+	xor	12, 12, 17
+	xor	14, 14, 18
+	bdnz	L_AES_ECB_encrypt_loop_nr
+	srdi	15, 12, 32
+	srdi	16, 14, 32
+	rlwinm	17, 12, 10, 22, 29
+	rlwinm	18, 14, 10, 22, 29
+	rlwinm	19, 15, 10, 22, 29
+	rlwinm	20, 16, 10, 22, 29
+	lwzx	17, 8, 17
+	lwzx	18, 8, 18
+	lwzx	19, 8, 19
+	lwzx	20, 8, 20
+	rlwinm	21, 15, 18, 22, 29
+	rlwinm	22, 14, 18, 22, 29
+	rlwinm	23, 16, 18, 22, 29
+	rlwinm	24, 12, 18, 22, 29
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	lwzx	24, 9, 24
+	xor	17, 17, 21
+	xor	19, 19, 22
+	xor	18, 18, 23
+	xor	20, 20, 24
+	rlwinm	21, 14, 26, 22, 29
+	rlwinm	22, 16, 26, 22, 29
+	rlwinm	23, 12, 26, 22, 29
+	rlwinm	24, 15, 26, 22, 29
+	lwzx	21, 10, 21
+	lwzx	22, 10, 22
+	lwzx	23, 10, 23
+	lwzx	24, 10, 24
+	xor	17, 17, 21
+	xor	19, 19, 22
+	xor	18, 18, 23
+	xor	20, 20, 24
+	rlwinm	21, 16, 2, 22, 29
+	rlwinm	22, 12, 2, 22, 29
+	rlwinm	23, 15, 2, 22, 29
+	rlwinm	24, 14, 2, 22, 29
+	lwzx	21, 11, 21
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	xor	17, 17, 21
+	xor	19, 19, 22
+	xor	18, 18, 23
+	xor	20, 20, 24
+	ld	12, 0(25)
+	ld	14, 8(25)
+	rldimi	17, 19, 32, 0
+	rldimi	18, 20, 32, 0
+	addi	25, 25, 16
+	#   XOR in Key Schedule
+	xor	17, 17, 12
+	xor	18, 18, 14
+	srdi	19, 17, 32
+	srdi	20, 18, 32
+	rlwinm	12, 17, 2, 22, 29
+	rlwinm	14, 18, 2, 22, 29
+	rlwinm	16, 19, 2, 22, 29
+	rlwinm	15, 20, 2, 22, 29
+	lwzx	12, 9, 12
+	lwzx	14, 9, 14
+	lwzx	16, 9, 16
+	lwzx	15, 9, 15
+	rlwinm	21, 18, 26, 22, 29
+	rlwinm	22, 20, 26, 22, 29
+	rldimi	14, 16, 32, 24
+	rldimi	12, 15, 32, 24
+	rlwinm	23, 17, 26, 22, 29
+	rlwinm	24, 19, 26, 22, 29
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	lwzx	24, 9, 24
+	rldimi	12, 21, 40, 16
+	rlwimi	12, 22, 8, 16, 23
+	rldimi	14, 23, 40, 16
+	rlwimi	14, 24, 8, 16, 23
+	rlwinm	21, 19, 18, 22, 29
+	rlwinm	22, 18, 18, 22, 29
+	rlwinm	23, 20, 18, 22, 29
+	rlwinm	24, 17, 18, 22, 29
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	lwzx	24, 9, 24
+	rldimi	12, 21, 48, 8
+	rlwimi	12, 22, 16, 8, 15
+	rldimi	14, 23, 48, 8
+	rlwimi	14, 24, 16, 8, 15
+	rlwinm	21, 17, 10, 22, 29
+	rlwinm	22, 19, 10, 22, 29
+	rlwinm	23, 18, 10, 22, 29
+	rlwinm	24, 20, 10, 22, 29
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	lwzx	24, 9, 24
+	ld	17, 0(25)
+	ld	18, 8(25)
+	rldimi	12, 21, 56, 0
+	rlwimi	12, 22, 24, 0, 7
+	rldimi	14, 23, 56, 0
+	rlwimi	14, 24, 24, 0, 7
+	addi	25, 25, 16
+	#   XOR in Key Schedule
+	xor	12, 12, 17
+	xor	14, 14, 18
+	std	12, 0(4)
+	std	14, 8(4)
+	addi	3, 3, 16
+	addi	4, 4, 16
+	addic.	5, 5, -16
+	bne	L_AES_ECB_encrypt_loop_block_128
+	ld	0, 96(1)
+	mtlr	0
+	ld	14, 0(1)
+	ld	15, 8(1)
+	ld	16, 16(1)
+	ld	17, 24(1)
+	ld	18, 32(1)
+	ld	19, 40(1)
+	ld	20, 48(1)
+	ld	21, 56(1)
+	ld	22, 64(1)
+	ld	23, 72(1)
+	ld	24, 80(1)
+	ld	25, 88(1)
+	addi	1, 1, 0x68
+	blr
+#ifndef __APPLE__
+	.size	AES_ECB_encrypt,.-AES_ECB_encrypt
+#endif /* __APPLE__ */
+#endif /* HAVE_AESCCM || HAVE_AESGCM || WOLFSSL_AES_DIRECT ||
+        * WOLFSSL_AES_COUNTER || HAVE_AES_ECB */
+#ifdef HAVE_AES_CBC
+#ifndef __APPLE__
+.text
+.globl	AES_CBC_encrypt
+.type	AES_CBC_encrypt,@function
+.align	16
+AES_CBC_encrypt:
+#else
+.section	__TEXT,__text
+.globl	_AES_CBC_encrypt
+.p2align	4
+_AES_CBC_encrypt:
+#endif /* __APPLE__ */
+	stdu	1, -112(1)
+	mflr	0
+	std	0, 104(1)
+	std	14, 0(1)
+	std	15, 8(1)
+	std	16, 16(1)
+	std	17, 24(1)
+	std	18, 32(1)
+	std	19, 40(1)
+	std	20, 48(1)
+	std	21, 56(1)
+	std	22, 64(1)
+	std	23, 72(1)
+	std	24, 80(1)
+	std	25, 88(1)
+	std	26, 96(1)
+#if defined(__ARCH_PWR8) || (defined(_ARCH_PWR8))
+	addis	9, 2, L_AES_PPC64_te4_0@toc@ha
+	addi	9, 9, L_AES_PPC64_te4_0@toc@l
+#else
+	lis	9, L_AES_PPC64_te4_0@ha
+	la	9, L_AES_PPC64_te4_0@l(9)
+#endif /* defined(__ARCH_PWR8) || (defined(_ARCH_PWR8)) */
+	ld	14, 0(8)
+	ld	15, 8(8)
+	addi	10, 9, 0x400
+	addi	11, 9, 0x800
+	addi	12, 9, 0xc00
+L_AES_CBC_encrypt_loop_block:
+	addi	26, 6, 0
+	ld	18, 0(3)
+	ld	19, 8(3)
+	xor	14, 14, 18
+	xor	15, 15, 19
+	ld	18, 0(26)
+	ld	19, 8(26)
+	addi	26, 26, 16
+	rldicr	14, 14, 32, 63
+	rldicr	15, 15, 32, 63
+	# Round: 0 - XOR in key schedule
+	xor	14, 14, 18
+	xor	15, 15, 19
+	addi	0, 7, -2
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+	ld	21, 0(9)
+	ld	21, 64(9)
+	ld	21, 128(9)
+	ld	21, 192(9)
+	ld	21, 256(9)
+	ld	21, 320(9)
+	ld	21, 384(9)
+	ld	21, 448(9)
+	ld	21, 512(9)
+	ld	21, 576(9)
+	ld	21, 640(9)
+	ld	21, 704(9)
+	ld	21, 768(9)
+	ld	21, 832(9)
+	ld	21, 896(9)
+	ld	21, 960(9)
+	ld	21, 1024(9)
+	ld	21, 1088(9)
+	ld	21, 1152(9)
+	ld	21, 1216(9)
+	ld	21, 1280(9)
+	ld	21, 1344(9)
+	ld	21, 1408(9)
+	ld	21, 1472(9)
+	ld	21, 1536(9)
+	ld	21, 1600(9)
+	ld	21, 1664(9)
+	ld	21, 1728(9)
+	ld	21, 1792(9)
+	ld	21, 1856(9)
+	ld	21, 1920(9)
+	ld	21, 1984(9)
+	ld	21, 2048(9)
+	ld	21, 2112(9)
+	ld	21, 2176(9)
+	ld	21, 2240(9)
+	ld	21, 2304(9)
+	ld	21, 2368(9)
+	ld	21, 2432(9)
+	ld	21, 2496(9)
+	ld	21, 2560(9)
+	ld	21, 2624(9)
+	ld	21, 2688(9)
+	ld	21, 2752(9)
+	ld	21, 2816(9)
+	ld	21, 2880(9)
+	ld	21, 2944(9)
+	ld	21, 3008(9)
+	ld	21, 3072(9)
+	ld	21, 3136(9)
+	ld	21, 3200(9)
+	ld	21, 3264(9)
+	ld	21, 3328(9)
+	ld	21, 3392(9)
+	ld	21, 3456(9)
+	ld	21, 3520(9)
+	ld	21, 3584(9)
+	ld	21, 3648(9)
+	ld	21, 3712(9)
+	ld	21, 3776(9)
+	ld	21, 3840(9)
+	ld	21, 3904(9)
+	ld	21, 3968(9)
+	ld	21, 4032(9)
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+	srwi	0, 0, 1
+	mtctr	0
+L_AES_CBC_encrypt_loop_nr:
+	srdi	16, 14, 32
+	srdi	17, 15, 32
+	rlwinm	18, 14, 10, 22, 29
+	rlwinm	19, 15, 10, 22, 29
+	rlwinm	20, 16, 10, 22, 29
+	rlwinm	21, 17, 10, 22, 29
+	lwzx	18, 9, 18
+	lwzx	19, 9, 19
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	rlwinm	22, 16, 18, 22, 29
+	rlwinm	23, 15, 18, 22, 29
+	rlwinm	24, 17, 18, 22, 29
+	rlwinm	25, 14, 18, 22, 29
+	lwzx	22, 10, 22
+	lwzx	23, 10, 23
+	lwzx	24, 10, 24
+	lwzx	25, 10, 25
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 15, 26, 22, 29
+	rlwinm	23, 17, 26, 22, 29
+	rlwinm	24, 14, 26, 22, 29
+	rlwinm	25, 16, 26, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 17, 2, 22, 29
+	rlwinm	23, 14, 2, 22, 29
+	rlwinm	24, 16, 2, 22, 29
+	rlwinm	25, 15, 2, 22, 29
+	lwzx	22, 12, 22
+	lwzx	23, 12, 23
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	ld	14, 0(26)
+	ld	15, 8(26)
+	rldimi	18, 20, 32, 0
+	rldimi	19, 21, 32, 0
+	addi	26, 26, 16
+	#   XOR in Key Schedule
+	xor	18, 18, 14
+	xor	19, 19, 15
+	srdi	20, 18, 32
+	srdi	21, 19, 32
+	rlwinm	14, 18, 10, 22, 29
+	rlwinm	15, 19, 10, 22, 29
+	rlwinm	16, 20, 10, 22, 29
+	rlwinm	17, 21, 10, 22, 29
+	lwzx	14, 9, 14
+	lwzx	15, 9, 15
+	lwzx	16, 9, 16
+	lwzx	17, 9, 17
+	rlwinm	22, 20, 18, 22, 29
+	rlwinm	23, 19, 18, 22, 29
+	rlwinm	24, 21, 18, 22, 29
+	rlwinm	25, 18, 18, 22, 29
+	lwzx	22, 10, 22
+	lwzx	23, 10, 23
+	lwzx	24, 10, 24
+	lwzx	25, 10, 25
+	xor	14, 14, 22
+	xor	16, 16, 23
+	xor	15, 15, 24
+	xor	17, 17, 25
+	rlwinm	22, 19, 26, 22, 29
+	rlwinm	23, 21, 26, 22, 29
+	rlwinm	24, 18, 26, 22, 29
+	rlwinm	25, 20, 26, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	xor	14, 14, 22
+	xor	16, 16, 23
+	xor	15, 15, 24
+	xor	17, 17, 25
+	rlwinm	22, 21, 2, 22, 29
+	rlwinm	23, 18, 2, 22, 29
+	rlwinm	24, 20, 2, 22, 29
+	rlwinm	25, 19, 2, 22, 29
+	lwzx	22, 12, 22
+	lwzx	23, 12, 23
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	xor	14, 14, 22
+	xor	16, 16, 23
+	xor	15, 15, 24
+	xor	17, 17, 25
+	ld	18, 0(26)
+	ld	19, 8(26)
+	rldimi	14, 16, 32, 0
+	rldimi	15, 17, 32, 0
+	addi	26, 26, 16
+	#   XOR in Key Schedule
+	xor	14, 14, 18
+	xor	15, 15, 19
+	bdnz	L_AES_CBC_encrypt_loop_nr
+	srdi	16, 14, 32
+	srdi	17, 15, 32
+	rlwinm	18, 14, 10, 22, 29
+	rlwinm	19, 15, 10, 22, 29
+	rlwinm	20, 16, 10, 22, 29
+	rlwinm	21, 17, 10, 22, 29
+	lwzx	18, 9, 18
+	lwzx	19, 9, 19
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	rlwinm	22, 16, 18, 22, 29
+	rlwinm	23, 15, 18, 22, 29
+	rlwinm	24, 17, 18, 22, 29
+	rlwinm	25, 14, 18, 22, 29
+	lwzx	22, 10, 22
+	lwzx	23, 10, 23
+	lwzx	24, 10, 24
+	lwzx	25, 10, 25
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 15, 26, 22, 29
+	rlwinm	23, 17, 26, 22, 29
+	rlwinm	24, 14, 26, 22, 29
+	rlwinm	25, 16, 26, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 17, 2, 22, 29
+	rlwinm	23, 14, 2, 22, 29
+	rlwinm	24, 16, 2, 22, 29
+	rlwinm	25, 15, 2, 22, 29
+	lwzx	22, 12, 22
+	lwzx	23, 12, 23
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	ld	14, 0(26)
+	ld	15, 8(26)
+	rldimi	18, 20, 32, 0
+	rldimi	19, 21, 32, 0
+	addi	26, 26, 16
+	#   XOR in Key Schedule
+	xor	18, 18, 14
+	xor	19, 19, 15
+	srdi	20, 18, 32
+	srdi	21, 19, 32
+	rlwinm	14, 18, 2, 22, 29
+	rlwinm	15, 19, 2, 22, 29
+	rlwinm	17, 20, 2, 22, 29
+	rlwinm	16, 21, 2, 22, 29
+	lwzx	14, 10, 14
+	lwzx	15, 10, 15
+	lwzx	17, 10, 17
+	lwzx	16, 10, 16
+	rlwinm	22, 19, 26, 22, 29
+	rlwinm	23, 21, 26, 22, 29
+	rldimi	15, 17, 32, 24
+	rldimi	14, 16, 32, 24
+	rlwinm	24, 18, 26, 22, 29
+	rlwinm	25, 20, 26, 22, 29
+	lwzx	22, 10, 22
+	lwzx	23, 10, 23
+	lwzx	24, 10, 24
+	lwzx	25, 10, 25
+	rldimi	14, 22, 40, 16
+	rlwimi	14, 23, 8, 16, 23
+	rldimi	15, 24, 40, 16
+	rlwimi	15, 25, 8, 16, 23
+	rlwinm	22, 20, 18, 22, 29
+	rlwinm	23, 19, 18, 22, 29
+	rlwinm	24, 21, 18, 22, 29
+	rlwinm	25, 18, 18, 22, 29
+	lwzx	22, 10, 22
+	lwzx	23, 10, 23
+	lwzx	24, 10, 24
+	lwzx	25, 10, 25
+	rldimi	14, 22, 48, 8
+	rlwimi	14, 23, 16, 8, 15
+	rldimi	15, 24, 48, 8
+	rlwimi	15, 25, 16, 8, 15
+	rlwinm	22, 18, 10, 22, 29
+	rlwinm	23, 20, 10, 22, 29
+	rlwinm	24, 19, 10, 22, 29
+	rlwinm	25, 21, 10, 22, 29
+	lwzx	22, 10, 22
+	lwzx	23, 10, 23
+	lwzx	24, 10, 24
+	lwzx	25, 10, 25
+	ld	18, 0(26)
+	ld	19, 8(26)
+	rldimi	14, 22, 56, 0
+	rlwimi	14, 23, 24, 0, 7
+	rldimi	15, 24, 56, 0
+	rlwimi	15, 25, 24, 0, 7
+	addi	26, 26, 16
+	#   XOR in Key Schedule
+	xor	14, 14, 18
+	xor	15, 15, 19
+	std	14, 0(4)
+	std	15, 8(4)
+	addi	3, 3, 16
+	addi	4, 4, 16
+	addic.	5, 5, -16
+	bne	L_AES_CBC_encrypt_loop_block
+	std	14, 0(8)
+	std	15, 8(8)
+	ld	0, 104(1)
+	mtlr	0
+	ld	14, 0(1)
+	ld	15, 8(1)
+	ld	16, 16(1)
+	ld	17, 24(1)
+	ld	18, 32(1)
+	ld	19, 40(1)
+	ld	20, 48(1)
+	ld	21, 56(1)
+	ld	22, 64(1)
+	ld	23, 72(1)
+	ld	24, 80(1)
+	ld	25, 88(1)
+	ld	26, 96(1)
+	addi	1, 1, 0x70
+	blr
+#ifndef __APPLE__
+	.size	AES_CBC_encrypt,.-AES_CBC_encrypt
+#endif /* __APPLE__ */
+#endif /* HAVE_AES_CBC */
+#ifdef WOLFSSL_AES_COUNTER
+#ifndef __APPLE__
+.text
+.globl	AES_CTR_encrypt
+.type	AES_CTR_encrypt,@function
+.align	16
+AES_CTR_encrypt:
+#else
+.section	__TEXT,__text
+.globl	_AES_CTR_encrypt
+.p2align	4
+_AES_CTR_encrypt:
+#endif /* __APPLE__ */
+	stdu	1, -128(1)
+	mflr	0
+	std	0, 120(1)
+	std	14, 0(1)
+	std	15, 8(1)
+	std	16, 16(1)
+	std	17, 24(1)
+	std	18, 32(1)
+	std	19, 40(1)
+	std	20, 48(1)
+	std	21, 56(1)
+	std	22, 64(1)
+	std	23, 72(1)
+	std	24, 80(1)
+	std	25, 88(1)
+	std	26, 96(1)
+	std	27, 104(1)
+	std	28, 112(1)
+#if defined(__ARCH_PWR8) || (defined(_ARCH_PWR8))
+	addis	9, 2, L_AES_PPC64_te4_0@toc@ha
+	addi	9, 9, L_AES_PPC64_te4_0@toc@l
+#else
+	lis	9, L_AES_PPC64_te4_0@ha
+	la	9, L_AES_PPC64_te4_0@l(9)
+#endif /* defined(__ARCH_PWR8) || (defined(_ARCH_PWR8)) */
+	ld	22, 0(8)
+	ld	23, 8(8)
+	addi	10, 9, 0x400
+	addi	11, 9, 0x800
+	addi	12, 9, 0xc00
+L_AES_CTR_encrypt_loop_block_128:
+	addi	28, 6, 0
+	ld	18, 0(28)
+	ld	19, 8(28)
+	addi	28, 28, 16
+	rldicr	14, 22, 32, 63
+	rldicr	15, 23, 32, 63
+	# Round: 0 - XOR in key schedule
+	xor	14, 14, 18
+	xor	15, 15, 19
+	addi	0, 7, -2
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+	ld	21, 0(9)
+	ld	21, 64(9)
+	ld	21, 128(9)
+	ld	21, 192(9)
+	ld	21, 256(9)
+	ld	21, 320(9)
+	ld	21, 384(9)
+	ld	21, 448(9)
+	ld	21, 512(9)
+	ld	21, 576(9)
+	ld	21, 640(9)
+	ld	21, 704(9)
+	ld	21, 768(9)
+	ld	21, 832(9)
+	ld	21, 896(9)
+	ld	21, 960(9)
+	ld	21, 1024(9)
+	ld	21, 1088(9)
+	ld	21, 1152(9)
+	ld	21, 1216(9)
+	ld	21, 1280(9)
+	ld	21, 1344(9)
+	ld	21, 1408(9)
+	ld	21, 1472(9)
+	ld	21, 1536(9)
+	ld	21, 1600(9)
+	ld	21, 1664(9)
+	ld	21, 1728(9)
+	ld	21, 1792(9)
+	ld	21, 1856(9)
+	ld	21, 1920(9)
+	ld	21, 1984(9)
+	ld	21, 2048(9)
+	ld	21, 2112(9)
+	ld	21, 2176(9)
+	ld	21, 2240(9)
+	ld	21, 2304(9)
+	ld	21, 2368(9)
+	ld	21, 2432(9)
+	ld	21, 2496(9)
+	ld	21, 2560(9)
+	ld	21, 2624(9)
+	ld	21, 2688(9)
+	ld	21, 2752(9)
+	ld	21, 2816(9)
+	ld	21, 2880(9)
+	ld	21, 2944(9)
+	ld	21, 3008(9)
+	ld	21, 3072(9)
+	ld	21, 3136(9)
+	ld	21, 3200(9)
+	ld	21, 3264(9)
+	ld	21, 3328(9)
+	ld	21, 3392(9)
+	ld	21, 3456(9)
+	ld	21, 3520(9)
+	ld	21, 3584(9)
+	ld	21, 3648(9)
+	ld	21, 3712(9)
+	ld	21, 3776(9)
+	ld	21, 3840(9)
+	ld	21, 3904(9)
+	ld	21, 3968(9)
+	ld	21, 4032(9)
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+	srwi	0, 0, 1
+	mtctr	0
+L_AES_CTR_encrypt_loop_nr:
+	srdi	16, 14, 32
+	srdi	17, 15, 32
+	rlwinm	18, 14, 10, 22, 29
+	rlwinm	19, 15, 10, 22, 29
+	rlwinm	20, 16, 10, 22, 29
+	rlwinm	21, 17, 10, 22, 29
+	lwzx	18, 9, 18
+	lwzx	19, 9, 19
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	rlwinm	24, 16, 18, 22, 29
+	rlwinm	25, 15, 18, 22, 29
+	rlwinm	26, 17, 18, 22, 29
+	rlwinm	27, 14, 18, 22, 29
+	lwzx	24, 10, 24
+	lwzx	25, 10, 25
+	lwzx	26, 10, 26
+	lwzx	27, 10, 27
+	xor	18, 18, 24
+	xor	20, 20, 25
+	xor	19, 19, 26
+	xor	21, 21, 27
+	rlwinm	24, 15, 26, 22, 29
+	rlwinm	25, 17, 26, 22, 29
+	rlwinm	26, 14, 26, 22, 29
+	rlwinm	27, 16, 26, 22, 29
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	lwzx	26, 11, 26
+	lwzx	27, 11, 27
+	xor	18, 18, 24
+	xor	20, 20, 25
+	xor	19, 19, 26
+	xor	21, 21, 27
+	rlwinm	24, 17, 2, 22, 29
+	rlwinm	25, 14, 2, 22, 29
+	rlwinm	26, 16, 2, 22, 29
+	rlwinm	27, 15, 2, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	xor	18, 18, 24
+	xor	20, 20, 25
+	xor	19, 19, 26
+	xor	21, 21, 27
+	ld	14, 0(28)
+	ld	15, 8(28)
+	rldimi	18, 20, 32, 0
+	rldimi	19, 21, 32, 0
+	addi	28, 28, 16
+	#   XOR in Key Schedule
+	xor	18, 18, 14
+	xor	19, 19, 15
+	srdi	20, 18, 32
+	srdi	21, 19, 32
+	rlwinm	14, 18, 10, 22, 29
+	rlwinm	15, 19, 10, 22, 29
+	rlwinm	16, 20, 10, 22, 29
+	rlwinm	17, 21, 10, 22, 29
+	lwzx	14, 9, 14
+	lwzx	15, 9, 15
+	lwzx	16, 9, 16
+	lwzx	17, 9, 17
+	rlwinm	24, 20, 18, 22, 29
+	rlwinm	25, 19, 18, 22, 29
+	rlwinm	26, 21, 18, 22, 29
+	rlwinm	27, 18, 18, 22, 29
+	lwzx	24, 10, 24
+	lwzx	25, 10, 25
+	lwzx	26, 10, 26
+	lwzx	27, 10, 27
+	xor	14, 14, 24
+	xor	16, 16, 25
+	xor	15, 15, 26
+	xor	17, 17, 27
+	rlwinm	24, 19, 26, 22, 29
+	rlwinm	25, 21, 26, 22, 29
+	rlwinm	26, 18, 26, 22, 29
+	rlwinm	27, 20, 26, 22, 29
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	lwzx	26, 11, 26
+	lwzx	27, 11, 27
+	xor	14, 14, 24
+	xor	16, 16, 25
+	xor	15, 15, 26
+	xor	17, 17, 27
+	rlwinm	24, 21, 2, 22, 29
+	rlwinm	25, 18, 2, 22, 29
+	rlwinm	26, 20, 2, 22, 29
+	rlwinm	27, 19, 2, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	xor	14, 14, 24
+	xor	16, 16, 25
+	xor	15, 15, 26
+	xor	17, 17, 27
+	ld	18, 0(28)
+	ld	19, 8(28)
+	rldimi	14, 16, 32, 0
+	rldimi	15, 17, 32, 0
+	addi	28, 28, 16
+	#   XOR in Key Schedule
+	xor	14, 14, 18
+	xor	15, 15, 19
+	bdnz	L_AES_CTR_encrypt_loop_nr
+	srdi	16, 14, 32
+	srdi	17, 15, 32
+	rlwinm	18, 14, 10, 22, 29
+	rlwinm	19, 15, 10, 22, 29
+	rlwinm	20, 16, 10, 22, 29
+	rlwinm	21, 17, 10, 22, 29
+	lwzx	18, 9, 18
+	lwzx	19, 9, 19
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	rlwinm	24, 16, 18, 22, 29
+	rlwinm	25, 15, 18, 22, 29
+	rlwinm	26, 17, 18, 22, 29
+	rlwinm	27, 14, 18, 22, 29
+	lwzx	24, 10, 24
+	lwzx	25, 10, 25
+	lwzx	26, 10, 26
+	lwzx	27, 10, 27
+	xor	18, 18, 24
+	xor	20, 20, 25
+	xor	19, 19, 26
+	xor	21, 21, 27
+	rlwinm	24, 15, 26, 22, 29
+	rlwinm	25, 17, 26, 22, 29
+	rlwinm	26, 14, 26, 22, 29
+	rlwinm	27, 16, 26, 22, 29
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	lwzx	26, 11, 26
+	lwzx	27, 11, 27
+	xor	18, 18, 24
+	xor	20, 20, 25
+	xor	19, 19, 26
+	xor	21, 21, 27
+	rlwinm	24, 17, 2, 22, 29
+	rlwinm	25, 14, 2, 22, 29
+	rlwinm	26, 16, 2, 22, 29
+	rlwinm	27, 15, 2, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	xor	18, 18, 24
+	xor	20, 20, 25
+	xor	19, 19, 26
+	xor	21, 21, 27
+	ld	14, 0(28)
+	ld	15, 8(28)
+	rldimi	18, 20, 32, 0
+	rldimi	19, 21, 32, 0
+	addi	28, 28, 16
+	#   XOR in Key Schedule
+	xor	18, 18, 14
+	xor	19, 19, 15
+	srdi	20, 18, 32
+	srdi	21, 19, 32
+	rlwinm	14, 18, 2, 22, 29
+	rlwinm	15, 19, 2, 22, 29
+	rlwinm	17, 20, 2, 22, 29
+	rlwinm	16, 21, 2, 22, 29
+	lwzx	14, 10, 14
+	lwzx	15, 10, 15
+	lwzx	17, 10, 17
+	lwzx	16, 10, 16
+	rlwinm	24, 19, 26, 22, 29
+	rlwinm	25, 21, 26, 22, 29
+	rldimi	15, 17, 32, 24
+	rldimi	14, 16, 32, 24
+	rlwinm	26, 18, 26, 22, 29
+	rlwinm	27, 20, 26, 22, 29
+	lwzx	24, 10, 24
+	lwzx	25, 10, 25
+	lwzx	26, 10, 26
+	lwzx	27, 10, 27
+	rldimi	14, 24, 40, 16
+	rlwimi	14, 25, 8, 16, 23
+	rldimi	15, 26, 40, 16
+	rlwimi	15, 27, 8, 16, 23
+	rlwinm	24, 20, 18, 22, 29
+	rlwinm	25, 19, 18, 22, 29
+	rlwinm	26, 21, 18, 22, 29
+	rlwinm	27, 18, 18, 22, 29
+	lwzx	24, 10, 24
+	lwzx	25, 10, 25
+	lwzx	26, 10, 26
+	lwzx	27, 10, 27
+	rldimi	14, 24, 48, 8
+	rlwimi	14, 25, 16, 8, 15
+	rldimi	15, 26, 48, 8
+	rlwimi	15, 27, 16, 8, 15
+	rlwinm	24, 18, 10, 22, 29
+	rlwinm	25, 20, 10, 22, 29
+	rlwinm	26, 19, 10, 22, 29
+	rlwinm	27, 21, 10, 22, 29
+	lwzx	24, 10, 24
+	lwzx	25, 10, 25
+	lwzx	26, 10, 26
+	lwzx	27, 10, 27
+	ld	18, 0(28)
+	ld	19, 8(28)
+	rldimi	14, 24, 56, 0
+	rlwimi	14, 25, 24, 0, 7
+	rldimi	15, 26, 56, 0
+	rlwimi	15, 27, 24, 0, 7
+	addi	28, 28, 16
+	#   XOR in Key Schedule
+	xor	14, 14, 18
+	xor	15, 15, 19
+	ld	18, 0(3)
+	ld	19, 8(3)
+	xor	14, 14, 18
+	xor	15, 15, 19
+	std	14, 0(4)
+	std	15, 8(4)
+	addic	23, 23, 1
+	addze	22, 22
+	addi	3, 3, 16
+	addi	4, 4, 16
+	addic.	5, 5, -16
+	bne	L_AES_CTR_encrypt_loop_block_128
+	std	22, 0(8)
+	std	23, 8(8)
+	ld	0, 120(1)
+	mtlr	0
+	ld	14, 0(1)
+	ld	15, 8(1)
+	ld	16, 16(1)
+	ld	17, 24(1)
+	ld	18, 32(1)
+	ld	19, 40(1)
+	ld	20, 48(1)
+	ld	21, 56(1)
+	ld	22, 64(1)
+	ld	23, 72(1)
+	ld	24, 80(1)
+	ld	25, 88(1)
+	ld	26, 96(1)
+	ld	27, 104(1)
+	ld	28, 112(1)
+	addi	1, 1, 0x80
+	blr
+#ifndef __APPLE__
+	.size	AES_CTR_encrypt,.-AES_CTR_encrypt
+#endif /* __APPLE__ */
+#endif /* WOLFSSL_AES_COUNTER */
+#ifdef HAVE_AESGCM
+#ifndef __APPLE__
+.text
+.globl	AES_GCM_encrypt
+.type	AES_GCM_encrypt,@function
+.align	16
+AES_GCM_encrypt:
+#else
+.section	__TEXT,__text
+.globl	_AES_GCM_encrypt
+.p2align	4
+_AES_GCM_encrypt:
+#endif /* __APPLE__ */
+	stdu	1, -128(1)
+	mflr	0
+	std	0, 120(1)
+	std	14, 0(1)
+	std	15, 8(1)
+	std	16, 16(1)
+	std	17, 24(1)
+	std	18, 32(1)
+	std	19, 40(1)
+	std	20, 48(1)
+	std	21, 56(1)
+	std	22, 64(1)
+	std	23, 72(1)
+	std	24, 80(1)
+	std	25, 88(1)
+	std	26, 96(1)
+	std	27, 104(1)
+	std	28, 112(1)
+#if defined(__ARCH_PWR8) || (defined(_ARCH_PWR8))
+	addis	9, 2, L_AES_PPC64_te4_0@toc@ha
+	addi	9, 9, L_AES_PPC64_te4_0@toc@l
+#else
+	lis	9, L_AES_PPC64_te4_0@ha
+	la	9, L_AES_PPC64_te4_0@l(9)
+#endif /* defined(__ARCH_PWR8) || (defined(_ARCH_PWR8)) */
+	ld	26, 0(8)
+	ld	27, 8(8)
+	addi	10, 9, 0x400
+	addi	11, 9, 0x800
+	addi	12, 9, 0xc00
+L_AES_GCM_encrypt_loop_block:
+	addi	28, 6, 0
+	addi	17, 27, 1
+	ld	18, 0(28)
+	ld	19, 8(28)
+	addi	28, 28, 16
+	rldimi	27, 17, 0, 32
+	rldicr	14, 26, 32, 63
+	rldicr	15, 27, 32, 63
+	# Round: 0 - XOR in key schedule
+	xor	14, 14, 18
+	xor	15, 15, 19
+	addi	0, 7, -2
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+	ld	21, 0(9)
+	ld	21, 64(9)
+	ld	21, 128(9)
+	ld	21, 192(9)
+	ld	21, 256(9)
+	ld	21, 320(9)
+	ld	21, 384(9)
+	ld	21, 448(9)
+	ld	21, 512(9)
+	ld	21, 576(9)
+	ld	21, 640(9)
+	ld	21, 704(9)
+	ld	21, 768(9)
+	ld	21, 832(9)
+	ld	21, 896(9)
+	ld	21, 960(9)
+	ld	21, 1024(9)
+	ld	21, 1088(9)
+	ld	21, 1152(9)
+	ld	21, 1216(9)
+	ld	21, 1280(9)
+	ld	21, 1344(9)
+	ld	21, 1408(9)
+	ld	21, 1472(9)
+	ld	21, 1536(9)
+	ld	21, 1600(9)
+	ld	21, 1664(9)
+	ld	21, 1728(9)
+	ld	21, 1792(9)
+	ld	21, 1856(9)
+	ld	21, 1920(9)
+	ld	21, 1984(9)
+	ld	21, 2048(9)
+	ld	21, 2112(9)
+	ld	21, 2176(9)
+	ld	21, 2240(9)
+	ld	21, 2304(9)
+	ld	21, 2368(9)
+	ld	21, 2432(9)
+	ld	21, 2496(9)
+	ld	21, 2560(9)
+	ld	21, 2624(9)
+	ld	21, 2688(9)
+	ld	21, 2752(9)
+	ld	21, 2816(9)
+	ld	21, 2880(9)
+	ld	21, 2944(9)
+	ld	21, 3008(9)
+	ld	21, 3072(9)
+	ld	21, 3136(9)
+	ld	21, 3200(9)
+	ld	21, 3264(9)
+	ld	21, 3328(9)
+	ld	21, 3392(9)
+	ld	21, 3456(9)
+	ld	21, 3520(9)
+	ld	21, 3584(9)
+	ld	21, 3648(9)
+	ld	21, 3712(9)
+	ld	21, 3776(9)
+	ld	21, 3840(9)
+	ld	21, 3904(9)
+	ld	21, 3968(9)
+	ld	21, 4032(9)
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+	srwi	0, 0, 1
+	mtctr	0
+L_AES_GCM_encrypt_loop_nr:
+	srdi	16, 14, 32
+	srdi	17, 15, 32
+	rlwinm	18, 14, 10, 22, 29
+	rlwinm	19, 15, 10, 22, 29
+	rlwinm	20, 16, 10, 22, 29
+	rlwinm	21, 17, 10, 22, 29
+	lwzx	18, 9, 18
+	lwzx	19, 9, 19
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	rlwinm	22, 16, 18, 22, 29
+	rlwinm	23, 15, 18, 22, 29
+	rlwinm	24, 17, 18, 22, 29
+	rlwinm	25, 14, 18, 22, 29
+	lwzx	22, 10, 22
+	lwzx	23, 10, 23
+	lwzx	24, 10, 24
+	lwzx	25, 10, 25
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 15, 26, 22, 29
+	rlwinm	23, 17, 26, 22, 29
+	rlwinm	24, 14, 26, 22, 29
+	rlwinm	25, 16, 26, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 17, 2, 22, 29
+	rlwinm	23, 14, 2, 22, 29
+	rlwinm	24, 16, 2, 22, 29
+	rlwinm	25, 15, 2, 22, 29
+	lwzx	22, 12, 22
+	lwzx	23, 12, 23
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	ld	14, 0(28)
+	ld	15, 8(28)
+	rldimi	18, 20, 32, 0
+	rldimi	19, 21, 32, 0
+	addi	28, 28, 16
+	#   XOR in Key Schedule
+	xor	18, 18, 14
+	xor	19, 19, 15
+	srdi	20, 18, 32
+	srdi	21, 19, 32
+	rlwinm	14, 18, 10, 22, 29
+	rlwinm	15, 19, 10, 22, 29
+	rlwinm	16, 20, 10, 22, 29
+	rlwinm	17, 21, 10, 22, 29
+	lwzx	14, 9, 14
+	lwzx	15, 9, 15
+	lwzx	16, 9, 16
+	lwzx	17, 9, 17
+	rlwinm	22, 20, 18, 22, 29
+	rlwinm	23, 19, 18, 22, 29
+	rlwinm	24, 21, 18, 22, 29
+	rlwinm	25, 18, 18, 22, 29
+	lwzx	22, 10, 22
+	lwzx	23, 10, 23
+	lwzx	24, 10, 24
+	lwzx	25, 10, 25
+	xor	14, 14, 22
+	xor	16, 16, 23
+	xor	15, 15, 24
+	xor	17, 17, 25
+	rlwinm	22, 19, 26, 22, 29
+	rlwinm	23, 21, 26, 22, 29
+	rlwinm	24, 18, 26, 22, 29
+	rlwinm	25, 20, 26, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	xor	14, 14, 22
+	xor	16, 16, 23
+	xor	15, 15, 24
+	xor	17, 17, 25
+	rlwinm	22, 21, 2, 22, 29
+	rlwinm	23, 18, 2, 22, 29
+	rlwinm	24, 20, 2, 22, 29
+	rlwinm	25, 19, 2, 22, 29
+	lwzx	22, 12, 22
+	lwzx	23, 12, 23
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	xor	14, 14, 22
+	xor	16, 16, 23
+	xor	15, 15, 24
+	xor	17, 17, 25
+	ld	18, 0(28)
+	ld	19, 8(28)
+	rldimi	14, 16, 32, 0
+	rldimi	15, 17, 32, 0
+	addi	28, 28, 16
+	#   XOR in Key Schedule
+	xor	14, 14, 18
+	xor	15, 15, 19
+	bdnz	L_AES_GCM_encrypt_loop_nr
+	srdi	16, 14, 32
+	srdi	17, 15, 32
+	rlwinm	18, 14, 10, 22, 29
+	rlwinm	19, 15, 10, 22, 29
+	rlwinm	20, 16, 10, 22, 29
+	rlwinm	21, 17, 10, 22, 29
+	lwzx	18, 9, 18
+	lwzx	19, 9, 19
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	rlwinm	22, 16, 18, 22, 29
+	rlwinm	23, 15, 18, 22, 29
+	rlwinm	24, 17, 18, 22, 29
+	rlwinm	25, 14, 18, 22, 29
+	lwzx	22, 10, 22
+	lwzx	23, 10, 23
+	lwzx	24, 10, 24
+	lwzx	25, 10, 25
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 15, 26, 22, 29
+	rlwinm	23, 17, 26, 22, 29
+	rlwinm	24, 14, 26, 22, 29
+	rlwinm	25, 16, 26, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 17, 2, 22, 29
+	rlwinm	23, 14, 2, 22, 29
+	rlwinm	24, 16, 2, 22, 29
+	rlwinm	25, 15, 2, 22, 29
+	lwzx	22, 12, 22
+	lwzx	23, 12, 23
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	ld	14, 0(28)
+	ld	15, 8(28)
+	rldimi	18, 20, 32, 0
+	rldimi	19, 21, 32, 0
+	addi	28, 28, 16
+	#   XOR in Key Schedule
+	xor	18, 18, 14
+	xor	19, 19, 15
+	srdi	20, 18, 32
+	srdi	21, 19, 32
+	rlwinm	14, 18, 2, 22, 29
+	rlwinm	15, 19, 2, 22, 29
+	rlwinm	17, 20, 2, 22, 29
+	rlwinm	16, 21, 2, 22, 29
+	lwzx	14, 10, 14
+	lwzx	15, 10, 15
+	lwzx	17, 10, 17
+	lwzx	16, 10, 16
+	rlwinm	22, 19, 26, 22, 29
+	rlwinm	23, 21, 26, 22, 29
+	rldimi	15, 17, 32, 24
+	rldimi	14, 16, 32, 24
+	rlwinm	24, 18, 26, 22, 29
+	rlwinm	25, 20, 26, 22, 29
+	lwzx	22, 10, 22
+	lwzx	23, 10, 23
+	lwzx	24, 10, 24
+	lwzx	25, 10, 25
+	rldimi	14, 22, 40, 16
+	rlwimi	14, 23, 8, 16, 23
+	rldimi	15, 24, 40, 16
+	rlwimi	15, 25, 8, 16, 23
+	rlwinm	22, 20, 18, 22, 29
+	rlwinm	23, 19, 18, 22, 29
+	rlwinm	24, 21, 18, 22, 29
+	rlwinm	25, 18, 18, 22, 29
+	lwzx	22, 10, 22
+	lwzx	23, 10, 23
+	lwzx	24, 10, 24
+	lwzx	25, 10, 25
+	rldimi	14, 22, 48, 8
+	rlwimi	14, 23, 16, 8, 15
+	rldimi	15, 24, 48, 8
+	rlwimi	15, 25, 16, 8, 15
+	rlwinm	22, 18, 10, 22, 29
+	rlwinm	23, 20, 10, 22, 29
+	rlwinm	24, 19, 10, 22, 29
+	rlwinm	25, 21, 10, 22, 29
+	lwzx	22, 10, 22
+	lwzx	23, 10, 23
+	lwzx	24, 10, 24
+	lwzx	25, 10, 25
+	ld	18, 0(28)
+	ld	19, 8(28)
+	rldimi	14, 22, 56, 0
+	rlwimi	14, 23, 24, 0, 7
+	rldimi	15, 24, 56, 0
+	rlwimi	15, 25, 24, 0, 7
+	addi	28, 28, 16
+	#   XOR in Key Schedule
+	xor	14, 14, 18
+	xor	15, 15, 19
+	ld	18, 0(3)
+	ld	19, 8(3)
+	addi	3, 3, 16
+	addic.	5, 5, -16
+	xor	14, 14, 18
+	xor	15, 15, 19
+	std	14, 0(4)
+	std	15, 8(4)
+	addi	4, 4, 16
+	bne	L_AES_GCM_encrypt_loop_block
+	std	26, 0(8)
+	std	27, 8(8)
+	ld	0, 120(1)
+	mtlr	0
+	ld	14, 0(1)
+	ld	15, 8(1)
+	ld	16, 16(1)
+	ld	17, 24(1)
+	ld	18, 32(1)
+	ld	19, 40(1)
+	ld	20, 48(1)
+	ld	21, 56(1)
+	ld	22, 64(1)
+	ld	23, 72(1)
+	ld	24, 80(1)
+	ld	25, 88(1)
+	ld	26, 96(1)
+	ld	27, 104(1)
+	ld	28, 112(1)
+	addi	1, 1, 0x80
+	blr
+#ifndef __APPLE__
+	.size	AES_GCM_encrypt,.-AES_GCM_encrypt
+#endif /* __APPLE__ */
+#endif /* HAVE_AESGCM */
+#ifdef WOLFSSL_AES_XTS
+#ifndef __APPLE__
+.text
+.globl	AES_XTS_encrypt
+.type	AES_XTS_encrypt,@function
+.align	16
+AES_XTS_encrypt:
+#else
+.section	__TEXT,__text
+.globl	_AES_XTS_encrypt
+.p2align	4
+_AES_XTS_encrypt:
+#endif /* __APPLE__ */
+	stdu	1, -136(1)
+	mflr	0
+	std	0, 128(1)
+	std	14, 0(1)
+	std	15, 8(1)
+	std	16, 16(1)
+	std	17, 24(1)
+	std	18, 32(1)
+	std	19, 40(1)
+	std	20, 48(1)
+	std	21, 56(1)
+	std	22, 64(1)
+	std	23, 72(1)
+	std	24, 80(1)
+	std	25, 88(1)
+	std	26, 96(1)
+	std	27, 104(1)
+	std	28, 112(1)
+	std	29, 120(1)
+#if defined(__ARCH_PWR8) || (defined(_ARCH_PWR8))
+	addis	11, 2, L_AES_PPC64_te4_0@toc@ha
+	addi	11, 11, L_AES_PPC64_te4_0@toc@l
+#else
+	lis	11, L_AES_PPC64_te4_0@ha
+	la	11, L_AES_PPC64_te4_0@l(11)
+#endif /* defined(__ARCH_PWR8) || (defined(_ARCH_PWR8)) */
+	addi	12, 11, 0x400
+	addi	14, 11, 0x800
+	addi	15, 11, 0xc00
+	ld	28, 0(6)
+	ld	29, 8(6)
+	ld	20, 0(8)
+	ld	21, 8(8)
+	addi	8, 8, 16
+	rldicr	28, 28, 32, 63
+	rldicr	29, 29, 32, 63
+	# Round: 0 - XOR in key schedule
+	xor	28, 28, 20
+	xor	29, 29, 21
+	addi	0, 10, -2
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+	ld	23, 0(11)
+	ld	23, 64(11)
+	ld	23, 128(11)
+	ld	23, 192(11)
+	ld	23, 256(11)
+	ld	23, 320(11)
+	ld	23, 384(11)
+	ld	23, 448(11)
+	ld	23, 512(11)
+	ld	23, 576(11)
+	ld	23, 640(11)
+	ld	23, 704(11)
+	ld	23, 768(11)
+	ld	23, 832(11)
+	ld	23, 896(11)
+	ld	23, 960(11)
+	ld	23, 1024(11)
+	ld	23, 1088(11)
+	ld	23, 1152(11)
+	ld	23, 1216(11)
+	ld	23, 1280(11)
+	ld	23, 1344(11)
+	ld	23, 1408(11)
+	ld	23, 1472(11)
+	ld	23, 1536(11)
+	ld	23, 1600(11)
+	ld	23, 1664(11)
+	ld	23, 1728(11)
+	ld	23, 1792(11)
+	ld	23, 1856(11)
+	ld	23, 1920(11)
+	ld	23, 1984(11)
+	ld	23, 2048(11)
+	ld	23, 2112(11)
+	ld	23, 2176(11)
+	ld	23, 2240(11)
+	ld	23, 2304(11)
+	ld	23, 2368(11)
+	ld	23, 2432(11)
+	ld	23, 2496(11)
+	ld	23, 2560(11)
+	ld	23, 2624(11)
+	ld	23, 2688(11)
+	ld	23, 2752(11)
+	ld	23, 2816(11)
+	ld	23, 2880(11)
+	ld	23, 2944(11)
+	ld	23, 3008(11)
+	ld	23, 3072(11)
+	ld	23, 3136(11)
+	ld	23, 3200(11)
+	ld	23, 3264(11)
+	ld	23, 3328(11)
+	ld	23, 3392(11)
+	ld	23, 3456(11)
+	ld	23, 3520(11)
+	ld	23, 3584(11)
+	ld	23, 3648(11)
+	ld	23, 3712(11)
+	ld	23, 3776(11)
+	ld	23, 3840(11)
+	ld	23, 3904(11)
+	ld	23, 3968(11)
+	ld	23, 4032(11)
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+	srwi	0, 0, 1
+	mtctr	0
+L_AES_XTS_encrypt_loop_nr_tweak:
+	srdi	18, 28, 32
+	srdi	19, 29, 32
+	rlwinm	20, 28, 10, 22, 29
+	rlwinm	21, 29, 10, 22, 29
+	rlwinm	22, 18, 10, 22, 29
+	rlwinm	23, 19, 10, 22, 29
+	lwzx	20, 11, 20
+	lwzx	21, 11, 21
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	rlwinm	24, 18, 18, 22, 29
+	rlwinm	25, 29, 18, 22, 29
+	rlwinm	26, 19, 18, 22, 29
+	rlwinm	27, 28, 18, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	xor	20, 20, 24
+	xor	22, 22, 25
+	xor	21, 21, 26
+	xor	23, 23, 27
+	rlwinm	24, 29, 26, 22, 29
+	rlwinm	25, 19, 26, 22, 29
+	rlwinm	26, 28, 26, 22, 29
+	rlwinm	27, 18, 26, 22, 29
+	lwzx	24, 14, 24
+	lwzx	25, 14, 25
+	lwzx	26, 14, 26
+	lwzx	27, 14, 27
+	xor	20, 20, 24
+	xor	22, 22, 25
+	xor	21, 21, 26
+	xor	23, 23, 27
+	rlwinm	24, 19, 2, 22, 29
+	rlwinm	25, 28, 2, 22, 29
+	rlwinm	26, 18, 2, 22, 29
+	rlwinm	27, 29, 2, 22, 29
+	lwzx	24, 15, 24
+	lwzx	25, 15, 25
+	lwzx	26, 15, 26
+	lwzx	27, 15, 27
+	xor	20, 20, 24
+	xor	22, 22, 25
+	xor	21, 21, 26
+	xor	23, 23, 27
+	ld	28, 0(8)
+	ld	29, 8(8)
+	rldimi	20, 22, 32, 0
+	rldimi	21, 23, 32, 0
+	addi	8, 8, 16
+	#   XOR in Key Schedule
+	xor	20, 20, 28
+	xor	21, 21, 29
+	srdi	22, 20, 32
+	srdi	23, 21, 32
+	rlwinm	28, 20, 10, 22, 29
+	rlwinm	29, 21, 10, 22, 29
+	rlwinm	18, 22, 10, 22, 29
+	rlwinm	19, 23, 10, 22, 29
+	lwzx	28, 11, 28
+	lwzx	29, 11, 29
+	lwzx	18, 11, 18
+	lwzx	19, 11, 19
+	rlwinm	24, 22, 18, 22, 29
+	rlwinm	25, 21, 18, 22, 29
+	rlwinm	26, 23, 18, 22, 29
+	rlwinm	27, 20, 18, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	xor	28, 28, 24
+	xor	18, 18, 25
+	xor	29, 29, 26
+	xor	19, 19, 27
+	rlwinm	24, 21, 26, 22, 29
+	rlwinm	25, 23, 26, 22, 29
+	rlwinm	26, 20, 26, 22, 29
+	rlwinm	27, 22, 26, 22, 29
+	lwzx	24, 14, 24
+	lwzx	25, 14, 25
+	lwzx	26, 14, 26
+	lwzx	27, 14, 27
+	xor	28, 28, 24
+	xor	18, 18, 25
+	xor	29, 29, 26
+	xor	19, 19, 27
+	rlwinm	24, 23, 2, 22, 29
+	rlwinm	25, 20, 2, 22, 29
+	rlwinm	26, 22, 2, 22, 29
+	rlwinm	27, 21, 2, 22, 29
+	lwzx	24, 15, 24
+	lwzx	25, 15, 25
+	lwzx	26, 15, 26
+	lwzx	27, 15, 27
+	xor	28, 28, 24
+	xor	18, 18, 25
+	xor	29, 29, 26
+	xor	19, 19, 27
+	ld	20, 0(8)
+	ld	21, 8(8)
+	rldimi	28, 18, 32, 0
+	rldimi	29, 19, 32, 0
+	addi	8, 8, 16
+	#   XOR in Key Schedule
+	xor	28, 28, 20
+	xor	29, 29, 21
+	bdnz	L_AES_XTS_encrypt_loop_nr_tweak
+	srdi	18, 28, 32
+	srdi	19, 29, 32
+	rlwinm	20, 28, 10, 22, 29
+	rlwinm	21, 29, 10, 22, 29
+	rlwinm	22, 18, 10, 22, 29
+	rlwinm	23, 19, 10, 22, 29
+	lwzx	20, 11, 20
+	lwzx	21, 11, 21
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	rlwinm	24, 18, 18, 22, 29
+	rlwinm	25, 29, 18, 22, 29
+	rlwinm	26, 19, 18, 22, 29
+	rlwinm	27, 28, 18, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	xor	20, 20, 24
+	xor	22, 22, 25
+	xor	21, 21, 26
+	xor	23, 23, 27
+	rlwinm	24, 29, 26, 22, 29
+	rlwinm	25, 19, 26, 22, 29
+	rlwinm	26, 28, 26, 22, 29
+	rlwinm	27, 18, 26, 22, 29
+	lwzx	24, 14, 24
+	lwzx	25, 14, 25
+	lwzx	26, 14, 26
+	lwzx	27, 14, 27
+	xor	20, 20, 24
+	xor	22, 22, 25
+	xor	21, 21, 26
+	xor	23, 23, 27
+	rlwinm	24, 19, 2, 22, 29
+	rlwinm	25, 28, 2, 22, 29
+	rlwinm	26, 18, 2, 22, 29
+	rlwinm	27, 29, 2, 22, 29
+	lwzx	24, 15, 24
+	lwzx	25, 15, 25
+	lwzx	26, 15, 26
+	lwzx	27, 15, 27
+	xor	20, 20, 24
+	xor	22, 22, 25
+	xor	21, 21, 26
+	xor	23, 23, 27
+	ld	28, 0(8)
+	ld	29, 8(8)
+	rldimi	20, 22, 32, 0
+	rldimi	21, 23, 32, 0
+	addi	8, 8, 16
+	#   XOR in Key Schedule
+	xor	20, 20, 28
+	xor	21, 21, 29
+	srdi	22, 20, 32
+	srdi	23, 21, 32
+	rlwinm	28, 20, 2, 22, 29
+	rlwinm	29, 21, 2, 22, 29
+	rlwinm	19, 22, 2, 22, 29
+	rlwinm	18, 23, 2, 22, 29
+	lwzx	28, 12, 28
+	lwzx	29, 12, 29
+	lwzx	19, 12, 19
+	lwzx	18, 12, 18
+	rlwinm	24, 21, 26, 22, 29
+	rlwinm	25, 23, 26, 22, 29
+	rldimi	29, 19, 32, 24
+	rldimi	28, 18, 32, 24
+	rlwinm	26, 20, 26, 22, 29
+	rlwinm	27, 22, 26, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	rldimi	28, 24, 40, 16
+	rlwimi	28, 25, 8, 16, 23
+	rldimi	29, 26, 40, 16
+	rlwimi	29, 27, 8, 16, 23
+	rlwinm	24, 22, 18, 22, 29
+	rlwinm	25, 21, 18, 22, 29
+	rlwinm	26, 23, 18, 22, 29
+	rlwinm	27, 20, 18, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	rldimi	28, 24, 48, 8
+	rlwimi	28, 25, 16, 8, 15
+	rldimi	29, 26, 48, 8
+	rlwimi	29, 27, 16, 8, 15
+	rlwinm	24, 20, 10, 22, 29
+	rlwinm	25, 22, 10, 22, 29
+	rlwinm	26, 21, 10, 22, 29
+	rlwinm	27, 23, 10, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	ld	20, 0(8)
+	ld	21, 8(8)
+	rldimi	28, 24, 56, 0
+	rlwimi	28, 25, 24, 0, 7
+	rldimi	29, 26, 56, 0
+	rlwimi	29, 27, 24, 0, 7
+	addi	8, 8, 16
+	#   XOR in Key Schedule
+	xor	28, 28, 20
+	xor	29, 29, 21
+	std	28, 0(9)
+	std	29, 8(9)
+L_AES_XTS_encrypt_loop_block:
+	addi	8, 7, 0
+	ld	16, 0(3)
+	ld	17, 8(3)
+	ld	20, 0(8)
+	ld	21, 8(8)
+	addi	8, 8, 16
+	xor	16, 16, 28
+	xor	17, 17, 29
+	rldicr	16, 16, 32, 63
+	rldicr	17, 17, 32, 63
+	# Round: 0 - XOR in key schedule
+	xor	16, 16, 20
+	xor	17, 17, 21
+	addi	0, 10, -2
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+	ld	23, 0(11)
+	ld	23, 64(11)
+	ld	23, 128(11)
+	ld	23, 192(11)
+	ld	23, 256(11)
+	ld	23, 320(11)
+	ld	23, 384(11)
+	ld	23, 448(11)
+	ld	23, 512(11)
+	ld	23, 576(11)
+	ld	23, 640(11)
+	ld	23, 704(11)
+	ld	23, 768(11)
+	ld	23, 832(11)
+	ld	23, 896(11)
+	ld	23, 960(11)
+	ld	23, 1024(11)
+	ld	23, 1088(11)
+	ld	23, 1152(11)
+	ld	23, 1216(11)
+	ld	23, 1280(11)
+	ld	23, 1344(11)
+	ld	23, 1408(11)
+	ld	23, 1472(11)
+	ld	23, 1536(11)
+	ld	23, 1600(11)
+	ld	23, 1664(11)
+	ld	23, 1728(11)
+	ld	23, 1792(11)
+	ld	23, 1856(11)
+	ld	23, 1920(11)
+	ld	23, 1984(11)
+	ld	23, 2048(11)
+	ld	23, 2112(11)
+	ld	23, 2176(11)
+	ld	23, 2240(11)
+	ld	23, 2304(11)
+	ld	23, 2368(11)
+	ld	23, 2432(11)
+	ld	23, 2496(11)
+	ld	23, 2560(11)
+	ld	23, 2624(11)
+	ld	23, 2688(11)
+	ld	23, 2752(11)
+	ld	23, 2816(11)
+	ld	23, 2880(11)
+	ld	23, 2944(11)
+	ld	23, 3008(11)
+	ld	23, 3072(11)
+	ld	23, 3136(11)
+	ld	23, 3200(11)
+	ld	23, 3264(11)
+	ld	23, 3328(11)
+	ld	23, 3392(11)
+	ld	23, 3456(11)
+	ld	23, 3520(11)
+	ld	23, 3584(11)
+	ld	23, 3648(11)
+	ld	23, 3712(11)
+	ld	23, 3776(11)
+	ld	23, 3840(11)
+	ld	23, 3904(11)
+	ld	23, 3968(11)
+	ld	23, 4032(11)
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+	srwi	0, 0, 1
+	mtctr	0
+L_AES_XTS_encrypt_loop_nr:
+	srdi	18, 16, 32
+	srdi	19, 17, 32
+	rlwinm	20, 16, 10, 22, 29
+	rlwinm	21, 17, 10, 22, 29
+	rlwinm	22, 18, 10, 22, 29
+	rlwinm	23, 19, 10, 22, 29
+	lwzx	20, 11, 20
+	lwzx	21, 11, 21
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	rlwinm	24, 18, 18, 22, 29
+	rlwinm	25, 17, 18, 22, 29
+	rlwinm	26, 19, 18, 22, 29
+	rlwinm	27, 16, 18, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	xor	20, 20, 24
+	xor	22, 22, 25
+	xor	21, 21, 26
+	xor	23, 23, 27
+	rlwinm	24, 17, 26, 22, 29
+	rlwinm	25, 19, 26, 22, 29
+	rlwinm	26, 16, 26, 22, 29
+	rlwinm	27, 18, 26, 22, 29
+	lwzx	24, 14, 24
+	lwzx	25, 14, 25
+	lwzx	26, 14, 26
+	lwzx	27, 14, 27
+	xor	20, 20, 24
+	xor	22, 22, 25
+	xor	21, 21, 26
+	xor	23, 23, 27
+	rlwinm	24, 19, 2, 22, 29
+	rlwinm	25, 16, 2, 22, 29
+	rlwinm	26, 18, 2, 22, 29
+	rlwinm	27, 17, 2, 22, 29
+	lwzx	24, 15, 24
+	lwzx	25, 15, 25
+	lwzx	26, 15, 26
+	lwzx	27, 15, 27
+	xor	20, 20, 24
+	xor	22, 22, 25
+	xor	21, 21, 26
+	xor	23, 23, 27
+	ld	16, 0(8)
+	ld	17, 8(8)
+	rldimi	20, 22, 32, 0
+	rldimi	21, 23, 32, 0
+	addi	8, 8, 16
+	#   XOR in Key Schedule
+	xor	20, 20, 16
+	xor	21, 21, 17
+	srdi	22, 20, 32
+	srdi	23, 21, 32
+	rlwinm	16, 20, 10, 22, 29
+	rlwinm	17, 21, 10, 22, 29
+	rlwinm	18, 22, 10, 22, 29
+	rlwinm	19, 23, 10, 22, 29
+	lwzx	16, 11, 16
+	lwzx	17, 11, 17
+	lwzx	18, 11, 18
+	lwzx	19, 11, 19
+	rlwinm	24, 22, 18, 22, 29
+	rlwinm	25, 21, 18, 22, 29
+	rlwinm	26, 23, 18, 22, 29
+	rlwinm	27, 20, 18, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	xor	16, 16, 24
+	xor	18, 18, 25
+	xor	17, 17, 26
+	xor	19, 19, 27
+	rlwinm	24, 21, 26, 22, 29
+	rlwinm	25, 23, 26, 22, 29
+	rlwinm	26, 20, 26, 22, 29
+	rlwinm	27, 22, 26, 22, 29
+	lwzx	24, 14, 24
+	lwzx	25, 14, 25
+	lwzx	26, 14, 26
+	lwzx	27, 14, 27
+	xor	16, 16, 24
+	xor	18, 18, 25
+	xor	17, 17, 26
+	xor	19, 19, 27
+	rlwinm	24, 23, 2, 22, 29
+	rlwinm	25, 20, 2, 22, 29
+	rlwinm	26, 22, 2, 22, 29
+	rlwinm	27, 21, 2, 22, 29
+	lwzx	24, 15, 24
+	lwzx	25, 15, 25
+	lwzx	26, 15, 26
+	lwzx	27, 15, 27
+	xor	16, 16, 24
+	xor	18, 18, 25
+	xor	17, 17, 26
+	xor	19, 19, 27
+	ld	20, 0(8)
+	ld	21, 8(8)
+	rldimi	16, 18, 32, 0
+	rldimi	17, 19, 32, 0
+	addi	8, 8, 16
+	#   XOR in Key Schedule
+	xor	16, 16, 20
+	xor	17, 17, 21
+	bdnz	L_AES_XTS_encrypt_loop_nr
+	srdi	18, 16, 32
+	srdi	19, 17, 32
+	rlwinm	20, 16, 10, 22, 29
+	rlwinm	21, 17, 10, 22, 29
+	rlwinm	22, 18, 10, 22, 29
+	rlwinm	23, 19, 10, 22, 29
+	lwzx	20, 11, 20
+	lwzx	21, 11, 21
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	rlwinm	24, 18, 18, 22, 29
+	rlwinm	25, 17, 18, 22, 29
+	rlwinm	26, 19, 18, 22, 29
+	rlwinm	27, 16, 18, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	xor	20, 20, 24
+	xor	22, 22, 25
+	xor	21, 21, 26
+	xor	23, 23, 27
+	rlwinm	24, 17, 26, 22, 29
+	rlwinm	25, 19, 26, 22, 29
+	rlwinm	26, 16, 26, 22, 29
+	rlwinm	27, 18, 26, 22, 29
+	lwzx	24, 14, 24
+	lwzx	25, 14, 25
+	lwzx	26, 14, 26
+	lwzx	27, 14, 27
+	xor	20, 20, 24
+	xor	22, 22, 25
+	xor	21, 21, 26
+	xor	23, 23, 27
+	rlwinm	24, 19, 2, 22, 29
+	rlwinm	25, 16, 2, 22, 29
+	rlwinm	26, 18, 2, 22, 29
+	rlwinm	27, 17, 2, 22, 29
+	lwzx	24, 15, 24
+	lwzx	25, 15, 25
+	lwzx	26, 15, 26
+	lwzx	27, 15, 27
+	xor	20, 20, 24
+	xor	22, 22, 25
+	xor	21, 21, 26
+	xor	23, 23, 27
+	ld	16, 0(8)
+	ld	17, 8(8)
+	rldimi	20, 22, 32, 0
+	rldimi	21, 23, 32, 0
+	addi	8, 8, 16
+	#   XOR in Key Schedule
+	xor	20, 20, 16
+	xor	21, 21, 17
+	srdi	22, 20, 32
+	srdi	23, 21, 32
+	rlwinm	16, 20, 2, 22, 29
+	rlwinm	17, 21, 2, 22, 29
+	rlwinm	19, 22, 2, 22, 29
+	rlwinm	18, 23, 2, 22, 29
+	lwzx	16, 12, 16
+	lwzx	17, 12, 17
+	lwzx	19, 12, 19
+	lwzx	18, 12, 18
+	rlwinm	24, 21, 26, 22, 29
+	rlwinm	25, 23, 26, 22, 29
+	rldimi	17, 19, 32, 24
+	rldimi	16, 18, 32, 24
+	rlwinm	26, 20, 26, 22, 29
+	rlwinm	27, 22, 26, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	rldimi	16, 24, 40, 16
+	rlwimi	16, 25, 8, 16, 23
+	rldimi	17, 26, 40, 16
+	rlwimi	17, 27, 8, 16, 23
+	rlwinm	24, 22, 18, 22, 29
+	rlwinm	25, 21, 18, 22, 29
+	rlwinm	26, 23, 18, 22, 29
+	rlwinm	27, 20, 18, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	rldimi	16, 24, 48, 8
+	rlwimi	16, 25, 16, 8, 15
+	rldimi	17, 26, 48, 8
+	rlwimi	17, 27, 16, 8, 15
+	rlwinm	24, 20, 10, 22, 29
+	rlwinm	25, 22, 10, 22, 29
+	rlwinm	26, 21, 10, 22, 29
+	rlwinm	27, 23, 10, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	ld	20, 0(8)
+	ld	21, 8(8)
+	rldimi	16, 24, 56, 0
+	rlwimi	16, 25, 24, 0, 7
+	rldimi	17, 26, 56, 0
+	rlwimi	17, 27, 24, 0, 7
+	addi	8, 8, 16
+	#   XOR in Key Schedule
+	xor	16, 16, 20
+	xor	17, 17, 21
+	xor	16, 16, 28
+	xor	17, 17, 29
+	li	19, 8
+	ldbrx	28, 0, 9
+	ldbrx	29, 9, 19
+	std	16, 0(4)
+	std	17, 8(4)
+	sradi	24, 29, 63
+	srdi	25, 28, 63
+	sldi	29, 29, 1
+	andi.	24, 24, 0x87
+	sldi	28, 28, 1
+	xor	29, 29, 25
+	xor	28, 28, 24
+	stdbrx	28, 0, 9
+	stdbrx	29, 9, 19
+	ld	28, 0(9)
+	ld	29, 8(9)
+	addi	5, 5, -16
+	addi	3, 3, 16
+	addi	4, 4, 16
+	cmpdi	5, 16
+	bge	L_AES_XTS_encrypt_loop_block
+	cmpdi	5, 0
+	beq	L_AES_XTS_encrypt_done_data
+	addi	8, 7, 0
+	addi	4, 4, -16
+	ld	16, 0(4)
+	ld	17, 8(4)
+	addi	4, 4, 16
+	std	16, 0(9)
+	std	17, 8(9)
+	addi	20, 5, 0
+L_AES_XTS_encrypt_start_byte:
+	lbz	24, 0(9)
+	lbz	25, 0(3)
+	addi	3, 3, 1
+	stb	24, 0(4)
+	addi	4, 4, 1
+	stb	25, 0(9)
+	addi	9, 9, 1
+	addic.	20, 20, -1
+	bne	L_AES_XTS_encrypt_start_byte
+	subf	4, 5, 4
+	subf	9, 5, 9
+	addi	4, 4, -16
+	ld	16, 0(9)
+	ld	17, 8(9)
+	ld	20, 0(8)
+	ld	21, 8(8)
+	addi	8, 8, 16
+	xor	16, 16, 28
+	xor	17, 17, 29
+	rldicr	16, 16, 32, 63
+	rldicr	17, 17, 32, 63
+	# Round: 0 - XOR in key schedule
+	xor	16, 16, 20
+	xor	17, 17, 21
+	addi	0, 10, -2
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+	ld	23, 0(11)
+	ld	23, 64(11)
+	ld	23, 128(11)
+	ld	23, 192(11)
+	ld	23, 256(11)
+	ld	23, 320(11)
+	ld	23, 384(11)
+	ld	23, 448(11)
+	ld	23, 512(11)
+	ld	23, 576(11)
+	ld	23, 640(11)
+	ld	23, 704(11)
+	ld	23, 768(11)
+	ld	23, 832(11)
+	ld	23, 896(11)
+	ld	23, 960(11)
+	ld	23, 1024(11)
+	ld	23, 1088(11)
+	ld	23, 1152(11)
+	ld	23, 1216(11)
+	ld	23, 1280(11)
+	ld	23, 1344(11)
+	ld	23, 1408(11)
+	ld	23, 1472(11)
+	ld	23, 1536(11)
+	ld	23, 1600(11)
+	ld	23, 1664(11)
+	ld	23, 1728(11)
+	ld	23, 1792(11)
+	ld	23, 1856(11)
+	ld	23, 1920(11)
+	ld	23, 1984(11)
+	ld	23, 2048(11)
+	ld	23, 2112(11)
+	ld	23, 2176(11)
+	ld	23, 2240(11)
+	ld	23, 2304(11)
+	ld	23, 2368(11)
+	ld	23, 2432(11)
+	ld	23, 2496(11)
+	ld	23, 2560(11)
+	ld	23, 2624(11)
+	ld	23, 2688(11)
+	ld	23, 2752(11)
+	ld	23, 2816(11)
+	ld	23, 2880(11)
+	ld	23, 2944(11)
+	ld	23, 3008(11)
+	ld	23, 3072(11)
+	ld	23, 3136(11)
+	ld	23, 3200(11)
+	ld	23, 3264(11)
+	ld	23, 3328(11)
+	ld	23, 3392(11)
+	ld	23, 3456(11)
+	ld	23, 3520(11)
+	ld	23, 3584(11)
+	ld	23, 3648(11)
+	ld	23, 3712(11)
+	ld	23, 3776(11)
+	ld	23, 3840(11)
+	ld	23, 3904(11)
+	ld	23, 3968(11)
+	ld	23, 4032(11)
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+	srwi	0, 0, 1
+	mtctr	0
+L_AES_XTS_encrypt_loop_nr_partial:
+	srdi	18, 16, 32
+	srdi	19, 17, 32
+	rlwinm	20, 16, 10, 22, 29
+	rlwinm	21, 17, 10, 22, 29
+	rlwinm	22, 18, 10, 22, 29
+	rlwinm	23, 19, 10, 22, 29
+	lwzx	20, 11, 20
+	lwzx	21, 11, 21
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	rlwinm	24, 18, 18, 22, 29
+	rlwinm	25, 17, 18, 22, 29
+	rlwinm	26, 19, 18, 22, 29
+	rlwinm	27, 16, 18, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	xor	20, 20, 24
+	xor	22, 22, 25
+	xor	21, 21, 26
+	xor	23, 23, 27
+	rlwinm	24, 17, 26, 22, 29
+	rlwinm	25, 19, 26, 22, 29
+	rlwinm	26, 16, 26, 22, 29
+	rlwinm	27, 18, 26, 22, 29
+	lwzx	24, 14, 24
+	lwzx	25, 14, 25
+	lwzx	26, 14, 26
+	lwzx	27, 14, 27
+	xor	20, 20, 24
+	xor	22, 22, 25
+	xor	21, 21, 26
+	xor	23, 23, 27
+	rlwinm	24, 19, 2, 22, 29
+	rlwinm	25, 16, 2, 22, 29
+	rlwinm	26, 18, 2, 22, 29
+	rlwinm	27, 17, 2, 22, 29
+	lwzx	24, 15, 24
+	lwzx	25, 15, 25
+	lwzx	26, 15, 26
+	lwzx	27, 15, 27
+	xor	20, 20, 24
+	xor	22, 22, 25
+	xor	21, 21, 26
+	xor	23, 23, 27
+	ld	16, 0(8)
+	ld	17, 8(8)
+	rldimi	20, 22, 32, 0
+	rldimi	21, 23, 32, 0
+	addi	8, 8, 16
+	#   XOR in Key Schedule
+	xor	20, 20, 16
+	xor	21, 21, 17
+	srdi	22, 20, 32
+	srdi	23, 21, 32
+	rlwinm	16, 20, 10, 22, 29
+	rlwinm	17, 21, 10, 22, 29
+	rlwinm	18, 22, 10, 22, 29
+	rlwinm	19, 23, 10, 22, 29
+	lwzx	16, 11, 16
+	lwzx	17, 11, 17
+	lwzx	18, 11, 18
+	lwzx	19, 11, 19
+	rlwinm	24, 22, 18, 22, 29
+	rlwinm	25, 21, 18, 22, 29
+	rlwinm	26, 23, 18, 22, 29
+	rlwinm	27, 20, 18, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	xor	16, 16, 24
+	xor	18, 18, 25
+	xor	17, 17, 26
+	xor	19, 19, 27
+	rlwinm	24, 21, 26, 22, 29
+	rlwinm	25, 23, 26, 22, 29
+	rlwinm	26, 20, 26, 22, 29
+	rlwinm	27, 22, 26, 22, 29
+	lwzx	24, 14, 24
+	lwzx	25, 14, 25
+	lwzx	26, 14, 26
+	lwzx	27, 14, 27
+	xor	16, 16, 24
+	xor	18, 18, 25
+	xor	17, 17, 26
+	xor	19, 19, 27
+	rlwinm	24, 23, 2, 22, 29
+	rlwinm	25, 20, 2, 22, 29
+	rlwinm	26, 22, 2, 22, 29
+	rlwinm	27, 21, 2, 22, 29
+	lwzx	24, 15, 24
+	lwzx	25, 15, 25
+	lwzx	26, 15, 26
+	lwzx	27, 15, 27
+	xor	16, 16, 24
+	xor	18, 18, 25
+	xor	17, 17, 26
+	xor	19, 19, 27
+	ld	20, 0(8)
+	ld	21, 8(8)
+	rldimi	16, 18, 32, 0
+	rldimi	17, 19, 32, 0
+	addi	8, 8, 16
+	#   XOR in Key Schedule
+	xor	16, 16, 20
+	xor	17, 17, 21
+	bdnz	L_AES_XTS_encrypt_loop_nr_partial
+	srdi	18, 16, 32
+	srdi	19, 17, 32
+	rlwinm	20, 16, 10, 22, 29
+	rlwinm	21, 17, 10, 22, 29
+	rlwinm	22, 18, 10, 22, 29
+	rlwinm	23, 19, 10, 22, 29
+	lwzx	20, 11, 20
+	lwzx	21, 11, 21
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	rlwinm	24, 18, 18, 22, 29
+	rlwinm	25, 17, 18, 22, 29
+	rlwinm	26, 19, 18, 22, 29
+	rlwinm	27, 16, 18, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	xor	20, 20, 24
+	xor	22, 22, 25
+	xor	21, 21, 26
+	xor	23, 23, 27
+	rlwinm	24, 17, 26, 22, 29
+	rlwinm	25, 19, 26, 22, 29
+	rlwinm	26, 16, 26, 22, 29
+	rlwinm	27, 18, 26, 22, 29
+	lwzx	24, 14, 24
+	lwzx	25, 14, 25
+	lwzx	26, 14, 26
+	lwzx	27, 14, 27
+	xor	20, 20, 24
+	xor	22, 22, 25
+	xor	21, 21, 26
+	xor	23, 23, 27
+	rlwinm	24, 19, 2, 22, 29
+	rlwinm	25, 16, 2, 22, 29
+	rlwinm	26, 18, 2, 22, 29
+	rlwinm	27, 17, 2, 22, 29
+	lwzx	24, 15, 24
+	lwzx	25, 15, 25
+	lwzx	26, 15, 26
+	lwzx	27, 15, 27
+	xor	20, 20, 24
+	xor	22, 22, 25
+	xor	21, 21, 26
+	xor	23, 23, 27
+	ld	16, 0(8)
+	ld	17, 8(8)
+	rldimi	20, 22, 32, 0
+	rldimi	21, 23, 32, 0
+	addi	8, 8, 16
+	#   XOR in Key Schedule
+	xor	20, 20, 16
+	xor	21, 21, 17
+	srdi	22, 20, 32
+	srdi	23, 21, 32
+	rlwinm	16, 20, 2, 22, 29
+	rlwinm	17, 21, 2, 22, 29
+	rlwinm	19, 22, 2, 22, 29
+	rlwinm	18, 23, 2, 22, 29
+	lwzx	16, 12, 16
+	lwzx	17, 12, 17
+	lwzx	19, 12, 19
+	lwzx	18, 12, 18
+	rlwinm	24, 21, 26, 22, 29
+	rlwinm	25, 23, 26, 22, 29
+	rldimi	17, 19, 32, 24
+	rldimi	16, 18, 32, 24
+	rlwinm	26, 20, 26, 22, 29
+	rlwinm	27, 22, 26, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	rldimi	16, 24, 40, 16
+	rlwimi	16, 25, 8, 16, 23
+	rldimi	17, 26, 40, 16
+	rlwimi	17, 27, 8, 16, 23
+	rlwinm	24, 22, 18, 22, 29
+	rlwinm	25, 21, 18, 22, 29
+	rlwinm	26, 23, 18, 22, 29
+	rlwinm	27, 20, 18, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	rldimi	16, 24, 48, 8
+	rlwimi	16, 25, 16, 8, 15
+	rldimi	17, 26, 48, 8
+	rlwimi	17, 27, 16, 8, 15
+	rlwinm	24, 20, 10, 22, 29
+	rlwinm	25, 22, 10, 22, 29
+	rlwinm	26, 21, 10, 22, 29
+	rlwinm	27, 23, 10, 22, 29
+	lwzx	24, 12, 24
+	lwzx	25, 12, 25
+	lwzx	26, 12, 26
+	lwzx	27, 12, 27
+	ld	20, 0(8)
+	ld	21, 8(8)
+	rldimi	16, 24, 56, 0
+	rlwimi	16, 25, 24, 0, 7
+	rldimi	17, 26, 56, 0
+	rlwimi	17, 27, 24, 0, 7
+	addi	8, 8, 16
+	#   XOR in Key Schedule
+	xor	16, 16, 20
+	xor	17, 17, 21
+	xor	16, 16, 28
+	xor	17, 17, 29
+	std	16, 0(4)
+	std	17, 8(4)
+L_AES_XTS_encrypt_done_data:
+	ld	0, 128(1)
+	mtlr	0
+	ld	14, 0(1)
+	ld	15, 8(1)
+	ld	16, 16(1)
+	ld	17, 24(1)
+	ld	18, 32(1)
+	ld	19, 40(1)
+	ld	20, 48(1)
+	ld	21, 56(1)
+	ld	22, 64(1)
+	ld	23, 72(1)
+	ld	24, 80(1)
+	ld	25, 88(1)
+	ld	26, 96(1)
+	ld	27, 104(1)
+	ld	28, 112(1)
+	ld	29, 120(1)
+	addi	1, 1, 0x88
+	blr
+#ifndef __APPLE__
+	.size	AES_XTS_encrypt,.-AES_XTS_encrypt
+#endif /* __APPLE__ */
+#endif /* WOLFSSL_AES_XTS */
+#ifdef HAVE_AES_DECRYPT
+#if defined(WOLFSSL_AES_DIRECT) || defined(WOLFSSL_AES_COUNTER) || \
+    defined(HAVE_AES_CBC) || defined(HAVE_AES_ECB)
+#ifndef __APPLE__
+	.text
+	.section	.rodata
+	.type	L_AES_PPC64_td4, @object
+	.size	L_AES_PPC64_td4, 256
+#else
+	.section	__DATA,__data
+#endif /* __APPLE__ */
+#ifndef __APPLE__
+	.align	4
+#else
+	.p2align	4
+#endif /* __APPLE__ */
+L_AES_PPC64_td4:
+	.byte	0x52
+	.byte	0x9
+	.byte	0x6a
+	.byte	0xd5
+	.byte	0x30
+	.byte	0x36
+	.byte	0xa5
+	.byte	0x38
+	.byte	0xbf
+	.byte	0x40
+	.byte	0xa3
+	.byte	0x9e
+	.byte	0x81
+	.byte	0xf3
+	.byte	0xd7
+	.byte	0xfb
+	.byte	0x7c
+	.byte	0xe3
+	.byte	0x39
+	.byte	0x82
+	.byte	0x9b
+	.byte	0x2f
+	.byte	0xff
+	.byte	0x87
+	.byte	0x34
+	.byte	0x8e
+	.byte	0x43
+	.byte	0x44
+	.byte	0xc4
+	.byte	0xde
+	.byte	0xe9
+	.byte	0xcb
+	.byte	0x54
+	.byte	0x7b
+	.byte	0x94
+	.byte	0x32
+	.byte	0xa6
+	.byte	0xc2
+	.byte	0x23
+	.byte	0x3d
+	.byte	0xee
+	.byte	0x4c
+	.byte	0x95
+	.byte	0xb
+	.byte	0x42
+	.byte	0xfa
+	.byte	0xc3
+	.byte	0x4e
+	.byte	0x8
+	.byte	0x2e
+	.byte	0xa1
+	.byte	0x66
+	.byte	0x28
+	.byte	0xd9
+	.byte	0x24
+	.byte	0xb2
+	.byte	0x76
+	.byte	0x5b
+	.byte	0xa2
+	.byte	0x49
+	.byte	0x6d
+	.byte	0x8b
+	.byte	0xd1
+	.byte	0x25
+	.byte	0x72
+	.byte	0xf8
+	.byte	0xf6
+	.byte	0x64
+	.byte	0x86
+	.byte	0x68
+	.byte	0x98
+	.byte	0x16
+	.byte	0xd4
+	.byte	0xa4
+	.byte	0x5c
+	.byte	0xcc
+	.byte	0x5d
+	.byte	0x65
+	.byte	0xb6
+	.byte	0x92
+	.byte	0x6c
+	.byte	0x70
+	.byte	0x48
+	.byte	0x50
+	.byte	0xfd
+	.byte	0xed
+	.byte	0xb9
+	.byte	0xda
+	.byte	0x5e
+	.byte	0x15
+	.byte	0x46
+	.byte	0x57
+	.byte	0xa7
+	.byte	0x8d
+	.byte	0x9d
+	.byte	0x84
+	.byte	0x90
+	.byte	0xd8
+	.byte	0xab
+	.byte	0x0
+	.byte	0x8c
+	.byte	0xbc
+	.byte	0xd3
+	.byte	0xa
+	.byte	0xf7
+	.byte	0xe4
+	.byte	0x58
+	.byte	0x5
+	.byte	0xb8
+	.byte	0xb3
+	.byte	0x45
+	.byte	0x6
+	.byte	0xd0
+	.byte	0x2c
+	.byte	0x1e
+	.byte	0x8f
+	.byte	0xca
+	.byte	0x3f
+	.byte	0xf
+	.byte	0x2
+	.byte	0xc1
+	.byte	0xaf
+	.byte	0xbd
+	.byte	0x3
+	.byte	0x1
+	.byte	0x13
+	.byte	0x8a
+	.byte	0x6b
+	.byte	0x3a
+	.byte	0x91
+	.byte	0x11
+	.byte	0x41
+	.byte	0x4f
+	.byte	0x67
+	.byte	0xdc
+	.byte	0xea
+	.byte	0x97
+	.byte	0xf2
+	.byte	0xcf
+	.byte	0xce
+	.byte	0xf0
+	.byte	0xb4
+	.byte	0xe6
+	.byte	0x73
+	.byte	0x96
+	.byte	0xac
+	.byte	0x74
+	.byte	0x22
+	.byte	0xe7
+	.byte	0xad
+	.byte	0x35
+	.byte	0x85
+	.byte	0xe2
+	.byte	0xf9
+	.byte	0x37
+	.byte	0xe8
+	.byte	0x1c
+	.byte	0x75
+	.byte	0xdf
+	.byte	0x6e
+	.byte	0x47
+	.byte	0xf1
+	.byte	0x1a
+	.byte	0x71
+	.byte	0x1d
+	.byte	0x29
+	.byte	0xc5
+	.byte	0x89
+	.byte	0x6f
+	.byte	0xb7
+	.byte	0x62
+	.byte	0xe
+	.byte	0xaa
+	.byte	0x18
+	.byte	0xbe
+	.byte	0x1b
+	.byte	0xfc
+	.byte	0x56
+	.byte	0x3e
+	.byte	0x4b
+	.byte	0xc6
+	.byte	0xd2
+	.byte	0x79
+	.byte	0x20
+	.byte	0x9a
+	.byte	0xdb
+	.byte	0xc0
+	.byte	0xfe
+	.byte	0x78
+	.byte	0xcd
+	.byte	0x5a
+	.byte	0xf4
+	.byte	0x1f
+	.byte	0xdd
+	.byte	0xa8
+	.byte	0x33
+	.byte	0x88
+	.byte	0x7
+	.byte	0xc7
+	.byte	0x31
+	.byte	0xb1
+	.byte	0x12
+	.byte	0x10
+	.byte	0x59
+	.byte	0x27
+	.byte	0x80
+	.byte	0xec
+	.byte	0x5f
+	.byte	0x60
+	.byte	0x51
+	.byte	0x7f
+	.byte	0xa9
+	.byte	0x19
+	.byte	0xb5
+	.byte	0x4a
+	.byte	0xd
+	.byte	0x2d
+	.byte	0xe5
+	.byte	0x7a
+	.byte	0x9f
+	.byte	0x93
+	.byte	0xc9
+	.byte	0x9c
+	.byte	0xef
+	.byte	0xa0
+	.byte	0xe0
+	.byte	0x3b
+	.byte	0x4d
+	.byte	0xae
+	.byte	0x2a
+	.byte	0xf5
+	.byte	0xb0
+	.byte	0xc8
+	.byte	0xeb
+	.byte	0xbb
+	.byte	0x3c
+	.byte	0x83
+	.byte	0x53
+	.byte	0x99
+	.byte	0x61
+	.byte	0x17
+	.byte	0x2b
+	.byte	0x4
+	.byte	0x7e
+	.byte	0xba
+	.byte	0x77
+	.byte	0xd6
+	.byte	0x26
+	.byte	0xe1
+	.byte	0x69
+	.byte	0x14
+	.byte	0x63
+	.byte	0x55
+	.byte	0x21
+	.byte	0xc
+	.byte	0x7d
+#if defined(WOLFSSL_AES_DIRECT) || defined(WOLFSSL_AES_COUNTER) || defined(HAVE_AES_ECB)
+#ifndef __APPLE__
+.text
+.globl	AES_ECB_decrypt
+.type	AES_ECB_decrypt,@function
+.align	16
+AES_ECB_decrypt:
+#else
+.section	__TEXT,__text
+.globl	_AES_ECB_decrypt
+.p2align	4
+_AES_ECB_decrypt:
+#endif /* __APPLE__ */
+	stdu	1, -88(1)
+	mflr	0
+	std	0, 80(1)
+	std	14, 0(1)
+	std	15, 8(1)
+	std	16, 16(1)
+	std	17, 24(1)
+	std	18, 32(1)
+	std	19, 40(1)
+	std	20, 48(1)
+	std	21, 56(1)
+	std	22, 64(1)
+	std	23, 72(1)
+#if defined(__ARCH_PWR8) || (defined(_ARCH_PWR8))
+	addis	8, 2, L_AES_PPC64_td@toc@ha
+	addi	8, 8, L_AES_PPC64_td@toc@l
+#else
+	lis	8, L_AES_PPC64_td@ha
+	la	8, L_AES_PPC64_td@l(8)
+#endif /* defined(__ARCH_PWR8) || (defined(_ARCH_PWR8)) */
+#if defined(__ARCH_PWR8) || (defined(_ARCH_PWR8))
+	addis	9, 2, L_AES_PPC64_td4@toc@ha
+	addi	9, 9, L_AES_PPC64_td4@toc@l
+#else
+	lis	9, L_AES_PPC64_td4@ha
+	la	9, L_AES_PPC64_td4@l(9)
+#endif /* defined(__ARCH_PWR8) || (defined(_ARCH_PWR8)) */
+L_AES_ECB_decrypt_loop_block:
+	addi	23, 6, 0
+	ld	10, 0(3)
+	ld	11, 8(3)
+	ld	15, 0(23)
+	ld	16, 8(23)
+	addi	23, 23, 16
+	rldicr	10, 10, 32, 63
+	rldicr	11, 11, 32, 63
+	# Round: 0 - XOR in key schedule
+	xor	10, 10, 15
+	xor	11, 11, 16
+	addi	0, 7, -2
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+	ld	18, 0(8)
+	ld	18, 64(8)
+	ld	18, 128(8)
+	ld	18, 192(8)
+	ld	18, 256(8)
+	ld	18, 320(8)
+	ld	18, 384(8)
+	ld	18, 448(8)
+	ld	18, 512(8)
+	ld	18, 576(8)
+	ld	18, 640(8)
+	ld	18, 704(8)
+	ld	18, 768(8)
+	ld	18, 832(8)
+	ld	18, 896(8)
+	ld	18, 960(8)
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+	srwi	0, 0, 1
+	mtctr	0
+L_AES_ECB_decrypt_loop_nr:
+	srdi	12, 10, 32
+	srdi	14, 11, 32
+	rlwinm	15, 14, 18, 22, 29
+	rlwinm	17, 10, 18, 22, 29
+	rlwinm	16, 12, 18, 22, 29
+	rlwinm	18, 11, 18, 22, 29
+	lwzx	15, 8, 15
+	lwzx	17, 8, 17
+	lwzx	16, 8, 16
+	lwzx	18, 8, 18
+	rlwinm	19, 10, 10, 22, 29
+	rlwinm	20, 12, 10, 22, 29
+	rlwinm	21, 11, 10, 22, 29
+	rlwinm	22, 14, 10, 22, 29
+	lwzx	19, 8, 19
+	lwzx	20, 8, 20
+	lwzx	21, 8, 21
+	lwzx	22, 8, 22
+	rlwimi	19, 19, 8, 0, 31
+	rlwimi	20, 20, 8, 0, 31
+	rlwimi	21, 21, 8, 0, 31
+	rlwimi	22, 22, 8, 0, 31
+	xor	15, 15, 19
+	xor	17, 17, 20
+	xor	16, 16, 21
+	xor	18, 18, 22
+	rlwinm	19, 11, 26, 22, 29
+	rlwinm	20, 14, 26, 22, 29
+	rlwinm	21, 10, 26, 22, 29
+	rlwinm	22, 12, 26, 22, 29
+	lwzx	19, 8, 19
+	lwzx	20, 8, 20
+	lwzx	21, 8, 21
+	lwzx	22, 8, 22
+	rlwimi	19, 19, 24, 0, 31
+	rlwimi	20, 20, 24, 0, 31
+	rlwimi	21, 21, 24, 0, 31
+	rlwimi	22, 22, 24, 0, 31
+	xor	15, 15, 19
+	xor	17, 17, 20
+	xor	16, 16, 21
+	xor	18, 18, 22
+	rlwinm	19, 12, 2, 22, 29
+	rlwinm	20, 11, 2, 22, 29
+	rlwinm	21, 14, 2, 22, 29
+	rlwinm	22, 10, 2, 22, 29
+	lwzx	19, 8, 19
+	lwzx	20, 8, 20
+	lwzx	21, 8, 21
+	lwzx	22, 8, 22
+	rlwimi	19, 19, 16, 0, 31
+	rlwimi	20, 20, 16, 0, 31
+	rlwimi	21, 21, 16, 0, 31
+	rlwimi	22, 22, 16, 0, 31
+	ld	10, 0(23)
+	ld	11, 8(23)
+	xor	15, 15, 19
+	xor	17, 17, 20
+	xor	16, 16, 21
+	xor	18, 18, 22
+	addi	23, 23, 16
+	rldimi	15, 17, 32, 0
+	rldimi	16, 18, 32, 0
+	#   XOR in Key Schedule
+	xor	15, 15, 10
+	xor	16, 16, 11
+	srdi	17, 15, 32
+	srdi	18, 16, 32
+	rlwinm	10, 18, 18, 22, 29
+	rlwinm	12, 15, 18, 22, 29
+	rlwinm	11, 17, 18, 22, 29
+	rlwinm	14, 16, 18, 22, 29
+	lwzx	10, 8, 10
+	lwzx	12, 8, 12
+	lwzx	11, 8, 11
+	lwzx	14, 8, 14
+	rlwinm	19, 15, 10, 22, 29
+	rlwinm	20, 17, 10, 22, 29
+	rlwinm	21, 16, 10, 22, 29
+	rlwinm	22, 18, 10, 22, 29
+	lwzx	19, 8, 19
+	lwzx	20, 8, 20
+	lwzx	21, 8, 21
+	lwzx	22, 8, 22
+	rlwimi	19, 19, 8, 0, 31
+	rlwimi	20, 20, 8, 0, 31
+	rlwimi	21, 21, 8, 0, 31
+	rlwimi	22, 22, 8, 0, 31
+	xor	10, 10, 19
+	xor	12, 12, 20
+	xor	11, 11, 21
+	xor	14, 14, 22
+	rlwinm	19, 16, 26, 22, 29
+	rlwinm	20, 18, 26, 22, 29
+	rlwinm	21, 15, 26, 22, 29
+	rlwinm	22, 17, 26, 22, 29
+	lwzx	19, 8, 19
+	lwzx	20, 8, 20
+	lwzx	21, 8, 21
+	lwzx	22, 8, 22
+	rlwimi	19, 19, 24, 0, 31
+	rlwimi	20, 20, 24, 0, 31
+	rlwimi	21, 21, 24, 0, 31
+	rlwimi	22, 22, 24, 0, 31
+	xor	10, 10, 19
+	xor	12, 12, 20
+	xor	11, 11, 21
+	xor	14, 14, 22
+	rlwinm	19, 17, 2, 22, 29
+	rlwinm	20, 16, 2, 22, 29
+	rlwinm	21, 18, 2, 22, 29
+	rlwinm	22, 15, 2, 22, 29
+	lwzx	19, 8, 19
+	lwzx	20, 8, 20
+	lwzx	21, 8, 21
+	lwzx	22, 8, 22
+	rlwimi	19, 19, 16, 0, 31
+	rlwimi	20, 20, 16, 0, 31
+	rlwimi	21, 21, 16, 0, 31
+	rlwimi	22, 22, 16, 0, 31
+	ld	15, 0(23)
+	ld	16, 8(23)
+	xor	10, 10, 19
+	xor	12, 12, 20
+	xor	11, 11, 21
+	xor	14, 14, 22
+	addi	23, 23, 16
+	rldimi	10, 12, 32, 0
+	rldimi	11, 14, 32, 0
+	#   XOR in Key Schedule
+	xor	10, 10, 15
+	xor	11, 11, 16
+	bdnz	L_AES_ECB_decrypt_loop_nr
+	srdi	12, 10, 32
+	srdi	14, 11, 32
+	rlwinm	15, 14, 18, 22, 29
+	rlwinm	17, 10, 18, 22, 29
+	rlwinm	16, 12, 18, 22, 29
+	rlwinm	18, 11, 18, 22, 29
+	lwzx	15, 8, 15
+	lwzx	17, 8, 17
+	lwzx	16, 8, 16
+	lwzx	18, 8, 18
+	rlwinm	19, 10, 10, 22, 29
+	rlwinm	20, 12, 10, 22, 29
+	rlwinm	21, 11, 10, 22, 29
+	rlwinm	22, 14, 10, 22, 29
+	lwzx	19, 8, 19
+	lwzx	20, 8, 20
+	lwzx	21, 8, 21
+	lwzx	22, 8, 22
+	rlwimi	19, 19, 8, 0, 31
+	rlwimi	20, 20, 8, 0, 31
+	rlwimi	21, 21, 8, 0, 31
+	rlwimi	22, 22, 8, 0, 31
+	xor	15, 15, 19
+	xor	17, 17, 20
+	xor	16, 16, 21
+	xor	18, 18, 22
+	rlwinm	19, 11, 26, 22, 29
+	rlwinm	20, 14, 26, 22, 29
+	rlwinm	21, 10, 26, 22, 29
+	rlwinm	22, 12, 26, 22, 29
+	lwzx	19, 8, 19
+	lwzx	20, 8, 20
+	lwzx	21, 8, 21
+	lwzx	22, 8, 22
+	rlwimi	19, 19, 24, 0, 31
+	rlwimi	20, 20, 24, 0, 31
+	rlwimi	21, 21, 24, 0, 31
+	rlwimi	22, 22, 24, 0, 31
+	xor	15, 15, 19
+	xor	17, 17, 20
+	xor	16, 16, 21
+	xor	18, 18, 22
+	rlwinm	19, 12, 2, 22, 29
+	rlwinm	20, 11, 2, 22, 29
+	rlwinm	21, 14, 2, 22, 29
+	rlwinm	22, 10, 2, 22, 29
+	lwzx	19, 8, 19
+	lwzx	20, 8, 20
+	lwzx	21, 8, 21
+	lwzx	22, 8, 22
+	rlwimi	19, 19, 16, 0, 31
+	rlwimi	20, 20, 16, 0, 31
+	rlwimi	21, 21, 16, 0, 31
+	rlwimi	22, 22, 16, 0, 31
+	ld	10, 0(23)
+	ld	11, 8(23)
+	xor	15, 15, 19
+	xor	17, 17, 20
+	xor	16, 16, 21
+	xor	18, 18, 22
+	addi	23, 23, 16
+	rldimi	15, 17, 32, 0
+	rldimi	16, 18, 32, 0
+	#   XOR in Key Schedule
+	xor	15, 15, 10
+	xor	16, 16, 11
+	ld	18, 0(9)
+	ld	18, 64(9)
+	ld	18, 128(9)
+	ld	18, 192(9)
+	ld	18, 256(9)
+	ld	18, 320(9)
+	ld	18, 384(9)
+	ld	18, 448(9)
+	ld	18, 512(9)
+	ld	18, 576(9)
+	ld	18, 640(9)
+	ld	18, 704(9)
+	ld	18, 768(9)
+	ld	18, 832(9)
+	ld	18, 896(9)
+	ld	18, 960(9)
+	rldicl	12, 15, 32, 56
+	andi.   10, 16, 255
+	rldicl	14, 16, 32, 56
+	andi.   11, 15, 255
+	lbzx	12, 9, 12
+	lbzx	10, 9, 10
+	lbzx	14, 9, 14
+	lbzx	11, 9, 11
+	rldimi	10, 12, 32, 24
+	rldimi	11, 14, 32, 24
+	rldicl	19, 16, 56, 56
+	rldicl	20, 16, 24, 56
+	rldicl	21, 15, 56, 56
+	rldicl	22, 15, 24, 56
+	lbzx	19, 9, 19
+	lbzx	20, 9, 20
+	lbzx	21, 9, 21
+	lbzx	22, 9, 22
+	rldimi	10, 19, 40, 16
+	rlwimi	10, 20, 8, 16, 23
+	rldimi	11, 21, 40, 16
+	rlwimi	11, 22, 8, 16, 23
+	rldicl	19, 16, 16, 56
+	rldicl	20, 15, 48, 56
+	rldicl	21, 15, 16, 56
+	rldicl	22, 16, 48, 56
+	lbzx	19, 9, 19
+	lbzx	20, 9, 20
+	lbzx	21, 9, 21
+	lbzx	22, 9, 22
+	rldimi	10, 19, 48, 8
+	rlwimi	10, 20, 16, 8, 15
+	rldimi	11, 21, 48, 8
+	rlwimi	11, 22, 16, 8, 15
+	rldicl	19, 15, 40, 56
+	rldicl	20, 15, 8, 56
+	rldicl	21, 16, 40, 56
+	rldicl	22, 16, 8, 56
+	lbzx	19, 9, 19
+	lbzx	20, 9, 20
+	lbzx	21, 9, 21
+	lbzx	22, 9, 22
+	ld	15, 0(23)
+	ld	16, 8(23)
+	rldimi	10, 19, 56, 0
+	rlwimi	10, 20, 24, 0, 7
+	rldimi	11, 21, 56, 0
+	rlwimi	11, 22, 24, 0, 7
+	addi	23, 23, 16
+	#   XOR in Key Schedule
+	xor	10, 10, 15
+	xor	11, 11, 16
+	std	10, 0(4)
+	std	11, 8(4)
+	addic.	5, 5, -16
+	addi	3, 3, 16
+	addi	4, 4, 16
+	bne	L_AES_ECB_decrypt_loop_block
+	ld	0, 80(1)
+	mtlr	0
+	ld	14, 0(1)
+	ld	15, 8(1)
+	ld	16, 16(1)
+	ld	17, 24(1)
+	ld	18, 32(1)
+	ld	19, 40(1)
+	ld	20, 48(1)
+	ld	21, 56(1)
+	ld	22, 64(1)
+	ld	23, 72(1)
+	addi	1, 1, 0x58
+	blr
+#ifndef __APPLE__
+	.size	AES_ECB_decrypt,.-AES_ECB_decrypt
+#endif /* __APPLE__ */
+#endif /* WOLFSSL_AES_DIRECT || WOLFSSL_AES_COUNTER || defined(HAVE_AES_ECB) */
+#ifdef HAVE_AES_CBC
+#ifndef __APPLE__
+.text
+.globl	AES_CBC_decrypt
+.type	AES_CBC_decrypt,@function
+.align	16
+AES_CBC_decrypt:
+#else
+.section	__TEXT,__text
+.globl	_AES_CBC_decrypt
+.p2align	4
+_AES_CBC_decrypt:
+#endif /* __APPLE__ */
+	stdu	1, -128(1)
+	mflr	0
+	std	0, 120(1)
+	std	14, 0(1)
+	std	15, 8(1)
+	std	16, 16(1)
+	std	17, 24(1)
+	std	18, 32(1)
+	std	19, 40(1)
+	std	20, 48(1)
+	std	21, 56(1)
+	std	22, 64(1)
+	std	23, 72(1)
+	std	24, 80(1)
+	std	25, 88(1)
+	std	26, 96(1)
+	std	27, 104(1)
+	std	28, 112(1)
+#if defined(__ARCH_PWR8) || (defined(_ARCH_PWR8))
+	addis	10, 2, L_AES_PPC64_td4@toc@ha
+	addi	10, 10, L_AES_PPC64_td4@toc@l
+#else
+	lis	10, L_AES_PPC64_td4@ha
+	la	10, L_AES_PPC64_td4@l(10)
+#endif /* defined(__ARCH_PWR8) || (defined(_ARCH_PWR8)) */
+#if defined(__ARCH_PWR8) || (defined(_ARCH_PWR8))
+	addis	9, 2, L_AES_PPC64_td@toc@ha
+	addi	9, 9, L_AES_PPC64_td@toc@l
+#else
+	lis	9, L_AES_PPC64_td@ha
+	la	9, L_AES_PPC64_td@l(9)
+#endif /* defined(__ARCH_PWR8) || (defined(_ARCH_PWR8)) */
+	ld	24, 0(8)
+	ld	25, 8(8)
+L_AES_CBC_decrypt_loop_block:
+	addi	28, 6, 0
+	ld	26, 0(3)
+	ld	27, 8(3)
+	ld	16, 0(28)
+	ld	17, 8(28)
+	addi	28, 28, 16
+	rldicr	11, 26, 32, 63
+	rldicr	12, 27, 32, 63
+	# Round: 0 - XOR in key schedule
+	xor	11, 11, 16
+	xor	12, 12, 17
+	addi	0, 7, -2
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+	ld	19, 0(9)
+	ld	19, 64(9)
+	ld	19, 128(9)
+	ld	19, 192(9)
+	ld	19, 256(9)
+	ld	19, 320(9)
+	ld	19, 384(9)
+	ld	19, 448(9)
+	ld	19, 512(9)
+	ld	19, 576(9)
+	ld	19, 640(9)
+	ld	19, 704(9)
+	ld	19, 768(9)
+	ld	19, 832(9)
+	ld	19, 896(9)
+	ld	19, 960(9)
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+	srwi	0, 0, 1
+	mtctr	0
+L_AES_CBC_decrypt_loop_nr_even:
+	srdi	14, 11, 32
+	srdi	15, 12, 32
+	rlwinm	16, 15, 18, 22, 29
+	rlwinm	18, 11, 18, 22, 29
+	rlwinm	17, 14, 18, 22, 29
+	rlwinm	19, 12, 18, 22, 29
+	lwzx	16, 9, 16
+	lwzx	18, 9, 18
+	lwzx	17, 9, 17
+	lwzx	19, 9, 19
+	rlwinm	20, 11, 10, 22, 29
+	rlwinm	21, 14, 10, 22, 29
+	rlwinm	22, 12, 10, 22, 29
+	rlwinm	23, 15, 10, 22, 29
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	rlwimi	20, 20, 8, 0, 31
+	rlwimi	21, 21, 8, 0, 31
+	rlwimi	22, 22, 8, 0, 31
+	rlwimi	23, 23, 8, 0, 31
+	xor	16, 16, 20
+	xor	18, 18, 21
+	xor	17, 17, 22
+	xor	19, 19, 23
+	rlwinm	20, 12, 26, 22, 29
+	rlwinm	21, 15, 26, 22, 29
+	rlwinm	22, 11, 26, 22, 29
+	rlwinm	23, 14, 26, 22, 29
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	rlwimi	20, 20, 24, 0, 31
+	rlwimi	21, 21, 24, 0, 31
+	rlwimi	22, 22, 24, 0, 31
+	rlwimi	23, 23, 24, 0, 31
+	xor	16, 16, 20
+	xor	18, 18, 21
+	xor	17, 17, 22
+	xor	19, 19, 23
+	rlwinm	20, 14, 2, 22, 29
+	rlwinm	21, 12, 2, 22, 29
+	rlwinm	22, 15, 2, 22, 29
+	rlwinm	23, 11, 2, 22, 29
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	rlwimi	20, 20, 16, 0, 31
+	rlwimi	21, 21, 16, 0, 31
+	rlwimi	22, 22, 16, 0, 31
+	rlwimi	23, 23, 16, 0, 31
+	ld	11, 0(28)
+	ld	12, 8(28)
+	xor	16, 16, 20
+	xor	18, 18, 21
+	xor	17, 17, 22
+	xor	19, 19, 23
+	addi	28, 28, 16
+	rldimi	16, 18, 32, 0
+	rldimi	17, 19, 32, 0
+	#   XOR in Key Schedule
+	xor	16, 16, 11
+	xor	17, 17, 12
+	srdi	18, 16, 32
+	srdi	19, 17, 32
+	rlwinm	11, 19, 18, 22, 29
+	rlwinm	14, 16, 18, 22, 29
+	rlwinm	12, 18, 18, 22, 29
+	rlwinm	15, 17, 18, 22, 29
+	lwzx	11, 9, 11
+	lwzx	14, 9, 14
+	lwzx	12, 9, 12
+	lwzx	15, 9, 15
+	rlwinm	20, 16, 10, 22, 29
+	rlwinm	21, 18, 10, 22, 29
+	rlwinm	22, 17, 10, 22, 29
+	rlwinm	23, 19, 10, 22, 29
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	rlwimi	20, 20, 8, 0, 31
+	rlwimi	21, 21, 8, 0, 31
+	rlwimi	22, 22, 8, 0, 31
+	rlwimi	23, 23, 8, 0, 31
+	xor	11, 11, 20
+	xor	14, 14, 21
+	xor	12, 12, 22
+	xor	15, 15, 23
+	rlwinm	20, 17, 26, 22, 29
+	rlwinm	21, 19, 26, 22, 29
+	rlwinm	22, 16, 26, 22, 29
+	rlwinm	23, 18, 26, 22, 29
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	rlwimi	20, 20, 24, 0, 31
+	rlwimi	21, 21, 24, 0, 31
+	rlwimi	22, 22, 24, 0, 31
+	rlwimi	23, 23, 24, 0, 31
+	xor	11, 11, 20
+	xor	14, 14, 21
+	xor	12, 12, 22
+	xor	15, 15, 23
+	rlwinm	20, 18, 2, 22, 29
+	rlwinm	21, 17, 2, 22, 29
+	rlwinm	22, 19, 2, 22, 29
+	rlwinm	23, 16, 2, 22, 29
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	rlwimi	20, 20, 16, 0, 31
+	rlwimi	21, 21, 16, 0, 31
+	rlwimi	22, 22, 16, 0, 31
+	rlwimi	23, 23, 16, 0, 31
+	ld	16, 0(28)
+	ld	17, 8(28)
+	xor	11, 11, 20
+	xor	14, 14, 21
+	xor	12, 12, 22
+	xor	15, 15, 23
+	addi	28, 28, 16
+	rldimi	11, 14, 32, 0
+	rldimi	12, 15, 32, 0
+	#   XOR in Key Schedule
+	xor	11, 11, 16
+	xor	12, 12, 17
+	bdnz	L_AES_CBC_decrypt_loop_nr_even
+	srdi	14, 11, 32
+	srdi	15, 12, 32
+	rlwinm	16, 15, 18, 22, 29
+	rlwinm	18, 11, 18, 22, 29
+	rlwinm	17, 14, 18, 22, 29
+	rlwinm	19, 12, 18, 22, 29
+	lwzx	16, 9, 16
+	lwzx	18, 9, 18
+	lwzx	17, 9, 17
+	lwzx	19, 9, 19
+	rlwinm	20, 11, 10, 22, 29
+	rlwinm	21, 14, 10, 22, 29
+	rlwinm	22, 12, 10, 22, 29
+	rlwinm	23, 15, 10, 22, 29
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	rlwimi	20, 20, 8, 0, 31
+	rlwimi	21, 21, 8, 0, 31
+	rlwimi	22, 22, 8, 0, 31
+	rlwimi	23, 23, 8, 0, 31
+	xor	16, 16, 20
+	xor	18, 18, 21
+	xor	17, 17, 22
+	xor	19, 19, 23
+	rlwinm	20, 12, 26, 22, 29
+	rlwinm	21, 15, 26, 22, 29
+	rlwinm	22, 11, 26, 22, 29
+	rlwinm	23, 14, 26, 22, 29
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	rlwimi	20, 20, 24, 0, 31
+	rlwimi	21, 21, 24, 0, 31
+	rlwimi	22, 22, 24, 0, 31
+	rlwimi	23, 23, 24, 0, 31
+	xor	16, 16, 20
+	xor	18, 18, 21
+	xor	17, 17, 22
+	xor	19, 19, 23
+	rlwinm	20, 14, 2, 22, 29
+	rlwinm	21, 12, 2, 22, 29
+	rlwinm	22, 15, 2, 22, 29
+	rlwinm	23, 11, 2, 22, 29
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	rlwimi	20, 20, 16, 0, 31
+	rlwimi	21, 21, 16, 0, 31
+	rlwimi	22, 22, 16, 0, 31
+	rlwimi	23, 23, 16, 0, 31
+	ld	11, 0(28)
+	ld	12, 8(28)
+	xor	16, 16, 20
+	xor	18, 18, 21
+	xor	17, 17, 22
+	xor	19, 19, 23
+	addi	28, 28, 16
+	rldimi	16, 18, 32, 0
+	rldimi	17, 19, 32, 0
+	#   XOR in Key Schedule
+	xor	16, 16, 11
+	xor	17, 17, 12
+	ld	19, 0(10)
+	ld	19, 64(10)
+	ld	19, 128(10)
+	ld	19, 192(10)
+	ld	19, 256(10)
+	ld	19, 320(10)
+	ld	19, 384(10)
+	ld	19, 448(10)
+	ld	19, 512(10)
+	ld	19, 576(10)
+	ld	19, 640(10)
+	ld	19, 704(10)
+	ld	19, 768(10)
+	ld	19, 832(10)
+	ld	19, 896(10)
+	ld	19, 960(10)
+	rldicl	14, 16, 32, 56
+	andi.   11, 17, 255
+	rldicl	15, 17, 32, 56
+	andi.   12, 16, 255
+	lbzx	14, 10, 14
+	lbzx	11, 10, 11
+	lbzx	15, 10, 15
+	lbzx	12, 10, 12
+	rldimi	11, 14, 32, 24
+	rldimi	12, 15, 32, 24
+	rldicl	20, 17, 56, 56
+	rldicl	21, 17, 24, 56
+	rldicl	22, 16, 56, 56
+	rldicl	23, 16, 24, 56
+	lbzx	20, 10, 20
+	lbzx	21, 10, 21
+	lbzx	22, 10, 22
+	lbzx	23, 10, 23
+	rldimi	11, 20, 40, 16
+	rlwimi	11, 21, 8, 16, 23
+	rldimi	12, 22, 40, 16
+	rlwimi	12, 23, 8, 16, 23
+	rldicl	20, 17, 16, 56
+	rldicl	21, 16, 48, 56
+	rldicl	22, 16, 16, 56
+	rldicl	23, 17, 48, 56
+	lbzx	20, 10, 20
+	lbzx	21, 10, 21
+	lbzx	22, 10, 22
+	lbzx	23, 10, 23
+	rldimi	11, 20, 48, 8
+	rlwimi	11, 21, 16, 8, 15
+	rldimi	12, 22, 48, 8
+	rlwimi	12, 23, 16, 8, 15
+	rldicl	20, 16, 40, 56
+	rldicl	21, 16, 8, 56
+	rldicl	22, 17, 40, 56
+	rldicl	23, 17, 8, 56
+	lbzx	20, 10, 20
+	lbzx	21, 10, 21
+	lbzx	22, 10, 22
+	lbzx	23, 10, 23
+	ld	16, 0(28)
+	ld	17, 8(28)
+	rldimi	11, 20, 56, 0
+	rlwimi	11, 21, 24, 0, 7
+	rldimi	12, 22, 56, 0
+	rlwimi	12, 23, 24, 0, 7
+	addi	28, 28, 16
+	#   XOR in Key Schedule
+	xor	11, 11, 16
+	xor	12, 12, 17
+	xor	11, 11, 24
+	xor	12, 12, 25
+	std	11, 0(4)
+	std	12, 8(4)
+	addic.	5, 5, -16
+	addi	3, 3, 16
+	addi	4, 4, 16
+	beq	L_AES_CBC_decrypt_end_dec_odd
+	addi	28, 6, 0
+	ld	24, 0(3)
+	ld	25, 8(3)
+	ld	16, 0(28)
+	ld	17, 8(28)
+	addi	28, 28, 16
+	rldicr	11, 24, 32, 63
+	rldicr	12, 25, 32, 63
+	# Round: 0 - XOR in key schedule
+	xor	11, 11, 16
+	xor	12, 12, 17
+	addi	0, 7, -2
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+	ld	19, 0(9)
+	ld	19, 64(9)
+	ld	19, 128(9)
+	ld	19, 192(9)
+	ld	19, 256(9)
+	ld	19, 320(9)
+	ld	19, 384(9)
+	ld	19, 448(9)
+	ld	19, 512(9)
+	ld	19, 576(9)
+	ld	19, 640(9)
+	ld	19, 704(9)
+	ld	19, 768(9)
+	ld	19, 832(9)
+	ld	19, 896(9)
+	ld	19, 960(9)
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+	srwi	0, 0, 1
+	mtctr	0
+L_AES_CBC_decrypt_loop_nr_odd:
+	srdi	14, 11, 32
+	srdi	15, 12, 32
+	rlwinm	16, 15, 18, 22, 29
+	rlwinm	18, 11, 18, 22, 29
+	rlwinm	17, 14, 18, 22, 29
+	rlwinm	19, 12, 18, 22, 29
+	lwzx	16, 9, 16
+	lwzx	18, 9, 18
+	lwzx	17, 9, 17
+	lwzx	19, 9, 19
+	rlwinm	20, 11, 10, 22, 29
+	rlwinm	21, 14, 10, 22, 29
+	rlwinm	22, 12, 10, 22, 29
+	rlwinm	23, 15, 10, 22, 29
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	rlwimi	20, 20, 8, 0, 31
+	rlwimi	21, 21, 8, 0, 31
+	rlwimi	22, 22, 8, 0, 31
+	rlwimi	23, 23, 8, 0, 31
+	xor	16, 16, 20
+	xor	18, 18, 21
+	xor	17, 17, 22
+	xor	19, 19, 23
+	rlwinm	20, 12, 26, 22, 29
+	rlwinm	21, 15, 26, 22, 29
+	rlwinm	22, 11, 26, 22, 29
+	rlwinm	23, 14, 26, 22, 29
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	rlwimi	20, 20, 24, 0, 31
+	rlwimi	21, 21, 24, 0, 31
+	rlwimi	22, 22, 24, 0, 31
+	rlwimi	23, 23, 24, 0, 31
+	xor	16, 16, 20
+	xor	18, 18, 21
+	xor	17, 17, 22
+	xor	19, 19, 23
+	rlwinm	20, 14, 2, 22, 29
+	rlwinm	21, 12, 2, 22, 29
+	rlwinm	22, 15, 2, 22, 29
+	rlwinm	23, 11, 2, 22, 29
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	rlwimi	20, 20, 16, 0, 31
+	rlwimi	21, 21, 16, 0, 31
+	rlwimi	22, 22, 16, 0, 31
+	rlwimi	23, 23, 16, 0, 31
+	ld	11, 0(28)
+	ld	12, 8(28)
+	xor	16, 16, 20
+	xor	18, 18, 21
+	xor	17, 17, 22
+	xor	19, 19, 23
+	addi	28, 28, 16
+	rldimi	16, 18, 32, 0
+	rldimi	17, 19, 32, 0
+	#   XOR in Key Schedule
+	xor	16, 16, 11
+	xor	17, 17, 12
+	srdi	18, 16, 32
+	srdi	19, 17, 32
+	rlwinm	11, 19, 18, 22, 29
+	rlwinm	14, 16, 18, 22, 29
+	rlwinm	12, 18, 18, 22, 29
+	rlwinm	15, 17, 18, 22, 29
+	lwzx	11, 9, 11
+	lwzx	14, 9, 14
+	lwzx	12, 9, 12
+	lwzx	15, 9, 15
+	rlwinm	20, 16, 10, 22, 29
+	rlwinm	21, 18, 10, 22, 29
+	rlwinm	22, 17, 10, 22, 29
+	rlwinm	23, 19, 10, 22, 29
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	rlwimi	20, 20, 8, 0, 31
+	rlwimi	21, 21, 8, 0, 31
+	rlwimi	22, 22, 8, 0, 31
+	rlwimi	23, 23, 8, 0, 31
+	xor	11, 11, 20
+	xor	14, 14, 21
+	xor	12, 12, 22
+	xor	15, 15, 23
+	rlwinm	20, 17, 26, 22, 29
+	rlwinm	21, 19, 26, 22, 29
+	rlwinm	22, 16, 26, 22, 29
+	rlwinm	23, 18, 26, 22, 29
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	rlwimi	20, 20, 24, 0, 31
+	rlwimi	21, 21, 24, 0, 31
+	rlwimi	22, 22, 24, 0, 31
+	rlwimi	23, 23, 24, 0, 31
+	xor	11, 11, 20
+	xor	14, 14, 21
+	xor	12, 12, 22
+	xor	15, 15, 23
+	rlwinm	20, 18, 2, 22, 29
+	rlwinm	21, 17, 2, 22, 29
+	rlwinm	22, 19, 2, 22, 29
+	rlwinm	23, 16, 2, 22, 29
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	rlwimi	20, 20, 16, 0, 31
+	rlwimi	21, 21, 16, 0, 31
+	rlwimi	22, 22, 16, 0, 31
+	rlwimi	23, 23, 16, 0, 31
+	ld	16, 0(28)
+	ld	17, 8(28)
+	xor	11, 11, 20
+	xor	14, 14, 21
+	xor	12, 12, 22
+	xor	15, 15, 23
+	addi	28, 28, 16
+	rldimi	11, 14, 32, 0
+	rldimi	12, 15, 32, 0
+	#   XOR in Key Schedule
+	xor	11, 11, 16
+	xor	12, 12, 17
+	bdnz	L_AES_CBC_decrypt_loop_nr_odd
+	srdi	14, 11, 32
+	srdi	15, 12, 32
+	rlwinm	16, 15, 18, 22, 29
+	rlwinm	18, 11, 18, 22, 29
+	rlwinm	17, 14, 18, 22, 29
+	rlwinm	19, 12, 18, 22, 29
+	lwzx	16, 9, 16
+	lwzx	18, 9, 18
+	lwzx	17, 9, 17
+	lwzx	19, 9, 19
+	rlwinm	20, 11, 10, 22, 29
+	rlwinm	21, 14, 10, 22, 29
+	rlwinm	22, 12, 10, 22, 29
+	rlwinm	23, 15, 10, 22, 29
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	rlwimi	20, 20, 8, 0, 31
+	rlwimi	21, 21, 8, 0, 31
+	rlwimi	22, 22, 8, 0, 31
+	rlwimi	23, 23, 8, 0, 31
+	xor	16, 16, 20
+	xor	18, 18, 21
+	xor	17, 17, 22
+	xor	19, 19, 23
+	rlwinm	20, 12, 26, 22, 29
+	rlwinm	21, 15, 26, 22, 29
+	rlwinm	22, 11, 26, 22, 29
+	rlwinm	23, 14, 26, 22, 29
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	rlwimi	20, 20, 24, 0, 31
+	rlwimi	21, 21, 24, 0, 31
+	rlwimi	22, 22, 24, 0, 31
+	rlwimi	23, 23, 24, 0, 31
+	xor	16, 16, 20
+	xor	18, 18, 21
+	xor	17, 17, 22
+	xor	19, 19, 23
+	rlwinm	20, 14, 2, 22, 29
+	rlwinm	21, 12, 2, 22, 29
+	rlwinm	22, 15, 2, 22, 29
+	rlwinm	23, 11, 2, 22, 29
+	lwzx	20, 9, 20
+	lwzx	21, 9, 21
+	lwzx	22, 9, 22
+	lwzx	23, 9, 23
+	rlwimi	20, 20, 16, 0, 31
+	rlwimi	21, 21, 16, 0, 31
+	rlwimi	22, 22, 16, 0, 31
+	rlwimi	23, 23, 16, 0, 31
+	ld	11, 0(28)
+	ld	12, 8(28)
+	xor	16, 16, 20
+	xor	18, 18, 21
+	xor	17, 17, 22
+	xor	19, 19, 23
+	addi	28, 28, 16
+	rldimi	16, 18, 32, 0
+	rldimi	17, 19, 32, 0
+	#   XOR in Key Schedule
+	xor	16, 16, 11
+	xor	17, 17, 12
+	ld	19, 0(10)
+	ld	19, 64(10)
+	ld	19, 128(10)
+	ld	19, 192(10)
+	ld	19, 256(10)
+	ld	19, 320(10)
+	ld	19, 384(10)
+	ld	19, 448(10)
+	ld	19, 512(10)
+	ld	19, 576(10)
+	ld	19, 640(10)
+	ld	19, 704(10)
+	ld	19, 768(10)
+	ld	19, 832(10)
+	ld	19, 896(10)
+	ld	19, 960(10)
+	rldicl	14, 16, 32, 56
+	andi.   11, 17, 255
+	rldicl	15, 17, 32, 56
+	andi.   12, 16, 255
+	lbzx	14, 10, 14
+	lbzx	11, 10, 11
+	lbzx	15, 10, 15
+	lbzx	12, 10, 12
+	rldimi	11, 14, 32, 24
+	rldimi	12, 15, 32, 24
+	rldicl	20, 17, 56, 56
+	rldicl	21, 17, 24, 56
+	rldicl	22, 16, 56, 56
+	rldicl	23, 16, 24, 56
+	lbzx	20, 10, 20
+	lbzx	21, 10, 21
+	lbzx	22, 10, 22
+	lbzx	23, 10, 23
+	rldimi	11, 20, 40, 16
+	rlwimi	11, 21, 8, 16, 23
+	rldimi	12, 22, 40, 16
+	rlwimi	12, 23, 8, 16, 23
+	rldicl	20, 17, 16, 56
+	rldicl	21, 16, 48, 56
+	rldicl	22, 16, 16, 56
+	rldicl	23, 17, 48, 56
+	lbzx	20, 10, 20
+	lbzx	21, 10, 21
+	lbzx	22, 10, 22
+	lbzx	23, 10, 23
+	rldimi	11, 20, 48, 8
+	rlwimi	11, 21, 16, 8, 15
+	rldimi	12, 22, 48, 8
+	rlwimi	12, 23, 16, 8, 15
+	rldicl	20, 16, 40, 56
+	rldicl	21, 16, 8, 56
+	rldicl	22, 17, 40, 56
+	rldicl	23, 17, 8, 56
+	lbzx	20, 10, 20
+	lbzx	21, 10, 21
+	lbzx	22, 10, 22
+	lbzx	23, 10, 23
+	ld	16, 0(28)
+	ld	17, 8(28)
+	rldimi	11, 20, 56, 0
+	rlwimi	11, 21, 24, 0, 7
+	rldimi	12, 22, 56, 0
+	rlwimi	12, 23, 24, 0, 7
+	addi	28, 28, 16
+	#   XOR in Key Schedule
+	xor	11, 11, 16
+	xor	12, 12, 17
+	xor	11, 11, 26
+	xor	12, 12, 27
+	std	11, 0(4)
+	std	12, 8(4)
+	addic.	5, 5, -16
+	addi	3, 3, 16
+	addi	4, 4, 16
+	bne	L_AES_CBC_decrypt_loop_block
+	std	24, 0(8)
+	std	25, 8(8)
+	b	L_AES_CBC_decrypt_end_dec
+L_AES_CBC_decrypt_end_dec_odd:
+	std	26, 0(8)
+	std	27, 8(8)
+L_AES_CBC_decrypt_end_dec:
+	ld	0, 120(1)
+	mtlr	0
+	ld	14, 0(1)
+	ld	15, 8(1)
+	ld	16, 16(1)
+	ld	17, 24(1)
+	ld	18, 32(1)
+	ld	19, 40(1)
+	ld	20, 48(1)
+	ld	21, 56(1)
+	ld	22, 64(1)
+	ld	23, 72(1)
+	ld	24, 80(1)
+	ld	25, 88(1)
+	ld	26, 96(1)
+	ld	27, 104(1)
+	ld	28, 112(1)
+	addi	1, 1, 0x80
+	blr
+#ifndef __APPLE__
+	.size	AES_CBC_decrypt,.-AES_CBC_decrypt
+#endif /* __APPLE__ */
+#endif /* HAVE_AES_CBC */
+#endif /* WOLFSSL_AES_DIRECT || WOLFSSL_AES_COUNTER || HAVE_AES_CBC
+        * HAVE_AES_ECB */
+#ifdef WOLFSSL_AES_XTS
+#ifndef __APPLE__
+.text
+.globl	AES_XTS_decrypt
+.type	AES_XTS_decrypt,@function
+.align	16
+AES_XTS_decrypt:
+#else
+.section	__TEXT,__text
+.globl	_AES_XTS_decrypt
+.p2align	4
+_AES_XTS_decrypt:
+#endif /* __APPLE__ */
+	stdu	1, -136(1)
+	mflr	0
+	std	0, 128(1)
+	std	14, 0(1)
+	std	15, 8(1)
+	std	16, 16(1)
+	std	17, 24(1)
+	std	18, 32(1)
+	std	19, 40(1)
+	std	20, 48(1)
+	std	21, 56(1)
+	std	22, 64(1)
+	std	23, 72(1)
+	std	24, 80(1)
+	std	25, 88(1)
+	std	26, 96(1)
+	std	27, 104(1)
+	std	28, 112(1)
+	std	29, 120(1)
+#if defined(__ARCH_PWR8) || (defined(_ARCH_PWR8))
+	addis	14, 2, L_AES_PPC64_te4_0@toc@ha
+	addi	14, 14, L_AES_PPC64_te4_0@toc@l
+#else
+	lis	14, L_AES_PPC64_te4_0@ha
+	la	14, L_AES_PPC64_te4_0@l(14)
+#endif /* defined(__ARCH_PWR8) || (defined(_ARCH_PWR8)) */
+	addi	15, 14, 0x400
+	addi	16, 14, 0x800
+	addi	17, 14, 0xc00
+	andi.   18, 5, 15
+	addi	18, 18, 15
+	andi.   18, 18, 16
+	subf	5, 18, 5
+	ld	26, 0(6)
+	ld	27, 8(6)
+	ld	18, 0(8)
+	ld	19, 8(8)
+	addi	8, 8, 16
+	rldicr	26, 26, 32, 63
+	rldicr	27, 27, 32, 63
+	# Round: 0 - XOR in key schedule
+	xor	26, 26, 18
+	xor	27, 27, 19
+	addi	0, 10, -2
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+	ld	21, 0(14)
+	ld	21, 32(14)
+	ld	21, 64(14)
+	ld	21, 96(14)
+	ld	21, 128(14)
+	ld	21, 160(14)
+	ld	21, 192(14)
+	ld	21, 224(14)
+	ld	21, 256(14)
+	ld	21, 288(14)
+	ld	21, 320(14)
+	ld	21, 352(14)
+	ld	21, 384(14)
+	ld	21, 416(14)
+	ld	21, 448(14)
+	ld	21, 480(14)
+	ld	21, 512(14)
+	ld	21, 544(14)
+	ld	21, 576(14)
+	ld	21, 608(14)
+	ld	21, 640(14)
+	ld	21, 672(14)
+	ld	21, 704(14)
+	ld	21, 736(14)
+	ld	21, 768(14)
+	ld	21, 800(14)
+	ld	21, 832(14)
+	ld	21, 864(14)
+	ld	21, 896(14)
+	ld	21, 928(14)
+	ld	21, 960(14)
+	ld	21, 992(14)
+	ld	21, 1024(14)
+	ld	21, 1056(14)
+	ld	21, 1088(14)
+	ld	21, 1120(14)
+	ld	21, 1152(14)
+	ld	21, 1184(14)
+	ld	21, 1216(14)
+	ld	21, 1248(14)
+	ld	21, 1280(14)
+	ld	21, 1312(14)
+	ld	21, 1344(14)
+	ld	21, 1376(14)
+	ld	21, 1408(14)
+	ld	21, 1440(14)
+	ld	21, 1472(14)
+	ld	21, 1504(14)
+	ld	21, 1536(14)
+	ld	21, 1568(14)
+	ld	21, 1600(14)
+	ld	21, 1632(14)
+	ld	21, 1664(14)
+	ld	21, 1696(14)
+	ld	21, 1728(14)
+	ld	21, 1760(14)
+	ld	21, 1792(14)
+	ld	21, 1824(14)
+	ld	21, 1856(14)
+	ld	21, 1888(14)
+	ld	21, 1920(14)
+	ld	21, 1952(14)
+	ld	21, 1984(14)
+	ld	21, 2016(14)
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+	srwi	0, 0, 1
+	mtctr	0
+L_AES_XTS_decrypt_loop_nr_tweak:
+	srdi	28, 26, 32
+	srdi	29, 27, 32
+	rlwinm	18, 26, 10, 22, 29
+	rlwinm	19, 27, 10, 22, 29
+	rlwinm	20, 28, 10, 22, 29
+	rlwinm	21, 29, 10, 22, 29
+	lwzx	18, 14, 18
+	lwzx	19, 14, 19
+	lwzx	20, 14, 20
+	lwzx	21, 14, 21
+	rlwinm	22, 28, 18, 22, 29
+	rlwinm	23, 27, 18, 22, 29
+	rlwinm	24, 29, 18, 22, 29
+	rlwinm	25, 26, 18, 22, 29
+	lwzx	22, 15, 22
+	lwzx	23, 15, 23
+	lwzx	24, 15, 24
+	lwzx	25, 15, 25
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 27, 26, 22, 29
+	rlwinm	23, 29, 26, 22, 29
+	rlwinm	24, 26, 26, 22, 29
+	rlwinm	25, 28, 26, 22, 29
+	lwzx	22, 16, 22
+	lwzx	23, 16, 23
+	lwzx	24, 16, 24
+	lwzx	25, 16, 25
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 29, 2, 22, 29
+	rlwinm	23, 26, 2, 22, 29
+	rlwinm	24, 28, 2, 22, 29
+	rlwinm	25, 27, 2, 22, 29
+	lwzx	22, 17, 22
+	lwzx	23, 17, 23
+	lwzx	24, 17, 24
+	lwzx	25, 17, 25
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	ld	26, 0(8)
+	ld	27, 8(8)
+	rldimi	18, 20, 32, 0
+	rldimi	19, 21, 32, 0
+	addi	8, 8, 16
+	#   XOR in Key Schedule
+	xor	18, 18, 26
+	xor	19, 19, 27
+	srdi	20, 18, 32
+	srdi	21, 19, 32
+	rlwinm	26, 18, 10, 22, 29
+	rlwinm	27, 19, 10, 22, 29
+	rlwinm	28, 20, 10, 22, 29
+	rlwinm	29, 21, 10, 22, 29
+	lwzx	26, 14, 26
+	lwzx	27, 14, 27
+	lwzx	28, 14, 28
+	lwzx	29, 14, 29
+	rlwinm	22, 20, 18, 22, 29
+	rlwinm	23, 19, 18, 22, 29
+	rlwinm	24, 21, 18, 22, 29
+	rlwinm	25, 18, 18, 22, 29
+	lwzx	22, 15, 22
+	lwzx	23, 15, 23
+	lwzx	24, 15, 24
+	lwzx	25, 15, 25
+	xor	26, 26, 22
+	xor	28, 28, 23
+	xor	27, 27, 24
+	xor	29, 29, 25
+	rlwinm	22, 19, 26, 22, 29
+	rlwinm	23, 21, 26, 22, 29
+	rlwinm	24, 18, 26, 22, 29
+	rlwinm	25, 20, 26, 22, 29
+	lwzx	22, 16, 22
+	lwzx	23, 16, 23
+	lwzx	24, 16, 24
+	lwzx	25, 16, 25
+	xor	26, 26, 22
+	xor	28, 28, 23
+	xor	27, 27, 24
+	xor	29, 29, 25
+	rlwinm	22, 21, 2, 22, 29
+	rlwinm	23, 18, 2, 22, 29
+	rlwinm	24, 20, 2, 22, 29
+	rlwinm	25, 19, 2, 22, 29
+	lwzx	22, 17, 22
+	lwzx	23, 17, 23
+	lwzx	24, 17, 24
+	lwzx	25, 17, 25
+	xor	26, 26, 22
+	xor	28, 28, 23
+	xor	27, 27, 24
+	xor	29, 29, 25
+	ld	18, 0(8)
+	ld	19, 8(8)
+	rldimi	26, 28, 32, 0
+	rldimi	27, 29, 32, 0
+	addi	8, 8, 16
+	#   XOR in Key Schedule
+	xor	26, 26, 18
+	xor	27, 27, 19
+	bdnz	L_AES_XTS_decrypt_loop_nr_tweak
+	srdi	28, 26, 32
+	srdi	29, 27, 32
+	rlwinm	18, 26, 10, 22, 29
+	rlwinm	19, 27, 10, 22, 29
+	rlwinm	20, 28, 10, 22, 29
+	rlwinm	21, 29, 10, 22, 29
+	lwzx	18, 14, 18
+	lwzx	19, 14, 19
+	lwzx	20, 14, 20
+	lwzx	21, 14, 21
+	rlwinm	22, 28, 18, 22, 29
+	rlwinm	23, 27, 18, 22, 29
+	rlwinm	24, 29, 18, 22, 29
+	rlwinm	25, 26, 18, 22, 29
+	lwzx	22, 15, 22
+	lwzx	23, 15, 23
+	lwzx	24, 15, 24
+	lwzx	25, 15, 25
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 27, 26, 22, 29
+	rlwinm	23, 29, 26, 22, 29
+	rlwinm	24, 26, 26, 22, 29
+	rlwinm	25, 28, 26, 22, 29
+	lwzx	22, 16, 22
+	lwzx	23, 16, 23
+	lwzx	24, 16, 24
+	lwzx	25, 16, 25
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 29, 2, 22, 29
+	rlwinm	23, 26, 2, 22, 29
+	rlwinm	24, 28, 2, 22, 29
+	rlwinm	25, 27, 2, 22, 29
+	lwzx	22, 17, 22
+	lwzx	23, 17, 23
+	lwzx	24, 17, 24
+	lwzx	25, 17, 25
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	ld	26, 0(8)
+	ld	27, 8(8)
+	rldimi	18, 20, 32, 0
+	rldimi	19, 21, 32, 0
+	addi	8, 8, 16
+	#   XOR in Key Schedule
+	xor	18, 18, 26
+	xor	19, 19, 27
+	srdi	20, 18, 32
+	srdi	21, 19, 32
+	rlwinm	26, 18, 2, 22, 29
+	rlwinm	27, 19, 2, 22, 29
+	rlwinm	29, 20, 2, 22, 29
+	rlwinm	28, 21, 2, 22, 29
+	lwzx	26, 15, 26
+	lwzx	27, 15, 27
+	lwzx	29, 15, 29
+	lwzx	28, 15, 28
+	rlwinm	22, 19, 26, 22, 29
+	rlwinm	23, 21, 26, 22, 29
+	rldimi	27, 29, 32, 24
+	rldimi	26, 28, 32, 24
+	rlwinm	24, 18, 26, 22, 29
+	rlwinm	25, 20, 26, 22, 29
+	lwzx	22, 15, 22
+	lwzx	23, 15, 23
+	lwzx	24, 15, 24
+	lwzx	25, 15, 25
+	rldimi	26, 22, 40, 16
+	rlwimi	26, 23, 8, 16, 23
+	rldimi	27, 24, 40, 16
+	rlwimi	27, 25, 8, 16, 23
+	rlwinm	22, 20, 18, 22, 29
+	rlwinm	23, 19, 18, 22, 29
+	rlwinm	24, 21, 18, 22, 29
+	rlwinm	25, 18, 18, 22, 29
+	lwzx	22, 15, 22
+	lwzx	23, 15, 23
+	lwzx	24, 15, 24
+	lwzx	25, 15, 25
+	rldimi	26, 22, 48, 8
+	rlwimi	26, 23, 16, 8, 15
+	rldimi	27, 24, 48, 8
+	rlwimi	27, 25, 16, 8, 15
+	rlwinm	22, 18, 10, 22, 29
+	rlwinm	23, 20, 10, 22, 29
+	rlwinm	24, 19, 10, 22, 29
+	rlwinm	25, 21, 10, 22, 29
+	lwzx	22, 15, 22
+	lwzx	23, 15, 23
+	lwzx	24, 15, 24
+	lwzx	25, 15, 25
+	ld	18, 0(8)
+	ld	19, 8(8)
+	rldimi	26, 22, 56, 0
+	rlwimi	26, 23, 24, 0, 7
+	rldimi	27, 24, 56, 0
+	rlwimi	27, 25, 24, 0, 7
+	addi	8, 8, 16
+	#   XOR in Key Schedule
+	xor	26, 26, 18
+	xor	27, 27, 19
+	std	26, 0(9)
+	std	27, 8(9)
+#if defined(__ARCH_PWR8) || (defined(_ARCH_PWR8))
+	addis	11, 2, L_AES_PPC64_td@toc@ha
+	addi	11, 11, L_AES_PPC64_td@toc@l
+#else
+	lis	11, L_AES_PPC64_td@ha
+	la	11, L_AES_PPC64_td@l(11)
+#endif /* defined(__ARCH_PWR8) || (defined(_ARCH_PWR8)) */
+#if defined(__ARCH_PWR8) || (defined(_ARCH_PWR8))
+	addis	12, 2, L_AES_PPC64_td4@toc@ha
+	addi	12, 12, L_AES_PPC64_td4@toc@l
+#else
+	lis	12, L_AES_PPC64_td4@ha
+	la	12, L_AES_PPC64_td4@l(12)
+#endif /* defined(__ARCH_PWR8) || (defined(_ARCH_PWR8)) */
+	cmpdi	5, 16
+	blt	L_AES_XTS_decrypt_start_partail
+L_AES_XTS_decrypt_loop_block:
+	addi	8, 7, 0
+	ld	14, 0(3)
+	ld	15, 8(3)
+	ld	18, 0(8)
+	ld	19, 8(8)
+	addi	8, 8, 16
+	xor	14, 14, 26
+	xor	15, 15, 27
+	rldicr	14, 14, 32, 63
+	rldicr	15, 15, 32, 63
+	# Round: 0 - XOR in key schedule
+	xor	14, 14, 18
+	xor	15, 15, 19
+	addi	0, 10, -2
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+	ld	21, 0(11)
+	ld	21, 64(11)
+	ld	21, 128(11)
+	ld	21, 192(11)
+	ld	21, 256(11)
+	ld	21, 320(11)
+	ld	21, 384(11)
+	ld	21, 448(11)
+	ld	21, 512(11)
+	ld	21, 576(11)
+	ld	21, 640(11)
+	ld	21, 704(11)
+	ld	21, 768(11)
+	ld	21, 832(11)
+	ld	21, 896(11)
+	ld	21, 960(11)
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+	srwi	0, 0, 1
+	mtctr	0
+L_AES_XTS_decrypt_loop_nr:
+	srdi	16, 14, 32
+	srdi	17, 15, 32
+	rlwinm	18, 17, 18, 22, 29
+	rlwinm	20, 14, 18, 22, 29
+	rlwinm	19, 16, 18, 22, 29
+	rlwinm	21, 15, 18, 22, 29
+	lwzx	18, 11, 18
+	lwzx	20, 11, 20
+	lwzx	19, 11, 19
+	lwzx	21, 11, 21
+	rlwinm	22, 14, 10, 22, 29
+	rlwinm	23, 16, 10, 22, 29
+	rlwinm	24, 15, 10, 22, 29
+	rlwinm	25, 17, 10, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 8, 0, 31
+	rlwimi	23, 23, 8, 0, 31
+	rlwimi	24, 24, 8, 0, 31
+	rlwimi	25, 25, 8, 0, 31
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 15, 26, 22, 29
+	rlwinm	23, 17, 26, 22, 29
+	rlwinm	24, 14, 26, 22, 29
+	rlwinm	25, 16, 26, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 24, 0, 31
+	rlwimi	23, 23, 24, 0, 31
+	rlwimi	24, 24, 24, 0, 31
+	rlwimi	25, 25, 24, 0, 31
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 16, 2, 22, 29
+	rlwinm	23, 15, 2, 22, 29
+	rlwinm	24, 17, 2, 22, 29
+	rlwinm	25, 14, 2, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 16, 0, 31
+	rlwimi	23, 23, 16, 0, 31
+	rlwimi	24, 24, 16, 0, 31
+	rlwimi	25, 25, 16, 0, 31
+	ld	14, 0(8)
+	ld	15, 8(8)
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	addi	8, 8, 16
+	rldimi	18, 20, 32, 0
+	rldimi	19, 21, 32, 0
+	#   XOR in Key Schedule
+	xor	18, 18, 14
+	xor	19, 19, 15
+	srdi	20, 18, 32
+	srdi	21, 19, 32
+	rlwinm	14, 21, 18, 22, 29
+	rlwinm	16, 18, 18, 22, 29
+	rlwinm	15, 20, 18, 22, 29
+	rlwinm	17, 19, 18, 22, 29
+	lwzx	14, 11, 14
+	lwzx	16, 11, 16
+	lwzx	15, 11, 15
+	lwzx	17, 11, 17
+	rlwinm	22, 18, 10, 22, 29
+	rlwinm	23, 20, 10, 22, 29
+	rlwinm	24, 19, 10, 22, 29
+	rlwinm	25, 21, 10, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 8, 0, 31
+	rlwimi	23, 23, 8, 0, 31
+	rlwimi	24, 24, 8, 0, 31
+	rlwimi	25, 25, 8, 0, 31
+	xor	14, 14, 22
+	xor	16, 16, 23
+	xor	15, 15, 24
+	xor	17, 17, 25
+	rlwinm	22, 19, 26, 22, 29
+	rlwinm	23, 21, 26, 22, 29
+	rlwinm	24, 18, 26, 22, 29
+	rlwinm	25, 20, 26, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 24, 0, 31
+	rlwimi	23, 23, 24, 0, 31
+	rlwimi	24, 24, 24, 0, 31
+	rlwimi	25, 25, 24, 0, 31
+	xor	14, 14, 22
+	xor	16, 16, 23
+	xor	15, 15, 24
+	xor	17, 17, 25
+	rlwinm	22, 20, 2, 22, 29
+	rlwinm	23, 19, 2, 22, 29
+	rlwinm	24, 21, 2, 22, 29
+	rlwinm	25, 18, 2, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 16, 0, 31
+	rlwimi	23, 23, 16, 0, 31
+	rlwimi	24, 24, 16, 0, 31
+	rlwimi	25, 25, 16, 0, 31
+	ld	18, 0(8)
+	ld	19, 8(8)
+	xor	14, 14, 22
+	xor	16, 16, 23
+	xor	15, 15, 24
+	xor	17, 17, 25
+	addi	8, 8, 16
+	rldimi	14, 16, 32, 0
+	rldimi	15, 17, 32, 0
+	#   XOR in Key Schedule
+	xor	14, 14, 18
+	xor	15, 15, 19
+	bdnz	L_AES_XTS_decrypt_loop_nr
+	srdi	16, 14, 32
+	srdi	17, 15, 32
+	rlwinm	18, 17, 18, 22, 29
+	rlwinm	20, 14, 18, 22, 29
+	rlwinm	19, 16, 18, 22, 29
+	rlwinm	21, 15, 18, 22, 29
+	lwzx	18, 11, 18
+	lwzx	20, 11, 20
+	lwzx	19, 11, 19
+	lwzx	21, 11, 21
+	rlwinm	22, 14, 10, 22, 29
+	rlwinm	23, 16, 10, 22, 29
+	rlwinm	24, 15, 10, 22, 29
+	rlwinm	25, 17, 10, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 8, 0, 31
+	rlwimi	23, 23, 8, 0, 31
+	rlwimi	24, 24, 8, 0, 31
+	rlwimi	25, 25, 8, 0, 31
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 15, 26, 22, 29
+	rlwinm	23, 17, 26, 22, 29
+	rlwinm	24, 14, 26, 22, 29
+	rlwinm	25, 16, 26, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 24, 0, 31
+	rlwimi	23, 23, 24, 0, 31
+	rlwimi	24, 24, 24, 0, 31
+	rlwimi	25, 25, 24, 0, 31
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 16, 2, 22, 29
+	rlwinm	23, 15, 2, 22, 29
+	rlwinm	24, 17, 2, 22, 29
+	rlwinm	25, 14, 2, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 16, 0, 31
+	rlwimi	23, 23, 16, 0, 31
+	rlwimi	24, 24, 16, 0, 31
+	rlwimi	25, 25, 16, 0, 31
+	ld	14, 0(8)
+	ld	15, 8(8)
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	addi	8, 8, 16
+	rldimi	18, 20, 32, 0
+	rldimi	19, 21, 32, 0
+	#   XOR in Key Schedule
+	xor	18, 18, 14
+	xor	19, 19, 15
+	ld	21, 0(12)
+	ld	21, 64(12)
+	ld	21, 128(12)
+	ld	21, 192(12)
+	ld	21, 256(12)
+	ld	21, 320(12)
+	ld	21, 384(12)
+	ld	21, 448(12)
+	ld	21, 512(12)
+	ld	21, 576(12)
+	ld	21, 640(12)
+	ld	21, 704(12)
+	ld	21, 768(12)
+	ld	21, 832(12)
+	ld	21, 896(12)
+	ld	21, 960(12)
+	rldicl	16, 18, 32, 56
+	andi.   14, 19, 255
+	rldicl	17, 19, 32, 56
+	andi.   15, 18, 255
+	lbzx	16, 12, 16
+	lbzx	14, 12, 14
+	lbzx	17, 12, 17
+	lbzx	15, 12, 15
+	rldimi	14, 16, 32, 24
+	rldimi	15, 17, 32, 24
+	rldicl	22, 19, 56, 56
+	rldicl	23, 19, 24, 56
+	rldicl	24, 18, 56, 56
+	rldicl	25, 18, 24, 56
+	lbzx	22, 12, 22
+	lbzx	23, 12, 23
+	lbzx	24, 12, 24
+	lbzx	25, 12, 25
+	rldimi	14, 22, 40, 16
+	rlwimi	14, 23, 8, 16, 23
+	rldimi	15, 24, 40, 16
+	rlwimi	15, 25, 8, 16, 23
+	rldicl	22, 19, 16, 56
+	rldicl	23, 18, 48, 56
+	rldicl	24, 18, 16, 56
+	rldicl	25, 19, 48, 56
+	lbzx	22, 12, 22
+	lbzx	23, 12, 23
+	lbzx	24, 12, 24
+	lbzx	25, 12, 25
+	rldimi	14, 22, 48, 8
+	rlwimi	14, 23, 16, 8, 15
+	rldimi	15, 24, 48, 8
+	rlwimi	15, 25, 16, 8, 15
+	rldicl	22, 18, 40, 56
+	rldicl	23, 18, 8, 56
+	rldicl	24, 19, 40, 56
+	rldicl	25, 19, 8, 56
+	lbzx	22, 12, 22
+	lbzx	23, 12, 23
+	lbzx	24, 12, 24
+	lbzx	25, 12, 25
+	ld	18, 0(8)
+	ld	19, 8(8)
+	rldimi	14, 22, 56, 0
+	rlwimi	14, 23, 24, 0, 7
+	rldimi	15, 24, 56, 0
+	rlwimi	15, 25, 24, 0, 7
+	addi	8, 8, 16
+	#   XOR in Key Schedule
+	xor	14, 14, 18
+	xor	15, 15, 19
+	xor	14, 14, 26
+	xor	15, 15, 27
+	li	17, 8
+	ldbrx	26, 0, 9
+	ldbrx	27, 9, 17
+	std	14, 0(4)
+	std	15, 8(4)
+	sradi	22, 27, 63
+	srdi	23, 26, 63
+	sldi	27, 27, 1
+	andi.	22, 22, 0x87
+	sldi	26, 26, 1
+	xor	27, 27, 23
+	xor	26, 26, 22
+	stdbrx	26, 0, 9
+	stdbrx	27, 9, 17
+	ld	26, 0(9)
+	ld	27, 8(9)
+	addi	5, 5, -16
+	addi	3, 3, 16
+	addi	4, 4, 16
+	cmpdi	5, 16
+	bge	L_AES_XTS_decrypt_loop_block
+	cmpdi	5, 0
+	beq	L_AES_XTS_decrypt_done_data
+L_AES_XTS_decrypt_start_partail:
+	li	17, 8
+	ldbrx	28, 0, 9
+	ldbrx	29, 9, 17
+	sradi	22, 29, 63
+	srdi	23, 28, 63
+	sldi	29, 29, 1
+	andi.	22, 22, 0x87
+	sldi	28, 28, 1
+	xor	29, 29, 23
+	xor	28, 28, 22
+	stdbrx	28, 0, 9
+	stdbrx	29, 9, 17
+	ld	28, 0(9)
+	ld	29, 8(9)
+	addi	8, 7, 0
+	ld	14, 0(3)
+	ld	15, 8(3)
+	addi	3, 3, 16
+	ld	18, 0(8)
+	ld	19, 8(8)
+	addi	8, 8, 16
+	xor	14, 14, 28
+	xor	15, 15, 29
+	rldicr	14, 14, 32, 63
+	rldicr	15, 15, 32, 63
+	# Round: 0 - XOR in key schedule
+	xor	14, 14, 18
+	xor	15, 15, 19
+	addi	0, 10, -2
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+	ld	21, 0(11)
+	ld	21, 64(11)
+	ld	21, 128(11)
+	ld	21, 192(11)
+	ld	21, 256(11)
+	ld	21, 320(11)
+	ld	21, 384(11)
+	ld	21, 448(11)
+	ld	21, 512(11)
+	ld	21, 576(11)
+	ld	21, 640(11)
+	ld	21, 704(11)
+	ld	21, 768(11)
+	ld	21, 832(11)
+	ld	21, 896(11)
+	ld	21, 960(11)
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+	srwi	0, 0, 1
+	mtctr	0
+L_AES_XTS_decrypt_loop_nr_partial_1:
+	srdi	16, 14, 32
+	srdi	17, 15, 32
+	rlwinm	18, 17, 18, 22, 29
+	rlwinm	20, 14, 18, 22, 29
+	rlwinm	19, 16, 18, 22, 29
+	rlwinm	21, 15, 18, 22, 29
+	lwzx	18, 11, 18
+	lwzx	20, 11, 20
+	lwzx	19, 11, 19
+	lwzx	21, 11, 21
+	rlwinm	22, 14, 10, 22, 29
+	rlwinm	23, 16, 10, 22, 29
+	rlwinm	24, 15, 10, 22, 29
+	rlwinm	25, 17, 10, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 8, 0, 31
+	rlwimi	23, 23, 8, 0, 31
+	rlwimi	24, 24, 8, 0, 31
+	rlwimi	25, 25, 8, 0, 31
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 15, 26, 22, 29
+	rlwinm	23, 17, 26, 22, 29
+	rlwinm	24, 14, 26, 22, 29
+	rlwinm	25, 16, 26, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 24, 0, 31
+	rlwimi	23, 23, 24, 0, 31
+	rlwimi	24, 24, 24, 0, 31
+	rlwimi	25, 25, 24, 0, 31
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 16, 2, 22, 29
+	rlwinm	23, 15, 2, 22, 29
+	rlwinm	24, 17, 2, 22, 29
+	rlwinm	25, 14, 2, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 16, 0, 31
+	rlwimi	23, 23, 16, 0, 31
+	rlwimi	24, 24, 16, 0, 31
+	rlwimi	25, 25, 16, 0, 31
+	ld	14, 0(8)
+	ld	15, 8(8)
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	addi	8, 8, 16
+	rldimi	18, 20, 32, 0
+	rldimi	19, 21, 32, 0
+	#   XOR in Key Schedule
+	xor	18, 18, 14
+	xor	19, 19, 15
+	srdi	20, 18, 32
+	srdi	21, 19, 32
+	rlwinm	14, 21, 18, 22, 29
+	rlwinm	16, 18, 18, 22, 29
+	rlwinm	15, 20, 18, 22, 29
+	rlwinm	17, 19, 18, 22, 29
+	lwzx	14, 11, 14
+	lwzx	16, 11, 16
+	lwzx	15, 11, 15
+	lwzx	17, 11, 17
+	rlwinm	22, 18, 10, 22, 29
+	rlwinm	23, 20, 10, 22, 29
+	rlwinm	24, 19, 10, 22, 29
+	rlwinm	25, 21, 10, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 8, 0, 31
+	rlwimi	23, 23, 8, 0, 31
+	rlwimi	24, 24, 8, 0, 31
+	rlwimi	25, 25, 8, 0, 31
+	xor	14, 14, 22
+	xor	16, 16, 23
+	xor	15, 15, 24
+	xor	17, 17, 25
+	rlwinm	22, 19, 26, 22, 29
+	rlwinm	23, 21, 26, 22, 29
+	rlwinm	24, 18, 26, 22, 29
+	rlwinm	25, 20, 26, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 24, 0, 31
+	rlwimi	23, 23, 24, 0, 31
+	rlwimi	24, 24, 24, 0, 31
+	rlwimi	25, 25, 24, 0, 31
+	xor	14, 14, 22
+	xor	16, 16, 23
+	xor	15, 15, 24
+	xor	17, 17, 25
+	rlwinm	22, 20, 2, 22, 29
+	rlwinm	23, 19, 2, 22, 29
+	rlwinm	24, 21, 2, 22, 29
+	rlwinm	25, 18, 2, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 16, 0, 31
+	rlwimi	23, 23, 16, 0, 31
+	rlwimi	24, 24, 16, 0, 31
+	rlwimi	25, 25, 16, 0, 31
+	ld	18, 0(8)
+	ld	19, 8(8)
+	xor	14, 14, 22
+	xor	16, 16, 23
+	xor	15, 15, 24
+	xor	17, 17, 25
+	addi	8, 8, 16
+	rldimi	14, 16, 32, 0
+	rldimi	15, 17, 32, 0
+	#   XOR in Key Schedule
+	xor	14, 14, 18
+	xor	15, 15, 19
+	bdnz	L_AES_XTS_decrypt_loop_nr_partial_1
+	srdi	16, 14, 32
+	srdi	17, 15, 32
+	rlwinm	18, 17, 18, 22, 29
+	rlwinm	20, 14, 18, 22, 29
+	rlwinm	19, 16, 18, 22, 29
+	rlwinm	21, 15, 18, 22, 29
+	lwzx	18, 11, 18
+	lwzx	20, 11, 20
+	lwzx	19, 11, 19
+	lwzx	21, 11, 21
+	rlwinm	22, 14, 10, 22, 29
+	rlwinm	23, 16, 10, 22, 29
+	rlwinm	24, 15, 10, 22, 29
+	rlwinm	25, 17, 10, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 8, 0, 31
+	rlwimi	23, 23, 8, 0, 31
+	rlwimi	24, 24, 8, 0, 31
+	rlwimi	25, 25, 8, 0, 31
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 15, 26, 22, 29
+	rlwinm	23, 17, 26, 22, 29
+	rlwinm	24, 14, 26, 22, 29
+	rlwinm	25, 16, 26, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 24, 0, 31
+	rlwimi	23, 23, 24, 0, 31
+	rlwimi	24, 24, 24, 0, 31
+	rlwimi	25, 25, 24, 0, 31
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 16, 2, 22, 29
+	rlwinm	23, 15, 2, 22, 29
+	rlwinm	24, 17, 2, 22, 29
+	rlwinm	25, 14, 2, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 16, 0, 31
+	rlwimi	23, 23, 16, 0, 31
+	rlwimi	24, 24, 16, 0, 31
+	rlwimi	25, 25, 16, 0, 31
+	ld	14, 0(8)
+	ld	15, 8(8)
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	addi	8, 8, 16
+	rldimi	18, 20, 32, 0
+	rldimi	19, 21, 32, 0
+	#   XOR in Key Schedule
+	xor	18, 18, 14
+	xor	19, 19, 15
+	ld	21, 0(12)
+	ld	21, 64(12)
+	ld	21, 128(12)
+	ld	21, 192(12)
+	ld	21, 256(12)
+	ld	21, 320(12)
+	ld	21, 384(12)
+	ld	21, 448(12)
+	ld	21, 512(12)
+	ld	21, 576(12)
+	ld	21, 640(12)
+	ld	21, 704(12)
+	ld	21, 768(12)
+	ld	21, 832(12)
+	ld	21, 896(12)
+	ld	21, 960(12)
+	rldicl	16, 18, 32, 56
+	andi.   14, 19, 255
+	rldicl	17, 19, 32, 56
+	andi.   15, 18, 255
+	lbzx	16, 12, 16
+	lbzx	14, 12, 14
+	lbzx	17, 12, 17
+	lbzx	15, 12, 15
+	rldimi	14, 16, 32, 24
+	rldimi	15, 17, 32, 24
+	rldicl	22, 19, 56, 56
+	rldicl	23, 19, 24, 56
+	rldicl	24, 18, 56, 56
+	rldicl	25, 18, 24, 56
+	lbzx	22, 12, 22
+	lbzx	23, 12, 23
+	lbzx	24, 12, 24
+	lbzx	25, 12, 25
+	rldimi	14, 22, 40, 16
+	rlwimi	14, 23, 8, 16, 23
+	rldimi	15, 24, 40, 16
+	rlwimi	15, 25, 8, 16, 23
+	rldicl	22, 19, 16, 56
+	rldicl	23, 18, 48, 56
+	rldicl	24, 18, 16, 56
+	rldicl	25, 19, 48, 56
+	lbzx	22, 12, 22
+	lbzx	23, 12, 23
+	lbzx	24, 12, 24
+	lbzx	25, 12, 25
+	rldimi	14, 22, 48, 8
+	rlwimi	14, 23, 16, 8, 15
+	rldimi	15, 24, 48, 8
+	rlwimi	15, 25, 16, 8, 15
+	rldicl	22, 18, 40, 56
+	rldicl	23, 18, 8, 56
+	rldicl	24, 19, 40, 56
+	rldicl	25, 19, 8, 56
+	lbzx	22, 12, 22
+	lbzx	23, 12, 23
+	lbzx	24, 12, 24
+	lbzx	25, 12, 25
+	ld	18, 0(8)
+	ld	19, 8(8)
+	rldimi	14, 22, 56, 0
+	rlwimi	14, 23, 24, 0, 7
+	rldimi	15, 24, 56, 0
+	rlwimi	15, 25, 24, 0, 7
+	addi	8, 8, 16
+	#   XOR in Key Schedule
+	xor	14, 14, 18
+	xor	15, 15, 19
+	xor	14, 14, 28
+	xor	15, 15, 29
+	std	14, 0(9)
+	std	15, 8(9)
+	addi	4, 4, 16
+	addi	18, 5, 0
+L_AES_XTS_decrypt_start_byte:
+	lbz	22, 0(9)
+	lbz	23, 0(3)
+	addi	3, 3, 1
+	stb	22, 0(4)
+	addi	4, 4, 1
+	stb	23, 0(9)
+	addi	9, 9, 1
+	addic.	18, 18, -1
+	bne	L_AES_XTS_decrypt_start_byte
+	subf	4, 5, 4
+	subf	9, 5, 9
+	addi	4, 4, -16
+	addi	8, 7, 0
+	ld	14, 0(9)
+	ld	15, 8(9)
+	ld	18, 0(8)
+	ld	19, 8(8)
+	addi	8, 8, 16
+	xor	14, 14, 26
+	xor	15, 15, 27
+	rldicr	14, 14, 32, 63
+	rldicr	15, 15, 32, 63
+	# Round: 0 - XOR in key schedule
+	xor	14, 14, 18
+	xor	15, 15, 19
+	addi	0, 10, -2
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+	ld	21, 0(11)
+	ld	21, 64(11)
+	ld	21, 128(11)
+	ld	21, 192(11)
+	ld	21, 256(11)
+	ld	21, 320(11)
+	ld	21, 384(11)
+	ld	21, 448(11)
+	ld	21, 512(11)
+	ld	21, 576(11)
+	ld	21, 640(11)
+	ld	21, 704(11)
+	ld	21, 768(11)
+	ld	21, 832(11)
+	ld	21, 896(11)
+	ld	21, 960(11)
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+	srwi	0, 0, 1
+	mtctr	0
+L_AES_XTS_decrypt_loop_nr_partial_2:
+	srdi	16, 14, 32
+	srdi	17, 15, 32
+	rlwinm	18, 17, 18, 22, 29
+	rlwinm	20, 14, 18, 22, 29
+	rlwinm	19, 16, 18, 22, 29
+	rlwinm	21, 15, 18, 22, 29
+	lwzx	18, 11, 18
+	lwzx	20, 11, 20
+	lwzx	19, 11, 19
+	lwzx	21, 11, 21
+	rlwinm	22, 14, 10, 22, 29
+	rlwinm	23, 16, 10, 22, 29
+	rlwinm	24, 15, 10, 22, 29
+	rlwinm	25, 17, 10, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 8, 0, 31
+	rlwimi	23, 23, 8, 0, 31
+	rlwimi	24, 24, 8, 0, 31
+	rlwimi	25, 25, 8, 0, 31
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 15, 26, 22, 29
+	rlwinm	23, 17, 26, 22, 29
+	rlwinm	24, 14, 26, 22, 29
+	rlwinm	25, 16, 26, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 24, 0, 31
+	rlwimi	23, 23, 24, 0, 31
+	rlwimi	24, 24, 24, 0, 31
+	rlwimi	25, 25, 24, 0, 31
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 16, 2, 22, 29
+	rlwinm	23, 15, 2, 22, 29
+	rlwinm	24, 17, 2, 22, 29
+	rlwinm	25, 14, 2, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 16, 0, 31
+	rlwimi	23, 23, 16, 0, 31
+	rlwimi	24, 24, 16, 0, 31
+	rlwimi	25, 25, 16, 0, 31
+	ld	14, 0(8)
+	ld	15, 8(8)
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	addi	8, 8, 16
+	rldimi	18, 20, 32, 0
+	rldimi	19, 21, 32, 0
+	#   XOR in Key Schedule
+	xor	18, 18, 14
+	xor	19, 19, 15
+	srdi	20, 18, 32
+	srdi	21, 19, 32
+	rlwinm	14, 21, 18, 22, 29
+	rlwinm	16, 18, 18, 22, 29
+	rlwinm	15, 20, 18, 22, 29
+	rlwinm	17, 19, 18, 22, 29
+	lwzx	14, 11, 14
+	lwzx	16, 11, 16
+	lwzx	15, 11, 15
+	lwzx	17, 11, 17
+	rlwinm	22, 18, 10, 22, 29
+	rlwinm	23, 20, 10, 22, 29
+	rlwinm	24, 19, 10, 22, 29
+	rlwinm	25, 21, 10, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 8, 0, 31
+	rlwimi	23, 23, 8, 0, 31
+	rlwimi	24, 24, 8, 0, 31
+	rlwimi	25, 25, 8, 0, 31
+	xor	14, 14, 22
+	xor	16, 16, 23
+	xor	15, 15, 24
+	xor	17, 17, 25
+	rlwinm	22, 19, 26, 22, 29
+	rlwinm	23, 21, 26, 22, 29
+	rlwinm	24, 18, 26, 22, 29
+	rlwinm	25, 20, 26, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 24, 0, 31
+	rlwimi	23, 23, 24, 0, 31
+	rlwimi	24, 24, 24, 0, 31
+	rlwimi	25, 25, 24, 0, 31
+	xor	14, 14, 22
+	xor	16, 16, 23
+	xor	15, 15, 24
+	xor	17, 17, 25
+	rlwinm	22, 20, 2, 22, 29
+	rlwinm	23, 19, 2, 22, 29
+	rlwinm	24, 21, 2, 22, 29
+	rlwinm	25, 18, 2, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 16, 0, 31
+	rlwimi	23, 23, 16, 0, 31
+	rlwimi	24, 24, 16, 0, 31
+	rlwimi	25, 25, 16, 0, 31
+	ld	18, 0(8)
+	ld	19, 8(8)
+	xor	14, 14, 22
+	xor	16, 16, 23
+	xor	15, 15, 24
+	xor	17, 17, 25
+	addi	8, 8, 16
+	rldimi	14, 16, 32, 0
+	rldimi	15, 17, 32, 0
+	#   XOR in Key Schedule
+	xor	14, 14, 18
+	xor	15, 15, 19
+	bdnz	L_AES_XTS_decrypt_loop_nr_partial_2
+	srdi	16, 14, 32
+	srdi	17, 15, 32
+	rlwinm	18, 17, 18, 22, 29
+	rlwinm	20, 14, 18, 22, 29
+	rlwinm	19, 16, 18, 22, 29
+	rlwinm	21, 15, 18, 22, 29
+	lwzx	18, 11, 18
+	lwzx	20, 11, 20
+	lwzx	19, 11, 19
+	lwzx	21, 11, 21
+	rlwinm	22, 14, 10, 22, 29
+	rlwinm	23, 16, 10, 22, 29
+	rlwinm	24, 15, 10, 22, 29
+	rlwinm	25, 17, 10, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 8, 0, 31
+	rlwimi	23, 23, 8, 0, 31
+	rlwimi	24, 24, 8, 0, 31
+	rlwimi	25, 25, 8, 0, 31
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 15, 26, 22, 29
+	rlwinm	23, 17, 26, 22, 29
+	rlwinm	24, 14, 26, 22, 29
+	rlwinm	25, 16, 26, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 24, 0, 31
+	rlwimi	23, 23, 24, 0, 31
+	rlwimi	24, 24, 24, 0, 31
+	rlwimi	25, 25, 24, 0, 31
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	rlwinm	22, 16, 2, 22, 29
+	rlwinm	23, 15, 2, 22, 29
+	rlwinm	24, 17, 2, 22, 29
+	rlwinm	25, 14, 2, 22, 29
+	lwzx	22, 11, 22
+	lwzx	23, 11, 23
+	lwzx	24, 11, 24
+	lwzx	25, 11, 25
+	rlwimi	22, 22, 16, 0, 31
+	rlwimi	23, 23, 16, 0, 31
+	rlwimi	24, 24, 16, 0, 31
+	rlwimi	25, 25, 16, 0, 31
+	ld	14, 0(8)
+	ld	15, 8(8)
+	xor	18, 18, 22
+	xor	20, 20, 23
+	xor	19, 19, 24
+	xor	21, 21, 25
+	addi	8, 8, 16
+	rldimi	18, 20, 32, 0
+	rldimi	19, 21, 32, 0
+	#   XOR in Key Schedule
+	xor	18, 18, 14
+	xor	19, 19, 15
+	ld	21, 0(12)
+	ld	21, 64(12)
+	ld	21, 128(12)
+	ld	21, 192(12)
+	ld	21, 256(12)
+	ld	21, 320(12)
+	ld	21, 384(12)
+	ld	21, 448(12)
+	ld	21, 512(12)
+	ld	21, 576(12)
+	ld	21, 640(12)
+	ld	21, 704(12)
+	ld	21, 768(12)
+	ld	21, 832(12)
+	ld	21, 896(12)
+	ld	21, 960(12)
+	rldicl	16, 18, 32, 56
+	andi.   14, 19, 255
+	rldicl	17, 19, 32, 56
+	andi.   15, 18, 255
+	lbzx	16, 12, 16
+	lbzx	14, 12, 14
+	lbzx	17, 12, 17
+	lbzx	15, 12, 15
+	rldimi	14, 16, 32, 24
+	rldimi	15, 17, 32, 24
+	rldicl	22, 19, 56, 56
+	rldicl	23, 19, 24, 56
+	rldicl	24, 18, 56, 56
+	rldicl	25, 18, 24, 56
+	lbzx	22, 12, 22
+	lbzx	23, 12, 23
+	lbzx	24, 12, 24
+	lbzx	25, 12, 25
+	rldimi	14, 22, 40, 16
+	rlwimi	14, 23, 8, 16, 23
+	rldimi	15, 24, 40, 16
+	rlwimi	15, 25, 8, 16, 23
+	rldicl	22, 19, 16, 56
+	rldicl	23, 18, 48, 56
+	rldicl	24, 18, 16, 56
+	rldicl	25, 19, 48, 56
+	lbzx	22, 12, 22
+	lbzx	23, 12, 23
+	lbzx	24, 12, 24
+	lbzx	25, 12, 25
+	rldimi	14, 22, 48, 8
+	rlwimi	14, 23, 16, 8, 15
+	rldimi	15, 24, 48, 8
+	rlwimi	15, 25, 16, 8, 15
+	rldicl	22, 18, 40, 56
+	rldicl	23, 18, 8, 56
+	rldicl	24, 19, 40, 56
+	rldicl	25, 19, 8, 56
+	lbzx	22, 12, 22
+	lbzx	23, 12, 23
+	lbzx	24, 12, 24
+	lbzx	25, 12, 25
+	ld	18, 0(8)
+	ld	19, 8(8)
+	rldimi	14, 22, 56, 0
+	rlwimi	14, 23, 24, 0, 7
+	rldimi	15, 24, 56, 0
+	rlwimi	15, 25, 24, 0, 7
+	addi	8, 8, 16
+	#   XOR in Key Schedule
+	xor	14, 14, 18
+	xor	15, 15, 19
+	xor	14, 14, 26
+	xor	15, 15, 27
+	std	14, 0(4)
+	std	15, 8(4)
+L_AES_XTS_decrypt_done_data:
+	ld	0, 128(1)
+	mtlr	0
+	ld	14, 0(1)
+	ld	15, 8(1)
+	ld	16, 16(1)
+	ld	17, 24(1)
+	ld	18, 32(1)
+	ld	19, 40(1)
+	ld	20, 48(1)
+	ld	21, 56(1)
+	ld	22, 64(1)
+	ld	23, 72(1)
+	ld	24, 80(1)
+	ld	25, 88(1)
+	ld	26, 96(1)
+	ld	27, 104(1)
+	ld	28, 112(1)
+	ld	29, 120(1)
+	addi	1, 1, 0x88
+	blr
+#ifndef __APPLE__
+	.size	AES_XTS_decrypt,.-AES_XTS_decrypt
+#endif /* __APPLE__ */
+#endif /* WOLFSSL_AES_XTS */
+#endif /* HAVE_AES_DECRYPT */
+#ifdef HAVE_AESGCM
+#ifdef GCM_TABLE_4BIT
+#ifndef __APPLE__
+	.text
+	.section	.rodata
+	.type	L_GCM_gmult_len_r, @object
+	.size	L_GCM_gmult_len_r, 128
+#else
+	.section	__DATA,__data
+#endif /* __APPLE__ */
+#ifndef __APPLE__
+	.align	4
+#else
+	.p2align	4
+#endif /* __APPLE__ */
+L_GCM_gmult_len_r:
+	.long	0x0
+	.long	0x1c200000
+	.long	0x38400000
+	.long	0x24600000
+	.long	0x70800000
+	.long	0x6ca00000
+	.long	0x48c00000
+	.long	0x54e00000
+	.long	0xe1000000
+	.long	0xfd200000
+	.long	0xd9400000
+	.long	0xc5600000
+	.long	0x91800000
+	.long	0x8da00000
+	.long	0xa9c00000
+	.long	0xb5e00000
+	.long	0x0
+	.long	0x1c20000
+	.long	0x3840000
+	.long	0x2460000
+	.long	0x7080000
+	.long	0x6ca0000
+	.long	0x48c0000
+	.long	0x54e0000
+	.long	0xe100000
+	.long	0xfd20000
+	.long	0xd940000
+	.long	0xc560000
+	.long	0x9180000
+	.long	0x8da0000
+	.long	0xa9c0000
+	.long	0xb5e0000
+#ifndef __APPLE__
+.text
+.globl	GCM_gmult_len
+.type	GCM_gmult_len,@function
+.align	16
+GCM_gmult_len:
+#else
+.section	__TEXT,__text
+.globl	_GCM_gmult_len
+.p2align	4
+_GCM_gmult_len:
+#endif /* __APPLE__ */
+	stdu	1, -104(1)
+	mflr	0
+	std	0, 96(1)
+	std	14, 0(1)
+	std	15, 8(1)
+	std	16, 16(1)
+	std	17, 24(1)
+	std	18, 32(1)
+	std	19, 40(1)
+	std	20, 48(1)
+	std	21, 56(1)
+	std	22, 64(1)
+	std	23, 72(1)
+	std	24, 80(1)
+	std	25, 88(1)
+	li	21, 0x100
+	li	25, 8
+#if defined(__ARCH_PWR8) || (defined(_ARCH_PWR8))
+	addis	7, 2, L_GCM_gmult_len_r@toc@ha
+	addi	7, 7, L_GCM_gmult_len_r@toc@l
+#else
+	lis	7, L_GCM_gmult_len_r@ha
+	la	7, L_GCM_gmult_len_r@l(7)
+#endif /* defined(__ARCH_PWR8) || (defined(_ARCH_PWR8)) */
+	add	21, 21, 4
+	addi	20, 4, 8
+	addi	22, 21, 8
+	addi	8, 7, 0x40
+L_GCM_gmult_len_start_block:
+	ldbrx	18, 0, 3
+	ldbrx	19, 25, 3
+	ldbrx	9, 0, 5
+	ldbrx	10, 25, 5
+	xor	18, 18, 9
+	xor	19, 19, 10
+	rldicr	0, 19, 32, 63
+	# Byte 15
+	rlwinm	23, 0, 12, 24, 27
+	rlwinm	24, 0, 8, 24, 27
+	ldx	14, 23, 4
+	ldx	15, 23, 20
+	ldx	11, 24, 21
+	ldx	12, 24, 22
+	ldx	10, 24, 20
+	rldicl	16, 15, 60, 60
+	rldic	17, 15, 2, 58
+	srdi	15, 15, 8
+	rldimi	15, 14, 56, 0
+	srdi	14, 14, 8
+	lwzx	9, 17, 8
+	xor	16, 16, 10
+	xor	14, 14, 11
+	rlwinm	16, 16, 2, 26, 29
+	lwzx	10, 16, 7
+	xor	15, 15, 12
+	sldi	9, 9, 32
+	xor	14, 14, 9
+	sldi	10, 10, 32
+	xor	14, 14, 10
+	# Byte 14
+	rlwinm	23, 0, 20, 24, 27
+	rlwinm	24, 0, 16, 24, 27
+	ldx	9, 23, 4
+	ldx	10, 23, 20
+	ldx	11, 24, 21
+	ldx	12, 24, 22
+	xor	14, 14, 9
+	xor	15, 15, 10
+	ldx	10, 24, 20
+	rldicl	16, 15, 60, 60
+	rldic	17, 15, 2, 58
+	srdi	15, 15, 8
+	rldimi	15, 14, 56, 0
+	srdi	14, 14, 8
+	lwzx	9, 17, 8
+	xor	16, 16, 10
+	xor	14, 14, 11
+	rlwinm	16, 16, 2, 26, 29
+	lwzx	10, 16, 7
+	xor	15, 15, 12
+	sldi	9, 9, 32
+	xor	14, 14, 9
+	sldi	10, 10, 32
+	xor	14, 14, 10
+	# Byte 13
+	rlwinm	23, 0, 28, 24, 27
+	rlwinm	24, 0, 24, 24, 27
+	ldx	9, 23, 4
+	ldx	10, 23, 20
+	ldx	11, 24, 21
+	ldx	12, 24, 22
+	xor	14, 14, 9
+	xor	15, 15, 10
+	ldx	10, 24, 20
+	rldicl	16, 15, 60, 60
+	rldic	17, 15, 2, 58
+	srdi	15, 15, 8
+	rldimi	15, 14, 56, 0
+	srdi	14, 14, 8
+	lwzx	9, 17, 8
+	xor	16, 16, 10
+	xor	14, 14, 11
+	rlwinm	16, 16, 2, 26, 29
+	lwzx	10, 16, 7
+	xor	15, 15, 12
+	sldi	9, 9, 32
+	xor	14, 14, 9
+	sldi	10, 10, 32
+	xor	14, 14, 10
+	# Byte 12
+	rlwinm	23, 0, 4, 24, 27
+	rlwinm	24, 0, 0, 24, 27
+	ldx	9, 23, 4
+	ldx	10, 23, 20
+	ldx	11, 24, 21
+	ldx	12, 24, 22
+	xor	14, 14, 9
+	xor	15, 15, 10
+	ldx	10, 24, 20
+	rldicl	16, 15, 60, 60
+	rldic	17, 15, 2, 58
+	srdi	15, 15, 8
+	rldimi	15, 14, 56, 0
+	srdi	14, 14, 8
+	lwzx	9, 17, 8
+	xor	16, 16, 10
+	xor	14, 14, 11
+	rlwinm	16, 16, 2, 26, 29
+	lwzx	10, 16, 7
+	xor	15, 15, 12
+	sldi	9, 9, 32
+	xor	14, 14, 9
+	sldi	10, 10, 32
+	xor	14, 14, 10
+	# Byte 11
+	rlwinm	23, 19, 12, 24, 27
+	rlwinm	24, 19, 8, 24, 27
+	ldx	9, 23, 4
+	ldx	10, 23, 20
+	ldx	11, 24, 21
+	ldx	12, 24, 22
+	xor	14, 14, 9
+	xor	15, 15, 10
+	ldx	10, 24, 20
+	rldicl	16, 15, 60, 60
+	rldic	17, 15, 2, 58
+	srdi	15, 15, 8
+	rldimi	15, 14, 56, 0
+	srdi	14, 14, 8
+	lwzx	9, 17, 8
+	xor	16, 16, 10
+	xor	14, 14, 11
+	rlwinm	16, 16, 2, 26, 29
+	lwzx	10, 16, 7
+	xor	15, 15, 12
+	sldi	9, 9, 32
+	xor	14, 14, 9
+	sldi	10, 10, 32
+	xor	14, 14, 10
+	# Byte 10
+	rlwinm	23, 19, 20, 24, 27
+	rlwinm	24, 19, 16, 24, 27
+	ldx	9, 23, 4
+	ldx	10, 23, 20
+	ldx	11, 24, 21
+	ldx	12, 24, 22
+	xor	14, 14, 9
+	xor	15, 15, 10
+	ldx	10, 24, 20
+	rldicl	16, 15, 60, 60
+	rldic	17, 15, 2, 58
+	srdi	15, 15, 8
+	rldimi	15, 14, 56, 0
+	srdi	14, 14, 8
+	lwzx	9, 17, 8
+	xor	16, 16, 10
+	xor	14, 14, 11
+	rlwinm	16, 16, 2, 26, 29
+	lwzx	10, 16, 7
+	xor	15, 15, 12
+	sldi	9, 9, 32
+	xor	14, 14, 9
+	sldi	10, 10, 32
+	xor	14, 14, 10
+	# Byte 9
+	rlwinm	23, 19, 28, 24, 27
+	rlwinm	24, 19, 24, 24, 27
+	ldx	9, 23, 4
+	ldx	10, 23, 20
+	ldx	11, 24, 21
+	ldx	12, 24, 22
+	xor	14, 14, 9
+	xor	15, 15, 10
+	ldx	10, 24, 20
+	rldicl	16, 15, 60, 60
+	rldic	17, 15, 2, 58
+	srdi	15, 15, 8
+	rldimi	15, 14, 56, 0
+	srdi	14, 14, 8
+	lwzx	9, 17, 8
+	xor	16, 16, 10
+	xor	14, 14, 11
+	rlwinm	16, 16, 2, 26, 29
+	lwzx	10, 16, 7
+	xor	15, 15, 12
+	sldi	9, 9, 32
+	xor	14, 14, 9
+	sldi	10, 10, 32
+	xor	14, 14, 10
+	# Byte 8
+	rlwinm	23, 19, 4, 24, 27
+	rlwinm	24, 19, 0, 24, 27
+	ldx	9, 23, 4
+	ldx	10, 23, 20
+	ldx	11, 24, 21
+	ldx	12, 24, 22
+	xor	14, 14, 9
+	xor	15, 15, 10
+	ldx	10, 24, 20
+	rldicl	16, 15, 60, 60
+	rldic	17, 15, 2, 58
+	srdi	15, 15, 8
+	rldimi	15, 14, 56, 0
+	srdi	14, 14, 8
+	lwzx	9, 17, 8
+	xor	16, 16, 10
+	xor	14, 14, 11
+	rlwinm	16, 16, 2, 26, 29
+	lwzx	10, 16, 7
+	xor	15, 15, 12
+	sldi	9, 9, 32
+	xor	14, 14, 9
+	sldi	10, 10, 32
+	xor	14, 14, 10
+	rldicr	0, 18, 32, 63
+	# Byte 7
+	rlwinm	23, 0, 12, 24, 27
+	rlwinm	24, 0, 8, 24, 27
+	ldx	9, 23, 4
+	ldx	10, 23, 20
+	ldx	11, 24, 21
+	ldx	12, 24, 22
+	xor	14, 14, 9
+	xor	15, 15, 10
+	ldx	10, 24, 20
+	rldicl	16, 15, 60, 60
+	rldic	17, 15, 2, 58
+	srdi	15, 15, 8
+	rldimi	15, 14, 56, 0
+	srdi	14, 14, 8
+	lwzx	9, 17, 8
+	xor	16, 16, 10
+	xor	14, 14, 11
+	rlwinm	16, 16, 2, 26, 29
+	lwzx	10, 16, 7
+	xor	15, 15, 12
+	sldi	9, 9, 32
+	xor	14, 14, 9
+	sldi	10, 10, 32
+	xor	14, 14, 10
+	# Byte 6
+	rlwinm	23, 0, 20, 24, 27
+	rlwinm	24, 0, 16, 24, 27
+	ldx	9, 23, 4
+	ldx	10, 23, 20
+	ldx	11, 24, 21
+	ldx	12, 24, 22
+	xor	14, 14, 9
+	xor	15, 15, 10
+	ldx	10, 24, 20
+	rldicl	16, 15, 60, 60
+	rldic	17, 15, 2, 58
+	srdi	15, 15, 8
+	rldimi	15, 14, 56, 0
+	srdi	14, 14, 8
+	lwzx	9, 17, 8
+	xor	16, 16, 10
+	xor	14, 14, 11
+	rlwinm	16, 16, 2, 26, 29
+	lwzx	10, 16, 7
+	xor	15, 15, 12
+	sldi	9, 9, 32
+	xor	14, 14, 9
+	sldi	10, 10, 32
+	xor	14, 14, 10
+	# Byte 5
+	rlwinm	23, 0, 28, 24, 27
+	rlwinm	24, 0, 24, 24, 27
+	ldx	9, 23, 4
+	ldx	10, 23, 20
+	ldx	11, 24, 21
+	ldx	12, 24, 22
+	xor	14, 14, 9
+	xor	15, 15, 10
+	ldx	10, 24, 20
+	rldicl	16, 15, 60, 60
+	rldic	17, 15, 2, 58
+	srdi	15, 15, 8
+	rldimi	15, 14, 56, 0
+	srdi	14, 14, 8
+	lwzx	9, 17, 8
+	xor	16, 16, 10
+	xor	14, 14, 11
+	rlwinm	16, 16, 2, 26, 29
+	lwzx	10, 16, 7
+	xor	15, 15, 12
+	sldi	9, 9, 32
+	xor	14, 14, 9
+	sldi	10, 10, 32
+	xor	14, 14, 10
+	# Byte 4
+	rlwinm	23, 0, 4, 24, 27
+	rlwinm	24, 0, 0, 24, 27
+	ldx	9, 23, 4
+	ldx	10, 23, 20
+	ldx	11, 24, 21
+	ldx	12, 24, 22
+	xor	14, 14, 9
+	xor	15, 15, 10
+	ldx	10, 24, 20
+	rldicl	16, 15, 60, 60
+	rldic	17, 15, 2, 58
+	srdi	15, 15, 8
+	rldimi	15, 14, 56, 0
+	srdi	14, 14, 8
+	lwzx	9, 17, 8
+	xor	16, 16, 10
+	xor	14, 14, 11
+	rlwinm	16, 16, 2, 26, 29
+	lwzx	10, 16, 7
+	xor	15, 15, 12
+	sldi	9, 9, 32
+	xor	14, 14, 9
+	sldi	10, 10, 32
+	xor	14, 14, 10
+	# Byte 3
+	rlwinm	23, 18, 12, 24, 27
+	rlwinm	24, 18, 8, 24, 27
+	ldx	9, 23, 4
+	ldx	10, 23, 20
+	ldx	11, 24, 21
+	ldx	12, 24, 22
+	xor	14, 14, 9
+	xor	15, 15, 10
+	ldx	10, 24, 20
+	rldicl	16, 15, 60, 60
+	rldic	17, 15, 2, 58
+	srdi	15, 15, 8
+	rldimi	15, 14, 56, 0
+	srdi	14, 14, 8
+	lwzx	9, 17, 8
+	xor	16, 16, 10
+	xor	14, 14, 11
+	rlwinm	16, 16, 2, 26, 29
+	lwzx	10, 16, 7
+	xor	15, 15, 12
+	sldi	9, 9, 32
+	xor	14, 14, 9
+	sldi	10, 10, 32
+	xor	14, 14, 10
+	# Byte 2
+	rlwinm	23, 18, 20, 24, 27
+	rlwinm	24, 18, 16, 24, 27
+	ldx	9, 23, 4
+	ldx	10, 23, 20
+	ldx	11, 24, 21
+	ldx	12, 24, 22
+	xor	14, 14, 9
+	xor	15, 15, 10
+	ldx	10, 24, 20
+	rldicl	16, 15, 60, 60
+	rldic	17, 15, 2, 58
+	srdi	15, 15, 8
+	rldimi	15, 14, 56, 0
+	srdi	14, 14, 8
+	lwzx	9, 17, 8
+	xor	16, 16, 10
+	xor	14, 14, 11
+	rlwinm	16, 16, 2, 26, 29
+	lwzx	10, 16, 7
+	xor	15, 15, 12
+	sldi	9, 9, 32
+	xor	14, 14, 9
+	sldi	10, 10, 32
+	xor	14, 14, 10
+	# Byte 1
+	rlwinm	23, 18, 28, 24, 27
+	rlwinm	24, 18, 24, 24, 27
+	ldx	9, 23, 4
+	ldx	10, 23, 20
+	ldx	11, 24, 21
+	ldx	12, 24, 22
+	xor	14, 14, 9
+	xor	15, 15, 10
+	ldx	10, 24, 20
+	rldicl	16, 15, 60, 60
+	rldic	17, 15, 2, 58
+	srdi	15, 15, 8
+	rldimi	15, 14, 56, 0
+	srdi	14, 14, 8
+	lwzx	9, 17, 8
+	xor	16, 16, 10
+	xor	14, 14, 11
+	rlwinm	16, 16, 2, 26, 29
+	lwzx	10, 16, 7
+	xor	15, 15, 12
+	sldi	9, 9, 32
+	xor	14, 14, 9
+	sldi	10, 10, 32
+	xor	14, 14, 10
+	# Byte 0
+	rlwinm	23, 18, 4, 24, 27
+	rlwinm	24, 18, 0, 24, 27
+	ldx	9, 23, 4
+	ldx	10, 23, 20
+	ldx	11, 24, 4
+	ldx	12, 24, 20
+	xor	14, 14, 9
+	xor	15, 15, 10
+	rldic	16, 15, 2, 58
+	srdi	15, 15, 4
+	lwzx	9, 16, 7
+	rldimi	15, 14, 60, 0
+	srdi	14, 14, 4
+	xor	19, 15, 12
+	xor	14, 14, 11
+	sldi	9, 9, 32
+	xor	18, 14, 9
+	addi	5, 5, 16
+	addic.	6, 6, -16
+	std	18, 0(3)
+	std	19, 8(3)
+	bne	L_GCM_gmult_len_start_block
+	ld	0, 96(1)
+	mtlr	0
+	ld	14, 0(1)
+	ld	15, 8(1)
+	ld	16, 16(1)
+	ld	17, 24(1)
+	ld	18, 32(1)
+	ld	19, 40(1)
+	ld	20, 48(1)
+	ld	21, 56(1)
+	ld	22, 64(1)
+	ld	23, 72(1)
+	ld	24, 80(1)
+	ld	25, 88(1)
+	addi	1, 1, 0x68
+	blr
+#ifndef __APPLE__
+	.size	GCM_gmult_len,.-GCM_gmult_len
+#endif /* __APPLE__ */
+#endif /* GCM_TABLE_4BIT */
+#ifdef GCM_TABLE
+#ifndef __APPLE__
+	.text
+	.section	.rodata
+	.type	L_GCM_gmult_len_r, @object
+	.size	L_GCM_gmult_len_r, 512
+#else
+	.section	__DATA,__data
+#endif /* __APPLE__ */
+#ifndef __APPLE__
+	.align	4
+#else
+	.p2align	4
+#endif /* __APPLE__ */
+L_GCM_gmult_len_r:
+	.byte	0x0
+	.byte	0x0
+	.byte	0xc2
+	.byte	0x1
+	.byte	0x84
+	.byte	0x3
+	.byte	0x46
+	.byte	0x2
+	.byte	0x8
+	.byte	0x7
+	.byte	0xca
+	.byte	0x6
+	.byte	0x8c
+	.byte	0x4
+	.byte	0x4e
+	.byte	0x5
+	.byte	0x10
+	.byte	0xe
+	.byte	0xd2
+	.byte	0xf
+	.byte	0x94
+	.byte	0xd
+	.byte	0x56
+	.byte	0xc
+	.byte	0x18
+	.byte	0x9
+	.byte	0xda
+	.byte	0x8
+	.byte	0x9c
+	.byte	0xa
+	.byte	0x5e
+	.byte	0xb
+	.byte	0x20
+	.byte	0x1c
+	.byte	0xe2
+	.byte	0x1d
+	.byte	0xa4
+	.byte	0x1f
+	.byte	0x66
+	.byte	0x1e
+	.byte	0x28
+	.byte	0x1b
+	.byte	0xea
+	.byte	0x1a
+	.byte	0xac
+	.byte	0x18
+	.byte	0x6e
+	.byte	0x19
+	.byte	0x30
+	.byte	0x12
+	.byte	0xf2
+	.byte	0x13
+	.byte	0xb4
+	.byte	0x11
+	.byte	0x76
+	.byte	0x10
+	.byte	0x38
+	.byte	0x15
+	.byte	0xfa
+	.byte	0x14
+	.byte	0xbc
+	.byte	0x16
+	.byte	0x7e
+	.byte	0x17
+	.byte	0x40
+	.byte	0x38
+	.byte	0x82
+	.byte	0x39
+	.byte	0xc4
+	.byte	0x3b
+	.byte	0x6
+	.byte	0x3a
+	.byte	0x48
+	.byte	0x3f
+	.byte	0x8a
+	.byte	0x3e
+	.byte	0xcc
+	.byte	0x3c
+	.byte	0xe
+	.byte	0x3d
+	.byte	0x50
+	.byte	0x36
+	.byte	0x92
+	.byte	0x37
+	.byte	0xd4
+	.byte	0x35
+	.byte	0x16
+	.byte	0x34
+	.byte	0x58
+	.byte	0x31
+	.byte	0x9a
+	.byte	0x30
+	.byte	0xdc
+	.byte	0x32
+	.byte	0x1e
+	.byte	0x33
+	.byte	0x60
+	.byte	0x24
+	.byte	0xa2
+	.byte	0x25
+	.byte	0xe4
+	.byte	0x27
+	.byte	0x26
+	.byte	0x26
+	.byte	0x68
+	.byte	0x23
+	.byte	0xaa
+	.byte	0x22
+	.byte	0xec
+	.byte	0x20
+	.byte	0x2e
+	.byte	0x21
+	.byte	0x70
+	.byte	0x2a
+	.byte	0xb2
+	.byte	0x2b
+	.byte	0xf4
+	.byte	0x29
+	.byte	0x36
+	.byte	0x28
+	.byte	0x78
+	.byte	0x2d
+	.byte	0xba
+	.byte	0x2c
+	.byte	0xfc
+	.byte	0x2e
+	.byte	0x3e
+	.byte	0x2f
+	.byte	0x80
+	.byte	0x70
+	.byte	0x42
+	.byte	0x71
+	.byte	0x4
+	.byte	0x73
+	.byte	0xc6
+	.byte	0x72
+	.byte	0x88
+	.byte	0x77
+	.byte	0x4a
+	.byte	0x76
+	.byte	0xc
+	.byte	0x74
+	.byte	0xce
+	.byte	0x75
+	.byte	0x90
+	.byte	0x7e
+	.byte	0x52
+	.byte	0x7f
+	.byte	0x14
+	.byte	0x7d
+	.byte	0xd6
+	.byte	0x7c
+	.byte	0x98
+	.byte	0x79
+	.byte	0x5a
+	.byte	0x78
+	.byte	0x1c
+	.byte	0x7a
+	.byte	0xde
+	.byte	0x7b
+	.byte	0xa0
+	.byte	0x6c
+	.byte	0x62
+	.byte	0x6d
+	.byte	0x24
+	.byte	0x6f
+	.byte	0xe6
+	.byte	0x6e
+	.byte	0xa8
+	.byte	0x6b
+	.byte	0x6a
+	.byte	0x6a
+	.byte	0x2c
+	.byte	0x68
+	.byte	0xee
+	.byte	0x69
+	.byte	0xb0
+	.byte	0x62
+	.byte	0x72
+	.byte	0x63
+	.byte	0x34
+	.byte	0x61
+	.byte	0xf6
+	.byte	0x60
+	.byte	0xb8
+	.byte	0x65
+	.byte	0x7a
+	.byte	0x64
+	.byte	0x3c
+	.byte	0x66
+	.byte	0xfe
+	.byte	0x67
+	.byte	0xc0
+	.byte	0x48
+	.byte	0x2
+	.byte	0x49
+	.byte	0x44
+	.byte	0x4b
+	.byte	0x86
+	.byte	0x4a
+	.byte	0xc8
+	.byte	0x4f
+	.byte	0xa
+	.byte	0x4e
+	.byte	0x4c
+	.byte	0x4c
+	.byte	0x8e
+	.byte	0x4d
+	.byte	0xd0
+	.byte	0x46
+	.byte	0x12
+	.byte	0x47
+	.byte	0x54
+	.byte	0x45
+	.byte	0x96
+	.byte	0x44
+	.byte	0xd8
+	.byte	0x41
+	.byte	0x1a
+	.byte	0x40
+	.byte	0x5c
+	.byte	0x42
+	.byte	0x9e
+	.byte	0x43
+	.byte	0xe0
+	.byte	0x54
+	.byte	0x22
+	.byte	0x55
+	.byte	0x64
+	.byte	0x57
+	.byte	0xa6
+	.byte	0x56
+	.byte	0xe8
+	.byte	0x53
+	.byte	0x2a
+	.byte	0x52
+	.byte	0x6c
+	.byte	0x50
+	.byte	0xae
+	.byte	0x51
+	.byte	0xf0
+	.byte	0x5a
+	.byte	0x32
+	.byte	0x5b
+	.byte	0x74
+	.byte	0x59
+	.byte	0xb6
+	.byte	0x58
+	.byte	0xf8
+	.byte	0x5d
+	.byte	0x3a
+	.byte	0x5c
+	.byte	0x7c
+	.byte	0x5e
+	.byte	0xbe
+	.byte	0x5f
+	.byte	0x0
+	.byte	0xe1
+	.byte	0xc2
+	.byte	0xe0
+	.byte	0x84
+	.byte	0xe2
+	.byte	0x46
+	.byte	0xe3
+	.byte	0x8
+	.byte	0xe6
+	.byte	0xca
+	.byte	0xe7
+	.byte	0x8c
+	.byte	0xe5
+	.byte	0x4e
+	.byte	0xe4
+	.byte	0x10
+	.byte	0xef
+	.byte	0xd2
+	.byte	0xee
+	.byte	0x94
+	.byte	0xec
+	.byte	0x56
+	.byte	0xed
+	.byte	0x18
+	.byte	0xe8
+	.byte	0xda
+	.byte	0xe9
+	.byte	0x9c
+	.byte	0xeb
+	.byte	0x5e
+	.byte	0xea
+	.byte	0x20
+	.byte	0xfd
+	.byte	0xe2
+	.byte	0xfc
+	.byte	0xa4
+	.byte	0xfe
+	.byte	0x66
+	.byte	0xff
+	.byte	0x28
+	.byte	0xfa
+	.byte	0xea
+	.byte	0xfb
+	.byte	0xac
+	.byte	0xf9
+	.byte	0x6e
+	.byte	0xf8
+	.byte	0x30
+	.byte	0xf3
+	.byte	0xf2
+	.byte	0xf2
+	.byte	0xb4
+	.byte	0xf0
+	.byte	0x76
+	.byte	0xf1
+	.byte	0x38
+	.byte	0xf4
+	.byte	0xfa
+	.byte	0xf5
+	.byte	0xbc
+	.byte	0xf7
+	.byte	0x7e
+	.byte	0xf6
+	.byte	0x40
+	.byte	0xd9
+	.byte	0x82
+	.byte	0xd8
+	.byte	0xc4
+	.byte	0xda
+	.byte	0x6
+	.byte	0xdb
+	.byte	0x48
+	.byte	0xde
+	.byte	0x8a
+	.byte	0xdf
+	.byte	0xcc
+	.byte	0xdd
+	.byte	0xe
+	.byte	0xdc
+	.byte	0x50
+	.byte	0xd7
+	.byte	0x92
+	.byte	0xd6
+	.byte	0xd4
+	.byte	0xd4
+	.byte	0x16
+	.byte	0xd5
+	.byte	0x58
+	.byte	0xd0
+	.byte	0x9a
+	.byte	0xd1
+	.byte	0xdc
+	.byte	0xd3
+	.byte	0x1e
+	.byte	0xd2
+	.byte	0x60
+	.byte	0xc5
+	.byte	0xa2
+	.byte	0xc4
+	.byte	0xe4
+	.byte	0xc6
+	.byte	0x26
+	.byte	0xc7
+	.byte	0x68
+	.byte	0xc2
+	.byte	0xaa
+	.byte	0xc3
+	.byte	0xec
+	.byte	0xc1
+	.byte	0x2e
+	.byte	0xc0
+	.byte	0x70
+	.byte	0xcb
+	.byte	0xb2
+	.byte	0xca
+	.byte	0xf4
+	.byte	0xc8
+	.byte	0x36
+	.byte	0xc9
+	.byte	0x78
+	.byte	0xcc
+	.byte	0xba
+	.byte	0xcd
+	.byte	0xfc
+	.byte	0xcf
+	.byte	0x3e
+	.byte	0xce
+	.byte	0x80
+	.byte	0x91
+	.byte	0x42
+	.byte	0x90
+	.byte	0x4
+	.byte	0x92
+	.byte	0xc6
+	.byte	0x93
+	.byte	0x88
+	.byte	0x96
+	.byte	0x4a
+	.byte	0x97
+	.byte	0xc
+	.byte	0x95
+	.byte	0xce
+	.byte	0x94
+	.byte	0x90
+	.byte	0x9f
+	.byte	0x52
+	.byte	0x9e
+	.byte	0x14
+	.byte	0x9c
+	.byte	0xd6
+	.byte	0x9d
+	.byte	0x98
+	.byte	0x98
+	.byte	0x5a
+	.byte	0x99
+	.byte	0x1c
+	.byte	0x9b
+	.byte	0xde
+	.byte	0x9a
+	.byte	0xa0
+	.byte	0x8d
+	.byte	0x62
+	.byte	0x8c
+	.byte	0x24
+	.byte	0x8e
+	.byte	0xe6
+	.byte	0x8f
+	.byte	0xa8
+	.byte	0x8a
+	.byte	0x6a
+	.byte	0x8b
+	.byte	0x2c
+	.byte	0x89
+	.byte	0xee
+	.byte	0x88
+	.byte	0xb0
+	.byte	0x83
+	.byte	0x72
+	.byte	0x82
+	.byte	0x34
+	.byte	0x80
+	.byte	0xf6
+	.byte	0x81
+	.byte	0xb8
+	.byte	0x84
+	.byte	0x7a
+	.byte	0x85
+	.byte	0x3c
+	.byte	0x87
+	.byte	0xfe
+	.byte	0x86
+	.byte	0xc0
+	.byte	0xa9
+	.byte	0x2
+	.byte	0xa8
+	.byte	0x44
+	.byte	0xaa
+	.byte	0x86
+	.byte	0xab
+	.byte	0xc8
+	.byte	0xae
+	.byte	0xa
+	.byte	0xaf
+	.byte	0x4c
+	.byte	0xad
+	.byte	0x8e
+	.byte	0xac
+	.byte	0xd0
+	.byte	0xa7
+	.byte	0x12
+	.byte	0xa6
+	.byte	0x54
+	.byte	0xa4
+	.byte	0x96
+	.byte	0xa5
+	.byte	0xd8
+	.byte	0xa0
+	.byte	0x1a
+	.byte	0xa1
+	.byte	0x5c
+	.byte	0xa3
+	.byte	0x9e
+	.byte	0xa2
+	.byte	0xe0
+	.byte	0xb5
+	.byte	0x22
+	.byte	0xb4
+	.byte	0x64
+	.byte	0xb6
+	.byte	0xa6
+	.byte	0xb7
+	.byte	0xe8
+	.byte	0xb2
+	.byte	0x2a
+	.byte	0xb3
+	.byte	0x6c
+	.byte	0xb1
+	.byte	0xae
+	.byte	0xb0
+	.byte	0xf0
+	.byte	0xbb
+	.byte	0x32
+	.byte	0xba
+	.byte	0x74
+	.byte	0xb8
+	.byte	0xb6
+	.byte	0xb9
+	.byte	0xf8
+	.byte	0xbc
+	.byte	0x3a
+	.byte	0xbd
+	.byte	0x7c
+	.byte	0xbf
+	.byte	0xbe
+	.byte	0xbe
+#ifndef __APPLE__
+.text
+.globl	GCM_gmult_len
+.type	GCM_gmult_len,@function
+.align	16
+GCM_gmult_len:
+#else
+.section	__TEXT,__text
+.globl	_GCM_gmult_len
+.p2align	4
+_GCM_gmult_len:
+#endif /* __APPLE__ */
+	stdu	1, -56(1)
+	mflr	0
+	std	0, 48(1)
+	std	14, 0(1)
+	std	15, 8(1)
+	std	16, 16(1)
+	std	17, 24(1)
+	std	18, 32(1)
+	std	19, 40(1)
+#if defined(__ARCH_PWR8) || (defined(_ARCH_PWR8))
+	addis	7, 2, L_GCM_gmult_len_r@toc@ha
+	addi	7, 7, L_GCM_gmult_len_r@toc@l
+#else
+	lis	7, L_GCM_gmult_len_r@ha
+	la	7, L_GCM_gmult_len_r@l(7)
+#endif /* defined(__ARCH_PWR8) || (defined(_ARCH_PWR8)) */
+	addi	0, 4, 8
+	li	17, 8
+L_GCM_gmult_len_start_block:
+	ldbrx	8, 0, 3
+	ldbrx	9, 17, 3
+	ldbrx	10, 0, 5
+	ldbrx	11, 17, 5
+	xor	8, 8, 10
+	xor	9, 9, 11
+	rldicr	19, 9, 32, 63
+	rldicr	18, 8, 32, 63
+	rlwinm	16, 19, 12, 20, 27
+	# Byte 15
+	ldx	12, 16, 4
+	ldx	14, 16, 0
+	rldicl	11, 12, 8, 56
+	rldicl	15, 14, 9, 55
+	sldi	14, 14, 8
+	andi.   15, 15, 510
+	sldi	12, 12, 8
+	lhzx	10, 15, 7
+	xor	14, 14, 11
+	rlwinm	16, 19, 20, 20, 27
+	xor	12, 12, 10
+	# Byte 14
+	ldx	10, 16, 4
+	ldx	11, 16, 0
+	xor	12, 12, 10
+	xor	14, 14, 11
+	rldicl	11, 12, 8, 56
+	rldicl	15, 14, 9, 55
+	sldi	14, 14, 8
+	andi.   15, 15, 510
+	sldi	12, 12, 8
+	lhzx	10, 15, 7
+	xor	14, 14, 11
+	rlwinm	16, 19, 28, 20, 27
+	xor	12, 12, 10
+	# Byte 13
+	ldx	10, 16, 4
+	ldx	11, 16, 0
+	xor	12, 12, 10
+	xor	14, 14, 11
+	rldicl	11, 12, 8, 56
+	rldicl	15, 14, 9, 55
+	sldi	14, 14, 8
+	andi.   15, 15, 510
+	sldi	12, 12, 8
+	lhzx	10, 15, 7
+	xor	14, 14, 11
+	rlwinm	16, 19, 4, 20, 27
+	xor	12, 12, 10
+	# Byte 12
+	ldx	10, 16, 4
+	ldx	11, 16, 0
+	xor	12, 12, 10
+	xor	14, 14, 11
+	rldicl	11, 12, 8, 56
+	rldicl	15, 14, 9, 55
+	sldi	14, 14, 8
+	andi.   15, 15, 510
+	sldi	12, 12, 8
+	lhzx	10, 15, 7
+	xor	14, 14, 11
+	rlwinm	16, 9, 12, 20, 27
+	xor	12, 12, 10
+	# Byte 11
+	ldx	10, 16, 4
+	ldx	11, 16, 0
+	xor	12, 12, 10
+	xor	14, 14, 11
+	rldicl	11, 12, 8, 56
+	rldicl	15, 14, 9, 55
+	sldi	14, 14, 8
+	andi.   15, 15, 510
+	sldi	12, 12, 8
+	lhzx	10, 15, 7
+	xor	14, 14, 11
+	rlwinm	16, 9, 20, 20, 27
+	xor	12, 12, 10
+	# Byte 10
+	ldx	10, 16, 4
+	ldx	11, 16, 0
+	xor	12, 12, 10
+	xor	14, 14, 11
+	rldicl	11, 12, 8, 56
+	rldicl	15, 14, 9, 55
+	sldi	14, 14, 8
+	andi.   15, 15, 510
+	sldi	12, 12, 8
+	lhzx	10, 15, 7
+	xor	14, 14, 11
+	rlwinm	16, 9, 28, 20, 27
+	xor	12, 12, 10
+	# Byte 9
+	ldx	10, 16, 4
+	ldx	11, 16, 0
+	xor	12, 12, 10
+	xor	14, 14, 11
+	rldicl	11, 12, 8, 56
+	rldicl	15, 14, 9, 55
+	sldi	14, 14, 8
+	andi.   15, 15, 510
+	sldi	12, 12, 8
+	lhzx	10, 15, 7
+	xor	14, 14, 11
+	rlwinm	16, 9, 4, 20, 27
+	xor	12, 12, 10
+	# Byte 8
+	ldx	10, 16, 4
+	ldx	11, 16, 0
+	xor	12, 12, 10
+	xor	14, 14, 11
+	rldicl	11, 12, 8, 56
+	rldicl	15, 14, 9, 55
+	sldi	14, 14, 8
+	andi.   15, 15, 510
+	sldi	12, 12, 8
+	lhzx	10, 15, 7
+	xor	14, 14, 11
+	rlwinm	16, 18, 12, 20, 27
+	xor	12, 12, 10
+	# Byte 7
+	ldx	10, 16, 4
+	ldx	11, 16, 0
+	xor	12, 12, 10
+	xor	14, 14, 11
+	rldicl	11, 12, 8, 56
+	rldicl	15, 14, 9, 55
+	sldi	14, 14, 8
+	andi.   15, 15, 510
+	sldi	12, 12, 8
+	lhzx	10, 15, 7
+	xor	14, 14, 11
+	rlwinm	16, 18, 20, 20, 27
+	xor	12, 12, 10
+	# Byte 6
+	ldx	10, 16, 4
+	ldx	11, 16, 0
+	xor	12, 12, 10
+	xor	14, 14, 11
+	rldicl	11, 12, 8, 56
+	rldicl	15, 14, 9, 55
+	sldi	14, 14, 8
+	andi.   15, 15, 510
+	sldi	12, 12, 8
+	lhzx	10, 15, 7
+	xor	14, 14, 11
+	rlwinm	16, 18, 28, 20, 27
+	xor	12, 12, 10
+	# Byte 5
+	ldx	10, 16, 4
+	ldx	11, 16, 0
+	xor	12, 12, 10
+	xor	14, 14, 11
+	rldicl	11, 12, 8, 56
+	rldicl	15, 14, 9, 55
+	sldi	14, 14, 8
+	andi.   15, 15, 510
+	sldi	12, 12, 8
+	lhzx	10, 15, 7
+	xor	14, 14, 11
+	rlwinm	16, 18, 4, 20, 27
+	xor	12, 12, 10
+	# Byte 4
+	ldx	10, 16, 4
+	ldx	11, 16, 0
+	xor	12, 12, 10
+	xor	14, 14, 11
+	rldicl	11, 12, 8, 56
+	rldicl	15, 14, 9, 55
+	sldi	14, 14, 8
+	andi.   15, 15, 510
+	sldi	12, 12, 8
+	lhzx	10, 15, 7
+	xor	14, 14, 11
+	rlwinm	16, 8, 12, 20, 27
+	xor	12, 12, 10
+	# Byte 3
+	ldx	10, 16, 4
+	ldx	11, 16, 0
+	xor	12, 12, 10
+	xor	14, 14, 11
+	rldicl	11, 12, 8, 56
+	rldicl	15, 14, 9, 55
+	sldi	14, 14, 8
+	andi.   15, 15, 510
+	sldi	12, 12, 8
+	lhzx	10, 15, 7
+	xor	14, 14, 11
+	rlwinm	16, 8, 20, 20, 27
+	xor	12, 12, 10
+	# Byte 2
+	ldx	10, 16, 4
+	ldx	11, 16, 0
+	xor	12, 12, 10
+	xor	14, 14, 11
+	rldicl	11, 12, 8, 56
+	rldicl	15, 14, 9, 55
+	sldi	14, 14, 8
+	andi.   15, 15, 510
+	sldi	12, 12, 8
+	lhzx	10, 15, 7
+	xor	14, 14, 11
+	rlwinm	16, 8, 28, 20, 27
+	xor	12, 12, 10
+	# Byte 1
+	ldx	10, 16, 4
+	ldx	11, 16, 0
+	xor	12, 12, 10
+	xor	14, 14, 11
+	rldicl	11, 12, 8, 56
+	rldicl	15, 14, 9, 55
+	sldi	14, 14, 8
+	andi.   15, 15, 510
+	sldi	12, 12, 8
+	lhzx	10, 15, 7
+	xor	14, 14, 11
+	rlwinm	16, 8, 4, 20, 27
+	xor	12, 12, 10
+	# Byte 0
+	ldx	10, 16, 4
+	ldx	11, 16, 0
+	addi	5, 5, 16
+	addic.	6, 6, -16
+	xor	8, 12, 10
+	xor	9, 14, 11
+	stdbrx	8, 0, 3
+	stdbrx	9, 17, 3
+	bne	L_GCM_gmult_len_start_block
+	ld	0, 48(1)
+	mtlr	0
+	ld	14, 0(1)
+	ld	15, 8(1)
+	ld	16, 16(1)
+	ld	17, 24(1)
+	ld	18, 32(1)
+	ld	19, 40(1)
+	addi	1, 1, 56
+	blr
+#ifndef __APPLE__
+	.size	GCM_gmult_len,.-GCM_gmult_len
+#endif /* __APPLE__ */
+#endif /* GCM_TABLE */
+#endif /* HAVE_AESGCM */
+#endif /* !defined(NO_AES) && defined(WOLFSSL_PPC64_ASM) */
+#endif /* WOLFSSL_PPC64_ASM */
+
+#if defined(__linux__) && defined(__ELF__)
+.section	.note.GNU-stack,"",%progbits
+#endif
+#endif /* !WOLFSSL_PPC64_ASM_INLINE */
diff --git a/wolfcrypt/src/port/ppc64/ppc64-aes-asm_c.c b/wolfcrypt/src/port/ppc64/ppc64-aes-asm_c.c
new file mode 100644
index 0000000000..d0c25755aa
--- /dev/null
+++ b/wolfcrypt/src/port/ppc64/ppc64-aes-asm_c.c
@@ -0,0 +1,6648 @@
+/* ppc64-aes-asm
+ *
+ * Copyright (C) 2006-2025 wolfSSL Inc.
+ *
+ * This file is part of wolfSSL.
+ *
+ * wolfSSL is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * wolfSSL is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1335, USA
+ */
+
+#include <wolfssl/wolfcrypt/libwolfssl_sources_asm.h>
+#include <wolfssl/wolfcrypt/error-crypt.h>
+
+/* Generated using (from wolfssl):
+ *   cd ../scripts
+ *   ruby ./aes/aes.rb ppc64 \
+ *       ../wolfssl/wolfcrypt/src/port/ppc64/ppc64-aes-asm.c
+ */
+#ifdef WOLFSSL_PPC64_ASM
+#include <stdint.h>
+#include <wolfssl/wolfcrypt/libwolfssl_sources.h>
+#ifdef WOLFSSL_PPC64_ASM_INLINE
+
+#ifdef __IAR_SYSTEMS_ICC__
+#define __asm__        asm
+#define __volatile__   volatile
+#define WOLFSSL_NO_VAR_ASSIGN_REG
+#endif /* __IAR_SYSTEMS_ICC__ */
+#ifdef __KEIL__
+#define __asm__        __asm
+#define __volatile__   volatile
+#endif /* __KEIL__ */
+#ifdef __ghs__
+#define __asm__        __asm
+#define __volatile__
+#define WOLFSSL_NO_VAR_ASSIGN_REG
+#endif /* __ghs__ */
+#include <wolfssl/wolfcrypt/aes.h>
+
+#if !defined(NO_AES) && defined(WOLFSSL_PPC64_ASM)
+#ifdef HAVE_AES_DECRYPT
+static const word32 L_AES_PPC64_td[] = {
+    0x5051f4a7, 0x537e4165, 0xc31a17a4, 0x963a275e,
+    0xcb3bab6b, 0xf11f9d45, 0xabacfa58, 0x934be303,
+    0x552030fa, 0xf6ad766d, 0x9188cc76, 0x25f5024c,
+    0xfc4fe5d7, 0xd7c52acb, 0x80263544, 0x8fb562a3,
+    0x49deb15a, 0x6725ba1b, 0x9845ea0e, 0xe15dfec0,
+    0x02c32f75, 0x12814cf0, 0xa38d4697, 0xc66bd3f9,
+    0xe7038f5f, 0x9515929c, 0xebbf6d7a, 0xda955259,
+    0x2dd4be83, 0xd3587421, 0x2949e069, 0x448ec9c8,
+    0x6a75c289, 0x78f48e79, 0x6b99583e, 0xdd27b971,
+    0xb6bee14f, 0x17f088ad, 0x66c920ac, 0xb47dce3a,
+    0x1863df4a, 0x82e51a31, 0x60975133, 0x4562537f,
+    0xe0b16477, 0x84bb6bae, 0x1cfe81a0, 0x94f9082b,
+    0x58704868, 0x198f45fd, 0x8794de6c, 0xb7527bf8,
+    0x23ab73d3, 0xe2724b02, 0x57e31f8f, 0x2a6655ab,
+    0x07b2eb28, 0x032fb5c2, 0x9a86c57b, 0xa5d33708,
+    0xf2302887, 0xb223bfa5, 0xba02036a, 0x5ced1682,
+    0x2b8acf1c, 0x92a779b4, 0xf0f307f2, 0xa14e69e2,
+    0xcd65daf4, 0xd50605be, 0x1fd13462, 0x8ac4a6fe,
+    0x9d342e53, 0xa0a2f355, 0x32058ae1, 0x75a4f6eb,
+    0x390b83ec, 0xaa4060ef, 0x065e719f, 0x51bd6e10,
+    0xf93e218a, 0x3d96dd06, 0xaedd3e05, 0x464de6bd,
+    0xb591548d, 0x0571c45d, 0x6f0406d4, 0xff605015,
+    0x241998fb, 0x97d6bde9, 0xcc894043, 0x7767d99e,
+    0xbdb0e842, 0x8807898b, 0x38e7195b, 0xdb79c8ee,
+    0x47a17c0a, 0xe97c420f, 0xc9f8841e, 0x00000000,
+    0x83098086, 0x48322bed, 0xac1e1170, 0x4e6c5a72,
+    0xfbfd0eff, 0x560f8538, 0x1e3daed5, 0x27362d39,
+    0x640a0fd9, 0x21685ca6, 0xd19b5b54, 0x3a24362e,
+    0xb10c0a67, 0x0f9357e7, 0xd2b4ee96, 0x9e1b9b91,
+    0x4f80c0c5, 0xa261dc20, 0x695a774b, 0x161c121a,
+    0x0ae293ba, 0xe5c0a02a, 0x433c22e0, 0x1d121b17,
+    0x0b0e090d, 0xadf28bc7, 0xb92db6a8, 0xc8141ea9,
+    0x8557f119, 0x4caf7507, 0xbbee99dd, 0xfda37f60,
+    0x9ff70126, 0xbc5c72f5, 0xc544663b, 0x345bfb7e,
+    0x768b4329, 0xdccb23c6, 0x68b6edfc, 0x63b8e4f1,
+    0xcad731dc, 0x10426385, 0x40139722, 0x2084c611,
+    0x7d854a24, 0xf8d2bb3d, 0x11aef932, 0x6dc729a1,
+    0x4b1d9e2f, 0xf3dcb230, 0xec0d8652, 0xd077c1e3,
+    0x6c2bb316, 0x99a970b9, 0xfa119448, 0x2247e964,
+    0xc4a8fc8c, 0x1aa0f03f, 0xd8567d2c, 0xef223390,
+    0xc787494e, 0xc1d938d1, 0xfe8ccaa2, 0x3698d40b,
+    0xcfa6f581, 0x28a57ade, 0x26dab78e, 0xa43fadbf,
+    0xe42c3a9d, 0x0d507892, 0x9b6a5fcc, 0x62547e46,
+    0xc2f68d13, 0xe890d8b8, 0x5e2e39f7, 0xf582c3af,
+    0xbe9f5d80, 0x7c69d093, 0xa96fd52d, 0xb3cf2512,
+    0x3bc8ac99, 0xa710187d, 0x6ee89c63, 0x7bdb3bbb,
+    0x09cd2678, 0xf46e5918, 0x01ec9ab7, 0xa8834f9a,
+    0x65e6956e, 0x7eaaffe6, 0x0821bccf, 0xe6ef15e8,
+    0xd9bae79b, 0xce4a6f36, 0xd4ea9f09, 0xd629b07c,
+    0xaf31a4b2, 0x312a3f23, 0x30c6a594, 0xc035a266,
+    0x37744ebc, 0xa6fc82ca, 0xb0e090d0, 0x1533a7d8,
+    0x4af10498, 0xf741ecda, 0x0e7fcd50, 0x2f1791f6,
+    0x8d764dd6, 0x4d43efb0, 0x54ccaa4d, 0xdfe49604,
+    0xe39ed1b5, 0x1b4c6a88, 0xb8c12c1f, 0x7f466551,
+    0x049d5eea, 0x5d018c35, 0x73fa8774, 0x2efb0b41,
+    0x5ab3671d, 0x5292dbd2, 0x33e91056, 0x136dd647,
+    0x8c9ad761, 0x7a37a10c, 0x8e59f814, 0x89eb133c,
+    0xeecea927, 0x35b761c9, 0xede11ce5, 0x3c7a47b1,
+    0x599cd2df, 0x3f55f273, 0x791814ce, 0xbf73c737,
+    0xea53f7cd, 0x5b5ffdaa, 0x14df3d6f, 0x867844db,
+    0x81caaff3, 0x3eb968c4, 0x2c382434, 0x5fc2a340,
+    0x72161dc3, 0x0cbce225, 0x8b283c49, 0x41ff0d95,
+    0x7139a801, 0xde080cb3, 0x9cd8b4e4, 0x906456c1,
+    0x617bcb84, 0x70d532b6, 0x74486c5c, 0x42d0b857,
+};
+
+#endif /* HAVE_AES_DECRYPT */
+#if defined(HAVE_AES_DECRYPT) || defined(HAVE_AES_CBC) || \
+    defined(HAVE_AESCCM) || defined(HAVE_AESGCM) || \
+    defined(WOLFSSL_AES_DIRECT) || defined(WOLFSSL_AES_COUNTER)
+static const word32 L_AES_PPC64_te[] = {
+    0xa5c66363, 0x84f87c7c, 0x99ee7777, 0x8df67b7b,
+    0x0dfff2f2, 0xbdd66b6b, 0xb1de6f6f, 0x5491c5c5,
+    0x50603030, 0x03020101, 0xa9ce6767, 0x7d562b2b,
+    0x19e7fefe, 0x62b5d7d7, 0xe64dabab, 0x9aec7676,
+    0x458fcaca, 0x9d1f8282, 0x4089c9c9, 0x87fa7d7d,
+    0x15effafa, 0xebb25959, 0xc98e4747, 0x0bfbf0f0,
+    0xec41adad, 0x67b3d4d4, 0xfd5fa2a2, 0xea45afaf,
+    0xbf239c9c, 0xf753a4a4, 0x96e47272, 0x5b9bc0c0,
+    0xc275b7b7, 0x1ce1fdfd, 0xae3d9393, 0x6a4c2626,
+    0x5a6c3636, 0x417e3f3f, 0x02f5f7f7, 0x4f83cccc,
+    0x5c683434, 0xf451a5a5, 0x34d1e5e5, 0x08f9f1f1,
+    0x93e27171, 0x73abd8d8, 0x53623131, 0x3f2a1515,
+    0x0c080404, 0x5295c7c7, 0x65462323, 0x5e9dc3c3,
+    0x28301818, 0xa1379696, 0x0f0a0505, 0xb52f9a9a,
+    0x090e0707, 0x36241212, 0x9b1b8080, 0x3ddfe2e2,
+    0x26cdebeb, 0x694e2727, 0xcd7fb2b2, 0x9fea7575,
+    0x1b120909, 0x9e1d8383, 0x74582c2c, 0x2e341a1a,
+    0x2d361b1b, 0xb2dc6e6e, 0xeeb45a5a, 0xfb5ba0a0,
+    0xf6a45252, 0x4d763b3b, 0x61b7d6d6, 0xce7db3b3,
+    0x7b522929, 0x3edde3e3, 0x715e2f2f, 0x97138484,
+    0xf5a65353, 0x68b9d1d1, 0x00000000, 0x2cc1eded,
+    0x60402020, 0x1fe3fcfc, 0xc879b1b1, 0xedb65b5b,
+    0xbed46a6a, 0x468dcbcb, 0xd967bebe, 0x4b723939,
+    0xde944a4a, 0xd4984c4c, 0xe8b05858, 0x4a85cfcf,
+    0x6bbbd0d0, 0x2ac5efef, 0xe54faaaa, 0x16edfbfb,
+    0xc5864343, 0xd79a4d4d, 0x55663333, 0x94118585,
+    0xcf8a4545, 0x10e9f9f9, 0x06040202, 0x81fe7f7f,
+    0xf0a05050, 0x44783c3c, 0xba259f9f, 0xe34ba8a8,
+    0xf3a25151, 0xfe5da3a3, 0xc0804040, 0x8a058f8f,
+    0xad3f9292, 0xbc219d9d, 0x48703838, 0x04f1f5f5,
+    0xdf63bcbc, 0xc177b6b6, 0x75afdada, 0x63422121,
+    0x30201010, 0x1ae5ffff, 0x0efdf3f3, 0x6dbfd2d2,
+    0x4c81cdcd, 0x14180c0c, 0x35261313, 0x2fc3ecec,
+    0xe1be5f5f, 0xa2359797, 0xcc884444, 0x392e1717,
+    0x5793c4c4, 0xf255a7a7, 0x82fc7e7e, 0x477a3d3d,
+    0xacc86464, 0xe7ba5d5d, 0x2b321919, 0x95e67373,
+    0xa0c06060, 0x98198181, 0xd19e4f4f, 0x7fa3dcdc,
+    0x66442222, 0x7e542a2a, 0xab3b9090, 0x830b8888,
+    0xca8c4646, 0x29c7eeee, 0xd36bb8b8, 0x3c281414,
+    0x79a7dede, 0xe2bc5e5e, 0x1d160b0b, 0x76addbdb,
+    0x3bdbe0e0, 0x56643232, 0x4e743a3a, 0x1e140a0a,
+    0xdb924949, 0x0a0c0606, 0x6c482424, 0xe4b85c5c,
+    0x5d9fc2c2, 0x6ebdd3d3, 0xef43acac, 0xa6c46262,
+    0xa8399191, 0xa4319595, 0x37d3e4e4, 0x8bf27979,
+    0x32d5e7e7, 0x438bc8c8, 0x596e3737, 0xb7da6d6d,
+    0x8c018d8d, 0x64b1d5d5, 0xd29c4e4e, 0xe049a9a9,
+    0xb4d86c6c, 0xfaac5656, 0x07f3f4f4, 0x25cfeaea,
+    0xafca6565, 0x8ef47a7a, 0xe947aeae, 0x18100808,
+    0xd56fbaba, 0x88f07878, 0x6f4a2525, 0x725c2e2e,
+    0x24381c1c, 0xf157a6a6, 0xc773b4b4, 0x5197c6c6,
+    0x23cbe8e8, 0x7ca1dddd, 0x9ce87474, 0x213e1f1f,
+    0xdd964b4b, 0xdc61bdbd, 0x860d8b8b, 0x850f8a8a,
+    0x90e07070, 0x427c3e3e, 0xc471b5b5, 0xaacc6666,
+    0xd8904848, 0x05060303, 0x01f7f6f6, 0x121c0e0e,
+    0xa3c26161, 0x5f6a3535, 0xf9ae5757, 0xd069b9b9,
+    0x91178686, 0x5899c1c1, 0x273a1d1d, 0xb9279e9e,
+    0x38d9e1e1, 0x13ebf8f8, 0xb32b9898, 0x33221111,
+    0xbbd26969, 0x70a9d9d9, 0x89078e8e, 0xa7339494,
+    0xb62d9b9b, 0x223c1e1e, 0x92158787, 0x20c9e9e9,
+    0x4987cece, 0xffaa5555, 0x78502828, 0x7aa5dfdf,
+    0x8f038c8c, 0xf859a1a1, 0x80098989, 0x171a0d0d,
+    0xda65bfbf, 0x31d7e6e6, 0xc6844242, 0xb8d06868,
+    0xc3824141, 0xb0299999, 0x775a2d2d, 0x111e0f0f,
+    0xcb7bb0b0, 0xfca85454, 0xd66dbbbb, 0x3a2c1616,
+};
+
+#endif /* HAVE_AES_DECRYPT || HAVE_AES_CBC || HAVE_AESCCM || HAVE_AESGCM ||
+        * WOLFSSL_AES_DIRECT || WOLFSSL_AES_COUNTER */
+#ifdef HAVE_AES_DECRYPT
+void AES_invert_key(unsigned char* ks, word32 rounds);
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+void AES_invert_key(unsigned char* ks_p, word32 rounds_p)
+#else
+void AES_invert_key(unsigned char* ks, word32 rounds)
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+{
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+    register unsigned char* ks asm ("3") = (unsigned char*)ks_p;
+    register word32 rounds asm ("4") = (word32)rounds_p;
+    register word32* L_AES_PPC64_te_c asm ("5") = (word32*)&L_AES_PPC64_te;
+    register word32* L_AES_PPC64_td_c asm ("6") = (word32*)&L_AES_PPC64_td;
+#else
+    register word32* L_AES_PPC64_te_c = (word32*)&L_AES_PPC64_te;
+    register word32* L_AES_PPC64_td_c = (word32*)&L_AES_PPC64_td;
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+
+    __asm__ __volatile__ (
+#ifdef WOLFSSL_NO_VAR_ASSIGN_REG
+        "mr      5, %[L_AES_PPC64_te]\n\t"
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+#ifdef WOLFSSL_NO_VAR_ASSIGN_REG
+        "mr      6, %[L_AES_PPC64_td]\n\t"
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+        "addi    5, 5, 3\n\t"
+        "sldi    16, %[rounds], 4\n\t"
+        "add     16, 16, %[ks]\n\t"
+        "ld      7, 0(%[ks])\n\t"
+        "ld      8, 8(%[ks])\n\t"
+        "ld      11, 0(16)\n\t"
+        "ld      12, 8(16)\n\t"
+        "rldicr  7, 7, 32, 63\n\t"
+        "rldicr  8, 8, 32, 63\n\t"
+        "rldicr  11, 11, 32, 63\n\t"
+        "rldicr  12, 12, 32, 63\n\t"
+        "std     7, 0(16)\n\t"
+        "std     8, 8(16)\n\t"
+        "std     11, 0(%[ks])\n\t"
+        "std     12, 8(%[ks])\n\t"
+        "addi    %[ks], %[ks], 16\n\t"
+        "addi    16, 16, -16\n\t"
+        "addi    0, %[rounds], -2\n\t"
+        "srdi    0, 0, 1\n\t"
+        "mtctr   0\n\t"
+        "\n"
+    "L_AES_invert_key_loop_%=: \n\t"
+        "ld      7, 0(%[ks])\n\t"
+        "ld      8, 8(%[ks])\n\t"
+        "ld      11, 0(16)\n\t"
+        "ld      12, 8(16)\n\t"
+        "std     7, 0(16)\n\t"
+        "std     8, 8(16)\n\t"
+        "std     11, 0(%[ks])\n\t"
+        "std     12, 8(%[ks])\n\t"
+        "addi    %[ks], %[ks], 16\n\t"
+        "addi    16, 16, -16\n\t"
+        "bdnz    L_AES_invert_key_loop_%=\n\t"
+        "sldi    16, %[rounds], 3\n\t"
+        "subf    %[ks], 16, %[ks]\n\t"
+        "addi    %[ks], %[ks], 16\n\t"
+        "addi    0, %[rounds], -1\n\t"
+        "mtctr   0\n\t"
+        "\n"
+    "L_AES_invert_key_mix_loop_%=: \n\t"
+        "lwz     8, 0(%[ks])\n\t"
+        "lwz     7, 4(%[ks])\n\t"
+        "lwz     10, 8(%[ks])\n\t"
+        "lwz     9, 12(%[ks])\n\t"
+        "rlwinm  11, 7, 2, 22, 29\n\t"
+        "rlwinm  12, 7, 26, 22, 29\n\t"
+        "rlwinm  14, 7, 18, 22, 29\n\t"
+        "rlwinm  15, 7, 10, 22, 29\n\t"
+        "lbzx    11, 5, 11\n\t"
+        "lbzx    12, 5, 12\n\t"
+        "lbzx    14, 5, 14\n\t"
+        "lbzx    15, 5, 15\n\t"
+        "slwi    11, 11, 2\n\t"
+        "slwi    12, 12, 2\n\t"
+        "slwi    14, 14, 2\n\t"
+        "slwi    15, 15, 2\n\t"
+        "lwzx    11, 6, 11\n\t"
+        "lwzx    12, 6, 12\n\t"
+        "lwzx    14, 6, 14\n\t"
+        "lwzx    15, 6, 15\n\t"
+        "rlwimi  11, 11, 16, 0, 31\n\t"
+        "rlwimi  12, 12, 24, 0, 31\n\t"
+        "rlwimi  15, 15, 8, 0, 31\n\t"
+        "xor     14, 14, 11\n\t"
+        "xor     14, 14, 12\n\t"
+        "xor     14, 14, 15\n\t"
+        "stw     14, 4(%[ks])\n\t"
+        "rlwinm  11, 8, 2, 22, 29\n\t"
+        "rlwinm  12, 8, 26, 22, 29\n\t"
+        "rlwinm  14, 8, 18, 22, 29\n\t"
+        "rlwinm  15, 8, 10, 22, 29\n\t"
+        "lbzx    11, 5, 11\n\t"
+        "lbzx    12, 5, 12\n\t"
+        "lbzx    14, 5, 14\n\t"
+        "lbzx    15, 5, 15\n\t"
+        "slwi    11, 11, 2\n\t"
+        "slwi    12, 12, 2\n\t"
+        "slwi    14, 14, 2\n\t"
+        "slwi    15, 15, 2\n\t"
+        "lwzx    11, 6, 11\n\t"
+        "lwzx    12, 6, 12\n\t"
+        "lwzx    14, 6, 14\n\t"
+        "lwzx    15, 6, 15\n\t"
+        "rlwimi  11, 11, 16, 0, 31\n\t"
+        "rlwimi  12, 12, 24, 0, 31\n\t"
+        "rlwimi  15, 15, 8, 0, 31\n\t"
+        "xor     14, 14, 11\n\t"
+        "xor     14, 14, 12\n\t"
+        "xor     14, 14, 15\n\t"
+        "stw     14, 0(%[ks])\n\t"
+        "addi    %[ks], %[ks], 8\n\t"
+        "rlwinm  11, 9, 2, 22, 29\n\t"
+        "rlwinm  12, 9, 26, 22, 29\n\t"
+        "rlwinm  14, 9, 18, 22, 29\n\t"
+        "rlwinm  15, 9, 10, 22, 29\n\t"
+        "lbzx    11, 5, 11\n\t"
+        "lbzx    12, 5, 12\n\t"
+        "lbzx    14, 5, 14\n\t"
+        "lbzx    15, 5, 15\n\t"
+        "slwi    11, 11, 2\n\t"
+        "slwi    12, 12, 2\n\t"
+        "slwi    14, 14, 2\n\t"
+        "slwi    15, 15, 2\n\t"
+        "lwzx    11, 6, 11\n\t"
+        "lwzx    12, 6, 12\n\t"
+        "lwzx    14, 6, 14\n\t"
+        "lwzx    15, 6, 15\n\t"
+        "rlwimi  11, 11, 16, 0, 31\n\t"
+        "rlwimi  12, 12, 24, 0, 31\n\t"
+        "rlwimi  15, 15, 8, 0, 31\n\t"
+        "xor     14, 14, 11\n\t"
+        "xor     14, 14, 12\n\t"
+        "xor     14, 14, 15\n\t"
+        "stw     14, 4(%[ks])\n\t"
+        "rlwinm  11, 10, 2, 22, 29\n\t"
+        "rlwinm  12, 10, 26, 22, 29\n\t"
+        "rlwinm  14, 10, 18, 22, 29\n\t"
+        "rlwinm  15, 10, 10, 22, 29\n\t"
+        "lbzx    11, 5, 11\n\t"
+        "lbzx    12, 5, 12\n\t"
+        "lbzx    14, 5, 14\n\t"
+        "lbzx    15, 5, 15\n\t"
+        "slwi    11, 11, 2\n\t"
+        "slwi    12, 12, 2\n\t"
+        "slwi    14, 14, 2\n\t"
+        "slwi    15, 15, 2\n\t"
+        "lwzx    11, 6, 11\n\t"
+        "lwzx    12, 6, 12\n\t"
+        "lwzx    14, 6, 14\n\t"
+        "lwzx    15, 6, 15\n\t"
+        "rlwimi  11, 11, 16, 0, 31\n\t"
+        "rlwimi  12, 12, 24, 0, 31\n\t"
+        "rlwimi  15, 15, 8, 0, 31\n\t"
+        "xor     14, 14, 11\n\t"
+        "xor     14, 14, 12\n\t"
+        "xor     14, 14, 15\n\t"
+        "stw     14, 0(%[ks])\n\t"
+        "addi    %[ks], %[ks], 8\n\t"
+        "bdnz    L_AES_invert_key_mix_loop_%=\n\t"
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+        : [ks] "+r" (ks), [rounds] "+r" (rounds),
+          [L_AES_PPC64_te] "+r" (L_AES_PPC64_te_c),
+          [L_AES_PPC64_td] "+r" (L_AES_PPC64_td_c)
+        :
+        : "memory", "cc", "0", "7", "8", "9", "10", "11", "12", "14", "15",
+            "16"
+#else
+        :
+        : [ks] "r" (ks), [rounds] "r" (rounds),
+          [L_AES_PPC64_te] "r" (L_AES_PPC64_te_c),
+          [L_AES_PPC64_td] "r" (L_AES_PPC64_td_c)
+        : "memory", "cc", "0", "5", "6", "7", "8", "9", "10", "11", "12", "14",
+            "15", "16"
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+    );
+}
+
+#endif /* HAVE_AES_DECRYPT */
+static const word32 L_AES_PPC64_rcon[] = {
+    0x01000000, 0x02000000, 0x04000000, 0x08000000,
+    0x10000000, 0x20000000, 0x40000000, 0x80000000,
+    0x1b000000, 0x36000000
+};
+
+void AES_set_encrypt_key(const unsigned char* key, word32 len,
+    unsigned char* ks);
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+void AES_set_encrypt_key(const unsigned char* key_p, word32 len_p,
+    unsigned char* ks_p)
+#else
+void AES_set_encrypt_key(const unsigned char* key, word32 len,
+    unsigned char* ks)
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+{
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+    register const unsigned char* key asm ("3") = (const unsigned char*)key_p;
+    register word32 len asm ("4") = (word32)len_p;
+    register unsigned char* ks asm ("5") = (unsigned char*)ks_p;
+    register word32* L_AES_PPC64_te_c asm ("6") = (word32*)&L_AES_PPC64_te;
+    register word32* L_AES_PPC64_rcon_c asm ("7") = (word32*)&L_AES_PPC64_rcon;
+#else
+    register word32* L_AES_PPC64_te_c = (word32*)&L_AES_PPC64_te;
+    register word32* L_AES_PPC64_rcon_c = (word32*)&L_AES_PPC64_rcon;
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+
+    __asm__ __volatile__ (
+#ifdef WOLFSSL_NO_VAR_ASSIGN_REG
+        "mr      6, %[L_AES_PPC64_te]\n\t"
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+#ifdef WOLFSSL_NO_VAR_ASSIGN_REG
+        "mr      7, %[L_AES_PPC64_rcon]\n\t"
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+        "addi    6, 6, 3\n\t"
+        "cmplwi  %[len], 0x80\n\t"
+        "beq     L_AES_set_encrypt_key_start_128_%=\n\t"
+        "cmplwi  %[len], 0xc0\n\t"
+        "beq     L_AES_set_encrypt_key_start_192_%=\n\t"
+        "ld      9, 0(%[key])\n\t"
+        "ld      10, 8(%[key])\n\t"
+        "ld      11, 16(%[key])\n\t"
+        "ld      12, 24(%[key])\n\t"
+        "rldicr  9, 9, 32, 63\n\t"
+        "rldicr  10, 10, 32, 63\n\t"
+        "rldicr  11, 11, 32, 63\n\t"
+        "rldicr  12, 12, 32, 63\n\t"
+        "std     9, 0(%[ks])\n\t"
+        "std     10, 8(%[ks])\n\t"
+        "std     11, 16(%[ks])\n\t"
+        "std     12, 24(%[ks])\n\t"
+        "rldicr  12, 12, 32, 63\n\t"
+        "li      8, 6\n\t"
+        "mtctr   8\n\t"
+        "\n"
+    "L_AES_set_encrypt_key_loop_256_%=: \n\t"
+        "rlwinm  9, 12, 2, 22, 29\n\t"
+        "rlwinm  10, 12, 26, 22, 29\n\t"
+        "rlwinm  11, 12, 18, 22, 29\n\t"
+        "rlwinm  12, 12, 10, 22, 29\n\t"
+        "lbzx    9, 6, 9\n\t"
+        "lbzx    10, 6, 10\n\t"
+        "lbzx    11, 6, 11\n\t"
+        "lbzx    0, 6, 12\n\t"
+        "rlwimi  0, 9, 8, 16, 23\n\t"
+        "rlwimi  0, 10, 16, 8, 15\n\t"
+        "rlwimi  0, 11, 24, 0, 7\n\t"
+        "lwz     10, 0(%[ks])\n\t"
+        "lwz     9, 4(%[ks])\n\t"
+        "lwz     12, 8(%[ks])\n\t"
+        "lwz     11, 12(%[ks])\n\t"
+        "addi    %[ks], %[ks], 16\n\t"
+        "xor     9, 9, 0\n\t"
+        "lwz     0, 0(7)\n\t"
+        "addi    7, 7, 4\n\t"
+        "xor     9, 9, 0\n\t"
+        "xor     10, 10, 9\n\t"
+        "xor     11, 11, 10\n\t"
+        "xor     12, 12, 11\n\t"
+        "stw     10, 16(%[ks])\n\t"
+        "stw     9, 20(%[ks])\n\t"
+        "stw     12, 24(%[ks])\n\t"
+        "stw     11, 28(%[ks])\n\t"
+        "rlwinm  9, 12, 26, 22, 29\n\t"
+        "rlwinm  10, 12, 18, 22, 29\n\t"
+        "rlwinm  11, 12, 10, 22, 29\n\t"
+        "rlwinm  12, 12, 2, 22, 29\n\t"
+        "lbzx    9, 6, 9\n\t"
+        "lbzx    10, 6, 10\n\t"
+        "lbzx    11, 6, 11\n\t"
+        "lbzx    0, 6, 12\n\t"
+        "rlwimi  0, 9, 8, 16, 23\n\t"
+        "rlwimi  0, 10, 16, 8, 15\n\t"
+        "rlwimi  0, 11, 24, 0, 7\n\t"
+        "lwz     10, 0(%[ks])\n\t"
+        "lwz     9, 4(%[ks])\n\t"
+        "lwz     12, 8(%[ks])\n\t"
+        "lwz     11, 12(%[ks])\n\t"
+        "addi    %[ks], %[ks], 16\n\t"
+        "xor     9, 9, 0\n\t"
+        "xor     10, 10, 9\n\t"
+        "xor     11, 11, 10\n\t"
+        "xor     12, 12, 11\n\t"
+        "stw     10, 16(%[ks])\n\t"
+        "stw     9, 20(%[ks])\n\t"
+        "stw     12, 24(%[ks])\n\t"
+        "stw     11, 28(%[ks])\n\t"
+        "bdnz    L_AES_set_encrypt_key_loop_256_%=\n\t"
+        "rlwinm  9, 12, 2, 22, 29\n\t"
+        "rlwinm  10, 12, 26, 22, 29\n\t"
+        "rlwinm  11, 12, 18, 22, 29\n\t"
+        "rlwinm  12, 12, 10, 22, 29\n\t"
+        "lbzx    9, 6, 9\n\t"
+        "lbzx    10, 6, 10\n\t"
+        "lbzx    11, 6, 11\n\t"
+        "lbzx    0, 6, 12\n\t"
+        "rlwimi  0, 9, 8, 16, 23\n\t"
+        "rlwimi  0, 10, 16, 8, 15\n\t"
+        "rlwimi  0, 11, 24, 0, 7\n\t"
+        "lwz     10, 0(%[ks])\n\t"
+        "lwz     9, 4(%[ks])\n\t"
+        "lwz     12, 8(%[ks])\n\t"
+        "lwz     11, 12(%[ks])\n\t"
+        "addi    %[ks], %[ks], 16\n\t"
+        "xor     9, 9, 0\n\t"
+        "lwz     0, 0(7)\n\t"
+        "addi    7, 7, 4\n\t"
+        "xor     9, 9, 0\n\t"
+        "xor     10, 10, 9\n\t"
+        "xor     11, 11, 10\n\t"
+        "xor     12, 12, 11\n\t"
+        "stw     9, 16(%[ks])\n\t"
+        "stw     10, 20(%[ks])\n\t"
+        "stw     11, 24(%[ks])\n\t"
+        "stw     12, 28(%[ks])\n\t"
+        "b       L_AES_set_encrypt_key_end_%=\n\t"
+        "\n"
+    "L_AES_set_encrypt_key_start_192_%=: \n\t"
+        "ld      12, 0(%[key])\n\t"
+        "ld      14, 8(%[key])\n\t"
+        "ld      15, 16(%[key])\n\t"
+        "rldicr  12, 12, 32, 63\n\t"
+        "rldicr  14, 14, 32, 63\n\t"
+        "rldicr  15, 15, 32, 63\n\t"
+        "std     12, 0(%[ks])\n\t"
+        "std     14, 8(%[ks])\n\t"
+        "std     15, 16(%[ks])\n\t"
+        "rldicr  15, 15, 32, 63\n\t"
+        "li      8, 7\n\t"
+        "mtctr   8\n\t"
+        "\n"
+    "L_AES_set_encrypt_key_loop_192_%=: \n\t"
+        "rlwinm  9, 15, 2, 22, 29\n\t"
+        "rlwinm  10, 15, 26, 22, 29\n\t"
+        "rlwinm  11, 15, 18, 22, 29\n\t"
+        "rlwinm  15, 15, 10, 22, 29\n\t"
+        "lbzx    9, 6, 9\n\t"
+        "lbzx    10, 6, 10\n\t"
+        "lbzx    11, 6, 11\n\t"
+        "lbzx    0, 6, 15\n\t"
+        "rlwimi  0, 9, 8, 16, 23\n\t"
+        "rlwimi  0, 10, 16, 8, 15\n\t"
+        "rlwimi  0, 11, 24, 0, 7\n\t"
+        "lwz     10, 0(%[ks])\n\t"
+        "lwz     9, 4(%[ks])\n\t"
+        "lwz     12, 8(%[ks])\n\t"
+        "lwz     11, 12(%[ks])\n\t"
+        "lwz     15, 16(%[ks])\n\t"
+        "lwz     14, 20(%[ks])\n\t"
+        "addi    %[ks], %[ks], 24\n\t"
+        "xor     9, 9, 0\n\t"
+        "lwz     0, 0(7)\n\t"
+        "addi    7, 7, 4\n\t"
+        "xor     9, 9, 0\n\t"
+        "xor     10, 10, 9\n\t"
+        "xor     11, 11, 10\n\t"
+        "xor     12, 12, 11\n\t"
+        "xor     14, 14, 12\n\t"
+        "xor     15, 15, 14\n\t"
+        "stw     10, 0(%[ks])\n\t"
+        "stw     9, 4(%[ks])\n\t"
+        "stw     12, 8(%[ks])\n\t"
+        "stw     11, 12(%[ks])\n\t"
+        "stw     15, 16(%[ks])\n\t"
+        "stw     14, 20(%[ks])\n\t"
+        "bdnz    L_AES_set_encrypt_key_loop_192_%=\n\t"
+        "rlwinm  9, 15, 2, 22, 29\n\t"
+        "rlwinm  10, 15, 26, 22, 29\n\t"
+        "rlwinm  11, 15, 18, 22, 29\n\t"
+        "rlwinm  15, 15, 10, 22, 29\n\t"
+        "lbzx    9, 6, 9\n\t"
+        "lbzx    10, 6, 10\n\t"
+        "lbzx    11, 6, 11\n\t"
+        "lbzx    0, 6, 15\n\t"
+        "rlwimi  0, 9, 8, 16, 23\n\t"
+        "rlwimi  0, 10, 16, 8, 15\n\t"
+        "rlwimi  0, 11, 24, 0, 7\n\t"
+        "lwz     10, 0(%[ks])\n\t"
+        "lwz     9, 4(%[ks])\n\t"
+        "lwz     12, 8(%[ks])\n\t"
+        "lwz     11, 12(%[ks])\n\t"
+        "lwz     15, 16(%[ks])\n\t"
+        "lwz     14, 20(%[ks])\n\t"
+        "addi    %[ks], %[ks], 24\n\t"
+        "xor     9, 9, 0\n\t"
+        "lwz     0, 0(7)\n\t"
+        "addi    7, 7, 4\n\t"
+        "xor     9, 9, 0\n\t"
+        "xor     10, 10, 9\n\t"
+        "xor     11, 11, 10\n\t"
+        "xor     12, 12, 11\n\t"
+        "stw     9, 0(%[ks])\n\t"
+        "stw     10, 4(%[ks])\n\t"
+        "stw     11, 8(%[ks])\n\t"
+        "stw     12, 12(%[ks])\n\t"
+        "b       L_AES_set_encrypt_key_end_%=\n\t"
+        "\n"
+    "L_AES_set_encrypt_key_start_128_%=: \n\t"
+        "li      8, 0\n\t"
+        "ld      11, 0(%[key])\n\t"
+        "ld      12, 8(%[key])\n\t"
+        "rldicr  11, 11, 32, 63\n\t"
+        "rldicr  12, 12, 32, 63\n\t"
+        "std     11, 0(%[ks])\n\t"
+        "std     12, 8(%[ks])\n\t"
+        "rldicr  12, 12, 32, 63\n\t"
+        "li      8, 9\n\t"
+        "mtctr   8\n\t"
+        "\n"
+    "L_AES_set_encrypt_key_loop_128_%=: \n\t"
+        "rlwinm  9, 12, 2, 22, 29\n\t"
+        "rlwinm  10, 12, 26, 22, 29\n\t"
+        "rlwinm  11, 12, 18, 22, 29\n\t"
+        "rlwinm  12, 12, 10, 22, 29\n\t"
+        "lbzx    9, 6, 9\n\t"
+        "lbzx    10, 6, 10\n\t"
+        "lbzx    11, 6, 11\n\t"
+        "lbzx    0, 6, 12\n\t"
+        "rlwimi  0, 9, 8, 16, 23\n\t"
+        "rlwimi  0, 10, 16, 8, 15\n\t"
+        "rlwimi  0, 11, 24, 0, 7\n\t"
+        "lwz     10, 0(%[ks])\n\t"
+        "lwz     9, 4(%[ks])\n\t"
+        "lwz     12, 8(%[ks])\n\t"
+        "lwz     11, 12(%[ks])\n\t"
+        "addi    %[ks], %[ks], 16\n\t"
+        "xor     9, 9, 0\n\t"
+        "lwz     0, 0(7)\n\t"
+        "addi    7, 7, 4\n\t"
+        "xor     9, 9, 0\n\t"
+        "xor     10, 10, 9\n\t"
+        "xor     11, 11, 10\n\t"
+        "xor     12, 12, 11\n\t"
+        "stw     10, 0(%[ks])\n\t"
+        "stw     9, 4(%[ks])\n\t"
+        "stw     12, 8(%[ks])\n\t"
+        "stw     11, 12(%[ks])\n\t"
+        "bdnz    L_AES_set_encrypt_key_loop_128_%=\n\t"
+        "rlwinm  9, 12, 2, 22, 29\n\t"
+        "rlwinm  10, 12, 26, 22, 29\n\t"
+        "rlwinm  11, 12, 18, 22, 29\n\t"
+        "rlwinm  12, 12, 10, 22, 29\n\t"
+        "lbzx    9, 6, 9\n\t"
+        "lbzx    10, 6, 10\n\t"
+        "lbzx    11, 6, 11\n\t"
+        "lbzx    0, 6, 12\n\t"
+        "rlwimi  0, 9, 8, 16, 23\n\t"
+        "rlwimi  0, 10, 16, 8, 15\n\t"
+        "rlwimi  0, 11, 24, 0, 7\n\t"
+        "lwz     10, 0(%[ks])\n\t"
+        "lwz     9, 4(%[ks])\n\t"
+        "lwz     12, 8(%[ks])\n\t"
+        "lwz     11, 12(%[ks])\n\t"
+        "addi    %[ks], %[ks], 16\n\t"
+        "xor     9, 9, 0\n\t"
+        "lwz     0, 0(7)\n\t"
+        "addi    7, 7, 4\n\t"
+        "xor     9, 9, 0\n\t"
+        "xor     10, 10, 9\n\t"
+        "xor     11, 11, 10\n\t"
+        "xor     12, 12, 11\n\t"
+        "stw     9, 0(%[ks])\n\t"
+        "stw     10, 4(%[ks])\n\t"
+        "stw     11, 8(%[ks])\n\t"
+        "stw     12, 12(%[ks])\n\t"
+        "\n"
+    "L_AES_set_encrypt_key_end_%=: \n\t"
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+        : [key] "+r" (key), [len] "+r" (len), [ks] "+r" (ks),
+          [L_AES_PPC64_te] "+r" (L_AES_PPC64_te_c),
+          [L_AES_PPC64_rcon] "+r" (L_AES_PPC64_rcon_c)
+        :
+        : "memory", "cc", "0", "8", "9", "10", "11", "12", "14", "15"
+#else
+        :
+        : [key] "r" (key), [len] "r" (len), [ks] "r" (ks),
+          [L_AES_PPC64_te] "r" (L_AES_PPC64_te_c),
+          [L_AES_PPC64_rcon] "r" (L_AES_PPC64_rcon_c)
+        : "memory", "cc", "0", "6", "7", "8", "9", "10", "11", "12", "14", "15"
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+    );
+}
+
+static const word32 L_AES_PPC64_te4_0[] = {
+    0xc66363a5, 0xf87c7c84, 0xee777799, 0xf67b7b8d,
+    0xfff2f20d, 0xd66b6bbd, 0xde6f6fb1, 0x91c5c554,
+    0x60303050, 0x02010103, 0xce6767a9, 0x562b2b7d,
+    0xe7fefe19, 0xb5d7d762, 0x4dababe6, 0xec76769a,
+    0x8fcaca45, 0x1f82829d, 0x89c9c940, 0xfa7d7d87,
+    0xeffafa15, 0xb25959eb, 0x8e4747c9, 0xfbf0f00b,
+    0x41adadec, 0xb3d4d467, 0x5fa2a2fd, 0x45afafea,
+    0x239c9cbf, 0x53a4a4f7, 0xe4727296, 0x9bc0c05b,
+    0x75b7b7c2, 0xe1fdfd1c, 0x3d9393ae, 0x4c26266a,
+    0x6c36365a, 0x7e3f3f41, 0xf5f7f702, 0x83cccc4f,
+    0x6834345c, 0x51a5a5f4, 0xd1e5e534, 0xf9f1f108,
+    0xe2717193, 0xabd8d873, 0x62313153, 0x2a15153f,
+    0x0804040c, 0x95c7c752, 0x46232365, 0x9dc3c35e,
+    0x30181828, 0x379696a1, 0x0a05050f, 0x2f9a9ab5,
+    0x0e070709, 0x24121236, 0x1b80809b, 0xdfe2e23d,
+    0xcdebeb26, 0x4e272769, 0x7fb2b2cd, 0xea75759f,
+    0x1209091b, 0x1d83839e, 0x582c2c74, 0x341a1a2e,
+    0x361b1b2d, 0xdc6e6eb2, 0xb45a5aee, 0x5ba0a0fb,
+    0xa45252f6, 0x763b3b4d, 0xb7d6d661, 0x7db3b3ce,
+    0x5229297b, 0xdde3e33e, 0x5e2f2f71, 0x13848497,
+    0xa65353f5, 0xb9d1d168, 0x00000000, 0xc1eded2c,
+    0x40202060, 0xe3fcfc1f, 0x79b1b1c8, 0xb65b5bed,
+    0xd46a6abe, 0x8dcbcb46, 0x67bebed9, 0x7239394b,
+    0x944a4ade, 0x984c4cd4, 0xb05858e8, 0x85cfcf4a,
+    0xbbd0d06b, 0xc5efef2a, 0x4faaaae5, 0xedfbfb16,
+    0x864343c5, 0x9a4d4dd7, 0x66333355, 0x11858594,
+    0x8a4545cf, 0xe9f9f910, 0x04020206, 0xfe7f7f81,
+    0xa05050f0, 0x783c3c44, 0x259f9fba, 0x4ba8a8e3,
+    0xa25151f3, 0x5da3a3fe, 0x804040c0, 0x058f8f8a,
+    0x3f9292ad, 0x219d9dbc, 0x70383848, 0xf1f5f504,
+    0x63bcbcdf, 0x77b6b6c1, 0xafdada75, 0x42212163,
+    0x20101030, 0xe5ffff1a, 0xfdf3f30e, 0xbfd2d26d,
+    0x81cdcd4c, 0x180c0c14, 0x26131335, 0xc3ecec2f,
+    0xbe5f5fe1, 0x359797a2, 0x884444cc, 0x2e171739,
+    0x93c4c457, 0x55a7a7f2, 0xfc7e7e82, 0x7a3d3d47,
+    0xc86464ac, 0xba5d5de7, 0x3219192b, 0xe6737395,
+    0xc06060a0, 0x19818198, 0x9e4f4fd1, 0xa3dcdc7f,
+    0x44222266, 0x542a2a7e, 0x3b9090ab, 0x0b888883,
+    0x8c4646ca, 0xc7eeee29, 0x6bb8b8d3, 0x2814143c,
+    0xa7dede79, 0xbc5e5ee2, 0x160b0b1d, 0xaddbdb76,
+    0xdbe0e03b, 0x64323256, 0x743a3a4e, 0x140a0a1e,
+    0x924949db, 0x0c06060a, 0x4824246c, 0xb85c5ce4,
+    0x9fc2c25d, 0xbdd3d36e, 0x43acacef, 0xc46262a6,
+    0x399191a8, 0x319595a4, 0xd3e4e437, 0xf279798b,
+    0xd5e7e732, 0x8bc8c843, 0x6e373759, 0xda6d6db7,
+    0x018d8d8c, 0xb1d5d564, 0x9c4e4ed2, 0x49a9a9e0,
+    0xd86c6cb4, 0xac5656fa, 0xf3f4f407, 0xcfeaea25,
+    0xca6565af, 0xf47a7a8e, 0x47aeaee9, 0x10080818,
+    0x6fbabad5, 0xf0787888, 0x4a25256f, 0x5c2e2e72,
+    0x381c1c24, 0x57a6a6f1, 0x73b4b4c7, 0x97c6c651,
+    0xcbe8e823, 0xa1dddd7c, 0xe874749c, 0x3e1f1f21,
+    0x964b4bdd, 0x61bdbddc, 0x0d8b8b86, 0x0f8a8a85,
+    0xe0707090, 0x7c3e3e42, 0x71b5b5c4, 0xcc6666aa,
+    0x904848d8, 0x06030305, 0xf7f6f601, 0x1c0e0e12,
+    0xc26161a3, 0x6a35355f, 0xae5757f9, 0x69b9b9d0,
+    0x17868691, 0x99c1c158, 0x3a1d1d27, 0x279e9eb9,
+    0xd9e1e138, 0xebf8f813, 0x2b9898b3, 0x22111133,
+    0xd26969bb, 0xa9d9d970, 0x078e8e89, 0x339494a7,
+    0x2d9b9bb6, 0x3c1e1e22, 0x15878792, 0xc9e9e920,
+    0x87cece49, 0xaa5555ff, 0x50282878, 0xa5dfdf7a,
+    0x038c8c8f, 0x59a1a1f8, 0x09898980, 0x1a0d0d17,
+    0x65bfbfda, 0xd7e6e631, 0x844242c6, 0xd06868b8,
+    0x824141c3, 0x299999b0, 0x5a2d2d77, 0x1e0f0f11,
+    0x7bb0b0cb, 0xa85454fc, 0x6dbbbbd6, 0x2c16163a,
+    0xa5c66363, 0x84f87c7c, 0x99ee7777, 0x8df67b7b,
+    0x0dfff2f2, 0xbdd66b6b, 0xb1de6f6f, 0x5491c5c5,
+    0x50603030, 0x03020101, 0xa9ce6767, 0x7d562b2b,
+    0x19e7fefe, 0x62b5d7d7, 0xe64dabab, 0x9aec7676,
+    0x458fcaca, 0x9d1f8282, 0x4089c9c9, 0x87fa7d7d,
+    0x15effafa, 0xebb25959, 0xc98e4747, 0x0bfbf0f0,
+    0xec41adad, 0x67b3d4d4, 0xfd5fa2a2, 0xea45afaf,
+    0xbf239c9c, 0xf753a4a4, 0x96e47272, 0x5b9bc0c0,
+    0xc275b7b7, 0x1ce1fdfd, 0xae3d9393, 0x6a4c2626,
+    0x5a6c3636, 0x417e3f3f, 0x02f5f7f7, 0x4f83cccc,
+    0x5c683434, 0xf451a5a5, 0x34d1e5e5, 0x08f9f1f1,
+    0x93e27171, 0x73abd8d8, 0x53623131, 0x3f2a1515,
+    0x0c080404, 0x5295c7c7, 0x65462323, 0x5e9dc3c3,
+    0x28301818, 0xa1379696, 0x0f0a0505, 0xb52f9a9a,
+    0x090e0707, 0x36241212, 0x9b1b8080, 0x3ddfe2e2,
+    0x26cdebeb, 0x694e2727, 0xcd7fb2b2, 0x9fea7575,
+    0x1b120909, 0x9e1d8383, 0x74582c2c, 0x2e341a1a,
+    0x2d361b1b, 0xb2dc6e6e, 0xeeb45a5a, 0xfb5ba0a0,
+    0xf6a45252, 0x4d763b3b, 0x61b7d6d6, 0xce7db3b3,
+    0x7b522929, 0x3edde3e3, 0x715e2f2f, 0x97138484,
+    0xf5a65353, 0x68b9d1d1, 0x00000000, 0x2cc1eded,
+    0x60402020, 0x1fe3fcfc, 0xc879b1b1, 0xedb65b5b,
+    0xbed46a6a, 0x468dcbcb, 0xd967bebe, 0x4b723939,
+    0xde944a4a, 0xd4984c4c, 0xe8b05858, 0x4a85cfcf,
+    0x6bbbd0d0, 0x2ac5efef, 0xe54faaaa, 0x16edfbfb,
+    0xc5864343, 0xd79a4d4d, 0x55663333, 0x94118585,
+    0xcf8a4545, 0x10e9f9f9, 0x06040202, 0x81fe7f7f,
+    0xf0a05050, 0x44783c3c, 0xba259f9f, 0xe34ba8a8,
+    0xf3a25151, 0xfe5da3a3, 0xc0804040, 0x8a058f8f,
+    0xad3f9292, 0xbc219d9d, 0x48703838, 0x04f1f5f5,
+    0xdf63bcbc, 0xc177b6b6, 0x75afdada, 0x63422121,
+    0x30201010, 0x1ae5ffff, 0x0efdf3f3, 0x6dbfd2d2,
+    0x4c81cdcd, 0x14180c0c, 0x35261313, 0x2fc3ecec,
+    0xe1be5f5f, 0xa2359797, 0xcc884444, 0x392e1717,
+    0x5793c4c4, 0xf255a7a7, 0x82fc7e7e, 0x477a3d3d,
+    0xacc86464, 0xe7ba5d5d, 0x2b321919, 0x95e67373,
+    0xa0c06060, 0x98198181, 0xd19e4f4f, 0x7fa3dcdc,
+    0x66442222, 0x7e542a2a, 0xab3b9090, 0x830b8888,
+    0xca8c4646, 0x29c7eeee, 0xd36bb8b8, 0x3c281414,
+    0x79a7dede, 0xe2bc5e5e, 0x1d160b0b, 0x76addbdb,
+    0x3bdbe0e0, 0x56643232, 0x4e743a3a, 0x1e140a0a,
+    0xdb924949, 0x0a0c0606, 0x6c482424, 0xe4b85c5c,
+    0x5d9fc2c2, 0x6ebdd3d3, 0xef43acac, 0xa6c46262,
+    0xa8399191, 0xa4319595, 0x37d3e4e4, 0x8bf27979,
+    0x32d5e7e7, 0x438bc8c8, 0x596e3737, 0xb7da6d6d,
+    0x8c018d8d, 0x64b1d5d5, 0xd29c4e4e, 0xe049a9a9,
+    0xb4d86c6c, 0xfaac5656, 0x07f3f4f4, 0x25cfeaea,
+    0xafca6565, 0x8ef47a7a, 0xe947aeae, 0x18100808,
+    0xd56fbaba, 0x88f07878, 0x6f4a2525, 0x725c2e2e,
+    0x24381c1c, 0xf157a6a6, 0xc773b4b4, 0x5197c6c6,
+    0x23cbe8e8, 0x7ca1dddd, 0x9ce87474, 0x213e1f1f,
+    0xdd964b4b, 0xdc61bdbd, 0x860d8b8b, 0x850f8a8a,
+    0x90e07070, 0x427c3e3e, 0xc471b5b5, 0xaacc6666,
+    0xd8904848, 0x05060303, 0x01f7f6f6, 0x121c0e0e,
+    0xa3c26161, 0x5f6a3535, 0xf9ae5757, 0xd069b9b9,
+    0x91178686, 0x5899c1c1, 0x273a1d1d, 0xb9279e9e,
+    0x38d9e1e1, 0x13ebf8f8, 0xb32b9898, 0x33221111,
+    0xbbd26969, 0x70a9d9d9, 0x89078e8e, 0xa7339494,
+    0xb62d9b9b, 0x223c1e1e, 0x92158787, 0x20c9e9e9,
+    0x4987cece, 0xffaa5555, 0x78502828, 0x7aa5dfdf,
+    0x8f038c8c, 0xf859a1a1, 0x80098989, 0x171a0d0d,
+    0xda65bfbf, 0x31d7e6e6, 0xc6844242, 0xb8d06868,
+    0xc3824141, 0xb0299999, 0x775a2d2d, 0x111e0f0f,
+    0xcb7bb0b0, 0xfca85454, 0xd66dbbbb, 0x3a2c1616,
+    0x63a5c663, 0x7c84f87c, 0x7799ee77, 0x7b8df67b,
+    0xf20dfff2, 0x6bbdd66b, 0x6fb1de6f, 0xc55491c5,
+    0x30506030, 0x01030201, 0x67a9ce67, 0x2b7d562b,
+    0xfe19e7fe, 0xd762b5d7, 0xabe64dab, 0x769aec76,
+    0xca458fca, 0x829d1f82, 0xc94089c9, 0x7d87fa7d,
+    0xfa15effa, 0x59ebb259, 0x47c98e47, 0xf00bfbf0,
+    0xadec41ad, 0xd467b3d4, 0xa2fd5fa2, 0xafea45af,
+    0x9cbf239c, 0xa4f753a4, 0x7296e472, 0xc05b9bc0,
+    0xb7c275b7, 0xfd1ce1fd, 0x93ae3d93, 0x266a4c26,
+    0x365a6c36, 0x3f417e3f, 0xf702f5f7, 0xcc4f83cc,
+    0x345c6834, 0xa5f451a5, 0xe534d1e5, 0xf108f9f1,
+    0x7193e271, 0xd873abd8, 0x31536231, 0x153f2a15,
+    0x040c0804, 0xc75295c7, 0x23654623, 0xc35e9dc3,
+    0x18283018, 0x96a13796, 0x050f0a05, 0x9ab52f9a,
+    0x07090e07, 0x12362412, 0x809b1b80, 0xe23ddfe2,
+    0xeb26cdeb, 0x27694e27, 0xb2cd7fb2, 0x759fea75,
+    0x091b1209, 0x839e1d83, 0x2c74582c, 0x1a2e341a,
+    0x1b2d361b, 0x6eb2dc6e, 0x5aeeb45a, 0xa0fb5ba0,
+    0x52f6a452, 0x3b4d763b, 0xd661b7d6, 0xb3ce7db3,
+    0x297b5229, 0xe33edde3, 0x2f715e2f, 0x84971384,
+    0x53f5a653, 0xd168b9d1, 0x00000000, 0xed2cc1ed,
+    0x20604020, 0xfc1fe3fc, 0xb1c879b1, 0x5bedb65b,
+    0x6abed46a, 0xcb468dcb, 0xbed967be, 0x394b7239,
+    0x4ade944a, 0x4cd4984c, 0x58e8b058, 0xcf4a85cf,
+    0xd06bbbd0, 0xef2ac5ef, 0xaae54faa, 0xfb16edfb,
+    0x43c58643, 0x4dd79a4d, 0x33556633, 0x85941185,
+    0x45cf8a45, 0xf910e9f9, 0x02060402, 0x7f81fe7f,
+    0x50f0a050, 0x3c44783c, 0x9fba259f, 0xa8e34ba8,
+    0x51f3a251, 0xa3fe5da3, 0x40c08040, 0x8f8a058f,
+    0x92ad3f92, 0x9dbc219d, 0x38487038, 0xf504f1f5,
+    0xbcdf63bc, 0xb6c177b6, 0xda75afda, 0x21634221,
+    0x10302010, 0xff1ae5ff, 0xf30efdf3, 0xd26dbfd2,
+    0xcd4c81cd, 0x0c14180c, 0x13352613, 0xec2fc3ec,
+    0x5fe1be5f, 0x97a23597, 0x44cc8844, 0x17392e17,
+    0xc45793c4, 0xa7f255a7, 0x7e82fc7e, 0x3d477a3d,
+    0x64acc864, 0x5de7ba5d, 0x192b3219, 0x7395e673,
+    0x60a0c060, 0x81981981, 0x4fd19e4f, 0xdc7fa3dc,
+    0x22664422, 0x2a7e542a, 0x90ab3b90, 0x88830b88,
+    0x46ca8c46, 0xee29c7ee, 0xb8d36bb8, 0x143c2814,
+    0xde79a7de, 0x5ee2bc5e, 0x0b1d160b, 0xdb76addb,
+    0xe03bdbe0, 0x32566432, 0x3a4e743a, 0x0a1e140a,
+    0x49db9249, 0x060a0c06, 0x246c4824, 0x5ce4b85c,
+    0xc25d9fc2, 0xd36ebdd3, 0xacef43ac, 0x62a6c462,
+    0x91a83991, 0x95a43195, 0xe437d3e4, 0x798bf279,
+    0xe732d5e7, 0xc8438bc8, 0x37596e37, 0x6db7da6d,
+    0x8d8c018d, 0xd564b1d5, 0x4ed29c4e, 0xa9e049a9,
+    0x6cb4d86c, 0x56faac56, 0xf407f3f4, 0xea25cfea,
+    0x65afca65, 0x7a8ef47a, 0xaee947ae, 0x08181008,
+    0xbad56fba, 0x7888f078, 0x256f4a25, 0x2e725c2e,
+    0x1c24381c, 0xa6f157a6, 0xb4c773b4, 0xc65197c6,
+    0xe823cbe8, 0xdd7ca1dd, 0x749ce874, 0x1f213e1f,
+    0x4bdd964b, 0xbddc61bd, 0x8b860d8b, 0x8a850f8a,
+    0x7090e070, 0x3e427c3e, 0xb5c471b5, 0x66aacc66,
+    0x48d89048, 0x03050603, 0xf601f7f6, 0x0e121c0e,
+    0x61a3c261, 0x355f6a35, 0x57f9ae57, 0xb9d069b9,
+    0x86911786, 0xc15899c1, 0x1d273a1d, 0x9eb9279e,
+    0xe138d9e1, 0xf813ebf8, 0x98b32b98, 0x11332211,
+    0x69bbd269, 0xd970a9d9, 0x8e89078e, 0x94a73394,
+    0x9bb62d9b, 0x1e223c1e, 0x87921587, 0xe920c9e9,
+    0xce4987ce, 0x55ffaa55, 0x28785028, 0xdf7aa5df,
+    0x8c8f038c, 0xa1f859a1, 0x89800989, 0x0d171a0d,
+    0xbfda65bf, 0xe631d7e6, 0x42c68442, 0x68b8d068,
+    0x41c38241, 0x99b02999, 0x2d775a2d, 0x0f111e0f,
+    0xb0cb7bb0, 0x54fca854, 0xbbd66dbb, 0x163a2c16,
+    0x6363a5c6, 0x7c7c84f8, 0x777799ee, 0x7b7b8df6,
+    0xf2f20dff, 0x6b6bbdd6, 0x6f6fb1de, 0xc5c55491,
+    0x30305060, 0x01010302, 0x6767a9ce, 0x2b2b7d56,
+    0xfefe19e7, 0xd7d762b5, 0xababe64d, 0x76769aec,
+    0xcaca458f, 0x82829d1f, 0xc9c94089, 0x7d7d87fa,
+    0xfafa15ef, 0x5959ebb2, 0x4747c98e, 0xf0f00bfb,
+    0xadadec41, 0xd4d467b3, 0xa2a2fd5f, 0xafafea45,
+    0x9c9cbf23, 0xa4a4f753, 0x727296e4, 0xc0c05b9b,
+    0xb7b7c275, 0xfdfd1ce1, 0x9393ae3d, 0x26266a4c,
+    0x36365a6c, 0x3f3f417e, 0xf7f702f5, 0xcccc4f83,
+    0x34345c68, 0xa5a5f451, 0xe5e534d1, 0xf1f108f9,
+    0x717193e2, 0xd8d873ab, 0x31315362, 0x15153f2a,
+    0x04040c08, 0xc7c75295, 0x23236546, 0xc3c35e9d,
+    0x18182830, 0x9696a137, 0x05050f0a, 0x9a9ab52f,
+    0x0707090e, 0x12123624, 0x80809b1b, 0xe2e23ddf,
+    0xebeb26cd, 0x2727694e, 0xb2b2cd7f, 0x75759fea,
+    0x09091b12, 0x83839e1d, 0x2c2c7458, 0x1a1a2e34,
+    0x1b1b2d36, 0x6e6eb2dc, 0x5a5aeeb4, 0xa0a0fb5b,
+    0x5252f6a4, 0x3b3b4d76, 0xd6d661b7, 0xb3b3ce7d,
+    0x29297b52, 0xe3e33edd, 0x2f2f715e, 0x84849713,
+    0x5353f5a6, 0xd1d168b9, 0x00000000, 0xeded2cc1,
+    0x20206040, 0xfcfc1fe3, 0xb1b1c879, 0x5b5bedb6,
+    0x6a6abed4, 0xcbcb468d, 0xbebed967, 0x39394b72,
+    0x4a4ade94, 0x4c4cd498, 0x5858e8b0, 0xcfcf4a85,
+    0xd0d06bbb, 0xefef2ac5, 0xaaaae54f, 0xfbfb16ed,
+    0x4343c586, 0x4d4dd79a, 0x33335566, 0x85859411,
+    0x4545cf8a, 0xf9f910e9, 0x02020604, 0x7f7f81fe,
+    0x5050f0a0, 0x3c3c4478, 0x9f9fba25, 0xa8a8e34b,
+    0x5151f3a2, 0xa3a3fe5d, 0x4040c080, 0x8f8f8a05,
+    0x9292ad3f, 0x9d9dbc21, 0x38384870, 0xf5f504f1,
+    0xbcbcdf63, 0xb6b6c177, 0xdada75af, 0x21216342,
+    0x10103020, 0xffff1ae5, 0xf3f30efd, 0xd2d26dbf,
+    0xcdcd4c81, 0x0c0c1418, 0x13133526, 0xecec2fc3,
+    0x5f5fe1be, 0x9797a235, 0x4444cc88, 0x1717392e,
+    0xc4c45793, 0xa7a7f255, 0x7e7e82fc, 0x3d3d477a,
+    0x6464acc8, 0x5d5de7ba, 0x19192b32, 0x737395e6,
+    0x6060a0c0, 0x81819819, 0x4f4fd19e, 0xdcdc7fa3,
+    0x22226644, 0x2a2a7e54, 0x9090ab3b, 0x8888830b,
+    0x4646ca8c, 0xeeee29c7, 0xb8b8d36b, 0x14143c28,
+    0xdede79a7, 0x5e5ee2bc, 0x0b0b1d16, 0xdbdb76ad,
+    0xe0e03bdb, 0x32325664, 0x3a3a4e74, 0x0a0a1e14,
+    0x4949db92, 0x06060a0c, 0x24246c48, 0x5c5ce4b8,
+    0xc2c25d9f, 0xd3d36ebd, 0xacacef43, 0x6262a6c4,
+    0x9191a839, 0x9595a431, 0xe4e437d3, 0x79798bf2,
+    0xe7e732d5, 0xc8c8438b, 0x3737596e, 0x6d6db7da,
+    0x8d8d8c01, 0xd5d564b1, 0x4e4ed29c, 0xa9a9e049,
+    0x6c6cb4d8, 0x5656faac, 0xf4f407f3, 0xeaea25cf,
+    0x6565afca, 0x7a7a8ef4, 0xaeaee947, 0x08081810,
+    0xbabad56f, 0x787888f0, 0x25256f4a, 0x2e2e725c,
+    0x1c1c2438, 0xa6a6f157, 0xb4b4c773, 0xc6c65197,
+    0xe8e823cb, 0xdddd7ca1, 0x74749ce8, 0x1f1f213e,
+    0x4b4bdd96, 0xbdbddc61, 0x8b8b860d, 0x8a8a850f,
+    0x707090e0, 0x3e3e427c, 0xb5b5c471, 0x6666aacc,
+    0x4848d890, 0x03030506, 0xf6f601f7, 0x0e0e121c,
+    0x6161a3c2, 0x35355f6a, 0x5757f9ae, 0xb9b9d069,
+    0x86869117, 0xc1c15899, 0x1d1d273a, 0x9e9eb927,
+    0xe1e138d9, 0xf8f813eb, 0x9898b32b, 0x11113322,
+    0x6969bbd2, 0xd9d970a9, 0x8e8e8907, 0x9494a733,
+    0x9b9bb62d, 0x1e1e223c, 0x87879215, 0xe9e920c9,
+    0xcece4987, 0x5555ffaa, 0x28287850, 0xdfdf7aa5,
+    0x8c8c8f03, 0xa1a1f859, 0x89898009, 0x0d0d171a,
+    0xbfbfda65, 0xe6e631d7, 0x4242c684, 0x6868b8d0,
+    0x4141c382, 0x9999b029, 0x2d2d775a, 0x0f0f111e,
+    0xb0b0cb7b, 0x5454fca8, 0xbbbbd66d, 0x16163a2c,
+};
+
+#if defined(HAVE_AESCCM) || defined(HAVE_AESGCM) || \
+    defined(WOLFSSL_AES_DIRECT) || defined(WOLFSSL_AES_COUNTER) || \
+    defined(HAVE_AES_ECB)
+void AES_ECB_encrypt(const unsigned char* in, unsigned char* out,
+    unsigned long len, const unsigned char* ks, int nr);
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+void AES_ECB_encrypt(const unsigned char* in_p, unsigned char* out_p,
+    unsigned long len_p, const unsigned char* ks_p, int nr_p)
+#else
+void AES_ECB_encrypt(const unsigned char* in, unsigned char* out,
+    unsigned long len, const unsigned char* ks, int nr)
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+{
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+    register const unsigned char* in asm ("3") = (const unsigned char*)in_p;
+    register unsigned char* out asm ("4") = (unsigned char*)out_p;
+    register unsigned long len asm ("5") = (unsigned long)len_p;
+    register const unsigned char* ks asm ("6") = (const unsigned char*)ks_p;
+    register int nr asm ("7") = (int)nr_p;
+    register word32* L_AES_PPC64_te4_0_c asm ("8") =
+        (word32*)&L_AES_PPC64_te4_0;
+#else
+    register word32* L_AES_PPC64_te4_0_c = (word32*)&L_AES_PPC64_te4_0;
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+
+    __asm__ __volatile__ (
+#ifdef WOLFSSL_NO_VAR_ASSIGN_REG
+        "mr      8, %[L_AES_PPC64_te4_0]\n\t"
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+        "addi    9, 8, 0x400\n\t"
+        "addi    10, 8, 0x800\n\t"
+        "addi    11, 8, 0xc00\n\t"
+        "\n"
+    "L_AES_ECB_encrypt_loop_block_128_%=: \n\t"
+        "addi    25, %[ks], 0\n\t"
+        "ld      12, 0(%[in])\n\t"
+        "ld      14, 8(%[in])\n\t"
+        "ld      17, 0(25)\n\t"
+        "ld      18, 8(25)\n\t"
+        "addi    25, 25, 16\n\t"
+        "rldicr  12, 12, 32, 63\n\t"
+        "rldicr  14, 14, 32, 63\n\t"
+        /* Round: 0 - XOR in key schedule */
+        "xor     12, 12, 17\n\t"
+        "xor     14, 14, 18\n\t"
+        "addi    0, %[nr], -2\n\t"
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+        "ld      20, 0(8)\n\t"
+        "ld      20, 64(8)\n\t"
+        "ld      20, 128(8)\n\t"
+        "ld      20, 192(8)\n\t"
+        "ld      20, 256(8)\n\t"
+        "ld      20, 320(8)\n\t"
+        "ld      20, 384(8)\n\t"
+        "ld      20, 448(8)\n\t"
+        "ld      20, 512(8)\n\t"
+        "ld      20, 576(8)\n\t"
+        "ld      20, 640(8)\n\t"
+        "ld      20, 704(8)\n\t"
+        "ld      20, 768(8)\n\t"
+        "ld      20, 832(8)\n\t"
+        "ld      20, 896(8)\n\t"
+        "ld      20, 960(8)\n\t"
+        "ld      20, 1024(8)\n\t"
+        "ld      20, 1088(8)\n\t"
+        "ld      20, 1152(8)\n\t"
+        "ld      20, 1216(8)\n\t"
+        "ld      20, 1280(8)\n\t"
+        "ld      20, 1344(8)\n\t"
+        "ld      20, 1408(8)\n\t"
+        "ld      20, 1472(8)\n\t"
+        "ld      20, 1536(8)\n\t"
+        "ld      20, 1600(8)\n\t"
+        "ld      20, 1664(8)\n\t"
+        "ld      20, 1728(8)\n\t"
+        "ld      20, 1792(8)\n\t"
+        "ld      20, 1856(8)\n\t"
+        "ld      20, 1920(8)\n\t"
+        "ld      20, 1984(8)\n\t"
+        "ld      20, 2048(8)\n\t"
+        "ld      20, 2112(8)\n\t"
+        "ld      20, 2176(8)\n\t"
+        "ld      20, 2240(8)\n\t"
+        "ld      20, 2304(8)\n\t"
+        "ld      20, 2368(8)\n\t"
+        "ld      20, 2432(8)\n\t"
+        "ld      20, 2496(8)\n\t"
+        "ld      20, 2560(8)\n\t"
+        "ld      20, 2624(8)\n\t"
+        "ld      20, 2688(8)\n\t"
+        "ld      20, 2752(8)\n\t"
+        "ld      20, 2816(8)\n\t"
+        "ld      20, 2880(8)\n\t"
+        "ld      20, 2944(8)\n\t"
+        "ld      20, 3008(8)\n\t"
+        "ld      20, 3072(8)\n\t"
+        "ld      20, 3136(8)\n\t"
+        "ld      20, 3200(8)\n\t"
+        "ld      20, 3264(8)\n\t"
+        "ld      20, 3328(8)\n\t"
+        "ld      20, 3392(8)\n\t"
+        "ld      20, 3456(8)\n\t"
+        "ld      20, 3520(8)\n\t"
+        "ld      20, 3584(8)\n\t"
+        "ld      20, 3648(8)\n\t"
+        "ld      20, 3712(8)\n\t"
+        "ld      20, 3776(8)\n\t"
+        "ld      20, 3840(8)\n\t"
+        "ld      20, 3904(8)\n\t"
+        "ld      20, 3968(8)\n\t"
+        "ld      20, 4032(8)\n\t"
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+        "srwi    0, 0, 1\n\t"
+        "mtctr   0\n\t"
+        "\n"
+    "L_AES_ECB_encrypt_loop_nr_%=: \n\t"
+        "srdi    15, 12, 32\n\t"
+        "srdi    16, 14, 32\n\t"
+        "rlwinm  17, 12, 10, 22, 29\n\t"
+        "rlwinm  18, 14, 10, 22, 29\n\t"
+        "rlwinm  19, 15, 10, 22, 29\n\t"
+        "rlwinm  20, 16, 10, 22, 29\n\t"
+        "lwzx    17, 8, 17\n\t"
+        "lwzx    18, 8, 18\n\t"
+        "lwzx    19, 8, 19\n\t"
+        "lwzx    20, 8, 20\n\t"
+        "rlwinm  21, 15, 18, 22, 29\n\t"
+        "rlwinm  22, 14, 18, 22, 29\n\t"
+        "rlwinm  23, 16, 18, 22, 29\n\t"
+        "rlwinm  24, 12, 18, 22, 29\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "lwzx    24, 9, 24\n\t"
+        "xor     17, 17, 21\n\t"
+        "xor     19, 19, 22\n\t"
+        "xor     18, 18, 23\n\t"
+        "xor     20, 20, 24\n\t"
+        "rlwinm  21, 14, 26, 22, 29\n\t"
+        "rlwinm  22, 16, 26, 22, 29\n\t"
+        "rlwinm  23, 12, 26, 22, 29\n\t"
+        "rlwinm  24, 15, 26, 22, 29\n\t"
+        "lwzx    21, 10, 21\n\t"
+        "lwzx    22, 10, 22\n\t"
+        "lwzx    23, 10, 23\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "xor     17, 17, 21\n\t"
+        "xor     19, 19, 22\n\t"
+        "xor     18, 18, 23\n\t"
+        "xor     20, 20, 24\n\t"
+        "rlwinm  21, 16, 2, 22, 29\n\t"
+        "rlwinm  22, 12, 2, 22, 29\n\t"
+        "rlwinm  23, 15, 2, 22, 29\n\t"
+        "rlwinm  24, 14, 2, 22, 29\n\t"
+        "lwzx    21, 11, 21\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "xor     17, 17, 21\n\t"
+        "xor     19, 19, 22\n\t"
+        "xor     18, 18, 23\n\t"
+        "xor     20, 20, 24\n\t"
+        "ld      12, 0(25)\n\t"
+        "ld      14, 8(25)\n\t"
+        "rldimi  17, 19, 32, 0\n\t"
+        "rldimi  18, 20, 32, 0\n\t"
+        "addi    25, 25, 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     17, 17, 12\n\t"
+        "xor     18, 18, 14\n\t"
+        "srdi    19, 17, 32\n\t"
+        "srdi    20, 18, 32\n\t"
+        "rlwinm  12, 17, 10, 22, 29\n\t"
+        "rlwinm  14, 18, 10, 22, 29\n\t"
+        "rlwinm  15, 19, 10, 22, 29\n\t"
+        "rlwinm  16, 20, 10, 22, 29\n\t"
+        "lwzx    12, 8, 12\n\t"
+        "lwzx    14, 8, 14\n\t"
+        "lwzx    15, 8, 15\n\t"
+        "lwzx    16, 8, 16\n\t"
+        "rlwinm  21, 19, 18, 22, 29\n\t"
+        "rlwinm  22, 18, 18, 22, 29\n\t"
+        "rlwinm  23, 20, 18, 22, 29\n\t"
+        "rlwinm  24, 17, 18, 22, 29\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "lwzx    24, 9, 24\n\t"
+        "xor     12, 12, 21\n\t"
+        "xor     15, 15, 22\n\t"
+        "xor     14, 14, 23\n\t"
+        "xor     16, 16, 24\n\t"
+        "rlwinm  21, 18, 26, 22, 29\n\t"
+        "rlwinm  22, 20, 26, 22, 29\n\t"
+        "rlwinm  23, 17, 26, 22, 29\n\t"
+        "rlwinm  24, 19, 26, 22, 29\n\t"
+        "lwzx    21, 10, 21\n\t"
+        "lwzx    22, 10, 22\n\t"
+        "lwzx    23, 10, 23\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "xor     12, 12, 21\n\t"
+        "xor     15, 15, 22\n\t"
+        "xor     14, 14, 23\n\t"
+        "xor     16, 16, 24\n\t"
+        "rlwinm  21, 20, 2, 22, 29\n\t"
+        "rlwinm  22, 17, 2, 22, 29\n\t"
+        "rlwinm  23, 19, 2, 22, 29\n\t"
+        "rlwinm  24, 18, 2, 22, 29\n\t"
+        "lwzx    21, 11, 21\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "xor     12, 12, 21\n\t"
+        "xor     15, 15, 22\n\t"
+        "xor     14, 14, 23\n\t"
+        "xor     16, 16, 24\n\t"
+        "ld      17, 0(25)\n\t"
+        "ld      18, 8(25)\n\t"
+        "rldimi  12, 15, 32, 0\n\t"
+        "rldimi  14, 16, 32, 0\n\t"
+        "addi    25, 25, 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     12, 12, 17\n\t"
+        "xor     14, 14, 18\n\t"
+        "bdnz    L_AES_ECB_encrypt_loop_nr_%=\n\t"
+        "srdi    15, 12, 32\n\t"
+        "srdi    16, 14, 32\n\t"
+        "rlwinm  17, 12, 10, 22, 29\n\t"
+        "rlwinm  18, 14, 10, 22, 29\n\t"
+        "rlwinm  19, 15, 10, 22, 29\n\t"
+        "rlwinm  20, 16, 10, 22, 29\n\t"
+        "lwzx    17, 8, 17\n\t"
+        "lwzx    18, 8, 18\n\t"
+        "lwzx    19, 8, 19\n\t"
+        "lwzx    20, 8, 20\n\t"
+        "rlwinm  21, 15, 18, 22, 29\n\t"
+        "rlwinm  22, 14, 18, 22, 29\n\t"
+        "rlwinm  23, 16, 18, 22, 29\n\t"
+        "rlwinm  24, 12, 18, 22, 29\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "lwzx    24, 9, 24\n\t"
+        "xor     17, 17, 21\n\t"
+        "xor     19, 19, 22\n\t"
+        "xor     18, 18, 23\n\t"
+        "xor     20, 20, 24\n\t"
+        "rlwinm  21, 14, 26, 22, 29\n\t"
+        "rlwinm  22, 16, 26, 22, 29\n\t"
+        "rlwinm  23, 12, 26, 22, 29\n\t"
+        "rlwinm  24, 15, 26, 22, 29\n\t"
+        "lwzx    21, 10, 21\n\t"
+        "lwzx    22, 10, 22\n\t"
+        "lwzx    23, 10, 23\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "xor     17, 17, 21\n\t"
+        "xor     19, 19, 22\n\t"
+        "xor     18, 18, 23\n\t"
+        "xor     20, 20, 24\n\t"
+        "rlwinm  21, 16, 2, 22, 29\n\t"
+        "rlwinm  22, 12, 2, 22, 29\n\t"
+        "rlwinm  23, 15, 2, 22, 29\n\t"
+        "rlwinm  24, 14, 2, 22, 29\n\t"
+        "lwzx    21, 11, 21\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "xor     17, 17, 21\n\t"
+        "xor     19, 19, 22\n\t"
+        "xor     18, 18, 23\n\t"
+        "xor     20, 20, 24\n\t"
+        "ld      12, 0(25)\n\t"
+        "ld      14, 8(25)\n\t"
+        "rldimi  17, 19, 32, 0\n\t"
+        "rldimi  18, 20, 32, 0\n\t"
+        "addi    25, 25, 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     17, 17, 12\n\t"
+        "xor     18, 18, 14\n\t"
+        "srdi    19, 17, 32\n\t"
+        "srdi    20, 18, 32\n\t"
+        "rlwinm  12, 17, 2, 22, 29\n\t"
+        "rlwinm  14, 18, 2, 22, 29\n\t"
+        "rlwinm  16, 19, 2, 22, 29\n\t"
+        "rlwinm  15, 20, 2, 22, 29\n\t"
+        "lwzx    12, 9, 12\n\t"
+        "lwzx    14, 9, 14\n\t"
+        "lwzx    16, 9, 16\n\t"
+        "lwzx    15, 9, 15\n\t"
+        "rlwinm  21, 18, 26, 22, 29\n\t"
+        "rlwinm  22, 20, 26, 22, 29\n\t"
+        "rldimi  14, 16, 32, 24\n\t"
+        "rldimi  12, 15, 32, 24\n\t"
+        "rlwinm  23, 17, 26, 22, 29\n\t"
+        "rlwinm  24, 19, 26, 22, 29\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "lwzx    24, 9, 24\n\t"
+        "rldimi  12, 21, 40, 16\n\t"
+        "rlwimi  12, 22, 8, 16, 23\n\t"
+        "rldimi  14, 23, 40, 16\n\t"
+        "rlwimi  14, 24, 8, 16, 23\n\t"
+        "rlwinm  21, 19, 18, 22, 29\n\t"
+        "rlwinm  22, 18, 18, 22, 29\n\t"
+        "rlwinm  23, 20, 18, 22, 29\n\t"
+        "rlwinm  24, 17, 18, 22, 29\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "lwzx    24, 9, 24\n\t"
+        "rldimi  12, 21, 48, 8\n\t"
+        "rlwimi  12, 22, 16, 8, 15\n\t"
+        "rldimi  14, 23, 48, 8\n\t"
+        "rlwimi  14, 24, 16, 8, 15\n\t"
+        "rlwinm  21, 17, 10, 22, 29\n\t"
+        "rlwinm  22, 19, 10, 22, 29\n\t"
+        "rlwinm  23, 18, 10, 22, 29\n\t"
+        "rlwinm  24, 20, 10, 22, 29\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "lwzx    24, 9, 24\n\t"
+        "ld      17, 0(25)\n\t"
+        "ld      18, 8(25)\n\t"
+        "rldimi  12, 21, 56, 0\n\t"
+        "rlwimi  12, 22, 24, 0, 7\n\t"
+        "rldimi  14, 23, 56, 0\n\t"
+        "rlwimi  14, 24, 24, 0, 7\n\t"
+        "addi    25, 25, 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     12, 12, 17\n\t"
+        "xor     14, 14, 18\n\t"
+        "std     12, 0(%[out])\n\t"
+        "std     14, 8(%[out])\n\t"
+        "addi    %[in], %[in], 16\n\t"
+        "addi    %[out], %[out], 16\n\t"
+        "addic.  %[len], %[len], -16\n\t"
+        "bne     L_AES_ECB_encrypt_loop_block_128_%=\n\t"
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+        : [in] "+r" (in), [out] "+r" (out), [len] "+r" (len), [ks] "+r" (ks),
+          [nr] "+r" (nr), [L_AES_PPC64_te4_0] "+r" (L_AES_PPC64_te4_0_c)
+        :
+        : "memory", "cc", "0", "9", "10", "11", "12", "14", "15", "16", "17",
+            "18", "19", "20", "21", "22", "23", "24", "25"
+#else
+        :
+        : [in] "r" (in), [out] "r" (out), [len] "r" (len), [ks] "r" (ks),
+          [nr] "r" (nr), [L_AES_PPC64_te4_0] "r" (L_AES_PPC64_te4_0_c)
+        : "memory", "cc", "0", "8", "9", "10", "11", "12", "14", "15", "16",
+            "17", "18", "19", "20", "21", "22", "23", "24", "25"
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+    );
+}
+
+#endif /* HAVE_AESCCM || HAVE_AESGCM || WOLFSSL_AES_DIRECT ||
+        * WOLFSSL_AES_COUNTER || HAVE_AES_ECB */
+#ifdef HAVE_AES_CBC
+void AES_CBC_encrypt(const unsigned char* in, unsigned char* out,
+    unsigned long len, const unsigned char* ks, int nr, unsigned char* iv);
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+void AES_CBC_encrypt(const unsigned char* in_p, unsigned char* out_p,
+    unsigned long len_p, const unsigned char* ks_p, int nr_p,
+    unsigned char* iv_p)
+#else
+void AES_CBC_encrypt(const unsigned char* in, unsigned char* out,
+    unsigned long len, const unsigned char* ks, int nr, unsigned char* iv)
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+{
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+    register const unsigned char* in asm ("3") = (const unsigned char*)in_p;
+    register unsigned char* out asm ("4") = (unsigned char*)out_p;
+    register unsigned long len asm ("5") = (unsigned long)len_p;
+    register const unsigned char* ks asm ("6") = (const unsigned char*)ks_p;
+    register int nr asm ("7") = (int)nr_p;
+    register unsigned char* iv asm ("8") = (unsigned char*)iv_p;
+    register word32* L_AES_PPC64_te4_0_c asm ("9") =
+        (word32*)&L_AES_PPC64_te4_0;
+#else
+    register word32* L_AES_PPC64_te4_0_c = (word32*)&L_AES_PPC64_te4_0;
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+
+    __asm__ __volatile__ (
+#ifdef WOLFSSL_NO_VAR_ASSIGN_REG
+        "mr      9, %[L_AES_PPC64_te4_0]\n\t"
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+        "ld      14, 0(%[iv])\n\t"
+        "ld      15, 8(%[iv])\n\t"
+        "addi    10, 9, 0x400\n\t"
+        "addi    11, 9, 0x800\n\t"
+        "addi    12, 9, 0xc00\n\t"
+        "\n"
+    "L_AES_CBC_encrypt_loop_block_%=: \n\t"
+        "addi    26, %[ks], 0\n\t"
+        "ld      18, 0(%[in])\n\t"
+        "ld      19, 8(%[in])\n\t"
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "ld      18, 0(26)\n\t"
+        "ld      19, 8(26)\n\t"
+        "addi    26, 26, 16\n\t"
+        "rldicr  14, 14, 32, 63\n\t"
+        "rldicr  15, 15, 32, 63\n\t"
+        /* Round: 0 - XOR in key schedule */
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "addi    0, %[nr], -2\n\t"
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+        "ld      21, 0(9)\n\t"
+        "ld      21, 64(9)\n\t"
+        "ld      21, 128(9)\n\t"
+        "ld      21, 192(9)\n\t"
+        "ld      21, 256(9)\n\t"
+        "ld      21, 320(9)\n\t"
+        "ld      21, 384(9)\n\t"
+        "ld      21, 448(9)\n\t"
+        "ld      21, 512(9)\n\t"
+        "ld      21, 576(9)\n\t"
+        "ld      21, 640(9)\n\t"
+        "ld      21, 704(9)\n\t"
+        "ld      21, 768(9)\n\t"
+        "ld      21, 832(9)\n\t"
+        "ld      21, 896(9)\n\t"
+        "ld      21, 960(9)\n\t"
+        "ld      21, 1024(9)\n\t"
+        "ld      21, 1088(9)\n\t"
+        "ld      21, 1152(9)\n\t"
+        "ld      21, 1216(9)\n\t"
+        "ld      21, 1280(9)\n\t"
+        "ld      21, 1344(9)\n\t"
+        "ld      21, 1408(9)\n\t"
+        "ld      21, 1472(9)\n\t"
+        "ld      21, 1536(9)\n\t"
+        "ld      21, 1600(9)\n\t"
+        "ld      21, 1664(9)\n\t"
+        "ld      21, 1728(9)\n\t"
+        "ld      21, 1792(9)\n\t"
+        "ld      21, 1856(9)\n\t"
+        "ld      21, 1920(9)\n\t"
+        "ld      21, 1984(9)\n\t"
+        "ld      21, 2048(9)\n\t"
+        "ld      21, 2112(9)\n\t"
+        "ld      21, 2176(9)\n\t"
+        "ld      21, 2240(9)\n\t"
+        "ld      21, 2304(9)\n\t"
+        "ld      21, 2368(9)\n\t"
+        "ld      21, 2432(9)\n\t"
+        "ld      21, 2496(9)\n\t"
+        "ld      21, 2560(9)\n\t"
+        "ld      21, 2624(9)\n\t"
+        "ld      21, 2688(9)\n\t"
+        "ld      21, 2752(9)\n\t"
+        "ld      21, 2816(9)\n\t"
+        "ld      21, 2880(9)\n\t"
+        "ld      21, 2944(9)\n\t"
+        "ld      21, 3008(9)\n\t"
+        "ld      21, 3072(9)\n\t"
+        "ld      21, 3136(9)\n\t"
+        "ld      21, 3200(9)\n\t"
+        "ld      21, 3264(9)\n\t"
+        "ld      21, 3328(9)\n\t"
+        "ld      21, 3392(9)\n\t"
+        "ld      21, 3456(9)\n\t"
+        "ld      21, 3520(9)\n\t"
+        "ld      21, 3584(9)\n\t"
+        "ld      21, 3648(9)\n\t"
+        "ld      21, 3712(9)\n\t"
+        "ld      21, 3776(9)\n\t"
+        "ld      21, 3840(9)\n\t"
+        "ld      21, 3904(9)\n\t"
+        "ld      21, 3968(9)\n\t"
+        "ld      21, 4032(9)\n\t"
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+        "srwi    0, 0, 1\n\t"
+        "mtctr   0\n\t"
+        "\n"
+    "L_AES_CBC_encrypt_loop_nr_%=: \n\t"
+        "srdi    16, 14, 32\n\t"
+        "srdi    17, 15, 32\n\t"
+        "rlwinm  18, 14, 10, 22, 29\n\t"
+        "rlwinm  19, 15, 10, 22, 29\n\t"
+        "rlwinm  20, 16, 10, 22, 29\n\t"
+        "rlwinm  21, 17, 10, 22, 29\n\t"
+        "lwzx    18, 9, 18\n\t"
+        "lwzx    19, 9, 19\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "rlwinm  22, 16, 18, 22, 29\n\t"
+        "rlwinm  23, 15, 18, 22, 29\n\t"
+        "rlwinm  24, 17, 18, 22, 29\n\t"
+        "rlwinm  25, 14, 18, 22, 29\n\t"
+        "lwzx    22, 10, 22\n\t"
+        "lwzx    23, 10, 23\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "lwzx    25, 10, 25\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 15, 26, 22, 29\n\t"
+        "rlwinm  23, 17, 26, 22, 29\n\t"
+        "rlwinm  24, 14, 26, 22, 29\n\t"
+        "rlwinm  25, 16, 26, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 17, 2, 22, 29\n\t"
+        "rlwinm  23, 14, 2, 22, 29\n\t"
+        "rlwinm  24, 16, 2, 22, 29\n\t"
+        "rlwinm  25, 15, 2, 22, 29\n\t"
+        "lwzx    22, 12, 22\n\t"
+        "lwzx    23, 12, 23\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "ld      14, 0(26)\n\t"
+        "ld      15, 8(26)\n\t"
+        "rldimi  18, 20, 32, 0\n\t"
+        "rldimi  19, 21, 32, 0\n\t"
+        "addi    26, 26, 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     18, 18, 14\n\t"
+        "xor     19, 19, 15\n\t"
+        "srdi    20, 18, 32\n\t"
+        "srdi    21, 19, 32\n\t"
+        "rlwinm  14, 18, 10, 22, 29\n\t"
+        "rlwinm  15, 19, 10, 22, 29\n\t"
+        "rlwinm  16, 20, 10, 22, 29\n\t"
+        "rlwinm  17, 21, 10, 22, 29\n\t"
+        "lwzx    14, 9, 14\n\t"
+        "lwzx    15, 9, 15\n\t"
+        "lwzx    16, 9, 16\n\t"
+        "lwzx    17, 9, 17\n\t"
+        "rlwinm  22, 20, 18, 22, 29\n\t"
+        "rlwinm  23, 19, 18, 22, 29\n\t"
+        "rlwinm  24, 21, 18, 22, 29\n\t"
+        "rlwinm  25, 18, 18, 22, 29\n\t"
+        "lwzx    22, 10, 22\n\t"
+        "lwzx    23, 10, 23\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "lwzx    25, 10, 25\n\t"
+        "xor     14, 14, 22\n\t"
+        "xor     16, 16, 23\n\t"
+        "xor     15, 15, 24\n\t"
+        "xor     17, 17, 25\n\t"
+        "rlwinm  22, 19, 26, 22, 29\n\t"
+        "rlwinm  23, 21, 26, 22, 29\n\t"
+        "rlwinm  24, 18, 26, 22, 29\n\t"
+        "rlwinm  25, 20, 26, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "xor     14, 14, 22\n\t"
+        "xor     16, 16, 23\n\t"
+        "xor     15, 15, 24\n\t"
+        "xor     17, 17, 25\n\t"
+        "rlwinm  22, 21, 2, 22, 29\n\t"
+        "rlwinm  23, 18, 2, 22, 29\n\t"
+        "rlwinm  24, 20, 2, 22, 29\n\t"
+        "rlwinm  25, 19, 2, 22, 29\n\t"
+        "lwzx    22, 12, 22\n\t"
+        "lwzx    23, 12, 23\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "xor     14, 14, 22\n\t"
+        "xor     16, 16, 23\n\t"
+        "xor     15, 15, 24\n\t"
+        "xor     17, 17, 25\n\t"
+        "ld      18, 0(26)\n\t"
+        "ld      19, 8(26)\n\t"
+        "rldimi  14, 16, 32, 0\n\t"
+        "rldimi  15, 17, 32, 0\n\t"
+        "addi    26, 26, 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "bdnz    L_AES_CBC_encrypt_loop_nr_%=\n\t"
+        "srdi    16, 14, 32\n\t"
+        "srdi    17, 15, 32\n\t"
+        "rlwinm  18, 14, 10, 22, 29\n\t"
+        "rlwinm  19, 15, 10, 22, 29\n\t"
+        "rlwinm  20, 16, 10, 22, 29\n\t"
+        "rlwinm  21, 17, 10, 22, 29\n\t"
+        "lwzx    18, 9, 18\n\t"
+        "lwzx    19, 9, 19\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "rlwinm  22, 16, 18, 22, 29\n\t"
+        "rlwinm  23, 15, 18, 22, 29\n\t"
+        "rlwinm  24, 17, 18, 22, 29\n\t"
+        "rlwinm  25, 14, 18, 22, 29\n\t"
+        "lwzx    22, 10, 22\n\t"
+        "lwzx    23, 10, 23\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "lwzx    25, 10, 25\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 15, 26, 22, 29\n\t"
+        "rlwinm  23, 17, 26, 22, 29\n\t"
+        "rlwinm  24, 14, 26, 22, 29\n\t"
+        "rlwinm  25, 16, 26, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 17, 2, 22, 29\n\t"
+        "rlwinm  23, 14, 2, 22, 29\n\t"
+        "rlwinm  24, 16, 2, 22, 29\n\t"
+        "rlwinm  25, 15, 2, 22, 29\n\t"
+        "lwzx    22, 12, 22\n\t"
+        "lwzx    23, 12, 23\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "ld      14, 0(26)\n\t"
+        "ld      15, 8(26)\n\t"
+        "rldimi  18, 20, 32, 0\n\t"
+        "rldimi  19, 21, 32, 0\n\t"
+        "addi    26, 26, 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     18, 18, 14\n\t"
+        "xor     19, 19, 15\n\t"
+        "srdi    20, 18, 32\n\t"
+        "srdi    21, 19, 32\n\t"
+        "rlwinm  14, 18, 2, 22, 29\n\t"
+        "rlwinm  15, 19, 2, 22, 29\n\t"
+        "rlwinm  17, 20, 2, 22, 29\n\t"
+        "rlwinm  16, 21, 2, 22, 29\n\t"
+        "lwzx    14, 10, 14\n\t"
+        "lwzx    15, 10, 15\n\t"
+        "lwzx    17, 10, 17\n\t"
+        "lwzx    16, 10, 16\n\t"
+        "rlwinm  22, 19, 26, 22, 29\n\t"
+        "rlwinm  23, 21, 26, 22, 29\n\t"
+        "rldimi  15, 17, 32, 24\n\t"
+        "rldimi  14, 16, 32, 24\n\t"
+        "rlwinm  24, 18, 26, 22, 29\n\t"
+        "rlwinm  25, 20, 26, 22, 29\n\t"
+        "lwzx    22, 10, 22\n\t"
+        "lwzx    23, 10, 23\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "lwzx    25, 10, 25\n\t"
+        "rldimi  14, 22, 40, 16\n\t"
+        "rlwimi  14, 23, 8, 16, 23\n\t"
+        "rldimi  15, 24, 40, 16\n\t"
+        "rlwimi  15, 25, 8, 16, 23\n\t"
+        "rlwinm  22, 20, 18, 22, 29\n\t"
+        "rlwinm  23, 19, 18, 22, 29\n\t"
+        "rlwinm  24, 21, 18, 22, 29\n\t"
+        "rlwinm  25, 18, 18, 22, 29\n\t"
+        "lwzx    22, 10, 22\n\t"
+        "lwzx    23, 10, 23\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "lwzx    25, 10, 25\n\t"
+        "rldimi  14, 22, 48, 8\n\t"
+        "rlwimi  14, 23, 16, 8, 15\n\t"
+        "rldimi  15, 24, 48, 8\n\t"
+        "rlwimi  15, 25, 16, 8, 15\n\t"
+        "rlwinm  22, 18, 10, 22, 29\n\t"
+        "rlwinm  23, 20, 10, 22, 29\n\t"
+        "rlwinm  24, 19, 10, 22, 29\n\t"
+        "rlwinm  25, 21, 10, 22, 29\n\t"
+        "lwzx    22, 10, 22\n\t"
+        "lwzx    23, 10, 23\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "lwzx    25, 10, 25\n\t"
+        "ld      18, 0(26)\n\t"
+        "ld      19, 8(26)\n\t"
+        "rldimi  14, 22, 56, 0\n\t"
+        "rlwimi  14, 23, 24, 0, 7\n\t"
+        "rldimi  15, 24, 56, 0\n\t"
+        "rlwimi  15, 25, 24, 0, 7\n\t"
+        "addi    26, 26, 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "std     14, 0(%[out])\n\t"
+        "std     15, 8(%[out])\n\t"
+        "addi    %[in], %[in], 16\n\t"
+        "addi    %[out], %[out], 16\n\t"
+        "addic.  %[len], %[len], -16\n\t"
+        "bne     L_AES_CBC_encrypt_loop_block_%=\n\t"
+        "std     14, 0(%[iv])\n\t"
+        "std     15, 8(%[iv])\n\t"
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+        : [in] "+r" (in), [out] "+r" (out), [len] "+r" (len), [ks] "+r" (ks),
+          [nr] "+r" (nr), [iv] "+r" (iv),
+          [L_AES_PPC64_te4_0] "+r" (L_AES_PPC64_te4_0_c)
+        :
+        : "memory", "cc", "0", "10", "11", "12", "14", "15", "16", "17", "18",
+            "19", "20", "21", "22", "23", "24", "25", "26"
+#else
+        :
+        : [in] "r" (in), [out] "r" (out), [len] "r" (len), [ks] "r" (ks),
+          [nr] "r" (nr), [iv] "r" (iv),
+          [L_AES_PPC64_te4_0] "r" (L_AES_PPC64_te4_0_c)
+        : "memory", "cc", "0", "9", "10", "11", "12", "14", "15", "16", "17",
+            "18", "19", "20", "21", "22", "23", "24", "25", "26"
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+    );
+}
+
+#endif /* HAVE_AES_CBC */
+#ifdef WOLFSSL_AES_COUNTER
+void AES_CTR_encrypt(const unsigned char* in, unsigned char* out,
+    unsigned long len, const unsigned char* ks, int nr, unsigned char* ctr);
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+void AES_CTR_encrypt(const unsigned char* in_p, unsigned char* out_p,
+    unsigned long len_p, const unsigned char* ks_p, int nr_p,
+    unsigned char* ctr_p)
+#else
+void AES_CTR_encrypt(const unsigned char* in, unsigned char* out,
+    unsigned long len, const unsigned char* ks, int nr, unsigned char* ctr)
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+{
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+    register const unsigned char* in asm ("3") = (const unsigned char*)in_p;
+    register unsigned char* out asm ("4") = (unsigned char*)out_p;
+    register unsigned long len asm ("5") = (unsigned long)len_p;
+    register const unsigned char* ks asm ("6") = (const unsigned char*)ks_p;
+    register int nr asm ("7") = (int)nr_p;
+    register unsigned char* ctr asm ("8") = (unsigned char*)ctr_p;
+    register word32* L_AES_PPC64_te4_0_c asm ("9") =
+        (word32*)&L_AES_PPC64_te4_0;
+#else
+    register word32* L_AES_PPC64_te4_0_c = (word32*)&L_AES_PPC64_te4_0;
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+
+    __asm__ __volatile__ (
+#ifdef WOLFSSL_NO_VAR_ASSIGN_REG
+        "mr      9, %[L_AES_PPC64_te4_0]\n\t"
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+        "ld      22, 0(%[ctr])\n\t"
+        "ld      23, 8(%[ctr])\n\t"
+        "addi    10, 9, 0x400\n\t"
+        "addi    11, 9, 0x800\n\t"
+        "addi    12, 9, 0xc00\n\t"
+        "\n"
+    "L_AES_CTR_encrypt_loop_block_128_%=: \n\t"
+        "addi    28, %[ks], 0\n\t"
+        "ld      18, 0(28)\n\t"
+        "ld      19, 8(28)\n\t"
+        "addi    28, 28, 16\n\t"
+        "rldicr  14, 22, 32, 63\n\t"
+        "rldicr  15, 23, 32, 63\n\t"
+        /* Round: 0 - XOR in key schedule */
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "addi    0, %[nr], -2\n\t"
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+        "ld      21, 0(9)\n\t"
+        "ld      21, 64(9)\n\t"
+        "ld      21, 128(9)\n\t"
+        "ld      21, 192(9)\n\t"
+        "ld      21, 256(9)\n\t"
+        "ld      21, 320(9)\n\t"
+        "ld      21, 384(9)\n\t"
+        "ld      21, 448(9)\n\t"
+        "ld      21, 512(9)\n\t"
+        "ld      21, 576(9)\n\t"
+        "ld      21, 640(9)\n\t"
+        "ld      21, 704(9)\n\t"
+        "ld      21, 768(9)\n\t"
+        "ld      21, 832(9)\n\t"
+        "ld      21, 896(9)\n\t"
+        "ld      21, 960(9)\n\t"
+        "ld      21, 1024(9)\n\t"
+        "ld      21, 1088(9)\n\t"
+        "ld      21, 1152(9)\n\t"
+        "ld      21, 1216(9)\n\t"
+        "ld      21, 1280(9)\n\t"
+        "ld      21, 1344(9)\n\t"
+        "ld      21, 1408(9)\n\t"
+        "ld      21, 1472(9)\n\t"
+        "ld      21, 1536(9)\n\t"
+        "ld      21, 1600(9)\n\t"
+        "ld      21, 1664(9)\n\t"
+        "ld      21, 1728(9)\n\t"
+        "ld      21, 1792(9)\n\t"
+        "ld      21, 1856(9)\n\t"
+        "ld      21, 1920(9)\n\t"
+        "ld      21, 1984(9)\n\t"
+        "ld      21, 2048(9)\n\t"
+        "ld      21, 2112(9)\n\t"
+        "ld      21, 2176(9)\n\t"
+        "ld      21, 2240(9)\n\t"
+        "ld      21, 2304(9)\n\t"
+        "ld      21, 2368(9)\n\t"
+        "ld      21, 2432(9)\n\t"
+        "ld      21, 2496(9)\n\t"
+        "ld      21, 2560(9)\n\t"
+        "ld      21, 2624(9)\n\t"
+        "ld      21, 2688(9)\n\t"
+        "ld      21, 2752(9)\n\t"
+        "ld      21, 2816(9)\n\t"
+        "ld      21, 2880(9)\n\t"
+        "ld      21, 2944(9)\n\t"
+        "ld      21, 3008(9)\n\t"
+        "ld      21, 3072(9)\n\t"
+        "ld      21, 3136(9)\n\t"
+        "ld      21, 3200(9)\n\t"
+        "ld      21, 3264(9)\n\t"
+        "ld      21, 3328(9)\n\t"
+        "ld      21, 3392(9)\n\t"
+        "ld      21, 3456(9)\n\t"
+        "ld      21, 3520(9)\n\t"
+        "ld      21, 3584(9)\n\t"
+        "ld      21, 3648(9)\n\t"
+        "ld      21, 3712(9)\n\t"
+        "ld      21, 3776(9)\n\t"
+        "ld      21, 3840(9)\n\t"
+        "ld      21, 3904(9)\n\t"
+        "ld      21, 3968(9)\n\t"
+        "ld      21, 4032(9)\n\t"
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+        "srwi    0, 0, 1\n\t"
+        "mtctr   0\n\t"
+        "\n"
+    "L_AES_CTR_encrypt_loop_nr_%=: \n\t"
+        "srdi    16, 14, 32\n\t"
+        "srdi    17, 15, 32\n\t"
+        "rlwinm  18, 14, 10, 22, 29\n\t"
+        "rlwinm  19, 15, 10, 22, 29\n\t"
+        "rlwinm  20, 16, 10, 22, 29\n\t"
+        "rlwinm  21, 17, 10, 22, 29\n\t"
+        "lwzx    18, 9, 18\n\t"
+        "lwzx    19, 9, 19\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "rlwinm  24, 16, 18, 22, 29\n\t"
+        "rlwinm  25, 15, 18, 22, 29\n\t"
+        "rlwinm  26, 17, 18, 22, 29\n\t"
+        "rlwinm  27, 14, 18, 22, 29\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "lwzx    25, 10, 25\n\t"
+        "lwzx    26, 10, 26\n\t"
+        "lwzx    27, 10, 27\n\t"
+        "xor     18, 18, 24\n\t"
+        "xor     20, 20, 25\n\t"
+        "xor     19, 19, 26\n\t"
+        "xor     21, 21, 27\n\t"
+        "rlwinm  24, 15, 26, 22, 29\n\t"
+        "rlwinm  25, 17, 26, 22, 29\n\t"
+        "rlwinm  26, 14, 26, 22, 29\n\t"
+        "rlwinm  27, 16, 26, 22, 29\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "lwzx    26, 11, 26\n\t"
+        "lwzx    27, 11, 27\n\t"
+        "xor     18, 18, 24\n\t"
+        "xor     20, 20, 25\n\t"
+        "xor     19, 19, 26\n\t"
+        "xor     21, 21, 27\n\t"
+        "rlwinm  24, 17, 2, 22, 29\n\t"
+        "rlwinm  25, 14, 2, 22, 29\n\t"
+        "rlwinm  26, 16, 2, 22, 29\n\t"
+        "rlwinm  27, 15, 2, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "xor     18, 18, 24\n\t"
+        "xor     20, 20, 25\n\t"
+        "xor     19, 19, 26\n\t"
+        "xor     21, 21, 27\n\t"
+        "ld      14, 0(28)\n\t"
+        "ld      15, 8(28)\n\t"
+        "rldimi  18, 20, 32, 0\n\t"
+        "rldimi  19, 21, 32, 0\n\t"
+        "addi    28, 28, 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     18, 18, 14\n\t"
+        "xor     19, 19, 15\n\t"
+        "srdi    20, 18, 32\n\t"
+        "srdi    21, 19, 32\n\t"
+        "rlwinm  14, 18, 10, 22, 29\n\t"
+        "rlwinm  15, 19, 10, 22, 29\n\t"
+        "rlwinm  16, 20, 10, 22, 29\n\t"
+        "rlwinm  17, 21, 10, 22, 29\n\t"
+        "lwzx    14, 9, 14\n\t"
+        "lwzx    15, 9, 15\n\t"
+        "lwzx    16, 9, 16\n\t"
+        "lwzx    17, 9, 17\n\t"
+        "rlwinm  24, 20, 18, 22, 29\n\t"
+        "rlwinm  25, 19, 18, 22, 29\n\t"
+        "rlwinm  26, 21, 18, 22, 29\n\t"
+        "rlwinm  27, 18, 18, 22, 29\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "lwzx    25, 10, 25\n\t"
+        "lwzx    26, 10, 26\n\t"
+        "lwzx    27, 10, 27\n\t"
+        "xor     14, 14, 24\n\t"
+        "xor     16, 16, 25\n\t"
+        "xor     15, 15, 26\n\t"
+        "xor     17, 17, 27\n\t"
+        "rlwinm  24, 19, 26, 22, 29\n\t"
+        "rlwinm  25, 21, 26, 22, 29\n\t"
+        "rlwinm  26, 18, 26, 22, 29\n\t"
+        "rlwinm  27, 20, 26, 22, 29\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "lwzx    26, 11, 26\n\t"
+        "lwzx    27, 11, 27\n\t"
+        "xor     14, 14, 24\n\t"
+        "xor     16, 16, 25\n\t"
+        "xor     15, 15, 26\n\t"
+        "xor     17, 17, 27\n\t"
+        "rlwinm  24, 21, 2, 22, 29\n\t"
+        "rlwinm  25, 18, 2, 22, 29\n\t"
+        "rlwinm  26, 20, 2, 22, 29\n\t"
+        "rlwinm  27, 19, 2, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "xor     14, 14, 24\n\t"
+        "xor     16, 16, 25\n\t"
+        "xor     15, 15, 26\n\t"
+        "xor     17, 17, 27\n\t"
+        "ld      18, 0(28)\n\t"
+        "ld      19, 8(28)\n\t"
+        "rldimi  14, 16, 32, 0\n\t"
+        "rldimi  15, 17, 32, 0\n\t"
+        "addi    28, 28, 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "bdnz    L_AES_CTR_encrypt_loop_nr_%=\n\t"
+        "srdi    16, 14, 32\n\t"
+        "srdi    17, 15, 32\n\t"
+        "rlwinm  18, 14, 10, 22, 29\n\t"
+        "rlwinm  19, 15, 10, 22, 29\n\t"
+        "rlwinm  20, 16, 10, 22, 29\n\t"
+        "rlwinm  21, 17, 10, 22, 29\n\t"
+        "lwzx    18, 9, 18\n\t"
+        "lwzx    19, 9, 19\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "rlwinm  24, 16, 18, 22, 29\n\t"
+        "rlwinm  25, 15, 18, 22, 29\n\t"
+        "rlwinm  26, 17, 18, 22, 29\n\t"
+        "rlwinm  27, 14, 18, 22, 29\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "lwzx    25, 10, 25\n\t"
+        "lwzx    26, 10, 26\n\t"
+        "lwzx    27, 10, 27\n\t"
+        "xor     18, 18, 24\n\t"
+        "xor     20, 20, 25\n\t"
+        "xor     19, 19, 26\n\t"
+        "xor     21, 21, 27\n\t"
+        "rlwinm  24, 15, 26, 22, 29\n\t"
+        "rlwinm  25, 17, 26, 22, 29\n\t"
+        "rlwinm  26, 14, 26, 22, 29\n\t"
+        "rlwinm  27, 16, 26, 22, 29\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "lwzx    26, 11, 26\n\t"
+        "lwzx    27, 11, 27\n\t"
+        "xor     18, 18, 24\n\t"
+        "xor     20, 20, 25\n\t"
+        "xor     19, 19, 26\n\t"
+        "xor     21, 21, 27\n\t"
+        "rlwinm  24, 17, 2, 22, 29\n\t"
+        "rlwinm  25, 14, 2, 22, 29\n\t"
+        "rlwinm  26, 16, 2, 22, 29\n\t"
+        "rlwinm  27, 15, 2, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "xor     18, 18, 24\n\t"
+        "xor     20, 20, 25\n\t"
+        "xor     19, 19, 26\n\t"
+        "xor     21, 21, 27\n\t"
+        "ld      14, 0(28)\n\t"
+        "ld      15, 8(28)\n\t"
+        "rldimi  18, 20, 32, 0\n\t"
+        "rldimi  19, 21, 32, 0\n\t"
+        "addi    28, 28, 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     18, 18, 14\n\t"
+        "xor     19, 19, 15\n\t"
+        "srdi    20, 18, 32\n\t"
+        "srdi    21, 19, 32\n\t"
+        "rlwinm  14, 18, 2, 22, 29\n\t"
+        "rlwinm  15, 19, 2, 22, 29\n\t"
+        "rlwinm  17, 20, 2, 22, 29\n\t"
+        "rlwinm  16, 21, 2, 22, 29\n\t"
+        "lwzx    14, 10, 14\n\t"
+        "lwzx    15, 10, 15\n\t"
+        "lwzx    17, 10, 17\n\t"
+        "lwzx    16, 10, 16\n\t"
+        "rlwinm  24, 19, 26, 22, 29\n\t"
+        "rlwinm  25, 21, 26, 22, 29\n\t"
+        "rldimi  15, 17, 32, 24\n\t"
+        "rldimi  14, 16, 32, 24\n\t"
+        "rlwinm  26, 18, 26, 22, 29\n\t"
+        "rlwinm  27, 20, 26, 22, 29\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "lwzx    25, 10, 25\n\t"
+        "lwzx    26, 10, 26\n\t"
+        "lwzx    27, 10, 27\n\t"
+        "rldimi  14, 24, 40, 16\n\t"
+        "rlwimi  14, 25, 8, 16, 23\n\t"
+        "rldimi  15, 26, 40, 16\n\t"
+        "rlwimi  15, 27, 8, 16, 23\n\t"
+        "rlwinm  24, 20, 18, 22, 29\n\t"
+        "rlwinm  25, 19, 18, 22, 29\n\t"
+        "rlwinm  26, 21, 18, 22, 29\n\t"
+        "rlwinm  27, 18, 18, 22, 29\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "lwzx    25, 10, 25\n\t"
+        "lwzx    26, 10, 26\n\t"
+        "lwzx    27, 10, 27\n\t"
+        "rldimi  14, 24, 48, 8\n\t"
+        "rlwimi  14, 25, 16, 8, 15\n\t"
+        "rldimi  15, 26, 48, 8\n\t"
+        "rlwimi  15, 27, 16, 8, 15\n\t"
+        "rlwinm  24, 18, 10, 22, 29\n\t"
+        "rlwinm  25, 20, 10, 22, 29\n\t"
+        "rlwinm  26, 19, 10, 22, 29\n\t"
+        "rlwinm  27, 21, 10, 22, 29\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "lwzx    25, 10, 25\n\t"
+        "lwzx    26, 10, 26\n\t"
+        "lwzx    27, 10, 27\n\t"
+        "ld      18, 0(28)\n\t"
+        "ld      19, 8(28)\n\t"
+        "rldimi  14, 24, 56, 0\n\t"
+        "rlwimi  14, 25, 24, 0, 7\n\t"
+        "rldimi  15, 26, 56, 0\n\t"
+        "rlwimi  15, 27, 24, 0, 7\n\t"
+        "addi    28, 28, 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "ld      18, 0(%[in])\n\t"
+        "ld      19, 8(%[in])\n\t"
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "std     14, 0(%[out])\n\t"
+        "std     15, 8(%[out])\n\t"
+        "addic   23, 23, 1\n\t"
+        "addze   22, 22\n\t"
+        "addi    %[in], %[in], 16\n\t"
+        "addi    %[out], %[out], 16\n\t"
+        "addic.  %[len], %[len], -16\n\t"
+        "bne     L_AES_CTR_encrypt_loop_block_128_%=\n\t"
+        "std     22, 0(%[ctr])\n\t"
+        "std     23, 8(%[ctr])\n\t"
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+        : [in] "+r" (in), [out] "+r" (out), [len] "+r" (len), [ks] "+r" (ks),
+          [nr] "+r" (nr), [ctr] "+r" (ctr),
+          [L_AES_PPC64_te4_0] "+r" (L_AES_PPC64_te4_0_c)
+        :
+        : "memory", "cc", "0", "10", "11", "12", "14", "15", "16", "17", "18",
+            "19", "20", "21", "22", "23", "24", "25", "26", "27", "28"
+#else
+        :
+        : [in] "r" (in), [out] "r" (out), [len] "r" (len), [ks] "r" (ks),
+          [nr] "r" (nr), [ctr] "r" (ctr),
+          [L_AES_PPC64_te4_0] "r" (L_AES_PPC64_te4_0_c)
+        : "memory", "cc", "0", "9", "10", "11", "12", "14", "15", "16", "17",
+            "18", "19", "20", "21", "22", "23", "24", "25", "26", "27", "28"
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+    );
+}
+
+#endif /* WOLFSSL_AES_COUNTER */
+#ifdef HAVE_AESGCM
+void AES_GCM_encrypt(const unsigned char* in, unsigned char* out,
+    unsigned long len, const unsigned char* ks, int nr, unsigned char* ctr);
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+void AES_GCM_encrypt(const unsigned char* in_p, unsigned char* out_p,
+    unsigned long len_p, const unsigned char* ks_p, int nr_p,
+    unsigned char* ctr_p)
+#else
+void AES_GCM_encrypt(const unsigned char* in, unsigned char* out,
+    unsigned long len, const unsigned char* ks, int nr, unsigned char* ctr)
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+{
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+    register const unsigned char* in asm ("3") = (const unsigned char*)in_p;
+    register unsigned char* out asm ("4") = (unsigned char*)out_p;
+    register unsigned long len asm ("5") = (unsigned long)len_p;
+    register const unsigned char* ks asm ("6") = (const unsigned char*)ks_p;
+    register int nr asm ("7") = (int)nr_p;
+    register unsigned char* ctr asm ("8") = (unsigned char*)ctr_p;
+    register word32* L_AES_PPC64_te4_0_c asm ("9") =
+        (word32*)&L_AES_PPC64_te4_0;
+#else
+    register word32* L_AES_PPC64_te4_0_c = (word32*)&L_AES_PPC64_te4_0;
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+
+    __asm__ __volatile__ (
+#ifdef WOLFSSL_NO_VAR_ASSIGN_REG
+        "mr      9, %[L_AES_PPC64_te4_0]\n\t"
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+        "ld      26, 0(%[ctr])\n\t"
+        "ld      27, 8(%[ctr])\n\t"
+        "addi    10, 9, 0x400\n\t"
+        "addi    11, 9, 0x800\n\t"
+        "addi    12, 9, 0xc00\n\t"
+        "\n"
+    "L_AES_GCM_encrypt_loop_block_%=: \n\t"
+        "addi    28, %[ks], 0\n\t"
+        "addi    17, 27, 1\n\t"
+        "ld      18, 0(28)\n\t"
+        "ld      19, 8(28)\n\t"
+        "addi    28, 28, 16\n\t"
+        "rldimi  27, 17, 0, 32\n\t"
+        "rldicr  14, 26, 32, 63\n\t"
+        "rldicr  15, 27, 32, 63\n\t"
+        /* Round: 0 - XOR in key schedule */
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "addi    0, %[nr], -2\n\t"
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+        "ld      21, 0(9)\n\t"
+        "ld      21, 64(9)\n\t"
+        "ld      21, 128(9)\n\t"
+        "ld      21, 192(9)\n\t"
+        "ld      21, 256(9)\n\t"
+        "ld      21, 320(9)\n\t"
+        "ld      21, 384(9)\n\t"
+        "ld      21, 448(9)\n\t"
+        "ld      21, 512(9)\n\t"
+        "ld      21, 576(9)\n\t"
+        "ld      21, 640(9)\n\t"
+        "ld      21, 704(9)\n\t"
+        "ld      21, 768(9)\n\t"
+        "ld      21, 832(9)\n\t"
+        "ld      21, 896(9)\n\t"
+        "ld      21, 960(9)\n\t"
+        "ld      21, 1024(9)\n\t"
+        "ld      21, 1088(9)\n\t"
+        "ld      21, 1152(9)\n\t"
+        "ld      21, 1216(9)\n\t"
+        "ld      21, 1280(9)\n\t"
+        "ld      21, 1344(9)\n\t"
+        "ld      21, 1408(9)\n\t"
+        "ld      21, 1472(9)\n\t"
+        "ld      21, 1536(9)\n\t"
+        "ld      21, 1600(9)\n\t"
+        "ld      21, 1664(9)\n\t"
+        "ld      21, 1728(9)\n\t"
+        "ld      21, 1792(9)\n\t"
+        "ld      21, 1856(9)\n\t"
+        "ld      21, 1920(9)\n\t"
+        "ld      21, 1984(9)\n\t"
+        "ld      21, 2048(9)\n\t"
+        "ld      21, 2112(9)\n\t"
+        "ld      21, 2176(9)\n\t"
+        "ld      21, 2240(9)\n\t"
+        "ld      21, 2304(9)\n\t"
+        "ld      21, 2368(9)\n\t"
+        "ld      21, 2432(9)\n\t"
+        "ld      21, 2496(9)\n\t"
+        "ld      21, 2560(9)\n\t"
+        "ld      21, 2624(9)\n\t"
+        "ld      21, 2688(9)\n\t"
+        "ld      21, 2752(9)\n\t"
+        "ld      21, 2816(9)\n\t"
+        "ld      21, 2880(9)\n\t"
+        "ld      21, 2944(9)\n\t"
+        "ld      21, 3008(9)\n\t"
+        "ld      21, 3072(9)\n\t"
+        "ld      21, 3136(9)\n\t"
+        "ld      21, 3200(9)\n\t"
+        "ld      21, 3264(9)\n\t"
+        "ld      21, 3328(9)\n\t"
+        "ld      21, 3392(9)\n\t"
+        "ld      21, 3456(9)\n\t"
+        "ld      21, 3520(9)\n\t"
+        "ld      21, 3584(9)\n\t"
+        "ld      21, 3648(9)\n\t"
+        "ld      21, 3712(9)\n\t"
+        "ld      21, 3776(9)\n\t"
+        "ld      21, 3840(9)\n\t"
+        "ld      21, 3904(9)\n\t"
+        "ld      21, 3968(9)\n\t"
+        "ld      21, 4032(9)\n\t"
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+        "srwi    0, 0, 1\n\t"
+        "mtctr   0\n\t"
+        "\n"
+    "L_AES_GCM_encrypt_loop_nr_%=: \n\t"
+        "srdi    16, 14, 32\n\t"
+        "srdi    17, 15, 32\n\t"
+        "rlwinm  18, 14, 10, 22, 29\n\t"
+        "rlwinm  19, 15, 10, 22, 29\n\t"
+        "rlwinm  20, 16, 10, 22, 29\n\t"
+        "rlwinm  21, 17, 10, 22, 29\n\t"
+        "lwzx    18, 9, 18\n\t"
+        "lwzx    19, 9, 19\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "rlwinm  22, 16, 18, 22, 29\n\t"
+        "rlwinm  23, 15, 18, 22, 29\n\t"
+        "rlwinm  24, 17, 18, 22, 29\n\t"
+        "rlwinm  25, 14, 18, 22, 29\n\t"
+        "lwzx    22, 10, 22\n\t"
+        "lwzx    23, 10, 23\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "lwzx    25, 10, 25\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 15, 26, 22, 29\n\t"
+        "rlwinm  23, 17, 26, 22, 29\n\t"
+        "rlwinm  24, 14, 26, 22, 29\n\t"
+        "rlwinm  25, 16, 26, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 17, 2, 22, 29\n\t"
+        "rlwinm  23, 14, 2, 22, 29\n\t"
+        "rlwinm  24, 16, 2, 22, 29\n\t"
+        "rlwinm  25, 15, 2, 22, 29\n\t"
+        "lwzx    22, 12, 22\n\t"
+        "lwzx    23, 12, 23\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "ld      14, 0(28)\n\t"
+        "ld      15, 8(28)\n\t"
+        "rldimi  18, 20, 32, 0\n\t"
+        "rldimi  19, 21, 32, 0\n\t"
+        "addi    28, 28, 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     18, 18, 14\n\t"
+        "xor     19, 19, 15\n\t"
+        "srdi    20, 18, 32\n\t"
+        "srdi    21, 19, 32\n\t"
+        "rlwinm  14, 18, 10, 22, 29\n\t"
+        "rlwinm  15, 19, 10, 22, 29\n\t"
+        "rlwinm  16, 20, 10, 22, 29\n\t"
+        "rlwinm  17, 21, 10, 22, 29\n\t"
+        "lwzx    14, 9, 14\n\t"
+        "lwzx    15, 9, 15\n\t"
+        "lwzx    16, 9, 16\n\t"
+        "lwzx    17, 9, 17\n\t"
+        "rlwinm  22, 20, 18, 22, 29\n\t"
+        "rlwinm  23, 19, 18, 22, 29\n\t"
+        "rlwinm  24, 21, 18, 22, 29\n\t"
+        "rlwinm  25, 18, 18, 22, 29\n\t"
+        "lwzx    22, 10, 22\n\t"
+        "lwzx    23, 10, 23\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "lwzx    25, 10, 25\n\t"
+        "xor     14, 14, 22\n\t"
+        "xor     16, 16, 23\n\t"
+        "xor     15, 15, 24\n\t"
+        "xor     17, 17, 25\n\t"
+        "rlwinm  22, 19, 26, 22, 29\n\t"
+        "rlwinm  23, 21, 26, 22, 29\n\t"
+        "rlwinm  24, 18, 26, 22, 29\n\t"
+        "rlwinm  25, 20, 26, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "xor     14, 14, 22\n\t"
+        "xor     16, 16, 23\n\t"
+        "xor     15, 15, 24\n\t"
+        "xor     17, 17, 25\n\t"
+        "rlwinm  22, 21, 2, 22, 29\n\t"
+        "rlwinm  23, 18, 2, 22, 29\n\t"
+        "rlwinm  24, 20, 2, 22, 29\n\t"
+        "rlwinm  25, 19, 2, 22, 29\n\t"
+        "lwzx    22, 12, 22\n\t"
+        "lwzx    23, 12, 23\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "xor     14, 14, 22\n\t"
+        "xor     16, 16, 23\n\t"
+        "xor     15, 15, 24\n\t"
+        "xor     17, 17, 25\n\t"
+        "ld      18, 0(28)\n\t"
+        "ld      19, 8(28)\n\t"
+        "rldimi  14, 16, 32, 0\n\t"
+        "rldimi  15, 17, 32, 0\n\t"
+        "addi    28, 28, 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "bdnz    L_AES_GCM_encrypt_loop_nr_%=\n\t"
+        "srdi    16, 14, 32\n\t"
+        "srdi    17, 15, 32\n\t"
+        "rlwinm  18, 14, 10, 22, 29\n\t"
+        "rlwinm  19, 15, 10, 22, 29\n\t"
+        "rlwinm  20, 16, 10, 22, 29\n\t"
+        "rlwinm  21, 17, 10, 22, 29\n\t"
+        "lwzx    18, 9, 18\n\t"
+        "lwzx    19, 9, 19\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "rlwinm  22, 16, 18, 22, 29\n\t"
+        "rlwinm  23, 15, 18, 22, 29\n\t"
+        "rlwinm  24, 17, 18, 22, 29\n\t"
+        "rlwinm  25, 14, 18, 22, 29\n\t"
+        "lwzx    22, 10, 22\n\t"
+        "lwzx    23, 10, 23\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "lwzx    25, 10, 25\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 15, 26, 22, 29\n\t"
+        "rlwinm  23, 17, 26, 22, 29\n\t"
+        "rlwinm  24, 14, 26, 22, 29\n\t"
+        "rlwinm  25, 16, 26, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 17, 2, 22, 29\n\t"
+        "rlwinm  23, 14, 2, 22, 29\n\t"
+        "rlwinm  24, 16, 2, 22, 29\n\t"
+        "rlwinm  25, 15, 2, 22, 29\n\t"
+        "lwzx    22, 12, 22\n\t"
+        "lwzx    23, 12, 23\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "ld      14, 0(28)\n\t"
+        "ld      15, 8(28)\n\t"
+        "rldimi  18, 20, 32, 0\n\t"
+        "rldimi  19, 21, 32, 0\n\t"
+        "addi    28, 28, 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     18, 18, 14\n\t"
+        "xor     19, 19, 15\n\t"
+        "srdi    20, 18, 32\n\t"
+        "srdi    21, 19, 32\n\t"
+        "rlwinm  14, 18, 2, 22, 29\n\t"
+        "rlwinm  15, 19, 2, 22, 29\n\t"
+        "rlwinm  17, 20, 2, 22, 29\n\t"
+        "rlwinm  16, 21, 2, 22, 29\n\t"
+        "lwzx    14, 10, 14\n\t"
+        "lwzx    15, 10, 15\n\t"
+        "lwzx    17, 10, 17\n\t"
+        "lwzx    16, 10, 16\n\t"
+        "rlwinm  22, 19, 26, 22, 29\n\t"
+        "rlwinm  23, 21, 26, 22, 29\n\t"
+        "rldimi  15, 17, 32, 24\n\t"
+        "rldimi  14, 16, 32, 24\n\t"
+        "rlwinm  24, 18, 26, 22, 29\n\t"
+        "rlwinm  25, 20, 26, 22, 29\n\t"
+        "lwzx    22, 10, 22\n\t"
+        "lwzx    23, 10, 23\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "lwzx    25, 10, 25\n\t"
+        "rldimi  14, 22, 40, 16\n\t"
+        "rlwimi  14, 23, 8, 16, 23\n\t"
+        "rldimi  15, 24, 40, 16\n\t"
+        "rlwimi  15, 25, 8, 16, 23\n\t"
+        "rlwinm  22, 20, 18, 22, 29\n\t"
+        "rlwinm  23, 19, 18, 22, 29\n\t"
+        "rlwinm  24, 21, 18, 22, 29\n\t"
+        "rlwinm  25, 18, 18, 22, 29\n\t"
+        "lwzx    22, 10, 22\n\t"
+        "lwzx    23, 10, 23\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "lwzx    25, 10, 25\n\t"
+        "rldimi  14, 22, 48, 8\n\t"
+        "rlwimi  14, 23, 16, 8, 15\n\t"
+        "rldimi  15, 24, 48, 8\n\t"
+        "rlwimi  15, 25, 16, 8, 15\n\t"
+        "rlwinm  22, 18, 10, 22, 29\n\t"
+        "rlwinm  23, 20, 10, 22, 29\n\t"
+        "rlwinm  24, 19, 10, 22, 29\n\t"
+        "rlwinm  25, 21, 10, 22, 29\n\t"
+        "lwzx    22, 10, 22\n\t"
+        "lwzx    23, 10, 23\n\t"
+        "lwzx    24, 10, 24\n\t"
+        "lwzx    25, 10, 25\n\t"
+        "ld      18, 0(28)\n\t"
+        "ld      19, 8(28)\n\t"
+        "rldimi  14, 22, 56, 0\n\t"
+        "rlwimi  14, 23, 24, 0, 7\n\t"
+        "rldimi  15, 24, 56, 0\n\t"
+        "rlwimi  15, 25, 24, 0, 7\n\t"
+        "addi    28, 28, 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "ld      18, 0(%[in])\n\t"
+        "ld      19, 8(%[in])\n\t"
+        "addi    %[in], %[in], 16\n\t"
+        "addic.  %[len], %[len], -16\n\t"
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "std     14, 0(%[out])\n\t"
+        "std     15, 8(%[out])\n\t"
+        "addi    %[out], %[out], 16\n\t"
+        "bne     L_AES_GCM_encrypt_loop_block_%=\n\t"
+        "std     26, 0(%[ctr])\n\t"
+        "std     27, 8(%[ctr])\n\t"
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+        : [in] "+r" (in), [out] "+r" (out), [len] "+r" (len), [ks] "+r" (ks),
+          [nr] "+r" (nr), [ctr] "+r" (ctr),
+          [L_AES_PPC64_te4_0] "+r" (L_AES_PPC64_te4_0_c)
+        :
+        : "memory", "cc", "0", "10", "11", "12", "14", "15", "16", "17", "18",
+            "19", "20", "21", "22", "23", "24", "25", "26", "27", "28"
+#else
+        :
+        : [in] "r" (in), [out] "r" (out), [len] "r" (len), [ks] "r" (ks),
+          [nr] "r" (nr), [ctr] "r" (ctr),
+          [L_AES_PPC64_te4_0] "r" (L_AES_PPC64_te4_0_c)
+        : "memory", "cc", "0", "9", "10", "11", "12", "14", "15", "16", "17",
+            "18", "19", "20", "21", "22", "23", "24", "25", "26", "27", "28"
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+    );
+}
+
+#endif /* HAVE_AESGCM */
+#ifdef WOLFSSL_AES_XTS
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+void AES_XTS_encrypt(const byte* in_p, byte* out_p, word32 sz_p,
+    const byte* i_p, byte* key_p, byte* key2_p, byte* tmp_p, int nr_p)
+#else
+void AES_XTS_encrypt(const byte* in, byte* out, word32 sz, const byte* i,
+    byte* key, byte* key2, byte* tmp, int nr)
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+{
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+    register const byte* in asm ("3") = (const byte*)in_p;
+    register byte* out asm ("4") = (byte*)out_p;
+    register word32 sz asm ("5") = (word32)sz_p;
+    register const byte* i asm ("6") = (const byte*)i_p;
+    register byte* key asm ("7") = (byte*)key_p;
+    register byte* key2 asm ("8") = (byte*)key2_p;
+    register byte* tmp asm ("9") = (byte*)tmp_p;
+    register int nr asm ("10") = (int)nr_p;
+    register word32* L_AES_PPC64_te4_0_c asm ("11") =
+        (word32*)&L_AES_PPC64_te4_0;
+#else
+    register word32* L_AES_PPC64_te4_0_c = (word32*)&L_AES_PPC64_te4_0;
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+
+    __asm__ __volatile__ (
+        "subi    1, 1, 8\n\t"
+#ifdef WOLFSSL_NO_VAR_ASSIGN_REG
+        "mr      11, %[L_AES_PPC64_te4_0]\n\t"
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+        "addi    12, 11, 0x400\n\t"
+        "addi    14, 11, 0x800\n\t"
+        "addi    15, 11, 0xc00\n\t"
+        "ld      28, 0(%[i])\n\t"
+        "ld      29, 8(%[i])\n\t"
+        "ld      20, 0(%[key2])\n\t"
+        "ld      21, 8(%[key2])\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        "rldicr  28, 28, 32, 63\n\t"
+        "rldicr  29, 29, 32, 63\n\t"
+        /* Round: 0 - XOR in key schedule */
+        "xor     28, 28, 20\n\t"
+        "xor     29, 29, 21\n\t"
+        "addi    0, %[nr], -2\n\t"
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+        "ld      23, 0(11)\n\t"
+        "ld      23, 64(11)\n\t"
+        "ld      23, 128(11)\n\t"
+        "ld      23, 192(11)\n\t"
+        "ld      23, 256(11)\n\t"
+        "ld      23, 320(11)\n\t"
+        "ld      23, 384(11)\n\t"
+        "ld      23, 448(11)\n\t"
+        "ld      23, 512(11)\n\t"
+        "ld      23, 576(11)\n\t"
+        "ld      23, 640(11)\n\t"
+        "ld      23, 704(11)\n\t"
+        "ld      23, 768(11)\n\t"
+        "ld      23, 832(11)\n\t"
+        "ld      23, 896(11)\n\t"
+        "ld      23, 960(11)\n\t"
+        "ld      23, 1024(11)\n\t"
+        "ld      23, 1088(11)\n\t"
+        "ld      23, 1152(11)\n\t"
+        "ld      23, 1216(11)\n\t"
+        "ld      23, 1280(11)\n\t"
+        "ld      23, 1344(11)\n\t"
+        "ld      23, 1408(11)\n\t"
+        "ld      23, 1472(11)\n\t"
+        "ld      23, 1536(11)\n\t"
+        "ld      23, 1600(11)\n\t"
+        "ld      23, 1664(11)\n\t"
+        "ld      23, 1728(11)\n\t"
+        "ld      23, 1792(11)\n\t"
+        "ld      23, 1856(11)\n\t"
+        "ld      23, 1920(11)\n\t"
+        "ld      23, 1984(11)\n\t"
+        "ld      23, 2048(11)\n\t"
+        "ld      23, 2112(11)\n\t"
+        "ld      23, 2176(11)\n\t"
+        "ld      23, 2240(11)\n\t"
+        "ld      23, 2304(11)\n\t"
+        "ld      23, 2368(11)\n\t"
+        "ld      23, 2432(11)\n\t"
+        "ld      23, 2496(11)\n\t"
+        "ld      23, 2560(11)\n\t"
+        "ld      23, 2624(11)\n\t"
+        "ld      23, 2688(11)\n\t"
+        "ld      23, 2752(11)\n\t"
+        "ld      23, 2816(11)\n\t"
+        "ld      23, 2880(11)\n\t"
+        "ld      23, 2944(11)\n\t"
+        "ld      23, 3008(11)\n\t"
+        "ld      23, 3072(11)\n\t"
+        "ld      23, 3136(11)\n\t"
+        "ld      23, 3200(11)\n\t"
+        "ld      23, 3264(11)\n\t"
+        "ld      23, 3328(11)\n\t"
+        "ld      23, 3392(11)\n\t"
+        "ld      23, 3456(11)\n\t"
+        "ld      23, 3520(11)\n\t"
+        "ld      23, 3584(11)\n\t"
+        "ld      23, 3648(11)\n\t"
+        "ld      23, 3712(11)\n\t"
+        "ld      23, 3776(11)\n\t"
+        "ld      23, 3840(11)\n\t"
+        "ld      23, 3904(11)\n\t"
+        "ld      23, 3968(11)\n\t"
+        "ld      23, 4032(11)\n\t"
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+        "srwi    0, 0, 1\n\t"
+        "mtctr   0\n\t"
+        "\n"
+    "L_AES_XTS_encrypt_loop_nr_tweak_%=: \n\t"
+        "srdi    18, 28, 32\n\t"
+        "srdi    19, 29, 32\n\t"
+        "rlwinm  20, 28, 10, 22, 29\n\t"
+        "rlwinm  21, 29, 10, 22, 29\n\t"
+        "rlwinm  22, 18, 10, 22, 29\n\t"
+        "rlwinm  23, 19, 10, 22, 29\n\t"
+        "lwzx    20, 11, 20\n\t"
+        "lwzx    21, 11, 21\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "rlwinm  24, 18, 18, 22, 29\n\t"
+        "rlwinm  25, 29, 18, 22, 29\n\t"
+        "rlwinm  26, 19, 18, 22, 29\n\t"
+        "rlwinm  27, 28, 18, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "xor     20, 20, 24\n\t"
+        "xor     22, 22, 25\n\t"
+        "xor     21, 21, 26\n\t"
+        "xor     23, 23, 27\n\t"
+        "rlwinm  24, 29, 26, 22, 29\n\t"
+        "rlwinm  25, 19, 26, 22, 29\n\t"
+        "rlwinm  26, 28, 26, 22, 29\n\t"
+        "rlwinm  27, 18, 26, 22, 29\n\t"
+        "lwzx    24, 14, 24\n\t"
+        "lwzx    25, 14, 25\n\t"
+        "lwzx    26, 14, 26\n\t"
+        "lwzx    27, 14, 27\n\t"
+        "xor     20, 20, 24\n\t"
+        "xor     22, 22, 25\n\t"
+        "xor     21, 21, 26\n\t"
+        "xor     23, 23, 27\n\t"
+        "rlwinm  24, 19, 2, 22, 29\n\t"
+        "rlwinm  25, 28, 2, 22, 29\n\t"
+        "rlwinm  26, 18, 2, 22, 29\n\t"
+        "rlwinm  27, 29, 2, 22, 29\n\t"
+        "lwzx    24, 15, 24\n\t"
+        "lwzx    25, 15, 25\n\t"
+        "lwzx    26, 15, 26\n\t"
+        "lwzx    27, 15, 27\n\t"
+        "xor     20, 20, 24\n\t"
+        "xor     22, 22, 25\n\t"
+        "xor     21, 21, 26\n\t"
+        "xor     23, 23, 27\n\t"
+        "ld      28, 0(%[key2])\n\t"
+        "ld      29, 8(%[key2])\n\t"
+        "rldimi  20, 22, 32, 0\n\t"
+        "rldimi  21, 23, 32, 0\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     20, 20, 28\n\t"
+        "xor     21, 21, 29\n\t"
+        "srdi    22, 20, 32\n\t"
+        "srdi    23, 21, 32\n\t"
+        "rlwinm  28, 20, 10, 22, 29\n\t"
+        "rlwinm  29, 21, 10, 22, 29\n\t"
+        "rlwinm  18, 22, 10, 22, 29\n\t"
+        "rlwinm  19, 23, 10, 22, 29\n\t"
+        "lwzx    28, 11, 28\n\t"
+        "lwzx    29, 11, 29\n\t"
+        "lwzx    18, 11, 18\n\t"
+        "lwzx    19, 11, 19\n\t"
+        "rlwinm  24, 22, 18, 22, 29\n\t"
+        "rlwinm  25, 21, 18, 22, 29\n\t"
+        "rlwinm  26, 23, 18, 22, 29\n\t"
+        "rlwinm  27, 20, 18, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "xor     28, 28, 24\n\t"
+        "xor     18, 18, 25\n\t"
+        "xor     29, 29, 26\n\t"
+        "xor     19, 19, 27\n\t"
+        "rlwinm  24, 21, 26, 22, 29\n\t"
+        "rlwinm  25, 23, 26, 22, 29\n\t"
+        "rlwinm  26, 20, 26, 22, 29\n\t"
+        "rlwinm  27, 22, 26, 22, 29\n\t"
+        "lwzx    24, 14, 24\n\t"
+        "lwzx    25, 14, 25\n\t"
+        "lwzx    26, 14, 26\n\t"
+        "lwzx    27, 14, 27\n\t"
+        "xor     28, 28, 24\n\t"
+        "xor     18, 18, 25\n\t"
+        "xor     29, 29, 26\n\t"
+        "xor     19, 19, 27\n\t"
+        "rlwinm  24, 23, 2, 22, 29\n\t"
+        "rlwinm  25, 20, 2, 22, 29\n\t"
+        "rlwinm  26, 22, 2, 22, 29\n\t"
+        "rlwinm  27, 21, 2, 22, 29\n\t"
+        "lwzx    24, 15, 24\n\t"
+        "lwzx    25, 15, 25\n\t"
+        "lwzx    26, 15, 26\n\t"
+        "lwzx    27, 15, 27\n\t"
+        "xor     28, 28, 24\n\t"
+        "xor     18, 18, 25\n\t"
+        "xor     29, 29, 26\n\t"
+        "xor     19, 19, 27\n\t"
+        "ld      20, 0(%[key2])\n\t"
+        "ld      21, 8(%[key2])\n\t"
+        "rldimi  28, 18, 32, 0\n\t"
+        "rldimi  29, 19, 32, 0\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     28, 28, 20\n\t"
+        "xor     29, 29, 21\n\t"
+        "bdnz    L_AES_XTS_encrypt_loop_nr_tweak_%=\n\t"
+        "srdi    18, 28, 32\n\t"
+        "srdi    19, 29, 32\n\t"
+        "rlwinm  20, 28, 10, 22, 29\n\t"
+        "rlwinm  21, 29, 10, 22, 29\n\t"
+        "rlwinm  22, 18, 10, 22, 29\n\t"
+        "rlwinm  23, 19, 10, 22, 29\n\t"
+        "lwzx    20, 11, 20\n\t"
+        "lwzx    21, 11, 21\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "rlwinm  24, 18, 18, 22, 29\n\t"
+        "rlwinm  25, 29, 18, 22, 29\n\t"
+        "rlwinm  26, 19, 18, 22, 29\n\t"
+        "rlwinm  27, 28, 18, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "xor     20, 20, 24\n\t"
+        "xor     22, 22, 25\n\t"
+        "xor     21, 21, 26\n\t"
+        "xor     23, 23, 27\n\t"
+        "rlwinm  24, 29, 26, 22, 29\n\t"
+        "rlwinm  25, 19, 26, 22, 29\n\t"
+        "rlwinm  26, 28, 26, 22, 29\n\t"
+        "rlwinm  27, 18, 26, 22, 29\n\t"
+        "lwzx    24, 14, 24\n\t"
+        "lwzx    25, 14, 25\n\t"
+        "lwzx    26, 14, 26\n\t"
+        "lwzx    27, 14, 27\n\t"
+        "xor     20, 20, 24\n\t"
+        "xor     22, 22, 25\n\t"
+        "xor     21, 21, 26\n\t"
+        "xor     23, 23, 27\n\t"
+        "rlwinm  24, 19, 2, 22, 29\n\t"
+        "rlwinm  25, 28, 2, 22, 29\n\t"
+        "rlwinm  26, 18, 2, 22, 29\n\t"
+        "rlwinm  27, 29, 2, 22, 29\n\t"
+        "lwzx    24, 15, 24\n\t"
+        "lwzx    25, 15, 25\n\t"
+        "lwzx    26, 15, 26\n\t"
+        "lwzx    27, 15, 27\n\t"
+        "xor     20, 20, 24\n\t"
+        "xor     22, 22, 25\n\t"
+        "xor     21, 21, 26\n\t"
+        "xor     23, 23, 27\n\t"
+        "ld      28, 0(%[key2])\n\t"
+        "ld      29, 8(%[key2])\n\t"
+        "rldimi  20, 22, 32, 0\n\t"
+        "rldimi  21, 23, 32, 0\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     20, 20, 28\n\t"
+        "xor     21, 21, 29\n\t"
+        "srdi    22, 20, 32\n\t"
+        "srdi    23, 21, 32\n\t"
+        "rlwinm  28, 20, 2, 22, 29\n\t"
+        "rlwinm  29, 21, 2, 22, 29\n\t"
+        "rlwinm  19, 22, 2, 22, 29\n\t"
+        "rlwinm  18, 23, 2, 22, 29\n\t"
+        "lwzx    28, 12, 28\n\t"
+        "lwzx    29, 12, 29\n\t"
+        "lwzx    19, 12, 19\n\t"
+        "lwzx    18, 12, 18\n\t"
+        "rlwinm  24, 21, 26, 22, 29\n\t"
+        "rlwinm  25, 23, 26, 22, 29\n\t"
+        "rldimi  29, 19, 32, 24\n\t"
+        "rldimi  28, 18, 32, 24\n\t"
+        "rlwinm  26, 20, 26, 22, 29\n\t"
+        "rlwinm  27, 22, 26, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "rldimi  28, 24, 40, 16\n\t"
+        "rlwimi  28, 25, 8, 16, 23\n\t"
+        "rldimi  29, 26, 40, 16\n\t"
+        "rlwimi  29, 27, 8, 16, 23\n\t"
+        "rlwinm  24, 22, 18, 22, 29\n\t"
+        "rlwinm  25, 21, 18, 22, 29\n\t"
+        "rlwinm  26, 23, 18, 22, 29\n\t"
+        "rlwinm  27, 20, 18, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "rldimi  28, 24, 48, 8\n\t"
+        "rlwimi  28, 25, 16, 8, 15\n\t"
+        "rldimi  29, 26, 48, 8\n\t"
+        "rlwimi  29, 27, 16, 8, 15\n\t"
+        "rlwinm  24, 20, 10, 22, 29\n\t"
+        "rlwinm  25, 22, 10, 22, 29\n\t"
+        "rlwinm  26, 21, 10, 22, 29\n\t"
+        "rlwinm  27, 23, 10, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "ld      20, 0(%[key2])\n\t"
+        "ld      21, 8(%[key2])\n\t"
+        "rldimi  28, 24, 56, 0\n\t"
+        "rlwimi  28, 25, 24, 0, 7\n\t"
+        "rldimi  29, 26, 56, 0\n\t"
+        "rlwimi  29, 27, 24, 0, 7\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     28, 28, 20\n\t"
+        "xor     29, 29, 21\n\t"
+        "std     28, 0(%[tmp])\n\t"
+        "std     29, 8(%[tmp])\n\t"
+        "\n"
+    "L_AES_XTS_encrypt_loop_block_%=: \n\t"
+        "addi    %[key2], %[key], 0\n\t"
+        "ld      16, 0(%[in])\n\t"
+        "ld      17, 8(%[in])\n\t"
+        "ld      20, 0(%[key2])\n\t"
+        "ld      21, 8(%[key2])\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        "xor     16, 16, 28\n\t"
+        "xor     17, 17, 29\n\t"
+        "rldicr  16, 16, 32, 63\n\t"
+        "rldicr  17, 17, 32, 63\n\t"
+        /* Round: 0 - XOR in key schedule */
+        "xor     16, 16, 20\n\t"
+        "xor     17, 17, 21\n\t"
+        "addi    0, %[nr], -2\n\t"
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+        "ld      23, 0(11)\n\t"
+        "ld      23, 64(11)\n\t"
+        "ld      23, 128(11)\n\t"
+        "ld      23, 192(11)\n\t"
+        "ld      23, 256(11)\n\t"
+        "ld      23, 320(11)\n\t"
+        "ld      23, 384(11)\n\t"
+        "ld      23, 448(11)\n\t"
+        "ld      23, 512(11)\n\t"
+        "ld      23, 576(11)\n\t"
+        "ld      23, 640(11)\n\t"
+        "ld      23, 704(11)\n\t"
+        "ld      23, 768(11)\n\t"
+        "ld      23, 832(11)\n\t"
+        "ld      23, 896(11)\n\t"
+        "ld      23, 960(11)\n\t"
+        "ld      23, 1024(11)\n\t"
+        "ld      23, 1088(11)\n\t"
+        "ld      23, 1152(11)\n\t"
+        "ld      23, 1216(11)\n\t"
+        "ld      23, 1280(11)\n\t"
+        "ld      23, 1344(11)\n\t"
+        "ld      23, 1408(11)\n\t"
+        "ld      23, 1472(11)\n\t"
+        "ld      23, 1536(11)\n\t"
+        "ld      23, 1600(11)\n\t"
+        "ld      23, 1664(11)\n\t"
+        "ld      23, 1728(11)\n\t"
+        "ld      23, 1792(11)\n\t"
+        "ld      23, 1856(11)\n\t"
+        "ld      23, 1920(11)\n\t"
+        "ld      23, 1984(11)\n\t"
+        "ld      23, 2048(11)\n\t"
+        "ld      23, 2112(11)\n\t"
+        "ld      23, 2176(11)\n\t"
+        "ld      23, 2240(11)\n\t"
+        "ld      23, 2304(11)\n\t"
+        "ld      23, 2368(11)\n\t"
+        "ld      23, 2432(11)\n\t"
+        "ld      23, 2496(11)\n\t"
+        "ld      23, 2560(11)\n\t"
+        "ld      23, 2624(11)\n\t"
+        "ld      23, 2688(11)\n\t"
+        "ld      23, 2752(11)\n\t"
+        "ld      23, 2816(11)\n\t"
+        "ld      23, 2880(11)\n\t"
+        "ld      23, 2944(11)\n\t"
+        "ld      23, 3008(11)\n\t"
+        "ld      23, 3072(11)\n\t"
+        "ld      23, 3136(11)\n\t"
+        "ld      23, 3200(11)\n\t"
+        "ld      23, 3264(11)\n\t"
+        "ld      23, 3328(11)\n\t"
+        "ld      23, 3392(11)\n\t"
+        "ld      23, 3456(11)\n\t"
+        "ld      23, 3520(11)\n\t"
+        "ld      23, 3584(11)\n\t"
+        "ld      23, 3648(11)\n\t"
+        "ld      23, 3712(11)\n\t"
+        "ld      23, 3776(11)\n\t"
+        "ld      23, 3840(11)\n\t"
+        "ld      23, 3904(11)\n\t"
+        "ld      23, 3968(11)\n\t"
+        "ld      23, 4032(11)\n\t"
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+        "srwi    0, 0, 1\n\t"
+        "mtctr   0\n\t"
+        "\n"
+    "L_AES_XTS_encrypt_loop_nr_%=: \n\t"
+        "srdi    18, 16, 32\n\t"
+        "srdi    19, 17, 32\n\t"
+        "rlwinm  20, 16, 10, 22, 29\n\t"
+        "rlwinm  21, 17, 10, 22, 29\n\t"
+        "rlwinm  22, 18, 10, 22, 29\n\t"
+        "rlwinm  23, 19, 10, 22, 29\n\t"
+        "lwzx    20, 11, 20\n\t"
+        "lwzx    21, 11, 21\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "rlwinm  24, 18, 18, 22, 29\n\t"
+        "rlwinm  25, 17, 18, 22, 29\n\t"
+        "rlwinm  26, 19, 18, 22, 29\n\t"
+        "rlwinm  27, 16, 18, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "xor     20, 20, 24\n\t"
+        "xor     22, 22, 25\n\t"
+        "xor     21, 21, 26\n\t"
+        "xor     23, 23, 27\n\t"
+        "rlwinm  24, 17, 26, 22, 29\n\t"
+        "rlwinm  25, 19, 26, 22, 29\n\t"
+        "rlwinm  26, 16, 26, 22, 29\n\t"
+        "rlwinm  27, 18, 26, 22, 29\n\t"
+        "lwzx    24, 14, 24\n\t"
+        "lwzx    25, 14, 25\n\t"
+        "lwzx    26, 14, 26\n\t"
+        "lwzx    27, 14, 27\n\t"
+        "xor     20, 20, 24\n\t"
+        "xor     22, 22, 25\n\t"
+        "xor     21, 21, 26\n\t"
+        "xor     23, 23, 27\n\t"
+        "rlwinm  24, 19, 2, 22, 29\n\t"
+        "rlwinm  25, 16, 2, 22, 29\n\t"
+        "rlwinm  26, 18, 2, 22, 29\n\t"
+        "rlwinm  27, 17, 2, 22, 29\n\t"
+        "lwzx    24, 15, 24\n\t"
+        "lwzx    25, 15, 25\n\t"
+        "lwzx    26, 15, 26\n\t"
+        "lwzx    27, 15, 27\n\t"
+        "xor     20, 20, 24\n\t"
+        "xor     22, 22, 25\n\t"
+        "xor     21, 21, 26\n\t"
+        "xor     23, 23, 27\n\t"
+        "ld      16, 0(%[key2])\n\t"
+        "ld      17, 8(%[key2])\n\t"
+        "rldimi  20, 22, 32, 0\n\t"
+        "rldimi  21, 23, 32, 0\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     20, 20, 16\n\t"
+        "xor     21, 21, 17\n\t"
+        "srdi    22, 20, 32\n\t"
+        "srdi    23, 21, 32\n\t"
+        "rlwinm  16, 20, 10, 22, 29\n\t"
+        "rlwinm  17, 21, 10, 22, 29\n\t"
+        "rlwinm  18, 22, 10, 22, 29\n\t"
+        "rlwinm  19, 23, 10, 22, 29\n\t"
+        "lwzx    16, 11, 16\n\t"
+        "lwzx    17, 11, 17\n\t"
+        "lwzx    18, 11, 18\n\t"
+        "lwzx    19, 11, 19\n\t"
+        "rlwinm  24, 22, 18, 22, 29\n\t"
+        "rlwinm  25, 21, 18, 22, 29\n\t"
+        "rlwinm  26, 23, 18, 22, 29\n\t"
+        "rlwinm  27, 20, 18, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "xor     16, 16, 24\n\t"
+        "xor     18, 18, 25\n\t"
+        "xor     17, 17, 26\n\t"
+        "xor     19, 19, 27\n\t"
+        "rlwinm  24, 21, 26, 22, 29\n\t"
+        "rlwinm  25, 23, 26, 22, 29\n\t"
+        "rlwinm  26, 20, 26, 22, 29\n\t"
+        "rlwinm  27, 22, 26, 22, 29\n\t"
+        "lwzx    24, 14, 24\n\t"
+        "lwzx    25, 14, 25\n\t"
+        "lwzx    26, 14, 26\n\t"
+        "lwzx    27, 14, 27\n\t"
+        "xor     16, 16, 24\n\t"
+        "xor     18, 18, 25\n\t"
+        "xor     17, 17, 26\n\t"
+        "xor     19, 19, 27\n\t"
+        "rlwinm  24, 23, 2, 22, 29\n\t"
+        "rlwinm  25, 20, 2, 22, 29\n\t"
+        "rlwinm  26, 22, 2, 22, 29\n\t"
+        "rlwinm  27, 21, 2, 22, 29\n\t"
+        "lwzx    24, 15, 24\n\t"
+        "lwzx    25, 15, 25\n\t"
+        "lwzx    26, 15, 26\n\t"
+        "lwzx    27, 15, 27\n\t"
+        "xor     16, 16, 24\n\t"
+        "xor     18, 18, 25\n\t"
+        "xor     17, 17, 26\n\t"
+        "xor     19, 19, 27\n\t"
+        "ld      20, 0(%[key2])\n\t"
+        "ld      21, 8(%[key2])\n\t"
+        "rldimi  16, 18, 32, 0\n\t"
+        "rldimi  17, 19, 32, 0\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     16, 16, 20\n\t"
+        "xor     17, 17, 21\n\t"
+        "bdnz    L_AES_XTS_encrypt_loop_nr_%=\n\t"
+        "srdi    18, 16, 32\n\t"
+        "srdi    19, 17, 32\n\t"
+        "rlwinm  20, 16, 10, 22, 29\n\t"
+        "rlwinm  21, 17, 10, 22, 29\n\t"
+        "rlwinm  22, 18, 10, 22, 29\n\t"
+        "rlwinm  23, 19, 10, 22, 29\n\t"
+        "lwzx    20, 11, 20\n\t"
+        "lwzx    21, 11, 21\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "rlwinm  24, 18, 18, 22, 29\n\t"
+        "rlwinm  25, 17, 18, 22, 29\n\t"
+        "rlwinm  26, 19, 18, 22, 29\n\t"
+        "rlwinm  27, 16, 18, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "xor     20, 20, 24\n\t"
+        "xor     22, 22, 25\n\t"
+        "xor     21, 21, 26\n\t"
+        "xor     23, 23, 27\n\t"
+        "rlwinm  24, 17, 26, 22, 29\n\t"
+        "rlwinm  25, 19, 26, 22, 29\n\t"
+        "rlwinm  26, 16, 26, 22, 29\n\t"
+        "rlwinm  27, 18, 26, 22, 29\n\t"
+        "lwzx    24, 14, 24\n\t"
+        "lwzx    25, 14, 25\n\t"
+        "lwzx    26, 14, 26\n\t"
+        "lwzx    27, 14, 27\n\t"
+        "xor     20, 20, 24\n\t"
+        "xor     22, 22, 25\n\t"
+        "xor     21, 21, 26\n\t"
+        "xor     23, 23, 27\n\t"
+        "rlwinm  24, 19, 2, 22, 29\n\t"
+        "rlwinm  25, 16, 2, 22, 29\n\t"
+        "rlwinm  26, 18, 2, 22, 29\n\t"
+        "rlwinm  27, 17, 2, 22, 29\n\t"
+        "lwzx    24, 15, 24\n\t"
+        "lwzx    25, 15, 25\n\t"
+        "lwzx    26, 15, 26\n\t"
+        "lwzx    27, 15, 27\n\t"
+        "xor     20, 20, 24\n\t"
+        "xor     22, 22, 25\n\t"
+        "xor     21, 21, 26\n\t"
+        "xor     23, 23, 27\n\t"
+        "ld      16, 0(%[key2])\n\t"
+        "ld      17, 8(%[key2])\n\t"
+        "rldimi  20, 22, 32, 0\n\t"
+        "rldimi  21, 23, 32, 0\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     20, 20, 16\n\t"
+        "xor     21, 21, 17\n\t"
+        "srdi    22, 20, 32\n\t"
+        "srdi    23, 21, 32\n\t"
+        "rlwinm  16, 20, 2, 22, 29\n\t"
+        "rlwinm  17, 21, 2, 22, 29\n\t"
+        "rlwinm  19, 22, 2, 22, 29\n\t"
+        "rlwinm  18, 23, 2, 22, 29\n\t"
+        "lwzx    16, 12, 16\n\t"
+        "lwzx    17, 12, 17\n\t"
+        "lwzx    19, 12, 19\n\t"
+        "lwzx    18, 12, 18\n\t"
+        "rlwinm  24, 21, 26, 22, 29\n\t"
+        "rlwinm  25, 23, 26, 22, 29\n\t"
+        "rldimi  17, 19, 32, 24\n\t"
+        "rldimi  16, 18, 32, 24\n\t"
+        "rlwinm  26, 20, 26, 22, 29\n\t"
+        "rlwinm  27, 22, 26, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "rldimi  16, 24, 40, 16\n\t"
+        "rlwimi  16, 25, 8, 16, 23\n\t"
+        "rldimi  17, 26, 40, 16\n\t"
+        "rlwimi  17, 27, 8, 16, 23\n\t"
+        "rlwinm  24, 22, 18, 22, 29\n\t"
+        "rlwinm  25, 21, 18, 22, 29\n\t"
+        "rlwinm  26, 23, 18, 22, 29\n\t"
+        "rlwinm  27, 20, 18, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "rldimi  16, 24, 48, 8\n\t"
+        "rlwimi  16, 25, 16, 8, 15\n\t"
+        "rldimi  17, 26, 48, 8\n\t"
+        "rlwimi  17, 27, 16, 8, 15\n\t"
+        "rlwinm  24, 20, 10, 22, 29\n\t"
+        "rlwinm  25, 22, 10, 22, 29\n\t"
+        "rlwinm  26, 21, 10, 22, 29\n\t"
+        "rlwinm  27, 23, 10, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "ld      20, 0(%[key2])\n\t"
+        "ld      21, 8(%[key2])\n\t"
+        "rldimi  16, 24, 56, 0\n\t"
+        "rlwimi  16, 25, 24, 0, 7\n\t"
+        "rldimi  17, 26, 56, 0\n\t"
+        "rlwimi  17, 27, 24, 0, 7\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     16, 16, 20\n\t"
+        "xor     17, 17, 21\n\t"
+        "xor     16, 16, 28\n\t"
+        "xor     17, 17, 29\n\t"
+        "li      19, 8\n\t"
+        "ldbrx   28, 0, %[tmp]\n\t"
+        "ldbrx   29, %[tmp], 19\n\t"
+        "std     16, 0(%[out])\n\t"
+        "std     17, 8(%[out])\n\t"
+        "sradi   24, 29, 63\n\t"
+        "srdi    25, 28, 63\n\t"
+        "sldi    29, 29, 1\n\t"
+        "andi.   24, 24, 0x87\n\t"
+        "sldi    28, 28, 1\n\t"
+        "xor     29, 29, 25\n\t"
+        "xor     28, 28, 24\n\t"
+        "stdbrx  28, 0, %[tmp]\n\t"
+        "stdbrx  29, %[tmp], 19\n\t"
+        "ld      28, 0(%[tmp])\n\t"
+        "ld      29, 8(%[tmp])\n\t"
+        "addi    %[sz], %[sz], -16\n\t"
+        "addi    %[in], %[in], 16\n\t"
+        "addi    %[out], %[out], 16\n\t"
+        "cmpdi   %[sz], 16\n\t"
+        "bge     L_AES_XTS_encrypt_loop_block_%=\n\t"
+        "cmpdi   %[sz], 0\n\t"
+        "beq     L_AES_XTS_encrypt_done_data_%=\n\t"
+        "addi    %[key2], %[key], 0\n\t"
+        "addi    %[out], %[out], -16\n\t"
+        "ld      16, 0(%[out])\n\t"
+        "ld      17, 8(%[out])\n\t"
+        "addi    %[out], %[out], 16\n\t"
+        "std     16, 0(%[tmp])\n\t"
+        "std     17, 8(%[tmp])\n\t"
+        "addi    20, %[sz], 0\n\t"
+        "\n"
+    "L_AES_XTS_encrypt_start_byte_%=: \n\t"
+        "lbz     24, 0(%[tmp])\n\t"
+        "lbz     25, 0(%[in])\n\t"
+        "addi    %[in], %[in], 1\n\t"
+        "stb     24, 0(%[out])\n\t"
+        "addi    %[out], %[out], 1\n\t"
+        "stb     25, 0(%[tmp])\n\t"
+        "addi    %[tmp], %[tmp], 1\n\t"
+        "addic.  20, 20, -1\n\t"
+        "bne     L_AES_XTS_encrypt_start_byte_%=\n\t"
+        "subf    %[out], %[sz], %[out]\n\t"
+        "subf    %[tmp], %[sz], %[tmp]\n\t"
+        "addi    %[out], %[out], -16\n\t"
+        "ld      16, 0(%[tmp])\n\t"
+        "ld      17, 8(%[tmp])\n\t"
+        "ld      20, 0(%[key2])\n\t"
+        "ld      21, 8(%[key2])\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        "xor     16, 16, 28\n\t"
+        "xor     17, 17, 29\n\t"
+        "rldicr  16, 16, 32, 63\n\t"
+        "rldicr  17, 17, 32, 63\n\t"
+        /* Round: 0 - XOR in key schedule */
+        "xor     16, 16, 20\n\t"
+        "xor     17, 17, 21\n\t"
+        "addi    0, %[nr], -2\n\t"
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+        "ld      23, 0(11)\n\t"
+        "ld      23, 64(11)\n\t"
+        "ld      23, 128(11)\n\t"
+        "ld      23, 192(11)\n\t"
+        "ld      23, 256(11)\n\t"
+        "ld      23, 320(11)\n\t"
+        "ld      23, 384(11)\n\t"
+        "ld      23, 448(11)\n\t"
+        "ld      23, 512(11)\n\t"
+        "ld      23, 576(11)\n\t"
+        "ld      23, 640(11)\n\t"
+        "ld      23, 704(11)\n\t"
+        "ld      23, 768(11)\n\t"
+        "ld      23, 832(11)\n\t"
+        "ld      23, 896(11)\n\t"
+        "ld      23, 960(11)\n\t"
+        "ld      23, 1024(11)\n\t"
+        "ld      23, 1088(11)\n\t"
+        "ld      23, 1152(11)\n\t"
+        "ld      23, 1216(11)\n\t"
+        "ld      23, 1280(11)\n\t"
+        "ld      23, 1344(11)\n\t"
+        "ld      23, 1408(11)\n\t"
+        "ld      23, 1472(11)\n\t"
+        "ld      23, 1536(11)\n\t"
+        "ld      23, 1600(11)\n\t"
+        "ld      23, 1664(11)\n\t"
+        "ld      23, 1728(11)\n\t"
+        "ld      23, 1792(11)\n\t"
+        "ld      23, 1856(11)\n\t"
+        "ld      23, 1920(11)\n\t"
+        "ld      23, 1984(11)\n\t"
+        "ld      23, 2048(11)\n\t"
+        "ld      23, 2112(11)\n\t"
+        "ld      23, 2176(11)\n\t"
+        "ld      23, 2240(11)\n\t"
+        "ld      23, 2304(11)\n\t"
+        "ld      23, 2368(11)\n\t"
+        "ld      23, 2432(11)\n\t"
+        "ld      23, 2496(11)\n\t"
+        "ld      23, 2560(11)\n\t"
+        "ld      23, 2624(11)\n\t"
+        "ld      23, 2688(11)\n\t"
+        "ld      23, 2752(11)\n\t"
+        "ld      23, 2816(11)\n\t"
+        "ld      23, 2880(11)\n\t"
+        "ld      23, 2944(11)\n\t"
+        "ld      23, 3008(11)\n\t"
+        "ld      23, 3072(11)\n\t"
+        "ld      23, 3136(11)\n\t"
+        "ld      23, 3200(11)\n\t"
+        "ld      23, 3264(11)\n\t"
+        "ld      23, 3328(11)\n\t"
+        "ld      23, 3392(11)\n\t"
+        "ld      23, 3456(11)\n\t"
+        "ld      23, 3520(11)\n\t"
+        "ld      23, 3584(11)\n\t"
+        "ld      23, 3648(11)\n\t"
+        "ld      23, 3712(11)\n\t"
+        "ld      23, 3776(11)\n\t"
+        "ld      23, 3840(11)\n\t"
+        "ld      23, 3904(11)\n\t"
+        "ld      23, 3968(11)\n\t"
+        "ld      23, 4032(11)\n\t"
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+        "srwi    0, 0, 1\n\t"
+        "mtctr   0\n\t"
+        "\n"
+    "L_AES_XTS_encrypt_loop_nr_partial_%=: \n\t"
+        "srdi    18, 16, 32\n\t"
+        "srdi    19, 17, 32\n\t"
+        "rlwinm  20, 16, 10, 22, 29\n\t"
+        "rlwinm  21, 17, 10, 22, 29\n\t"
+        "rlwinm  22, 18, 10, 22, 29\n\t"
+        "rlwinm  23, 19, 10, 22, 29\n\t"
+        "lwzx    20, 11, 20\n\t"
+        "lwzx    21, 11, 21\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "rlwinm  24, 18, 18, 22, 29\n\t"
+        "rlwinm  25, 17, 18, 22, 29\n\t"
+        "rlwinm  26, 19, 18, 22, 29\n\t"
+        "rlwinm  27, 16, 18, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "xor     20, 20, 24\n\t"
+        "xor     22, 22, 25\n\t"
+        "xor     21, 21, 26\n\t"
+        "xor     23, 23, 27\n\t"
+        "rlwinm  24, 17, 26, 22, 29\n\t"
+        "rlwinm  25, 19, 26, 22, 29\n\t"
+        "rlwinm  26, 16, 26, 22, 29\n\t"
+        "rlwinm  27, 18, 26, 22, 29\n\t"
+        "lwzx    24, 14, 24\n\t"
+        "lwzx    25, 14, 25\n\t"
+        "lwzx    26, 14, 26\n\t"
+        "lwzx    27, 14, 27\n\t"
+        "xor     20, 20, 24\n\t"
+        "xor     22, 22, 25\n\t"
+        "xor     21, 21, 26\n\t"
+        "xor     23, 23, 27\n\t"
+        "rlwinm  24, 19, 2, 22, 29\n\t"
+        "rlwinm  25, 16, 2, 22, 29\n\t"
+        "rlwinm  26, 18, 2, 22, 29\n\t"
+        "rlwinm  27, 17, 2, 22, 29\n\t"
+        "lwzx    24, 15, 24\n\t"
+        "lwzx    25, 15, 25\n\t"
+        "lwzx    26, 15, 26\n\t"
+        "lwzx    27, 15, 27\n\t"
+        "xor     20, 20, 24\n\t"
+        "xor     22, 22, 25\n\t"
+        "xor     21, 21, 26\n\t"
+        "xor     23, 23, 27\n\t"
+        "ld      16, 0(%[key2])\n\t"
+        "ld      17, 8(%[key2])\n\t"
+        "rldimi  20, 22, 32, 0\n\t"
+        "rldimi  21, 23, 32, 0\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     20, 20, 16\n\t"
+        "xor     21, 21, 17\n\t"
+        "srdi    22, 20, 32\n\t"
+        "srdi    23, 21, 32\n\t"
+        "rlwinm  16, 20, 10, 22, 29\n\t"
+        "rlwinm  17, 21, 10, 22, 29\n\t"
+        "rlwinm  18, 22, 10, 22, 29\n\t"
+        "rlwinm  19, 23, 10, 22, 29\n\t"
+        "lwzx    16, 11, 16\n\t"
+        "lwzx    17, 11, 17\n\t"
+        "lwzx    18, 11, 18\n\t"
+        "lwzx    19, 11, 19\n\t"
+        "rlwinm  24, 22, 18, 22, 29\n\t"
+        "rlwinm  25, 21, 18, 22, 29\n\t"
+        "rlwinm  26, 23, 18, 22, 29\n\t"
+        "rlwinm  27, 20, 18, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "xor     16, 16, 24\n\t"
+        "xor     18, 18, 25\n\t"
+        "xor     17, 17, 26\n\t"
+        "xor     19, 19, 27\n\t"
+        "rlwinm  24, 21, 26, 22, 29\n\t"
+        "rlwinm  25, 23, 26, 22, 29\n\t"
+        "rlwinm  26, 20, 26, 22, 29\n\t"
+        "rlwinm  27, 22, 26, 22, 29\n\t"
+        "lwzx    24, 14, 24\n\t"
+        "lwzx    25, 14, 25\n\t"
+        "lwzx    26, 14, 26\n\t"
+        "lwzx    27, 14, 27\n\t"
+        "xor     16, 16, 24\n\t"
+        "xor     18, 18, 25\n\t"
+        "xor     17, 17, 26\n\t"
+        "xor     19, 19, 27\n\t"
+        "rlwinm  24, 23, 2, 22, 29\n\t"
+        "rlwinm  25, 20, 2, 22, 29\n\t"
+        "rlwinm  26, 22, 2, 22, 29\n\t"
+        "rlwinm  27, 21, 2, 22, 29\n\t"
+        "lwzx    24, 15, 24\n\t"
+        "lwzx    25, 15, 25\n\t"
+        "lwzx    26, 15, 26\n\t"
+        "lwzx    27, 15, 27\n\t"
+        "xor     16, 16, 24\n\t"
+        "xor     18, 18, 25\n\t"
+        "xor     17, 17, 26\n\t"
+        "xor     19, 19, 27\n\t"
+        "ld      20, 0(%[key2])\n\t"
+        "ld      21, 8(%[key2])\n\t"
+        "rldimi  16, 18, 32, 0\n\t"
+        "rldimi  17, 19, 32, 0\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     16, 16, 20\n\t"
+        "xor     17, 17, 21\n\t"
+        "bdnz    L_AES_XTS_encrypt_loop_nr_partial_%=\n\t"
+        "srdi    18, 16, 32\n\t"
+        "srdi    19, 17, 32\n\t"
+        "rlwinm  20, 16, 10, 22, 29\n\t"
+        "rlwinm  21, 17, 10, 22, 29\n\t"
+        "rlwinm  22, 18, 10, 22, 29\n\t"
+        "rlwinm  23, 19, 10, 22, 29\n\t"
+        "lwzx    20, 11, 20\n\t"
+        "lwzx    21, 11, 21\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "rlwinm  24, 18, 18, 22, 29\n\t"
+        "rlwinm  25, 17, 18, 22, 29\n\t"
+        "rlwinm  26, 19, 18, 22, 29\n\t"
+        "rlwinm  27, 16, 18, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "xor     20, 20, 24\n\t"
+        "xor     22, 22, 25\n\t"
+        "xor     21, 21, 26\n\t"
+        "xor     23, 23, 27\n\t"
+        "rlwinm  24, 17, 26, 22, 29\n\t"
+        "rlwinm  25, 19, 26, 22, 29\n\t"
+        "rlwinm  26, 16, 26, 22, 29\n\t"
+        "rlwinm  27, 18, 26, 22, 29\n\t"
+        "lwzx    24, 14, 24\n\t"
+        "lwzx    25, 14, 25\n\t"
+        "lwzx    26, 14, 26\n\t"
+        "lwzx    27, 14, 27\n\t"
+        "xor     20, 20, 24\n\t"
+        "xor     22, 22, 25\n\t"
+        "xor     21, 21, 26\n\t"
+        "xor     23, 23, 27\n\t"
+        "rlwinm  24, 19, 2, 22, 29\n\t"
+        "rlwinm  25, 16, 2, 22, 29\n\t"
+        "rlwinm  26, 18, 2, 22, 29\n\t"
+        "rlwinm  27, 17, 2, 22, 29\n\t"
+        "lwzx    24, 15, 24\n\t"
+        "lwzx    25, 15, 25\n\t"
+        "lwzx    26, 15, 26\n\t"
+        "lwzx    27, 15, 27\n\t"
+        "xor     20, 20, 24\n\t"
+        "xor     22, 22, 25\n\t"
+        "xor     21, 21, 26\n\t"
+        "xor     23, 23, 27\n\t"
+        "ld      16, 0(%[key2])\n\t"
+        "ld      17, 8(%[key2])\n\t"
+        "rldimi  20, 22, 32, 0\n\t"
+        "rldimi  21, 23, 32, 0\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     20, 20, 16\n\t"
+        "xor     21, 21, 17\n\t"
+        "srdi    22, 20, 32\n\t"
+        "srdi    23, 21, 32\n\t"
+        "rlwinm  16, 20, 2, 22, 29\n\t"
+        "rlwinm  17, 21, 2, 22, 29\n\t"
+        "rlwinm  19, 22, 2, 22, 29\n\t"
+        "rlwinm  18, 23, 2, 22, 29\n\t"
+        "lwzx    16, 12, 16\n\t"
+        "lwzx    17, 12, 17\n\t"
+        "lwzx    19, 12, 19\n\t"
+        "lwzx    18, 12, 18\n\t"
+        "rlwinm  24, 21, 26, 22, 29\n\t"
+        "rlwinm  25, 23, 26, 22, 29\n\t"
+        "rldimi  17, 19, 32, 24\n\t"
+        "rldimi  16, 18, 32, 24\n\t"
+        "rlwinm  26, 20, 26, 22, 29\n\t"
+        "rlwinm  27, 22, 26, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "rldimi  16, 24, 40, 16\n\t"
+        "rlwimi  16, 25, 8, 16, 23\n\t"
+        "rldimi  17, 26, 40, 16\n\t"
+        "rlwimi  17, 27, 8, 16, 23\n\t"
+        "rlwinm  24, 22, 18, 22, 29\n\t"
+        "rlwinm  25, 21, 18, 22, 29\n\t"
+        "rlwinm  26, 23, 18, 22, 29\n\t"
+        "rlwinm  27, 20, 18, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "rldimi  16, 24, 48, 8\n\t"
+        "rlwimi  16, 25, 16, 8, 15\n\t"
+        "rldimi  17, 26, 48, 8\n\t"
+        "rlwimi  17, 27, 16, 8, 15\n\t"
+        "rlwinm  24, 20, 10, 22, 29\n\t"
+        "rlwinm  25, 22, 10, 22, 29\n\t"
+        "rlwinm  26, 21, 10, 22, 29\n\t"
+        "rlwinm  27, 23, 10, 22, 29\n\t"
+        "lwzx    24, 12, 24\n\t"
+        "lwzx    25, 12, 25\n\t"
+        "lwzx    26, 12, 26\n\t"
+        "lwzx    27, 12, 27\n\t"
+        "ld      20, 0(%[key2])\n\t"
+        "ld      21, 8(%[key2])\n\t"
+        "rldimi  16, 24, 56, 0\n\t"
+        "rlwimi  16, 25, 24, 0, 7\n\t"
+        "rldimi  17, 26, 56, 0\n\t"
+        "rlwimi  17, 27, 24, 0, 7\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     16, 16, 20\n\t"
+        "xor     17, 17, 21\n\t"
+        "xor     16, 16, 28\n\t"
+        "xor     17, 17, 29\n\t"
+        "std     16, 0(%[out])\n\t"
+        "std     17, 8(%[out])\n\t"
+        "\n"
+    "L_AES_XTS_encrypt_done_data_%=: \n\t"
+        "addi    1, 1, 8\n\t"
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+        : [in] "+r" (in), [out] "+r" (out), [sz] "+r" (sz), [i] "+r" (i),
+          [key] "+r" (key), [key2] "+r" (key2), [tmp] "+r" (tmp),
+          [nr] "+r" (nr), [L_AES_PPC64_te4_0] "+r" (L_AES_PPC64_te4_0_c)
+        :
+        : "memory", "cc", "0", "12", "14", "15", "16", "17", "18", "19", "20",
+            "21", "22", "23", "24", "25", "26", "27", "28", "29"
+#else
+        :
+        : [in] "r" (in), [out] "r" (out), [sz] "r" (sz), [i] "r" (i),
+          [key] "r" (key), [key2] "r" (key2), [tmp] "r" (tmp), [nr] "r" (nr),
+          [L_AES_PPC64_te4_0] "r" (L_AES_PPC64_te4_0_c)
+        : "memory", "cc", "0", "11", "12", "14", "15", "16", "17", "18", "19",
+            "20", "21", "22", "23", "24", "25", "26", "27", "28", "29"
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+    );
+}
+
+#endif /* WOLFSSL_AES_XTS */
+#ifdef HAVE_AES_DECRYPT
+#if defined(WOLFSSL_AES_DIRECT) || defined(WOLFSSL_AES_COUNTER) || \
+    defined(HAVE_AES_CBC) || defined(HAVE_AES_ECB)
+static const byte L_AES_PPC64_td4[] = {
+    0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38,
+    0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb,
+    0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87,
+    0x34, 0x8e, 0x43, 0x44, 0xc4, 0xde, 0xe9, 0xcb,
+    0x54, 0x7b, 0x94, 0x32, 0xa6, 0xc2, 0x23, 0x3d,
+    0xee, 0x4c, 0x95, 0x0b, 0x42, 0xfa, 0xc3, 0x4e,
+    0x08, 0x2e, 0xa1, 0x66, 0x28, 0xd9, 0x24, 0xb2,
+    0x76, 0x5b, 0xa2, 0x49, 0x6d, 0x8b, 0xd1, 0x25,
+    0x72, 0xf8, 0xf6, 0x64, 0x86, 0x68, 0x98, 0x16,
+    0xd4, 0xa4, 0x5c, 0xcc, 0x5d, 0x65, 0xb6, 0x92,
+    0x6c, 0x70, 0x48, 0x50, 0xfd, 0xed, 0xb9, 0xda,
+    0x5e, 0x15, 0x46, 0x57, 0xa7, 0x8d, 0x9d, 0x84,
+    0x90, 0xd8, 0xab, 0x00, 0x8c, 0xbc, 0xd3, 0x0a,
+    0xf7, 0xe4, 0x58, 0x05, 0xb8, 0xb3, 0x45, 0x06,
+    0xd0, 0x2c, 0x1e, 0x8f, 0xca, 0x3f, 0x0f, 0x02,
+    0xc1, 0xaf, 0xbd, 0x03, 0x01, 0x13, 0x8a, 0x6b,
+    0x3a, 0x91, 0x11, 0x41, 0x4f, 0x67, 0xdc, 0xea,
+    0x97, 0xf2, 0xcf, 0xce, 0xf0, 0xb4, 0xe6, 0x73,
+    0x96, 0xac, 0x74, 0x22, 0xe7, 0xad, 0x35, 0x85,
+    0xe2, 0xf9, 0x37, 0xe8, 0x1c, 0x75, 0xdf, 0x6e,
+    0x47, 0xf1, 0x1a, 0x71, 0x1d, 0x29, 0xc5, 0x89,
+    0x6f, 0xb7, 0x62, 0x0e, 0xaa, 0x18, 0xbe, 0x1b,
+    0xfc, 0x56, 0x3e, 0x4b, 0xc6, 0xd2, 0x79, 0x20,
+    0x9a, 0xdb, 0xc0, 0xfe, 0x78, 0xcd, 0x5a, 0xf4,
+    0x1f, 0xdd, 0xa8, 0x33, 0x88, 0x07, 0xc7, 0x31,
+    0xb1, 0x12, 0x10, 0x59, 0x27, 0x80, 0xec, 0x5f,
+    0x60, 0x51, 0x7f, 0xa9, 0x19, 0xb5, 0x4a, 0x0d,
+    0x2d, 0xe5, 0x7a, 0x9f, 0x93, 0xc9, 0x9c, 0xef,
+    0xa0, 0xe0, 0x3b, 0x4d, 0xae, 0x2a, 0xf5, 0xb0,
+    0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61,
+    0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26,
+    0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d,
+};
+
+#if defined(WOLFSSL_AES_DIRECT) || defined(WOLFSSL_AES_COUNTER) || \
+        defined(HAVE_AES_ECB)
+void AES_ECB_decrypt(const unsigned char* in, unsigned char* out,
+    unsigned long len, const unsigned char* ks, int nr);
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+void AES_ECB_decrypt(const unsigned char* in_p, unsigned char* out_p,
+    unsigned long len_p, const unsigned char* ks_p, int nr_p)
+#else
+void AES_ECB_decrypt(const unsigned char* in, unsigned char* out,
+    unsigned long len, const unsigned char* ks, int nr)
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+{
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+    register const unsigned char* in asm ("3") = (const unsigned char*)in_p;
+    register unsigned char* out asm ("4") = (unsigned char*)out_p;
+    register unsigned long len asm ("5") = (unsigned long)len_p;
+    register const unsigned char* ks asm ("6") = (const unsigned char*)ks_p;
+    register int nr asm ("7") = (int)nr_p;
+    register word32* L_AES_PPC64_td_c asm ("8") = (word32*)&L_AES_PPC64_td;
+    register byte* L_AES_PPC64_td4_c asm ("9") = (byte*)&L_AES_PPC64_td4;
+#else
+    register word32* L_AES_PPC64_td_c = (word32*)&L_AES_PPC64_td;
+    register byte* L_AES_PPC64_td4_c = (byte*)&L_AES_PPC64_td4;
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+
+    __asm__ __volatile__ (
+#ifdef WOLFSSL_NO_VAR_ASSIGN_REG
+        "mr      8, %[L_AES_PPC64_td]\n\t"
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+#ifdef WOLFSSL_NO_VAR_ASSIGN_REG
+        "mr      9, %[L_AES_PPC64_td4]\n\t"
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+        "\n"
+    "L_AES_ECB_decrypt_loop_block_%=: \n\t"
+        "addi    23, %[ks], 0\n\t"
+        "ld      10, 0(%[in])\n\t"
+        "ld      11, 8(%[in])\n\t"
+        "ld      15, 0(23)\n\t"
+        "ld      16, 8(23)\n\t"
+        "addi    23, 23, 16\n\t"
+        "rldicr  10, 10, 32, 63\n\t"
+        "rldicr  11, 11, 32, 63\n\t"
+        /* Round: 0 - XOR in key schedule */
+        "xor     10, 10, 15\n\t"
+        "xor     11, 11, 16\n\t"
+        "addi    0, %[nr], -2\n\t"
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+        "ld      18, 0(8)\n\t"
+        "ld      18, 64(8)\n\t"
+        "ld      18, 128(8)\n\t"
+        "ld      18, 192(8)\n\t"
+        "ld      18, 256(8)\n\t"
+        "ld      18, 320(8)\n\t"
+        "ld      18, 384(8)\n\t"
+        "ld      18, 448(8)\n\t"
+        "ld      18, 512(8)\n\t"
+        "ld      18, 576(8)\n\t"
+        "ld      18, 640(8)\n\t"
+        "ld      18, 704(8)\n\t"
+        "ld      18, 768(8)\n\t"
+        "ld      18, 832(8)\n\t"
+        "ld      18, 896(8)\n\t"
+        "ld      18, 960(8)\n\t"
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+        "srwi    0, 0, 1\n\t"
+        "mtctr   0\n\t"
+        "\n"
+    "L_AES_ECB_decrypt_loop_nr_%=: \n\t"
+        "srdi    12, 10, 32\n\t"
+        "srdi    14, 11, 32\n\t"
+        "rlwinm  15, 14, 18, 22, 29\n\t"
+        "rlwinm  17, 10, 18, 22, 29\n\t"
+        "rlwinm  16, 12, 18, 22, 29\n\t"
+        "rlwinm  18, 11, 18, 22, 29\n\t"
+        "lwzx    15, 8, 15\n\t"
+        "lwzx    17, 8, 17\n\t"
+        "lwzx    16, 8, 16\n\t"
+        "lwzx    18, 8, 18\n\t"
+        "rlwinm  19, 10, 10, 22, 29\n\t"
+        "rlwinm  20, 12, 10, 22, 29\n\t"
+        "rlwinm  21, 11, 10, 22, 29\n\t"
+        "rlwinm  22, 14, 10, 22, 29\n\t"
+        "lwzx    19, 8, 19\n\t"
+        "lwzx    20, 8, 20\n\t"
+        "lwzx    21, 8, 21\n\t"
+        "lwzx    22, 8, 22\n\t"
+        "rlwimi  19, 19, 8, 0, 31\n\t"
+        "rlwimi  20, 20, 8, 0, 31\n\t"
+        "rlwimi  21, 21, 8, 0, 31\n\t"
+        "rlwimi  22, 22, 8, 0, 31\n\t"
+        "xor     15, 15, 19\n\t"
+        "xor     17, 17, 20\n\t"
+        "xor     16, 16, 21\n\t"
+        "xor     18, 18, 22\n\t"
+        "rlwinm  19, 11, 26, 22, 29\n\t"
+        "rlwinm  20, 14, 26, 22, 29\n\t"
+        "rlwinm  21, 10, 26, 22, 29\n\t"
+        "rlwinm  22, 12, 26, 22, 29\n\t"
+        "lwzx    19, 8, 19\n\t"
+        "lwzx    20, 8, 20\n\t"
+        "lwzx    21, 8, 21\n\t"
+        "lwzx    22, 8, 22\n\t"
+        "rlwimi  19, 19, 24, 0, 31\n\t"
+        "rlwimi  20, 20, 24, 0, 31\n\t"
+        "rlwimi  21, 21, 24, 0, 31\n\t"
+        "rlwimi  22, 22, 24, 0, 31\n\t"
+        "xor     15, 15, 19\n\t"
+        "xor     17, 17, 20\n\t"
+        "xor     16, 16, 21\n\t"
+        "xor     18, 18, 22\n\t"
+        "rlwinm  19, 12, 2, 22, 29\n\t"
+        "rlwinm  20, 11, 2, 22, 29\n\t"
+        "rlwinm  21, 14, 2, 22, 29\n\t"
+        "rlwinm  22, 10, 2, 22, 29\n\t"
+        "lwzx    19, 8, 19\n\t"
+        "lwzx    20, 8, 20\n\t"
+        "lwzx    21, 8, 21\n\t"
+        "lwzx    22, 8, 22\n\t"
+        "rlwimi  19, 19, 16, 0, 31\n\t"
+        "rlwimi  20, 20, 16, 0, 31\n\t"
+        "rlwimi  21, 21, 16, 0, 31\n\t"
+        "rlwimi  22, 22, 16, 0, 31\n\t"
+        "ld      10, 0(23)\n\t"
+        "ld      11, 8(23)\n\t"
+        "xor     15, 15, 19\n\t"
+        "xor     17, 17, 20\n\t"
+        "xor     16, 16, 21\n\t"
+        "xor     18, 18, 22\n\t"
+        "addi    23, 23, 16\n\t"
+        "rldimi  15, 17, 32, 0\n\t"
+        "rldimi  16, 18, 32, 0\n\t"
+        /*   XOR in Key Schedule */
+        "xor     15, 15, 10\n\t"
+        "xor     16, 16, 11\n\t"
+        "srdi    17, 15, 32\n\t"
+        "srdi    18, 16, 32\n\t"
+        "rlwinm  10, 18, 18, 22, 29\n\t"
+        "rlwinm  12, 15, 18, 22, 29\n\t"
+        "rlwinm  11, 17, 18, 22, 29\n\t"
+        "rlwinm  14, 16, 18, 22, 29\n\t"
+        "lwzx    10, 8, 10\n\t"
+        "lwzx    12, 8, 12\n\t"
+        "lwzx    11, 8, 11\n\t"
+        "lwzx    14, 8, 14\n\t"
+        "rlwinm  19, 15, 10, 22, 29\n\t"
+        "rlwinm  20, 17, 10, 22, 29\n\t"
+        "rlwinm  21, 16, 10, 22, 29\n\t"
+        "rlwinm  22, 18, 10, 22, 29\n\t"
+        "lwzx    19, 8, 19\n\t"
+        "lwzx    20, 8, 20\n\t"
+        "lwzx    21, 8, 21\n\t"
+        "lwzx    22, 8, 22\n\t"
+        "rlwimi  19, 19, 8, 0, 31\n\t"
+        "rlwimi  20, 20, 8, 0, 31\n\t"
+        "rlwimi  21, 21, 8, 0, 31\n\t"
+        "rlwimi  22, 22, 8, 0, 31\n\t"
+        "xor     10, 10, 19\n\t"
+        "xor     12, 12, 20\n\t"
+        "xor     11, 11, 21\n\t"
+        "xor     14, 14, 22\n\t"
+        "rlwinm  19, 16, 26, 22, 29\n\t"
+        "rlwinm  20, 18, 26, 22, 29\n\t"
+        "rlwinm  21, 15, 26, 22, 29\n\t"
+        "rlwinm  22, 17, 26, 22, 29\n\t"
+        "lwzx    19, 8, 19\n\t"
+        "lwzx    20, 8, 20\n\t"
+        "lwzx    21, 8, 21\n\t"
+        "lwzx    22, 8, 22\n\t"
+        "rlwimi  19, 19, 24, 0, 31\n\t"
+        "rlwimi  20, 20, 24, 0, 31\n\t"
+        "rlwimi  21, 21, 24, 0, 31\n\t"
+        "rlwimi  22, 22, 24, 0, 31\n\t"
+        "xor     10, 10, 19\n\t"
+        "xor     12, 12, 20\n\t"
+        "xor     11, 11, 21\n\t"
+        "xor     14, 14, 22\n\t"
+        "rlwinm  19, 17, 2, 22, 29\n\t"
+        "rlwinm  20, 16, 2, 22, 29\n\t"
+        "rlwinm  21, 18, 2, 22, 29\n\t"
+        "rlwinm  22, 15, 2, 22, 29\n\t"
+        "lwzx    19, 8, 19\n\t"
+        "lwzx    20, 8, 20\n\t"
+        "lwzx    21, 8, 21\n\t"
+        "lwzx    22, 8, 22\n\t"
+        "rlwimi  19, 19, 16, 0, 31\n\t"
+        "rlwimi  20, 20, 16, 0, 31\n\t"
+        "rlwimi  21, 21, 16, 0, 31\n\t"
+        "rlwimi  22, 22, 16, 0, 31\n\t"
+        "ld      15, 0(23)\n\t"
+        "ld      16, 8(23)\n\t"
+        "xor     10, 10, 19\n\t"
+        "xor     12, 12, 20\n\t"
+        "xor     11, 11, 21\n\t"
+        "xor     14, 14, 22\n\t"
+        "addi    23, 23, 16\n\t"
+        "rldimi  10, 12, 32, 0\n\t"
+        "rldimi  11, 14, 32, 0\n\t"
+        /*   XOR in Key Schedule */
+        "xor     10, 10, 15\n\t"
+        "xor     11, 11, 16\n\t"
+        "bdnz    L_AES_ECB_decrypt_loop_nr_%=\n\t"
+        "srdi    12, 10, 32\n\t"
+        "srdi    14, 11, 32\n\t"
+        "rlwinm  15, 14, 18, 22, 29\n\t"
+        "rlwinm  17, 10, 18, 22, 29\n\t"
+        "rlwinm  16, 12, 18, 22, 29\n\t"
+        "rlwinm  18, 11, 18, 22, 29\n\t"
+        "lwzx    15, 8, 15\n\t"
+        "lwzx    17, 8, 17\n\t"
+        "lwzx    16, 8, 16\n\t"
+        "lwzx    18, 8, 18\n\t"
+        "rlwinm  19, 10, 10, 22, 29\n\t"
+        "rlwinm  20, 12, 10, 22, 29\n\t"
+        "rlwinm  21, 11, 10, 22, 29\n\t"
+        "rlwinm  22, 14, 10, 22, 29\n\t"
+        "lwzx    19, 8, 19\n\t"
+        "lwzx    20, 8, 20\n\t"
+        "lwzx    21, 8, 21\n\t"
+        "lwzx    22, 8, 22\n\t"
+        "rlwimi  19, 19, 8, 0, 31\n\t"
+        "rlwimi  20, 20, 8, 0, 31\n\t"
+        "rlwimi  21, 21, 8, 0, 31\n\t"
+        "rlwimi  22, 22, 8, 0, 31\n\t"
+        "xor     15, 15, 19\n\t"
+        "xor     17, 17, 20\n\t"
+        "xor     16, 16, 21\n\t"
+        "xor     18, 18, 22\n\t"
+        "rlwinm  19, 11, 26, 22, 29\n\t"
+        "rlwinm  20, 14, 26, 22, 29\n\t"
+        "rlwinm  21, 10, 26, 22, 29\n\t"
+        "rlwinm  22, 12, 26, 22, 29\n\t"
+        "lwzx    19, 8, 19\n\t"
+        "lwzx    20, 8, 20\n\t"
+        "lwzx    21, 8, 21\n\t"
+        "lwzx    22, 8, 22\n\t"
+        "rlwimi  19, 19, 24, 0, 31\n\t"
+        "rlwimi  20, 20, 24, 0, 31\n\t"
+        "rlwimi  21, 21, 24, 0, 31\n\t"
+        "rlwimi  22, 22, 24, 0, 31\n\t"
+        "xor     15, 15, 19\n\t"
+        "xor     17, 17, 20\n\t"
+        "xor     16, 16, 21\n\t"
+        "xor     18, 18, 22\n\t"
+        "rlwinm  19, 12, 2, 22, 29\n\t"
+        "rlwinm  20, 11, 2, 22, 29\n\t"
+        "rlwinm  21, 14, 2, 22, 29\n\t"
+        "rlwinm  22, 10, 2, 22, 29\n\t"
+        "lwzx    19, 8, 19\n\t"
+        "lwzx    20, 8, 20\n\t"
+        "lwzx    21, 8, 21\n\t"
+        "lwzx    22, 8, 22\n\t"
+        "rlwimi  19, 19, 16, 0, 31\n\t"
+        "rlwimi  20, 20, 16, 0, 31\n\t"
+        "rlwimi  21, 21, 16, 0, 31\n\t"
+        "rlwimi  22, 22, 16, 0, 31\n\t"
+        "ld      10, 0(23)\n\t"
+        "ld      11, 8(23)\n\t"
+        "xor     15, 15, 19\n\t"
+        "xor     17, 17, 20\n\t"
+        "xor     16, 16, 21\n\t"
+        "xor     18, 18, 22\n\t"
+        "addi    23, 23, 16\n\t"
+        "rldimi  15, 17, 32, 0\n\t"
+        "rldimi  16, 18, 32, 0\n\t"
+        /*   XOR in Key Schedule */
+        "xor     15, 15, 10\n\t"
+        "xor     16, 16, 11\n\t"
+        "ld      18, 0(9)\n\t"
+        "ld      18, 64(9)\n\t"
+        "ld      18, 128(9)\n\t"
+        "ld      18, 192(9)\n\t"
+        "ld      18, 256(9)\n\t"
+        "ld      18, 320(9)\n\t"
+        "ld      18, 384(9)\n\t"
+        "ld      18, 448(9)\n\t"
+        "ld      18, 512(9)\n\t"
+        "ld      18, 576(9)\n\t"
+        "ld      18, 640(9)\n\t"
+        "ld      18, 704(9)\n\t"
+        "ld      18, 768(9)\n\t"
+        "ld      18, 832(9)\n\t"
+        "ld      18, 896(9)\n\t"
+        "ld      18, 960(9)\n\t"
+        "rldicl  12, 15, 32, 56\n\t"
+        "andi.   10, 16, 255\n\t"
+        "rldicl  14, 16, 32, 56\n\t"
+        "andi.   11, 15, 255\n\t"
+        "lbzx    12, 9, 12\n\t"
+        "lbzx    10, 9, 10\n\t"
+        "lbzx    14, 9, 14\n\t"
+        "lbzx    11, 9, 11\n\t"
+        "rldimi  10, 12, 32, 24\n\t"
+        "rldimi  11, 14, 32, 24\n\t"
+        "rldicl  19, 16, 56, 56\n\t"
+        "rldicl  20, 16, 24, 56\n\t"
+        "rldicl  21, 15, 56, 56\n\t"
+        "rldicl  22, 15, 24, 56\n\t"
+        "lbzx    19, 9, 19\n\t"
+        "lbzx    20, 9, 20\n\t"
+        "lbzx    21, 9, 21\n\t"
+        "lbzx    22, 9, 22\n\t"
+        "rldimi  10, 19, 40, 16\n\t"
+        "rlwimi  10, 20, 8, 16, 23\n\t"
+        "rldimi  11, 21, 40, 16\n\t"
+        "rlwimi  11, 22, 8, 16, 23\n\t"
+        "rldicl  19, 16, 16, 56\n\t"
+        "rldicl  20, 15, 48, 56\n\t"
+        "rldicl  21, 15, 16, 56\n\t"
+        "rldicl  22, 16, 48, 56\n\t"
+        "lbzx    19, 9, 19\n\t"
+        "lbzx    20, 9, 20\n\t"
+        "lbzx    21, 9, 21\n\t"
+        "lbzx    22, 9, 22\n\t"
+        "rldimi  10, 19, 48, 8\n\t"
+        "rlwimi  10, 20, 16, 8, 15\n\t"
+        "rldimi  11, 21, 48, 8\n\t"
+        "rlwimi  11, 22, 16, 8, 15\n\t"
+        "rldicl  19, 15, 40, 56\n\t"
+        "rldicl  20, 15, 8, 56\n\t"
+        "rldicl  21, 16, 40, 56\n\t"
+        "rldicl  22, 16, 8, 56\n\t"
+        "lbzx    19, 9, 19\n\t"
+        "lbzx    20, 9, 20\n\t"
+        "lbzx    21, 9, 21\n\t"
+        "lbzx    22, 9, 22\n\t"
+        "ld      15, 0(23)\n\t"
+        "ld      16, 8(23)\n\t"
+        "rldimi  10, 19, 56, 0\n\t"
+        "rlwimi  10, 20, 24, 0, 7\n\t"
+        "rldimi  11, 21, 56, 0\n\t"
+        "rlwimi  11, 22, 24, 0, 7\n\t"
+        "addi    23, 23, 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     10, 10, 15\n\t"
+        "xor     11, 11, 16\n\t"
+        "std     10, 0(%[out])\n\t"
+        "std     11, 8(%[out])\n\t"
+        "addic.  %[len], %[len], -16\n\t"
+        "addi    %[in], %[in], 16\n\t"
+        "addi    %[out], %[out], 16\n\t"
+        "bne     L_AES_ECB_decrypt_loop_block_%=\n\t"
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+        : [in] "+r" (in), [out] "+r" (out), [len] "+r" (len), [ks] "+r" (ks),
+          [nr] "+r" (nr), [L_AES_PPC64_td] "+r" (L_AES_PPC64_td_c),
+          [L_AES_PPC64_td4] "+r" (L_AES_PPC64_td4_c)
+        :
+        : "memory", "cc", "0", "10", "11", "12", "14", "15", "16", "17", "18",
+            "19", "20", "21", "22", "23"
+#else
+        :
+        : [in] "r" (in), [out] "r" (out), [len] "r" (len), [ks] "r" (ks),
+          [nr] "r" (nr), [L_AES_PPC64_td] "r" (L_AES_PPC64_td_c),
+          [L_AES_PPC64_td4] "r" (L_AES_PPC64_td4_c)
+        : "memory", "cc", "0", "8", "9", "10", "11", "12", "14", "15", "16",
+            "17", "18", "19", "20", "21", "22", "23"
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+    );
+}
+
+#endif /* WOLFSSL_AES_DIRECT || WOLFSSL_AES_COUNTER || defined(HAVE_AES_ECB) */
+#ifdef HAVE_AES_CBC
+void AES_CBC_decrypt(const unsigned char* in, unsigned char* out,
+    unsigned long len, const unsigned char* ks, int nr, unsigned char* iv);
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+void AES_CBC_decrypt(const unsigned char* in_p, unsigned char* out_p,
+    unsigned long len_p, const unsigned char* ks_p, int nr_p,
+    unsigned char* iv_p)
+#else
+void AES_CBC_decrypt(const unsigned char* in, unsigned char* out,
+    unsigned long len, const unsigned char* ks, int nr, unsigned char* iv)
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+{
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+    register const unsigned char* in asm ("3") = (const unsigned char*)in_p;
+    register unsigned char* out asm ("4") = (unsigned char*)out_p;
+    register unsigned long len asm ("5") = (unsigned long)len_p;
+    register const unsigned char* ks asm ("6") = (const unsigned char*)ks_p;
+    register int nr asm ("7") = (int)nr_p;
+    register unsigned char* iv asm ("8") = (unsigned char*)iv_p;
+    register word32* L_AES_PPC64_td_c asm ("9") = (word32*)&L_AES_PPC64_td;
+    register byte* L_AES_PPC64_td4_c asm ("10") = (byte*)&L_AES_PPC64_td4;
+#else
+    register word32* L_AES_PPC64_td_c = (word32*)&L_AES_PPC64_td;
+    register byte* L_AES_PPC64_td4_c = (byte*)&L_AES_PPC64_td4;
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+
+    __asm__ __volatile__ (
+#ifdef WOLFSSL_NO_VAR_ASSIGN_REG
+        "mr      10, %[L_AES_PPC64_td4]\n\t"
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+#ifdef WOLFSSL_NO_VAR_ASSIGN_REG
+        "mr      9, %[L_AES_PPC64_td]\n\t"
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+        "ld      24, 0(%[iv])\n\t"
+        "ld      25, 8(%[iv])\n\t"
+        "\n"
+    "L_AES_CBC_decrypt_loop_block_%=: \n\t"
+        "addi    28, %[ks], 0\n\t"
+        "ld      26, 0(%[in])\n\t"
+        "ld      27, 8(%[in])\n\t"
+        "ld      16, 0(28)\n\t"
+        "ld      17, 8(28)\n\t"
+        "addi    28, 28, 16\n\t"
+        "rldicr  11, 26, 32, 63\n\t"
+        "rldicr  12, 27, 32, 63\n\t"
+        /* Round: 0 - XOR in key schedule */
+        "xor     11, 11, 16\n\t"
+        "xor     12, 12, 17\n\t"
+        "addi    0, %[nr], -2\n\t"
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+        "ld      19, 0(9)\n\t"
+        "ld      19, 64(9)\n\t"
+        "ld      19, 128(9)\n\t"
+        "ld      19, 192(9)\n\t"
+        "ld      19, 256(9)\n\t"
+        "ld      19, 320(9)\n\t"
+        "ld      19, 384(9)\n\t"
+        "ld      19, 448(9)\n\t"
+        "ld      19, 512(9)\n\t"
+        "ld      19, 576(9)\n\t"
+        "ld      19, 640(9)\n\t"
+        "ld      19, 704(9)\n\t"
+        "ld      19, 768(9)\n\t"
+        "ld      19, 832(9)\n\t"
+        "ld      19, 896(9)\n\t"
+        "ld      19, 960(9)\n\t"
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+        "srwi    0, 0, 1\n\t"
+        "mtctr   0\n\t"
+        "\n"
+    "L_AES_CBC_decrypt_loop_nr_even_%=: \n\t"
+        "srdi    14, 11, 32\n\t"
+        "srdi    15, 12, 32\n\t"
+        "rlwinm  16, 15, 18, 22, 29\n\t"
+        "rlwinm  18, 11, 18, 22, 29\n\t"
+        "rlwinm  17, 14, 18, 22, 29\n\t"
+        "rlwinm  19, 12, 18, 22, 29\n\t"
+        "lwzx    16, 9, 16\n\t"
+        "lwzx    18, 9, 18\n\t"
+        "lwzx    17, 9, 17\n\t"
+        "lwzx    19, 9, 19\n\t"
+        "rlwinm  20, 11, 10, 22, 29\n\t"
+        "rlwinm  21, 14, 10, 22, 29\n\t"
+        "rlwinm  22, 12, 10, 22, 29\n\t"
+        "rlwinm  23, 15, 10, 22, 29\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "rlwimi  20, 20, 8, 0, 31\n\t"
+        "rlwimi  21, 21, 8, 0, 31\n\t"
+        "rlwimi  22, 22, 8, 0, 31\n\t"
+        "rlwimi  23, 23, 8, 0, 31\n\t"
+        "xor     16, 16, 20\n\t"
+        "xor     18, 18, 21\n\t"
+        "xor     17, 17, 22\n\t"
+        "xor     19, 19, 23\n\t"
+        "rlwinm  20, 12, 26, 22, 29\n\t"
+        "rlwinm  21, 15, 26, 22, 29\n\t"
+        "rlwinm  22, 11, 26, 22, 29\n\t"
+        "rlwinm  23, 14, 26, 22, 29\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "rlwimi  20, 20, 24, 0, 31\n\t"
+        "rlwimi  21, 21, 24, 0, 31\n\t"
+        "rlwimi  22, 22, 24, 0, 31\n\t"
+        "rlwimi  23, 23, 24, 0, 31\n\t"
+        "xor     16, 16, 20\n\t"
+        "xor     18, 18, 21\n\t"
+        "xor     17, 17, 22\n\t"
+        "xor     19, 19, 23\n\t"
+        "rlwinm  20, 14, 2, 22, 29\n\t"
+        "rlwinm  21, 12, 2, 22, 29\n\t"
+        "rlwinm  22, 15, 2, 22, 29\n\t"
+        "rlwinm  23, 11, 2, 22, 29\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "rlwimi  20, 20, 16, 0, 31\n\t"
+        "rlwimi  21, 21, 16, 0, 31\n\t"
+        "rlwimi  22, 22, 16, 0, 31\n\t"
+        "rlwimi  23, 23, 16, 0, 31\n\t"
+        "ld      11, 0(28)\n\t"
+        "ld      12, 8(28)\n\t"
+        "xor     16, 16, 20\n\t"
+        "xor     18, 18, 21\n\t"
+        "xor     17, 17, 22\n\t"
+        "xor     19, 19, 23\n\t"
+        "addi    28, 28, 16\n\t"
+        "rldimi  16, 18, 32, 0\n\t"
+        "rldimi  17, 19, 32, 0\n\t"
+        /*   XOR in Key Schedule */
+        "xor     16, 16, 11\n\t"
+        "xor     17, 17, 12\n\t"
+        "srdi    18, 16, 32\n\t"
+        "srdi    19, 17, 32\n\t"
+        "rlwinm  11, 19, 18, 22, 29\n\t"
+        "rlwinm  14, 16, 18, 22, 29\n\t"
+        "rlwinm  12, 18, 18, 22, 29\n\t"
+        "rlwinm  15, 17, 18, 22, 29\n\t"
+        "lwzx    11, 9, 11\n\t"
+        "lwzx    14, 9, 14\n\t"
+        "lwzx    12, 9, 12\n\t"
+        "lwzx    15, 9, 15\n\t"
+        "rlwinm  20, 16, 10, 22, 29\n\t"
+        "rlwinm  21, 18, 10, 22, 29\n\t"
+        "rlwinm  22, 17, 10, 22, 29\n\t"
+        "rlwinm  23, 19, 10, 22, 29\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "rlwimi  20, 20, 8, 0, 31\n\t"
+        "rlwimi  21, 21, 8, 0, 31\n\t"
+        "rlwimi  22, 22, 8, 0, 31\n\t"
+        "rlwimi  23, 23, 8, 0, 31\n\t"
+        "xor     11, 11, 20\n\t"
+        "xor     14, 14, 21\n\t"
+        "xor     12, 12, 22\n\t"
+        "xor     15, 15, 23\n\t"
+        "rlwinm  20, 17, 26, 22, 29\n\t"
+        "rlwinm  21, 19, 26, 22, 29\n\t"
+        "rlwinm  22, 16, 26, 22, 29\n\t"
+        "rlwinm  23, 18, 26, 22, 29\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "rlwimi  20, 20, 24, 0, 31\n\t"
+        "rlwimi  21, 21, 24, 0, 31\n\t"
+        "rlwimi  22, 22, 24, 0, 31\n\t"
+        "rlwimi  23, 23, 24, 0, 31\n\t"
+        "xor     11, 11, 20\n\t"
+        "xor     14, 14, 21\n\t"
+        "xor     12, 12, 22\n\t"
+        "xor     15, 15, 23\n\t"
+        "rlwinm  20, 18, 2, 22, 29\n\t"
+        "rlwinm  21, 17, 2, 22, 29\n\t"
+        "rlwinm  22, 19, 2, 22, 29\n\t"
+        "rlwinm  23, 16, 2, 22, 29\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "rlwimi  20, 20, 16, 0, 31\n\t"
+        "rlwimi  21, 21, 16, 0, 31\n\t"
+        "rlwimi  22, 22, 16, 0, 31\n\t"
+        "rlwimi  23, 23, 16, 0, 31\n\t"
+        "ld      16, 0(28)\n\t"
+        "ld      17, 8(28)\n\t"
+        "xor     11, 11, 20\n\t"
+        "xor     14, 14, 21\n\t"
+        "xor     12, 12, 22\n\t"
+        "xor     15, 15, 23\n\t"
+        "addi    28, 28, 16\n\t"
+        "rldimi  11, 14, 32, 0\n\t"
+        "rldimi  12, 15, 32, 0\n\t"
+        /*   XOR in Key Schedule */
+        "xor     11, 11, 16\n\t"
+        "xor     12, 12, 17\n\t"
+        "bdnz    L_AES_CBC_decrypt_loop_nr_even_%=\n\t"
+        "srdi    14, 11, 32\n\t"
+        "srdi    15, 12, 32\n\t"
+        "rlwinm  16, 15, 18, 22, 29\n\t"
+        "rlwinm  18, 11, 18, 22, 29\n\t"
+        "rlwinm  17, 14, 18, 22, 29\n\t"
+        "rlwinm  19, 12, 18, 22, 29\n\t"
+        "lwzx    16, 9, 16\n\t"
+        "lwzx    18, 9, 18\n\t"
+        "lwzx    17, 9, 17\n\t"
+        "lwzx    19, 9, 19\n\t"
+        "rlwinm  20, 11, 10, 22, 29\n\t"
+        "rlwinm  21, 14, 10, 22, 29\n\t"
+        "rlwinm  22, 12, 10, 22, 29\n\t"
+        "rlwinm  23, 15, 10, 22, 29\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "rlwimi  20, 20, 8, 0, 31\n\t"
+        "rlwimi  21, 21, 8, 0, 31\n\t"
+        "rlwimi  22, 22, 8, 0, 31\n\t"
+        "rlwimi  23, 23, 8, 0, 31\n\t"
+        "xor     16, 16, 20\n\t"
+        "xor     18, 18, 21\n\t"
+        "xor     17, 17, 22\n\t"
+        "xor     19, 19, 23\n\t"
+        "rlwinm  20, 12, 26, 22, 29\n\t"
+        "rlwinm  21, 15, 26, 22, 29\n\t"
+        "rlwinm  22, 11, 26, 22, 29\n\t"
+        "rlwinm  23, 14, 26, 22, 29\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "rlwimi  20, 20, 24, 0, 31\n\t"
+        "rlwimi  21, 21, 24, 0, 31\n\t"
+        "rlwimi  22, 22, 24, 0, 31\n\t"
+        "rlwimi  23, 23, 24, 0, 31\n\t"
+        "xor     16, 16, 20\n\t"
+        "xor     18, 18, 21\n\t"
+        "xor     17, 17, 22\n\t"
+        "xor     19, 19, 23\n\t"
+        "rlwinm  20, 14, 2, 22, 29\n\t"
+        "rlwinm  21, 12, 2, 22, 29\n\t"
+        "rlwinm  22, 15, 2, 22, 29\n\t"
+        "rlwinm  23, 11, 2, 22, 29\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "rlwimi  20, 20, 16, 0, 31\n\t"
+        "rlwimi  21, 21, 16, 0, 31\n\t"
+        "rlwimi  22, 22, 16, 0, 31\n\t"
+        "rlwimi  23, 23, 16, 0, 31\n\t"
+        "ld      11, 0(28)\n\t"
+        "ld      12, 8(28)\n\t"
+        "xor     16, 16, 20\n\t"
+        "xor     18, 18, 21\n\t"
+        "xor     17, 17, 22\n\t"
+        "xor     19, 19, 23\n\t"
+        "addi    28, 28, 16\n\t"
+        "rldimi  16, 18, 32, 0\n\t"
+        "rldimi  17, 19, 32, 0\n\t"
+        /*   XOR in Key Schedule */
+        "xor     16, 16, 11\n\t"
+        "xor     17, 17, 12\n\t"
+        "ld      19, 0(10)\n\t"
+        "ld      19, 64(10)\n\t"
+        "ld      19, 128(10)\n\t"
+        "ld      19, 192(10)\n\t"
+        "ld      19, 256(10)\n\t"
+        "ld      19, 320(10)\n\t"
+        "ld      19, 384(10)\n\t"
+        "ld      19, 448(10)\n\t"
+        "ld      19, 512(10)\n\t"
+        "ld      19, 576(10)\n\t"
+        "ld      19, 640(10)\n\t"
+        "ld      19, 704(10)\n\t"
+        "ld      19, 768(10)\n\t"
+        "ld      19, 832(10)\n\t"
+        "ld      19, 896(10)\n\t"
+        "ld      19, 960(10)\n\t"
+        "rldicl  14, 16, 32, 56\n\t"
+        "andi.   11, 17, 255\n\t"
+        "rldicl  15, 17, 32, 56\n\t"
+        "andi.   12, 16, 255\n\t"
+        "lbzx    14, 10, 14\n\t"
+        "lbzx    11, 10, 11\n\t"
+        "lbzx    15, 10, 15\n\t"
+        "lbzx    12, 10, 12\n\t"
+        "rldimi  11, 14, 32, 24\n\t"
+        "rldimi  12, 15, 32, 24\n\t"
+        "rldicl  20, 17, 56, 56\n\t"
+        "rldicl  21, 17, 24, 56\n\t"
+        "rldicl  22, 16, 56, 56\n\t"
+        "rldicl  23, 16, 24, 56\n\t"
+        "lbzx    20, 10, 20\n\t"
+        "lbzx    21, 10, 21\n\t"
+        "lbzx    22, 10, 22\n\t"
+        "lbzx    23, 10, 23\n\t"
+        "rldimi  11, 20, 40, 16\n\t"
+        "rlwimi  11, 21, 8, 16, 23\n\t"
+        "rldimi  12, 22, 40, 16\n\t"
+        "rlwimi  12, 23, 8, 16, 23\n\t"
+        "rldicl  20, 17, 16, 56\n\t"
+        "rldicl  21, 16, 48, 56\n\t"
+        "rldicl  22, 16, 16, 56\n\t"
+        "rldicl  23, 17, 48, 56\n\t"
+        "lbzx    20, 10, 20\n\t"
+        "lbzx    21, 10, 21\n\t"
+        "lbzx    22, 10, 22\n\t"
+        "lbzx    23, 10, 23\n\t"
+        "rldimi  11, 20, 48, 8\n\t"
+        "rlwimi  11, 21, 16, 8, 15\n\t"
+        "rldimi  12, 22, 48, 8\n\t"
+        "rlwimi  12, 23, 16, 8, 15\n\t"
+        "rldicl  20, 16, 40, 56\n\t"
+        "rldicl  21, 16, 8, 56\n\t"
+        "rldicl  22, 17, 40, 56\n\t"
+        "rldicl  23, 17, 8, 56\n\t"
+        "lbzx    20, 10, 20\n\t"
+        "lbzx    21, 10, 21\n\t"
+        "lbzx    22, 10, 22\n\t"
+        "lbzx    23, 10, 23\n\t"
+        "ld      16, 0(28)\n\t"
+        "ld      17, 8(28)\n\t"
+        "rldimi  11, 20, 56, 0\n\t"
+        "rlwimi  11, 21, 24, 0, 7\n\t"
+        "rldimi  12, 22, 56, 0\n\t"
+        "rlwimi  12, 23, 24, 0, 7\n\t"
+        "addi    28, 28, 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     11, 11, 16\n\t"
+        "xor     12, 12, 17\n\t"
+        "xor     11, 11, 24\n\t"
+        "xor     12, 12, 25\n\t"
+        "std     11, 0(%[out])\n\t"
+        "std     12, 8(%[out])\n\t"
+        "addic.  %[len], %[len], -16\n\t"
+        "addi    %[in], %[in], 16\n\t"
+        "addi    %[out], %[out], 16\n\t"
+        "beq     L_AES_CBC_decrypt_end_dec_odd_%=\n\t"
+        "addi    28, %[ks], 0\n\t"
+        "ld      24, 0(%[in])\n\t"
+        "ld      25, 8(%[in])\n\t"
+        "ld      16, 0(28)\n\t"
+        "ld      17, 8(28)\n\t"
+        "addi    28, 28, 16\n\t"
+        "rldicr  11, 24, 32, 63\n\t"
+        "rldicr  12, 25, 32, 63\n\t"
+        /* Round: 0 - XOR in key schedule */
+        "xor     11, 11, 16\n\t"
+        "xor     12, 12, 17\n\t"
+        "addi    0, %[nr], -2\n\t"
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+        "ld      19, 0(9)\n\t"
+        "ld      19, 64(9)\n\t"
+        "ld      19, 128(9)\n\t"
+        "ld      19, 192(9)\n\t"
+        "ld      19, 256(9)\n\t"
+        "ld      19, 320(9)\n\t"
+        "ld      19, 384(9)\n\t"
+        "ld      19, 448(9)\n\t"
+        "ld      19, 512(9)\n\t"
+        "ld      19, 576(9)\n\t"
+        "ld      19, 640(9)\n\t"
+        "ld      19, 704(9)\n\t"
+        "ld      19, 768(9)\n\t"
+        "ld      19, 832(9)\n\t"
+        "ld      19, 896(9)\n\t"
+        "ld      19, 960(9)\n\t"
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+        "srwi    0, 0, 1\n\t"
+        "mtctr   0\n\t"
+        "\n"
+    "L_AES_CBC_decrypt_loop_nr_odd_%=: \n\t"
+        "srdi    14, 11, 32\n\t"
+        "srdi    15, 12, 32\n\t"
+        "rlwinm  16, 15, 18, 22, 29\n\t"
+        "rlwinm  18, 11, 18, 22, 29\n\t"
+        "rlwinm  17, 14, 18, 22, 29\n\t"
+        "rlwinm  19, 12, 18, 22, 29\n\t"
+        "lwzx    16, 9, 16\n\t"
+        "lwzx    18, 9, 18\n\t"
+        "lwzx    17, 9, 17\n\t"
+        "lwzx    19, 9, 19\n\t"
+        "rlwinm  20, 11, 10, 22, 29\n\t"
+        "rlwinm  21, 14, 10, 22, 29\n\t"
+        "rlwinm  22, 12, 10, 22, 29\n\t"
+        "rlwinm  23, 15, 10, 22, 29\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "rlwimi  20, 20, 8, 0, 31\n\t"
+        "rlwimi  21, 21, 8, 0, 31\n\t"
+        "rlwimi  22, 22, 8, 0, 31\n\t"
+        "rlwimi  23, 23, 8, 0, 31\n\t"
+        "xor     16, 16, 20\n\t"
+        "xor     18, 18, 21\n\t"
+        "xor     17, 17, 22\n\t"
+        "xor     19, 19, 23\n\t"
+        "rlwinm  20, 12, 26, 22, 29\n\t"
+        "rlwinm  21, 15, 26, 22, 29\n\t"
+        "rlwinm  22, 11, 26, 22, 29\n\t"
+        "rlwinm  23, 14, 26, 22, 29\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "rlwimi  20, 20, 24, 0, 31\n\t"
+        "rlwimi  21, 21, 24, 0, 31\n\t"
+        "rlwimi  22, 22, 24, 0, 31\n\t"
+        "rlwimi  23, 23, 24, 0, 31\n\t"
+        "xor     16, 16, 20\n\t"
+        "xor     18, 18, 21\n\t"
+        "xor     17, 17, 22\n\t"
+        "xor     19, 19, 23\n\t"
+        "rlwinm  20, 14, 2, 22, 29\n\t"
+        "rlwinm  21, 12, 2, 22, 29\n\t"
+        "rlwinm  22, 15, 2, 22, 29\n\t"
+        "rlwinm  23, 11, 2, 22, 29\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "rlwimi  20, 20, 16, 0, 31\n\t"
+        "rlwimi  21, 21, 16, 0, 31\n\t"
+        "rlwimi  22, 22, 16, 0, 31\n\t"
+        "rlwimi  23, 23, 16, 0, 31\n\t"
+        "ld      11, 0(28)\n\t"
+        "ld      12, 8(28)\n\t"
+        "xor     16, 16, 20\n\t"
+        "xor     18, 18, 21\n\t"
+        "xor     17, 17, 22\n\t"
+        "xor     19, 19, 23\n\t"
+        "addi    28, 28, 16\n\t"
+        "rldimi  16, 18, 32, 0\n\t"
+        "rldimi  17, 19, 32, 0\n\t"
+        /*   XOR in Key Schedule */
+        "xor     16, 16, 11\n\t"
+        "xor     17, 17, 12\n\t"
+        "srdi    18, 16, 32\n\t"
+        "srdi    19, 17, 32\n\t"
+        "rlwinm  11, 19, 18, 22, 29\n\t"
+        "rlwinm  14, 16, 18, 22, 29\n\t"
+        "rlwinm  12, 18, 18, 22, 29\n\t"
+        "rlwinm  15, 17, 18, 22, 29\n\t"
+        "lwzx    11, 9, 11\n\t"
+        "lwzx    14, 9, 14\n\t"
+        "lwzx    12, 9, 12\n\t"
+        "lwzx    15, 9, 15\n\t"
+        "rlwinm  20, 16, 10, 22, 29\n\t"
+        "rlwinm  21, 18, 10, 22, 29\n\t"
+        "rlwinm  22, 17, 10, 22, 29\n\t"
+        "rlwinm  23, 19, 10, 22, 29\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "rlwimi  20, 20, 8, 0, 31\n\t"
+        "rlwimi  21, 21, 8, 0, 31\n\t"
+        "rlwimi  22, 22, 8, 0, 31\n\t"
+        "rlwimi  23, 23, 8, 0, 31\n\t"
+        "xor     11, 11, 20\n\t"
+        "xor     14, 14, 21\n\t"
+        "xor     12, 12, 22\n\t"
+        "xor     15, 15, 23\n\t"
+        "rlwinm  20, 17, 26, 22, 29\n\t"
+        "rlwinm  21, 19, 26, 22, 29\n\t"
+        "rlwinm  22, 16, 26, 22, 29\n\t"
+        "rlwinm  23, 18, 26, 22, 29\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "rlwimi  20, 20, 24, 0, 31\n\t"
+        "rlwimi  21, 21, 24, 0, 31\n\t"
+        "rlwimi  22, 22, 24, 0, 31\n\t"
+        "rlwimi  23, 23, 24, 0, 31\n\t"
+        "xor     11, 11, 20\n\t"
+        "xor     14, 14, 21\n\t"
+        "xor     12, 12, 22\n\t"
+        "xor     15, 15, 23\n\t"
+        "rlwinm  20, 18, 2, 22, 29\n\t"
+        "rlwinm  21, 17, 2, 22, 29\n\t"
+        "rlwinm  22, 19, 2, 22, 29\n\t"
+        "rlwinm  23, 16, 2, 22, 29\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "rlwimi  20, 20, 16, 0, 31\n\t"
+        "rlwimi  21, 21, 16, 0, 31\n\t"
+        "rlwimi  22, 22, 16, 0, 31\n\t"
+        "rlwimi  23, 23, 16, 0, 31\n\t"
+        "ld      16, 0(28)\n\t"
+        "ld      17, 8(28)\n\t"
+        "xor     11, 11, 20\n\t"
+        "xor     14, 14, 21\n\t"
+        "xor     12, 12, 22\n\t"
+        "xor     15, 15, 23\n\t"
+        "addi    28, 28, 16\n\t"
+        "rldimi  11, 14, 32, 0\n\t"
+        "rldimi  12, 15, 32, 0\n\t"
+        /*   XOR in Key Schedule */
+        "xor     11, 11, 16\n\t"
+        "xor     12, 12, 17\n\t"
+        "bdnz    L_AES_CBC_decrypt_loop_nr_odd_%=\n\t"
+        "srdi    14, 11, 32\n\t"
+        "srdi    15, 12, 32\n\t"
+        "rlwinm  16, 15, 18, 22, 29\n\t"
+        "rlwinm  18, 11, 18, 22, 29\n\t"
+        "rlwinm  17, 14, 18, 22, 29\n\t"
+        "rlwinm  19, 12, 18, 22, 29\n\t"
+        "lwzx    16, 9, 16\n\t"
+        "lwzx    18, 9, 18\n\t"
+        "lwzx    17, 9, 17\n\t"
+        "lwzx    19, 9, 19\n\t"
+        "rlwinm  20, 11, 10, 22, 29\n\t"
+        "rlwinm  21, 14, 10, 22, 29\n\t"
+        "rlwinm  22, 12, 10, 22, 29\n\t"
+        "rlwinm  23, 15, 10, 22, 29\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "rlwimi  20, 20, 8, 0, 31\n\t"
+        "rlwimi  21, 21, 8, 0, 31\n\t"
+        "rlwimi  22, 22, 8, 0, 31\n\t"
+        "rlwimi  23, 23, 8, 0, 31\n\t"
+        "xor     16, 16, 20\n\t"
+        "xor     18, 18, 21\n\t"
+        "xor     17, 17, 22\n\t"
+        "xor     19, 19, 23\n\t"
+        "rlwinm  20, 12, 26, 22, 29\n\t"
+        "rlwinm  21, 15, 26, 22, 29\n\t"
+        "rlwinm  22, 11, 26, 22, 29\n\t"
+        "rlwinm  23, 14, 26, 22, 29\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "rlwimi  20, 20, 24, 0, 31\n\t"
+        "rlwimi  21, 21, 24, 0, 31\n\t"
+        "rlwimi  22, 22, 24, 0, 31\n\t"
+        "rlwimi  23, 23, 24, 0, 31\n\t"
+        "xor     16, 16, 20\n\t"
+        "xor     18, 18, 21\n\t"
+        "xor     17, 17, 22\n\t"
+        "xor     19, 19, 23\n\t"
+        "rlwinm  20, 14, 2, 22, 29\n\t"
+        "rlwinm  21, 12, 2, 22, 29\n\t"
+        "rlwinm  22, 15, 2, 22, 29\n\t"
+        "rlwinm  23, 11, 2, 22, 29\n\t"
+        "lwzx    20, 9, 20\n\t"
+        "lwzx    21, 9, 21\n\t"
+        "lwzx    22, 9, 22\n\t"
+        "lwzx    23, 9, 23\n\t"
+        "rlwimi  20, 20, 16, 0, 31\n\t"
+        "rlwimi  21, 21, 16, 0, 31\n\t"
+        "rlwimi  22, 22, 16, 0, 31\n\t"
+        "rlwimi  23, 23, 16, 0, 31\n\t"
+        "ld      11, 0(28)\n\t"
+        "ld      12, 8(28)\n\t"
+        "xor     16, 16, 20\n\t"
+        "xor     18, 18, 21\n\t"
+        "xor     17, 17, 22\n\t"
+        "xor     19, 19, 23\n\t"
+        "addi    28, 28, 16\n\t"
+        "rldimi  16, 18, 32, 0\n\t"
+        "rldimi  17, 19, 32, 0\n\t"
+        /*   XOR in Key Schedule */
+        "xor     16, 16, 11\n\t"
+        "xor     17, 17, 12\n\t"
+        "ld      19, 0(10)\n\t"
+        "ld      19, 64(10)\n\t"
+        "ld      19, 128(10)\n\t"
+        "ld      19, 192(10)\n\t"
+        "ld      19, 256(10)\n\t"
+        "ld      19, 320(10)\n\t"
+        "ld      19, 384(10)\n\t"
+        "ld      19, 448(10)\n\t"
+        "ld      19, 512(10)\n\t"
+        "ld      19, 576(10)\n\t"
+        "ld      19, 640(10)\n\t"
+        "ld      19, 704(10)\n\t"
+        "ld      19, 768(10)\n\t"
+        "ld      19, 832(10)\n\t"
+        "ld      19, 896(10)\n\t"
+        "ld      19, 960(10)\n\t"
+        "rldicl  14, 16, 32, 56\n\t"
+        "andi.   11, 17, 255\n\t"
+        "rldicl  15, 17, 32, 56\n\t"
+        "andi.   12, 16, 255\n\t"
+        "lbzx    14, 10, 14\n\t"
+        "lbzx    11, 10, 11\n\t"
+        "lbzx    15, 10, 15\n\t"
+        "lbzx    12, 10, 12\n\t"
+        "rldimi  11, 14, 32, 24\n\t"
+        "rldimi  12, 15, 32, 24\n\t"
+        "rldicl  20, 17, 56, 56\n\t"
+        "rldicl  21, 17, 24, 56\n\t"
+        "rldicl  22, 16, 56, 56\n\t"
+        "rldicl  23, 16, 24, 56\n\t"
+        "lbzx    20, 10, 20\n\t"
+        "lbzx    21, 10, 21\n\t"
+        "lbzx    22, 10, 22\n\t"
+        "lbzx    23, 10, 23\n\t"
+        "rldimi  11, 20, 40, 16\n\t"
+        "rlwimi  11, 21, 8, 16, 23\n\t"
+        "rldimi  12, 22, 40, 16\n\t"
+        "rlwimi  12, 23, 8, 16, 23\n\t"
+        "rldicl  20, 17, 16, 56\n\t"
+        "rldicl  21, 16, 48, 56\n\t"
+        "rldicl  22, 16, 16, 56\n\t"
+        "rldicl  23, 17, 48, 56\n\t"
+        "lbzx    20, 10, 20\n\t"
+        "lbzx    21, 10, 21\n\t"
+        "lbzx    22, 10, 22\n\t"
+        "lbzx    23, 10, 23\n\t"
+        "rldimi  11, 20, 48, 8\n\t"
+        "rlwimi  11, 21, 16, 8, 15\n\t"
+        "rldimi  12, 22, 48, 8\n\t"
+        "rlwimi  12, 23, 16, 8, 15\n\t"
+        "rldicl  20, 16, 40, 56\n\t"
+        "rldicl  21, 16, 8, 56\n\t"
+        "rldicl  22, 17, 40, 56\n\t"
+        "rldicl  23, 17, 8, 56\n\t"
+        "lbzx    20, 10, 20\n\t"
+        "lbzx    21, 10, 21\n\t"
+        "lbzx    22, 10, 22\n\t"
+        "lbzx    23, 10, 23\n\t"
+        "ld      16, 0(28)\n\t"
+        "ld      17, 8(28)\n\t"
+        "rldimi  11, 20, 56, 0\n\t"
+        "rlwimi  11, 21, 24, 0, 7\n\t"
+        "rldimi  12, 22, 56, 0\n\t"
+        "rlwimi  12, 23, 24, 0, 7\n\t"
+        "addi    28, 28, 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     11, 11, 16\n\t"
+        "xor     12, 12, 17\n\t"
+        "xor     11, 11, 26\n\t"
+        "xor     12, 12, 27\n\t"
+        "std     11, 0(%[out])\n\t"
+        "std     12, 8(%[out])\n\t"
+        "addic.  %[len], %[len], -16\n\t"
+        "addi    %[in], %[in], 16\n\t"
+        "addi    %[out], %[out], 16\n\t"
+        "bne     L_AES_CBC_decrypt_loop_block_%=\n\t"
+        "std     24, 0(%[iv])\n\t"
+        "std     25, 8(%[iv])\n\t"
+        "b       L_AES_CBC_decrypt_end_dec_%=\n\t"
+        "\n"
+    "L_AES_CBC_decrypt_end_dec_odd_%=: \n\t"
+        "std     26, 0(%[iv])\n\t"
+        "std     27, 8(%[iv])\n\t"
+        "\n"
+    "L_AES_CBC_decrypt_end_dec_%=: \n\t"
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+        : [in] "+r" (in), [out] "+r" (out), [len] "+r" (len), [ks] "+r" (ks),
+          [nr] "+r" (nr), [iv] "+r" (iv),
+          [L_AES_PPC64_td] "+r" (L_AES_PPC64_td_c),
+          [L_AES_PPC64_td4] "+r" (L_AES_PPC64_td4_c)
+        :
+        : "memory", "cc", "0", "11", "12", "14", "15", "16", "17", "18", "19",
+            "20", "21", "22", "23", "24", "25", "26", "27", "28"
+#else
+        :
+        : [in] "r" (in), [out] "r" (out), [len] "r" (len), [ks] "r" (ks),
+          [nr] "r" (nr), [iv] "r" (iv), [L_AES_PPC64_td] "r" (L_AES_PPC64_td_c),
+          [L_AES_PPC64_td4] "r" (L_AES_PPC64_td4_c)
+        : "memory", "cc", "0", "9", "10", "11", "12", "14", "15", "16", "17",
+            "18", "19", "20", "21", "22", "23", "24", "25", "26", "27", "28"
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+    );
+}
+
+#endif /* HAVE_AES_CBC */
+#endif /* WOLFSSL_AES_DIRECT || WOLFSSL_AES_COUNTER || HAVE_AES_CBC
+        * HAVE_AES_ECB */
+#ifdef WOLFSSL_AES_XTS
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+void AES_XTS_decrypt(const byte* in_p, byte* out_p, word32 sz_p,
+    const byte* i_p, byte* key_p, byte* key2_p, byte* tmp_p, int nr_p)
+#else
+void AES_XTS_decrypt(const byte* in, byte* out, word32 sz, const byte* i,
+    byte* key, byte* key2, byte* tmp, int nr)
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+{
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+    register const byte* in asm ("3") = (const byte*)in_p;
+    register byte* out asm ("4") = (byte*)out_p;
+    register word32 sz asm ("5") = (word32)sz_p;
+    register const byte* i asm ("6") = (const byte*)i_p;
+    register byte* key asm ("7") = (byte*)key_p;
+    register byte* key2 asm ("8") = (byte*)key2_p;
+    register byte* tmp asm ("9") = (byte*)tmp_p;
+    register int nr asm ("10") = (int)nr_p;
+    register word32* L_AES_PPC64_td_c asm ("11") = (word32*)&L_AES_PPC64_td;
+    register byte* L_AES_PPC64_td4_c asm ("12") = (byte*)&L_AES_PPC64_td4;
+    register word32* L_AES_PPC64_te4_0_c asm ("14") =
+        (word32*)&L_AES_PPC64_te4_0;
+#else
+    register word32* L_AES_PPC64_td_c = (word32*)&L_AES_PPC64_td;
+    register byte* L_AES_PPC64_td4_c = (byte*)&L_AES_PPC64_td4;
+    register word32* L_AES_PPC64_te4_0_c = (word32*)&L_AES_PPC64_te4_0;
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+
+    __asm__ __volatile__ (
+        "subi    1, 1, 8\n\t"
+#ifdef WOLFSSL_NO_VAR_ASSIGN_REG
+        "mr      14, %[L_AES_PPC64_te4_0]\n\t"
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+        "addi    15, 14, 0x400\n\t"
+        "addi    16, 14, 0x800\n\t"
+        "addi    17, 14, 0xc00\n\t"
+        "andi.   18, %[sz], 15\n\t"
+        "addi    18, 18, 15\n\t"
+        "andi.   18, 18, 16\n\t"
+        "subf    %[sz], 18, %[sz]\n\t"
+        "ld      26, 0(%[i])\n\t"
+        "ld      27, 8(%[i])\n\t"
+        "ld      18, 0(%[key2])\n\t"
+        "ld      19, 8(%[key2])\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        "rldicr  26, 26, 32, 63\n\t"
+        "rldicr  27, 27, 32, 63\n\t"
+        /* Round: 0 - XOR in key schedule */
+        "xor     26, 26, 18\n\t"
+        "xor     27, 27, 19\n\t"
+        "addi    0, %[nr], -2\n\t"
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+        "ld      21, 0(14)\n\t"
+        "ld      21, 32(14)\n\t"
+        "ld      21, 64(14)\n\t"
+        "ld      21, 96(14)\n\t"
+        "ld      21, 128(14)\n\t"
+        "ld      21, 160(14)\n\t"
+        "ld      21, 192(14)\n\t"
+        "ld      21, 224(14)\n\t"
+        "ld      21, 256(14)\n\t"
+        "ld      21, 288(14)\n\t"
+        "ld      21, 320(14)\n\t"
+        "ld      21, 352(14)\n\t"
+        "ld      21, 384(14)\n\t"
+        "ld      21, 416(14)\n\t"
+        "ld      21, 448(14)\n\t"
+        "ld      21, 480(14)\n\t"
+        "ld      21, 512(14)\n\t"
+        "ld      21, 544(14)\n\t"
+        "ld      21, 576(14)\n\t"
+        "ld      21, 608(14)\n\t"
+        "ld      21, 640(14)\n\t"
+        "ld      21, 672(14)\n\t"
+        "ld      21, 704(14)\n\t"
+        "ld      21, 736(14)\n\t"
+        "ld      21, 768(14)\n\t"
+        "ld      21, 800(14)\n\t"
+        "ld      21, 832(14)\n\t"
+        "ld      21, 864(14)\n\t"
+        "ld      21, 896(14)\n\t"
+        "ld      21, 928(14)\n\t"
+        "ld      21, 960(14)\n\t"
+        "ld      21, 992(14)\n\t"
+        "ld      21, 1024(14)\n\t"
+        "ld      21, 1056(14)\n\t"
+        "ld      21, 1088(14)\n\t"
+        "ld      21, 1120(14)\n\t"
+        "ld      21, 1152(14)\n\t"
+        "ld      21, 1184(14)\n\t"
+        "ld      21, 1216(14)\n\t"
+        "ld      21, 1248(14)\n\t"
+        "ld      21, 1280(14)\n\t"
+        "ld      21, 1312(14)\n\t"
+        "ld      21, 1344(14)\n\t"
+        "ld      21, 1376(14)\n\t"
+        "ld      21, 1408(14)\n\t"
+        "ld      21, 1440(14)\n\t"
+        "ld      21, 1472(14)\n\t"
+        "ld      21, 1504(14)\n\t"
+        "ld      21, 1536(14)\n\t"
+        "ld      21, 1568(14)\n\t"
+        "ld      21, 1600(14)\n\t"
+        "ld      21, 1632(14)\n\t"
+        "ld      21, 1664(14)\n\t"
+        "ld      21, 1696(14)\n\t"
+        "ld      21, 1728(14)\n\t"
+        "ld      21, 1760(14)\n\t"
+        "ld      21, 1792(14)\n\t"
+        "ld      21, 1824(14)\n\t"
+        "ld      21, 1856(14)\n\t"
+        "ld      21, 1888(14)\n\t"
+        "ld      21, 1920(14)\n\t"
+        "ld      21, 1952(14)\n\t"
+        "ld      21, 1984(14)\n\t"
+        "ld      21, 2016(14)\n\t"
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+        "srwi    0, 0, 1\n\t"
+        "mtctr   0\n\t"
+        "\n"
+    "L_AES_XTS_decrypt_loop_nr_tweak_%=: \n\t"
+        "srdi    28, 26, 32\n\t"
+        "srdi    29, 27, 32\n\t"
+        "rlwinm  18, 26, 10, 22, 29\n\t"
+        "rlwinm  19, 27, 10, 22, 29\n\t"
+        "rlwinm  20, 28, 10, 22, 29\n\t"
+        "rlwinm  21, 29, 10, 22, 29\n\t"
+        "lwzx    18, 14, 18\n\t"
+        "lwzx    19, 14, 19\n\t"
+        "lwzx    20, 14, 20\n\t"
+        "lwzx    21, 14, 21\n\t"
+        "rlwinm  22, 28, 18, 22, 29\n\t"
+        "rlwinm  23, 27, 18, 22, 29\n\t"
+        "rlwinm  24, 29, 18, 22, 29\n\t"
+        "rlwinm  25, 26, 18, 22, 29\n\t"
+        "lwzx    22, 15, 22\n\t"
+        "lwzx    23, 15, 23\n\t"
+        "lwzx    24, 15, 24\n\t"
+        "lwzx    25, 15, 25\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 27, 26, 22, 29\n\t"
+        "rlwinm  23, 29, 26, 22, 29\n\t"
+        "rlwinm  24, 26, 26, 22, 29\n\t"
+        "rlwinm  25, 28, 26, 22, 29\n\t"
+        "lwzx    22, 16, 22\n\t"
+        "lwzx    23, 16, 23\n\t"
+        "lwzx    24, 16, 24\n\t"
+        "lwzx    25, 16, 25\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 29, 2, 22, 29\n\t"
+        "rlwinm  23, 26, 2, 22, 29\n\t"
+        "rlwinm  24, 28, 2, 22, 29\n\t"
+        "rlwinm  25, 27, 2, 22, 29\n\t"
+        "lwzx    22, 17, 22\n\t"
+        "lwzx    23, 17, 23\n\t"
+        "lwzx    24, 17, 24\n\t"
+        "lwzx    25, 17, 25\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "ld      26, 0(%[key2])\n\t"
+        "ld      27, 8(%[key2])\n\t"
+        "rldimi  18, 20, 32, 0\n\t"
+        "rldimi  19, 21, 32, 0\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     18, 18, 26\n\t"
+        "xor     19, 19, 27\n\t"
+        "srdi    20, 18, 32\n\t"
+        "srdi    21, 19, 32\n\t"
+        "rlwinm  26, 18, 10, 22, 29\n\t"
+        "rlwinm  27, 19, 10, 22, 29\n\t"
+        "rlwinm  28, 20, 10, 22, 29\n\t"
+        "rlwinm  29, 21, 10, 22, 29\n\t"
+        "lwzx    26, 14, 26\n\t"
+        "lwzx    27, 14, 27\n\t"
+        "lwzx    28, 14, 28\n\t"
+        "lwzx    29, 14, 29\n\t"
+        "rlwinm  22, 20, 18, 22, 29\n\t"
+        "rlwinm  23, 19, 18, 22, 29\n\t"
+        "rlwinm  24, 21, 18, 22, 29\n\t"
+        "rlwinm  25, 18, 18, 22, 29\n\t"
+        "lwzx    22, 15, 22\n\t"
+        "lwzx    23, 15, 23\n\t"
+        "lwzx    24, 15, 24\n\t"
+        "lwzx    25, 15, 25\n\t"
+        "xor     26, 26, 22\n\t"
+        "xor     28, 28, 23\n\t"
+        "xor     27, 27, 24\n\t"
+        "xor     29, 29, 25\n\t"
+        "rlwinm  22, 19, 26, 22, 29\n\t"
+        "rlwinm  23, 21, 26, 22, 29\n\t"
+        "rlwinm  24, 18, 26, 22, 29\n\t"
+        "rlwinm  25, 20, 26, 22, 29\n\t"
+        "lwzx    22, 16, 22\n\t"
+        "lwzx    23, 16, 23\n\t"
+        "lwzx    24, 16, 24\n\t"
+        "lwzx    25, 16, 25\n\t"
+        "xor     26, 26, 22\n\t"
+        "xor     28, 28, 23\n\t"
+        "xor     27, 27, 24\n\t"
+        "xor     29, 29, 25\n\t"
+        "rlwinm  22, 21, 2, 22, 29\n\t"
+        "rlwinm  23, 18, 2, 22, 29\n\t"
+        "rlwinm  24, 20, 2, 22, 29\n\t"
+        "rlwinm  25, 19, 2, 22, 29\n\t"
+        "lwzx    22, 17, 22\n\t"
+        "lwzx    23, 17, 23\n\t"
+        "lwzx    24, 17, 24\n\t"
+        "lwzx    25, 17, 25\n\t"
+        "xor     26, 26, 22\n\t"
+        "xor     28, 28, 23\n\t"
+        "xor     27, 27, 24\n\t"
+        "xor     29, 29, 25\n\t"
+        "ld      18, 0(%[key2])\n\t"
+        "ld      19, 8(%[key2])\n\t"
+        "rldimi  26, 28, 32, 0\n\t"
+        "rldimi  27, 29, 32, 0\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     26, 26, 18\n\t"
+        "xor     27, 27, 19\n\t"
+        "bdnz    L_AES_XTS_decrypt_loop_nr_tweak_%=\n\t"
+        "srdi    28, 26, 32\n\t"
+        "srdi    29, 27, 32\n\t"
+        "rlwinm  18, 26, 10, 22, 29\n\t"
+        "rlwinm  19, 27, 10, 22, 29\n\t"
+        "rlwinm  20, 28, 10, 22, 29\n\t"
+        "rlwinm  21, 29, 10, 22, 29\n\t"
+        "lwzx    18, 14, 18\n\t"
+        "lwzx    19, 14, 19\n\t"
+        "lwzx    20, 14, 20\n\t"
+        "lwzx    21, 14, 21\n\t"
+        "rlwinm  22, 28, 18, 22, 29\n\t"
+        "rlwinm  23, 27, 18, 22, 29\n\t"
+        "rlwinm  24, 29, 18, 22, 29\n\t"
+        "rlwinm  25, 26, 18, 22, 29\n\t"
+        "lwzx    22, 15, 22\n\t"
+        "lwzx    23, 15, 23\n\t"
+        "lwzx    24, 15, 24\n\t"
+        "lwzx    25, 15, 25\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 27, 26, 22, 29\n\t"
+        "rlwinm  23, 29, 26, 22, 29\n\t"
+        "rlwinm  24, 26, 26, 22, 29\n\t"
+        "rlwinm  25, 28, 26, 22, 29\n\t"
+        "lwzx    22, 16, 22\n\t"
+        "lwzx    23, 16, 23\n\t"
+        "lwzx    24, 16, 24\n\t"
+        "lwzx    25, 16, 25\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 29, 2, 22, 29\n\t"
+        "rlwinm  23, 26, 2, 22, 29\n\t"
+        "rlwinm  24, 28, 2, 22, 29\n\t"
+        "rlwinm  25, 27, 2, 22, 29\n\t"
+        "lwzx    22, 17, 22\n\t"
+        "lwzx    23, 17, 23\n\t"
+        "lwzx    24, 17, 24\n\t"
+        "lwzx    25, 17, 25\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "ld      26, 0(%[key2])\n\t"
+        "ld      27, 8(%[key2])\n\t"
+        "rldimi  18, 20, 32, 0\n\t"
+        "rldimi  19, 21, 32, 0\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     18, 18, 26\n\t"
+        "xor     19, 19, 27\n\t"
+        "srdi    20, 18, 32\n\t"
+        "srdi    21, 19, 32\n\t"
+        "rlwinm  26, 18, 2, 22, 29\n\t"
+        "rlwinm  27, 19, 2, 22, 29\n\t"
+        "rlwinm  29, 20, 2, 22, 29\n\t"
+        "rlwinm  28, 21, 2, 22, 29\n\t"
+        "lwzx    26, 15, 26\n\t"
+        "lwzx    27, 15, 27\n\t"
+        "lwzx    29, 15, 29\n\t"
+        "lwzx    28, 15, 28\n\t"
+        "rlwinm  22, 19, 26, 22, 29\n\t"
+        "rlwinm  23, 21, 26, 22, 29\n\t"
+        "rldimi  27, 29, 32, 24\n\t"
+        "rldimi  26, 28, 32, 24\n\t"
+        "rlwinm  24, 18, 26, 22, 29\n\t"
+        "rlwinm  25, 20, 26, 22, 29\n\t"
+        "lwzx    22, 15, 22\n\t"
+        "lwzx    23, 15, 23\n\t"
+        "lwzx    24, 15, 24\n\t"
+        "lwzx    25, 15, 25\n\t"
+        "rldimi  26, 22, 40, 16\n\t"
+        "rlwimi  26, 23, 8, 16, 23\n\t"
+        "rldimi  27, 24, 40, 16\n\t"
+        "rlwimi  27, 25, 8, 16, 23\n\t"
+        "rlwinm  22, 20, 18, 22, 29\n\t"
+        "rlwinm  23, 19, 18, 22, 29\n\t"
+        "rlwinm  24, 21, 18, 22, 29\n\t"
+        "rlwinm  25, 18, 18, 22, 29\n\t"
+        "lwzx    22, 15, 22\n\t"
+        "lwzx    23, 15, 23\n\t"
+        "lwzx    24, 15, 24\n\t"
+        "lwzx    25, 15, 25\n\t"
+        "rldimi  26, 22, 48, 8\n\t"
+        "rlwimi  26, 23, 16, 8, 15\n\t"
+        "rldimi  27, 24, 48, 8\n\t"
+        "rlwimi  27, 25, 16, 8, 15\n\t"
+        "rlwinm  22, 18, 10, 22, 29\n\t"
+        "rlwinm  23, 20, 10, 22, 29\n\t"
+        "rlwinm  24, 19, 10, 22, 29\n\t"
+        "rlwinm  25, 21, 10, 22, 29\n\t"
+        "lwzx    22, 15, 22\n\t"
+        "lwzx    23, 15, 23\n\t"
+        "lwzx    24, 15, 24\n\t"
+        "lwzx    25, 15, 25\n\t"
+        "ld      18, 0(%[key2])\n\t"
+        "ld      19, 8(%[key2])\n\t"
+        "rldimi  26, 22, 56, 0\n\t"
+        "rlwimi  26, 23, 24, 0, 7\n\t"
+        "rldimi  27, 24, 56, 0\n\t"
+        "rlwimi  27, 25, 24, 0, 7\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     26, 26, 18\n\t"
+        "xor     27, 27, 19\n\t"
+        "std     26, 0(%[tmp])\n\t"
+        "std     27, 8(%[tmp])\n\t"
+#ifdef WOLFSSL_NO_VAR_ASSIGN_REG
+        "mr      11, %[L_AES_PPC64_td]\n\t"
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+#ifdef WOLFSSL_NO_VAR_ASSIGN_REG
+        "mr      12, %[L_AES_PPC64_td4]\n\t"
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+        "cmpdi   %[sz], 16\n\t"
+        "blt     L_AES_XTS_decrypt_start_partail_%=\n\t"
+        "\n"
+    "L_AES_XTS_decrypt_loop_block_%=: \n\t"
+        "addi    %[key2], %[key], 0\n\t"
+        "ld      14, 0(%[in])\n\t"
+        "ld      15, 8(%[in])\n\t"
+        "ld      18, 0(%[key2])\n\t"
+        "ld      19, 8(%[key2])\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        "xor     14, 14, 26\n\t"
+        "xor     15, 15, 27\n\t"
+        "rldicr  14, 14, 32, 63\n\t"
+        "rldicr  15, 15, 32, 63\n\t"
+        /* Round: 0 - XOR in key schedule */
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "addi    0, %[nr], -2\n\t"
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+        "ld      21, 0(11)\n\t"
+        "ld      21, 64(11)\n\t"
+        "ld      21, 128(11)\n\t"
+        "ld      21, 192(11)\n\t"
+        "ld      21, 256(11)\n\t"
+        "ld      21, 320(11)\n\t"
+        "ld      21, 384(11)\n\t"
+        "ld      21, 448(11)\n\t"
+        "ld      21, 512(11)\n\t"
+        "ld      21, 576(11)\n\t"
+        "ld      21, 640(11)\n\t"
+        "ld      21, 704(11)\n\t"
+        "ld      21, 768(11)\n\t"
+        "ld      21, 832(11)\n\t"
+        "ld      21, 896(11)\n\t"
+        "ld      21, 960(11)\n\t"
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+        "srwi    0, 0, 1\n\t"
+        "mtctr   0\n\t"
+        "\n"
+    "L_AES_XTS_decrypt_loop_nr_%=: \n\t"
+        "srdi    16, 14, 32\n\t"
+        "srdi    17, 15, 32\n\t"
+        "rlwinm  18, 17, 18, 22, 29\n\t"
+        "rlwinm  20, 14, 18, 22, 29\n\t"
+        "rlwinm  19, 16, 18, 22, 29\n\t"
+        "rlwinm  21, 15, 18, 22, 29\n\t"
+        "lwzx    18, 11, 18\n\t"
+        "lwzx    20, 11, 20\n\t"
+        "lwzx    19, 11, 19\n\t"
+        "lwzx    21, 11, 21\n\t"
+        "rlwinm  22, 14, 10, 22, 29\n\t"
+        "rlwinm  23, 16, 10, 22, 29\n\t"
+        "rlwinm  24, 15, 10, 22, 29\n\t"
+        "rlwinm  25, 17, 10, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 8, 0, 31\n\t"
+        "rlwimi  23, 23, 8, 0, 31\n\t"
+        "rlwimi  24, 24, 8, 0, 31\n\t"
+        "rlwimi  25, 25, 8, 0, 31\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 15, 26, 22, 29\n\t"
+        "rlwinm  23, 17, 26, 22, 29\n\t"
+        "rlwinm  24, 14, 26, 22, 29\n\t"
+        "rlwinm  25, 16, 26, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 24, 0, 31\n\t"
+        "rlwimi  23, 23, 24, 0, 31\n\t"
+        "rlwimi  24, 24, 24, 0, 31\n\t"
+        "rlwimi  25, 25, 24, 0, 31\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 16, 2, 22, 29\n\t"
+        "rlwinm  23, 15, 2, 22, 29\n\t"
+        "rlwinm  24, 17, 2, 22, 29\n\t"
+        "rlwinm  25, 14, 2, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 16, 0, 31\n\t"
+        "rlwimi  23, 23, 16, 0, 31\n\t"
+        "rlwimi  24, 24, 16, 0, 31\n\t"
+        "rlwimi  25, 25, 16, 0, 31\n\t"
+        "ld      14, 0(%[key2])\n\t"
+        "ld      15, 8(%[key2])\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        "rldimi  18, 20, 32, 0\n\t"
+        "rldimi  19, 21, 32, 0\n\t"
+        /*   XOR in Key Schedule */
+        "xor     18, 18, 14\n\t"
+        "xor     19, 19, 15\n\t"
+        "srdi    20, 18, 32\n\t"
+        "srdi    21, 19, 32\n\t"
+        "rlwinm  14, 21, 18, 22, 29\n\t"
+        "rlwinm  16, 18, 18, 22, 29\n\t"
+        "rlwinm  15, 20, 18, 22, 29\n\t"
+        "rlwinm  17, 19, 18, 22, 29\n\t"
+        "lwzx    14, 11, 14\n\t"
+        "lwzx    16, 11, 16\n\t"
+        "lwzx    15, 11, 15\n\t"
+        "lwzx    17, 11, 17\n\t"
+        "rlwinm  22, 18, 10, 22, 29\n\t"
+        "rlwinm  23, 20, 10, 22, 29\n\t"
+        "rlwinm  24, 19, 10, 22, 29\n\t"
+        "rlwinm  25, 21, 10, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 8, 0, 31\n\t"
+        "rlwimi  23, 23, 8, 0, 31\n\t"
+        "rlwimi  24, 24, 8, 0, 31\n\t"
+        "rlwimi  25, 25, 8, 0, 31\n\t"
+        "xor     14, 14, 22\n\t"
+        "xor     16, 16, 23\n\t"
+        "xor     15, 15, 24\n\t"
+        "xor     17, 17, 25\n\t"
+        "rlwinm  22, 19, 26, 22, 29\n\t"
+        "rlwinm  23, 21, 26, 22, 29\n\t"
+        "rlwinm  24, 18, 26, 22, 29\n\t"
+        "rlwinm  25, 20, 26, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 24, 0, 31\n\t"
+        "rlwimi  23, 23, 24, 0, 31\n\t"
+        "rlwimi  24, 24, 24, 0, 31\n\t"
+        "rlwimi  25, 25, 24, 0, 31\n\t"
+        "xor     14, 14, 22\n\t"
+        "xor     16, 16, 23\n\t"
+        "xor     15, 15, 24\n\t"
+        "xor     17, 17, 25\n\t"
+        "rlwinm  22, 20, 2, 22, 29\n\t"
+        "rlwinm  23, 19, 2, 22, 29\n\t"
+        "rlwinm  24, 21, 2, 22, 29\n\t"
+        "rlwinm  25, 18, 2, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 16, 0, 31\n\t"
+        "rlwimi  23, 23, 16, 0, 31\n\t"
+        "rlwimi  24, 24, 16, 0, 31\n\t"
+        "rlwimi  25, 25, 16, 0, 31\n\t"
+        "ld      18, 0(%[key2])\n\t"
+        "ld      19, 8(%[key2])\n\t"
+        "xor     14, 14, 22\n\t"
+        "xor     16, 16, 23\n\t"
+        "xor     15, 15, 24\n\t"
+        "xor     17, 17, 25\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        "rldimi  14, 16, 32, 0\n\t"
+        "rldimi  15, 17, 32, 0\n\t"
+        /*   XOR in Key Schedule */
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "bdnz    L_AES_XTS_decrypt_loop_nr_%=\n\t"
+        "srdi    16, 14, 32\n\t"
+        "srdi    17, 15, 32\n\t"
+        "rlwinm  18, 17, 18, 22, 29\n\t"
+        "rlwinm  20, 14, 18, 22, 29\n\t"
+        "rlwinm  19, 16, 18, 22, 29\n\t"
+        "rlwinm  21, 15, 18, 22, 29\n\t"
+        "lwzx    18, 11, 18\n\t"
+        "lwzx    20, 11, 20\n\t"
+        "lwzx    19, 11, 19\n\t"
+        "lwzx    21, 11, 21\n\t"
+        "rlwinm  22, 14, 10, 22, 29\n\t"
+        "rlwinm  23, 16, 10, 22, 29\n\t"
+        "rlwinm  24, 15, 10, 22, 29\n\t"
+        "rlwinm  25, 17, 10, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 8, 0, 31\n\t"
+        "rlwimi  23, 23, 8, 0, 31\n\t"
+        "rlwimi  24, 24, 8, 0, 31\n\t"
+        "rlwimi  25, 25, 8, 0, 31\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 15, 26, 22, 29\n\t"
+        "rlwinm  23, 17, 26, 22, 29\n\t"
+        "rlwinm  24, 14, 26, 22, 29\n\t"
+        "rlwinm  25, 16, 26, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 24, 0, 31\n\t"
+        "rlwimi  23, 23, 24, 0, 31\n\t"
+        "rlwimi  24, 24, 24, 0, 31\n\t"
+        "rlwimi  25, 25, 24, 0, 31\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 16, 2, 22, 29\n\t"
+        "rlwinm  23, 15, 2, 22, 29\n\t"
+        "rlwinm  24, 17, 2, 22, 29\n\t"
+        "rlwinm  25, 14, 2, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 16, 0, 31\n\t"
+        "rlwimi  23, 23, 16, 0, 31\n\t"
+        "rlwimi  24, 24, 16, 0, 31\n\t"
+        "rlwimi  25, 25, 16, 0, 31\n\t"
+        "ld      14, 0(%[key2])\n\t"
+        "ld      15, 8(%[key2])\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        "rldimi  18, 20, 32, 0\n\t"
+        "rldimi  19, 21, 32, 0\n\t"
+        /*   XOR in Key Schedule */
+        "xor     18, 18, 14\n\t"
+        "xor     19, 19, 15\n\t"
+        "ld      21, 0(12)\n\t"
+        "ld      21, 64(12)\n\t"
+        "ld      21, 128(12)\n\t"
+        "ld      21, 192(12)\n\t"
+        "ld      21, 256(12)\n\t"
+        "ld      21, 320(12)\n\t"
+        "ld      21, 384(12)\n\t"
+        "ld      21, 448(12)\n\t"
+        "ld      21, 512(12)\n\t"
+        "ld      21, 576(12)\n\t"
+        "ld      21, 640(12)\n\t"
+        "ld      21, 704(12)\n\t"
+        "ld      21, 768(12)\n\t"
+        "ld      21, 832(12)\n\t"
+        "ld      21, 896(12)\n\t"
+        "ld      21, 960(12)\n\t"
+        "rldicl  16, 18, 32, 56\n\t"
+        "andi.   14, 19, 255\n\t"
+        "rldicl  17, 19, 32, 56\n\t"
+        "andi.   15, 18, 255\n\t"
+        "lbzx    16, 12, 16\n\t"
+        "lbzx    14, 12, 14\n\t"
+        "lbzx    17, 12, 17\n\t"
+        "lbzx    15, 12, 15\n\t"
+        "rldimi  14, 16, 32, 24\n\t"
+        "rldimi  15, 17, 32, 24\n\t"
+        "rldicl  22, 19, 56, 56\n\t"
+        "rldicl  23, 19, 24, 56\n\t"
+        "rldicl  24, 18, 56, 56\n\t"
+        "rldicl  25, 18, 24, 56\n\t"
+        "lbzx    22, 12, 22\n\t"
+        "lbzx    23, 12, 23\n\t"
+        "lbzx    24, 12, 24\n\t"
+        "lbzx    25, 12, 25\n\t"
+        "rldimi  14, 22, 40, 16\n\t"
+        "rlwimi  14, 23, 8, 16, 23\n\t"
+        "rldimi  15, 24, 40, 16\n\t"
+        "rlwimi  15, 25, 8, 16, 23\n\t"
+        "rldicl  22, 19, 16, 56\n\t"
+        "rldicl  23, 18, 48, 56\n\t"
+        "rldicl  24, 18, 16, 56\n\t"
+        "rldicl  25, 19, 48, 56\n\t"
+        "lbzx    22, 12, 22\n\t"
+        "lbzx    23, 12, 23\n\t"
+        "lbzx    24, 12, 24\n\t"
+        "lbzx    25, 12, 25\n\t"
+        "rldimi  14, 22, 48, 8\n\t"
+        "rlwimi  14, 23, 16, 8, 15\n\t"
+        "rldimi  15, 24, 48, 8\n\t"
+        "rlwimi  15, 25, 16, 8, 15\n\t"
+        "rldicl  22, 18, 40, 56\n\t"
+        "rldicl  23, 18, 8, 56\n\t"
+        "rldicl  24, 19, 40, 56\n\t"
+        "rldicl  25, 19, 8, 56\n\t"
+        "lbzx    22, 12, 22\n\t"
+        "lbzx    23, 12, 23\n\t"
+        "lbzx    24, 12, 24\n\t"
+        "lbzx    25, 12, 25\n\t"
+        "ld      18, 0(%[key2])\n\t"
+        "ld      19, 8(%[key2])\n\t"
+        "rldimi  14, 22, 56, 0\n\t"
+        "rlwimi  14, 23, 24, 0, 7\n\t"
+        "rldimi  15, 24, 56, 0\n\t"
+        "rlwimi  15, 25, 24, 0, 7\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "xor     14, 14, 26\n\t"
+        "xor     15, 15, 27\n\t"
+        "li      17, 8\n\t"
+        "ldbrx   26, 0, %[tmp]\n\t"
+        "ldbrx   27, %[tmp], 17\n\t"
+        "std     14, 0(%[out])\n\t"
+        "std     15, 8(%[out])\n\t"
+        "sradi   22, 27, 63\n\t"
+        "srdi    23, 26, 63\n\t"
+        "sldi    27, 27, 1\n\t"
+        "andi.   22, 22, 0x87\n\t"
+        "sldi    26, 26, 1\n\t"
+        "xor     27, 27, 23\n\t"
+        "xor     26, 26, 22\n\t"
+        "stdbrx  26, 0, %[tmp]\n\t"
+        "stdbrx  27, %[tmp], 17\n\t"
+        "ld      26, 0(%[tmp])\n\t"
+        "ld      27, 8(%[tmp])\n\t"
+        "addi    %[sz], %[sz], -16\n\t"
+        "addi    %[in], %[in], 16\n\t"
+        "addi    %[out], %[out], 16\n\t"
+        "cmpdi   %[sz], 16\n\t"
+        "bge     L_AES_XTS_decrypt_loop_block_%=\n\t"
+        "cmpdi   %[sz], 0\n\t"
+        "beq     L_AES_XTS_decrypt_done_data_%=\n\t"
+        "\n"
+    "L_AES_XTS_decrypt_start_partail_%=: \n\t"
+        "li      17, 8\n\t"
+        "ldbrx   28, 0, %[tmp]\n\t"
+        "ldbrx   29, %[tmp], 17\n\t"
+        "sradi   22, 29, 63\n\t"
+        "srdi    23, 28, 63\n\t"
+        "sldi    29, 29, 1\n\t"
+        "andi.   22, 22, 0x87\n\t"
+        "sldi    28, 28, 1\n\t"
+        "xor     29, 29, 23\n\t"
+        "xor     28, 28, 22\n\t"
+        "stdbrx  28, 0, %[tmp]\n\t"
+        "stdbrx  29, %[tmp], 17\n\t"
+        "ld      28, 0(%[tmp])\n\t"
+        "ld      29, 8(%[tmp])\n\t"
+        "addi    %[key2], %[key], 0\n\t"
+        "ld      14, 0(%[in])\n\t"
+        "ld      15, 8(%[in])\n\t"
+        "addi    %[in], %[in], 16\n\t"
+        "ld      18, 0(%[key2])\n\t"
+        "ld      19, 8(%[key2])\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        "xor     14, 14, 28\n\t"
+        "xor     15, 15, 29\n\t"
+        "rldicr  14, 14, 32, 63\n\t"
+        "rldicr  15, 15, 32, 63\n\t"
+        /* Round: 0 - XOR in key schedule */
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "addi    0, %[nr], -2\n\t"
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+        "ld      21, 0(11)\n\t"
+        "ld      21, 64(11)\n\t"
+        "ld      21, 128(11)\n\t"
+        "ld      21, 192(11)\n\t"
+        "ld      21, 256(11)\n\t"
+        "ld      21, 320(11)\n\t"
+        "ld      21, 384(11)\n\t"
+        "ld      21, 448(11)\n\t"
+        "ld      21, 512(11)\n\t"
+        "ld      21, 576(11)\n\t"
+        "ld      21, 640(11)\n\t"
+        "ld      21, 704(11)\n\t"
+        "ld      21, 768(11)\n\t"
+        "ld      21, 832(11)\n\t"
+        "ld      21, 896(11)\n\t"
+        "ld      21, 960(11)\n\t"
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+        "srwi    0, 0, 1\n\t"
+        "mtctr   0\n\t"
+        "\n"
+    "L_AES_XTS_decrypt_loop_nr_partial_1_%=: \n\t"
+        "srdi    16, 14, 32\n\t"
+        "srdi    17, 15, 32\n\t"
+        "rlwinm  18, 17, 18, 22, 29\n\t"
+        "rlwinm  20, 14, 18, 22, 29\n\t"
+        "rlwinm  19, 16, 18, 22, 29\n\t"
+        "rlwinm  21, 15, 18, 22, 29\n\t"
+        "lwzx    18, 11, 18\n\t"
+        "lwzx    20, 11, 20\n\t"
+        "lwzx    19, 11, 19\n\t"
+        "lwzx    21, 11, 21\n\t"
+        "rlwinm  22, 14, 10, 22, 29\n\t"
+        "rlwinm  23, 16, 10, 22, 29\n\t"
+        "rlwinm  24, 15, 10, 22, 29\n\t"
+        "rlwinm  25, 17, 10, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 8, 0, 31\n\t"
+        "rlwimi  23, 23, 8, 0, 31\n\t"
+        "rlwimi  24, 24, 8, 0, 31\n\t"
+        "rlwimi  25, 25, 8, 0, 31\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 15, 26, 22, 29\n\t"
+        "rlwinm  23, 17, 26, 22, 29\n\t"
+        "rlwinm  24, 14, 26, 22, 29\n\t"
+        "rlwinm  25, 16, 26, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 24, 0, 31\n\t"
+        "rlwimi  23, 23, 24, 0, 31\n\t"
+        "rlwimi  24, 24, 24, 0, 31\n\t"
+        "rlwimi  25, 25, 24, 0, 31\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 16, 2, 22, 29\n\t"
+        "rlwinm  23, 15, 2, 22, 29\n\t"
+        "rlwinm  24, 17, 2, 22, 29\n\t"
+        "rlwinm  25, 14, 2, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 16, 0, 31\n\t"
+        "rlwimi  23, 23, 16, 0, 31\n\t"
+        "rlwimi  24, 24, 16, 0, 31\n\t"
+        "rlwimi  25, 25, 16, 0, 31\n\t"
+        "ld      14, 0(%[key2])\n\t"
+        "ld      15, 8(%[key2])\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        "rldimi  18, 20, 32, 0\n\t"
+        "rldimi  19, 21, 32, 0\n\t"
+        /*   XOR in Key Schedule */
+        "xor     18, 18, 14\n\t"
+        "xor     19, 19, 15\n\t"
+        "srdi    20, 18, 32\n\t"
+        "srdi    21, 19, 32\n\t"
+        "rlwinm  14, 21, 18, 22, 29\n\t"
+        "rlwinm  16, 18, 18, 22, 29\n\t"
+        "rlwinm  15, 20, 18, 22, 29\n\t"
+        "rlwinm  17, 19, 18, 22, 29\n\t"
+        "lwzx    14, 11, 14\n\t"
+        "lwzx    16, 11, 16\n\t"
+        "lwzx    15, 11, 15\n\t"
+        "lwzx    17, 11, 17\n\t"
+        "rlwinm  22, 18, 10, 22, 29\n\t"
+        "rlwinm  23, 20, 10, 22, 29\n\t"
+        "rlwinm  24, 19, 10, 22, 29\n\t"
+        "rlwinm  25, 21, 10, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 8, 0, 31\n\t"
+        "rlwimi  23, 23, 8, 0, 31\n\t"
+        "rlwimi  24, 24, 8, 0, 31\n\t"
+        "rlwimi  25, 25, 8, 0, 31\n\t"
+        "xor     14, 14, 22\n\t"
+        "xor     16, 16, 23\n\t"
+        "xor     15, 15, 24\n\t"
+        "xor     17, 17, 25\n\t"
+        "rlwinm  22, 19, 26, 22, 29\n\t"
+        "rlwinm  23, 21, 26, 22, 29\n\t"
+        "rlwinm  24, 18, 26, 22, 29\n\t"
+        "rlwinm  25, 20, 26, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 24, 0, 31\n\t"
+        "rlwimi  23, 23, 24, 0, 31\n\t"
+        "rlwimi  24, 24, 24, 0, 31\n\t"
+        "rlwimi  25, 25, 24, 0, 31\n\t"
+        "xor     14, 14, 22\n\t"
+        "xor     16, 16, 23\n\t"
+        "xor     15, 15, 24\n\t"
+        "xor     17, 17, 25\n\t"
+        "rlwinm  22, 20, 2, 22, 29\n\t"
+        "rlwinm  23, 19, 2, 22, 29\n\t"
+        "rlwinm  24, 21, 2, 22, 29\n\t"
+        "rlwinm  25, 18, 2, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 16, 0, 31\n\t"
+        "rlwimi  23, 23, 16, 0, 31\n\t"
+        "rlwimi  24, 24, 16, 0, 31\n\t"
+        "rlwimi  25, 25, 16, 0, 31\n\t"
+        "ld      18, 0(%[key2])\n\t"
+        "ld      19, 8(%[key2])\n\t"
+        "xor     14, 14, 22\n\t"
+        "xor     16, 16, 23\n\t"
+        "xor     15, 15, 24\n\t"
+        "xor     17, 17, 25\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        "rldimi  14, 16, 32, 0\n\t"
+        "rldimi  15, 17, 32, 0\n\t"
+        /*   XOR in Key Schedule */
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "bdnz    L_AES_XTS_decrypt_loop_nr_partial_1_%=\n\t"
+        "srdi    16, 14, 32\n\t"
+        "srdi    17, 15, 32\n\t"
+        "rlwinm  18, 17, 18, 22, 29\n\t"
+        "rlwinm  20, 14, 18, 22, 29\n\t"
+        "rlwinm  19, 16, 18, 22, 29\n\t"
+        "rlwinm  21, 15, 18, 22, 29\n\t"
+        "lwzx    18, 11, 18\n\t"
+        "lwzx    20, 11, 20\n\t"
+        "lwzx    19, 11, 19\n\t"
+        "lwzx    21, 11, 21\n\t"
+        "rlwinm  22, 14, 10, 22, 29\n\t"
+        "rlwinm  23, 16, 10, 22, 29\n\t"
+        "rlwinm  24, 15, 10, 22, 29\n\t"
+        "rlwinm  25, 17, 10, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 8, 0, 31\n\t"
+        "rlwimi  23, 23, 8, 0, 31\n\t"
+        "rlwimi  24, 24, 8, 0, 31\n\t"
+        "rlwimi  25, 25, 8, 0, 31\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 15, 26, 22, 29\n\t"
+        "rlwinm  23, 17, 26, 22, 29\n\t"
+        "rlwinm  24, 14, 26, 22, 29\n\t"
+        "rlwinm  25, 16, 26, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 24, 0, 31\n\t"
+        "rlwimi  23, 23, 24, 0, 31\n\t"
+        "rlwimi  24, 24, 24, 0, 31\n\t"
+        "rlwimi  25, 25, 24, 0, 31\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 16, 2, 22, 29\n\t"
+        "rlwinm  23, 15, 2, 22, 29\n\t"
+        "rlwinm  24, 17, 2, 22, 29\n\t"
+        "rlwinm  25, 14, 2, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 16, 0, 31\n\t"
+        "rlwimi  23, 23, 16, 0, 31\n\t"
+        "rlwimi  24, 24, 16, 0, 31\n\t"
+        "rlwimi  25, 25, 16, 0, 31\n\t"
+        "ld      14, 0(%[key2])\n\t"
+        "ld      15, 8(%[key2])\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        "rldimi  18, 20, 32, 0\n\t"
+        "rldimi  19, 21, 32, 0\n\t"
+        /*   XOR in Key Schedule */
+        "xor     18, 18, 14\n\t"
+        "xor     19, 19, 15\n\t"
+        "ld      21, 0(12)\n\t"
+        "ld      21, 64(12)\n\t"
+        "ld      21, 128(12)\n\t"
+        "ld      21, 192(12)\n\t"
+        "ld      21, 256(12)\n\t"
+        "ld      21, 320(12)\n\t"
+        "ld      21, 384(12)\n\t"
+        "ld      21, 448(12)\n\t"
+        "ld      21, 512(12)\n\t"
+        "ld      21, 576(12)\n\t"
+        "ld      21, 640(12)\n\t"
+        "ld      21, 704(12)\n\t"
+        "ld      21, 768(12)\n\t"
+        "ld      21, 832(12)\n\t"
+        "ld      21, 896(12)\n\t"
+        "ld      21, 960(12)\n\t"
+        "rldicl  16, 18, 32, 56\n\t"
+        "andi.   14, 19, 255\n\t"
+        "rldicl  17, 19, 32, 56\n\t"
+        "andi.   15, 18, 255\n\t"
+        "lbzx    16, 12, 16\n\t"
+        "lbzx    14, 12, 14\n\t"
+        "lbzx    17, 12, 17\n\t"
+        "lbzx    15, 12, 15\n\t"
+        "rldimi  14, 16, 32, 24\n\t"
+        "rldimi  15, 17, 32, 24\n\t"
+        "rldicl  22, 19, 56, 56\n\t"
+        "rldicl  23, 19, 24, 56\n\t"
+        "rldicl  24, 18, 56, 56\n\t"
+        "rldicl  25, 18, 24, 56\n\t"
+        "lbzx    22, 12, 22\n\t"
+        "lbzx    23, 12, 23\n\t"
+        "lbzx    24, 12, 24\n\t"
+        "lbzx    25, 12, 25\n\t"
+        "rldimi  14, 22, 40, 16\n\t"
+        "rlwimi  14, 23, 8, 16, 23\n\t"
+        "rldimi  15, 24, 40, 16\n\t"
+        "rlwimi  15, 25, 8, 16, 23\n\t"
+        "rldicl  22, 19, 16, 56\n\t"
+        "rldicl  23, 18, 48, 56\n\t"
+        "rldicl  24, 18, 16, 56\n\t"
+        "rldicl  25, 19, 48, 56\n\t"
+        "lbzx    22, 12, 22\n\t"
+        "lbzx    23, 12, 23\n\t"
+        "lbzx    24, 12, 24\n\t"
+        "lbzx    25, 12, 25\n\t"
+        "rldimi  14, 22, 48, 8\n\t"
+        "rlwimi  14, 23, 16, 8, 15\n\t"
+        "rldimi  15, 24, 48, 8\n\t"
+        "rlwimi  15, 25, 16, 8, 15\n\t"
+        "rldicl  22, 18, 40, 56\n\t"
+        "rldicl  23, 18, 8, 56\n\t"
+        "rldicl  24, 19, 40, 56\n\t"
+        "rldicl  25, 19, 8, 56\n\t"
+        "lbzx    22, 12, 22\n\t"
+        "lbzx    23, 12, 23\n\t"
+        "lbzx    24, 12, 24\n\t"
+        "lbzx    25, 12, 25\n\t"
+        "ld      18, 0(%[key2])\n\t"
+        "ld      19, 8(%[key2])\n\t"
+        "rldimi  14, 22, 56, 0\n\t"
+        "rlwimi  14, 23, 24, 0, 7\n\t"
+        "rldimi  15, 24, 56, 0\n\t"
+        "rlwimi  15, 25, 24, 0, 7\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "xor     14, 14, 28\n\t"
+        "xor     15, 15, 29\n\t"
+        "std     14, 0(%[tmp])\n\t"
+        "std     15, 8(%[tmp])\n\t"
+        "addi    %[out], %[out], 16\n\t"
+        "addi    18, %[sz], 0\n\t"
+        "\n"
+    "L_AES_XTS_decrypt_start_byte_%=: \n\t"
+        "lbz     22, 0(%[tmp])\n\t"
+        "lbz     23, 0(%[in])\n\t"
+        "addi    %[in], %[in], 1\n\t"
+        "stb     22, 0(%[out])\n\t"
+        "addi    %[out], %[out], 1\n\t"
+        "stb     23, 0(%[tmp])\n\t"
+        "addi    %[tmp], %[tmp], 1\n\t"
+        "addic.  18, 18, -1\n\t"
+        "bne     L_AES_XTS_decrypt_start_byte_%=\n\t"
+        "subf    %[out], %[sz], %[out]\n\t"
+        "subf    %[tmp], %[sz], %[tmp]\n\t"
+        "addi    %[out], %[out], -16\n\t"
+        "addi    %[key2], %[key], 0\n\t"
+        "ld      14, 0(%[tmp])\n\t"
+        "ld      15, 8(%[tmp])\n\t"
+        "ld      18, 0(%[key2])\n\t"
+        "ld      19, 8(%[key2])\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        "xor     14, 14, 26\n\t"
+        "xor     15, 15, 27\n\t"
+        "rldicr  14, 14, 32, 63\n\t"
+        "rldicr  15, 15, 32, 63\n\t"
+        /* Round: 0 - XOR in key schedule */
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "addi    0, %[nr], -2\n\t"
+#ifndef WOLFSSL_PPC64_ASM_AES_NO_HARDEN
+        "ld      21, 0(11)\n\t"
+        "ld      21, 64(11)\n\t"
+        "ld      21, 128(11)\n\t"
+        "ld      21, 192(11)\n\t"
+        "ld      21, 256(11)\n\t"
+        "ld      21, 320(11)\n\t"
+        "ld      21, 384(11)\n\t"
+        "ld      21, 448(11)\n\t"
+        "ld      21, 512(11)\n\t"
+        "ld      21, 576(11)\n\t"
+        "ld      21, 640(11)\n\t"
+        "ld      21, 704(11)\n\t"
+        "ld      21, 768(11)\n\t"
+        "ld      21, 832(11)\n\t"
+        "ld      21, 896(11)\n\t"
+        "ld      21, 960(11)\n\t"
+#endif /* !WOLFSSL_PPC64_ASM_AES_NO_HARDEN */
+        "srwi    0, 0, 1\n\t"
+        "mtctr   0\n\t"
+        "\n"
+    "L_AES_XTS_decrypt_loop_nr_partial_2_%=: \n\t"
+        "srdi    16, 14, 32\n\t"
+        "srdi    17, 15, 32\n\t"
+        "rlwinm  18, 17, 18, 22, 29\n\t"
+        "rlwinm  20, 14, 18, 22, 29\n\t"
+        "rlwinm  19, 16, 18, 22, 29\n\t"
+        "rlwinm  21, 15, 18, 22, 29\n\t"
+        "lwzx    18, 11, 18\n\t"
+        "lwzx    20, 11, 20\n\t"
+        "lwzx    19, 11, 19\n\t"
+        "lwzx    21, 11, 21\n\t"
+        "rlwinm  22, 14, 10, 22, 29\n\t"
+        "rlwinm  23, 16, 10, 22, 29\n\t"
+        "rlwinm  24, 15, 10, 22, 29\n\t"
+        "rlwinm  25, 17, 10, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 8, 0, 31\n\t"
+        "rlwimi  23, 23, 8, 0, 31\n\t"
+        "rlwimi  24, 24, 8, 0, 31\n\t"
+        "rlwimi  25, 25, 8, 0, 31\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 15, 26, 22, 29\n\t"
+        "rlwinm  23, 17, 26, 22, 29\n\t"
+        "rlwinm  24, 14, 26, 22, 29\n\t"
+        "rlwinm  25, 16, 26, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 24, 0, 31\n\t"
+        "rlwimi  23, 23, 24, 0, 31\n\t"
+        "rlwimi  24, 24, 24, 0, 31\n\t"
+        "rlwimi  25, 25, 24, 0, 31\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 16, 2, 22, 29\n\t"
+        "rlwinm  23, 15, 2, 22, 29\n\t"
+        "rlwinm  24, 17, 2, 22, 29\n\t"
+        "rlwinm  25, 14, 2, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 16, 0, 31\n\t"
+        "rlwimi  23, 23, 16, 0, 31\n\t"
+        "rlwimi  24, 24, 16, 0, 31\n\t"
+        "rlwimi  25, 25, 16, 0, 31\n\t"
+        "ld      14, 0(%[key2])\n\t"
+        "ld      15, 8(%[key2])\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        "rldimi  18, 20, 32, 0\n\t"
+        "rldimi  19, 21, 32, 0\n\t"
+        /*   XOR in Key Schedule */
+        "xor     18, 18, 14\n\t"
+        "xor     19, 19, 15\n\t"
+        "srdi    20, 18, 32\n\t"
+        "srdi    21, 19, 32\n\t"
+        "rlwinm  14, 21, 18, 22, 29\n\t"
+        "rlwinm  16, 18, 18, 22, 29\n\t"
+        "rlwinm  15, 20, 18, 22, 29\n\t"
+        "rlwinm  17, 19, 18, 22, 29\n\t"
+        "lwzx    14, 11, 14\n\t"
+        "lwzx    16, 11, 16\n\t"
+        "lwzx    15, 11, 15\n\t"
+        "lwzx    17, 11, 17\n\t"
+        "rlwinm  22, 18, 10, 22, 29\n\t"
+        "rlwinm  23, 20, 10, 22, 29\n\t"
+        "rlwinm  24, 19, 10, 22, 29\n\t"
+        "rlwinm  25, 21, 10, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 8, 0, 31\n\t"
+        "rlwimi  23, 23, 8, 0, 31\n\t"
+        "rlwimi  24, 24, 8, 0, 31\n\t"
+        "rlwimi  25, 25, 8, 0, 31\n\t"
+        "xor     14, 14, 22\n\t"
+        "xor     16, 16, 23\n\t"
+        "xor     15, 15, 24\n\t"
+        "xor     17, 17, 25\n\t"
+        "rlwinm  22, 19, 26, 22, 29\n\t"
+        "rlwinm  23, 21, 26, 22, 29\n\t"
+        "rlwinm  24, 18, 26, 22, 29\n\t"
+        "rlwinm  25, 20, 26, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 24, 0, 31\n\t"
+        "rlwimi  23, 23, 24, 0, 31\n\t"
+        "rlwimi  24, 24, 24, 0, 31\n\t"
+        "rlwimi  25, 25, 24, 0, 31\n\t"
+        "xor     14, 14, 22\n\t"
+        "xor     16, 16, 23\n\t"
+        "xor     15, 15, 24\n\t"
+        "xor     17, 17, 25\n\t"
+        "rlwinm  22, 20, 2, 22, 29\n\t"
+        "rlwinm  23, 19, 2, 22, 29\n\t"
+        "rlwinm  24, 21, 2, 22, 29\n\t"
+        "rlwinm  25, 18, 2, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 16, 0, 31\n\t"
+        "rlwimi  23, 23, 16, 0, 31\n\t"
+        "rlwimi  24, 24, 16, 0, 31\n\t"
+        "rlwimi  25, 25, 16, 0, 31\n\t"
+        "ld      18, 0(%[key2])\n\t"
+        "ld      19, 8(%[key2])\n\t"
+        "xor     14, 14, 22\n\t"
+        "xor     16, 16, 23\n\t"
+        "xor     15, 15, 24\n\t"
+        "xor     17, 17, 25\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        "rldimi  14, 16, 32, 0\n\t"
+        "rldimi  15, 17, 32, 0\n\t"
+        /*   XOR in Key Schedule */
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "bdnz    L_AES_XTS_decrypt_loop_nr_partial_2_%=\n\t"
+        "srdi    16, 14, 32\n\t"
+        "srdi    17, 15, 32\n\t"
+        "rlwinm  18, 17, 18, 22, 29\n\t"
+        "rlwinm  20, 14, 18, 22, 29\n\t"
+        "rlwinm  19, 16, 18, 22, 29\n\t"
+        "rlwinm  21, 15, 18, 22, 29\n\t"
+        "lwzx    18, 11, 18\n\t"
+        "lwzx    20, 11, 20\n\t"
+        "lwzx    19, 11, 19\n\t"
+        "lwzx    21, 11, 21\n\t"
+        "rlwinm  22, 14, 10, 22, 29\n\t"
+        "rlwinm  23, 16, 10, 22, 29\n\t"
+        "rlwinm  24, 15, 10, 22, 29\n\t"
+        "rlwinm  25, 17, 10, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 8, 0, 31\n\t"
+        "rlwimi  23, 23, 8, 0, 31\n\t"
+        "rlwimi  24, 24, 8, 0, 31\n\t"
+        "rlwimi  25, 25, 8, 0, 31\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 15, 26, 22, 29\n\t"
+        "rlwinm  23, 17, 26, 22, 29\n\t"
+        "rlwinm  24, 14, 26, 22, 29\n\t"
+        "rlwinm  25, 16, 26, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 24, 0, 31\n\t"
+        "rlwimi  23, 23, 24, 0, 31\n\t"
+        "rlwimi  24, 24, 24, 0, 31\n\t"
+        "rlwimi  25, 25, 24, 0, 31\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "rlwinm  22, 16, 2, 22, 29\n\t"
+        "rlwinm  23, 15, 2, 22, 29\n\t"
+        "rlwinm  24, 17, 2, 22, 29\n\t"
+        "rlwinm  25, 14, 2, 22, 29\n\t"
+        "lwzx    22, 11, 22\n\t"
+        "lwzx    23, 11, 23\n\t"
+        "lwzx    24, 11, 24\n\t"
+        "lwzx    25, 11, 25\n\t"
+        "rlwimi  22, 22, 16, 0, 31\n\t"
+        "rlwimi  23, 23, 16, 0, 31\n\t"
+        "rlwimi  24, 24, 16, 0, 31\n\t"
+        "rlwimi  25, 25, 16, 0, 31\n\t"
+        "ld      14, 0(%[key2])\n\t"
+        "ld      15, 8(%[key2])\n\t"
+        "xor     18, 18, 22\n\t"
+        "xor     20, 20, 23\n\t"
+        "xor     19, 19, 24\n\t"
+        "xor     21, 21, 25\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        "rldimi  18, 20, 32, 0\n\t"
+        "rldimi  19, 21, 32, 0\n\t"
+        /*   XOR in Key Schedule */
+        "xor     18, 18, 14\n\t"
+        "xor     19, 19, 15\n\t"
+        "ld      21, 0(12)\n\t"
+        "ld      21, 64(12)\n\t"
+        "ld      21, 128(12)\n\t"
+        "ld      21, 192(12)\n\t"
+        "ld      21, 256(12)\n\t"
+        "ld      21, 320(12)\n\t"
+        "ld      21, 384(12)\n\t"
+        "ld      21, 448(12)\n\t"
+        "ld      21, 512(12)\n\t"
+        "ld      21, 576(12)\n\t"
+        "ld      21, 640(12)\n\t"
+        "ld      21, 704(12)\n\t"
+        "ld      21, 768(12)\n\t"
+        "ld      21, 832(12)\n\t"
+        "ld      21, 896(12)\n\t"
+        "ld      21, 960(12)\n\t"
+        "rldicl  16, 18, 32, 56\n\t"
+        "andi.   14, 19, 255\n\t"
+        "rldicl  17, 19, 32, 56\n\t"
+        "andi.   15, 18, 255\n\t"
+        "lbzx    16, 12, 16\n\t"
+        "lbzx    14, 12, 14\n\t"
+        "lbzx    17, 12, 17\n\t"
+        "lbzx    15, 12, 15\n\t"
+        "rldimi  14, 16, 32, 24\n\t"
+        "rldimi  15, 17, 32, 24\n\t"
+        "rldicl  22, 19, 56, 56\n\t"
+        "rldicl  23, 19, 24, 56\n\t"
+        "rldicl  24, 18, 56, 56\n\t"
+        "rldicl  25, 18, 24, 56\n\t"
+        "lbzx    22, 12, 22\n\t"
+        "lbzx    23, 12, 23\n\t"
+        "lbzx    24, 12, 24\n\t"
+        "lbzx    25, 12, 25\n\t"
+        "rldimi  14, 22, 40, 16\n\t"
+        "rlwimi  14, 23, 8, 16, 23\n\t"
+        "rldimi  15, 24, 40, 16\n\t"
+        "rlwimi  15, 25, 8, 16, 23\n\t"
+        "rldicl  22, 19, 16, 56\n\t"
+        "rldicl  23, 18, 48, 56\n\t"
+        "rldicl  24, 18, 16, 56\n\t"
+        "rldicl  25, 19, 48, 56\n\t"
+        "lbzx    22, 12, 22\n\t"
+        "lbzx    23, 12, 23\n\t"
+        "lbzx    24, 12, 24\n\t"
+        "lbzx    25, 12, 25\n\t"
+        "rldimi  14, 22, 48, 8\n\t"
+        "rlwimi  14, 23, 16, 8, 15\n\t"
+        "rldimi  15, 24, 48, 8\n\t"
+        "rlwimi  15, 25, 16, 8, 15\n\t"
+        "rldicl  22, 18, 40, 56\n\t"
+        "rldicl  23, 18, 8, 56\n\t"
+        "rldicl  24, 19, 40, 56\n\t"
+        "rldicl  25, 19, 8, 56\n\t"
+        "lbzx    22, 12, 22\n\t"
+        "lbzx    23, 12, 23\n\t"
+        "lbzx    24, 12, 24\n\t"
+        "lbzx    25, 12, 25\n\t"
+        "ld      18, 0(%[key2])\n\t"
+        "ld      19, 8(%[key2])\n\t"
+        "rldimi  14, 22, 56, 0\n\t"
+        "rlwimi  14, 23, 24, 0, 7\n\t"
+        "rldimi  15, 24, 56, 0\n\t"
+        "rlwimi  15, 25, 24, 0, 7\n\t"
+        "addi    %[key2], %[key2], 16\n\t"
+        /*   XOR in Key Schedule */
+        "xor     14, 14, 18\n\t"
+        "xor     15, 15, 19\n\t"
+        "xor     14, 14, 26\n\t"
+        "xor     15, 15, 27\n\t"
+        "std     14, 0(%[out])\n\t"
+        "std     15, 8(%[out])\n\t"
+        "\n"
+    "L_AES_XTS_decrypt_done_data_%=: \n\t"
+        "addi    1, 1, 8\n\t"
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+        : [in] "+r" (in), [out] "+r" (out), [sz] "+r" (sz), [i] "+r" (i),
+          [key] "+r" (key), [key2] "+r" (key2), [tmp] "+r" (tmp),
+          [nr] "+r" (nr), [L_AES_PPC64_td] "+r" (L_AES_PPC64_td_c),
+          [L_AES_PPC64_td4] "+r" (L_AES_PPC64_td4_c),
+          [L_AES_PPC64_te4_0] "+r" (L_AES_PPC64_te4_0_c)
+        :
+        : "memory", "cc", "0", "15", "16", "17", "18", "19", "20", "21", "22",
+            "23", "24", "25", "26", "27", "28", "29"
+#else
+        :
+        : [in] "r" (in), [out] "r" (out), [sz] "r" (sz), [i] "r" (i),
+          [key] "r" (key), [key2] "r" (key2), [tmp] "r" (tmp), [nr] "r" (nr),
+          [L_AES_PPC64_td] "r" (L_AES_PPC64_td_c),
+          [L_AES_PPC64_td4] "r" (L_AES_PPC64_td4_c),
+          [L_AES_PPC64_te4_0] "r" (L_AES_PPC64_te4_0_c)
+        : "memory", "cc", "0", "11", "12", "14", "15", "16", "17", "18", "19",
+            "20", "21", "22", "23", "24", "25", "26", "27", "28", "29"
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+    );
+}
+
+#endif /* WOLFSSL_AES_XTS */
+#endif /* HAVE_AES_DECRYPT */
+#ifdef HAVE_AESGCM
+#ifdef GCM_TABLE_4BIT
+static const word32 L_GCM_gmult_len_r[] = {
+    0x00000000, 0x1c200000, 0x38400000, 0x24600000,
+    0x70800000, 0x6ca00000, 0x48c00000, 0x54e00000,
+    0xe1000000, 0xfd200000, 0xd9400000, 0xc5600000,
+    0x91800000, 0x8da00000, 0xa9c00000, 0xb5e00000,
+    0x00000000, 0x01c20000, 0x03840000, 0x02460000,
+    0x07080000, 0x06ca0000, 0x048c0000, 0x054e0000,
+    0x0e100000, 0x0fd20000, 0x0d940000, 0x0c560000,
+    0x09180000, 0x08da0000, 0x0a9c0000, 0x0b5e0000,
+};
+
+void GCM_gmult_len(unsigned char* x, const unsigned char** m,
+    const unsigned char* data, unsigned long len);
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+void GCM_gmult_len(unsigned char* x_p, const unsigned char** m_p,
+    const unsigned char* data_p, unsigned long len_p)
+#else
+void GCM_gmult_len(unsigned char* x, const unsigned char** m,
+    const unsigned char* data, unsigned long len)
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+{
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+    register unsigned char* x asm ("3") = (unsigned char*)x_p;
+    register const unsigned char** m asm ("4") = (const unsigned char**)m_p;
+    register const unsigned char* data asm ("5") =
+        (const unsigned char*)data_p;
+    register unsigned long len asm ("6") = (unsigned long)len_p;
+    register word32* L_GCM_gmult_len_r_c asm ("7") =
+        (word32*)&L_GCM_gmult_len_r;
+#else
+    register word32* L_GCM_gmult_len_r_c = (word32*)&L_GCM_gmult_len_r;
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+
+    __asm__ __volatile__ (
+        "li      21, 0x100\n\t"
+        "li      25, 8\n\t"
+#ifdef WOLFSSL_NO_VAR_ASSIGN_REG
+        "mr      7, %[L_GCM_gmult_len_r]\n\t"
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+        "add     21, 21, %[m]\n\t"
+        "addi    20, %[m], 8\n\t"
+        "addi    22, 21, 8\n\t"
+        "addi    8, 7, 0x40\n\t"
+        "\n"
+    "L_GCM_gmult_len_start_block_%=: \n\t"
+        "ldbrx   18, 0, %[x]\n\t"
+        "ldbrx   19, 25, %[x]\n\t"
+        "ldbrx   9, 0, %[data]\n\t"
+        "ldbrx   10, 25, %[data]\n\t"
+        "xor     18, 18, 9\n\t"
+        "xor     19, 19, 10\n\t"
+        "rldicr  0, 19, 32, 63\n\t"
+        /* Byte 15 */
+        "rlwinm  23, 0, 12, 24, 27\n\t"
+        "rlwinm  24, 0, 8, 24, 27\n\t"
+        "ldx     14, 23, %[m]\n\t"
+        "ldx     15, 23, 20\n\t"
+        "ldx     11, 24, 21\n\t"
+        "ldx     12, 24, 22\n\t"
+        "ldx     10, 24, 20\n\t"
+        "rldicl  16, 15, 60, 60\n\t"
+        "rldic   17, 15, 2, 58\n\t"
+        "srdi    15, 15, 8\n\t"
+        "rldimi  15, 14, 56, 0\n\t"
+        "srdi    14, 14, 8\n\t"
+        "lwzx    9, 17, 8\n\t"
+        "xor     16, 16, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 16, 2, 26, 29\n\t"
+        "lwzx    10, 16, 7\n\t"
+        "xor     15, 15, 12\n\t"
+        "sldi    9, 9, 32\n\t"
+        "xor     14, 14, 9\n\t"
+        "sldi    10, 10, 32\n\t"
+        "xor     14, 14, 10\n\t"
+        /* Byte 14 */
+        "rlwinm  23, 0, 20, 24, 27\n\t"
+        "rlwinm  24, 0, 16, 24, 27\n\t"
+        "ldx     9, 23, %[m]\n\t"
+        "ldx     10, 23, 20\n\t"
+        "ldx     11, 24, 21\n\t"
+        "ldx     12, 24, 22\n\t"
+        "xor     14, 14, 9\n\t"
+        "xor     15, 15, 10\n\t"
+        "ldx     10, 24, 20\n\t"
+        "rldicl  16, 15, 60, 60\n\t"
+        "rldic   17, 15, 2, 58\n\t"
+        "srdi    15, 15, 8\n\t"
+        "rldimi  15, 14, 56, 0\n\t"
+        "srdi    14, 14, 8\n\t"
+        "lwzx    9, 17, 8\n\t"
+        "xor     16, 16, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 16, 2, 26, 29\n\t"
+        "lwzx    10, 16, 7\n\t"
+        "xor     15, 15, 12\n\t"
+        "sldi    9, 9, 32\n\t"
+        "xor     14, 14, 9\n\t"
+        "sldi    10, 10, 32\n\t"
+        "xor     14, 14, 10\n\t"
+        /* Byte 13 */
+        "rlwinm  23, 0, 28, 24, 27\n\t"
+        "rlwinm  24, 0, 24, 24, 27\n\t"
+        "ldx     9, 23, %[m]\n\t"
+        "ldx     10, 23, 20\n\t"
+        "ldx     11, 24, 21\n\t"
+        "ldx     12, 24, 22\n\t"
+        "xor     14, 14, 9\n\t"
+        "xor     15, 15, 10\n\t"
+        "ldx     10, 24, 20\n\t"
+        "rldicl  16, 15, 60, 60\n\t"
+        "rldic   17, 15, 2, 58\n\t"
+        "srdi    15, 15, 8\n\t"
+        "rldimi  15, 14, 56, 0\n\t"
+        "srdi    14, 14, 8\n\t"
+        "lwzx    9, 17, 8\n\t"
+        "xor     16, 16, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 16, 2, 26, 29\n\t"
+        "lwzx    10, 16, 7\n\t"
+        "xor     15, 15, 12\n\t"
+        "sldi    9, 9, 32\n\t"
+        "xor     14, 14, 9\n\t"
+        "sldi    10, 10, 32\n\t"
+        "xor     14, 14, 10\n\t"
+        /* Byte 12 */
+        "rlwinm  23, 0, 4, 24, 27\n\t"
+        "rlwinm  24, 0, 0, 24, 27\n\t"
+        "ldx     9, 23, %[m]\n\t"
+        "ldx     10, 23, 20\n\t"
+        "ldx     11, 24, 21\n\t"
+        "ldx     12, 24, 22\n\t"
+        "xor     14, 14, 9\n\t"
+        "xor     15, 15, 10\n\t"
+        "ldx     10, 24, 20\n\t"
+        "rldicl  16, 15, 60, 60\n\t"
+        "rldic   17, 15, 2, 58\n\t"
+        "srdi    15, 15, 8\n\t"
+        "rldimi  15, 14, 56, 0\n\t"
+        "srdi    14, 14, 8\n\t"
+        "lwzx    9, 17, 8\n\t"
+        "xor     16, 16, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 16, 2, 26, 29\n\t"
+        "lwzx    10, 16, 7\n\t"
+        "xor     15, 15, 12\n\t"
+        "sldi    9, 9, 32\n\t"
+        "xor     14, 14, 9\n\t"
+        "sldi    10, 10, 32\n\t"
+        "xor     14, 14, 10\n\t"
+        /* Byte 11 */
+        "rlwinm  23, 19, 12, 24, 27\n\t"
+        "rlwinm  24, 19, 8, 24, 27\n\t"
+        "ldx     9, 23, %[m]\n\t"
+        "ldx     10, 23, 20\n\t"
+        "ldx     11, 24, 21\n\t"
+        "ldx     12, 24, 22\n\t"
+        "xor     14, 14, 9\n\t"
+        "xor     15, 15, 10\n\t"
+        "ldx     10, 24, 20\n\t"
+        "rldicl  16, 15, 60, 60\n\t"
+        "rldic   17, 15, 2, 58\n\t"
+        "srdi    15, 15, 8\n\t"
+        "rldimi  15, 14, 56, 0\n\t"
+        "srdi    14, 14, 8\n\t"
+        "lwzx    9, 17, 8\n\t"
+        "xor     16, 16, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 16, 2, 26, 29\n\t"
+        "lwzx    10, 16, 7\n\t"
+        "xor     15, 15, 12\n\t"
+        "sldi    9, 9, 32\n\t"
+        "xor     14, 14, 9\n\t"
+        "sldi    10, 10, 32\n\t"
+        "xor     14, 14, 10\n\t"
+        /* Byte 10 */
+        "rlwinm  23, 19, 20, 24, 27\n\t"
+        "rlwinm  24, 19, 16, 24, 27\n\t"
+        "ldx     9, 23, %[m]\n\t"
+        "ldx     10, 23, 20\n\t"
+        "ldx     11, 24, 21\n\t"
+        "ldx     12, 24, 22\n\t"
+        "xor     14, 14, 9\n\t"
+        "xor     15, 15, 10\n\t"
+        "ldx     10, 24, 20\n\t"
+        "rldicl  16, 15, 60, 60\n\t"
+        "rldic   17, 15, 2, 58\n\t"
+        "srdi    15, 15, 8\n\t"
+        "rldimi  15, 14, 56, 0\n\t"
+        "srdi    14, 14, 8\n\t"
+        "lwzx    9, 17, 8\n\t"
+        "xor     16, 16, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 16, 2, 26, 29\n\t"
+        "lwzx    10, 16, 7\n\t"
+        "xor     15, 15, 12\n\t"
+        "sldi    9, 9, 32\n\t"
+        "xor     14, 14, 9\n\t"
+        "sldi    10, 10, 32\n\t"
+        "xor     14, 14, 10\n\t"
+        /* Byte 9 */
+        "rlwinm  23, 19, 28, 24, 27\n\t"
+        "rlwinm  24, 19, 24, 24, 27\n\t"
+        "ldx     9, 23, %[m]\n\t"
+        "ldx     10, 23, 20\n\t"
+        "ldx     11, 24, 21\n\t"
+        "ldx     12, 24, 22\n\t"
+        "xor     14, 14, 9\n\t"
+        "xor     15, 15, 10\n\t"
+        "ldx     10, 24, 20\n\t"
+        "rldicl  16, 15, 60, 60\n\t"
+        "rldic   17, 15, 2, 58\n\t"
+        "srdi    15, 15, 8\n\t"
+        "rldimi  15, 14, 56, 0\n\t"
+        "srdi    14, 14, 8\n\t"
+        "lwzx    9, 17, 8\n\t"
+        "xor     16, 16, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 16, 2, 26, 29\n\t"
+        "lwzx    10, 16, 7\n\t"
+        "xor     15, 15, 12\n\t"
+        "sldi    9, 9, 32\n\t"
+        "xor     14, 14, 9\n\t"
+        "sldi    10, 10, 32\n\t"
+        "xor     14, 14, 10\n\t"
+        /* Byte 8 */
+        "rlwinm  23, 19, 4, 24, 27\n\t"
+        "rlwinm  24, 19, 0, 24, 27\n\t"
+        "ldx     9, 23, %[m]\n\t"
+        "ldx     10, 23, 20\n\t"
+        "ldx     11, 24, 21\n\t"
+        "ldx     12, 24, 22\n\t"
+        "xor     14, 14, 9\n\t"
+        "xor     15, 15, 10\n\t"
+        "ldx     10, 24, 20\n\t"
+        "rldicl  16, 15, 60, 60\n\t"
+        "rldic   17, 15, 2, 58\n\t"
+        "srdi    15, 15, 8\n\t"
+        "rldimi  15, 14, 56, 0\n\t"
+        "srdi    14, 14, 8\n\t"
+        "lwzx    9, 17, 8\n\t"
+        "xor     16, 16, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 16, 2, 26, 29\n\t"
+        "lwzx    10, 16, 7\n\t"
+        "xor     15, 15, 12\n\t"
+        "sldi    9, 9, 32\n\t"
+        "xor     14, 14, 9\n\t"
+        "sldi    10, 10, 32\n\t"
+        "xor     14, 14, 10\n\t"
+        "rldicr  0, 18, 32, 63\n\t"
+        /* Byte 7 */
+        "rlwinm  23, 0, 12, 24, 27\n\t"
+        "rlwinm  24, 0, 8, 24, 27\n\t"
+        "ldx     9, 23, %[m]\n\t"
+        "ldx     10, 23, 20\n\t"
+        "ldx     11, 24, 21\n\t"
+        "ldx     12, 24, 22\n\t"
+        "xor     14, 14, 9\n\t"
+        "xor     15, 15, 10\n\t"
+        "ldx     10, 24, 20\n\t"
+        "rldicl  16, 15, 60, 60\n\t"
+        "rldic   17, 15, 2, 58\n\t"
+        "srdi    15, 15, 8\n\t"
+        "rldimi  15, 14, 56, 0\n\t"
+        "srdi    14, 14, 8\n\t"
+        "lwzx    9, 17, 8\n\t"
+        "xor     16, 16, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 16, 2, 26, 29\n\t"
+        "lwzx    10, 16, 7\n\t"
+        "xor     15, 15, 12\n\t"
+        "sldi    9, 9, 32\n\t"
+        "xor     14, 14, 9\n\t"
+        "sldi    10, 10, 32\n\t"
+        "xor     14, 14, 10\n\t"
+        /* Byte 6 */
+        "rlwinm  23, 0, 20, 24, 27\n\t"
+        "rlwinm  24, 0, 16, 24, 27\n\t"
+        "ldx     9, 23, %[m]\n\t"
+        "ldx     10, 23, 20\n\t"
+        "ldx     11, 24, 21\n\t"
+        "ldx     12, 24, 22\n\t"
+        "xor     14, 14, 9\n\t"
+        "xor     15, 15, 10\n\t"
+        "ldx     10, 24, 20\n\t"
+        "rldicl  16, 15, 60, 60\n\t"
+        "rldic   17, 15, 2, 58\n\t"
+        "srdi    15, 15, 8\n\t"
+        "rldimi  15, 14, 56, 0\n\t"
+        "srdi    14, 14, 8\n\t"
+        "lwzx    9, 17, 8\n\t"
+        "xor     16, 16, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 16, 2, 26, 29\n\t"
+        "lwzx    10, 16, 7\n\t"
+        "xor     15, 15, 12\n\t"
+        "sldi    9, 9, 32\n\t"
+        "xor     14, 14, 9\n\t"
+        "sldi    10, 10, 32\n\t"
+        "xor     14, 14, 10\n\t"
+        /* Byte 5 */
+        "rlwinm  23, 0, 28, 24, 27\n\t"
+        "rlwinm  24, 0, 24, 24, 27\n\t"
+        "ldx     9, 23, %[m]\n\t"
+        "ldx     10, 23, 20\n\t"
+        "ldx     11, 24, 21\n\t"
+        "ldx     12, 24, 22\n\t"
+        "xor     14, 14, 9\n\t"
+        "xor     15, 15, 10\n\t"
+        "ldx     10, 24, 20\n\t"
+        "rldicl  16, 15, 60, 60\n\t"
+        "rldic   17, 15, 2, 58\n\t"
+        "srdi    15, 15, 8\n\t"
+        "rldimi  15, 14, 56, 0\n\t"
+        "srdi    14, 14, 8\n\t"
+        "lwzx    9, 17, 8\n\t"
+        "xor     16, 16, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 16, 2, 26, 29\n\t"
+        "lwzx    10, 16, 7\n\t"
+        "xor     15, 15, 12\n\t"
+        "sldi    9, 9, 32\n\t"
+        "xor     14, 14, 9\n\t"
+        "sldi    10, 10, 32\n\t"
+        "xor     14, 14, 10\n\t"
+        /* Byte 4 */
+        "rlwinm  23, 0, 4, 24, 27\n\t"
+        "rlwinm  24, 0, 0, 24, 27\n\t"
+        "ldx     9, 23, %[m]\n\t"
+        "ldx     10, 23, 20\n\t"
+        "ldx     11, 24, 21\n\t"
+        "ldx     12, 24, 22\n\t"
+        "xor     14, 14, 9\n\t"
+        "xor     15, 15, 10\n\t"
+        "ldx     10, 24, 20\n\t"
+        "rldicl  16, 15, 60, 60\n\t"
+        "rldic   17, 15, 2, 58\n\t"
+        "srdi    15, 15, 8\n\t"
+        "rldimi  15, 14, 56, 0\n\t"
+        "srdi    14, 14, 8\n\t"
+        "lwzx    9, 17, 8\n\t"
+        "xor     16, 16, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 16, 2, 26, 29\n\t"
+        "lwzx    10, 16, 7\n\t"
+        "xor     15, 15, 12\n\t"
+        "sldi    9, 9, 32\n\t"
+        "xor     14, 14, 9\n\t"
+        "sldi    10, 10, 32\n\t"
+        "xor     14, 14, 10\n\t"
+        /* Byte 3 */
+        "rlwinm  23, 18, 12, 24, 27\n\t"
+        "rlwinm  24, 18, 8, 24, 27\n\t"
+        "ldx     9, 23, %[m]\n\t"
+        "ldx     10, 23, 20\n\t"
+        "ldx     11, 24, 21\n\t"
+        "ldx     12, 24, 22\n\t"
+        "xor     14, 14, 9\n\t"
+        "xor     15, 15, 10\n\t"
+        "ldx     10, 24, 20\n\t"
+        "rldicl  16, 15, 60, 60\n\t"
+        "rldic   17, 15, 2, 58\n\t"
+        "srdi    15, 15, 8\n\t"
+        "rldimi  15, 14, 56, 0\n\t"
+        "srdi    14, 14, 8\n\t"
+        "lwzx    9, 17, 8\n\t"
+        "xor     16, 16, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 16, 2, 26, 29\n\t"
+        "lwzx    10, 16, 7\n\t"
+        "xor     15, 15, 12\n\t"
+        "sldi    9, 9, 32\n\t"
+        "xor     14, 14, 9\n\t"
+        "sldi    10, 10, 32\n\t"
+        "xor     14, 14, 10\n\t"
+        /* Byte 2 */
+        "rlwinm  23, 18, 20, 24, 27\n\t"
+        "rlwinm  24, 18, 16, 24, 27\n\t"
+        "ldx     9, 23, %[m]\n\t"
+        "ldx     10, 23, 20\n\t"
+        "ldx     11, 24, 21\n\t"
+        "ldx     12, 24, 22\n\t"
+        "xor     14, 14, 9\n\t"
+        "xor     15, 15, 10\n\t"
+        "ldx     10, 24, 20\n\t"
+        "rldicl  16, 15, 60, 60\n\t"
+        "rldic   17, 15, 2, 58\n\t"
+        "srdi    15, 15, 8\n\t"
+        "rldimi  15, 14, 56, 0\n\t"
+        "srdi    14, 14, 8\n\t"
+        "lwzx    9, 17, 8\n\t"
+        "xor     16, 16, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 16, 2, 26, 29\n\t"
+        "lwzx    10, 16, 7\n\t"
+        "xor     15, 15, 12\n\t"
+        "sldi    9, 9, 32\n\t"
+        "xor     14, 14, 9\n\t"
+        "sldi    10, 10, 32\n\t"
+        "xor     14, 14, 10\n\t"
+        /* Byte 1 */
+        "rlwinm  23, 18, 28, 24, 27\n\t"
+        "rlwinm  24, 18, 24, 24, 27\n\t"
+        "ldx     9, 23, %[m]\n\t"
+        "ldx     10, 23, 20\n\t"
+        "ldx     11, 24, 21\n\t"
+        "ldx     12, 24, 22\n\t"
+        "xor     14, 14, 9\n\t"
+        "xor     15, 15, 10\n\t"
+        "ldx     10, 24, 20\n\t"
+        "rldicl  16, 15, 60, 60\n\t"
+        "rldic   17, 15, 2, 58\n\t"
+        "srdi    15, 15, 8\n\t"
+        "rldimi  15, 14, 56, 0\n\t"
+        "srdi    14, 14, 8\n\t"
+        "lwzx    9, 17, 8\n\t"
+        "xor     16, 16, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 16, 2, 26, 29\n\t"
+        "lwzx    10, 16, 7\n\t"
+        "xor     15, 15, 12\n\t"
+        "sldi    9, 9, 32\n\t"
+        "xor     14, 14, 9\n\t"
+        "sldi    10, 10, 32\n\t"
+        "xor     14, 14, 10\n\t"
+        /* Byte 0 */
+        "rlwinm  23, 18, 4, 24, 27\n\t"
+        "rlwinm  24, 18, 0, 24, 27\n\t"
+        "ldx     9, 23, %[m]\n\t"
+        "ldx     10, 23, 20\n\t"
+        "ldx     11, 24, %[m]\n\t"
+        "ldx     12, 24, 20\n\t"
+        "xor     14, 14, 9\n\t"
+        "xor     15, 15, 10\n\t"
+        "rldic   16, 15, 2, 58\n\t"
+        "srdi    15, 15, 4\n\t"
+        "lwzx    9, 16, 7\n\t"
+        "rldimi  15, 14, 60, 0\n\t"
+        "srdi    14, 14, 4\n\t"
+        "xor     19, 15, 12\n\t"
+        "xor     14, 14, 11\n\t"
+        "sldi    9, 9, 32\n\t"
+        "xor     18, 14, 9\n\t"
+        "addi    %[data], %[data], 16\n\t"
+        "addic.  %[len], %[len], -16\n\t"
+        "std     18, 0(%[x])\n\t"
+        "std     19, 8(%[x])\n\t"
+        "bne     L_GCM_gmult_len_start_block_%=\n\t"
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+        : [x] "+r" (x), [m] "+r" (m), [data] "+r" (data), [len] "+r" (len),
+          [L_GCM_gmult_len_r] "+r" (L_GCM_gmult_len_r_c)
+        :
+        : "memory", "cc", "0", "8", "9", "10", "11", "12", "14", "15", "16",
+            "17", "18", "19", "20", "21", "22", "23", "24", "25"
+#else
+        :
+        : [x] "r" (x), [m] "r" (m), [data] "r" (data), [len] "r" (len),
+          [L_GCM_gmult_len_r] "r" (L_GCM_gmult_len_r_c)
+        : "memory", "cc", "0", "7", "8", "9", "10", "11", "12", "14", "15",
+            "16", "17", "18", "19", "20", "21", "22", "23", "24", "25"
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+    );
+}
+
+#endif /* GCM_TABLE_4BIT */
+#ifdef GCM_TABLE
+static const byte L_GCM_gmult_len_r[] = {
+    0x00, 0x00, 0xc2, 0x01, 0x84, 0x03, 0x46, 0x02,
+    0x08, 0x07, 0xca, 0x06, 0x8c, 0x04, 0x4e, 0x05,
+    0x10, 0x0e, 0xd2, 0x0f, 0x94, 0x0d, 0x56, 0x0c,
+    0x18, 0x09, 0xda, 0x08, 0x9c, 0x0a, 0x5e, 0x0b,
+    0x20, 0x1c, 0xe2, 0x1d, 0xa4, 0x1f, 0x66, 0x1e,
+    0x28, 0x1b, 0xea, 0x1a, 0xac, 0x18, 0x6e, 0x19,
+    0x30, 0x12, 0xf2, 0x13, 0xb4, 0x11, 0x76, 0x10,
+    0x38, 0x15, 0xfa, 0x14, 0xbc, 0x16, 0x7e, 0x17,
+    0x40, 0x38, 0x82, 0x39, 0xc4, 0x3b, 0x06, 0x3a,
+    0x48, 0x3f, 0x8a, 0x3e, 0xcc, 0x3c, 0x0e, 0x3d,
+    0x50, 0x36, 0x92, 0x37, 0xd4, 0x35, 0x16, 0x34,
+    0x58, 0x31, 0x9a, 0x30, 0xdc, 0x32, 0x1e, 0x33,
+    0x60, 0x24, 0xa2, 0x25, 0xe4, 0x27, 0x26, 0x26,
+    0x68, 0x23, 0xaa, 0x22, 0xec, 0x20, 0x2e, 0x21,
+    0x70, 0x2a, 0xb2, 0x2b, 0xf4, 0x29, 0x36, 0x28,
+    0x78, 0x2d, 0xba, 0x2c, 0xfc, 0x2e, 0x3e, 0x2f,
+    0x80, 0x70, 0x42, 0x71, 0x04, 0x73, 0xc6, 0x72,
+    0x88, 0x77, 0x4a, 0x76, 0x0c, 0x74, 0xce, 0x75,
+    0x90, 0x7e, 0x52, 0x7f, 0x14, 0x7d, 0xd6, 0x7c,
+    0x98, 0x79, 0x5a, 0x78, 0x1c, 0x7a, 0xde, 0x7b,
+    0xa0, 0x6c, 0x62, 0x6d, 0x24, 0x6f, 0xe6, 0x6e,
+    0xa8, 0x6b, 0x6a, 0x6a, 0x2c, 0x68, 0xee, 0x69,
+    0xb0, 0x62, 0x72, 0x63, 0x34, 0x61, 0xf6, 0x60,
+    0xb8, 0x65, 0x7a, 0x64, 0x3c, 0x66, 0xfe, 0x67,
+    0xc0, 0x48, 0x02, 0x49, 0x44, 0x4b, 0x86, 0x4a,
+    0xc8, 0x4f, 0x0a, 0x4e, 0x4c, 0x4c, 0x8e, 0x4d,
+    0xd0, 0x46, 0x12, 0x47, 0x54, 0x45, 0x96, 0x44,
+    0xd8, 0x41, 0x1a, 0x40, 0x5c, 0x42, 0x9e, 0x43,
+    0xe0, 0x54, 0x22, 0x55, 0x64, 0x57, 0xa6, 0x56,
+    0xe8, 0x53, 0x2a, 0x52, 0x6c, 0x50, 0xae, 0x51,
+    0xf0, 0x5a, 0x32, 0x5b, 0x74, 0x59, 0xb6, 0x58,
+    0xf8, 0x5d, 0x3a, 0x5c, 0x7c, 0x5e, 0xbe, 0x5f,
+    0x00, 0xe1, 0xc2, 0xe0, 0x84, 0xe2, 0x46, 0xe3,
+    0x08, 0xe6, 0xca, 0xe7, 0x8c, 0xe5, 0x4e, 0xe4,
+    0x10, 0xef, 0xd2, 0xee, 0x94, 0xec, 0x56, 0xed,
+    0x18, 0xe8, 0xda, 0xe9, 0x9c, 0xeb, 0x5e, 0xea,
+    0x20, 0xfd, 0xe2, 0xfc, 0xa4, 0xfe, 0x66, 0xff,
+    0x28, 0xfa, 0xea, 0xfb, 0xac, 0xf9, 0x6e, 0xf8,
+    0x30, 0xf3, 0xf2, 0xf2, 0xb4, 0xf0, 0x76, 0xf1,
+    0x38, 0xf4, 0xfa, 0xf5, 0xbc, 0xf7, 0x7e, 0xf6,
+    0x40, 0xd9, 0x82, 0xd8, 0xc4, 0xda, 0x06, 0xdb,
+    0x48, 0xde, 0x8a, 0xdf, 0xcc, 0xdd, 0x0e, 0xdc,
+    0x50, 0xd7, 0x92, 0xd6, 0xd4, 0xd4, 0x16, 0xd5,
+    0x58, 0xd0, 0x9a, 0xd1, 0xdc, 0xd3, 0x1e, 0xd2,
+    0x60, 0xc5, 0xa2, 0xc4, 0xe4, 0xc6, 0x26, 0xc7,
+    0x68, 0xc2, 0xaa, 0xc3, 0xec, 0xc1, 0x2e, 0xc0,
+    0x70, 0xcb, 0xb2, 0xca, 0xf4, 0xc8, 0x36, 0xc9,
+    0x78, 0xcc, 0xba, 0xcd, 0xfc, 0xcf, 0x3e, 0xce,
+    0x80, 0x91, 0x42, 0x90, 0x04, 0x92, 0xc6, 0x93,
+    0x88, 0x96, 0x4a, 0x97, 0x0c, 0x95, 0xce, 0x94,
+    0x90, 0x9f, 0x52, 0x9e, 0x14, 0x9c, 0xd6, 0x9d,
+    0x98, 0x98, 0x5a, 0x99, 0x1c, 0x9b, 0xde, 0x9a,
+    0xa0, 0x8d, 0x62, 0x8c, 0x24, 0x8e, 0xe6, 0x8f,
+    0xa8, 0x8a, 0x6a, 0x8b, 0x2c, 0x89, 0xee, 0x88,
+    0xb0, 0x83, 0x72, 0x82, 0x34, 0x80, 0xf6, 0x81,
+    0xb8, 0x84, 0x7a, 0x85, 0x3c, 0x87, 0xfe, 0x86,
+    0xc0, 0xa9, 0x02, 0xa8, 0x44, 0xaa, 0x86, 0xab,
+    0xc8, 0xae, 0x0a, 0xaf, 0x4c, 0xad, 0x8e, 0xac,
+    0xd0, 0xa7, 0x12, 0xa6, 0x54, 0xa4, 0x96, 0xa5,
+    0xd8, 0xa0, 0x1a, 0xa1, 0x5c, 0xa3, 0x9e, 0xa2,
+    0xe0, 0xb5, 0x22, 0xb4, 0x64, 0xb6, 0xa6, 0xb7,
+    0xe8, 0xb2, 0x2a, 0xb3, 0x6c, 0xb1, 0xae, 0xb0,
+    0xf0, 0xbb, 0x32, 0xba, 0x74, 0xb8, 0xb6, 0xb9,
+    0xf8, 0xbc, 0x3a, 0xbd, 0x7c, 0xbf, 0xbe, 0xbe,
+};
+
+void GCM_gmult_len(unsigned char* x, const unsigned char** m,
+    const unsigned char* data, unsigned long len);
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+void GCM_gmult_len(unsigned char* x_p, const unsigned char** m_p,
+    const unsigned char* data_p, unsigned long len_p)
+#else
+void GCM_gmult_len(unsigned char* x, const unsigned char** m,
+    const unsigned char* data, unsigned long len)
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+{
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+    register unsigned char* x asm ("3") = (unsigned char*)x_p;
+    register const unsigned char** m asm ("4") = (const unsigned char**)m_p;
+    register const unsigned char* data asm ("5") =
+        (const unsigned char*)data_p;
+    register unsigned long len asm ("6") = (unsigned long)len_p;
+    register byte* L_GCM_gmult_len_r_c asm ("7") = (byte*)&L_GCM_gmult_len_r;
+#else
+    register byte* L_GCM_gmult_len_r_c = (byte*)&L_GCM_gmult_len_r;
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+
+    __asm__ __volatile__ (
+#ifdef WOLFSSL_NO_VAR_ASSIGN_REG
+        "mr      7, %[L_GCM_gmult_len_r]\n\t"
+#endif /* WOLFSSL_NO_VAR_ASSIGN_REG */
+        "addi    0, %[m], 8\n\t"
+        "li      17, 8\n\t"
+        "\n"
+    "L_GCM_gmult_len_start_block_%=: \n\t"
+        "ldbrx   8, 0, %[x]\n\t"
+        "ldbrx   9, 17, %[x]\n\t"
+        "ldbrx   10, 0, %[data]\n\t"
+        "ldbrx   11, 17, %[data]\n\t"
+        "xor     8, 8, 10\n\t"
+        "xor     9, 9, 11\n\t"
+        "rldicr  19, 9, 32, 63\n\t"
+        "rldicr  18, 8, 32, 63\n\t"
+        "rlwinm  16, 19, 12, 20, 27\n\t"
+        /* Byte 15 */
+        "ldx     12, 16, %[m]\n\t"
+        "ldx     14, 16, 0\n\t"
+        "rldicl  11, 12, 8, 56\n\t"
+        "rldicl  15, 14, 9, 55\n\t"
+        "sldi    14, 14, 8\n\t"
+        "andi.   15, 15, 510\n\t"
+        "sldi    12, 12, 8\n\t"
+        "lhzx    10, 15, 7\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 19, 20, 20, 27\n\t"
+        "xor     12, 12, 10\n\t"
+        /* Byte 14 */
+        "ldx     10, 16, %[m]\n\t"
+        "ldx     11, 16, 0\n\t"
+        "xor     12, 12, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rldicl  11, 12, 8, 56\n\t"
+        "rldicl  15, 14, 9, 55\n\t"
+        "sldi    14, 14, 8\n\t"
+        "andi.   15, 15, 510\n\t"
+        "sldi    12, 12, 8\n\t"
+        "lhzx    10, 15, 7\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 19, 28, 20, 27\n\t"
+        "xor     12, 12, 10\n\t"
+        /* Byte 13 */
+        "ldx     10, 16, %[m]\n\t"
+        "ldx     11, 16, 0\n\t"
+        "xor     12, 12, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rldicl  11, 12, 8, 56\n\t"
+        "rldicl  15, 14, 9, 55\n\t"
+        "sldi    14, 14, 8\n\t"
+        "andi.   15, 15, 510\n\t"
+        "sldi    12, 12, 8\n\t"
+        "lhzx    10, 15, 7\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 19, 4, 20, 27\n\t"
+        "xor     12, 12, 10\n\t"
+        /* Byte 12 */
+        "ldx     10, 16, %[m]\n\t"
+        "ldx     11, 16, 0\n\t"
+        "xor     12, 12, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rldicl  11, 12, 8, 56\n\t"
+        "rldicl  15, 14, 9, 55\n\t"
+        "sldi    14, 14, 8\n\t"
+        "andi.   15, 15, 510\n\t"
+        "sldi    12, 12, 8\n\t"
+        "lhzx    10, 15, 7\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 9, 12, 20, 27\n\t"
+        "xor     12, 12, 10\n\t"
+        /* Byte 11 */
+        "ldx     10, 16, %[m]\n\t"
+        "ldx     11, 16, 0\n\t"
+        "xor     12, 12, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rldicl  11, 12, 8, 56\n\t"
+        "rldicl  15, 14, 9, 55\n\t"
+        "sldi    14, 14, 8\n\t"
+        "andi.   15, 15, 510\n\t"
+        "sldi    12, 12, 8\n\t"
+        "lhzx    10, 15, 7\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 9, 20, 20, 27\n\t"
+        "xor     12, 12, 10\n\t"
+        /* Byte 10 */
+        "ldx     10, 16, %[m]\n\t"
+        "ldx     11, 16, 0\n\t"
+        "xor     12, 12, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rldicl  11, 12, 8, 56\n\t"
+        "rldicl  15, 14, 9, 55\n\t"
+        "sldi    14, 14, 8\n\t"
+        "andi.   15, 15, 510\n\t"
+        "sldi    12, 12, 8\n\t"
+        "lhzx    10, 15, 7\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 9, 28, 20, 27\n\t"
+        "xor     12, 12, 10\n\t"
+        /* Byte 9 */
+        "ldx     10, 16, %[m]\n\t"
+        "ldx     11, 16, 0\n\t"
+        "xor     12, 12, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rldicl  11, 12, 8, 56\n\t"
+        "rldicl  15, 14, 9, 55\n\t"
+        "sldi    14, 14, 8\n\t"
+        "andi.   15, 15, 510\n\t"
+        "sldi    12, 12, 8\n\t"
+        "lhzx    10, 15, 7\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 9, 4, 20, 27\n\t"
+        "xor     12, 12, 10\n\t"
+        /* Byte 8 */
+        "ldx     10, 16, %[m]\n\t"
+        "ldx     11, 16, 0\n\t"
+        "xor     12, 12, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rldicl  11, 12, 8, 56\n\t"
+        "rldicl  15, 14, 9, 55\n\t"
+        "sldi    14, 14, 8\n\t"
+        "andi.   15, 15, 510\n\t"
+        "sldi    12, 12, 8\n\t"
+        "lhzx    10, 15, 7\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 18, 12, 20, 27\n\t"
+        "xor     12, 12, 10\n\t"
+        /* Byte 7 */
+        "ldx     10, 16, %[m]\n\t"
+        "ldx     11, 16, 0\n\t"
+        "xor     12, 12, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rldicl  11, 12, 8, 56\n\t"
+        "rldicl  15, 14, 9, 55\n\t"
+        "sldi    14, 14, 8\n\t"
+        "andi.   15, 15, 510\n\t"
+        "sldi    12, 12, 8\n\t"
+        "lhzx    10, 15, 7\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 18, 20, 20, 27\n\t"
+        "xor     12, 12, 10\n\t"
+        /* Byte 6 */
+        "ldx     10, 16, %[m]\n\t"
+        "ldx     11, 16, 0\n\t"
+        "xor     12, 12, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rldicl  11, 12, 8, 56\n\t"
+        "rldicl  15, 14, 9, 55\n\t"
+        "sldi    14, 14, 8\n\t"
+        "andi.   15, 15, 510\n\t"
+        "sldi    12, 12, 8\n\t"
+        "lhzx    10, 15, 7\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 18, 28, 20, 27\n\t"
+        "xor     12, 12, 10\n\t"
+        /* Byte 5 */
+        "ldx     10, 16, %[m]\n\t"
+        "ldx     11, 16, 0\n\t"
+        "xor     12, 12, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rldicl  11, 12, 8, 56\n\t"
+        "rldicl  15, 14, 9, 55\n\t"
+        "sldi    14, 14, 8\n\t"
+        "andi.   15, 15, 510\n\t"
+        "sldi    12, 12, 8\n\t"
+        "lhzx    10, 15, 7\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 18, 4, 20, 27\n\t"
+        "xor     12, 12, 10\n\t"
+        /* Byte 4 */
+        "ldx     10, 16, %[m]\n\t"
+        "ldx     11, 16, 0\n\t"
+        "xor     12, 12, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rldicl  11, 12, 8, 56\n\t"
+        "rldicl  15, 14, 9, 55\n\t"
+        "sldi    14, 14, 8\n\t"
+        "andi.   15, 15, 510\n\t"
+        "sldi    12, 12, 8\n\t"
+        "lhzx    10, 15, 7\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 8, 12, 20, 27\n\t"
+        "xor     12, 12, 10\n\t"
+        /* Byte 3 */
+        "ldx     10, 16, %[m]\n\t"
+        "ldx     11, 16, 0\n\t"
+        "xor     12, 12, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rldicl  11, 12, 8, 56\n\t"
+        "rldicl  15, 14, 9, 55\n\t"
+        "sldi    14, 14, 8\n\t"
+        "andi.   15, 15, 510\n\t"
+        "sldi    12, 12, 8\n\t"
+        "lhzx    10, 15, 7\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 8, 20, 20, 27\n\t"
+        "xor     12, 12, 10\n\t"
+        /* Byte 2 */
+        "ldx     10, 16, %[m]\n\t"
+        "ldx     11, 16, 0\n\t"
+        "xor     12, 12, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rldicl  11, 12, 8, 56\n\t"
+        "rldicl  15, 14, 9, 55\n\t"
+        "sldi    14, 14, 8\n\t"
+        "andi.   15, 15, 510\n\t"
+        "sldi    12, 12, 8\n\t"
+        "lhzx    10, 15, 7\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 8, 28, 20, 27\n\t"
+        "xor     12, 12, 10\n\t"
+        /* Byte 1 */
+        "ldx     10, 16, %[m]\n\t"
+        "ldx     11, 16, 0\n\t"
+        "xor     12, 12, 10\n\t"
+        "xor     14, 14, 11\n\t"
+        "rldicl  11, 12, 8, 56\n\t"
+        "rldicl  15, 14, 9, 55\n\t"
+        "sldi    14, 14, 8\n\t"
+        "andi.   15, 15, 510\n\t"
+        "sldi    12, 12, 8\n\t"
+        "lhzx    10, 15, 7\n\t"
+        "xor     14, 14, 11\n\t"
+        "rlwinm  16, 8, 4, 20, 27\n\t"
+        "xor     12, 12, 10\n\t"
+        /* Byte 0 */
+        "ldx     10, 16, %[m]\n\t"
+        "ldx     11, 16, 0\n\t"
+        "addi    %[data], %[data], 16\n\t"
+        "addic.  %[len], %[len], -16\n\t"
+        "xor     8, 12, 10\n\t"
+        "xor     9, 14, 11\n\t"
+        "stdbrx  8, 0, %[x]\n\t"
+        "stdbrx  9, 17, %[x]\n\t"
+        "bne     L_GCM_gmult_len_start_block_%=\n\t"
+#ifndef WOLFSSL_NO_VAR_ASSIGN_REG
+        : [x] "+r" (x), [m] "+r" (m), [data] "+r" (data), [len] "+r" (len),
+          [L_GCM_gmult_len_r] "+r" (L_GCM_gmult_len_r_c)
+        :
+        : "memory", "cc", "0", "8", "9", "10", "11", "12", "14", "15", "16",
+            "17", "18", "19"
+#else
+        :
+        : [x] "r" (x), [m] "r" (m), [data] "r" (data), [len] "r" (len),
+          [L_GCM_gmult_len_r] "r" (L_GCM_gmult_len_r_c)
+        : "memory", "cc", "0", "7", "8", "9", "10", "11", "12", "14", "15",
+            "16", "17", "18", "19"
+#endif /* !WOLFSSL_NO_VAR_ASSIGN_REG */
+    );
+}
+
+#endif /* GCM_TABLE */
+#endif /* HAVE_AESGCM */
+#endif /* !defined(NO_AES) && defined(WOLFSSL_PPC64_ASM) */
+#endif /* WOLFSSL_PPC64_ASM */
+
+#endif /* WOLFSSL_PPC64_ASM_INLINE */
diff --git a/wolfcrypt/src/port/riscv/riscv-64-aes.c b/wolfcrypt/src/port/riscv/riscv-64-aes.c
index 37918217e9..64d77a3630 100644
--- a/wolfcrypt/src/port/riscv/riscv-64-aes.c
+++ b/wolfcrypt/src/port/riscv/riscv-64-aes.c
@@ -1871,8 +1871,7 @@ int wc_AesSetKey(Aes* aes, const byte* key, word32 keyLen, const byte* iv,
 static void wc_AesEncrypt(Aes* aes, const byte* in, byte* out)
 {
     __asm__ __volatile__ (
-        "ld          t2, 0(%[in])\n\t"
-        "ld          t3, 8(%[in])\n\t"
+        UNALIGNED_LD2(t2, t3, 0, %[in], t0)
         "ld          a3, 0(%[key])\n\t"
         "ld          a4, 8(%[key])\n\t"
         "ld          a5, 16(%[key])\n\t"
@@ -1897,8 +1896,7 @@ static void wc_AesEncrypt(Aes* aes, const byte* in, byte* out)
         AESENC_2_ROUNDS(208, 216, 224, 232)
       "L_aes_encrypt_done:\n\t"
         AESENC_LAST_ROUND()
-        "sd          t2, 0(%[out])\n\t"
-        "sd          t3, 8(%[out])\n\t"
+        UNALIGNED_SD2(t2, t3, 0, %[out], t0)
         :
         : [in] "r" (in), [out] "r" (out), [key] "r" (aes->key),
           [rounds] "r" (aes->rounds)
@@ -1918,8 +1916,7 @@ static void wc_AesEncrypt(Aes* aes, const byte* in, byte* out)
 static void wc_AesDecrypt(Aes* aes, const byte* in, byte* out)
 {
     __asm__ __volatile__ (
-        "ld          t2, 0(%[in])\n\t"
-        "ld          t3, 8(%[in])\n\t"
+        UNALIGNED_LD2(t2, t3, 0, %[in], t0)
         "ld          a3, 0(%[key])\n\t"
         "ld          a4, 8(%[key])\n\t"
         "ld          a5, 16(%[key])\n\t"
@@ -1944,8 +1941,7 @@ static void wc_AesDecrypt(Aes* aes, const byte* in, byte* out)
         AESDEC_2_ROUNDS(208, 216, 224, 232)
       "L_aes_decrypt_done:\n\t"
         AESDEC_LAST_ROUND()
-        "sd          t2, 0(%[out])\n\t"
-        "sd          t3, 8(%[out])\n\t"
+        UNALIGNED_SD2(t2, t3, 0, %[out], t0)
         :
         : [in] "r" (in), [out] "r" (out), [key] "r" (aes->key),
           [rounds] "r" (aes->rounds)
@@ -3209,8 +3205,7 @@ static void wc_AesEncrypt(Aes* aes, const byte* in, byte* out)
         LOAD_WORD_REV(t2, 8, %[in])
         LOAD_WORD_REV(t3, 12, %[in])
 #else
-        "ld         t1,  0(%[in])\n\t"
-        "ld         t3,  8(%[in])\n\t"
+        UNALIGNED_LD2(t1, t3, 0, %[in], t0)
         REV8(REG_T1, REG_T1)
         REV8(REG_T3, REG_T3)
         "srli       t0, t1, 32\n\t"
@@ -3376,16 +3371,14 @@ static void wc_AesEncrypt(Aes* aes, const byte* in, byte* out)
         REV8(REG_T1, REG_T1)
         REV8(REG_T3, REG_T3)
         /* Write encrypted block to output. */
-        "sd         t1,  0(%[out])\n\t"
-        "sd         t3,  8(%[out])\n\t"
+        UNALIGNED_SD2(t1, t3, 0, %[out], t0)
 #else
         PACK(REG_T1, REG_A5, REG_A4)
         PACK(REG_T3, REG_A7, REG_A6)
         REV8(REG_T1, REG_T1)
         REV8(REG_T3, REG_T3)
         /* Write encrypted block to output. */
-        "sd         t1,  0(%[out])\n\t"
-        "sd         t3,  8(%[out])\n\t"
+        UNALIGNED_SD2(t1, t3, 0, %[out], t0)
 #endif
 
         :
@@ -3641,8 +3634,7 @@ static void wc_AesDecrypt(Aes* aes, const byte* in, byte* out)
         LOAD_WORD_REV(t2, 8, %[in])
         LOAD_WORD_REV(t3, 12, %[in])
 #else
-        "ld         t1,  0(%[in])\n\t"
-        "ld         t3,  8(%[in])\n\t"
+        UNALIGNED_LD2(t1, t3, 0, %[in], t0)
         REV8(REG_T1, REG_T1)
         REV8(REG_T3, REG_T3)
         "srli       t0, t1, 32\n\t"
@@ -3793,16 +3785,14 @@ static void wc_AesDecrypt(Aes* aes, const byte* in, byte* out)
         REV8(REG_T1, REG_T1)
         REV8(REG_T3, REG_T3)
         /* Write encrypted block to output. */
-        "sd         t1,  0(%[out])\n\t"
-        "sd         t3,  8(%[out])\n\t"
+        UNALIGNED_SD2(t1, t3, 0, %[out], t0)
 #else
         PACK(REG_T1, REG_A5, REG_A4)
         PACK(REG_T3, REG_A7, REG_A6)
         REV8(REG_T1, REG_T1)
         REV8(REG_T3, REG_T3)
         /* Write encrypted block to output. */
-        "sd         t1,  0(%[out])\n\t"
-        "sd         t3,  8(%[out])\n\t"
+        UNALIGNED_SD2(t1, t3, 0, %[out], t0)
 #endif
 
         :
@@ -4113,7 +4103,7 @@ static WC_INLINE void IncrementAesCounter(byte* inOutCtr)
  */
 int wc_AesCtrEncrypt(Aes* aes, byte* out, const byte* in, word32 sz)
 {
-    byte scratch[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte scratch[WC_AES_BLOCK_SIZE];
     word32 processed;
     int ret = 0;
 
@@ -4563,8 +4553,8 @@ void GHASH(Gcm* gcm, const byte* a, word32 aSz, const byte* c, word32 cSz,
     byte* s, word32 sSz)
 {
     if (gcm != NULL) {
-        byte x[WC_AES_BLOCK_SIZE];
-        byte scratch[WC_AES_BLOCK_SIZE];
+        ALIGN8 byte x[WC_AES_BLOCK_SIZE];
+        ALIGN8 byte scratch[WC_AES_BLOCK_SIZE];
         byte* h = gcm->H;
 
         __asm__ __volatile__ (
@@ -4896,8 +4886,8 @@ static void GMULT(byte* x, byte* y)
 void GHASH(Gcm* gcm, const byte* a, word32 aSz, const byte* c, word32 cSz,
     byte* s, word32 sSz)
 {
-    byte x[WC_AES_BLOCK_SIZE];
-    byte scratch[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte x[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte scratch[WC_AES_BLOCK_SIZE];
     word32 blocks, partial;
     byte* h;
 
@@ -5163,8 +5153,7 @@ static void ghash_blocks(byte* x, byte* y, const byte* in, word32 blocks)
 
     "L_ghash_loop:\n\t"
         /* Load input block. */
-        "ld          t5, 0(%[in])\n\t"
-        "ld          a5, 8(%[in])\n\t"
+        UNALIGNED_LD2(t5, a5, 0, %[in], t4)
         /* Reverse bits to match x. */
 #ifdef WOLFSSL_RISCV_BIT_MANIPULATION
         BREV8(REG_T5, REG_T5)
@@ -5307,8 +5296,8 @@ void GHASH(Gcm* gcm, const byte* a, word32 aSz, const byte* c, word32 cSz,
     byte* s, word32 sSz)
 {
     if (gcm != NULL) {
-        byte x[WC_AES_BLOCK_SIZE];
-        byte scratch[WC_AES_BLOCK_SIZE];
+        ALIGN8 byte x[WC_AES_BLOCK_SIZE];
+        ALIGN8 byte scratch[WC_AES_BLOCK_SIZE];
         word32 blocks, partial;
         byte* h = gcm->H;
 
@@ -5388,8 +5377,8 @@ static void Aes128GcmEncrypt(Aes* aes, byte* out, const byte* in, word32 sz,
     const byte* nonce, word32 nonceSz, byte* tag, word32 tagSz,
     const byte* aad, word32 aadSz)
 {
-    byte counter[WC_AES_BLOCK_SIZE];
-    byte scratch[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte counter[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte scratch[WC_AES_BLOCK_SIZE];
     /* Noticed different optimization levels treated head of array different.
      * Some cases was stack pointer plus offset others was a register containing
      * address. To make uniform for passing in to inline assembly code am using
@@ -5886,8 +5875,8 @@ static void Aes192GcmEncrypt(Aes* aes, byte* out, const byte* in, word32 sz,
     const byte* nonce, word32 nonceSz, byte* tag, word32 tagSz,
     const byte* aad, word32 aadSz)
 {
-    byte counter[WC_AES_BLOCK_SIZE];
-    byte scratch[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte counter[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte scratch[WC_AES_BLOCK_SIZE];
     /* Noticed different optimization levels treated head of array different.
      * Some cases was stack pointer plus offset others was a register containing
      * address. To make uniform for passing in to inline assembly code am using
@@ -6398,8 +6387,8 @@ static void Aes256GcmEncrypt(Aes* aes, byte* out, const byte* in, word32 sz,
     const byte* nonce, word32 nonceSz, byte* tag, word32 tagSz,
     const byte* aad, word32 aadSz)
 {
-    byte counter[WC_AES_BLOCK_SIZE];
-    byte scratch[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte counter[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte scratch[WC_AES_BLOCK_SIZE];
     /* Noticed different optimization levels treated head of array different.
      * Some cases was stack pointer plus offset others was a register containing
      * address. To make uniform for passing in to inline assembly code am using
@@ -7003,8 +6992,8 @@ static int Aes128GcmDecrypt(Aes* aes, byte* out, const byte* in, word32 sz,
     const byte* aad, word32 aadSz)
 {
     int ret = 0;
-    byte counter[WC_AES_BLOCK_SIZE];
-    byte scratch[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte counter[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte scratch[WC_AES_BLOCK_SIZE];
     /* Noticed different optimization levels treated head of array different.
      * Some cases was stack pointer plus offset others was a register containing
      * address. To make uniform for passing in to inline assembly code am using
@@ -7512,8 +7501,8 @@ static int Aes192GcmDecrypt(Aes* aes, byte* out, const byte* in, word32 sz,
     const byte* aad, word32 aadSz)
 {
     int ret = 0;
-    byte counter[WC_AES_BLOCK_SIZE];
-    byte scratch[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte counter[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte scratch[WC_AES_BLOCK_SIZE];
     /* Noticed different optimization levels treated head of array different.
      * Some cases was stack pointer plus offset others was a register containing
      * address. To make uniform for passing in to inline assembly code am using
@@ -8035,8 +8024,8 @@ static int Aes256GcmDecrypt(Aes* aes, byte* out, const byte* in, word32 sz,
     const byte* aad, word32 aadSz)
 {
     int ret = 0;
-    byte counter[WC_AES_BLOCK_SIZE];
-    byte scratch[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte counter[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte scratch[WC_AES_BLOCK_SIZE];
     /* Noticed different optimization levels treated head of array different.
      * Some cases was stack pointer plus offset others was a register containing
      * address. To make uniform for passing in to inline assembly code am using
@@ -8733,8 +8722,8 @@ void GHASH(Gcm* gcm, const byte* a, word32 aSz, const byte* c, word32 cSz,
     byte* s, word32 sSz)
 {
     if (gcm != NULL) {
-        byte x[WC_AES_BLOCK_SIZE];
-        byte scratch[WC_AES_BLOCK_SIZE];
+        ALIGN8 byte x[WC_AES_BLOCK_SIZE];
+        ALIGN8 byte scratch[WC_AES_BLOCK_SIZE];
         word32 blocks, partial;
 
         XMEMSET(x, 0, WC_AES_BLOCK_SIZE);
@@ -8834,9 +8823,9 @@ int wc_AesGcmEncrypt(Aes* aes, byte* out, const byte* in, word32 sz,
     word32 partial = sz % WC_AES_BLOCK_SIZE;
     const byte* p = in;
     byte* c = out;
-    ALIGN16 byte counter[WC_AES_BLOCK_SIZE];
-    ALIGN16 byte initialCounter[WC_AES_BLOCK_SIZE];
-    ALIGN16 byte scratch[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte counter[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte initialCounter[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte scratch[WC_AES_BLOCK_SIZE];
 
     /* Validate parameters. */
     if ((aes == NULL) || (nonce == NULL) || (nonceSz == 0) || (tag == NULL) ||
@@ -8934,10 +8923,10 @@ int wc_AesGcmDecrypt(Aes* aes, byte* out, const byte* in, word32 sz,
     word32 partial = sz % WC_AES_BLOCK_SIZE;
     const byte* c = in;
     byte* p = out;
-    ALIGN16 byte counter[WC_AES_BLOCK_SIZE];
-    ALIGN16 byte scratch[WC_AES_BLOCK_SIZE];
-    ALIGN16 byte Tprime[WC_AES_BLOCK_SIZE];
-    ALIGN16 byte EKY0[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte counter[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte scratch[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte Tprime[WC_AES_BLOCK_SIZE];
+    ALIGN8 byte EKY0[WC_AES_BLOCK_SIZE];
     sword32 res;
 
     /* Validate parameters. */
diff --git a/wolfcrypt/src/port/riscv/riscv-64-chacha.c b/wolfcrypt/src/port/riscv/riscv-64-chacha.c
index c150e4bb7f..ba0a8b21d3 100644
--- a/wolfcrypt/src/port/riscv/riscv-64-chacha.c
+++ b/wolfcrypt/src/port/riscv/riscv-64-chacha.c
@@ -1825,9 +1825,9 @@ static WC_INLINE void wc_chacha_encrypt_64(const word32* input, const byte* m,
         VSETIVLI(REG_X0, 2, 1, 1, 0b011, 0b000)
         VMV_X_S(REG_T0, REG_V0)
         VSETIVLI(REG_X0, 4, 1, 1, 0b010, 0b000)
-        "ld     t1, (%[m])\n\t"
+        UNALIGNED_LD(t1, 0, %[m], t2)
         "xor    t1, t1, t0\n\t"
-        "sd     t1, (%[c])\n\t"
+        UNALIGNED_SD(t1, 0, %[c], t2)
         "addi   %[bytes], %[bytes], -8\n\t"
         "addi   %[c], %[c], 8\n\t"
         "addi   %[m], %[m], 8\n\t"
@@ -2155,10 +2155,7 @@ static WC_INLINE void wc_chacha_encrypt(const word32* input, const byte* m,
         "bltz   %[bytes], L_chacha20_riscv_over\n\t"
 
 #if !defined(WOLFSSL_RISCV_BIT_MANIPULATION)
-        "ld     t0, 0(%[m])\n\t"
-        "ld     t1, 8(%[m])\n\t"
-        "ld     t2, 16(%[m])\n\t"
-        "ld     s1, 24(%[m])\n\t"
+        UNALIGNED_LD4(t0, t1, t2, s1, 0, %[m], a3)
         "xor    a4, a4, t0\n\t"
         "xor    a6, a6, t1\n\t"
         "xor    t3, t3, t2\n\t"
@@ -2171,10 +2168,7 @@ static WC_INLINE void wc_chacha_encrypt(const word32* input, const byte* m,
         "xor    a7, a7, t1\n\t"
         "xor    t4, t4, t2\n\t"
         "xor    t6, t6, s1\n\t"
-        "ld     t0, 32(%[m])\n\t"
-        "ld     t1, 40(%[m])\n\t"
-        "ld     t2, 48(%[m])\n\t"
-        "ld     s1, 56(%[m])\n\t"
+        UNALIGNED_LD4(t0, t1, t2, s1, 32, %[m], a3)
         "xor    s2, s2, t0\n\t"
         "xor    s4, s4, t1\n\t"
         "xor    s6, s6, t2\n\t"
@@ -2187,22 +2181,8 @@ static WC_INLINE void wc_chacha_encrypt(const word32* input, const byte* m,
         "xor    s5, s5, t1\n\t"
         "xor    s7, s7, t2\n\t"
         "xor    s9, s9, s1\n\t"
-        "sw     a4, 0(%[c])\n\t"
-        "sw     a5, 4(%[c])\n\t"
-        "sw     a6, 8(%[c])\n\t"
-        "sw     a7, 12(%[c])\n\t"
-        "sw     t3, 16(%[c])\n\t"
-        "sw     t4, 20(%[c])\n\t"
-        "sw     t5, 24(%[c])\n\t"
-        "sw     t6, 28(%[c])\n\t"
-        "sw     s2, 32(%[c])\n\t"
-        "sw     s3, 36(%[c])\n\t"
-        "sw     s4, 40(%[c])\n\t"
-        "sw     s5, 44(%[c])\n\t"
-        "sw     s6, 48(%[c])\n\t"
-        "sw     s7, 52(%[c])\n\t"
-        "sw     s8, 56(%[c])\n\t"
-        "sw     s9, 60(%[c])\n\t"
+        UNALIGNED_SW8(a4, a5, a6, a7, t3, t4, t5, t6, 0, %[c], t0)
+        UNALIGNED_SW8(s2, s3, s4, s5, s6, s7, s8, s9, 32, %[c], t0)
 #else
         PACK(REG_A4, REG_A4, REG_A5)
         PACK(REG_A6, REG_A6, REG_A7)
@@ -2212,14 +2192,7 @@ static WC_INLINE void wc_chacha_encrypt(const word32* input, const byte* m,
         PACK(REG_S4, REG_S4, REG_S5)
         PACK(REG_S6, REG_S6, REG_S7)
         PACK(REG_S8, REG_S8, REG_S9)
-        "ld     a5, 0(%[m])\n\t"
-        "ld     a7, 8(%[m])\n\t"
-        "ld     t4, 16(%[m])\n\t"
-        "ld     t6, 24(%[m])\n\t"
-        "ld     s3, 32(%[m])\n\t"
-        "ld     s5, 40(%[m])\n\t"
-        "ld     s7, 48(%[m])\n\t"
-        "ld     s9, 56(%[m])\n\t"
+        UNALIGNED_LD8(a5, a7, t4, t6, s3, s5, s7, s9, 0, %[m], t0)
         "xor    a4, a4, a5\n\t"
         "xor    a6, a6, a7\n\t"
         "xor    t3, t3, t4\n\t"
@@ -2228,14 +2201,7 @@ static WC_INLINE void wc_chacha_encrypt(const word32* input, const byte* m,
         "xor    s4, s4, s5\n\t"
         "xor    s6, s6, s7\n\t"
         "xor    s8, s8, s9\n\t"
-        "sd     a4, 0(%[c])\n\t"
-        "sd     a6, 8(%[c])\n\t"
-        "sd     t3, 16(%[c])\n\t"
-        "sd     t5, 24(%[c])\n\t"
-        "sd     s2, 32(%[c])\n\t"
-        "sd     s4, 40(%[c])\n\t"
-        "sd     s6, 48(%[c])\n\t"
-        "sd     s8, 56(%[c])\n\t"
+        UNALIGNED_SD8(a4, a6, t3, t5, s2, s4, s6, s8, 0, %[c], t0)
 #endif
 
         "addi   %[m], %[m], 64\n\t"
@@ -2268,10 +2234,10 @@ static WC_INLINE void wc_chacha_encrypt(const word32* input, const byte* m,
         "bltz   t0, L_chacha20_riscv_32bit\n\t"
         "addi   a3, a3, -1\n\t"
      "L_chacha20_riscv_64bit_loop:\n\t"
-        "ld     t0, (%[m])\n\t"
+        UNALIGNED_LD(t0, 0, %[m], t2)
         "ld     t1, (%[over])\n\t"
         "xor    t0, t0, t1\n\t"
-        "sd     t0, (%[c])\n\t"
+        UNALIGNED_SD(t0, 0, %[c], t2)
         "addi   %[m], %[m], 8\n\t"
         "addi   %[c], %[c], 8\n\t"
         "addi   %[over], %[over], 8\n\t"
@@ -2282,10 +2248,10 @@ static WC_INLINE void wc_chacha_encrypt(const word32* input, const byte* m,
      "L_chacha20_riscv_32bit:\n\t"
         "addi   t0, a3, -4\n\t"
         "bltz   t0, L_chacha20_riscv_16bit\n\t"
-        "lw     t0, (%[m])\n\t"
+        UNALIGNED_LW(t0, 0, %[m], t2)
         "lw     t1, (%[over])\n\t"
         "xor    t0, t0, t1\n\t"
-        "sw     t0, (%[c])\n\t"
+        UNALIGNED_SW(t0, 0, %[c], t2)
         "addi   %[m], %[m], 4\n\t"
         "addi   %[c], %[c], 4\n\t"
         "addi   %[over], %[over], 4\n\t"
@@ -2293,10 +2259,10 @@ static WC_INLINE void wc_chacha_encrypt(const word32* input, const byte* m,
      "L_chacha20_riscv_16bit:\n\t"
         "addi   t0, a3, -2\n\t"
         "bltz   t0, L_chacha20_riscv_8bit\n\t"
-        "lh     t0, (%[m])\n\t"
+        UNALIGNED_LH(t0, 0, %[m], t2)
         "lh     t1, (%[over])\n\t"
         "xor    t0, t0, t1\n\t"
-        "sh     t0, (%[c])\n\t"
+        UNALIGNED_SH(t0, 0, %[c], t2)
         "addi   %[m], %[m], 2\n\t"
         "addi   %[c], %[c], 2\n\t"
         "addi   %[over], %[over], 2\n\t"
diff --git a/wolfcrypt/src/port/riscv/riscv-64-poly1305.c b/wolfcrypt/src/port/riscv/riscv-64-poly1305.c
index de846ffde8..5733b75bc7 100644
--- a/wolfcrypt/src/port/riscv/riscv-64-poly1305.c
+++ b/wolfcrypt/src/port/riscv/riscv-64-poly1305.c
@@ -145,8 +145,7 @@ static WC_INLINE void poly1305_blocks_riscv64_16(Poly1305* ctx,
 
     "L_poly1305_riscv64_16_64_loop_%=:\n\t"
         /* Load m */
-        "ld     t0, (%[m])\n\t"
-        "ld     t1, 8(%[m])\n\t"
+        UNALIGNED_LD2(t0, t1, 0, %[m], t5)
         /* Split m into 26, 52, 52 */
         SPLIT_130(t2, t3, t4, t0, t1, %[notLast], t5)
 
@@ -285,8 +284,7 @@ void poly1305_blocks_riscv64(Poly1305* ctx, const unsigned char *m,
 
     "L_poly1305_riscv64_vec_loop_%=:\n\t"
         /* m0 + nfin */
-        "ld     t0, 0(%[m])\n\t"
-        "ld     t1, 8(%[m])\n\t"
+        UNALIGNED_LD2(t0, t1, 0, %[m], t5)
         "li     t6, 1\n\t"
         /* Split m into 24, 52, 52 */
         SPLIT_130(t2, t3, t4, t0, t1, t6, t5)
@@ -294,8 +292,7 @@ void poly1305_blocks_riscv64(Poly1305* ctx, const unsigned char *m,
         VMV_S_X(REG_V12, REG_T3)
         VMV_S_X(REG_V13, REG_T4)
         /* m1+ nfin */
-        "ld     t0, 16(%[m])\n\t"
-        "ld     t1, 24(%[m])\n\t"
+        UNALIGNED_LD2(t0, t1, 16, %[m], t5)
         /* Split m into 24, 52, 52 */
         SPLIT_130(t2, t3, t4, t0, t1, t6, t5)
         VMV_S_X(REG_V14, REG_T2)
@@ -464,10 +461,7 @@ int wc_Poly1305SetKey(Poly1305* ctx, const byte* key, word32 keySz)
 
     __asm__ __volatile__ (
         /* Load key material */
-        "ld     t0, 0(%[key])\n\t"
-        "ld     t1, 8(%[key])\n\t"
-        "ld     t2, 16(%[key])\n\t"
-        "ld     t3, 24(%[key])\n\t"
+        UNALIGNED_LD4(t0, t1, t2, t3, 0, %[key], t4)
         /* Load clamp */
         "ld     t4, 0(%[clamp])\n\t"
         "ld     t5, 8(%[clamp])\n\t"
@@ -636,8 +630,7 @@ int wc_Poly1305Final(Poly1305* ctx, byte* mac)
         "sltu   t3, t1, t3\n\t"
         "add    t2, t2, t3\n\t"
         "andi   t2, t2, 3\n\t"
-        "sd     t0, 0(%[mac])\n\t"
-        "sd     t1, 8(%[mac])\n\t"
+        UNALIGNED_SD2(t0, t1, 0, %[mac], t2)
         /* Zero out h. */
         "sd     x0, %[ctx_h_0]\n\t"
         "sd     x0, %[ctx_h_1]\n\t"
diff --git a/wolfcrypt/src/port/riscv/riscv-64-sha256.c b/wolfcrypt/src/port/riscv/riscv-64-sha256.c
index 44af0b80c3..f672db2739 100644
--- a/wolfcrypt/src/port/riscv/riscv-64-sha256.c
+++ b/wolfcrypt/src/port/riscv/riscv-64-sha256.c
@@ -484,14 +484,7 @@ static WC_OMIT_FRAME_POINTER WC_INLINE void Sha256Transform(wc_Sha256* sha256,
         LOAD_DWORD_REV(s6, 48, %[data], a4, a5, a6, a7)
         LOAD_DWORD_REV(s7, 56, %[data], a4, a5, a6, a7)
 #else
-        "lwu    a4, 0(%[data])\n\t"
-        "lwu    s0, 4(%[data])\n\t"
-        "lwu    a5, 8(%[data])\n\t"
-        "lwu    s1, 12(%[data])\n\t"
-        "lwu    a6, 16(%[data])\n\t"
-        "lwu    s2, 20(%[data])\n\t"
-        "lwu    a7, 24(%[data])\n\t"
-        "lwu    s3, 28(%[data])\n\t"
+        UNALIGNED_LWU8(a4, s0, a5, s1, a6, s2, a7, s3, 0, %[data], t4)
         PACK_BB(s0, s0, a4, REG_S0, REG_S0, REG_A4)
         PACK_BB(s1, s1, a5, REG_S1, REG_S1, REG_A5)
         PACK_BB(s2, s2, a6, REG_S2, REG_S2, REG_A6)
@@ -500,14 +493,7 @@ static WC_OMIT_FRAME_POINTER WC_INLINE void Sha256Transform(wc_Sha256* sha256,
         REV8(REG_S1, REG_S1)
         REV8(REG_S2, REG_S2)
         REV8(REG_S3, REG_S3)
-        "lwu    a4, 32(%[data])\n\t"
-        "lwu    s4, 36(%[data])\n\t"
-        "lwu    a5, 40(%[data])\n\t"
-        "lwu    s5, 44(%[data])\n\t"
-        "lwu    a6, 48(%[data])\n\t"
-        "lwu    s6, 52(%[data])\n\t"
-        "lwu    a7, 56(%[data])\n\t"
-        "lwu    s7, 60(%[data])\n\t"
+        UNALIGNED_LWU8(a4, s4, a5, s5, a6, s6, a7, s7, 32, %[data], t4)
         PACK_BB(s4, s4, a4, REG_S4, REG_S4, REG_A4)
         PACK_BB(s5, s5, a5, REG_S5, REG_S5, REG_A5)
         PACK_BB(s6, s6, a6, REG_S6, REG_S6, REG_A6)
@@ -840,31 +826,18 @@ static WC_INLINE void Sha256Final(wc_Sha256* sha256, byte* hash)
         "srli   t2, t3, 32\n\t"
         "srli   a4, a5, 32\n\t"
         "srli   a6, a7, 32\n\t"
-        "sw     t0, 0(%[hash])\n\t"
-        "sw     t1, 4(%[hash])\n\t"
-        "sw     t2, 8(%[hash])\n\t"
-        "sw     t3, 12(%[hash])\n\t"
-        "sw     a4, 16(%[hash])\n\t"
-        "sw     a5, 20(%[hash])\n\t"
-        "sw     a6, 24(%[hash])\n\t"
-        "sw     a7, 28(%[hash])\n\t"
+        UNALIGNED_SW8(t0, t1, t2, t3, a4, a5, a6, a7, 0, %[hash], t4)
 #else
         LOAD_WORD_REV(t0, 0, %[digest], t2, t3, t4)
         LOAD_WORD_REV(t1, 4, %[digest], t2, t3, t4)
         LOAD_WORD_REV(a4, 8, %[digest], t2, t3, t4)
         LOAD_WORD_REV(a5, 12, %[digest], t2, t3, t4)
-        "sw     t0, 0(%[hash])\n\t"
-        "sw     t1, 4(%[hash])\n\t"
-        "sw     a4, 8(%[hash])\n\t"
-        "sw     a5, 12(%[hash])\n\t"
+        UNALIGNED_SW4(t0, t1, a4, a5, 0, %[hash], t2)
         LOAD_WORD_REV(t0, 16, %[digest], t2, t3, t4)
         LOAD_WORD_REV(t1, 20, %[digest], t2, t3, t4)
         LOAD_WORD_REV(a4, 24, %[digest], t2, t3, t4)
         LOAD_WORD_REV(a5, 28, %[digest], t2, t3, t4)
-        "sw     t0, 16(%[hash])\n\t"
-        "sw     t1, 20(%[hash])\n\t"
-        "sw     a4, 24(%[hash])\n\t"
-        "sw     a5, 28(%[hash])\n\t"
+        UNALIGNED_SW4(t0, t1, a4, a5, 16, %[hash], t2)
 #endif
         :
         : [digest] "r" (sha256->digest), [hash] "r" (hash)
diff --git a/wolfcrypt/src/port/riscv/riscv-64-sha3.c b/wolfcrypt/src/port/riscv/riscv-64-sha3.c
index 9d2a959afe..3ecb68d065 100644
--- a/wolfcrypt/src/port/riscv/riscv-64-sha3.c
+++ b/wolfcrypt/src/port/riscv/riscv-64-sha3.c
@@ -139,7 +139,7 @@ static const word64 hash_keccak_r[24] =
 
 #endif
 
-void BlockSha3(word64* s)
+WC_OMIT_FRAME_POINTER void BlockSha3(word64* s)
 {
     const word64* r = hash_keccak_r;
 
diff --git a/wolfcrypt/src/port/riscv/riscv-64-sha512.c b/wolfcrypt/src/port/riscv/riscv-64-sha512.c
index 473b115dc0..8a2fd23e26 100644
--- a/wolfcrypt/src/port/riscv/riscv-64-sha512.c
+++ b/wolfcrypt/src/port/riscv/riscv-64-sha512.c
@@ -554,14 +554,7 @@ static WC_INLINE void Sha512Transform(wc_Sha512* sha512, const byte* data,
         LOAD_DWORD_REV(s6, 48, %[data], a4, a5, a6, a7)
         LOAD_DWORD_REV(s7, 56, %[data], a4, a5, a6, a7)
 #else
-        "ld     t4,  0(%[data])\n\t"
-        "ld     s1,  8(%[data])\n\t"
-        "ld     s2, 16(%[data])\n\t"
-        "ld     s3, 24(%[data])\n\t"
-        "ld     s4, 32(%[data])\n\t"
-        "ld     s5, 40(%[data])\n\t"
-        "ld     s6, 48(%[data])\n\t"
-        "ld     s7, 56(%[data])\n\t"
+        UNALIGNED_LD8(t4, s1, s2, s3, s4, s5, s6, s7, 0, %[data], t5)
         REV8(REG_T4, REG_T4)
         REV8(REG_S1, REG_S1)
         REV8(REG_S2, REG_S2)
@@ -946,14 +939,7 @@ static WC_INLINE void Sha512Final(wc_Sha512* sha512, byte* hash, int hashLen)
         REV8(REG_S9, REG_S9)
         REV8(REG_S10, REG_S10)
         REV8(REG_S11, REG_S11)
-        "sd     t0, 0(%[hash])\n\t"
-        "sd     t1, 8(%[hash])\n\t"
-        "sd     t2, 16(%[hash])\n\t"
-        "sd     t3, 24(%[hash])\n\t"
-        "sd     s8, 32(%[hash])\n\t"
-        "sd     s9, 40(%[hash])\n\t"
-        "sd     s10, 48(%[hash])\n\t"
-        "sd     s11, 56(%[hash])\n\t"
+        UNALIGNED_SD8(t0, t1, t2, t3, s8, s9, s10, s11, 0, %[hash], t4)
 #else
         LOAD_DWORD_REV(t0,  0, %[digest], a4, a5, a6, a7)
         LOAD_DWORD_REV(t1,  8, %[digest], a4, a5, a6, a7)
@@ -963,14 +949,7 @@ static WC_INLINE void Sha512Final(wc_Sha512* sha512, byte* hash, int hashLen)
         LOAD_DWORD_REV(s9,  40, %[digest], a4, a5, a6, a7)
         LOAD_DWORD_REV(s10,  48, %[digest], a4, a5, a6, a7)
         LOAD_DWORD_REV(s11,  56, %[digest], a4, a5, a6, a7)
-        "sd     t0, 0(%[hash])\n\t"
-        "sd     t1, 8(%[hash])\n\t"
-        "sd     t2, 16(%[hash])\n\t"
-        "sd     t3, 24(%[hash])\n\t"
-        "sd     s8, 32(%[hash])\n\t"
-        "sd     s9, 40(%[hash])\n\t"
-        "sd     s10, 48(%[hash])\n\t"
-        "sd     s11, 56(%[hash])\n\t"
+        UNALIGNED_SD8(t0, t1, t2, t3, s8, s9, s10, s11, 0, %[hash], t4)
 #endif
         :
         : [digest] "r" (sha512->digest), [hash] "r" (hashRes)
diff --git a/wolfcrypt/src/random.c b/wolfcrypt/src/random.c
index 4631734d7c..7b85d481d7 100644
--- a/wolfcrypt/src/random.c
+++ b/wolfcrypt/src/random.c
@@ -1720,11 +1720,11 @@ static int _InitRng(WC_RNG* rng, byte* nonce, word32 nonceSz,
     word32 seedSz = SEED_SZ;
 #else
     word32 seedSz = SEED_SZ + SEED_BLOCK_SZ;
+#endif
     WC_DECLARE_VAR(seed, byte, MAX_SEED_SZ, rng->heap);
 #ifdef WOLFSSL_SMALL_STACK_CACHE
     int drbg_scratch_instantiated = 0;
 #endif
-#endif
 #endif
 
     (void)nonce;
diff --git a/wolfcrypt/src/rsa.c b/wolfcrypt/src/rsa.c
index 5fbbd6bdf8..e265d69cf4 100644
--- a/wolfcrypt/src/rsa.c
+++ b/wolfcrypt/src/rsa.c
@@ -60,14 +60,6 @@ RSA keys can be used to encrypt, decrypt, sign and verify data.
 #include <wolfssl/wolfcrypt/port/nxp/casper_port.h>
 #endif
 
-#if defined(WOLFSSL_USE_SAVE_VECTOR_REGISTERS) && !defined(WOLFSSL_SP_ASM)
-    /* force off unneeded vector register save/restore. */
-    #undef SAVE_VECTOR_REGISTERS
-    #define SAVE_VECTOR_REGISTERS(fail_clause) SAVE_NO_VECTOR_REGISTERS(fail_clause)
-    #undef RESTORE_VECTOR_REGISTERS
-    #define RESTORE_VECTOR_REGISTERS() RESTORE_NO_VECTOR_REGISTERS()
-#endif
-
 /*
  * RSA Build Options:
  *
@@ -868,8 +860,6 @@ int wc_CheckRsaKey(RsaKey* key)
         }
     }
 
-    SAVE_VECTOR_REGISTERS(ret = _svr_ret;);
-
     if (ret == 0) {
         if (INIT_MP_INT_SIZE(tmp, mp_bitsused(&key->n)) != MP_OKAY)
             ret = MP_INIT_E;
@@ -982,8 +972,6 @@ int wc_CheckRsaKey(RsaKey* key)
 
     mp_forcezero(tmp);
 
-    RESTORE_VECTOR_REGISTERS();
-
     if ((rng != NULL) && (rng != key->rng)) {
         wc_FreeRng(rng);
 #ifdef WOLFSSL_SMALL_STACK
@@ -3571,7 +3559,6 @@ static int wc_RsaFunction_ex(const byte* in, word32 inLen, byte* out,
 #ifdef WOLF_CRYPTO_CB_ONLY_RSA
     return NO_VALID_DEVID;
 #else /* !WOLF_CRYPTO_CB_ONLY_RSA */
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
 
 #if !defined(WOLFSSL_RSA_VERIFY_ONLY) && !defined(TEST_UNPAD_CONSTANT_TIME) && \
     !defined(NO_RSA_BOUNDS_CHECK)
@@ -3580,7 +3567,6 @@ static int wc_RsaFunction_ex(const byte* in, word32 inLen, byte* out,
 
         ret = RsaFunctionCheckIn(in, inLen, key, checkSmallCt);
         if (ret != 0) {
-            RESTORE_VECTOR_REGISTERS();
             return ret;
         }
     }
@@ -3592,7 +3578,6 @@ static int wc_RsaFunction_ex(const byte* in, word32 inLen, byte* out,
 
         ret = RsaFunctionCheckIn(in, inLen, key, checkSmallCt);
         if (ret != 0) {
-            RESTORE_VECTOR_REGISTERS();
             return ret;
         }
     }
@@ -3623,8 +3608,6 @@ static int wc_RsaFunction_ex(const byte* in, word32 inLen, byte* out,
         ret = wc_RsaFunctionSync(in, inLen, out, outLen, type, key, rng);
     }
 
-    RESTORE_VECTOR_REGISTERS();
-
     /* handle error */
     if (ret < 0 && ret != WC_NO_ERR_TRACE(WC_PENDING_E)
     #ifdef WC_RSA_NONBLOCK
@@ -4172,11 +4155,9 @@ int wc_RsaPublicEncrypt(const byte* in, word32 inLen, byte* out, word32 outLen,
                                                      RsaKey* key, WC_RNG* rng)
 {
     int ret;
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
     ret = RsaPublicEncryptEx(in, inLen, out, outLen, key,
         RSA_PUBLIC_ENCRYPT, RSA_BLOCK_TYPE_2, WC_RSA_PKCSV15_PAD,
         WC_HASH_TYPE_NONE, WC_MGF1NONE, NULL, 0, 0, rng);
-    RESTORE_VECTOR_REGISTERS();
     return ret;
 }
 
@@ -4188,10 +4169,8 @@ int wc_RsaPublicEncrypt_ex(const byte* in, word32 inLen, byte* out,
                     word32 labelSz)
 {
     int ret;
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
     ret = RsaPublicEncryptEx(in, inLen, out, outLen, key, RSA_PUBLIC_ENCRYPT,
         RSA_BLOCK_TYPE_2, type, hash, mgf, label, labelSz, 0, rng);
-    RESTORE_VECTOR_REGISTERS();
     return ret;
 }
 #endif /* WC_NO_RSA_OAEP */
@@ -4211,11 +4190,9 @@ int wc_RsaPrivateDecryptInline(byte* in, word32 inLen, byte** out, RsaKey* key)
 #else
     rng = NULL;
 #endif
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
     ret = RsaPrivateDecryptEx(in, inLen, in, inLen, out, key,
         RSA_PRIVATE_DECRYPT, RSA_BLOCK_TYPE_2, WC_RSA_PKCSV15_PAD,
         WC_HASH_TYPE_NONE, WC_MGF1NONE, NULL, 0, 0, rng);
-    RESTORE_VECTOR_REGISTERS();
     return ret;
 }
 
@@ -4235,11 +4212,9 @@ int wc_RsaPrivateDecryptInline_ex(byte* in, word32 inLen, byte** out,
 #else
     rng = NULL;
 #endif
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
     ret = RsaPrivateDecryptEx(in, inLen, in, inLen, out, key,
         RSA_PRIVATE_DECRYPT, RSA_BLOCK_TYPE_2, type, hash,
         mgf, label, labelSz, 0, rng);
-    RESTORE_VECTOR_REGISTERS();
     return ret;
 }
 #endif /* WC_NO_RSA_OAEP */
@@ -4258,11 +4233,9 @@ int wc_RsaPrivateDecrypt(const byte* in, word32 inLen, byte* out,
 #else
     rng = NULL;
 #endif
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
     ret = RsaPrivateDecryptEx(in, inLen, out, outLen, NULL, key,
         RSA_PRIVATE_DECRYPT, RSA_BLOCK_TYPE_2, WC_RSA_PKCSV15_PAD,
         WC_HASH_TYPE_NONE, WC_MGF1NONE, NULL, 0, 0, rng);
-    RESTORE_VECTOR_REGISTERS();
     return ret;
 }
 
@@ -4282,11 +4255,9 @@ int wc_RsaPrivateDecrypt_ex(const byte* in, word32 inLen, byte* out,
 #else
     rng = NULL;
 #endif
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
     ret = RsaPrivateDecryptEx(in, inLen, out, outLen, NULL, key,
         RSA_PRIVATE_DECRYPT, RSA_BLOCK_TYPE_2, type, hash, mgf, label,
         labelSz, 0, rng);
-    RESTORE_VECTOR_REGISTERS();
     return ret;
 }
 #endif /* WC_NO_RSA_OAEP || WC_RSA_NO_PADDING */
@@ -4305,11 +4276,9 @@ int wc_RsaSSL_VerifyInline(byte* in, word32 inLen, byte** out, RsaKey* key)
 #else
     rng = NULL;
 #endif
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
     ret = RsaPrivateDecryptEx(in, inLen, in, inLen, out, key,
         RSA_PUBLIC_DECRYPT, RSA_BLOCK_TYPE_1, WC_RSA_PKCSV15_PAD,
         WC_HASH_TYPE_NONE, WC_MGF1NONE, NULL, 0, 0, rng);
-    RESTORE_VECTOR_REGISTERS();
     return ret;
 }
 #endif
@@ -4325,10 +4294,8 @@ int  wc_RsaSSL_Verify_ex(const byte* in, word32 inLen, byte* out, word32 outLen,
                          RsaKey* key, int pad_type)
 {
     int ret;
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
     ret = wc_RsaSSL_Verify_ex2(in, inLen, out, outLen, key, pad_type,
             WC_HASH_TYPE_NONE);
-    RESTORE_VECTOR_REGISTERS();
     return ret;
 }
 
@@ -4348,7 +4315,6 @@ int  wc_RsaSSL_Verify_ex2(const byte* in, word32 inLen, byte* out, word32 outLen
     rng = NULL;
 #endif
 
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
 #ifndef WOLFSSL_PSS_SALT_LEN_DISCOVER
     ret = RsaPrivateDecryptEx(in, inLen, out, outLen, NULL, key,
         RSA_PUBLIC_DECRYPT, RSA_BLOCK_TYPE_1, pad_type,
@@ -4358,7 +4324,6 @@ int  wc_RsaSSL_Verify_ex2(const byte* in, word32 inLen, byte* out, word32 outLen
         RSA_PUBLIC_DECRYPT, RSA_BLOCK_TYPE_1, pad_type,
         hash, wc_hash2mgf(hash), NULL, 0, RSA_PSS_SALT_LEN_DISCOVER, rng);
 #endif
-    RESTORE_VECTOR_REGISTERS();
     return ret;
 }
 #endif
@@ -4416,11 +4381,9 @@ int wc_RsaPSS_VerifyInline_ex(byte* in, word32 inLen, byte** out,
 #else
     rng = NULL;
 #endif
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
     ret = RsaPrivateDecryptEx(in, inLen, in, inLen, out, key,
         RSA_PUBLIC_DECRYPT, RSA_BLOCK_TYPE_1, WC_RSA_PSS_PAD,
         hash, mgf, NULL, 0, saltLen, rng);
-    RESTORE_VECTOR_REGISTERS();
     return ret;
 }
 
@@ -4474,11 +4437,9 @@ int wc_RsaPSS_Verify_ex(const byte* in, word32 inLen, byte* out, word32 outLen,
 #else
     rng = NULL;
 #endif
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
     ret = RsaPrivateDecryptEx(in, inLen, out, outLen, NULL, key,
         RSA_PUBLIC_DECRYPT, RSA_BLOCK_TYPE_1, WC_RSA_PSS_PAD,
         hash, mgf, NULL, 0, saltLen, rng);
-    RESTORE_VECTOR_REGISTERS();
     return ret;
 }
 
@@ -4754,11 +4715,9 @@ int wc_RsaSSL_Sign(const byte* in, word32 inLen, byte* out, word32 outLen,
                                                    RsaKey* key, WC_RNG* rng)
 {
     int ret;
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
     ret = RsaPublicEncryptEx(in, inLen, out, outLen, key,
         RSA_PRIVATE_ENCRYPT, RSA_BLOCK_TYPE_1, WC_RSA_PKCSV15_PAD,
         WC_HASH_TYPE_NONE, WC_MGF1NONE, NULL, 0, 0, rng);
-    RESTORE_VECTOR_REGISTERS();
     return ret;
 }
 
@@ -4805,11 +4764,9 @@ int wc_RsaPSS_Sign_ex(const byte* in, word32 inLen, byte* out, word32 outLen,
                       WC_RNG* rng)
 {
     int ret;
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
     ret = RsaPublicEncryptEx(in, inLen, out, outLen, key,
         RSA_PRIVATE_ENCRYPT, RSA_BLOCK_TYPE_1, WC_RSA_PSS_PAD,
         hash, mgf, NULL, 0, saltLen, rng);
-    RESTORE_VECTOR_REGISTERS();
     return ret;
 }
 #endif
@@ -5350,12 +5307,8 @@ int wc_CheckProbablePrime_ex(const byte* pRaw, word32 pRawSz,
     if (ret == MP_OKAY)
         ret = mp_read_unsigned_bin(e, eRaw, eRawSz);
 
-    if (ret == MP_OKAY)
-        SAVE_VECTOR_REGISTERS(ret = _svr_ret;);
-
     if (ret == 0) {
         ret = _CheckProbablePrime(p, Q, e, nlen, isPrime, rng);
-        RESTORE_VECTOR_REGISTERS();
     }
 
     ret = (ret == MP_OKAY) ? 0 : PRIME_GEN_E;
@@ -5549,8 +5502,6 @@ int wc_MakeRsaKey(RsaKey* key, int size, long e, WC_RNG* rng)
     }
 #endif
 
-    SAVE_VECTOR_REGISTERS(err = _svr_ret;);
-
     /* make p */
     if (err == MP_OKAY) {
     #ifdef WOLFSSL_CHECK_MEM_ZERO
@@ -5591,9 +5542,10 @@ int wc_MakeRsaKey(RsaKey* key, int size, long e, WC_RNG* rng)
             if (err != MP_OKAY || isPrime || i >= failCount)
                 break;
 
-            /* linuxkm: release the kernel for a moment before iterating. */
-            RESTORE_VECTOR_REGISTERS();
-            SAVE_VECTOR_REGISTERS(err = _svr_ret; break;);
+            err = WC_CHECK_FOR_INTR_SIGNALS();
+            if (err != 0)
+                break;
+            WC_RELAX_LONG_LOOP();
         };
     }
 
@@ -5644,6 +5596,12 @@ int wc_MakeRsaKey(RsaKey* key, int size, long e, WC_RNG* rng)
             /* Keep the old retry behavior in non-FIPS build. */
             (void)i;
 #endif
+
+            err = WC_CHECK_FOR_INTR_SIGNALS();
+            if (err != 0)
+                break;
+            WC_RELAX_LONG_LOOP();
+
         } while (err == MP_OKAY && !isPrime && i < failCount);
     }
 
@@ -5769,9 +5727,6 @@ int wc_MakeRsaKey(RsaKey* key, int size, long e, WC_RNG* rng)
     }
 #endif
 
-    if (err != WC_NO_ERR_TRACE(WC_ACCEL_INHIBIT_E))
-        RESTORE_VECTOR_REGISTERS();
-
     /* Last value p - 1. */
     mp_forcezero(tmp1);
     /* Last value q - 1. */
diff --git a/wolfcrypt/src/sakke.c b/wolfcrypt/src/sakke.c
index 4fc1a85b48..a7b64e32a7 100644
--- a/wolfcrypt/src/sakke.c
+++ b/wolfcrypt/src/sakke.c
@@ -37,14 +37,6 @@
 #include <wolfssl/wolfcrypt/sakke.h>
 #include <wolfssl/wolfcrypt/asn_public.h>
 
-#if defined(WOLFSSL_USE_SAVE_VECTOR_REGISTERS) && !defined(WOLFSSL_SP_ASM)
-    /* force off unneeded vector register save/restore. */
-    #undef SAVE_VECTOR_REGISTERS
-    #define SAVE_VECTOR_REGISTERS(fail_clause) SAVE_NO_VECTOR_REGISTERS(fail_clause)
-    #undef RESTORE_VECTOR_REGISTERS
-    #define RESTORE_VECTOR_REGISTERS() RESTORE_NO_VECTOR_REGISTERS()
-#endif
-
 #ifndef WOLFSSL_HAVE_ECC_KEY_GET_PRIV
     /* FIPS build has replaced ecc.h. */
     #define wc_ecc_key_get_priv(key) (&((key)->k))
@@ -1328,13 +1320,11 @@ int wc_GenerateSakkeRskTable(const SakkeKey* key, const ecc_point* rsk,
         err = BAD_FUNC_ARG;
     }
     if (err == 0) {
-        SAVE_VECTOR_REGISTERS(return _svr_ret;);
 #ifdef WOLFSSL_SP_1024
         err = sp_Pairing_gen_precomp_1024(rsk, table, len);
 #else
         err = NOT_COMPILED_IN;
 #endif
-        RESTORE_VECTOR_REGISTERS();
     }
 
     return err;
@@ -2441,8 +2431,6 @@ int wc_ValidateSakkeRsk(SakkeKey* key, const byte* id, word16 idSz,
         err = BAD_FUNC_ARG;
     }
 
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
-
     /* Load elliptic curve parameters */
     if (err == 0) {
         err = sakke_load_params(key);
@@ -2478,8 +2466,6 @@ int wc_ValidateSakkeRsk(SakkeKey* key, const byte* id, word16 idSz,
         *valid = ((err == 0) && (mp_cmp(a, &key->params.g) == MP_EQ));
     }
 
-    RESTORE_VECTOR_REGISTERS();
-
     return err;
 }
 
@@ -2622,6 +2608,22 @@ static int sakke_modexp_loop(SakkeKey* key, mp_int* b, mp_int* e, mp_proj* r,
     mp_int* by = key->tmp.p1->z;
     mp_int* prime = &key->params.prime;
     int i;
+#ifdef WC_NO_GLOBAL_OBJECT_POINTERS
+    static const wc_ptr_t wc_off_on_addr[2] =
+    {
+    #if defined(WC_64BIT_CPU)
+        W64LIT(0x0000000000000000),
+        W64LIT(0xffffffffffffffff)
+    #elif defined(WC_16BIT_CPU)
+        0x0000U,
+        0xffffU
+    #else
+        /* 32 bit */
+        0x00000000U,
+        0xffffffffU
+    #endif
+    };
+#endif
 
 #ifdef WC_NO_CACHE_RESISTANT
     c[0] = r;
@@ -6387,8 +6389,6 @@ int wc_MakeSakkePointI(SakkeKey* key, const byte* id, word16 idSz)
         err = BAD_FUNC_ARG;
     }
 
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
-
     if (err == 0) {
         err = sakke_load_params(key);
     }
@@ -6401,8 +6401,6 @@ int wc_MakeSakkePointI(SakkeKey* key, const byte* id, word16 idSz)
         key->i.idSz = idSz;
     }
 
-    RESTORE_VECTOR_REGISTERS();
-
     return err;
 }
 
@@ -6532,9 +6530,7 @@ int wc_GenerateSakkePointITable(SakkeKey* key, byte* table, word32* len)
 
 #ifdef WOLFSSL_HAVE_SP_ECC
     if (err == 0) {
-        SAVE_VECTOR_REGISTERS(return _svr_ret;);
         err = sp_ecc_gen_table_1024(key->i.i, table, len, key->heap);
-        RESTORE_VECTOR_REGISTERS();
     }
     if (err == 0) {
         key->i.table = table;
@@ -6722,8 +6718,6 @@ int wc_MakeSakkeEncapsulatedSSV(SakkeKey* key, enum wc_HashType hashType,
         err = BAD_STATE_E;
     }
 
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
-
     /* Load parameters */
     if (err == 0) {
         err = sakke_load_params(key);
@@ -6799,8 +6793,6 @@ int wc_MakeSakkeEncapsulatedSSV(SakkeKey* key, enum wc_HashType hashType,
 
     /* Step 6: Output SSV - already encoded in buffer */
 
-    RESTORE_VECTOR_REGISTERS();
-
     return err;
 }
 
@@ -6898,7 +6890,9 @@ int wc_DeriveSakkeSSV(SakkeKey* key, enum wc_HashType hashType, byte* ssv,
     mp_int* ri = NULL;
     byte* wb = NULL;
     byte* test = NULL;
-    byte a[WC_MAX_DIGEST_SIZE] = {0};
+    byte a[WC_MAX_DIGEST_SIZE];
+
+    XMEMSET(a, 0, sizeof(a));
 
     if ((key == NULL) || (ssv == NULL) || (auth == NULL) || (ssvSz == 0)) {
         err = BAD_FUNC_ARG;
@@ -6907,8 +6901,6 @@ int wc_DeriveSakkeSSV(SakkeKey* key, enum wc_HashType hashType, byte* ssv,
         err = BAD_STATE_E;
     }
 
-    SAVE_VECTOR_REGISTERS(return _svr_ret;);
-
     /* Load parameters */
     if (err == 0) {
         err = sakke_load_params(key);
@@ -6979,8 +6971,6 @@ int wc_DeriveSakkeSSV(SakkeKey* key, enum wc_HashType hashType, byte* ssv,
         err = SAKKE_VERIFY_FAIL_E;
     }
 
-    RESTORE_VECTOR_REGISTERS();
-
     return err;
 }
 #endif /* WOLFCRYPT_SAKKE_CLIENT */
diff --git a/wolfcrypt/src/sp_arm32.c b/wolfcrypt/src/sp_arm32.c
index 41c7d9ce1e..dd10cc64ad 100644
--- a/wolfcrypt/src/sp_arm32.c
+++ b/wolfcrypt/src/sp_arm32.c
@@ -47,6 +47,17 @@
 
 #include <wolfssl/wolfcrypt/sp.h>
 
+#if defined(WOLFSSL_USE_SAVE_VECTOR_REGISTERS) && !defined(WOLFSSL_SP_ASM) && \
+        !defined(DEBUG_VECTOR_REGISTER_ACCESS)
+    /* force off unneeded vector register save/restore. */
+    #undef SAVE_VECTOR_REGISTERS
+    #define SAVE_VECTOR_REGISTERS(fail_clause) SAVE_NO_VECTOR_REGISTERS(fail_clause)
+    #undef SAVE_VECTOR_REGISTERS2
+    #define SAVE_VECTOR_REGISTERS2() SAVE_NO_VECTOR_REGISTERS2()
+    #undef RESTORE_VECTOR_REGISTERS
+    #define RESTORE_VECTOR_REGISTERS() RESTORE_NO_VECTOR_REGISTERS()
+#endif
+
 #ifdef __IAR_SYSTEMS_ICC__
 #define __asm__        asm
 #define __volatile__   volatile
@@ -141,10 +152,10 @@ static void sp_2048_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
@@ -199,7 +210,7 @@ static void sp_2048_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -234,7 +245,7 @@ static void sp_2048_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -6583,17 +6594,17 @@ WC_OMIT_FRAME_POINTER static void sp_2048_sqr_32(sp_digit* r, const sp_digit* a)
  */
 static void sp_2048_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint32 x;
+    sp_uint32 b;
 
-    b = a[0];
+    b = (sp_uint32)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**32 */
 
     /* rho = -1/m mod b */
-    *rho = (sp_digit)0 - x;
+    *rho = (sp_digit)((sp_int32)0 - (sp_int32)x);
 }
 
 #ifdef WOLFSSL_SP_SMALL
@@ -12498,10 +12509,10 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -12510,14 +12521,14 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 32);
         for (; i>=0 || c>=4; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 28);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 28;
             }
             else if (c < 4) {
@@ -12525,12 +12536,12 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 4 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
 
@@ -12651,10 +12662,10 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -12663,14 +12674,14 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 32);
         for (; i>=0 || c>=5; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 27);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 27;
             }
             else if (c < 5) {
@@ -12678,12 +12689,12 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 5 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
 
@@ -17364,10 +17375,10 @@ static int sp_2048_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -17376,14 +17387,14 @@ static int sp_2048_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 64);
         for (; i>=0 || c>=3; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 29);
-                n <<= 3;
+                n = (sp_uint32)n << 3;
                 c = 29;
             }
             else if (c < 3) {
@@ -17391,12 +17402,12 @@ static int sp_2048_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 3 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 29) & 0x7);
-                n <<= 3;
+                n = (sp_uint32)n << 3;
                 c -= 3;
             }
 
@@ -17500,10 +17511,10 @@ static int sp_2048_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -17512,14 +17523,14 @@ static int sp_2048_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 64);
         for (; i>=0 || c>=4; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 28);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 28;
             }
             else if (c < 4) {
@@ -17527,12 +17538,12 @@ static int sp_2048_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 4 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
 
@@ -17605,7 +17616,7 @@ int sp_RsaPublic_2048(const byte* in, word32 inLen, const mp_int* em,
 #else
         e[0] = em->dp[0];
         if (em->used > 1) {
-            e[0] |= ((sp_digit)em->dp[1]) << DIGIT_BIT;
+            e[0] |= ((sp_uint32)em->dp[1]) << DIGIT_BIT;
         }
 #endif
         if (e[0] == 0) {
@@ -18087,7 +18098,7 @@ static int sp_2048_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 64; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -18112,7 +18123,7 @@ static int sp_2048_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 64; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -18646,10 +18657,10 @@ static int sp_2048_mod_exp_2_64(sp_digit* r, const sp_digit* e, int bits,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -18658,14 +18669,14 @@ static int sp_2048_mod_exp_2_64(sp_digit* r, const sp_digit* e, int bits,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         sp_2048_lshift_64(r, norm, y);
         for (; i>=0 || c>=5; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 27);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 27;
             }
             else if (c < 5) {
@@ -18673,12 +18684,12 @@ static int sp_2048_mod_exp_2_64(sp_digit* r, const sp_digit* e, int bits,
                 n = e[i--];
                 c = 5 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
 
@@ -18846,10 +18857,10 @@ static void sp_3072_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
@@ -18904,7 +18915,7 @@ static void sp_3072_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -18939,7 +18950,7 @@ static void sp_3072_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -29998,17 +30009,17 @@ WC_OMIT_FRAME_POINTER static void sp_3072_sqr_48(sp_digit* r, const sp_digit* a)
  */
 static void sp_3072_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint32 x;
+    sp_uint32 b;
 
-    b = a[0];
+    b = (sp_uint32)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**32 */
 
     /* rho = -1/m mod b */
-    *rho = (sp_digit)0 - x;
+    *rho = (sp_digit)((sp_int32)0 - (sp_int32)x);
 }
 
 #ifdef WOLFSSL_SP_SMALL
@@ -38353,10 +38364,10 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -38365,14 +38376,14 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 48);
         for (; i>=0 || c>=4; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 28);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 28;
             }
             else if (c < 4) {
@@ -38380,12 +38391,12 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 4 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
 
@@ -38506,10 +38517,10 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -38518,14 +38529,14 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 48);
         for (; i>=0 || c>=5; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 27);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 27;
             }
             else if (c < 5) {
@@ -38533,12 +38544,12 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 5 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
 
@@ -45088,10 +45099,10 @@ static int sp_3072_mod_exp_96(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -45100,14 +45111,14 @@ static int sp_3072_mod_exp_96(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 96);
         for (; i>=0 || c>=3; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 29);
-                n <<= 3;
+                n = (sp_uint32)n << 3;
                 c = 29;
             }
             else if (c < 3) {
@@ -45115,12 +45126,12 @@ static int sp_3072_mod_exp_96(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 3 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 29) & 0x7);
-                n <<= 3;
+                n = (sp_uint32)n << 3;
                 c -= 3;
             }
 
@@ -45224,10 +45235,10 @@ static int sp_3072_mod_exp_96(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -45236,14 +45247,14 @@ static int sp_3072_mod_exp_96(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 96);
         for (; i>=0 || c>=4; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 28);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 28;
             }
             else if (c < 4) {
@@ -45251,12 +45262,12 @@ static int sp_3072_mod_exp_96(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 4 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
 
@@ -45329,7 +45340,7 @@ int sp_RsaPublic_3072(const byte* in, word32 inLen, const mp_int* em,
 #else
         e[0] = em->dp[0];
         if (em->used > 1) {
-            e[0] |= ((sp_digit)em->dp[1]) << DIGIT_BIT;
+            e[0] |= ((sp_uint32)em->dp[1]) << DIGIT_BIT;
         }
 #endif
         if (e[0] == 0) {
@@ -45867,7 +45878,7 @@ static int sp_3072_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 96; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -45892,7 +45903,7 @@ static int sp_3072_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 96; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -46618,10 +46629,10 @@ static int sp_3072_mod_exp_2_96(sp_digit* r, const sp_digit* e, int bits,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -46630,14 +46641,14 @@ static int sp_3072_mod_exp_2_96(sp_digit* r, const sp_digit* e, int bits,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         sp_3072_lshift_96(r, norm, y);
         for (; i>=0 || c>=5; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 27);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 27;
             }
             else if (c < 5) {
@@ -46645,12 +46656,12 @@ static int sp_3072_mod_exp_2_96(sp_digit* r, const sp_digit* e, int bits,
                 n = e[i--];
                 c = 5 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
 
@@ -46818,10 +46829,10 @@ static void sp_4096_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
@@ -46876,7 +46887,7 @@ static void sp_4096_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -46911,7 +46922,7 @@ static void sp_4096_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -48059,17 +48070,17 @@ WC_OMIT_FRAME_POINTER static void sp_4096_sqr_128(sp_digit* r,
  */
 static void sp_4096_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint32 x;
+    sp_uint32 b;
 
-    b = a[0];
+    b = (sp_uint32)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**32 */
 
     /* rho = -1/m mod b */
-    *rho = (sp_digit)0 - x;
+    *rho = (sp_digit)((sp_int32)0 - (sp_int32)x);
 }
 
 #ifdef WOLFSSL_SP_SMALL
@@ -60696,10 +60707,10 @@ static int sp_4096_mod_exp_128(sp_digit* r, const sp_digit* a, const sp_digit* e
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -60708,14 +60719,14 @@ static int sp_4096_mod_exp_128(sp_digit* r, const sp_digit* a, const sp_digit* e
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 128);
         for (; i>=0 || c>=3; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 29);
-                n <<= 3;
+                n = (sp_uint32)n << 3;
                 c = 29;
             }
             else if (c < 3) {
@@ -60723,12 +60734,12 @@ static int sp_4096_mod_exp_128(sp_digit* r, const sp_digit* a, const sp_digit* e
                 n = e[i--];
                 c = 3 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 29) & 0x7);
-                n <<= 3;
+                n = (sp_uint32)n << 3;
                 c -= 3;
             }
 
@@ -60832,10 +60843,10 @@ static int sp_4096_mod_exp_128(sp_digit* r, const sp_digit* a, const sp_digit* e
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -60844,14 +60855,14 @@ static int sp_4096_mod_exp_128(sp_digit* r, const sp_digit* a, const sp_digit* e
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 128);
         for (; i>=0 || c>=4; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 28);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 28;
             }
             else if (c < 4) {
@@ -60859,12 +60870,12 @@ static int sp_4096_mod_exp_128(sp_digit* r, const sp_digit* a, const sp_digit* e
                 n = e[i--];
                 c = 4 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
 
@@ -60937,7 +60948,7 @@ int sp_RsaPublic_4096(const byte* in, word32 inLen, const mp_int* em,
 #else
         e[0] = em->dp[0];
         if (em->used > 1) {
-            e[0] |= ((sp_digit)em->dp[1]) << DIGIT_BIT;
+            e[0] |= ((sp_uint32)em->dp[1]) << DIGIT_BIT;
         }
 #endif
         if (e[0] == 0) {
@@ -61531,7 +61542,7 @@ static int sp_4096_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 128; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -61556,7 +61567,7 @@ static int sp_4096_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 128; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -62474,10 +62485,10 @@ static int sp_4096_mod_exp_2_128(sp_digit* r, const sp_digit* e, int bits,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -62486,14 +62497,14 @@ static int sp_4096_mod_exp_2_128(sp_digit* r, const sp_digit* e, int bits,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         sp_4096_lshift_128(r, norm, y);
         for (; i>=0 || c>=5; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 27);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 27;
             }
             else if (c < 5) {
@@ -62501,12 +62512,12 @@ static int sp_4096_mod_exp_2_128(sp_digit* r, const sp_digit* e, int bits,
                 n = e[i--];
                 c = 5 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
 
@@ -67597,7 +67608,7 @@ static void sp_256_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -67632,7 +67643,7 @@ static void sp_256_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -67698,7 +67709,7 @@ static int sp_256_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 8; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -67723,7 +67734,7 @@ static int sp_256_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 8; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -72601,7 +72612,7 @@ static void sp_256_mont_inv_8(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 8);
     for (i=254; i>=0; i--) {
         sp_256_mont_sqr_8(t, t, p256_mod, p256_mp_mod);
-        if (p256_mod_minus_2[i / 32] & ((sp_digit)1 << (i % 32)))
+        if (p256_mod_minus_2[i / 32] & ((sp_uint32)1 << (i % 32)))
             sp_256_mont_mul_8(t, t, a, p256_mod, p256_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 8);
@@ -75246,7 +75257,7 @@ static int sp_256_ecc_mulmod_fast_8(sp_point_256* r, const sp_point_256* g, cons
         t[15].infinity = 0;
 
         i = 6;
-        n = k[i+1] << 0;
+        n = (sp_uint32)k[i+1] << 0;
         c = 28;
         y = (int)(n >> 28);
     #ifndef WC_NO_CACHE_RESISTANT
@@ -75259,14 +75270,14 @@ static int sp_256_ecc_mulmod_fast_8(sp_point_256* r, const sp_point_256* g, cons
         {
             XMEMCPY(rt, &t[y], sizeof(sp_point_256));
         }
-        n <<= 4;
+        n = (sp_uint32)n << (4);
         for (; i>=0 || c>=4; ) {
             if (c < 4) {
                 n |= k[i--];
                 c += 32;
             }
             y = (n >> 28) & 0xf;
-            n <<= 4;
+            n = (sp_uint32)n << 4;
             c -= 4;
 
             sp_256_proj_point_dbl_8(rt, rt, tmp);
@@ -77932,10 +77943,10 @@ static void sp_256_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
@@ -90478,7 +90489,7 @@ static void sp_384_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -90513,7 +90524,7 @@ static void sp_384_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -90579,7 +90590,7 @@ static int sp_384_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 12; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -90604,7 +90615,7 @@ static int sp_384_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 12; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -91540,7 +91551,7 @@ static void sp_384_mont_inv_12(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 12);
     for (i=382; i>=0; i--) {
         sp_384_mont_sqr_12(t, t, p384_mod, p384_mp_mod);
-        if (p384_mod_minus_2[i / 32] & ((sp_digit)1 << (i % 32)))
+        if (p384_mod_minus_2[i / 32] & ((sp_uint32)1 << (i % 32)))
             sp_384_mont_mul_12(t, t, a, p384_mod, p384_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 12);
@@ -93263,7 +93274,7 @@ static int sp_384_ecc_mulmod_fast_12(sp_point_384* r, const sp_point_384* g, con
         t[15].infinity = 0;
 
         i = 10;
-        n = k[i+1] << 0;
+        n = (sp_uint32)k[i+1] << 0;
         c = 28;
         y = (int)(n >> 28);
     #ifndef WC_NO_CACHE_RESISTANT
@@ -93276,14 +93287,14 @@ static int sp_384_ecc_mulmod_fast_12(sp_point_384* r, const sp_point_384* g, con
         {
             XMEMCPY(rt, &t[y], sizeof(sp_point_384));
         }
-        n <<= 4;
+        n = (sp_uint32)n << (4);
         for (; i>=0 || c>=4; ) {
             if (c < 4) {
                 n |= k[i--];
                 c += 32;
             }
             y = (n >> 28) & 0xf;
-            n <<= 4;
+            n = (sp_uint32)n << 4;
             c -= 4;
 
             sp_384_proj_point_dbl_12(rt, rt, tmp);
@@ -95987,10 +95998,10 @@ static void sp_384_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
@@ -117084,7 +117095,7 @@ static void sp_521_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -117119,7 +117130,7 @@ static void sp_521_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -117185,7 +117196,7 @@ static int sp_521_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 17; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -117210,7 +117221,7 @@ static int sp_521_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 17; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -118769,7 +118780,7 @@ static void sp_521_mont_inv_17(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 17);
     for (i=519; i>=0; i--) {
         sp_521_mont_sqr_17(t, t, p521_mod, p521_mp_mod);
-        if (p521_mod_minus_2[i / 32] & ((sp_digit)1 << (i % 32)))
+        if (p521_mod_minus_2[i / 32] & ((sp_uint32)1 << (i % 32)))
             sp_521_mont_mul_17(t, t, a, p521_mod, p521_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 17);
@@ -119658,7 +119669,7 @@ static void sp_521_mont_div2_17(sp_digit* r, const sp_digit* a, const sp_digit*
 
     (void)m;
 
-    sp_521_rshift1_17(r, r);
+    sp_521_rshift1_17(r, a);
     r[16] |= o << 8;
 }
 
@@ -120388,7 +120399,7 @@ static int sp_521_ecc_mulmod_fast_17(sp_point_521* r, const sp_point_521* g, con
         t[15].infinity = 0;
 
         i = 15;
-        n = k[i+1] << 0;
+        n = (sp_uint32)k[i+1] << 0;
         c = 5;
         y = (int)(n >> 5);
     #ifndef WC_NO_CACHE_RESISTANT
@@ -120401,15 +120412,15 @@ static int sp_521_ecc_mulmod_fast_17(sp_point_521* r, const sp_point_521* g, con
         {
             XMEMCPY(rt, &t[y], sizeof(sp_point_521));
         }
-        n <<= 27;
+        n = (sp_uint32)n << (27);
         for (; i>=0 || c>=4; ) {
             if (c < 4) {
-                n = (k[i+1] << 31) | (k[i] >> 1);
+                n = ((sp_uint32)k[i+1] << 31) | (k[i] >> 1);
                 i--;
                 c += 32;
             }
             y = (n >> 28) & 0xf;
-            n <<= 4;
+            n = (sp_uint32)n << 4;
             c -= 4;
 
             sp_521_proj_point_dbl_17(rt, rt, tmp);
@@ -123719,10 +123730,10 @@ static void sp_521_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
@@ -125606,8 +125617,6 @@ static WC_INLINE int sp_521_div_17(const sp_digit* a, const sp_digit* d,
     sp_digit r1;
     int i;
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)m;
     div = (d[16] << 23) | (d[15] >> 9);
     XMEMCPY(t1, a, sizeof(*t1) * 2 * 17);
@@ -128335,7 +128344,7 @@ static int sp_521_mont_sqrt_17(sp_digit* y)
             XMEMCPY(t, y, sizeof(sp_digit) * 17);
             for (i=518; i>=0; i--) {
                 sp_521_mont_sqr_17(t, t, p521_mod, p521_mp_mod);
-                if (p521_sqrt_power[i / 32] & ((sp_digit)1 << (i % 32)))
+                if (p521_sqrt_power[i / 32] & ((sp_uint32)1 << (i % 32)))
                     sp_521_mont_mul_17(t, t, y, p521_mod, p521_mp_mod);
             }
             XMEMCPY(y, t, sizeof(sp_digit) * 17);
@@ -146613,7 +146622,7 @@ static void sp_1024_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -146648,7 +146657,7 @@ static void sp_1024_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -146714,7 +146723,7 @@ static int sp_1024_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 32; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -146739,7 +146748,7 @@ static int sp_1024_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 32; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -150295,18 +150304,18 @@ static int sp_1024_ecc_mulmod_fast_32(sp_point_1024* r, const sp_point_1024* g,
         t[15].infinity = 0;
 
         i = 30;
-        n = k[i+1] << 0;
+        n = (sp_uint32)k[i+1] << 0;
         c = 28;
         y = (int)(n >> 28);
         XMEMCPY(rt, &t[y], sizeof(sp_point_1024));
-        n <<= 4;
+        n = (sp_uint32)n << (4);
         for (; i>=0 || c>=4; ) {
             if (c < 4) {
                 n |= k[i--];
                 c += 32;
             }
             y = (n >> 28) & 0xf;
-            n <<= 4;
+            n = (sp_uint32)n << 4;
             c -= 4;
 
             sp_1024_proj_point_dbl_32(rt, rt, tmp);
@@ -158411,10 +158420,10 @@ static void sp_1024_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
diff --git a/wolfcrypt/src/sp_arm64.c b/wolfcrypt/src/sp_arm64.c
index 402e75a6db..59751ebe0c 100644
--- a/wolfcrypt/src/sp_arm64.c
+++ b/wolfcrypt/src/sp_arm64.c
@@ -47,6 +47,17 @@
 
 #include <wolfssl/wolfcrypt/sp.h>
 
+#if defined(WOLFSSL_USE_SAVE_VECTOR_REGISTERS) && !defined(WOLFSSL_SP_ASM) && \
+        !defined(DEBUG_VECTOR_REGISTER_ACCESS)
+    /* force off unneeded vector register save/restore. */
+    #undef SAVE_VECTOR_REGISTERS
+    #define SAVE_VECTOR_REGISTERS(fail_clause) SAVE_NO_VECTOR_REGISTERS(fail_clause)
+    #undef SAVE_VECTOR_REGISTERS2
+    #define SAVE_VECTOR_REGISTERS2() SAVE_NO_VECTOR_REGISTERS2()
+    #undef RESTORE_VECTOR_REGISTERS
+    #define RESTORE_VECTOR_REGISTERS() RESTORE_NO_VECTOR_REGISTERS()
+#endif
+
 #ifdef __IAR_SYSTEMS_ICC__
 #define __asm__        asm
 #define __volatile__   volatile
@@ -265,7 +276,7 @@ static void sp_2048_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0xffffffffffffffffl;
         s = 64U - s;
         if (j + 1 >= size) {
@@ -300,7 +311,7 @@ static void sp_2048_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 64) {
             r[j] &= 0xffffffffffffffffl;
             if (j + 1 >= size) {
@@ -2911,10 +2922,10 @@ static void sp_2048_sqr_16(sp_digit* r, const sp_digit* a)
  */
 static void sp_2048_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint64 x;
+    sp_uint64 b;
 
-    b = a[0];
+    b = (sp_uint64)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
@@ -2922,7 +2933,7 @@ static void sp_2048_mont_setup(const sp_digit* a, sp_digit* rho)
     x *= 2 - b * x;               /* here x*a==1 mod 2**64 */
 
     /* rho = -1/m mod b */
-    *rho = (sp_digit)0 - x;
+    *rho = (sp_digit)((sp_int64)0 - (sp_int64)x);
 }
 
 /* Mul a by digit b into r. (r = a * b)
@@ -4151,10 +4162,10 @@ static int sp_2048_mod_exp_16(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint64)n << c);
             n = e[i--];
             y |= (byte)(n >> (64 - c));
-            n <<= c;
+            n = (sp_uint64)n << c;
             c = 64 - c;
         }
         else if (c == 0) {
@@ -4163,14 +4174,14 @@ static int sp_2048_mod_exp_16(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 64 - c;
+            n = (sp_uint64)n << (64 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 16);
         for (; i>=0 || c>=4; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 60);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c = 60;
             }
             else if (c < 4) {
@@ -4178,12 +4189,12 @@ static int sp_2048_mod_exp_16(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 4 - c;
                 y |= (byte)(n >> (64 - c));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 64 - c;
             }
             else {
                 y = (byte)((n >> 60) & 0xf);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c -= 4;
             }
 
@@ -4304,10 +4315,10 @@ static int sp_2048_mod_exp_16(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint64)n << c);
             n = e[i--];
             y |= (byte)(n >> (64 - c));
-            n <<= c;
+            n = (sp_uint64)n << c;
             c = 64 - c;
         }
         else if (c == 0) {
@@ -4316,14 +4327,14 @@ static int sp_2048_mod_exp_16(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 64 - c;
+            n = (sp_uint64)n << (64 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 16);
         for (; i>=0 || c>=5; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 59);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c = 59;
             }
             else if (c < 5) {
@@ -4331,12 +4342,12 @@ static int sp_2048_mod_exp_16(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 5 - c;
                 y |= (byte)(n >> (64 - c));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 64 - c;
             }
             else {
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c -= 5;
             }
 
@@ -5746,10 +5757,10 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint64)n << c);
             n = e[i--];
             y |= (byte)(n >> (64 - c));
-            n <<= c;
+            n = (sp_uint64)n << c;
             c = 64 - c;
         }
         else if (c == 0) {
@@ -5758,14 +5769,14 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 64 - c;
+            n = (sp_uint64)n << (64 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 32);
         for (; i>=0 || c>=5; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 59);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c = 59;
             }
             else if (c < 5) {
@@ -5773,12 +5784,12 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 5 - c;
                 y |= (byte)(n >> (64 - c));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 64 - c;
             }
             else {
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c -= 5;
             }
 
@@ -5932,10 +5943,10 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint64)n << c);
             n = e[i--];
             y |= (byte)(n >> (64 - c));
-            n <<= c;
+            n = (sp_uint64)n << c;
             c = 64 - c;
         }
         else if (c == 0) {
@@ -5944,14 +5955,14 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 64 - c;
+            n = (sp_uint64)n << (64 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 32);
         for (; i>=0 || c>=6; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 58);
-                n <<= 6;
+                n = (sp_uint64)n << 6;
                 c = 58;
             }
             else if (c < 6) {
@@ -5959,12 +5970,12 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 6 - c;
                 y |= (byte)(n >> (64 - c));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 64 - c;
             }
             else {
                 y = (byte)((n >> 58) & 0x3f);
-                n <<= 6;
+                n = (sp_uint64)n << 6;
                 c -= 6;
             }
 
@@ -6039,7 +6050,7 @@ int sp_RsaPublic_2048(const byte* in, word32 inLen, const mp_int* em,
 #else
         e[0] = em->dp[0];
         if (em->used > 1) {
-            e[0] |= ((sp_digit)em->dp[1]) << DIGIT_BIT;
+            e[0] |= ((sp_uint64)em->dp[1]) << DIGIT_BIT;
         }
 #endif
         if (e[0] == 0) {
@@ -6352,7 +6363,7 @@ static int sp_2048_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 32; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -6377,7 +6388,7 @@ static int sp_2048_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 32; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 64 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -6703,10 +6714,10 @@ static int sp_2048_mod_exp_2_32(sp_digit* r, const sp_digit* e, int bits,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint64)n << c);
             n = e[i--];
             y |= (byte)(n >> (64 - c));
-            n <<= c;
+            n = (sp_uint64)n << c;
             c = 64 - c;
         }
         else if (c == 0) {
@@ -6715,14 +6726,14 @@ static int sp_2048_mod_exp_2_32(sp_digit* r, const sp_digit* e, int bits,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 64 - c;
+            n = (sp_uint64)n << (64 - c);
         }
         sp_2048_lshift_32(r, norm, y);
         for (; i>=0 || c>=6; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 58);
-                n <<= 6;
+                n = (sp_uint64)n << 6;
                 c = 58;
             }
             else if (c < 6) {
@@ -6730,12 +6741,12 @@ static int sp_2048_mod_exp_2_32(sp_digit* r, const sp_digit* e, int bits,
                 n = e[i--];
                 c = 6 - c;
                 y |= (byte)(n >> (64 - c));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 64 - c;
             }
             else {
                 y = (byte)((n >> 58) & 0x3f);
-                n <<= 6;
+                n = (sp_uint64)n << 6;
                 c -= 6;
             }
 
@@ -7028,7 +7039,7 @@ static void sp_3072_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0xffffffffffffffffl;
         s = 64U - s;
         if (j + 1 >= size) {
@@ -7063,7 +7074,7 @@ static void sp_3072_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 64) {
             r[j] &= 0xffffffffffffffffl;
             if (j + 1 >= size) {
@@ -11584,10 +11595,10 @@ static void sp_3072_sqr_24(sp_digit* r, const sp_digit* a)
  */
 static void sp_3072_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint64 x;
+    sp_uint64 b;
 
-    b = a[0];
+    b = (sp_uint64)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
@@ -11595,7 +11606,7 @@ static void sp_3072_mont_setup(const sp_digit* a, sp_digit* rho)
     x *= 2 - b * x;               /* here x*a==1 mod 2**64 */
 
     /* rho = -1/m mod b */
-    *rho = (sp_digit)0 - x;
+    *rho = (sp_digit)((sp_int64)0 - (sp_int64)x);
 }
 
 /* Mul a by digit b into r. (r = a * b)
@@ -13212,10 +13223,10 @@ static int sp_3072_mod_exp_24(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint64)n << c);
             n = e[i--];
             y |= (byte)(n >> (64 - c));
-            n <<= c;
+            n = (sp_uint64)n << c;
             c = 64 - c;
         }
         else if (c == 0) {
@@ -13224,14 +13235,14 @@ static int sp_3072_mod_exp_24(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 64 - c;
+            n = (sp_uint64)n << (64 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 24);
         for (; i>=0 || c>=4; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 60);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c = 60;
             }
             else if (c < 4) {
@@ -13239,12 +13250,12 @@ static int sp_3072_mod_exp_24(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 4 - c;
                 y |= (byte)(n >> (64 - c));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 64 - c;
             }
             else {
                 y = (byte)((n >> 60) & 0xf);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c -= 4;
             }
 
@@ -13365,10 +13376,10 @@ static int sp_3072_mod_exp_24(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint64)n << c);
             n = e[i--];
             y |= (byte)(n >> (64 - c));
-            n <<= c;
+            n = (sp_uint64)n << c;
             c = 64 - c;
         }
         else if (c == 0) {
@@ -13377,14 +13388,14 @@ static int sp_3072_mod_exp_24(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 64 - c;
+            n = (sp_uint64)n << (64 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 24);
         for (; i>=0 || c>=5; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 59);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c = 59;
             }
             else if (c < 5) {
@@ -13392,12 +13403,12 @@ static int sp_3072_mod_exp_24(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 5 - c;
                 y |= (byte)(n >> (64 - c));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 64 - c;
             }
             else {
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c -= 5;
             }
 
@@ -15175,10 +15186,10 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint64)n << c);
             n = e[i--];
             y |= (byte)(n >> (64 - c));
-            n <<= c;
+            n = (sp_uint64)n << c;
             c = 64 - c;
         }
         else if (c == 0) {
@@ -15187,14 +15198,14 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 64 - c;
+            n = (sp_uint64)n << (64 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 48);
         for (; i>=0 || c>=3; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 61);
-                n <<= 3;
+                n = (sp_uint64)n << 3;
                 c = 61;
             }
             else if (c < 3) {
@@ -15202,12 +15213,12 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 3 - c;
                 y |= (byte)(n >> (64 - c));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 64 - c;
             }
             else {
                 y = (byte)((n >> 61) & 0x7);
-                n <<= 3;
+                n = (sp_uint64)n << 3;
                 c -= 3;
             }
 
@@ -15311,10 +15322,10 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint64)n << c);
             n = e[i--];
             y |= (byte)(n >> (64 - c));
-            n <<= c;
+            n = (sp_uint64)n << c;
             c = 64 - c;
         }
         else if (c == 0) {
@@ -15323,14 +15334,14 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 64 - c;
+            n = (sp_uint64)n << (64 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 48);
         for (; i>=0 || c>=4; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 60);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c = 60;
             }
             else if (c < 4) {
@@ -15338,12 +15349,12 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 4 - c;
                 y |= (byte)(n >> (64 - c));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 64 - c;
             }
             else {
                 y = (byte)((n >> 60) & 0xf);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c -= 4;
             }
 
@@ -15416,7 +15427,7 @@ int sp_RsaPublic_3072(const byte* in, word32 inLen, const mp_int* em,
 #else
         e[0] = em->dp[0];
         if (em->used > 1) {
-            e[0] |= ((sp_digit)em->dp[1]) << DIGIT_BIT;
+            e[0] |= ((sp_uint64)em->dp[1]) << DIGIT_BIT;
         }
 #endif
         if (e[0] == 0) {
@@ -15729,7 +15740,7 @@ static int sp_3072_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 48; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -15754,7 +15765,7 @@ static int sp_3072_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 48; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 64 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -16176,10 +16187,10 @@ static int sp_3072_mod_exp_2_48(sp_digit* r, const sp_digit* e, int bits,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint64)n << c);
             n = e[i--];
             y |= (byte)(n >> (64 - c));
-            n <<= c;
+            n = (sp_uint64)n << c;
             c = 64 - c;
         }
         else if (c == 0) {
@@ -16188,14 +16199,14 @@ static int sp_3072_mod_exp_2_48(sp_digit* r, const sp_digit* e, int bits,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 64 - c;
+            n = (sp_uint64)n << (64 - c);
         }
         sp_3072_lshift_48(r, norm, y);
         for (; i>=0 || c>=6; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 58);
-                n <<= 6;
+                n = (sp_uint64)n << 6;
                 c = 58;
             }
             else if (c < 6) {
@@ -16203,12 +16214,12 @@ static int sp_3072_mod_exp_2_48(sp_digit* r, const sp_digit* e, int bits,
                 n = e[i--];
                 c = 6 - c;
                 y |= (byte)(n >> (64 - c));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 64 - c;
             }
             else {
                 y = (byte)((n >> 58) & 0x3f);
-                n <<= 6;
+                n = (sp_uint64)n << 6;
                 c -= 6;
             }
 
@@ -16501,7 +16512,7 @@ static void sp_4096_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0xffffffffffffffffl;
         s = 64U - s;
         if (j + 1 >= size) {
@@ -16536,7 +16547,7 @@ static void sp_4096_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 64) {
             r[j] &= 0xffffffffffffffffl;
             if (j + 1 >= size) {
@@ -17444,10 +17455,10 @@ static void sp_4096_sqr_64(sp_digit* r, const sp_digit* a)
  */
 static void sp_4096_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint64 x;
+    sp_uint64 b;
 
-    b = a[0];
+    b = (sp_uint64)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
@@ -17455,7 +17466,7 @@ static void sp_4096_mont_setup(const sp_digit* a, sp_digit* rho)
     x *= 2 - b * x;               /* here x*a==1 mod 2**64 */
 
     /* rho = -1/m mod b */
-    *rho = (sp_digit)0 - x;
+    *rho = (sp_digit)((sp_int64)0 - (sp_int64)x);
 }
 
 /* Mul a by digit b into r. (r = a * b)
@@ -20186,10 +20197,10 @@ static int sp_4096_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint64)n << c);
             n = e[i--];
             y |= (byte)(n >> (64 - c));
-            n <<= c;
+            n = (sp_uint64)n << c;
             c = 64 - c;
         }
         else if (c == 0) {
@@ -20198,14 +20209,14 @@ static int sp_4096_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 64 - c;
+            n = (sp_uint64)n << (64 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 64);
         for (; i>=0 || c>=3; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 61);
-                n <<= 3;
+                n = (sp_uint64)n << 3;
                 c = 61;
             }
             else if (c < 3) {
@@ -20213,12 +20224,12 @@ static int sp_4096_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 3 - c;
                 y |= (byte)(n >> (64 - c));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 64 - c;
             }
             else {
                 y = (byte)((n >> 61) & 0x7);
-                n <<= 3;
+                n = (sp_uint64)n << 3;
                 c -= 3;
             }
 
@@ -20322,10 +20333,10 @@ static int sp_4096_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint64)n << c);
             n = e[i--];
             y |= (byte)(n >> (64 - c));
-            n <<= c;
+            n = (sp_uint64)n << c;
             c = 64 - c;
         }
         else if (c == 0) {
@@ -20334,14 +20345,14 @@ static int sp_4096_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 64 - c;
+            n = (sp_uint64)n << (64 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 64);
         for (; i>=0 || c>=4; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 60);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c = 60;
             }
             else if (c < 4) {
@@ -20349,12 +20360,12 @@ static int sp_4096_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 4 - c;
                 y |= (byte)(n >> (64 - c));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 64 - c;
             }
             else {
                 y = (byte)((n >> 60) & 0xf);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c -= 4;
             }
 
@@ -20427,7 +20438,7 @@ int sp_RsaPublic_4096(const byte* in, word32 inLen, const mp_int* em,
 #else
         e[0] = em->dp[0];
         if (em->used > 1) {
-            e[0] |= ((sp_digit)em->dp[1]) << DIGIT_BIT;
+            e[0] |= ((sp_uint64)em->dp[1]) << DIGIT_BIT;
         }
 #endif
         if (e[0] == 0) {
@@ -20740,7 +20751,7 @@ static int sp_4096_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 64; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -20765,7 +20776,7 @@ static int sp_4096_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 64; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 64 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -21283,10 +21294,10 @@ static int sp_4096_mod_exp_2_64(sp_digit* r, const sp_digit* e, int bits,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint64)n << c);
             n = e[i--];
             y |= (byte)(n >> (64 - c));
-            n <<= c;
+            n = (sp_uint64)n << c;
             c = 64 - c;
         }
         else if (c == 0) {
@@ -21295,14 +21306,14 @@ static int sp_4096_mod_exp_2_64(sp_digit* r, const sp_digit* e, int bits,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 64 - c;
+            n = (sp_uint64)n << (64 - c);
         }
         sp_4096_lshift_64(r, norm, y);
         for (; i>=0 || c>=6; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 58);
-                n <<= 6;
+                n = (sp_uint64)n << 6;
                 c = 58;
             }
             else if (c < 6) {
@@ -21310,12 +21321,12 @@ static int sp_4096_mod_exp_2_64(sp_digit* r, const sp_digit* e, int bits,
                 n = e[i--];
                 c = 6 - c;
                 y |= (byte)(n >> (64 - c));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 64 - c;
             }
             else {
                 y = (byte)((n >> 58) & 0x3f);
-                n <<= 6;
+                n = (sp_uint64)n << 6;
                 c -= 6;
             }
 
@@ -21865,10 +21876,10 @@ static int sp_256_mod_mul_norm_4(sp_digit* r, const sp_digit* a, const sp_digit*
     t[5] += t[4] >> 32; t[4] &= 0xffffffff;
     t[6] += t[5] >> 32; t[5] &= 0xffffffff;
     t[7] += t[6] >> 32; t[6] &= 0xffffffff;
-    r[0] = (sp_digit)((t[1] << 32) | t[0]);
-    r[1] = (sp_digit)((t[3] << 32) | t[2]);
-    r[2] = (sp_digit)((t[5] << 32) | t[4]);
-    r[3] = (sp_digit)((t[7] << 32) | t[6]);
+    r[0] = (sp_digit)(((sp_uint64)t[1] << 32) | (sp_uint64)t[0]);
+    r[1] = (sp_digit)(((sp_uint64)t[3] << 32) | (sp_uint64)t[2]);
+    r[2] = (sp_digit)(((sp_uint64)t[5] << 32) | (sp_uint64)t[4]);
+    r[3] = (sp_digit)(((sp_uint64)t[7] << 32) | (sp_uint64)t[6]);
 
     return MP_OKAY;
 }
@@ -21899,7 +21910,7 @@ static void sp_256_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0xffffffffffffffffl;
         s = 64U - s;
         if (j + 1 >= size) {
@@ -21934,7 +21945,7 @@ static void sp_256_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 64) {
             r[j] &= 0xffffffffffffffffl;
             if (j + 1 >= size) {
@@ -22000,7 +22011,7 @@ static int sp_256_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 4; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -22025,7 +22036,7 @@ static int sp_256_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 4; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 64 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -22465,7 +22476,7 @@ static void sp_256_mont_inv_4(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 4);
     for (i=254; i>=0; i--) {
         sp_256_mont_sqr_4(t, t, p256_mod, p256_mp_mod);
-        if (p256_mod_minus_2[i / 64] & ((sp_digit)1 << (i % 64)))
+        if (p256_mod_minus_2[i / 64] & ((sp_uint64)1 << (i % 64)))
             sp_256_mont_mul_4(t, t, a, p256_mod, p256_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 4);
@@ -23959,7 +23970,7 @@ static void sp_256_ecc_recode_6_4(const sp_digit* k, ecc_recode_256* v)
         }
         else if (++j < 4) {
             n = k[j];
-            y |= (word8)((n << (64 - o)) & 0x3f);
+            y |= (word8)(((sp_uint64)n << (64 - o)) & 0x3f);
             o -= 58;
             n >>= o;
         }
@@ -26903,7 +26914,7 @@ static void sp_256_ecc_recode_7_4(const sp_digit* k, ecc_recode_256* v)
         }
         else if (++j < 4) {
             n = k[j];
-            y |= (word8)((n << (64 - o)) & 0x7f);
+            y |= (word8)(((sp_uint64)n << (64 - o)) & 0x7f);
             o -= 57;
             n >>= o;
         }
@@ -42693,12 +42704,12 @@ static int sp_384_mod_mul_norm_6(sp_digit* r, const sp_digit* a, const sp_digit*
         t[10] += t[9] >> 32; t[9] &= 0xffffffff;
         t[11] += t[10] >> 32; t[10] &= 0xffffffff;
 
-        r[0] = (sp_digit)((t[1] << 32) | t[0]);
-        r[1] = (sp_digit)((t[3] << 32) | t[2]);
-        r[2] = (sp_digit)((t[5] << 32) | t[4]);
-        r[3] = (sp_digit)((t[7] << 32) | t[6]);
-        r[4] = (sp_digit)((t[9] << 32) | t[8]);
-        r[5] = (sp_digit)((t[11] << 32) | t[10]);
+        r[0] = (sp_digit)(((sp_uint64)t[1] << 32) | (sp_uint64)t[0]);
+        r[1] = (sp_digit)(((sp_uint64)t[3] << 32) | (sp_uint64)t[2]);
+        r[2] = (sp_digit)(((sp_uint64)t[5] << 32) | (sp_uint64)t[4]);
+        r[3] = (sp_digit)(((sp_uint64)t[7] << 32) | (sp_uint64)t[6]);
+        r[4] = (sp_digit)(((sp_uint64)t[9] << 32) | (sp_uint64)t[8]);
+        r[5] = (sp_digit)(((sp_uint64)t[11] << 32) | (sp_uint64)t[10]);
     }
 
     SP_FREE_VAR(t, NULL, DYNAMIC_TYPE_ECC);
@@ -42732,7 +42743,7 @@ static void sp_384_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0xffffffffffffffffl;
         s = 64U - s;
         if (j + 1 >= size) {
@@ -42767,7 +42778,7 @@ static void sp_384_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 64) {
             r[j] &= 0xffffffffffffffffl;
             if (j + 1 >= size) {
@@ -42833,7 +42844,7 @@ static int sp_384_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 6; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -42858,7 +42869,7 @@ static int sp_384_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 6; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 64 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -43247,7 +43258,7 @@ static void sp_384_mont_inv_6(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 6);
     for (i=382; i>=0; i--) {
         sp_384_mont_sqr_6(t, t, p384_mod, p384_mp_mod);
-        if (p384_mod_minus_2[i / 64] & ((sp_digit)1 << (i % 64)))
+        if (p384_mod_minus_2[i / 64] & ((sp_uint64)1 << (i % 64)))
             sp_384_mont_mul_6(t, t, a, p384_mod, p384_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 6);
@@ -44497,7 +44508,7 @@ static void sp_384_ecc_recode_6_6(const sp_digit* k, ecc_recode_384* v)
         }
         else if (++j < 6) {
             n = k[j];
-            y |= (word8)((n << (64 - o)) & 0x3f);
+            y |= (word8)(((sp_uint64)n << (64 - o)) & 0x3f);
             o -= 58;
             n >>= o;
         }
@@ -47405,7 +47416,7 @@ static void sp_384_ecc_recode_7_6(const sp_digit* k, ecc_recode_384* v)
         }
         else if (++j < 6) {
             n = k[j];
-            y |= (word8)((n << (64 - o)) & 0x7f);
+            y |= (word8)(((sp_uint64)n << (64 - o)) & 0x7f);
             o -= 57;
             n >>= o;
         }
@@ -69631,8 +69642,6 @@ static WC_INLINE int sp_521_div_9(const sp_digit* a, const sp_digit* d,
     sp_digit r1;
     int i;
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)m;
     div = (d[8] << 55) | (d[7] >> 9);
     XMEMCPY(t1, a, sizeof(*t1) * 2 * 9);
@@ -69713,7 +69722,7 @@ static void sp_521_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0xffffffffffffffffl;
         s = 64U - s;
         if (j + 1 >= size) {
@@ -69748,7 +69757,7 @@ static void sp_521_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 64) {
             r[j] &= 0xffffffffffffffffl;
             if (j + 1 >= size) {
@@ -69814,7 +69823,7 @@ static int sp_521_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 9; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -69839,7 +69848,7 @@ static int sp_521_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 9; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 64 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -71004,7 +71013,7 @@ static void sp_521_mont_inv_9(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 9);
     for (i=519; i>=0; i--) {
         sp_521_mont_sqr_9(t, t, p521_mod, p521_mp_mod);
-        if (p521_mod_minus_2[i / 64] & ((sp_digit)1 << (i % 64)))
+        if (p521_mod_minus_2[i / 64] & ((sp_uint64)1 << (i % 64)))
             sp_521_mont_mul_9(t, t, a, p521_mod, p521_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 9);
@@ -72442,7 +72451,7 @@ static void sp_521_ecc_recode_6_9(const sp_digit* k, ecc_recode_521* v)
         }
         else if (++j < 9) {
             n = k[j];
-            y |= (word8)((n << (64 - o)) & 0x3f);
+            y |= (word8)(((sp_uint64)n << (64 - o)) & 0x3f);
             o -= 58;
             n >>= o;
         }
@@ -76049,7 +76058,7 @@ static void sp_521_ecc_recode_7_9(const sp_digit* k, ecc_recode_521* v)
         }
         else if (++j < 9) {
             n = k[j];
-            y |= (word8)((n << (64 - o)) & 0x7f);
+            y |= (word8)(((sp_uint64)n << (64 - o)) & 0x7f);
             o -= 57;
             n >>= o;
         }
@@ -111994,7 +112003,7 @@ static int sp_521_mont_sqrt_9(sp_digit* y)
             XMEMCPY(t, y, sizeof(sp_digit) * 9);
             for (i=518; i>=0; i--) {
                 sp_521_mont_sqr_9(t, t, p521_mod, p521_mp_mod);
-                if (p521_sqrt_power[i / 64] & ((sp_digit)1 << (i % 64)))
+                if (p521_sqrt_power[i / 64] & ((sp_uint64)1 << (i % 64)))
                     sp_521_mont_mul_9(t, t, y, p521_mod, p521_mp_mod);
             }
             XMEMCPY(y, t, sizeof(sp_digit) * 9);
@@ -114032,7 +114041,7 @@ static void sp_1024_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0xffffffffffffffffl;
         s = 64U - s;
         if (j + 1 >= size) {
@@ -114067,7 +114076,7 @@ static void sp_1024_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 64) {
             r[j] &= 0xffffffffffffffffl;
             if (j + 1 >= size) {
@@ -114133,7 +114142,7 @@ static int sp_1024_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 16; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -114158,7 +114167,7 @@ static int sp_1024_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 16; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 64 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -116114,7 +116123,7 @@ static void sp_1024_ecc_recode_7_16(const sp_digit* k, ecc_recode_1024* v)
         }
         else if (++j < 16) {
             n = k[j];
-            y |= (word8)((n << (64 - o)) & 0x7f);
+            y |= (word8)(((sp_uint64)n << (64 - o)) & 0x7f);
             o -= 57;
             n >>= o;
         }
diff --git a/wolfcrypt/src/sp_armthumb.c b/wolfcrypt/src/sp_armthumb.c
index b479c93b9b..f56e1d9091 100644
--- a/wolfcrypt/src/sp_armthumb.c
+++ b/wolfcrypt/src/sp_armthumb.c
@@ -47,6 +47,17 @@
 
 #include <wolfssl/wolfcrypt/sp.h>
 
+#if defined(WOLFSSL_USE_SAVE_VECTOR_REGISTERS) && !defined(WOLFSSL_SP_ASM) && \
+        !defined(DEBUG_VECTOR_REGISTER_ACCESS)
+    /* force off unneeded vector register save/restore. */
+    #undef SAVE_VECTOR_REGISTERS
+    #define SAVE_VECTOR_REGISTERS(fail_clause) SAVE_NO_VECTOR_REGISTERS(fail_clause)
+    #undef SAVE_VECTOR_REGISTERS2
+    #define SAVE_VECTOR_REGISTERS2() SAVE_NO_VECTOR_REGISTERS2()
+    #undef RESTORE_VECTOR_REGISTERS
+    #define RESTORE_VECTOR_REGISTERS() RESTORE_NO_VECTOR_REGISTERS()
+#endif
+
 #ifdef __IAR_SYSTEMS_ICC__
 #define __asm__        asm
 #define __volatile__   volatile
@@ -141,10 +152,10 @@ static void sp_2048_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
@@ -199,7 +210,7 @@ static void sp_2048_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -234,7 +245,7 @@ static void sp_2048_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -21740,17 +21751,17 @@ WC_OMIT_FRAME_POINTER static SP_NOINLINE void sp_2048_sqr_32(sp_digit* r,
  */
 static void sp_2048_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint32 x;
+    sp_uint32 b;
 
-    b = a[0];
+    b = (sp_uint32)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**32 */
 
     /* rho = -1/m mod b */
-    *rho = (sp_digit)0 - x;
+    *rho = (sp_digit)((sp_int32)0 - (sp_int32)x);
 }
 
 /* Mul a by digit b into r. (r = a * b)
@@ -24177,10 +24188,10 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -24189,14 +24200,14 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 32);
         for (; i>=0 || c>=4; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 28);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 28;
             }
             else if (c < 4) {
@@ -24204,12 +24215,12 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 4 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
 
@@ -24330,10 +24341,10 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -24342,14 +24353,14 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 32);
         for (; i>=0 || c>=5; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 27);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 27;
             }
             else if (c < 5) {
@@ -24357,12 +24368,12 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 5 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
 
@@ -27737,10 +27748,10 @@ static int sp_2048_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -27749,14 +27760,14 @@ static int sp_2048_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 64);
         for (; i>=0 || c>=3; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 29);
-                n <<= 3;
+                n = (sp_uint32)n << 3;
                 c = 29;
             }
             else if (c < 3) {
@@ -27764,12 +27775,12 @@ static int sp_2048_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 3 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 29) & 0x7);
-                n <<= 3;
+                n = (sp_uint32)n << 3;
                 c -= 3;
             }
 
@@ -27873,10 +27884,10 @@ static int sp_2048_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -27885,14 +27896,14 @@ static int sp_2048_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 64);
         for (; i>=0 || c>=4; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 28);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 28;
             }
             else if (c < 4) {
@@ -27900,12 +27911,12 @@ static int sp_2048_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 4 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
 
@@ -27978,7 +27989,7 @@ int sp_RsaPublic_2048(const byte* in, word32 inLen, const mp_int* em,
 #else
         e[0] = em->dp[0];
         if (em->used > 1) {
-            e[0] |= ((sp_digit)em->dp[1]) << DIGIT_BIT;
+            e[0] |= ((sp_uint32)em->dp[1]) << DIGIT_BIT;
         }
 #endif
         if (e[0] == 0) {
@@ -28325,7 +28336,7 @@ static int sp_2048_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 64; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -28350,7 +28361,7 @@ static int sp_2048_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 64; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -30066,10 +30077,10 @@ static int sp_2048_mod_exp_2_64(sp_digit* r, const sp_digit* e, int bits,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -30078,14 +30089,14 @@ static int sp_2048_mod_exp_2_64(sp_digit* r, const sp_digit* e, int bits,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         sp_2048_lshift_64(r, norm, y);
         for (; i>=0 || c>=5; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 27);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 27;
             }
             else if (c < 5) {
@@ -30093,12 +30104,12 @@ static int sp_2048_mod_exp_2_64(sp_digit* r, const sp_digit* e, int bits,
                 n = e[i--];
                 c = 5 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
 
@@ -30266,10 +30277,10 @@ static void sp_3072_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
@@ -30324,7 +30335,7 @@ static void sp_3072_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -30359,7 +30370,7 @@ static void sp_3072_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -72867,17 +72878,17 @@ WC_OMIT_FRAME_POINTER static SP_NOINLINE void sp_3072_sqr_48(sp_digit* r,
  */
 static void sp_3072_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint32 x;
+    sp_uint32 b;
 
-    b = a[0];
+    b = (sp_uint32)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**32 */
 
     /* rho = -1/m mod b */
-    *rho = (sp_digit)0 - x;
+    *rho = (sp_digit)((sp_int32)0 - (sp_int32)x);
 }
 
 /* Mul a by digit b into r. (r = a * b)
@@ -75576,10 +75587,10 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -75588,14 +75599,14 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 48);
         for (; i>=0 || c>=4; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 28);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 28;
             }
             else if (c < 4) {
@@ -75603,12 +75614,12 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 4 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
 
@@ -75729,10 +75740,10 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -75741,14 +75752,14 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 48);
         for (; i>=0 || c>=5; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 27);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 27;
             }
             else if (c < 5) {
@@ -75756,12 +75767,12 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 5 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
 
@@ -79967,10 +79978,10 @@ static int sp_3072_mod_exp_96(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -79979,14 +79990,14 @@ static int sp_3072_mod_exp_96(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 96);
         for (; i>=0 || c>=3; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 29);
-                n <<= 3;
+                n = (sp_uint32)n << 3;
                 c = 29;
             }
             else if (c < 3) {
@@ -79994,12 +80005,12 @@ static int sp_3072_mod_exp_96(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 3 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 29) & 0x7);
-                n <<= 3;
+                n = (sp_uint32)n << 3;
                 c -= 3;
             }
 
@@ -80103,10 +80114,10 @@ static int sp_3072_mod_exp_96(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -80115,14 +80126,14 @@ static int sp_3072_mod_exp_96(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 96);
         for (; i>=0 || c>=4; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 28);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 28;
             }
             else if (c < 4) {
@@ -80130,12 +80141,12 @@ static int sp_3072_mod_exp_96(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 4 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
 
@@ -80208,7 +80219,7 @@ int sp_RsaPublic_3072(const byte* in, word32 inLen, const mp_int* em,
 #else
         e[0] = em->dp[0];
         if (em->used > 1) {
-            e[0] |= ((sp_digit)em->dp[1]) << DIGIT_BIT;
+            e[0] |= ((sp_uint32)em->dp[1]) << DIGIT_BIT;
         }
 #endif
         if (e[0] == 0) {
@@ -80555,7 +80566,7 @@ static int sp_3072_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 96; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -80580,7 +80591,7 @@ static int sp_3072_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 96; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -83094,10 +83105,10 @@ static int sp_3072_mod_exp_2_96(sp_digit* r, const sp_digit* e, int bits,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -83106,14 +83117,14 @@ static int sp_3072_mod_exp_2_96(sp_digit* r, const sp_digit* e, int bits,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         sp_3072_lshift_96(r, norm, y);
         for (; i>=0 || c>=5; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 27);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 27;
             }
             else if (c < 5) {
@@ -83121,12 +83132,12 @@ static int sp_3072_mod_exp_2_96(sp_digit* r, const sp_digit* e, int bits,
                 n = e[i--];
                 c = 5 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
 
@@ -83294,10 +83305,10 @@ static void sp_4096_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
@@ -83352,7 +83363,7 @@ static void sp_4096_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -83387,7 +83398,7 @@ static void sp_4096_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -87248,17 +87259,17 @@ WC_OMIT_FRAME_POINTER static SP_NOINLINE void sp_4096_sqr_128(sp_digit* r,
  */
 static void sp_4096_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint32 x;
+    sp_uint32 b;
 
-    b = a[0];
+    b = (sp_uint32)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**32 */
 
     /* rho = -1/m mod b */
-    *rho = (sp_digit)0 - x;
+    *rho = (sp_digit)((sp_int32)0 - (sp_int32)x);
 }
 
 /* Mul a by digit b into r. (r = a * b)
@@ -92447,10 +92458,10 @@ static int sp_4096_mod_exp_128(sp_digit* r, const sp_digit* a, const sp_digit* e
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -92459,14 +92470,14 @@ static int sp_4096_mod_exp_128(sp_digit* r, const sp_digit* a, const sp_digit* e
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 128);
         for (; i>=0 || c>=3; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 29);
-                n <<= 3;
+                n = (sp_uint32)n << 3;
                 c = 29;
             }
             else if (c < 3) {
@@ -92474,12 +92485,12 @@ static int sp_4096_mod_exp_128(sp_digit* r, const sp_digit* a, const sp_digit* e
                 n = e[i--];
                 c = 3 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 29) & 0x7);
-                n <<= 3;
+                n = (sp_uint32)n << 3;
                 c -= 3;
             }
 
@@ -92583,10 +92594,10 @@ static int sp_4096_mod_exp_128(sp_digit* r, const sp_digit* a, const sp_digit* e
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -92595,14 +92606,14 @@ static int sp_4096_mod_exp_128(sp_digit* r, const sp_digit* a, const sp_digit* e
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 128);
         for (; i>=0 || c>=4; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 28);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 28;
             }
             else if (c < 4) {
@@ -92610,12 +92621,12 @@ static int sp_4096_mod_exp_128(sp_digit* r, const sp_digit* a, const sp_digit* e
                 n = e[i--];
                 c = 4 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
 
@@ -92688,7 +92699,7 @@ int sp_RsaPublic_4096(const byte* in, word32 inLen, const mp_int* em,
 #else
         e[0] = em->dp[0];
         if (em->used > 1) {
-            e[0] |= ((sp_digit)em->dp[1]) << DIGIT_BIT;
+            e[0] |= ((sp_uint32)em->dp[1]) << DIGIT_BIT;
         }
 #endif
         if (e[0] == 0) {
@@ -93040,7 +93051,7 @@ static int sp_4096_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 128; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -93065,7 +93076,7 @@ static int sp_4096_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 128; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -96367,10 +96378,10 @@ static int sp_4096_mod_exp_2_128(sp_digit* r, const sp_digit* e, int bits,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -96379,14 +96390,14 @@ static int sp_4096_mod_exp_2_128(sp_digit* r, const sp_digit* e, int bits,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         sp_4096_lshift_128(r, norm, y);
         for (; i>=0 || c>=5; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 27);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 27;
             }
             else if (c < 5) {
@@ -96394,12 +96405,12 @@ static int sp_4096_mod_exp_2_128(sp_digit* r, const sp_digit* e, int bits,
                 n = e[i--];
                 c = 5 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
 
@@ -97538,7 +97549,7 @@ static void sp_256_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -97573,7 +97584,7 @@ static void sp_256_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -97639,7 +97650,7 @@ static int sp_256_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 8; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -97664,7 +97675,7 @@ static int sp_256_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 8; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -98763,7 +98774,7 @@ static void sp_256_mont_inv_8(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 8);
     for (i=254; i>=0; i--) {
         sp_256_mont_sqr_8(t, t, p256_mod, p256_mp_mod);
-        if (p256_mod_minus_2[i / 32] & ((sp_digit)1 << (i % 32)))
+        if (p256_mod_minus_2[i / 32] & ((sp_uint32)1 << (i % 32)))
             sp_256_mont_mul_8(t, t, a, p256_mod, p256_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 8);
@@ -100866,7 +100877,7 @@ static int sp_256_ecc_mulmod_fast_8(sp_point_256* r, const sp_point_256* g, cons
         t[15].infinity = 0;
 
         i = 6;
-        n = k[i+1] << 0;
+        n = (sp_uint32)k[i+1] << 0;
         c = 28;
         y = (int)(n >> 28);
     #ifndef WC_NO_CACHE_RESISTANT
@@ -100879,14 +100890,14 @@ static int sp_256_ecc_mulmod_fast_8(sp_point_256* r, const sp_point_256* g, cons
         {
             XMEMCPY(rt, &t[y], sizeof(sp_point_256));
         }
-        n <<= 4;
+        n = (sp_uint32)n << (4);
         for (; i>=0 || c>=4; ) {
             if (c < 4) {
                 n |= k[i--];
                 c += 32;
             }
             y = (n >> 28) & 0xf;
-            n <<= 4;
+            n = (sp_uint32)n << 4;
             c -= 4;
 
             sp_256_proj_point_dbl_8(rt, rt, tmp);
@@ -103600,10 +103611,10 @@ static void sp_256_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
@@ -108780,7 +108791,7 @@ static void sp_384_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -108815,7 +108826,7 @@ static void sp_384_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -108881,7 +108892,7 @@ static int sp_384_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 12; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -108906,7 +108917,7 @@ static int sp_384_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 12; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -109816,7 +109827,7 @@ static void sp_384_mont_inv_12(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 12);
     for (i=382; i>=0; i--) {
         sp_384_mont_sqr_12(t, t, p384_mod, p384_mp_mod);
-        if (p384_mod_minus_2[i / 32] & ((sp_digit)1 << (i % 32)))
+        if (p384_mod_minus_2[i / 32] & ((sp_uint32)1 << (i % 32)))
             sp_384_mont_mul_12(t, t, a, p384_mod, p384_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 12);
@@ -111318,7 +111329,7 @@ static int sp_384_ecc_mulmod_fast_12(sp_point_384* r, const sp_point_384* g, con
         t[15].infinity = 0;
 
         i = 10;
-        n = k[i+1] << 0;
+        n = (sp_uint32)k[i+1] << 0;
         c = 28;
         y = (int)(n >> 28);
     #ifndef WC_NO_CACHE_RESISTANT
@@ -111331,14 +111342,14 @@ static int sp_384_ecc_mulmod_fast_12(sp_point_384* r, const sp_point_384* g, con
         {
             XMEMCPY(rt, &t[y], sizeof(sp_point_384));
         }
-        n <<= 4;
+        n = (sp_uint32)n << (4);
         for (; i>=0 || c>=4; ) {
             if (c < 4) {
                 n |= k[i--];
                 c += 32;
             }
             y = (n >> 28) & 0xf;
-            n <<= 4;
+            n = (sp_uint32)n << 4;
             c -= 4;
 
             sp_384_proj_point_dbl_12(rt, rt, tmp);
@@ -114120,10 +114131,10 @@ static void sp_384_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
@@ -119587,7 +119598,7 @@ static void sp_521_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -119622,7 +119633,7 @@ static void sp_521_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -119688,7 +119699,7 @@ static int sp_521_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 17; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -119713,7 +119724,7 @@ static int sp_521_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 17; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -121713,7 +121724,7 @@ static void sp_521_mont_inv_17(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 17);
     for (i=519; i>=0; i--) {
         sp_521_mont_sqr_17(t, t, p521_mod, p521_mp_mod);
-        if (p521_mod_minus_2[i / 32] & ((sp_digit)1 << (i % 32)))
+        if (p521_mod_minus_2[i / 32] & ((sp_uint32)1 << (i % 32)))
             sp_521_mont_mul_17(t, t, a, p521_mod, p521_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 17);
@@ -124547,7 +124558,7 @@ static int sp_521_ecc_mulmod_fast_17(sp_point_521* r, const sp_point_521* g, con
         t[15].infinity = 0;
 
         i = 15;
-        n = k[i+1] << 0;
+        n = (sp_uint32)k[i+1] << 0;
         c = 5;
         y = (int)(n >> 5);
     #ifndef WC_NO_CACHE_RESISTANT
@@ -124560,15 +124571,15 @@ static int sp_521_ecc_mulmod_fast_17(sp_point_521* r, const sp_point_521* g, con
         {
             XMEMCPY(rt, &t[y], sizeof(sp_point_521));
         }
-        n <<= 27;
+        n = (sp_uint32)n << (27);
         for (; i>=0 || c>=4; ) {
             if (c < 4) {
-                n = (k[i+1] << 31) | (k[i] >> 1);
+                n = ((sp_uint32)k[i+1] << 31) | (k[i] >> 1);
                 i--;
                 c += 32;
             }
             y = (n >> 28) & 0xf;
-            n <<= 4;
+            n = (sp_uint32)n << 4;
             c -= 4;
 
             sp_521_proj_point_dbl_17(rt, rt, tmp);
@@ -127992,10 +128003,10 @@ static void sp_521_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
@@ -131069,8 +131080,6 @@ static WC_INLINE int sp_521_div_17(const sp_digit* a, const sp_digit* d,
     sp_digit r1;
     int i;
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)m;
     div = (d[16] << 23) | (d[15] >> 9);
     XMEMCPY(t1, a, sizeof(*t1) * 2 * 17);
@@ -134979,7 +134988,7 @@ static int sp_521_mont_sqrt_17(sp_digit* y)
             XMEMCPY(t, y, sizeof(sp_digit) * 17);
             for (i=518; i>=0; i--) {
                 sp_521_mont_sqr_17(t, t, p521_mod, p521_mp_mod);
-                if (p521_sqrt_power[i / 32] & ((sp_digit)1 << (i % 32)))
+                if (p521_sqrt_power[i / 32] & ((sp_uint32)1 << (i % 32)))
                     sp_521_mont_mul_17(t, t, y, p521_mod, p521_mp_mod);
             }
             XMEMCPY(y, t, sizeof(sp_digit) * 17);
@@ -201254,7 +201263,7 @@ static void sp_1024_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -201289,7 +201298,7 @@ static void sp_1024_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -201355,7 +201364,7 @@ static int sp_1024_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 32; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -201380,7 +201389,7 @@ static int sp_1024_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 32; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -208949,18 +208958,18 @@ static int sp_1024_ecc_mulmod_fast_32(sp_point_1024* r, const sp_point_1024* g,
         t[15].infinity = 0;
 
         i = 30;
-        n = k[i+1] << 0;
+        n = (sp_uint32)k[i+1] << 0;
         c = 28;
         y = (int)(n >> 28);
         XMEMCPY(rt, &t[y], sizeof(sp_point_1024));
-        n <<= 4;
+        n = (sp_uint32)n << (4);
         for (; i>=0 || c>=4; ) {
             if (c < 4) {
                 n |= k[i--];
                 c += 32;
             }
             y = (n >> 28) & 0xf;
-            n <<= 4;
+            n = (sp_uint32)n << 4;
             c -= 4;
 
             sp_1024_proj_point_dbl_32(rt, rt, tmp);
@@ -217065,10 +217074,10 @@ static void sp_1024_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
diff --git a/wolfcrypt/src/sp_c32.c b/wolfcrypt/src/sp_c32.c
index 907c04c61a..8262fb2f73 100644
--- a/wolfcrypt/src/sp_c32.c
+++ b/wolfcrypt/src/sp_c32.c
@@ -51,6 +51,17 @@
 
 #include <wolfssl/wolfcrypt/sp.h>
 
+#if defined(WOLFSSL_USE_SAVE_VECTOR_REGISTERS) && !defined(WOLFSSL_SP_ASM) && \
+        !defined(DEBUG_VECTOR_REGISTER_ACCESS)
+    /* force off unneeded vector register save/restore. */
+    #undef SAVE_VECTOR_REGISTERS
+    #define SAVE_VECTOR_REGISTERS(fail_clause) SAVE_NO_VECTOR_REGISTERS(fail_clause)
+    #undef SAVE_VECTOR_REGISTERS2
+    #define SAVE_VECTOR_REGISTERS2() SAVE_NO_VECTOR_REGISTERS2()
+    #undef RESTORE_VECTOR_REGISTERS
+    #define RESTORE_VECTOR_REGISTERS() RESTORE_NO_VECTOR_REGISTERS()
+#endif
+
 #ifdef __IAR_SYSTEMS_ICC__
 #define __asm__        asm
 #define __volatile__   volatile
@@ -239,7 +250,7 @@ static void sp_2048_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     r[0] = 0;
     for (i = n-1; i >= 0; i--) {
-        r[j] |= (((sp_digit)a[i]) << s);
+        r[j] |= (((sp_uint32)a[i]) << s);
         if (s >= 21U) {
             r[j] &= 0x1fffffff;
             s = 29U - s;
@@ -285,7 +296,7 @@ static void sp_2048_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0x1fffffff;
         s = 29U - s;
         if (j + 1 >= size) {
@@ -320,7 +331,7 @@ static void sp_2048_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 29) {
             r[j] &= 0x1fffffff;
             if (j + 1 >= size) {
@@ -369,7 +380,7 @@ static void sp_2048_to_bin_72(sp_digit* r, byte* a)
     for (i=0; i<71 && j>=0; i++) {
         b = 0;
         /* lint allow cast of mismatch sp_digit and int */
-        a[j--] |= (byte)(r[i] << s); /*lint !e9033*/
+        a[j--] |= (byte)((sp_uint32)r[i] << s); /*lint !e9033*/
         b += 8 - s;
         if (j < 0) {
             break;
@@ -1526,10 +1537,10 @@ SP_NOINLINE static void sp_2048_sqr_36(sp_digit* r, const sp_digit* a)
  */
 static void sp_2048_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint32 x;
+    sp_uint32 b;
 
-    b = a[0];
+    b = (sp_uint32)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
@@ -1537,7 +1548,7 @@ static void sp_2048_mont_setup(const sp_digit* a, sp_digit* rho)
     x &= 0x1fffffff;
 
     /* rho = -1/m mod b */
-    *rho = ((sp_digit)1 << 29) - x;
+    *rho = (sp_digit)(((sp_uint32)1 << 29) - (sp_digit)x);
 }
 
 /* Multiply a by scalar b into r. (r = a * b)
@@ -2030,22 +2041,22 @@ SP_NOINLINE static void sp_2048_rshift_36(sp_digit* r, const sp_digit* a,
 
 #ifdef WOLFSSL_SP_SMALL
     for (i=0; i<35; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (29 - n))) & 0x1fffffff);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint32)a[i + 1] << (29 - n))) & 0x1fffffff);
     }
 #else
     for (i=0; i<32; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (29 - n)) & 0x1fffffff);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (29 - n)) & 0x1fffffff);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (29 - n)) & 0x1fffffff);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (29 - n)) & 0x1fffffff);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (29 - n)) & 0x1fffffff);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (29 - n)) & 0x1fffffff);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (29 - n)) & 0x1fffffff);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (29 - n)) & 0x1fffffff);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint32)a[i+1] << (29 - n)) & 0x1fffffff);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint32)a[i+2] << (29 - n)) & 0x1fffffff);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint32)a[i+3] << (29 - n)) & 0x1fffffff);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint32)a[i+4] << (29 - n)) & 0x1fffffff);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint32)a[i+5] << (29 - n)) & 0x1fffffff);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint32)a[i+6] << (29 - n)) & 0x1fffffff);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint32)a[i+7] << (29 - n)) & 0x1fffffff);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint32)a[i+8] << (29 - n)) & 0x1fffffff);
     }
-    r[32] = (a[32] >> n) | (sp_digit)((a[33] << (29 - n)) & 0x1fffffff);
-    r[33] = (a[33] >> n) | (sp_digit)((a[34] << (29 - n)) & 0x1fffffff);
-    r[34] = (a[34] >> n) | (sp_digit)((a[35] << (29 - n)) & 0x1fffffff);
+    r[32] = (a[32] >> n) | (sp_digit)(((sp_uint32)a[33] << (29 - n)) & 0x1fffffff);
+    r[33] = (a[33] >> n) | (sp_digit)(((sp_uint32)a[34] << (29 - n)) & 0x1fffffff);
+    r[34] = (a[34] >> n) | (sp_digit)(((sp_uint32)a[35] << (29 - n)) & 0x1fffffff);
 #endif /* WOLFSSL_SP_SMALL */
     r[35] = a[35] >> n;
 }
@@ -2085,7 +2096,7 @@ static WC_INLINE sp_digit sp_2048_div_word_36(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 27; i >= 1; i--) {
         t1 += t1 + (((sp_uint32)t0 >> 28) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint32)t0 << 1;
         t2 = (sp_digit)(((sp_uint32)(dv - t1)) >> 31);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -2277,9 +2288,9 @@ static int sp_2048_mod_exp_36(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 29;
-        c = bits % 29;
-        n = e[i--] << (29 - c);
+        i = (bits - 1) / 29;
+        c = ((bits - 1) % 29) + 1;
+        n = (sp_uint32)e[i--] << (29 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -2291,7 +2302,7 @@ static int sp_2048_mod_exp_36(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 28) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_2048_mont_mul_36(t[y^1], t[0], t[1], m, mp);
 
@@ -2353,9 +2364,9 @@ static int sp_2048_mod_exp_36(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 29;
-        c = bits % 29;
-        n = e[i--] << (29 - c);
+        i = (bits - 1) / 29;
+        c = ((bits - 1) % 29) + 1;
+        n = (sp_uint32)e[i--] << (29 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -2367,7 +2378,7 @@ static int sp_2048_mod_exp_36(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 28) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_2048_mont_mul_36(t[y^1], t[0], t[1], m, mp);
 
@@ -2468,38 +2479,38 @@ static int sp_2048_mod_exp_36(sp_digit* r, const sp_digit* a, const sp_digit* e,
             c = 29;
         }
         if (i < 36) {
-            n = e[i--] << (32 - c);
+            n = (sp_uint32)e[i--] << (32 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 5) {
-            n |= e[i--] << (3 - c);
+            n |= (sp_uint32)e[i--] << (3 - c);
             c += 29;
         }
         y = (int)((n >> 27) & 0x1f);
-        n <<= 5;
+        n = (sp_uint32)n << 5;
         c -= 5;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 72);
         while ((i >= 0) || (c >= 5)) {
             if (c >= 5) {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
             else if (c == 0) {
-                n = e[i--] << 3;
+                n = (sp_uint32)e[i--] << 3;
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 24;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n = e[i--] << 3;
+                n = (sp_uint32)e[i--] << 3;
                 c = 5 - c;
                 y |= (byte)((n >> (32 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 29 - c;
             }
 
@@ -3027,26 +3038,26 @@ SP_NOINLINE static void sp_2048_rshift_72(sp_digit* r, const sp_digit* a,
 
 #ifdef WOLFSSL_SP_SMALL
     for (i=0; i<71; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (29 - n))) & 0x1fffffff);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint32)a[i + 1] << (29 - n))) & 0x1fffffff);
     }
 #else
     for (i=0; i<64; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (29 - n)) & 0x1fffffff);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (29 - n)) & 0x1fffffff);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (29 - n)) & 0x1fffffff);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (29 - n)) & 0x1fffffff);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (29 - n)) & 0x1fffffff);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (29 - n)) & 0x1fffffff);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (29 - n)) & 0x1fffffff);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (29 - n)) & 0x1fffffff);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint32)a[i+1] << (29 - n)) & 0x1fffffff);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint32)a[i+2] << (29 - n)) & 0x1fffffff);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint32)a[i+3] << (29 - n)) & 0x1fffffff);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint32)a[i+4] << (29 - n)) & 0x1fffffff);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint32)a[i+5] << (29 - n)) & 0x1fffffff);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint32)a[i+6] << (29 - n)) & 0x1fffffff);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint32)a[i+7] << (29 - n)) & 0x1fffffff);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint32)a[i+8] << (29 - n)) & 0x1fffffff);
     }
-    r[64] = (a[64] >> n) | (sp_digit)((a[65] << (29 - n)) & 0x1fffffff);
-    r[65] = (a[65] >> n) | (sp_digit)((a[66] << (29 - n)) & 0x1fffffff);
-    r[66] = (a[66] >> n) | (sp_digit)((a[67] << (29 - n)) & 0x1fffffff);
-    r[67] = (a[67] >> n) | (sp_digit)((a[68] << (29 - n)) & 0x1fffffff);
-    r[68] = (a[68] >> n) | (sp_digit)((a[69] << (29 - n)) & 0x1fffffff);
-    r[69] = (a[69] >> n) | (sp_digit)((a[70] << (29 - n)) & 0x1fffffff);
-    r[70] = (a[70] >> n) | (sp_digit)((a[71] << (29 - n)) & 0x1fffffff);
+    r[64] = (a[64] >> n) | (sp_digit)(((sp_uint32)a[65] << (29 - n)) & 0x1fffffff);
+    r[65] = (a[65] >> n) | (sp_digit)(((sp_uint32)a[66] << (29 - n)) & 0x1fffffff);
+    r[66] = (a[66] >> n) | (sp_digit)(((sp_uint32)a[67] << (29 - n)) & 0x1fffffff);
+    r[67] = (a[67] >> n) | (sp_digit)(((sp_uint32)a[68] << (29 - n)) & 0x1fffffff);
+    r[68] = (a[68] >> n) | (sp_digit)(((sp_uint32)a[69] << (29 - n)) & 0x1fffffff);
+    r[69] = (a[69] >> n) | (sp_digit)(((sp_uint32)a[70] << (29 - n)) & 0x1fffffff);
+    r[70] = (a[70] >> n) | (sp_digit)(((sp_uint32)a[71] << (29 - n)) & 0x1fffffff);
 #endif /* WOLFSSL_SP_SMALL */
     r[71] = a[71] >> n;
 }
@@ -3086,7 +3097,7 @@ static WC_INLINE sp_digit sp_2048_div_word_72(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 27; i >= 1; i--) {
         t1 += t1 + (((sp_uint32)t0 >> 28) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint32)t0 << 1;
         t2 = (sp_digit)(((sp_uint32)(dv - t1)) >> 31);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -3281,9 +3292,9 @@ static int sp_2048_mod_exp_72(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 29;
-        c = bits % 29;
-        n = e[i--] << (29 - c);
+        i = (bits - 1) / 29;
+        c = ((bits - 1) % 29) + 1;
+        n = (sp_uint32)e[i--] << (29 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -3295,7 +3306,7 @@ static int sp_2048_mod_exp_72(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 28) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_2048_mont_mul_72(t[y^1], t[0], t[1], m, mp);
 
@@ -3357,9 +3368,9 @@ static int sp_2048_mod_exp_72(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 29;
-        c = bits % 29;
-        n = e[i--] << (29 - c);
+        i = (bits - 1) / 29;
+        c = ((bits - 1) % 29) + 1;
+        n = (sp_uint32)e[i--] << (29 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -3371,7 +3382,7 @@ static int sp_2048_mod_exp_72(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 28) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_2048_mont_mul_72(t[y^1], t[0], t[1], m, mp);
 
@@ -3456,38 +3467,38 @@ static int sp_2048_mod_exp_72(sp_digit* r, const sp_digit* a, const sp_digit* e,
             c = 29;
         }
         if (i < 72) {
-            n = e[i--] << (32 - c);
+            n = (sp_uint32)e[i--] << (32 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 4) {
-            n |= e[i--] << (3 - c);
+            n |= (sp_uint32)e[i--] << (3 - c);
             c += 29;
         }
         y = (int)((n >> 28) & 0xf);
-        n <<= 4;
+        n = (sp_uint32)n << 4;
         c -= 4;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 144);
         while ((i >= 0) || (c >= 4)) {
             if (c >= 4) {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
             else if (c == 0) {
-                n = e[i--] << 3;
+                n = (sp_uint32)e[i--] << 3;
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 25;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n = e[i--] << 3;
+                n = (sp_uint32)e[i--] << 3;
                 c = 4 - c;
                 y |= (byte)((n >> (32 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 29 - c;
             }
 
@@ -3583,10 +3594,10 @@ static int sp_2048_mod_exp_72_nb(sp_2048_mod_exp_72_ctx* ctx,
     case 4: /* BIT_INIT: index the most-significant exponent limb without
               reading off the end when bits is an exact multiple of the
               limb width. (bits-1) keeps i within the populated range and
-              c = (bits-1) % @bits + 1 keeps the shift in [1, @bits]. */
+              c = (bits-1) % 29 + 1 keeps the shift in [1, 29]. */
         ctx->i = (ctx->bits - 1) / 29;
         ctx->c = ((ctx->bits - 1) % 29) + 1;
-        ctx->n = e[ctx->i--] << (29 - ctx->c);
+        ctx->n = (sp_uint32)e[ctx->i--] << (29 - ctx->c);
         ctx->state = 5;
         break;
     case 5: /* BIT_NEXT: refill on word boundary, peel one exponent bit */
@@ -3599,7 +3610,7 @@ static int sp_2048_mod_exp_72_nb(sp_2048_mod_exp_72_ctx* ctx,
             ctx->c = 29;
         }
         ctx->y = (byte)((ctx->n >> 28) & 1);
-        ctx->n <<= 1;
+        ctx->n = (sp_uint32)ctx->n << 1;
         ctx->state = 6;
         break;
     case 6: /* MUL: t[y^1] = t[0] * t[1] in Montgomery form */
@@ -4372,7 +4383,7 @@ static int sp_2048_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 71; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -4397,7 +4408,7 @@ static int sp_2048_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 71; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 29 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -4614,7 +4625,7 @@ SP_NOINLINE static void sp_2048_lshift_72(sp_digit* r, const sp_digit* a,
 
     r[72] = a[71] >> (29 - n);
     for (i=71; i>0; i--) {
-        r[i] = (sp_digit)(((a[i] << n) | (a[i-1] >> (29 - n))) & 0x1fffffff);
+        r[i] = (sp_digit)((((sp_uint32)a[i] << n) | (a[i-1] >> (29 - n))) & 0x1fffffff);
     }
 #else
     sp_int_digit s;
@@ -4765,7 +4776,7 @@ SP_NOINLINE static void sp_2048_lshift_72(sp_digit* r, const sp_digit* a,
     s = (sp_int_digit)(a[1]); t = (sp_int_digit)(a[0]);
     r[1] = (sp_digit)(((s << n) | (t >> (29U - n))) & 0x1fffffff);
 #endif /* WOLFSSL_SP_SMALL */
-    r[0] = (sp_digit)((a[0] << n) & 0x1fffffff);
+    r[0] = (sp_digit)(((sp_uint32)a[0] << n) & 0x1fffffff);
 }
 
 /* Modular exponentiate 2 to the e mod m. (r = 2^e mod m)
@@ -4811,38 +4822,38 @@ static int sp_2048_mod_exp_2_72(sp_digit* r, const sp_digit* e, int bits, const
             c = 29;
         }
         if (i < 72) {
-            n = e[i--] << (32 - c);
+            n = (sp_uint32)e[i--] << (32 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 4) {
-            n |= e[i--] << (3 - c);
+            n |= (sp_uint32)e[i--] << (3 - c);
             c += 29;
         }
         y = (int)((n >> 28) & 0xf);
-        n <<= 4;
+        n = (sp_uint32)n << 4;
         c -= 4;
         sp_2048_lshift_72(r, norm, (byte)y);
         while ((i >= 0) || (c >= 4)) {
             if (c >= 4) {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
             else if (c == 0) {
-                n = e[i--] << 3;
+                n = (sp_uint32)e[i--] << 3;
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 25;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n = e[i--] << 3;
+                n = (sp_uint32)e[i--] << 3;
                 c = 4 - c;
                 y |= (byte)((n >> (32 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 29 - c;
             }
 
@@ -5172,7 +5183,7 @@ static void sp_3072_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     r[0] = 0;
     for (i = n-1; i >= 0; i--) {
-        r[j] |= (((sp_digit)a[i]) << s);
+        r[j] |= (((sp_uint32)a[i]) << s);
         if (s >= 21U) {
             r[j] &= 0x1fffffff;
             s = 29U - s;
@@ -5218,7 +5229,7 @@ static void sp_3072_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0x1fffffff;
         s = 29U - s;
         if (j + 1 >= size) {
@@ -5253,7 +5264,7 @@ static void sp_3072_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 29) {
             r[j] &= 0x1fffffff;
             if (j + 1 >= size) {
@@ -5302,7 +5313,7 @@ static void sp_3072_to_bin_106(sp_digit* r, byte* a)
     for (i=0; i<106 && j>=0; i++) {
         b = 0;
         /* lint allow cast of mismatch sp_digit and int */
-        a[j--] |= (byte)(r[i] << s); /*lint !e9033*/
+        a[j--] |= (byte)((sp_uint32)r[i] << s); /*lint !e9033*/
         b += 8 - s;
         if (j < 0) {
             break;
@@ -5478,10 +5489,10 @@ SP_NOINLINE static void sp_3072_sqr_106(sp_digit* r, const sp_digit* a)
  */
 static void sp_3072_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint32 x;
+    sp_uint32 b;
 
-    b = a[0];
+    b = (sp_uint32)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
@@ -5489,7 +5500,7 @@ static void sp_3072_mont_setup(const sp_digit* a, sp_digit* rho)
     x &= 0x1fffffff;
 
     /* rho = -1/m mod b */
-    *rho = ((sp_digit)1 << 29) - x;
+    *rho = (sp_digit)(((sp_uint32)1 << 29) - (sp_digit)x);
 }
 
 /* Multiply a by scalar b into r. (r = a * b)
@@ -5902,7 +5913,7 @@ SP_NOINLINE static void sp_3072_rshift_53(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<52; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (29 - n))) & 0x1fffffff);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint32)a[i + 1] << (29 - n))) & 0x1fffffff);
     }
     r[52] = a[52] >> n;
 }
@@ -5942,7 +5953,7 @@ static WC_INLINE sp_digit sp_3072_div_word_53(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 27; i >= 1; i--) {
         t1 += t1 + (((sp_uint32)t0 >> 28) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint32)t0 << 1;
         t2 = (sp_digit)(((sp_uint32)(dv - t1)) >> 31);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -6134,9 +6145,9 @@ static int sp_3072_mod_exp_53(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 29;
-        c = bits % 29;
-        n = e[i--] << (29 - c);
+        i = (bits - 1) / 29;
+        c = ((bits - 1) % 29) + 1;
+        n = (sp_uint32)e[i--] << (29 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -6148,7 +6159,7 @@ static int sp_3072_mod_exp_53(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 28) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_3072_mont_mul_53(t[y^1], t[0], t[1], m, mp);
 
@@ -6210,9 +6221,9 @@ static int sp_3072_mod_exp_53(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 29;
-        c = bits % 29;
-        n = e[i--] << (29 - c);
+        i = (bits - 1) / 29;
+        c = ((bits - 1) % 29) + 1;
+        n = (sp_uint32)e[i--] << (29 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -6224,7 +6235,7 @@ static int sp_3072_mod_exp_53(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 28) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_3072_mont_mul_53(t[y^1], t[0], t[1], m, mp);
 
@@ -6325,38 +6336,38 @@ static int sp_3072_mod_exp_53(sp_digit* r, const sp_digit* a, const sp_digit* e,
             c = 29;
         }
         if (i < 53) {
-            n = e[i--] << (32 - c);
+            n = (sp_uint32)e[i--] << (32 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 5) {
-            n |= e[i--] << (3 - c);
+            n |= (sp_uint32)e[i--] << (3 - c);
             c += 29;
         }
         y = (int)((n >> 27) & 0x1f);
-        n <<= 5;
+        n = (sp_uint32)n << 5;
         c -= 5;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 106);
         while ((i >= 0) || (c >= 5)) {
             if (c >= 5) {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
             else if (c == 0) {
-                n = e[i--] << 3;
+                n = (sp_uint32)e[i--] << 3;
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 24;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n = e[i--] << 3;
+                n = (sp_uint32)e[i--] << 3;
                 c = 5 - c;
                 y |= (byte)((n >> (32 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 29 - c;
             }
 
@@ -6680,7 +6691,7 @@ SP_NOINLINE static void sp_3072_rshift_106(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<105; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (29 - n))) & 0x1fffffff);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint32)a[i + 1] << (29 - n))) & 0x1fffffff);
     }
     r[105] = a[105] >> n;
 }
@@ -6720,7 +6731,7 @@ static WC_INLINE sp_digit sp_3072_div_word_106(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 27; i >= 1; i--) {
         t1 += t1 + (((sp_uint32)t0 >> 28) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint32)t0 << 1;
         t2 = (sp_digit)(((sp_uint32)(dv - t1)) >> 31);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -6913,9 +6924,9 @@ static int sp_3072_mod_exp_106(sp_digit* r, const sp_digit* a, const sp_digit* e
     }
 
     if (err == MP_OKAY) {
-        i = bits / 29;
-        c = bits % 29;
-        n = e[i--] << (29 - c);
+        i = (bits - 1) / 29;
+        c = ((bits - 1) % 29) + 1;
+        n = (sp_uint32)e[i--] << (29 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -6927,7 +6938,7 @@ static int sp_3072_mod_exp_106(sp_digit* r, const sp_digit* a, const sp_digit* e
             }
 
             y = (int)((n >> 28) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_3072_mont_mul_106(t[y^1], t[0], t[1], m, mp);
 
@@ -6989,9 +7000,9 @@ static int sp_3072_mod_exp_106(sp_digit* r, const sp_digit* a, const sp_digit* e
     }
 
     if (err == MP_OKAY) {
-        i = bits / 29;
-        c = bits % 29;
-        n = e[i--] << (29 - c);
+        i = (bits - 1) / 29;
+        c = ((bits - 1) % 29) + 1;
+        n = (sp_uint32)e[i--] << (29 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -7003,7 +7014,7 @@ static int sp_3072_mod_exp_106(sp_digit* r, const sp_digit* a, const sp_digit* e
             }
 
             y = (int)((n >> 28) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_3072_mont_mul_106(t[y^1], t[0], t[1], m, mp);
 
@@ -7088,38 +7099,38 @@ static int sp_3072_mod_exp_106(sp_digit* r, const sp_digit* a, const sp_digit* e
             c = 29;
         }
         if (i < 106) {
-            n = e[i--] << (32 - c);
+            n = (sp_uint32)e[i--] << (32 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 4) {
-            n |= e[i--] << (3 - c);
+            n |= (sp_uint32)e[i--] << (3 - c);
             c += 29;
         }
         y = (int)((n >> 28) & 0xf);
-        n <<= 4;
+        n = (sp_uint32)n << 4;
         c -= 4;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 212);
         while ((i >= 0) || (c >= 4)) {
             if (c >= 4) {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
             else if (c == 0) {
-                n = e[i--] << 3;
+                n = (sp_uint32)e[i--] << 3;
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 25;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n = e[i--] << 3;
+                n = (sp_uint32)e[i--] << 3;
                 c = 4 - c;
                 y |= (byte)((n >> (32 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 29 - c;
             }
 
@@ -7215,10 +7226,10 @@ static int sp_3072_mod_exp_106_nb(sp_3072_mod_exp_106_ctx* ctx,
     case 4: /* BIT_INIT: index the most-significant exponent limb without
               reading off the end when bits is an exact multiple of the
               limb width. (bits-1) keeps i within the populated range and
-              c = (bits-1) % @bits + 1 keeps the shift in [1, @bits]. */
+              c = (bits-1) % 29 + 1 keeps the shift in [1, 29]. */
         ctx->i = (ctx->bits - 1) / 29;
         ctx->c = ((ctx->bits - 1) % 29) + 1;
-        ctx->n = e[ctx->i--] << (29 - ctx->c);
+        ctx->n = (sp_uint32)e[ctx->i--] << (29 - ctx->c);
         ctx->state = 5;
         break;
     case 5: /* BIT_NEXT: refill on word boundary, peel one exponent bit */
@@ -7231,7 +7242,7 @@ static int sp_3072_mod_exp_106_nb(sp_3072_mod_exp_106_ctx* ctx,
             ctx->c = 29;
         }
         ctx->y = (byte)((ctx->n >> 28) & 1);
-        ctx->n <<= 1;
+        ctx->n = (sp_uint32)ctx->n << 1;
         ctx->state = 6;
         break;
     case 6: /* MUL: t[y^1] = t[0] * t[1] in Montgomery form */
@@ -8002,7 +8013,7 @@ static int sp_3072_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 106; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -8027,7 +8038,7 @@ static int sp_3072_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 106; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 29 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -8243,9 +8254,9 @@ SP_NOINLINE static void sp_3072_lshift_106(sp_digit* r, const sp_digit* a,
 
     r[106] = a[105] >> (29 - n);
     for (i=105; i>0; i--) {
-        r[i] = (sp_digit)(((a[i] << n) | (a[i-1] >> (29 - n))) & 0x1fffffff);
+        r[i] = (sp_digit)((((sp_uint32)a[i] << n) | (a[i-1] >> (29 - n))) & 0x1fffffff);
     }
-    r[0] = (sp_digit)((a[0] << n) & 0x1fffffff);
+    r[0] = (sp_digit)(((sp_uint32)a[0] << n) & 0x1fffffff);
 }
 
 /* Modular exponentiate 2 to the e mod m. (r = 2^e mod m)
@@ -8291,38 +8302,38 @@ static int sp_3072_mod_exp_2_106(sp_digit* r, const sp_digit* e, int bits, const
             c = 29;
         }
         if (i < 106) {
-            n = e[i--] << (32 - c);
+            n = (sp_uint32)e[i--] << (32 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 4) {
-            n |= e[i--] << (3 - c);
+            n |= (sp_uint32)e[i--] << (3 - c);
             c += 29;
         }
         y = (int)((n >> 28) & 0xf);
-        n <<= 4;
+        n = (sp_uint32)n << 4;
         c -= 4;
         sp_3072_lshift_106(r, norm, (byte)y);
         while ((i >= 0) || (c >= 4)) {
             if (c >= 4) {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
             else if (c == 0) {
-                n = e[i--] << 3;
+                n = (sp_uint32)e[i--] << 3;
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 25;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n = e[i--] << 3;
+                n = (sp_uint32)e[i--] << 3;
                 c = 4 - c;
                 y |= (byte)((n >> (32 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 29 - c;
             }
 
@@ -8649,7 +8660,7 @@ static void sp_3072_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     r[0] = 0;
     for (i = n-1; i >= 0; i--) {
-        r[j] |= (((sp_digit)a[i]) << s);
+        r[j] |= (((sp_uint32)a[i]) << s);
         if (s >= 20U) {
             r[j] &= 0xfffffff;
             s = 28U - s;
@@ -8695,7 +8706,7 @@ static void sp_3072_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xfffffff;
         s = 28U - s;
         if (j + 1 >= size) {
@@ -8730,7 +8741,7 @@ static void sp_3072_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 28) {
             r[j] &= 0xfffffff;
             if (j + 1 >= size) {
@@ -8779,7 +8790,7 @@ static void sp_3072_to_bin_112(sp_digit* r, byte* a)
     for (i=0; i<110 && j>=0; i++) {
         b = 0;
         /* lint allow cast of mismatch sp_digit and int */
-        a[j--] |= (byte)(r[i] << s); /*lint !e9033*/
+        a[j--] |= (byte)((sp_uint32)r[i] << s); /*lint !e9033*/
         b += 8 - s;
         if (j < 0) {
             break;
@@ -9701,10 +9712,10 @@ SP_NOINLINE static void sp_3072_sqr_112(sp_digit* r, const sp_digit* a)
  */
 static void sp_3072_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint32 x;
+    sp_uint32 b;
 
-    b = a[0];
+    b = (sp_uint32)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
@@ -9712,7 +9723,7 @@ static void sp_3072_mont_setup(const sp_digit* a, sp_digit* rho)
     x &= 0xfffffff;
 
     /* rho = -1/m mod b */
-    *rho = ((sp_digit)1 << 28) - x;
+    *rho = (sp_digit)(((sp_uint32)1 << 28) - (sp_digit)x);
 }
 
 /* Multiply a by scalar b into r. (r = a * b)
@@ -10094,22 +10105,22 @@ SP_NOINLINE static void sp_3072_rshift_56(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<48; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (28 - n)) & 0xfffffff);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (28 - n)) & 0xfffffff);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (28 - n)) & 0xfffffff);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (28 - n)) & 0xfffffff);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (28 - n)) & 0xfffffff);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (28 - n)) & 0xfffffff);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (28 - n)) & 0xfffffff);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (28 - n)) & 0xfffffff);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint32)a[i+1] << (28 - n)) & 0xfffffff);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint32)a[i+2] << (28 - n)) & 0xfffffff);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint32)a[i+3] << (28 - n)) & 0xfffffff);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint32)a[i+4] << (28 - n)) & 0xfffffff);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint32)a[i+5] << (28 - n)) & 0xfffffff);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint32)a[i+6] << (28 - n)) & 0xfffffff);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint32)a[i+7] << (28 - n)) & 0xfffffff);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint32)a[i+8] << (28 - n)) & 0xfffffff);
     }
-    r[48] = (a[48] >> n) | (sp_digit)((a[49] << (28 - n)) & 0xfffffff);
-    r[49] = (a[49] >> n) | (sp_digit)((a[50] << (28 - n)) & 0xfffffff);
-    r[50] = (a[50] >> n) | (sp_digit)((a[51] << (28 - n)) & 0xfffffff);
-    r[51] = (a[51] >> n) | (sp_digit)((a[52] << (28 - n)) & 0xfffffff);
-    r[52] = (a[52] >> n) | (sp_digit)((a[53] << (28 - n)) & 0xfffffff);
-    r[53] = (a[53] >> n) | (sp_digit)((a[54] << (28 - n)) & 0xfffffff);
-    r[54] = (a[54] >> n) | (sp_digit)((a[55] << (28 - n)) & 0xfffffff);
+    r[48] = (a[48] >> n) | (sp_digit)(((sp_uint32)a[49] << (28 - n)) & 0xfffffff);
+    r[49] = (a[49] >> n) | (sp_digit)(((sp_uint32)a[50] << (28 - n)) & 0xfffffff);
+    r[50] = (a[50] >> n) | (sp_digit)(((sp_uint32)a[51] << (28 - n)) & 0xfffffff);
+    r[51] = (a[51] >> n) | (sp_digit)(((sp_uint32)a[52] << (28 - n)) & 0xfffffff);
+    r[52] = (a[52] >> n) | (sp_digit)(((sp_uint32)a[53] << (28 - n)) & 0xfffffff);
+    r[53] = (a[53] >> n) | (sp_digit)(((sp_uint32)a[54] << (28 - n)) & 0xfffffff);
+    r[54] = (a[54] >> n) | (sp_digit)(((sp_uint32)a[55] << (28 - n)) & 0xfffffff);
     r[55] = a[55] >> n;
 }
 
@@ -10148,7 +10159,7 @@ static WC_INLINE sp_digit sp_3072_div_word_56(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 26; i >= 1; i--) {
         t1 += t1 + (((sp_uint32)t0 >> 27) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint32)t0 << 1;
         t2 = (sp_digit)(((sp_uint32)(dv - t1)) >> 31);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -10340,9 +10351,9 @@ static int sp_3072_mod_exp_56(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 28;
-        c = bits % 28;
-        n = e[i--] << (28 - c);
+        i = (bits - 1) / 28;
+        c = ((bits - 1) % 28) + 1;
+        n = (sp_uint32)e[i--] << (28 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -10354,7 +10365,7 @@ static int sp_3072_mod_exp_56(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 27) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_3072_mont_mul_56(t[y^1], t[0], t[1], m, mp);
 
@@ -10416,9 +10427,9 @@ static int sp_3072_mod_exp_56(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 28;
-        c = bits % 28;
-        n = e[i--] << (28 - c);
+        i = (bits - 1) / 28;
+        c = ((bits - 1) % 28) + 1;
+        n = (sp_uint32)e[i--] << (28 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -10430,7 +10441,7 @@ static int sp_3072_mod_exp_56(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 27) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_3072_mont_mul_56(t[y^1], t[0], t[1], m, mp);
 
@@ -10531,38 +10542,38 @@ static int sp_3072_mod_exp_56(sp_digit* r, const sp_digit* a, const sp_digit* e,
             c = 28;
         }
         if (i < 56) {
-            n = e[i--] << (32 - c);
+            n = (sp_uint32)e[i--] << (32 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 5) {
-            n |= e[i--] << (4 - c);
+            n |= (sp_uint32)e[i--] << (4 - c);
             c += 28;
         }
         y = (int)((n >> 27) & 0x1f);
-        n <<= 5;
+        n = (sp_uint32)n << 5;
         c -= 5;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 112);
         while ((i >= 0) || (c >= 5)) {
             if (c >= 5) {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
             else if (c == 0) {
-                n = e[i--] << 4;
+                n = (sp_uint32)e[i--] << 4;
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 23;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n = e[i--] << 4;
+                n = (sp_uint32)e[i--] << 4;
                 c = 5 - c;
                 y |= (byte)((n >> (32 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 28 - c;
             }
 
@@ -10951,22 +10962,22 @@ SP_NOINLINE static void sp_3072_rshift_112(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<104; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (28 - n)) & 0xfffffff);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (28 - n)) & 0xfffffff);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (28 - n)) & 0xfffffff);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (28 - n)) & 0xfffffff);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (28 - n)) & 0xfffffff);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (28 - n)) & 0xfffffff);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (28 - n)) & 0xfffffff);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (28 - n)) & 0xfffffff);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint32)a[i+1] << (28 - n)) & 0xfffffff);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint32)a[i+2] << (28 - n)) & 0xfffffff);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint32)a[i+3] << (28 - n)) & 0xfffffff);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint32)a[i+4] << (28 - n)) & 0xfffffff);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint32)a[i+5] << (28 - n)) & 0xfffffff);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint32)a[i+6] << (28 - n)) & 0xfffffff);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint32)a[i+7] << (28 - n)) & 0xfffffff);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint32)a[i+8] << (28 - n)) & 0xfffffff);
     }
-    r[104] = (a[104] >> n) | (sp_digit)((a[105] << (28 - n)) & 0xfffffff);
-    r[105] = (a[105] >> n) | (sp_digit)((a[106] << (28 - n)) & 0xfffffff);
-    r[106] = (a[106] >> n) | (sp_digit)((a[107] << (28 - n)) & 0xfffffff);
-    r[107] = (a[107] >> n) | (sp_digit)((a[108] << (28 - n)) & 0xfffffff);
-    r[108] = (a[108] >> n) | (sp_digit)((a[109] << (28 - n)) & 0xfffffff);
-    r[109] = (a[109] >> n) | (sp_digit)((a[110] << (28 - n)) & 0xfffffff);
-    r[110] = (a[110] >> n) | (sp_digit)((a[111] << (28 - n)) & 0xfffffff);
+    r[104] = (a[104] >> n) | (sp_digit)(((sp_uint32)a[105] << (28 - n)) & 0xfffffff);
+    r[105] = (a[105] >> n) | (sp_digit)(((sp_uint32)a[106] << (28 - n)) & 0xfffffff);
+    r[106] = (a[106] >> n) | (sp_digit)(((sp_uint32)a[107] << (28 - n)) & 0xfffffff);
+    r[107] = (a[107] >> n) | (sp_digit)(((sp_uint32)a[108] << (28 - n)) & 0xfffffff);
+    r[108] = (a[108] >> n) | (sp_digit)(((sp_uint32)a[109] << (28 - n)) & 0xfffffff);
+    r[109] = (a[109] >> n) | (sp_digit)(((sp_uint32)a[110] << (28 - n)) & 0xfffffff);
+    r[110] = (a[110] >> n) | (sp_digit)(((sp_uint32)a[111] << (28 - n)) & 0xfffffff);
     r[111] = a[111] >> n;
 }
 
@@ -11005,7 +11016,7 @@ static WC_INLINE sp_digit sp_3072_div_word_112(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 26; i >= 1; i--) {
         t1 += t1 + (((sp_uint32)t0 >> 27) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint32)t0 << 1;
         t2 = (sp_digit)(((sp_uint32)(dv - t1)) >> 31);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -11201,9 +11212,9 @@ static int sp_3072_mod_exp_112(sp_digit* r, const sp_digit* a, const sp_digit* e
     }
 
     if (err == MP_OKAY) {
-        i = bits / 28;
-        c = bits % 28;
-        n = e[i--] << (28 - c);
+        i = (bits - 1) / 28;
+        c = ((bits - 1) % 28) + 1;
+        n = (sp_uint32)e[i--] << (28 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -11215,7 +11226,7 @@ static int sp_3072_mod_exp_112(sp_digit* r, const sp_digit* a, const sp_digit* e
             }
 
             y = (int)((n >> 27) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_3072_mont_mul_112(t[y^1], t[0], t[1], m, mp);
 
@@ -11277,9 +11288,9 @@ static int sp_3072_mod_exp_112(sp_digit* r, const sp_digit* a, const sp_digit* e
     }
 
     if (err == MP_OKAY) {
-        i = bits / 28;
-        c = bits % 28;
-        n = e[i--] << (28 - c);
+        i = (bits - 1) / 28;
+        c = ((bits - 1) % 28) + 1;
+        n = (sp_uint32)e[i--] << (28 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -11291,7 +11302,7 @@ static int sp_3072_mod_exp_112(sp_digit* r, const sp_digit* a, const sp_digit* e
             }
 
             y = (int)((n >> 27) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_3072_mont_mul_112(t[y^1], t[0], t[1], m, mp);
 
@@ -11376,38 +11387,38 @@ static int sp_3072_mod_exp_112(sp_digit* r, const sp_digit* a, const sp_digit* e
             c = 28;
         }
         if (i < 112) {
-            n = e[i--] << (32 - c);
+            n = (sp_uint32)e[i--] << (32 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 4) {
-            n |= e[i--] << (4 - c);
+            n |= (sp_uint32)e[i--] << (4 - c);
             c += 28;
         }
         y = (int)((n >> 28) & 0xf);
-        n <<= 4;
+        n = (sp_uint32)n << 4;
         c -= 4;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 224);
         while ((i >= 0) || (c >= 4)) {
             if (c >= 4) {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
             else if (c == 0) {
-                n = e[i--] << 4;
+                n = (sp_uint32)e[i--] << 4;
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 24;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n = e[i--] << 4;
+                n = (sp_uint32)e[i--] << 4;
                 c = 4 - c;
                 y |= (byte)((n >> (32 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 28 - c;
             }
 
@@ -11503,10 +11514,10 @@ static int sp_3072_mod_exp_112_nb(sp_3072_mod_exp_112_ctx* ctx,
     case 4: /* BIT_INIT: index the most-significant exponent limb without
               reading off the end when bits is an exact multiple of the
               limb width. (bits-1) keeps i within the populated range and
-              c = (bits-1) % @bits + 1 keeps the shift in [1, @bits]. */
+              c = (bits-1) % 28 + 1 keeps the shift in [1, 28]. */
         ctx->i = (ctx->bits - 1) / 28;
         ctx->c = ((ctx->bits - 1) % 28) + 1;
-        ctx->n = e[ctx->i--] << (28 - ctx->c);
+        ctx->n = (sp_uint32)e[ctx->i--] << (28 - ctx->c);
         ctx->state = 5;
         break;
     case 5: /* BIT_NEXT: refill on word boundary, peel one exponent bit */
@@ -11519,7 +11530,7 @@ static int sp_3072_mod_exp_112_nb(sp_3072_mod_exp_112_ctx* ctx,
             ctx->c = 28;
         }
         ctx->y = (byte)((ctx->n >> 27) & 1);
-        ctx->n <<= 1;
+        ctx->n = (sp_uint32)ctx->n << 1;
         ctx->state = 6;
         break;
     case 6: /* MUL: t[y^1] = t[0] * t[1] in Montgomery form */
@@ -12090,7 +12101,7 @@ static int sp_3072_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 110; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -12115,7 +12126,7 @@ static int sp_3072_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 110; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 28 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -12468,7 +12479,7 @@ SP_NOINLINE static void sp_3072_lshift_112(sp_digit* r, const sp_digit* a,
     r[2] = (sp_digit)(((s << n) | (t >> (28U - n))) & 0xfffffff);
     s = (sp_int_digit)(a[1]); t = (sp_int_digit)(a[0]);
     r[1] = (sp_digit)(((s << n) | (t >> (28U - n))) & 0xfffffff);
-    r[0] = (sp_digit)((a[0] << n) & 0xfffffff);
+    r[0] = (sp_digit)(((sp_uint32)a[0] << n) & 0xfffffff);
 }
 
 /* Modular exponentiate 2 to the e mod m. (r = 2^e mod m)
@@ -12514,38 +12525,38 @@ static int sp_3072_mod_exp_2_112(sp_digit* r, const sp_digit* e, int bits, const
             c = 28;
         }
         if (i < 112) {
-            n = e[i--] << (32 - c);
+            n = (sp_uint32)e[i--] << (32 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 4) {
-            n |= e[i--] << (4 - c);
+            n |= (sp_uint32)e[i--] << (4 - c);
             c += 28;
         }
         y = (int)((n >> 28) & 0xf);
-        n <<= 4;
+        n = (sp_uint32)n << 4;
         c -= 4;
         sp_3072_lshift_112(r, norm, (byte)y);
         while ((i >= 0) || (c >= 4)) {
             if (c >= 4) {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
             else if (c == 0) {
-                n = e[i--] << 4;
+                n = (sp_uint32)e[i--] << 4;
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 24;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n = e[i--] << 4;
+                n = (sp_uint32)e[i--] << 4;
                 c = 4 - c;
                 y |= (byte)((n >> (32 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 28 - c;
             }
 
@@ -12774,7 +12785,7 @@ static void sp_4096_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     r[0] = 0;
     for (i = n-1; i >= 0; i--) {
-        r[j] |= (((sp_digit)a[i]) << s);
+        r[j] |= (((sp_uint32)a[i]) << s);
         if (s >= 21U) {
             r[j] &= 0x1fffffff;
             s = 29U - s;
@@ -12820,7 +12831,7 @@ static void sp_4096_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0x1fffffff;
         s = 29U - s;
         if (j + 1 >= size) {
@@ -12855,7 +12866,7 @@ static void sp_4096_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 29) {
             r[j] &= 0x1fffffff;
             if (j + 1 >= size) {
@@ -12904,7 +12915,7 @@ static void sp_4096_to_bin_142(sp_digit* r, byte* a)
     for (i=0; i<142 && j>=0; i++) {
         b = 0;
         /* lint allow cast of mismatch sp_digit and int */
-        a[j--] |= (byte)(r[i] << s); /*lint !e9033*/
+        a[j--] |= (byte)((sp_uint32)r[i] << s); /*lint !e9033*/
         b += 8 - s;
         if (j < 0) {
             break;
@@ -13082,10 +13093,10 @@ SP_NOINLINE static void sp_4096_sqr_142(sp_digit* r, const sp_digit* a)
  */
 static void sp_4096_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint32 x;
+    sp_uint32 b;
 
-    b = a[0];
+    b = (sp_uint32)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
@@ -13093,7 +13104,7 @@ static void sp_4096_mont_setup(const sp_digit* a, sp_digit* rho)
     x &= 0x1fffffff;
 
     /* rho = -1/m mod b */
-    *rho = ((sp_digit)1 << 29) - x;
+    *rho = (sp_digit)(((sp_uint32)1 << 29) - (sp_digit)x);
 }
 
 /* Multiply a by scalar b into r. (r = a * b)
@@ -13513,7 +13524,7 @@ SP_NOINLINE static void sp_4096_rshift_71(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<70; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (29 - n))) & 0x1fffffff);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint32)a[i + 1] << (29 - n))) & 0x1fffffff);
     }
     r[70] = a[70] >> n;
 }
@@ -13553,7 +13564,7 @@ static WC_INLINE sp_digit sp_4096_div_word_71(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 27; i >= 1; i--) {
         t1 += t1 + (((sp_uint32)t0 >> 28) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint32)t0 << 1;
         t2 = (sp_digit)(((sp_uint32)(dv - t1)) >> 31);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -13745,9 +13756,9 @@ static int sp_4096_mod_exp_71(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 29;
-        c = bits % 29;
-        n = e[i--] << (29 - c);
+        i = (bits - 1) / 29;
+        c = ((bits - 1) % 29) + 1;
+        n = (sp_uint32)e[i--] << (29 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -13759,7 +13770,7 @@ static int sp_4096_mod_exp_71(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 28) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_4096_mont_mul_71(t[y^1], t[0], t[1], m, mp);
 
@@ -13821,9 +13832,9 @@ static int sp_4096_mod_exp_71(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 29;
-        c = bits % 29;
-        n = e[i--] << (29 - c);
+        i = (bits - 1) / 29;
+        c = ((bits - 1) % 29) + 1;
+        n = (sp_uint32)e[i--] << (29 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -13835,7 +13846,7 @@ static int sp_4096_mod_exp_71(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 28) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_4096_mont_mul_71(t[y^1], t[0], t[1], m, mp);
 
@@ -13936,38 +13947,38 @@ static int sp_4096_mod_exp_71(sp_digit* r, const sp_digit* a, const sp_digit* e,
             c = 29;
         }
         if (i < 71) {
-            n = e[i--] << (32 - c);
+            n = (sp_uint32)e[i--] << (32 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 5) {
-            n |= e[i--] << (3 - c);
+            n |= (sp_uint32)e[i--] << (3 - c);
             c += 29;
         }
         y = (int)((n >> 27) & 0x1f);
-        n <<= 5;
+        n = (sp_uint32)n << 5;
         c -= 5;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 142);
         while ((i >= 0) || (c >= 5)) {
             if (c >= 5) {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
             else if (c == 0) {
-                n = e[i--] << 3;
+                n = (sp_uint32)e[i--] << 3;
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 24;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n = e[i--] << 3;
+                n = (sp_uint32)e[i--] << 3;
                 c = 5 - c;
                 y |= (byte)((n >> (32 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 29 - c;
             }
 
@@ -14292,7 +14303,7 @@ SP_NOINLINE static void sp_4096_rshift_142(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<141; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (29 - n))) & 0x1fffffff);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint32)a[i + 1] << (29 - n))) & 0x1fffffff);
     }
     r[141] = a[141] >> n;
 }
@@ -14332,7 +14343,7 @@ static WC_INLINE sp_digit sp_4096_div_word_142(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 27; i >= 1; i--) {
         t1 += t1 + (((sp_uint32)t0 >> 28) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint32)t0 << 1;
         t2 = (sp_digit)(((sp_uint32)(dv - t1)) >> 31);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -14525,9 +14536,9 @@ static int sp_4096_mod_exp_142(sp_digit* r, const sp_digit* a, const sp_digit* e
     }
 
     if (err == MP_OKAY) {
-        i = bits / 29;
-        c = bits % 29;
-        n = e[i--] << (29 - c);
+        i = (bits - 1) / 29;
+        c = ((bits - 1) % 29) + 1;
+        n = (sp_uint32)e[i--] << (29 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -14539,7 +14550,7 @@ static int sp_4096_mod_exp_142(sp_digit* r, const sp_digit* a, const sp_digit* e
             }
 
             y = (int)((n >> 28) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_4096_mont_mul_142(t[y^1], t[0], t[1], m, mp);
 
@@ -14601,9 +14612,9 @@ static int sp_4096_mod_exp_142(sp_digit* r, const sp_digit* a, const sp_digit* e
     }
 
     if (err == MP_OKAY) {
-        i = bits / 29;
-        c = bits % 29;
-        n = e[i--] << (29 - c);
+        i = (bits - 1) / 29;
+        c = ((bits - 1) % 29) + 1;
+        n = (sp_uint32)e[i--] << (29 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -14615,7 +14626,7 @@ static int sp_4096_mod_exp_142(sp_digit* r, const sp_digit* a, const sp_digit* e
             }
 
             y = (int)((n >> 28) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_4096_mont_mul_142(t[y^1], t[0], t[1], m, mp);
 
@@ -14700,38 +14711,38 @@ static int sp_4096_mod_exp_142(sp_digit* r, const sp_digit* a, const sp_digit* e
             c = 29;
         }
         if (i < 142) {
-            n = e[i--] << (32 - c);
+            n = (sp_uint32)e[i--] << (32 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 4) {
-            n |= e[i--] << (3 - c);
+            n |= (sp_uint32)e[i--] << (3 - c);
             c += 29;
         }
         y = (int)((n >> 28) & 0xf);
-        n <<= 4;
+        n = (sp_uint32)n << 4;
         c -= 4;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 284);
         while ((i >= 0) || (c >= 4)) {
             if (c >= 4) {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
             else if (c == 0) {
-                n = e[i--] << 3;
+                n = (sp_uint32)e[i--] << 3;
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 25;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n = e[i--] << 3;
+                n = (sp_uint32)e[i--] << 3;
                 c = 4 - c;
                 y |= (byte)((n >> (32 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 29 - c;
             }
 
@@ -14827,10 +14838,10 @@ static int sp_4096_mod_exp_142_nb(sp_4096_mod_exp_142_ctx* ctx,
     case 4: /* BIT_INIT: index the most-significant exponent limb without
               reading off the end when bits is an exact multiple of the
               limb width. (bits-1) keeps i within the populated range and
-              c = (bits-1) % @bits + 1 keeps the shift in [1, @bits]. */
+              c = (bits-1) % 29 + 1 keeps the shift in [1, 29]. */
         ctx->i = (ctx->bits - 1) / 29;
         ctx->c = ((ctx->bits - 1) % 29) + 1;
-        ctx->n = e[ctx->i--] << (29 - ctx->c);
+        ctx->n = (sp_uint32)e[ctx->i--] << (29 - ctx->c);
         ctx->state = 5;
         break;
     case 5: /* BIT_NEXT: refill on word boundary, peel one exponent bit */
@@ -14843,7 +14854,7 @@ static int sp_4096_mod_exp_142_nb(sp_4096_mod_exp_142_ctx* ctx,
             ctx->c = 29;
         }
         ctx->y = (byte)((ctx->n >> 28) & 1);
-        ctx->n <<= 1;
+        ctx->n = (sp_uint32)ctx->n << 1;
         ctx->state = 6;
         break;
     case 6: /* MUL: t[y^1] = t[0] * t[1] in Montgomery form */
@@ -15614,7 +15625,7 @@ static int sp_4096_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 142; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -15639,7 +15650,7 @@ static int sp_4096_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 142; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 29 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -15855,9 +15866,9 @@ SP_NOINLINE static void sp_4096_lshift_142(sp_digit* r, const sp_digit* a,
 
     r[142] = a[141] >> (29 - n);
     for (i=141; i>0; i--) {
-        r[i] = (sp_digit)(((a[i] << n) | (a[i-1] >> (29 - n))) & 0x1fffffff);
+        r[i] = (sp_digit)((((sp_uint32)a[i] << n) | (a[i-1] >> (29 - n))) & 0x1fffffff);
     }
-    r[0] = (sp_digit)((a[0] << n) & 0x1fffffff);
+    r[0] = (sp_digit)(((sp_uint32)a[0] << n) & 0x1fffffff);
 }
 
 /* Modular exponentiate 2 to the e mod m. (r = 2^e mod m)
@@ -15903,38 +15914,38 @@ static int sp_4096_mod_exp_2_142(sp_digit* r, const sp_digit* e, int bits, const
             c = 29;
         }
         if (i < 142) {
-            n = e[i--] << (32 - c);
+            n = (sp_uint32)e[i--] << (32 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 4) {
-            n |= e[i--] << (3 - c);
+            n |= (sp_uint32)e[i--] << (3 - c);
             c += 29;
         }
         y = (int)((n >> 28) & 0xf);
-        n <<= 4;
+        n = (sp_uint32)n << 4;
         c -= 4;
         sp_4096_lshift_142(r, norm, (byte)y);
         while ((i >= 0) || (c >= 4)) {
             if (c >= 4) {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
             else if (c == 0) {
-                n = e[i--] << 3;
+                n = (sp_uint32)e[i--] << 3;
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 25;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n = e[i--] << 3;
+                n = (sp_uint32)e[i--] << 3;
                 c = 4 - c;
                 y |= (byte)((n >> (32 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 29 - c;
             }
 
@@ -16160,7 +16171,7 @@ static void sp_4096_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     r[0] = 0;
     for (i = n-1; i >= 0; i--) {
-        r[j] |= (((sp_digit)a[i]) << s);
+        r[j] |= (((sp_uint32)a[i]) << s);
         if (s >= 18U) {
             r[j] &= 0x3ffffff;
             s = 26U - s;
@@ -16206,7 +16217,7 @@ static void sp_4096_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0x3ffffff;
         s = 26U - s;
         if (j + 1 >= size) {
@@ -16241,7 +16252,7 @@ static void sp_4096_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 26) {
             r[j] &= 0x3ffffff;
             if (j + 1 >= size) {
@@ -16290,7 +16301,7 @@ static void sp_4096_to_bin_162(sp_digit* r, byte* a)
     for (i=0; i<158 && j>=0; i++) {
         b = 0;
         /* lint allow cast of mismatch sp_digit and int */
-        a[j--] |= (byte)(r[i] << s); /*lint !e9033*/
+        a[j--] |= (byte)((sp_uint32)r[i] << s); /*lint !e9033*/
         b += 8 - s;
         if (j < 0) {
             break;
@@ -17221,10 +17232,10 @@ SP_NOINLINE static void sp_4096_sqr_162(sp_digit* r, const sp_digit* a)
  */
 static void sp_4096_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint32 x;
+    sp_uint32 b;
 
-    b = a[0];
+    b = (sp_uint32)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
@@ -17232,7 +17243,7 @@ static void sp_4096_mont_setup(const sp_digit* a, sp_digit* rho)
     x &= 0x3ffffff;
 
     /* rho = -1/m mod b */
-    *rho = ((sp_digit)1 << 26) - x;
+    *rho = (sp_digit)(((sp_uint32)1 << 26) - (sp_digit)x);
 }
 
 /* Multiply a by scalar b into r. (r = a * b)
@@ -17633,14 +17644,14 @@ SP_NOINLINE static void sp_4096_rshift_81(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<80; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (26 - n)) & 0x3ffffff);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (26 - n)) & 0x3ffffff);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (26 - n)) & 0x3ffffff);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (26 - n)) & 0x3ffffff);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (26 - n)) & 0x3ffffff);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (26 - n)) & 0x3ffffff);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (26 - n)) & 0x3ffffff);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (26 - n)) & 0x3ffffff);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint32)a[i+1] << (26 - n)) & 0x3ffffff);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint32)a[i+2] << (26 - n)) & 0x3ffffff);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint32)a[i+3] << (26 - n)) & 0x3ffffff);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint32)a[i+4] << (26 - n)) & 0x3ffffff);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint32)a[i+5] << (26 - n)) & 0x3ffffff);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint32)a[i+6] << (26 - n)) & 0x3ffffff);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint32)a[i+7] << (26 - n)) & 0x3ffffff);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint32)a[i+8] << (26 - n)) & 0x3ffffff);
     }
     r[80] = a[80] >> n;
 }
@@ -17680,7 +17691,7 @@ static WC_INLINE sp_digit sp_4096_div_word_81(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 24; i >= 1; i--) {
         t1 += t1 + (((sp_uint32)t0 >> 25) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint32)t0 << 1;
         t2 = (sp_digit)(((sp_uint32)(dv - t1)) >> 31);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -17873,9 +17884,9 @@ static int sp_4096_mod_exp_81(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 26;
-        c = bits % 26;
-        n = e[i--] << (26 - c);
+        i = (bits - 1) / 26;
+        c = ((bits - 1) % 26) + 1;
+        n = (sp_uint32)e[i--] << (26 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -17887,7 +17898,7 @@ static int sp_4096_mod_exp_81(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 25) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_4096_mont_mul_81(t[y^1], t[0], t[1], m, mp);
 
@@ -17949,9 +17960,9 @@ static int sp_4096_mod_exp_81(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 26;
-        c = bits % 26;
-        n = e[i--] << (26 - c);
+        i = (bits - 1) / 26;
+        c = ((bits - 1) % 26) + 1;
+        n = (sp_uint32)e[i--] << (26 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -17963,7 +17974,7 @@ static int sp_4096_mod_exp_81(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 25) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_4096_mont_mul_81(t[y^1], t[0], t[1], m, mp);
 
@@ -18064,38 +18075,38 @@ static int sp_4096_mod_exp_81(sp_digit* r, const sp_digit* a, const sp_digit* e,
             c = 26;
         }
         if (i < 81) {
-            n = e[i--] << (32 - c);
+            n = (sp_uint32)e[i--] << (32 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 5) {
-            n |= e[i--] << (6 - c);
+            n |= (sp_uint32)e[i--] << (6 - c);
             c += 26;
         }
         y = (int)((n >> 27) & 0x1f);
-        n <<= 5;
+        n = (sp_uint32)n << 5;
         c -= 5;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 162);
         while ((i >= 0) || (c >= 5)) {
             if (c >= 5) {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
             else if (c == 0) {
-                n = e[i--] << 6;
+                n = (sp_uint32)e[i--] << 6;
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 21;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n = e[i--] << 6;
+                n = (sp_uint32)e[i--] << 6;
                 c = 5 - c;
                 y |= (byte)((n >> (32 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 26 - c;
             }
 
@@ -18475,16 +18486,16 @@ SP_NOINLINE static void sp_4096_rshift_162(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<160; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (26 - n)) & 0x3ffffff);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (26 - n)) & 0x3ffffff);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (26 - n)) & 0x3ffffff);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (26 - n)) & 0x3ffffff);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (26 - n)) & 0x3ffffff);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (26 - n)) & 0x3ffffff);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (26 - n)) & 0x3ffffff);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (26 - n)) & 0x3ffffff);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint32)a[i+1] << (26 - n)) & 0x3ffffff);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint32)a[i+2] << (26 - n)) & 0x3ffffff);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint32)a[i+3] << (26 - n)) & 0x3ffffff);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint32)a[i+4] << (26 - n)) & 0x3ffffff);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint32)a[i+5] << (26 - n)) & 0x3ffffff);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint32)a[i+6] << (26 - n)) & 0x3ffffff);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint32)a[i+7] << (26 - n)) & 0x3ffffff);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint32)a[i+8] << (26 - n)) & 0x3ffffff);
     }
-    r[160] = (a[160] >> n) | (sp_digit)((a[161] << (26 - n)) & 0x3ffffff);
+    r[160] = (a[160] >> n) | (sp_digit)(((sp_uint32)a[161] << (26 - n)) & 0x3ffffff);
     r[161] = a[161] >> n;
 }
 
@@ -18523,7 +18534,7 @@ static WC_INLINE sp_digit sp_4096_div_word_162(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 24; i >= 1; i--) {
         t1 += t1 + (((sp_uint32)t0 >> 25) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint32)t0 << 1;
         t2 = (sp_digit)(((sp_uint32)(dv - t1)) >> 31);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -18721,9 +18732,9 @@ static int sp_4096_mod_exp_162(sp_digit* r, const sp_digit* a, const sp_digit* e
     }
 
     if (err == MP_OKAY) {
-        i = bits / 26;
-        c = bits % 26;
-        n = e[i--] << (26 - c);
+        i = (bits - 1) / 26;
+        c = ((bits - 1) % 26) + 1;
+        n = (sp_uint32)e[i--] << (26 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -18735,7 +18746,7 @@ static int sp_4096_mod_exp_162(sp_digit* r, const sp_digit* a, const sp_digit* e
             }
 
             y = (int)((n >> 25) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_4096_mont_mul_162(t[y^1], t[0], t[1], m, mp);
 
@@ -18797,9 +18808,9 @@ static int sp_4096_mod_exp_162(sp_digit* r, const sp_digit* a, const sp_digit* e
     }
 
     if (err == MP_OKAY) {
-        i = bits / 26;
-        c = bits % 26;
-        n = e[i--] << (26 - c);
+        i = (bits - 1) / 26;
+        c = ((bits - 1) % 26) + 1;
+        n = (sp_uint32)e[i--] << (26 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -18811,7 +18822,7 @@ static int sp_4096_mod_exp_162(sp_digit* r, const sp_digit* a, const sp_digit* e
             }
 
             y = (int)((n >> 25) & 1);
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_4096_mont_mul_162(t[y^1], t[0], t[1], m, mp);
 
@@ -18896,38 +18907,38 @@ static int sp_4096_mod_exp_162(sp_digit* r, const sp_digit* a, const sp_digit* e
             c = 26;
         }
         if (i < 162) {
-            n = e[i--] << (32 - c);
+            n = (sp_uint32)e[i--] << (32 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 4) {
-            n |= e[i--] << (6 - c);
+            n |= (sp_uint32)e[i--] << (6 - c);
             c += 26;
         }
         y = (int)((n >> 28) & 0xf);
-        n <<= 4;
+        n = (sp_uint32)n << 4;
         c -= 4;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 324);
         while ((i >= 0) || (c >= 4)) {
             if (c >= 4) {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
             else if (c == 0) {
-                n = e[i--] << 6;
+                n = (sp_uint32)e[i--] << 6;
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 22;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n = e[i--] << 6;
+                n = (sp_uint32)e[i--] << 6;
                 c = 4 - c;
                 y |= (byte)((n >> (32 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 26 - c;
             }
 
@@ -19023,10 +19034,10 @@ static int sp_4096_mod_exp_162_nb(sp_4096_mod_exp_162_ctx* ctx,
     case 4: /* BIT_INIT: index the most-significant exponent limb without
               reading off the end when bits is an exact multiple of the
               limb width. (bits-1) keeps i within the populated range and
-              c = (bits-1) % @bits + 1 keeps the shift in [1, @bits]. */
+              c = (bits-1) % 26 + 1 keeps the shift in [1, 26]. */
         ctx->i = (ctx->bits - 1) / 26;
         ctx->c = ((ctx->bits - 1) % 26) + 1;
-        ctx->n = e[ctx->i--] << (26 - ctx->c);
+        ctx->n = (sp_uint32)e[ctx->i--] << (26 - ctx->c);
         ctx->state = 5;
         break;
     case 5: /* BIT_NEXT: refill on word boundary, peel one exponent bit */
@@ -19039,7 +19050,7 @@ static int sp_4096_mod_exp_162_nb(sp_4096_mod_exp_162_ctx* ctx,
             ctx->c = 26;
         }
         ctx->y = (byte)((ctx->n >> 25) & 1);
-        ctx->n <<= 1;
+        ctx->n = (sp_uint32)ctx->n << 1;
         ctx->state = 6;
         break;
     case 6: /* MUL: t[y^1] = t[0] * t[1] in Montgomery form */
@@ -19610,7 +19621,7 @@ static int sp_4096_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 158; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -19635,7 +19646,7 @@ static int sp_4096_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 158; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 26 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -20088,7 +20099,7 @@ SP_NOINLINE static void sp_4096_lshift_162(sp_digit* r, const sp_digit* a,
     r[2] = (sp_digit)(((s << n) | (t >> (26U - n))) & 0x3ffffff);
     s = (sp_int_digit)(a[1]); t = (sp_int_digit)(a[0]);
     r[1] = (sp_digit)(((s << n) | (t >> (26U - n))) & 0x3ffffff);
-    r[0] = (sp_digit)((a[0] << n) & 0x3ffffff);
+    r[0] = (sp_digit)(((sp_uint32)a[0] << n) & 0x3ffffff);
 }
 
 /* Modular exponentiate 2 to the e mod m. (r = 2^e mod m)
@@ -20134,38 +20145,38 @@ static int sp_4096_mod_exp_2_162(sp_digit* r, const sp_digit* e, int bits, const
             c = 26;
         }
         if (i < 162) {
-            n = e[i--] << (32 - c);
+            n = (sp_uint32)e[i--] << (32 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 4) {
-            n |= e[i--] << (6 - c);
+            n |= (sp_uint32)e[i--] << (6 - c);
             c += 26;
         }
         y = (int)((n >> 28) & 0xf);
-        n <<= 4;
+        n = (sp_uint32)n << 4;
         c -= 4;
         sp_4096_lshift_162(r, norm, (byte)y);
         while ((i >= 0) || (c >= 4)) {
             if (c >= 4) {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
             else if (c == 0) {
-                n = e[i--] << 6;
+                n = (sp_uint32)e[i--] << 6;
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 22;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n = e[i--] << 6;
+                n = (sp_uint32)e[i--] << 6;
                 c = 4 - c;
                 y |= (byte)((n >> (32 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 26 - c;
             }
 
@@ -20753,7 +20764,7 @@ static void sp_256_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0x1fffffff;
         s = 29U - s;
         if (j + 1 >= size) {
@@ -20788,7 +20799,7 @@ static void sp_256_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 29) {
             r[j] &= 0x1fffffff;
             if (j + 1 >= size) {
@@ -20854,7 +20865,7 @@ static int sp_256_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 9; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -20879,7 +20890,7 @@ static int sp_256_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 9; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 29 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -21170,32 +21181,32 @@ static void sp_256_mont_reduce_9(sp_digit* a, const sp_digit* m, sp_digit mp)
 
     for (i = 0; i < 8; i++) {
         am = (sp_digit)(a[i] & 0x1fffffff);
-        a[i + 3] += (sp_digit)((am << 9) & 0x1fffffff);
+        a[i + 3] += (sp_digit)(((sp_uint32)am <<  9) & 0x1fffffff);
         a[i + 4] += am >> 20;
-        a[i + 6] += (sp_digit)((am << 18) & 0x1fffffff);
-        a[i + 7] += (am >> 11) - (sp_digit)((am << 21) & 0x1fffffff);
-        a[i + 8] += -(am >> 8) + (sp_digit)((am << 24) & 0x1fffffff);
+        a[i + 6] += (sp_digit)(((sp_uint32)am << 18) & 0x1fffffff);
+        a[i + 7] += (am >> 11) - (sp_digit)(((sp_uint32)am << 21) & 0x1fffffff);
+        a[i + 8] += -(am >> 8) + (sp_digit)(((sp_uint32)am << 24) & 0x1fffffff);
         a[i + 9] += am >> 5;
 
         a[i + 1] += a[i] >> 29;
     }
     am = (sp_digit)(a[8] & 0xffffff);
-    a[8 + 3] += (sp_digit)((am << 9) & 0x1fffffff);
+    a[8 + 3] += (sp_digit)(((sp_uint32)am <<  9) & 0x1fffffff);
     a[8 + 4] += am >> 20;
-    a[8 + 6] += (sp_digit)((am << 18) & 0x1fffffff);
-    a[8 + 7] += (am >> 11) - (sp_digit)((am << 21) & 0x1fffffff);
-    a[8 + 8] += -(am >> 8) + (sp_digit)((am << 24) & 0x1fffffff);
+    a[8 + 6] += (sp_digit)(((sp_uint32)am << 18) & 0x1fffffff);
+    a[8 + 7] += (am >> 11) - (sp_digit)(((sp_uint32)am << 21) & 0x1fffffff);
+    a[8 + 8] += -(am >> 8) + (sp_digit)(((sp_uint32)am << 24) & 0x1fffffff);
     a[8 + 9] += am >> 5;
 
-    a[0] = (a[ 8] >> 24) + (sp_digit)((a[ 9] << 5) & 0x1fffffff);
-    a[1] = (a[ 9] >> 24) + (sp_digit)((a[10] << 5) & 0x1fffffff);
-    a[2] = (a[10] >> 24) + (sp_digit)((a[11] << 5) & 0x1fffffff);
-    a[3] = (a[11] >> 24) + (sp_digit)((a[12] << 5) & 0x1fffffff);
-    a[4] = (a[12] >> 24) + (sp_digit)((a[13] << 5) & 0x1fffffff);
-    a[5] = (a[13] >> 24) + (sp_digit)((a[14] << 5) & 0x1fffffff);
-    a[6] = (a[14] >> 24) + (sp_digit)((a[15] << 5) & 0x1fffffff);
-    a[7] = (a[15] >> 24) + (sp_digit)((a[16] << 5) & 0x1fffffff);
-    a[8] = (a[16] >> 24) +  (a[17] << 5);
+    a[0] = (a[ 8] >> 24) + (sp_digit)(((sp_uint32)a[ 9] << 5) & 0x1fffffff);
+    a[1] = (a[ 9] >> 24) + (sp_digit)(((sp_uint32)a[10] << 5) & 0x1fffffff);
+    a[2] = (a[10] >> 24) + (sp_digit)(((sp_uint32)a[11] << 5) & 0x1fffffff);
+    a[3] = (a[11] >> 24) + (sp_digit)(((sp_uint32)a[12] << 5) & 0x1fffffff);
+    a[4] = (a[12] >> 24) + (sp_digit)(((sp_uint32)a[13] << 5) & 0x1fffffff);
+    a[5] = (a[13] >> 24) + (sp_digit)(((sp_uint32)a[14] << 5) & 0x1fffffff);
+    a[6] = (a[14] >> 24) + (sp_digit)(((sp_uint32)a[15] << 5) & 0x1fffffff);
+    a[7] = (a[15] >> 24) + (sp_digit)(((sp_uint32)a[16] << 5) & 0x1fffffff);
+    a[8] = (a[16] >> 24) +  (sp_digit)((sp_uint32)a[17] << 5);
 
     a[1] += a[0] >> 29; a[0] &= 0x1fffffff;
     a[2] += a[1] >> 29; a[1] &= 0x1fffffff;
@@ -21304,7 +21315,7 @@ static void sp_256_mont_inv_9(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 9);
     for (i=254; i>=0; i--) {
         sp_256_mont_sqr_9(t, t, p256_mod, p256_mp_mod);
-        if (p256_mod_minus_2[i / 32] & ((sp_digit)1 << (i % 32)))
+        if (p256_mod_minus_2[i / 32] & ((sp_uint32)1 << (i % 32)))
             sp_256_mont_mul_9(t, t, a, p256_mod, p256_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 9);
@@ -21524,17 +21535,17 @@ SP_NOINLINE static void sp_256_rshift1_9(sp_digit* r, const sp_digit* a)
     int i;
 
     for (i=0; i<8; i++) {
-        r[i] = (a[i] >> 1) + (sp_digit)((a[i + 1] << 28) & 0x1fffffff);
+        r[i] = (a[i] >> 1) + (sp_digit)(((sp_uint32)a[i + 1] << 28) & 0x1fffffff);
     }
 #else
-    r[0] = (a[0] >> 1) + (sp_digit)((a[1] << 28) & 0x1fffffff);
-    r[1] = (a[1] >> 1) + (sp_digit)((a[2] << 28) & 0x1fffffff);
-    r[2] = (a[2] >> 1) + (sp_digit)((a[3] << 28) & 0x1fffffff);
-    r[3] = (a[3] >> 1) + (sp_digit)((a[4] << 28) & 0x1fffffff);
-    r[4] = (a[4] >> 1) + (sp_digit)((a[5] << 28) & 0x1fffffff);
-    r[5] = (a[5] >> 1) + (sp_digit)((a[6] << 28) & 0x1fffffff);
-    r[6] = (a[6] >> 1) + (sp_digit)((a[7] << 28) & 0x1fffffff);
-    r[7] = (a[7] >> 1) + (sp_digit)((a[8] << 28) & 0x1fffffff);
+    r[0] = (a[0] >> 1) + (sp_digit)(((sp_uint32)a[1] << 28) & 0x1fffffff);
+    r[1] = (a[1] >> 1) + (sp_digit)(((sp_uint32)a[2] << 28) & 0x1fffffff);
+    r[2] = (a[2] >> 1) + (sp_digit)(((sp_uint32)a[3] << 28) & 0x1fffffff);
+    r[3] = (a[3] >> 1) + (sp_digit)(((sp_uint32)a[4] << 28) & 0x1fffffff);
+    r[4] = (a[4] >> 1) + (sp_digit)(((sp_uint32)a[5] << 28) & 0x1fffffff);
+    r[5] = (a[5] >> 1) + (sp_digit)(((sp_uint32)a[6] << 28) & 0x1fffffff);
+    r[6] = (a[6] >> 1) + (sp_digit)(((sp_uint32)a[7] << 28) & 0x1fffffff);
+    r[7] = (a[7] >> 1) + (sp_digit)(((sp_uint32)a[8] << 28) & 0x1fffffff);
 #endif
     r[8] = a[8] >> 1;
 }
@@ -22156,25 +22167,25 @@ static int sp_256_mod_mul_norm_9(sp_digit* r, const sp_digit* a, const sp_digit*
 
         r[0] = (sp_digit)(t[0]) & 0x1fffffffL;
         r[1] = (sp_digit)(t[0] >> 29U);
-        r[1] |= (sp_digit)(t[1] << 3U);
+        r[1] |= (sp_digit)((sp_uint32)t[1] << 3U);
         r[1] &= 0x1fffffffL;
         r[2] = (sp_digit)(t[1] >> 26U);
-        r[2] |= (sp_digit)(t[2] << 6U);
+        r[2] |= (sp_digit)((sp_uint32)t[2] << 6U);
         r[2] &= 0x1fffffffL;
         r[3] = (sp_digit)(t[2] >> 23U);
-        r[3] |= (sp_digit)(t[3] << 9U);
+        r[3] |= (sp_digit)((sp_uint32)t[3] << 9U);
         r[3] &= 0x1fffffffL;
         r[4] = (sp_digit)(t[3] >> 20U);
-        r[4] |= (sp_digit)(t[4] << 12U);
+        r[4] |= (sp_digit)((sp_uint32)t[4] << 12U);
         r[4] &= 0x1fffffffL;
         r[5] = (sp_digit)(t[4] >> 17U);
-        r[5] |= (sp_digit)(t[5] << 15U);
+        r[5] |= (sp_digit)((sp_uint32)t[5] << 15U);
         r[5] &= 0x1fffffffL;
         r[6] = (sp_digit)(t[5] >> 14U);
-        r[6] |= (sp_digit)(t[6] << 18U);
+        r[6] |= (sp_digit)((sp_uint32)t[6] << 18U);
         r[6] &= 0x1fffffffL;
         r[7] = (sp_digit)(t[6] >> 11U);
-        r[7] |= (sp_digit)(t[7] << 21U);
+        r[7] |= (sp_digit)((sp_uint32)t[7] << 21U);
         r[7] &= 0x1fffffffL;
         r[8] = (sp_digit)(t[7] >> 8U);
     }
@@ -22234,7 +22245,7 @@ static int sp_256_ecc_mulmod_9(sp_point_256* r, const sp_point_256* g,
     if (err == MP_OKAY) {
         i = 8;
         c = 24;
-        n = k[i--] << (29 - c);
+        n = (sp_uint32)k[i--] << (29 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1)
@@ -22245,7 +22256,7 @@ static int sp_256_ecc_mulmod_9(sp_point_256* r, const sp_point_256* g,
             }
 
             y = (n >> 28) & 1;
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_256_proj_point_add_9(&t[y^1], &t[0], &t[1], tmp);
 
@@ -22304,7 +22315,7 @@ static int sp_256_ecc_mulmod_9_nb(sp_ecc_ctx_t* sp_ctx, sp_point_256* r,
         XMEMSET(ctx->t, 0, sizeof(sp_point_256) * 3);
         ctx->i = 8;
         ctx->c = 24;
-        ctx->n = k[ctx->i--] << (29 - ctx->c);
+        ctx->n = (sp_uint32)k[ctx->i--] << (29 - ctx->c);
 
         /* t[0] = {0, 0, 1} * norm */
         ctx->t[0].infinity = 1;
@@ -22334,7 +22345,7 @@ static int sp_256_ecc_mulmod_9_nb(sp_ecc_ctx_t* sp_ctx, sp_point_256* r,
             ctx->c = 29;
         }
         ctx->y = (ctx->n >> 28) & 1;
-        ctx->n <<= 1;
+        ctx->n = (sp_uint32)ctx->n << 1;
         XMEMSET(&ctx->add_ctx, 0, sizeof(ctx->add_ctx));
         ctx->state = 5;
         break;
@@ -22737,7 +22748,7 @@ static void sp_256_ecc_recode_6_9(const sp_digit* k, ecc_recode_256* v)
         }
         else if (++j < 9) {
             n = k[j];
-            y |= (word8)((n << (29 - o)) & 0x3f);
+            y |= (word8)(((sp_uint32)n << (29 - o)) & 0x3f);
             o -= 23;
             n >>= o;
         }
@@ -24987,7 +24998,7 @@ static void sp_256_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     r[0] = 0;
     for (i = n-1; i >= 0; i--) {
-        r[j] |= (((sp_digit)a[i]) << s);
+        r[j] |= (((sp_uint32)a[i]) << s);
         if (s >= 21U) {
             r[j] &= 0x1fffffff;
             s = 29U - s;
@@ -25208,7 +25219,7 @@ static void sp_256_to_bin_9(sp_digit* r, byte* a)
     for (i=0; i<9 && j>=0; i++) {
         b = 0;
         /* lint allow cast of mismatch sp_digit and int */
-        a[j--] |= (byte)(r[i] << s); /*lint !e9033*/
+        a[j--] |= (byte)((sp_uint32)r[i] << s); /*lint !e9033*/
         b += 8 - s;
         if (j < 0) {
             break;
@@ -25332,18 +25343,18 @@ SP_NOINLINE static void sp_256_rshift_9(sp_digit* r, const sp_digit* a,
 
 #ifdef WOLFSSL_SP_SMALL
     for (i=0; i<8; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (29 - n))) & 0x1fffffff);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint32)a[i + 1] << (29 - n))) & 0x1fffffff);
     }
 #else
     for (i=0; i<8; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (29 - n)) & 0x1fffffff);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (29 - n)) & 0x1fffffff);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (29 - n)) & 0x1fffffff);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (29 - n)) & 0x1fffffff);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (29 - n)) & 0x1fffffff);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (29 - n)) & 0x1fffffff);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (29 - n)) & 0x1fffffff);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (29 - n)) & 0x1fffffff);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint32)a[i+1] << (29 - n)) & 0x1fffffff);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint32)a[i+2] << (29 - n)) & 0x1fffffff);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint32)a[i+3] << (29 - n)) & 0x1fffffff);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint32)a[i+4] << (29 - n)) & 0x1fffffff);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint32)a[i+5] << (29 - n)) & 0x1fffffff);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint32)a[i+6] << (29 - n)) & 0x1fffffff);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint32)a[i+7] << (29 - n)) & 0x1fffffff);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint32)a[i+8] << (29 - n)) & 0x1fffffff);
     }
 #endif /* WOLFSSL_SP_SMALL */
     r[8] = a[8] >> n;
@@ -25403,7 +25414,7 @@ SP_NOINLINE static void sp_256_lshift_18(sp_digit* r, const sp_digit* a,
 
     r[18] = a[17] >> (29 - n);
     for (i=17; i>0; i--) {
-        r[i] = (sp_digit)(((a[i] << n) | (a[i-1] >> (29 - n))) & 0x1fffffff);
+        r[i] = (sp_digit)((((sp_uint32)a[i] << n) | (a[i-1] >> (29 - n))) & 0x1fffffff);
     }
 #else
     sp_int_digit s;
@@ -25446,7 +25457,7 @@ SP_NOINLINE static void sp_256_lshift_18(sp_digit* r, const sp_digit* a,
     s = (sp_int_digit)(a[1]); t = (sp_int_digit)(a[0]);
     r[1] = (sp_digit)(((s << n) | (t >> (29U - n))) & 0x1fffffff);
 #endif /* WOLFSSL_SP_SMALL */
-    r[0] = (sp_digit)((a[0] << n) & 0x1fffffff);
+    r[0] = (sp_digit)(((sp_uint32)a[0] << n) & 0x1fffffff);
 }
 
 /* Divide d in a and put remainder into r (m*d + r = a)
@@ -27681,7 +27692,7 @@ static void sp_384_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0x3ffffff;
         s = 26U - s;
         if (j + 1 >= size) {
@@ -27716,7 +27727,7 @@ static void sp_384_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 26) {
             r[j] &= 0x3ffffff;
             if (j + 1 >= size) {
@@ -27782,7 +27793,7 @@ static int sp_384_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 15; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -27807,7 +27818,7 @@ static int sp_384_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 15; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 26 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -28113,43 +28124,43 @@ static void sp_384_mont_reduce_15(sp_digit* a, const sp_digit* m, sp_digit mp)
     (void)mp;
 
     for (i = 0; i < 14; i++) {
-        am = (sp_digit)((a[i] * 0x1) & 0x3ffffff);
-        a[i +  1] += (sp_digit)((am << 6) & 0x3ffffff);
+        am = (sp_digit)((((sp_uint32)a[i]) * 0x1) & 0x3ffffff);
+        a[i +  1] += (sp_digit)(((sp_uint32)am <<  6) & 0x3ffffff);
         a[i +  2] += am >> 20;
-        a[i +  3] -= (sp_digit)((am << 18) & 0x3ffffff);
+        a[i +  3] -= (sp_digit)(((sp_uint32)am << 18) & 0x3ffffff);
         a[i +  4] -= am >> 8;
-        a[i +  4] -= (sp_digit)((am << 24) & 0x3ffffff);
+        a[i +  4] -= (sp_digit)(((sp_uint32)am << 24) & 0x3ffffff);
         a[i +  5] -= am >> 2;
-        a[i + 14] += (sp_digit)((am << 20) & 0x3ffffff);
+        a[i + 14] += (sp_digit)(((sp_uint32)am << 20) & 0x3ffffff);
         a[i + 15] += am >> 6;
 
         a[i +  1] += a[i] >> 26;
     }
-    am = (sp_digit)((a[14] * 0x1) & 0xfffff);
-    a[14 +  1] += (sp_digit)((am << 6) & 0x3ffffff);
+    am = (sp_digit)((((sp_uint32)a[14]) * 0x1) & 0xfffff);
+    a[14 +  1] += (sp_digit)(((sp_uint32)am <<  6) & 0x3ffffff);
     a[14 +  2] += am >> 20;
-    a[14 +  3] -= (sp_digit)((am << 18) & 0x3ffffff);
+    a[14 +  3] -= (sp_digit)(((sp_uint32)am << 18) & 0x3ffffff);
     a[14 +  4] -= am >> 8;
-    a[14 +  4] -= (sp_digit)((am << 24) & 0x3ffffff);
+    a[14 +  4] -= (sp_digit)(((sp_uint32)am << 24) & 0x3ffffff);
     a[14 +  5] -= am >> 2;
-    a[14 + 14] += (sp_digit)((am << 20) & 0x3ffffff);
+    a[14 + 14] += (sp_digit)(((sp_uint32)am << 20) & 0x3ffffff);
     a[14 + 15] += am >> 6;
 
-    a[0] = (a[14] >> 20) + (sp_digit)((a[15] << 6) & 0x3ffffff);
-    a[1] = (a[15] >> 20) + (sp_digit)((a[16] << 6) & 0x3ffffff);
-    a[2] = (a[16] >> 20) + (sp_digit)((a[17] << 6) & 0x3ffffff);
-    a[3] = (a[17] >> 20) + (sp_digit)((a[18] << 6) & 0x3ffffff);
-    a[4] = (a[18] >> 20) + (sp_digit)((a[19] << 6) & 0x3ffffff);
-    a[5] = (a[19] >> 20) + (sp_digit)((a[20] << 6) & 0x3ffffff);
-    a[6] = (a[20] >> 20) + (sp_digit)((a[21] << 6) & 0x3ffffff);
-    a[7] = (a[21] >> 20) + (sp_digit)((a[22] << 6) & 0x3ffffff);
-    a[8] = (a[22] >> 20) + (sp_digit)((a[23] << 6) & 0x3ffffff);
-    a[9] = (a[23] >> 20) + (sp_digit)((a[24] << 6) & 0x3ffffff);
-    a[10] = (a[24] >> 20) + (sp_digit)((a[25] << 6) & 0x3ffffff);
-    a[11] = (a[25] >> 20) + (sp_digit)((a[26] << 6) & 0x3ffffff);
-    a[12] = (a[26] >> 20) + (sp_digit)((a[27] << 6) & 0x3ffffff);
-    a[13] = (a[27] >> 20) + (sp_digit)((a[28] << 6) & 0x3ffffff);
-    a[14] = (a[14 + 14] >> 20) +  (a[29] << 6);
+    a[0] = (a[14] >> 20) + (sp_digit)(((sp_uint32)a[15] << 6) & 0x3ffffff);
+    a[1] = (a[15] >> 20) + (sp_digit)(((sp_uint32)a[16] << 6) & 0x3ffffff);
+    a[2] = (a[16] >> 20) + (sp_digit)(((sp_uint32)a[17] << 6) & 0x3ffffff);
+    a[3] = (a[17] >> 20) + (sp_digit)(((sp_uint32)a[18] << 6) & 0x3ffffff);
+    a[4] = (a[18] >> 20) + (sp_digit)(((sp_uint32)a[19] << 6) & 0x3ffffff);
+    a[5] = (a[19] >> 20) + (sp_digit)(((sp_uint32)a[20] << 6) & 0x3ffffff);
+    a[6] = (a[20] >> 20) + (sp_digit)(((sp_uint32)a[21] << 6) & 0x3ffffff);
+    a[7] = (a[21] >> 20) + (sp_digit)(((sp_uint32)a[22] << 6) & 0x3ffffff);
+    a[8] = (a[22] >> 20) + (sp_digit)(((sp_uint32)a[23] << 6) & 0x3ffffff);
+    a[9] = (a[23] >> 20) + (sp_digit)(((sp_uint32)a[24] << 6) & 0x3ffffff);
+    a[10] = (a[24] >> 20) + (sp_digit)(((sp_uint32)a[25] << 6) & 0x3ffffff);
+    a[11] = (a[25] >> 20) + (sp_digit)(((sp_uint32)a[26] << 6) & 0x3ffffff);
+    a[12] = (a[26] >> 20) + (sp_digit)(((sp_uint32)a[27] << 6) & 0x3ffffff);
+    a[13] = (a[27] >> 20) + (sp_digit)(((sp_uint32)a[28] << 6) & 0x3ffffff);
+    a[14] = (a[14 + 14] >> 20) +  (sp_digit)((sp_uint32)a[29] << 6);
 
     a[1] += a[0] >> 26; a[0] &= 0x3ffffff;
     a[2] += a[1] >> 26; a[1] &= 0x3ffffff;
@@ -28276,7 +28287,7 @@ static void sp_384_mont_inv_15(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 15);
     for (i=382; i>=0; i--) {
         sp_384_mont_sqr_15(t, t, p384_mod, p384_mp_mod);
-        if (p384_mod_minus_2[i / 32] & ((sp_digit)1 << (i % 32)))
+        if (p384_mod_minus_2[i / 32] & ((sp_uint32)1 << (i % 32)))
             sp_384_mont_mul_15(t, t, a, p384_mod, p384_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 15);
@@ -28518,23 +28529,23 @@ SP_NOINLINE static void sp_384_rshift1_15(sp_digit* r, const sp_digit* a)
     int i;
 
     for (i=0; i<14; i++) {
-        r[i] = (a[i] >> 1) + (sp_digit)((a[i + 1] << 25) & 0x3ffffff);
+        r[i] = (a[i] >> 1) + (sp_digit)(((sp_uint32)a[i + 1] << 25) & 0x3ffffff);
     }
 #else
-    r[0] = (a[0] >> 1) + (sp_digit)((a[1] << 25) & 0x3ffffff);
-    r[1] = (a[1] >> 1) + (sp_digit)((a[2] << 25) & 0x3ffffff);
-    r[2] = (a[2] >> 1) + (sp_digit)((a[3] << 25) & 0x3ffffff);
-    r[3] = (a[3] >> 1) + (sp_digit)((a[4] << 25) & 0x3ffffff);
-    r[4] = (a[4] >> 1) + (sp_digit)((a[5] << 25) & 0x3ffffff);
-    r[5] = (a[5] >> 1) + (sp_digit)((a[6] << 25) & 0x3ffffff);
-    r[6] = (a[6] >> 1) + (sp_digit)((a[7] << 25) & 0x3ffffff);
-    r[7] = (a[7] >> 1) + (sp_digit)((a[8] << 25) & 0x3ffffff);
-    r[8] = (a[8] >> 1) + (sp_digit)((a[9] << 25) & 0x3ffffff);
-    r[9] = (a[9] >> 1) + (sp_digit)((a[10] << 25) & 0x3ffffff);
-    r[10] = (a[10] >> 1) + (sp_digit)((a[11] << 25) & 0x3ffffff);
-    r[11] = (a[11] >> 1) + (sp_digit)((a[12] << 25) & 0x3ffffff);
-    r[12] = (a[12] >> 1) + (sp_digit)((a[13] << 25) & 0x3ffffff);
-    r[13] = (a[13] >> 1) + (sp_digit)((a[14] << 25) & 0x3ffffff);
+    r[0] = (a[0] >> 1) + (sp_digit)(((sp_uint32)a[1] << 25) & 0x3ffffff);
+    r[1] = (a[1] >> 1) + (sp_digit)(((sp_uint32)a[2] << 25) & 0x3ffffff);
+    r[2] = (a[2] >> 1) + (sp_digit)(((sp_uint32)a[3] << 25) & 0x3ffffff);
+    r[3] = (a[3] >> 1) + (sp_digit)(((sp_uint32)a[4] << 25) & 0x3ffffff);
+    r[4] = (a[4] >> 1) + (sp_digit)(((sp_uint32)a[5] << 25) & 0x3ffffff);
+    r[5] = (a[5] >> 1) + (sp_digit)(((sp_uint32)a[6] << 25) & 0x3ffffff);
+    r[6] = (a[6] >> 1) + (sp_digit)(((sp_uint32)a[7] << 25) & 0x3ffffff);
+    r[7] = (a[7] >> 1) + (sp_digit)(((sp_uint32)a[8] << 25) & 0x3ffffff);
+    r[8] = (a[8] >> 1) + (sp_digit)(((sp_uint32)a[9] << 25) & 0x3ffffff);
+    r[9] = (a[9] >> 1) + (sp_digit)(((sp_uint32)a[10] << 25) & 0x3ffffff);
+    r[10] = (a[10] >> 1) + (sp_digit)(((sp_uint32)a[11] << 25) & 0x3ffffff);
+    r[11] = (a[11] >> 1) + (sp_digit)(((sp_uint32)a[12] << 25) & 0x3ffffff);
+    r[12] = (a[12] >> 1) + (sp_digit)(((sp_uint32)a[13] << 25) & 0x3ffffff);
+    r[13] = (a[13] >> 1) + (sp_digit)(((sp_uint32)a[14] << 25) & 0x3ffffff);
 #endif
     r[14] = a[14] >> 1;
 }
@@ -29188,39 +29199,39 @@ static int sp_384_mod_mul_norm_15(sp_digit* r, const sp_digit* a, const sp_digit
 
         r[0] = (sp_digit)(t[0]) & 0x3ffffffL;
         r[1] = (sp_digit)(t[0] >> 26U);
-        r[1] |= (sp_digit)(t[1] << 6U);
+        r[1] |= (sp_digit)((sp_uint32)t[1] << 6U);
         r[1] &= 0x3ffffffL;
         r[2] = (sp_digit)(t[1] >> 20U);
-        r[2] |= (sp_digit)(t[2] << 12U);
+        r[2] |= (sp_digit)((sp_uint32)t[2] << 12U);
         r[2] &= 0x3ffffffL;
         r[3] = (sp_digit)(t[2] >> 14U);
-        r[3] |= (sp_digit)(t[3] << 18U);
+        r[3] |= (sp_digit)((sp_uint32)t[3] << 18U);
         r[3] &= 0x3ffffffL;
         r[4] = (sp_digit)(t[3] >> 8U);
-        r[4] |= (sp_digit)(t[4] << 24U);
+        r[4] |= (sp_digit)((sp_uint32)t[4] << 24U);
         r[4] &= 0x3ffffffL;
         r[5] = (sp_digit)(t[4] >> 2U) & 0x3ffffffL;
         r[6] = (sp_digit)(t[4] >> 28U);
-        r[6] |= (sp_digit)(t[5] << 4U);
+        r[6] |= (sp_digit)((sp_uint32)t[5] << 4U);
         r[6] &= 0x3ffffffL;
         r[7] = (sp_digit)(t[5] >> 22U);
-        r[7] |= (sp_digit)(t[6] << 10U);
+        r[7] |= (sp_digit)((sp_uint32)t[6] << 10U);
         r[7] &= 0x3ffffffL;
         r[8] = (sp_digit)(t[6] >> 16U);
-        r[8] |= (sp_digit)(t[7] << 16U);
+        r[8] |= (sp_digit)((sp_uint32)t[7] << 16U);
         r[8] &= 0x3ffffffL;
         r[9] = (sp_digit)(t[7] >> 10U);
-        r[9] |= (sp_digit)(t[8] << 22U);
+        r[9] |= (sp_digit)((sp_uint32)t[8] << 22U);
         r[9] &= 0x3ffffffL;
         r[10] = (sp_digit)(t[8] >> 4U) & 0x3ffffffL;
         r[11] = (sp_digit)(t[8] >> 30U);
-        r[11] |= (sp_digit)(t[9] << 2U);
+        r[11] |= (sp_digit)((sp_uint32)t[9] << 2U);
         r[11] &= 0x3ffffffL;
         r[12] = (sp_digit)(t[9] >> 24U);
-        r[12] |= (sp_digit)(t[10] << 8U);
+        r[12] |= (sp_digit)((sp_uint32)t[10] << 8U);
         r[12] &= 0x3ffffffL;
         r[13] = (sp_digit)(t[10] >> 18U);
-        r[13] |= (sp_digit)(t[11] << 14U);
+        r[13] |= (sp_digit)((sp_uint32)t[11] << 14U);
         r[13] &= 0x3ffffffL;
         r[14] = (sp_digit)(t[11] >> 12U);
     }
@@ -29280,7 +29291,7 @@ static int sp_384_ecc_mulmod_15(sp_point_384* r, const sp_point_384* g,
     if (err == MP_OKAY) {
         i = 14;
         c = 20;
-        n = k[i--] << (26 - c);
+        n = (sp_uint32)k[i--] << (26 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1)
@@ -29291,7 +29302,7 @@ static int sp_384_ecc_mulmod_15(sp_point_384* r, const sp_point_384* g,
             }
 
             y = (n >> 25) & 1;
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_384_proj_point_add_15(&t[y^1], &t[0], &t[1], tmp);
 
@@ -29350,7 +29361,7 @@ static int sp_384_ecc_mulmod_15_nb(sp_ecc_ctx_t* sp_ctx, sp_point_384* r,
         XMEMSET(ctx->t, 0, sizeof(sp_point_384) * 3);
         ctx->i = 14;
         ctx->c = 20;
-        ctx->n = k[ctx->i--] << (26 - ctx->c);
+        ctx->n = (sp_uint32)k[ctx->i--] << (26 - ctx->c);
 
         /* t[0] = {0, 0, 1} * norm */
         ctx->t[0].infinity = 1;
@@ -29380,7 +29391,7 @@ static int sp_384_ecc_mulmod_15_nb(sp_ecc_ctx_t* sp_ctx, sp_point_384* r,
             ctx->c = 26;
         }
         ctx->y = (ctx->n >> 25) & 1;
-        ctx->n <<= 1;
+        ctx->n = (sp_uint32)ctx->n << 1;
         XMEMSET(&ctx->add_ctx, 0, sizeof(ctx->add_ctx));
         ctx->state = 5;
         break;
@@ -29795,7 +29806,7 @@ static void sp_384_ecc_recode_6_15(const sp_digit* k, ecc_recode_384* v)
         }
         else if (++j < 15) {
             n = k[j];
-            y |= (word8)((n << (26 - o)) & 0x3f);
+            y |= (word8)(((sp_uint32)n << (26 - o)) & 0x3f);
             o -= 20;
             n >>= o;
         }
@@ -32617,7 +32628,7 @@ static void sp_384_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     r[0] = 0;
     for (i = n-1; i >= 0; i--) {
-        r[j] |= (((sp_digit)a[i]) << s);
+        r[j] |= (((sp_uint32)a[i]) << s);
         if (s >= 18U) {
             r[j] &= 0x3ffffff;
             s = 26U - s;
@@ -32838,7 +32849,7 @@ static void sp_384_to_bin_15(sp_digit* r, byte* a)
     for (i=0; i<15 && j>=0; i++) {
         b = 0;
         /* lint allow cast of mismatch sp_digit and int */
-        a[j--] |= (byte)(r[i] << s); /*lint !e9033*/
+        a[j--] |= (byte)((sp_uint32)r[i] << s); /*lint !e9033*/
         b += 8 - s;
         if (j < 0) {
             break;
@@ -32962,25 +32973,25 @@ SP_NOINLINE static void sp_384_rshift_15(sp_digit* r, const sp_digit* a,
 
 #ifdef WOLFSSL_SP_SMALL
     for (i=0; i<14; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (26 - n))) & 0x3ffffff);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint32)a[i + 1] << (26 - n))) & 0x3ffffff);
     }
 #else
     for (i=0; i<8; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (26 - n)) & 0x3ffffff);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (26 - n)) & 0x3ffffff);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (26 - n)) & 0x3ffffff);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (26 - n)) & 0x3ffffff);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (26 - n)) & 0x3ffffff);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (26 - n)) & 0x3ffffff);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (26 - n)) & 0x3ffffff);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (26 - n)) & 0x3ffffff);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint32)a[i+1] << (26 - n)) & 0x3ffffff);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint32)a[i+2] << (26 - n)) & 0x3ffffff);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint32)a[i+3] << (26 - n)) & 0x3ffffff);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint32)a[i+4] << (26 - n)) & 0x3ffffff);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint32)a[i+5] << (26 - n)) & 0x3ffffff);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint32)a[i+6] << (26 - n)) & 0x3ffffff);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint32)a[i+7] << (26 - n)) & 0x3ffffff);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint32)a[i+8] << (26 - n)) & 0x3ffffff);
     }
-    r[8] = (a[8] >> n) | (sp_digit)((a[9] << (26 - n)) & 0x3ffffff);
-    r[9] = (a[9] >> n) | (sp_digit)((a[10] << (26 - n)) & 0x3ffffff);
-    r[10] = (a[10] >> n) | (sp_digit)((a[11] << (26 - n)) & 0x3ffffff);
-    r[11] = (a[11] >> n) | (sp_digit)((a[12] << (26 - n)) & 0x3ffffff);
-    r[12] = (a[12] >> n) | (sp_digit)((a[13] << (26 - n)) & 0x3ffffff);
-    r[13] = (a[13] >> n) | (sp_digit)((a[14] << (26 - n)) & 0x3ffffff);
+    r[8] = (a[8] >> n) | (sp_digit)(((sp_uint32)a[9] << (26 - n)) & 0x3ffffff);
+    r[9] = (a[9] >> n) | (sp_digit)(((sp_uint32)a[10] << (26 - n)) & 0x3ffffff);
+    r[10] = (a[10] >> n) | (sp_digit)(((sp_uint32)a[11] << (26 - n)) & 0x3ffffff);
+    r[11] = (a[11] >> n) | (sp_digit)(((sp_uint32)a[12] << (26 - n)) & 0x3ffffff);
+    r[12] = (a[12] >> n) | (sp_digit)(((sp_uint32)a[13] << (26 - n)) & 0x3ffffff);
+    r[13] = (a[13] >> n) | (sp_digit)(((sp_uint32)a[14] << (26 - n)) & 0x3ffffff);
 #endif /* WOLFSSL_SP_SMALL */
     r[14] = a[14] >> n;
 }
@@ -33051,7 +33062,7 @@ SP_NOINLINE static void sp_384_lshift_30(sp_digit* r, const sp_digit* a,
 
     r[30] = a[29] >> (26 - n);
     for (i=29; i>0; i--) {
-        r[i] = (sp_digit)(((a[i] << n) | (a[i-1] >> (26 - n))) & 0x3ffffff);
+        r[i] = (sp_digit)((((sp_uint32)a[i] << n) | (a[i-1] >> (26 - n))) & 0x3ffffff);
     }
 #else
     sp_int_digit s;
@@ -33118,7 +33129,7 @@ SP_NOINLINE static void sp_384_lshift_30(sp_digit* r, const sp_digit* a,
     s = (sp_int_digit)(a[1]); t = (sp_int_digit)(a[0]);
     r[1] = (sp_digit)(((s << n) | (t >> (26U - n))) & 0x3ffffff);
 #endif /* WOLFSSL_SP_SMALL */
-    r[0] = (sp_digit)((a[0] << n) & 0x3ffffff);
+    r[0] = (sp_digit)(((sp_uint32)a[0] << n) & 0x3ffffff);
 }
 
 /* Divide d in a and put remainder into r (m*d + r = a)
@@ -34983,7 +34994,7 @@ static void sp_521_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0x1ffffff;
         s = 25U - s;
         if (j + 1 >= size) {
@@ -35018,7 +35029,7 @@ static void sp_521_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 25) {
             r[j] &= 0x1ffffff;
             if (j + 1 >= size) {
@@ -35084,7 +35095,7 @@ static int sp_521_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 21; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -35109,7 +35120,7 @@ static int sp_521_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 21; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 25 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -35197,10 +35208,10 @@ static void sp_521_mont_reduce_21(sp_digit* a, const sp_digit* m, sp_digit mp)
     (void)mp;
 
     for (i = 0; i < 20; i++) {
-        a[i] += (sp_digit)(((a[20 + i] >> 21) + (a[20 + i + 1] << 4)) & 0x1ffffff);
+        a[i] += (sp_digit)(((a[20 + i] >> 21) + ((sp_uint32)a[20 + i + 1] << 4)) & 0x1ffffff);
     }
     a[20] &= 0x1fffff;
-    a[20] += (sp_digit)(((a[40] >> 21) + (a[41] << 4)) & 0x1ffffff);
+    a[20] += (sp_digit)(((a[40] >> 21) + ((sp_uint32)a[41] << 4)) & 0x1ffffff);
 
     sp_521_norm_21(a);
 
@@ -35368,12 +35379,12 @@ static void sp_521_mont_shift_21(sp_digit* r, const sp_digit* a)
     s = a[21];
     n = a[20] >> 21;
     for (i = 0; i < 20; i++) {
-        n += (sp_digit)((s & 0x1ffffff) << 4);
+        n += (sp_digit)((sp_uint32)(s & 0x1ffffff) << 4);
         r[i] = (sp_digit)(n & 0x1ffffff);
         n >>= 25;
         s = a[22 + i] + (s >> 25);
     }
-    n += s << 4;
+    n += (sp_uint32)s << 4;
     r[20] = n;
 #else
     sp_digit n;
@@ -35382,32 +35393,32 @@ static void sp_521_mont_shift_21(sp_digit* r, const sp_digit* a)
 
     s = a[21]; n = a[20] >> 21;
     for (i = 0; i < 16; i += 8) {
-        n += (sp_digit)((s & 0x1ffffff) << 4); r[i+0] = (sp_digit)(n & 0x1ffffff);
+        n += (sp_digit)((sp_uint32)(s & 0x1ffffff) << 4); r[i+0] = (sp_digit)(n & 0x1ffffff);
         n >>= 25; s = a[i+22] + (s >> 25);
-        n += (sp_digit)((s & 0x1ffffff) << 4); r[i+1] = (sp_digit)(n & 0x1ffffff);
+        n += (sp_digit)((sp_uint32)(s & 0x1ffffff) << 4); r[i+1] = (sp_digit)(n & 0x1ffffff);
         n >>= 25; s = a[i+23] + (s >> 25);
-        n += (sp_digit)((s & 0x1ffffff) << 4); r[i+2] = (sp_digit)(n & 0x1ffffff);
+        n += (sp_digit)((sp_uint32)(s & 0x1ffffff) << 4); r[i+2] = (sp_digit)(n & 0x1ffffff);
         n >>= 25; s = a[i+24] + (s >> 25);
-        n += (sp_digit)((s & 0x1ffffff) << 4); r[i+3] = (sp_digit)(n & 0x1ffffff);
+        n += (sp_digit)((sp_uint32)(s & 0x1ffffff) << 4); r[i+3] = (sp_digit)(n & 0x1ffffff);
         n >>= 25; s = a[i+25] + (s >> 25);
-        n += (sp_digit)((s & 0x1ffffff) << 4); r[i+4] = (sp_digit)(n & 0x1ffffff);
+        n += (sp_digit)((sp_uint32)(s & 0x1ffffff) << 4); r[i+4] = (sp_digit)(n & 0x1ffffff);
         n >>= 25; s = a[i+26] + (s >> 25);
-        n += (sp_digit)((s & 0x1ffffff) << 4); r[i+5] = (sp_digit)(n & 0x1ffffff);
+        n += (sp_digit)((sp_uint32)(s & 0x1ffffff) << 4); r[i+5] = (sp_digit)(n & 0x1ffffff);
         n >>= 25; s = a[i+27] + (s >> 25);
-        n += (sp_digit)((s & 0x1ffffff) << 4); r[i+6] = (sp_digit)(n & 0x1ffffff);
+        n += (sp_digit)((sp_uint32)(s & 0x1ffffff) << 4); r[i+6] = (sp_digit)(n & 0x1ffffff);
         n >>= 25; s = a[i+28] + (s >> 25);
-        n += (sp_digit)((s & 0x1ffffff) << 4); r[i+7] = (sp_digit)(n & 0x1ffffff);
+        n += (sp_digit)((sp_uint32)(s & 0x1ffffff) << 4); r[i+7] = (sp_digit)(n & 0x1ffffff);
         n >>= 25; s = a[i+29] + (s >> 25);
     }
-    n += (sp_digit)((s & 0x1ffffff) << 4); r[16] = (sp_digit)(n & 0x1ffffff);
+    n += (sp_digit)((sp_uint32)(s & 0x1ffffff) << 4); r[16] = (sp_digit)(n & 0x1ffffff);
     n >>= 25; s = a[38] + (s >> 25);
-    n += (sp_digit)((s & 0x1ffffff) << 4); r[17] = (sp_digit)(n & 0x1ffffff);
+    n += (sp_digit)((sp_uint32)(s & 0x1ffffff) << 4); r[17] = (sp_digit)(n & 0x1ffffff);
     n >>= 25; s = a[39] + (s >> 25);
-    n += (sp_digit)((s & 0x1ffffff) << 4); r[18] = (sp_digit)(n & 0x1ffffff);
+    n += (sp_digit)((sp_uint32)(s & 0x1ffffff) << 4); r[18] = (sp_digit)(n & 0x1ffffff);
     n >>= 25; s = a[40] + (s >> 25);
-    n += (sp_digit)((s & 0x1ffffff) << 4); r[19] = (sp_digit)(n & 0x1ffffff);
+    n += (sp_digit)((sp_uint32)(s & 0x1ffffff) << 4); r[19] = (sp_digit)(n & 0x1ffffff);
     n >>= 25; s = a[41] + (s >> 25);
-    n += s << 4;              r[20] = n;
+    n += (sp_uint32)s << 4;              r[20] = n;
 #endif /* WOLFSSL_SP_SMALL */
     XMEMSET(&r[21], 0, sizeof(*r) * 21U);
 }
@@ -35515,7 +35526,7 @@ static void sp_521_mont_inv_21(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 21);
     for (i=519; i>=0; i--) {
         sp_521_mont_sqr_21(t, t, p521_mod, p521_mp_mod);
-        if (p521_mod_minus_2[i / 32] & ((sp_digit)1 << (i % 32)))
+        if (p521_mod_minus_2[i / 32] & ((sp_uint32)1 << (i % 32)))
             sp_521_mont_mul_21(t, t, a, p521_mod, p521_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 21);
@@ -35755,29 +35766,29 @@ SP_NOINLINE static void sp_521_rshift1_21(sp_digit* r, const sp_digit* a)
     int i;
 
     for (i=0; i<20; i++) {
-        r[i] = (a[i] >> 1) + (sp_digit)((a[i + 1] << 24) & 0x1ffffff);
+        r[i] = (a[i] >> 1) + (sp_digit)(((sp_uint32)a[i + 1] << 24) & 0x1ffffff);
     }
 #else
-    r[0] = (a[0] >> 1) + (sp_digit)((a[1] << 24) & 0x1ffffff);
-    r[1] = (a[1] >> 1) + (sp_digit)((a[2] << 24) & 0x1ffffff);
-    r[2] = (a[2] >> 1) + (sp_digit)((a[3] << 24) & 0x1ffffff);
-    r[3] = (a[3] >> 1) + (sp_digit)((a[4] << 24) & 0x1ffffff);
-    r[4] = (a[4] >> 1) + (sp_digit)((a[5] << 24) & 0x1ffffff);
-    r[5] = (a[5] >> 1) + (sp_digit)((a[6] << 24) & 0x1ffffff);
-    r[6] = (a[6] >> 1) + (sp_digit)((a[7] << 24) & 0x1ffffff);
-    r[7] = (a[7] >> 1) + (sp_digit)((a[8] << 24) & 0x1ffffff);
-    r[8] = (a[8] >> 1) + (sp_digit)((a[9] << 24) & 0x1ffffff);
-    r[9] = (a[9] >> 1) + (sp_digit)((a[10] << 24) & 0x1ffffff);
-    r[10] = (a[10] >> 1) + (sp_digit)((a[11] << 24) & 0x1ffffff);
-    r[11] = (a[11] >> 1) + (sp_digit)((a[12] << 24) & 0x1ffffff);
-    r[12] = (a[12] >> 1) + (sp_digit)((a[13] << 24) & 0x1ffffff);
-    r[13] = (a[13] >> 1) + (sp_digit)((a[14] << 24) & 0x1ffffff);
-    r[14] = (a[14] >> 1) + (sp_digit)((a[15] << 24) & 0x1ffffff);
-    r[15] = (a[15] >> 1) + (sp_digit)((a[16] << 24) & 0x1ffffff);
-    r[16] = (a[16] >> 1) + (sp_digit)((a[17] << 24) & 0x1ffffff);
-    r[17] = (a[17] >> 1) + (sp_digit)((a[18] << 24) & 0x1ffffff);
-    r[18] = (a[18] >> 1) + (sp_digit)((a[19] << 24) & 0x1ffffff);
-    r[19] = (a[19] >> 1) + (sp_digit)((a[20] << 24) & 0x1ffffff);
+    r[0] = (a[0] >> 1) + (sp_digit)(((sp_uint32)a[1] << 24) & 0x1ffffff);
+    r[1] = (a[1] >> 1) + (sp_digit)(((sp_uint32)a[2] << 24) & 0x1ffffff);
+    r[2] = (a[2] >> 1) + (sp_digit)(((sp_uint32)a[3] << 24) & 0x1ffffff);
+    r[3] = (a[3] >> 1) + (sp_digit)(((sp_uint32)a[4] << 24) & 0x1ffffff);
+    r[4] = (a[4] >> 1) + (sp_digit)(((sp_uint32)a[5] << 24) & 0x1ffffff);
+    r[5] = (a[5] >> 1) + (sp_digit)(((sp_uint32)a[6] << 24) & 0x1ffffff);
+    r[6] = (a[6] >> 1) + (sp_digit)(((sp_uint32)a[7] << 24) & 0x1ffffff);
+    r[7] = (a[7] >> 1) + (sp_digit)(((sp_uint32)a[8] << 24) & 0x1ffffff);
+    r[8] = (a[8] >> 1) + (sp_digit)(((sp_uint32)a[9] << 24) & 0x1ffffff);
+    r[9] = (a[9] >> 1) + (sp_digit)(((sp_uint32)a[10] << 24) & 0x1ffffff);
+    r[10] = (a[10] >> 1) + (sp_digit)(((sp_uint32)a[11] << 24) & 0x1ffffff);
+    r[11] = (a[11] >> 1) + (sp_digit)(((sp_uint32)a[12] << 24) & 0x1ffffff);
+    r[12] = (a[12] >> 1) + (sp_digit)(((sp_uint32)a[13] << 24) & 0x1ffffff);
+    r[13] = (a[13] >> 1) + (sp_digit)(((sp_uint32)a[14] << 24) & 0x1ffffff);
+    r[14] = (a[14] >> 1) + (sp_digit)(((sp_uint32)a[15] << 24) & 0x1ffffff);
+    r[15] = (a[15] >> 1) + (sp_digit)(((sp_uint32)a[16] << 24) & 0x1ffffff);
+    r[16] = (a[16] >> 1) + (sp_digit)(((sp_uint32)a[17] << 24) & 0x1ffffff);
+    r[17] = (a[17] >> 1) + (sp_digit)(((sp_uint32)a[18] << 24) & 0x1ffffff);
+    r[18] = (a[18] >> 1) + (sp_digit)(((sp_uint32)a[19] << 24) & 0x1ffffff);
+    r[19] = (a[19] >> 1) + (sp_digit)(((sp_uint32)a[20] << 24) & 0x1ffffff);
 #endif
     r[20] = a[20] >> 1;
 }
@@ -36388,7 +36399,7 @@ static int sp_521_ecc_mulmod_21(sp_point_521* r, const sp_point_521* g,
     if (err == MP_OKAY) {
         i = 20;
         c = 21;
-        n = k[i--] << (25 - c);
+        n = (sp_uint32)k[i--] << (25 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1)
@@ -36399,7 +36410,7 @@ static int sp_521_ecc_mulmod_21(sp_point_521* r, const sp_point_521* g,
             }
 
             y = (n >> 24) & 1;
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_521_proj_point_add_21(&t[y^1], &t[0], &t[1], tmp);
 
@@ -36458,7 +36469,7 @@ static int sp_521_ecc_mulmod_21_nb(sp_ecc_ctx_t* sp_ctx, sp_point_521* r,
         XMEMSET(ctx->t, 0, sizeof(sp_point_521) * 3);
         ctx->i = 20;
         ctx->c = 21;
-        ctx->n = k[ctx->i--] << (25 - ctx->c);
+        ctx->n = (sp_uint32)k[ctx->i--] << (25 - ctx->c);
 
         /* t[0] = {0, 0, 1} * norm */
         ctx->t[0].infinity = 1;
@@ -36488,7 +36499,7 @@ static int sp_521_ecc_mulmod_21_nb(sp_ecc_ctx_t* sp_ctx, sp_point_521* r,
             ctx->c = 25;
         }
         ctx->y = (ctx->n >> 24) & 1;
-        ctx->n <<= 1;
+        ctx->n = (sp_uint32)ctx->n << 1;
         XMEMSET(&ctx->add_ctx, 0, sizeof(ctx->add_ctx));
         ctx->state = 5;
         break;
@@ -36915,7 +36926,7 @@ static void sp_521_ecc_recode_6_21(const sp_digit* k, ecc_recode_521* v)
         }
         else if (++j < 21) {
             n = k[j];
-            y |= (word8)((n << (25 - o)) & 0x3f);
+            y |= (word8)(((sp_uint32)n << (25 - o)) & 0x3f);
             o -= 19;
             n >>= o;
         }
@@ -40307,7 +40318,7 @@ static void sp_521_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     r[0] = 0;
     for (i = n-1; i >= 0; i--) {
-        r[j] |= (((sp_digit)a[i]) << s);
+        r[j] |= (((sp_uint32)a[i]) << s);
         if (s >= 17U) {
             r[j] &= 0x1ffffff;
             s = 25U - s;
@@ -40529,7 +40540,7 @@ static void sp_521_to_bin_21(sp_digit* r, byte* a)
     for (i=0; i<21 && j>=0; i++) {
         b = 0;
         /* lint allow cast of mismatch sp_digit and int */
-        a[j--] |= (byte)(r[i] << s); /*lint !e9033*/
+        a[j--] |= (byte)((sp_uint32)r[i] << s); /*lint !e9033*/
         b += 8 - s;
         if (j < 0) {
             break;
@@ -40651,23 +40662,23 @@ SP_NOINLINE static void sp_521_rshift_21(sp_digit* r, const sp_digit* a,
 
 #ifdef WOLFSSL_SP_SMALL
     for (i=0; i<20; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (25 - n))) & 0x1ffffff);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint32)a[i + 1] << (25 - n))) & 0x1ffffff);
     }
 #else
     for (i=0; i<16; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (25 - n)) & 0x1ffffff);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (25 - n)) & 0x1ffffff);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (25 - n)) & 0x1ffffff);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (25 - n)) & 0x1ffffff);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (25 - n)) & 0x1ffffff);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (25 - n)) & 0x1ffffff);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (25 - n)) & 0x1ffffff);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (25 - n)) & 0x1ffffff);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint32)a[i+1] << (25 - n)) & 0x1ffffff);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint32)a[i+2] << (25 - n)) & 0x1ffffff);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint32)a[i+3] << (25 - n)) & 0x1ffffff);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint32)a[i+4] << (25 - n)) & 0x1ffffff);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint32)a[i+5] << (25 - n)) & 0x1ffffff);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint32)a[i+6] << (25 - n)) & 0x1ffffff);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint32)a[i+7] << (25 - n)) & 0x1ffffff);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint32)a[i+8] << (25 - n)) & 0x1ffffff);
     }
-    r[16] = (a[16] >> n) | (sp_digit)((a[17] << (25 - n)) & 0x1ffffff);
-    r[17] = (a[17] >> n) | (sp_digit)((a[18] << (25 - n)) & 0x1ffffff);
-    r[18] = (a[18] >> n) | (sp_digit)((a[19] << (25 - n)) & 0x1ffffff);
-    r[19] = (a[19] >> n) | (sp_digit)((a[20] << (25 - n)) & 0x1ffffff);
+    r[16] = (a[16] >> n) | (sp_digit)(((sp_uint32)a[17] << (25 - n)) & 0x1ffffff);
+    r[17] = (a[17] >> n) | (sp_digit)(((sp_uint32)a[18] << (25 - n)) & 0x1ffffff);
+    r[18] = (a[18] >> n) | (sp_digit)(((sp_uint32)a[19] << (25 - n)) & 0x1ffffff);
+    r[19] = (a[19] >> n) | (sp_digit)(((sp_uint32)a[20] << (25 - n)) & 0x1ffffff);
 #endif /* WOLFSSL_SP_SMALL */
     r[20] = a[20] >> n;
 }
@@ -40738,7 +40749,7 @@ SP_NOINLINE static void sp_521_lshift_42(sp_digit* r, const sp_digit* a,
 
     r[42] = a[41] >> (25 - n);
     for (i=41; i>0; i--) {
-        r[i] = (sp_digit)(((a[i] << n) | (a[i-1] >> (25 - n))) & 0x1ffffff);
+        r[i] = (sp_digit)((((sp_uint32)a[i] << n) | (a[i-1] >> (25 - n))) & 0x1ffffff);
     }
 #else
     sp_int_digit s;
@@ -40829,7 +40840,7 @@ SP_NOINLINE static void sp_521_lshift_42(sp_digit* r, const sp_digit* a,
     s = (sp_int_digit)(a[1]); t = (sp_int_digit)(a[0]);
     r[1] = (sp_digit)(((s << n) | (t >> (25U - n))) & 0x1ffffff);
 #endif /* WOLFSSL_SP_SMALL */
-    r[0] = (sp_digit)((a[0] << n) & 0x1ffffff);
+    r[0] = (sp_digit)(((sp_uint32)a[0] << n) & 0x1ffffff);
 }
 
 /* Divide d in a and put remainder into r (m*d + r = a)
@@ -42253,7 +42264,7 @@ static int sp_521_mont_sqrt_21(sp_digit* y)
             XMEMCPY(t, y, sizeof(sp_digit) * 21);
             for (i=518; i>=0; i--) {
                 sp_521_mont_sqr_21(t, t, p521_mod, p521_mp_mod);
-                if (p521_sqrt_power[i / 32] & ((sp_digit)1 << (i % 32)))
+                if (p521_sqrt_power[i / 32] & ((sp_uint32)1 << (i % 32)))
                     sp_521_mont_mul_21(t, t, y, p521_mod, p521_mp_mod);
             }
             XMEMCPY(y, t, sizeof(sp_digit) * 21);
@@ -43165,20 +43176,20 @@ SP_NOINLINE static void sp_1024_rshift_42(sp_digit* r, const sp_digit* a,
 
 #ifdef WOLFSSL_SP_SMALL
     for (i=0; i<41; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (25 - n))) & 0x1ffffff);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint32)a[i + 1] << (25 - n))) & 0x1ffffff);
     }
 #else
     for (i=0; i<40; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (25 - n)) & 0x1ffffff);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (25 - n)) & 0x1ffffff);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (25 - n)) & 0x1ffffff);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (25 - n)) & 0x1ffffff);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (25 - n)) & 0x1ffffff);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (25 - n)) & 0x1ffffff);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (25 - n)) & 0x1ffffff);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (25 - n)) & 0x1ffffff);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint32)a[i+1] << (25 - n)) & 0x1ffffff);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint32)a[i+2] << (25 - n)) & 0x1ffffff);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint32)a[i+3] << (25 - n)) & 0x1ffffff);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint32)a[i+4] << (25 - n)) & 0x1ffffff);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint32)a[i+5] << (25 - n)) & 0x1ffffff);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint32)a[i+6] << (25 - n)) & 0x1ffffff);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint32)a[i+7] << (25 - n)) & 0x1ffffff);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint32)a[i+8] << (25 - n)) & 0x1ffffff);
     }
-    r[40] = (a[40] >> n) | (sp_digit)((a[41] << (25 - n)) & 0x1ffffff);
+    r[40] = (a[40] >> n) | (sp_digit)(((sp_uint32)a[41] << (25 - n)) & 0x1ffffff);
 #endif /* WOLFSSL_SP_SMALL */
     r[41] = a[41] >> n;
 }
@@ -43218,7 +43229,7 @@ static WC_INLINE sp_digit sp_1024_div_word_42(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 23; i >= 1; i--) {
         t1 += t1 + (((sp_uint32)t0 >> 24) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint32)t0 << 1;
         t2 = (sp_digit)(((sp_uint32)(dv - t1)) >> 31);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -43461,7 +43472,7 @@ static void sp_1024_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0x1ffffff;
         s = 25U - s;
         if (j + 1 >= size) {
@@ -43496,7 +43507,7 @@ static void sp_1024_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 25) {
             r[j] &= 0x1ffffff;
             if (j + 1 >= size) {
@@ -43562,7 +43573,7 @@ static int sp_1024_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 41; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -43587,7 +43598,7 @@ static int sp_1024_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 41; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 25 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -43850,11 +43861,11 @@ static void sp_1024_mont_reduce_42(sp_digit* a, const sp_digit* m, sp_digit mp)
 
     if (mp != 1) {
         for (i=0; i<40; i++) {
-            mu = (sp_digit)((a[i] * mp) & 0x1ffffff);
+            mu = (sp_digit)(((sp_uint32)a[i] * (sp_uint32)mp) & 0x1ffffff);
             sp_1024_mul_add_42(a+i, m, mu);
             a[i+1] += a[i] >> 25;
         }
-        mu = (sp_digit)((a[i] * mp) & 0xffffffL);
+        mu = (sp_digit)(((sp_uint32)a[i] * (sp_uint32)mp) & 0xffffffL);
         sp_1024_mul_add_42(a+i, m, mu);
         a[i+1] += a[i] >> 25;
         a[i] &= 0x1ffffff;
@@ -44091,50 +44102,50 @@ SP_NOINLINE static void sp_1024_rshift1_42(sp_digit* r, const sp_digit* a)
     int i;
 
     for (i=0; i<41; i++) {
-        r[i] = (a[i] >> 1) + (sp_digit)((a[i + 1] << 24) & 0x1ffffff);
+        r[i] = (a[i] >> 1) + (sp_digit)(((sp_uint32)a[i + 1] << 24) & 0x1ffffff);
     }
 #else
-    r[0] = (a[0] >> 1) + (sp_digit)((a[1] << 24) & 0x1ffffff);
-    r[1] = (a[1] >> 1) + (sp_digit)((a[2] << 24) & 0x1ffffff);
-    r[2] = (a[2] >> 1) + (sp_digit)((a[3] << 24) & 0x1ffffff);
-    r[3] = (a[3] >> 1) + (sp_digit)((a[4] << 24) & 0x1ffffff);
-    r[4] = (a[4] >> 1) + (sp_digit)((a[5] << 24) & 0x1ffffff);
-    r[5] = (a[5] >> 1) + (sp_digit)((a[6] << 24) & 0x1ffffff);
-    r[6] = (a[6] >> 1) + (sp_digit)((a[7] << 24) & 0x1ffffff);
-    r[7] = (a[7] >> 1) + (sp_digit)((a[8] << 24) & 0x1ffffff);
-    r[8] = (a[8] >> 1) + (sp_digit)((a[9] << 24) & 0x1ffffff);
-    r[9] = (a[9] >> 1) + (sp_digit)((a[10] << 24) & 0x1ffffff);
-    r[10] = (a[10] >> 1) + (sp_digit)((a[11] << 24) & 0x1ffffff);
-    r[11] = (a[11] >> 1) + (sp_digit)((a[12] << 24) & 0x1ffffff);
-    r[12] = (a[12] >> 1) + (sp_digit)((a[13] << 24) & 0x1ffffff);
-    r[13] = (a[13] >> 1) + (sp_digit)((a[14] << 24) & 0x1ffffff);
-    r[14] = (a[14] >> 1) + (sp_digit)((a[15] << 24) & 0x1ffffff);
-    r[15] = (a[15] >> 1) + (sp_digit)((a[16] << 24) & 0x1ffffff);
-    r[16] = (a[16] >> 1) + (sp_digit)((a[17] << 24) & 0x1ffffff);
-    r[17] = (a[17] >> 1) + (sp_digit)((a[18] << 24) & 0x1ffffff);
-    r[18] = (a[18] >> 1) + (sp_digit)((a[19] << 24) & 0x1ffffff);
-    r[19] = (a[19] >> 1) + (sp_digit)((a[20] << 24) & 0x1ffffff);
-    r[20] = (a[20] >> 1) + (sp_digit)((a[21] << 24) & 0x1ffffff);
-    r[21] = (a[21] >> 1) + (sp_digit)((a[22] << 24) & 0x1ffffff);
-    r[22] = (a[22] >> 1) + (sp_digit)((a[23] << 24) & 0x1ffffff);
-    r[23] = (a[23] >> 1) + (sp_digit)((a[24] << 24) & 0x1ffffff);
-    r[24] = (a[24] >> 1) + (sp_digit)((a[25] << 24) & 0x1ffffff);
-    r[25] = (a[25] >> 1) + (sp_digit)((a[26] << 24) & 0x1ffffff);
-    r[26] = (a[26] >> 1) + (sp_digit)((a[27] << 24) & 0x1ffffff);
-    r[27] = (a[27] >> 1) + (sp_digit)((a[28] << 24) & 0x1ffffff);
-    r[28] = (a[28] >> 1) + (sp_digit)((a[29] << 24) & 0x1ffffff);
-    r[29] = (a[29] >> 1) + (sp_digit)((a[30] << 24) & 0x1ffffff);
-    r[30] = (a[30] >> 1) + (sp_digit)((a[31] << 24) & 0x1ffffff);
-    r[31] = (a[31] >> 1) + (sp_digit)((a[32] << 24) & 0x1ffffff);
-    r[32] = (a[32] >> 1) + (sp_digit)((a[33] << 24) & 0x1ffffff);
-    r[33] = (a[33] >> 1) + (sp_digit)((a[34] << 24) & 0x1ffffff);
-    r[34] = (a[34] >> 1) + (sp_digit)((a[35] << 24) & 0x1ffffff);
-    r[35] = (a[35] >> 1) + (sp_digit)((a[36] << 24) & 0x1ffffff);
-    r[36] = (a[36] >> 1) + (sp_digit)((a[37] << 24) & 0x1ffffff);
-    r[37] = (a[37] >> 1) + (sp_digit)((a[38] << 24) & 0x1ffffff);
-    r[38] = (a[38] >> 1) + (sp_digit)((a[39] << 24) & 0x1ffffff);
-    r[39] = (a[39] >> 1) + (sp_digit)((a[40] << 24) & 0x1ffffff);
-    r[40] = (a[40] >> 1) + (sp_digit)((a[41] << 24) & 0x1ffffff);
+    r[0] = (a[0] >> 1) + (sp_digit)(((sp_uint32)a[1] << 24) & 0x1ffffff);
+    r[1] = (a[1] >> 1) + (sp_digit)(((sp_uint32)a[2] << 24) & 0x1ffffff);
+    r[2] = (a[2] >> 1) + (sp_digit)(((sp_uint32)a[3] << 24) & 0x1ffffff);
+    r[3] = (a[3] >> 1) + (sp_digit)(((sp_uint32)a[4] << 24) & 0x1ffffff);
+    r[4] = (a[4] >> 1) + (sp_digit)(((sp_uint32)a[5] << 24) & 0x1ffffff);
+    r[5] = (a[5] >> 1) + (sp_digit)(((sp_uint32)a[6] << 24) & 0x1ffffff);
+    r[6] = (a[6] >> 1) + (sp_digit)(((sp_uint32)a[7] << 24) & 0x1ffffff);
+    r[7] = (a[7] >> 1) + (sp_digit)(((sp_uint32)a[8] << 24) & 0x1ffffff);
+    r[8] = (a[8] >> 1) + (sp_digit)(((sp_uint32)a[9] << 24) & 0x1ffffff);
+    r[9] = (a[9] >> 1) + (sp_digit)(((sp_uint32)a[10] << 24) & 0x1ffffff);
+    r[10] = (a[10] >> 1) + (sp_digit)(((sp_uint32)a[11] << 24) & 0x1ffffff);
+    r[11] = (a[11] >> 1) + (sp_digit)(((sp_uint32)a[12] << 24) & 0x1ffffff);
+    r[12] = (a[12] >> 1) + (sp_digit)(((sp_uint32)a[13] << 24) & 0x1ffffff);
+    r[13] = (a[13] >> 1) + (sp_digit)(((sp_uint32)a[14] << 24) & 0x1ffffff);
+    r[14] = (a[14] >> 1) + (sp_digit)(((sp_uint32)a[15] << 24) & 0x1ffffff);
+    r[15] = (a[15] >> 1) + (sp_digit)(((sp_uint32)a[16] << 24) & 0x1ffffff);
+    r[16] = (a[16] >> 1) + (sp_digit)(((sp_uint32)a[17] << 24) & 0x1ffffff);
+    r[17] = (a[17] >> 1) + (sp_digit)(((sp_uint32)a[18] << 24) & 0x1ffffff);
+    r[18] = (a[18] >> 1) + (sp_digit)(((sp_uint32)a[19] << 24) & 0x1ffffff);
+    r[19] = (a[19] >> 1) + (sp_digit)(((sp_uint32)a[20] << 24) & 0x1ffffff);
+    r[20] = (a[20] >> 1) + (sp_digit)(((sp_uint32)a[21] << 24) & 0x1ffffff);
+    r[21] = (a[21] >> 1) + (sp_digit)(((sp_uint32)a[22] << 24) & 0x1ffffff);
+    r[22] = (a[22] >> 1) + (sp_digit)(((sp_uint32)a[23] << 24) & 0x1ffffff);
+    r[23] = (a[23] >> 1) + (sp_digit)(((sp_uint32)a[24] << 24) & 0x1ffffff);
+    r[24] = (a[24] >> 1) + (sp_digit)(((sp_uint32)a[25] << 24) & 0x1ffffff);
+    r[25] = (a[25] >> 1) + (sp_digit)(((sp_uint32)a[26] << 24) & 0x1ffffff);
+    r[26] = (a[26] >> 1) + (sp_digit)(((sp_uint32)a[27] << 24) & 0x1ffffff);
+    r[27] = (a[27] >> 1) + (sp_digit)(((sp_uint32)a[28] << 24) & 0x1ffffff);
+    r[28] = (a[28] >> 1) + (sp_digit)(((sp_uint32)a[29] << 24) & 0x1ffffff);
+    r[29] = (a[29] >> 1) + (sp_digit)(((sp_uint32)a[30] << 24) & 0x1ffffff);
+    r[30] = (a[30] >> 1) + (sp_digit)(((sp_uint32)a[31] << 24) & 0x1ffffff);
+    r[31] = (a[31] >> 1) + (sp_digit)(((sp_uint32)a[32] << 24) & 0x1ffffff);
+    r[32] = (a[32] >> 1) + (sp_digit)(((sp_uint32)a[33] << 24) & 0x1ffffff);
+    r[33] = (a[33] >> 1) + (sp_digit)(((sp_uint32)a[34] << 24) & 0x1ffffff);
+    r[34] = (a[34] >> 1) + (sp_digit)(((sp_uint32)a[35] << 24) & 0x1ffffff);
+    r[35] = (a[35] >> 1) + (sp_digit)(((sp_uint32)a[36] << 24) & 0x1ffffff);
+    r[36] = (a[36] >> 1) + (sp_digit)(((sp_uint32)a[37] << 24) & 0x1ffffff);
+    r[37] = (a[37] >> 1) + (sp_digit)(((sp_uint32)a[38] << 24) & 0x1ffffff);
+    r[38] = (a[38] >> 1) + (sp_digit)(((sp_uint32)a[39] << 24) & 0x1ffffff);
+    r[39] = (a[39] >> 1) + (sp_digit)(((sp_uint32)a[40] << 24) & 0x1ffffff);
+    r[40] = (a[40] >> 1) + (sp_digit)(((sp_uint32)a[41] << 24) & 0x1ffffff);
 #endif
     r[41] = a[41] >> 1;
 }
@@ -44737,7 +44748,7 @@ static int sp_1024_ecc_mulmod_42(sp_point_1024* r, const sp_point_1024* g,
     if (err == MP_OKAY) {
         i = 40;
         c = 24;
-        n = k[i--] << (25 - c);
+        n = (sp_uint32)k[i--] << (25 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1)
@@ -44748,7 +44759,7 @@ static int sp_1024_ecc_mulmod_42(sp_point_1024* r, const sp_point_1024* g,
             }
 
             y = (n >> 24) & 1;
-            n <<= 1;
+            n = (sp_uint32)n << 1;
 
             sp_1024_proj_point_add_42(&t[y^1], &t[0], &t[1], tmp);
 
@@ -44807,7 +44818,7 @@ static int sp_1024_ecc_mulmod_42_nb(sp_ecc_ctx_t* sp_ctx, sp_point_1024* r,
         XMEMSET(ctx->t, 0, sizeof(sp_point_1024) * 3);
         ctx->i = 40;
         ctx->c = 24;
-        ctx->n = k[ctx->i--] << (25 - ctx->c);
+        ctx->n = (sp_uint32)k[ctx->i--] << (25 - ctx->c);
 
         /* t[0] = {0, 0, 1} * norm */
         ctx->t[0].infinity = 1;
@@ -44837,7 +44848,7 @@ static int sp_1024_ecc_mulmod_42_nb(sp_ecc_ctx_t* sp_ctx, sp_point_1024* r,
             ctx->c = 25;
         }
         ctx->y = (ctx->n >> 24) & 1;
-        ctx->n <<= 1;
+        ctx->n = (sp_uint32)ctx->n << 1;
         XMEMSET(&ctx->add_ctx, 0, sizeof(ctx->add_ctx));
         ctx->state = 5;
         break;
@@ -45314,7 +45325,7 @@ static void sp_1024_ecc_recode_7_42(const sp_digit* k, ecc_recode_1024* v)
         }
         else if (++j < 42) {
             n = k[j];
-            y |= (word8)((n << (25 - o)) & 0x7f);
+            y |= (word8)(((sp_uint32)n << (25 - o)) & 0x7f);
             o -= 18;
             n >>= o;
         }
@@ -53427,7 +53438,7 @@ static void sp_1024_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     r[0] = 0;
     for (i = n-1; i >= 0; i--) {
-        r[j] |= (((sp_digit)a[i]) << s);
+        r[j] |= (((sp_uint32)a[i]) << s);
         if (s >= 17U) {
             r[j] &= 0x1ffffff;
             s = 25U - s;
diff --git a/wolfcrypt/src/sp_c64.c b/wolfcrypt/src/sp_c64.c
index ae32ae2ce7..b0dd93cbd9 100644
--- a/wolfcrypt/src/sp_c64.c
+++ b/wolfcrypt/src/sp_c64.c
@@ -51,6 +51,17 @@
 
 #include <wolfssl/wolfcrypt/sp.h>
 
+#if defined(WOLFSSL_USE_SAVE_VECTOR_REGISTERS) && !defined(WOLFSSL_SP_ASM) && \
+        !defined(DEBUG_VECTOR_REGISTER_ACCESS)
+    /* force off unneeded vector register save/restore. */
+    #undef SAVE_VECTOR_REGISTERS
+    #define SAVE_VECTOR_REGISTERS(fail_clause) SAVE_NO_VECTOR_REGISTERS(fail_clause)
+    #undef SAVE_VECTOR_REGISTERS2
+    #define SAVE_VECTOR_REGISTERS2() SAVE_NO_VECTOR_REGISTERS2()
+    #undef RESTORE_VECTOR_REGISTERS
+    #define RESTORE_VECTOR_REGISTERS() RESTORE_NO_VECTOR_REGISTERS()
+#endif
+
 #ifdef __IAR_SYSTEMS_ICC__
 #define __asm__        asm
 #define __volatile__   volatile
@@ -175,7 +186,7 @@ static void sp_2048_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     r[0] = 0;
     for (i = n-1; i >= 0; i--) {
-        r[j] |= (((sp_digit)a[i]) << s);
+        r[j] |= (((sp_uint64)a[i]) << s);
         if (s >= 53U) {
             r[j] &= 0x1fffffffffffffffL;
             s = 61U - s;
@@ -221,7 +232,7 @@ static void sp_2048_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0x1fffffffffffffffL;
         s = 61U - s;
         if (j + 1 >= size) {
@@ -256,7 +267,7 @@ static void sp_2048_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 61) {
             r[j] &= 0x1fffffffffffffffL;
             if (j + 1 >= size) {
@@ -305,7 +316,7 @@ static void sp_2048_to_bin_34(sp_digit* r, byte* a)
     for (i=0; i<34 && j>=0; i++) {
         b = 0;
         /* lint allow cast of mismatch sp_digit and int */
-        a[j--] |= (byte)(r[i] << s); /*lint !e9033*/
+        a[j--] |= (byte)((sp_uint64)r[i] << s); /*lint !e9033*/
         b += 8 - s;
         if (j < 0) {
             break;
@@ -481,10 +492,10 @@ SP_NOINLINE static void sp_2048_sqr_34(sp_digit* r, const sp_digit* a)
  */
 static void sp_2048_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint64 x;
+    sp_uint64 b;
 
-    b = a[0];
+    b = (sp_uint64)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
@@ -493,7 +504,7 @@ static void sp_2048_mont_setup(const sp_digit* a, sp_digit* rho)
     x &= 0x1fffffffffffffffL;
 
     /* rho = -1/m mod b */
-    *rho = ((sp_digit)1 << 61) - x;
+    *rho = (sp_digit)(((sp_uint64)1 << 61) - (sp_digit)x);
 }
 
 /* Multiply a by scalar b into r. (r = a * b)
@@ -892,7 +903,7 @@ SP_NOINLINE static void sp_2048_rshift_17(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<16; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (61 - n))) & 0x1fffffffffffffffL);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint64)a[i + 1] << (61 - n))) & 0x1fffffffffffffffL);
     }
     r[16] = a[16] >> n;
 }
@@ -932,7 +943,7 @@ static WC_INLINE sp_digit sp_2048_div_word_17(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 59; i >= 1; i--) {
         t1 += t1 + (((sp_uint64)t0 >> 60) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint64)t0 << 1;
         t2 = (sp_digit)(((sp_uint64)(dv - t1)) >> 63);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -1124,9 +1135,9 @@ static int sp_2048_mod_exp_17(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 61;
-        c = bits % 61;
-        n = e[i--] << (61 - c);
+        i = (bits - 1) / 61;
+        c = ((bits - 1) % 61) + 1;
+        n = (sp_uint64)e[i--] << (61 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -1138,7 +1149,7 @@ static int sp_2048_mod_exp_17(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 60) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_2048_mont_mul_17(t[y^1], t[0], t[1], m, mp);
 
@@ -1200,9 +1211,9 @@ static int sp_2048_mod_exp_17(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 61;
-        c = bits % 61;
-        n = e[i--] << (61 - c);
+        i = (bits - 1) / 61;
+        c = ((bits - 1) % 61) + 1;
+        n = (sp_uint64)e[i--] << (61 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -1214,7 +1225,7 @@ static int sp_2048_mod_exp_17(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 60) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_2048_mont_mul_17(t[y^1], t[0], t[1], m, mp);
 
@@ -1315,38 +1326,38 @@ static int sp_2048_mod_exp_17(sp_digit* r, const sp_digit* a, const sp_digit* e,
             c = 61;
         }
         if (i < 17) {
-            n = e[i--] << (64 - c);
+            n = (sp_uint64)e[i--] << (64 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 5) {
-            n |= e[i--] << (3 - c);
+            n |= (sp_uint64)e[i--] << (3 - c);
             c += 61;
         }
         y = (int)((n >> 59) & 0x1f);
-        n <<= 5;
+        n = (sp_uint64)n << 5;
         c -= 5;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 34);
         while ((i >= 0) || (c >= 5)) {
             if (c >= 5) {
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c -= 5;
             }
             else if (c == 0) {
-                n = e[i--] << 3;
+                n = (sp_uint64)e[i--] << 3;
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c = 56;
             }
             else {
                 y = (byte)((n >> 59) & 0x1f);
-                n = e[i--] << 3;
+                n = (sp_uint64)e[i--] << 3;
                 c = 5 - c;
                 y |= (byte)((n >> (64 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 61 - c;
             }
 
@@ -1656,7 +1667,7 @@ SP_NOINLINE static void sp_2048_rshift_34(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<33; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (61 - n))) & 0x1fffffffffffffffL);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint64)a[i + 1] << (61 - n))) & 0x1fffffffffffffffL);
     }
     r[33] = a[33] >> n;
 }
@@ -1696,7 +1707,7 @@ static WC_INLINE sp_digit sp_2048_div_word_34(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 59; i >= 1; i--) {
         t1 += t1 + (((sp_uint64)t0 >> 60) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint64)t0 << 1;
         t2 = (sp_digit)(((sp_uint64)(dv - t1)) >> 63);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -1889,9 +1900,9 @@ static int sp_2048_mod_exp_34(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 61;
-        c = bits % 61;
-        n = e[i--] << (61 - c);
+        i = (bits - 1) / 61;
+        c = ((bits - 1) % 61) + 1;
+        n = (sp_uint64)e[i--] << (61 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -1903,7 +1914,7 @@ static int sp_2048_mod_exp_34(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 60) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_2048_mont_mul_34(t[y^1], t[0], t[1], m, mp);
 
@@ -1965,9 +1976,9 @@ static int sp_2048_mod_exp_34(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 61;
-        c = bits % 61;
-        n = e[i--] << (61 - c);
+        i = (bits - 1) / 61;
+        c = ((bits - 1) % 61) + 1;
+        n = (sp_uint64)e[i--] << (61 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -1979,7 +1990,7 @@ static int sp_2048_mod_exp_34(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 60) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_2048_mont_mul_34(t[y^1], t[0], t[1], m, mp);
 
@@ -2064,38 +2075,38 @@ static int sp_2048_mod_exp_34(sp_digit* r, const sp_digit* a, const sp_digit* e,
             c = 61;
         }
         if (i < 34) {
-            n = e[i--] << (64 - c);
+            n = (sp_uint64)e[i--] << (64 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 4) {
-            n |= e[i--] << (3 - c);
+            n |= (sp_uint64)e[i--] << (3 - c);
             c += 61;
         }
         y = (int)((n >> 60) & 0xf);
-        n <<= 4;
+        n = (sp_uint64)n << 4;
         c -= 4;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 68);
         while ((i >= 0) || (c >= 4)) {
             if (c >= 4) {
                 y = (byte)((n >> 60) & 0xf);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c -= 4;
             }
             else if (c == 0) {
-                n = e[i--] << 3;
+                n = (sp_uint64)e[i--] << 3;
                 y = (byte)((n >> 60) & 0xf);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c = 57;
             }
             else {
                 y = (byte)((n >> 60) & 0xf);
-                n = e[i--] << 3;
+                n = (sp_uint64)e[i--] << 3;
                 c = 4 - c;
                 y |= (byte)((n >> (64 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 61 - c;
             }
 
@@ -2191,10 +2202,10 @@ static int sp_2048_mod_exp_34_nb(sp_2048_mod_exp_34_ctx* ctx,
     case 4: /* BIT_INIT: index the most-significant exponent limb without
               reading off the end when bits is an exact multiple of the
               limb width. (bits-1) keeps i within the populated range and
-              c = (bits-1) % @bits + 1 keeps the shift in [1, @bits]. */
+              c = (bits-1) % 61 + 1 keeps the shift in [1, 61]. */
         ctx->i = (ctx->bits - 1) / 61;
         ctx->c = ((ctx->bits - 1) % 61) + 1;
-        ctx->n = e[ctx->i--] << (61 - ctx->c);
+        ctx->n = (sp_uint64)e[ctx->i--] << (61 - ctx->c);
         ctx->state = 5;
         break;
     case 5: /* BIT_NEXT: refill on word boundary, peel one exponent bit */
@@ -2207,7 +2218,7 @@ static int sp_2048_mod_exp_34_nb(sp_2048_mod_exp_34_ctx* ctx,
             ctx->c = 61;
         }
         ctx->y = (byte)((ctx->n >> 60) & 1);
-        ctx->n <<= 1;
+        ctx->n = (sp_uint64)ctx->n << 1;
         ctx->state = 6;
         break;
     case 6: /* MUL: t[y^1] = t[0] * t[1] in Montgomery form */
@@ -2978,7 +2989,7 @@ static int sp_2048_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 34; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -3003,7 +3014,7 @@ static int sp_2048_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 34; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 61 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -3219,9 +3230,9 @@ SP_NOINLINE static void sp_2048_lshift_34(sp_digit* r, const sp_digit* a,
 
     r[34] = a[33] >> (61 - n);
     for (i=33; i>0; i--) {
-        r[i] = (sp_digit)(((a[i] << n) | (a[i-1] >> (61 - n))) & 0x1fffffffffffffffL);
+        r[i] = (sp_digit)((((sp_uint64)a[i] << n) | (a[i-1] >> (61 - n))) & 0x1fffffffffffffffL);
     }
-    r[0] = (sp_digit)((a[0] << n) & 0x1fffffffffffffffL);
+    r[0] = (sp_digit)(((sp_uint64)a[0] << n) & 0x1fffffffffffffffL);
 }
 
 /* Modular exponentiate 2 to the e mod m. (r = 2^e mod m)
@@ -3267,38 +3278,38 @@ static int sp_2048_mod_exp_2_34(sp_digit* r, const sp_digit* e, int bits, const
             c = 61;
         }
         if (i < 34) {
-            n = e[i--] << (64 - c);
+            n = (sp_uint64)e[i--] << (64 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 5) {
-            n |= e[i--] << (3 - c);
+            n |= (sp_uint64)e[i--] << (3 - c);
             c += 61;
         }
         y = (int)((n >> 59) & 0x1f);
-        n <<= 5;
+        n = (sp_uint64)n << 5;
         c -= 5;
         sp_2048_lshift_34(r, norm, (byte)y);
         while ((i >= 0) || (c >= 5)) {
             if (c >= 5) {
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c -= 5;
             }
             else if (c == 0) {
-                n = e[i--] << 3;
+                n = (sp_uint64)e[i--] << 3;
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c = 56;
             }
             else {
                 y = (byte)((n >> 59) & 0x1f);
-                n = e[i--] << 3;
+                n = (sp_uint64)e[i--] << 3;
                 c = 5 - c;
                 y |= (byte)((n >> (64 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 61 - c;
             }
 
@@ -3626,7 +3637,7 @@ static void sp_2048_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     r[0] = 0;
     for (i = n-1; i >= 0; i--) {
-        r[j] |= (((sp_digit)a[i]) << s);
+        r[j] |= (((sp_uint64)a[i]) << s);
         if (s >= 49U) {
             r[j] &= 0x1ffffffffffffffL;
             s = 57U - s;
@@ -3672,7 +3683,7 @@ static void sp_2048_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0x1ffffffffffffffL;
         s = 57U - s;
         if (j + 1 >= size) {
@@ -3707,7 +3718,7 @@ static void sp_2048_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 57) {
             r[j] &= 0x1ffffffffffffffL;
             if (j + 1 >= size) {
@@ -3756,7 +3767,7 @@ static void sp_2048_to_bin_36(sp_digit* r, byte* a)
     for (i=0; i<36 && j>=0; i++) {
         b = 0;
         /* lint allow cast of mismatch sp_digit and int */
-        a[j--] |= (byte)(r[i] << s); /*lint !e9033*/
+        a[j--] |= (byte)((sp_uint64)r[i] << s); /*lint !e9033*/
         b += 8 - s;
         if (j < 0) {
             break;
@@ -4245,10 +4256,10 @@ SP_NOINLINE static void sp_2048_sqr_36(sp_digit* r, const sp_digit* a)
  */
 static void sp_2048_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint64 x;
+    sp_uint64 b;
 
-    b = a[0];
+    b = (sp_uint64)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
@@ -4257,7 +4268,7 @@ static void sp_2048_mont_setup(const sp_digit* a, sp_digit* rho)
     x &= 0x1ffffffffffffffL;
 
     /* rho = -1/m mod b */
-    *rho = ((sp_digit)1 << 57) - x;
+    *rho = (sp_digit)(((sp_uint64)1 << 57) - (sp_digit)x);
 }
 
 /* Multiply a by scalar b into r. (r = a * b)
@@ -4591,16 +4602,16 @@ SP_NOINLINE static void sp_2048_rshift_18(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<16; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint64)a[i+1] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint64)a[i+2] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint64)a[i+3] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint64)a[i+4] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint64)a[i+5] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint64)a[i+6] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint64)a[i+7] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint64)a[i+8] << (57 - n)) & 0x1ffffffffffffffL);
     }
-    r[16] = (a[16] >> n) | (sp_digit)((a[17] << (57 - n)) & 0x1ffffffffffffffL);
+    r[16] = (a[16] >> n) | (sp_digit)(((sp_uint64)a[17] << (57 - n)) & 0x1ffffffffffffffL);
     r[17] = a[17] >> n;
 }
 
@@ -4639,7 +4650,7 @@ static WC_INLINE sp_digit sp_2048_div_word_18(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 55; i >= 1; i--) {
         t1 += t1 + (((sp_uint64)t0 >> 56) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint64)t0 << 1;
         t2 = (sp_digit)(((sp_uint64)(dv - t1)) >> 63);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -4831,9 +4842,9 @@ static int sp_2048_mod_exp_18(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 57;
-        c = bits % 57;
-        n = e[i--] << (57 - c);
+        i = (bits - 1) / 57;
+        c = ((bits - 1) % 57) + 1;
+        n = (sp_uint64)e[i--] << (57 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -4845,7 +4856,7 @@ static int sp_2048_mod_exp_18(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 56) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_2048_mont_mul_18(t[y^1], t[0], t[1], m, mp);
 
@@ -4907,9 +4918,9 @@ static int sp_2048_mod_exp_18(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 57;
-        c = bits % 57;
-        n = e[i--] << (57 - c);
+        i = (bits - 1) / 57;
+        c = ((bits - 1) % 57) + 1;
+        n = (sp_uint64)e[i--] << (57 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -4921,7 +4932,7 @@ static int sp_2048_mod_exp_18(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 56) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_2048_mont_mul_18(t[y^1], t[0], t[1], m, mp);
 
@@ -5022,38 +5033,38 @@ static int sp_2048_mod_exp_18(sp_digit* r, const sp_digit* a, const sp_digit* e,
             c = 57;
         }
         if (i < 18) {
-            n = e[i--] << (64 - c);
+            n = (sp_uint64)e[i--] << (64 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 5) {
-            n |= e[i--] << (7 - c);
+            n |= (sp_uint64)e[i--] << (7 - c);
             c += 57;
         }
         y = (int)((n >> 59) & 0x1f);
-        n <<= 5;
+        n = (sp_uint64)n << 5;
         c -= 5;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 36);
         while ((i >= 0) || (c >= 5)) {
             if (c >= 5) {
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c -= 5;
             }
             else if (c == 0) {
-                n = e[i--] << 7;
+                n = (sp_uint64)e[i--] << 7;
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c = 52;
             }
             else {
                 y = (byte)((n >> 59) & 0x1f);
-                n = e[i--] << 7;
+                n = (sp_uint64)e[i--] << 7;
                 c = 5 - c;
                 y |= (byte)((n >> (64 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 57 - c;
             }
 
@@ -5226,30 +5237,30 @@ static void sp_2048_mont_shift_36(sp_digit* r, const sp_digit* a)
 
     s = a[36]; n = a[35] >> 53;
     for (i = 0; i < 32; i += 8) {
-        n += (sp_digit)((s & 0x1ffffffffffffffL) << 4); r[i+0] = (sp_digit)(n & 0x1ffffffffffffffL);
+        n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 4); r[i+0] = (sp_digit)(n & 0x1ffffffffffffffL);
         n >>= 57; s = a[i+37] + (s >> 57);
-        n += (sp_digit)((s & 0x1ffffffffffffffL) << 4); r[i+1] = (sp_digit)(n & 0x1ffffffffffffffL);
+        n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 4); r[i+1] = (sp_digit)(n & 0x1ffffffffffffffL);
         n >>= 57; s = a[i+38] + (s >> 57);
-        n += (sp_digit)((s & 0x1ffffffffffffffL) << 4); r[i+2] = (sp_digit)(n & 0x1ffffffffffffffL);
+        n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 4); r[i+2] = (sp_digit)(n & 0x1ffffffffffffffL);
         n >>= 57; s = a[i+39] + (s >> 57);
-        n += (sp_digit)((s & 0x1ffffffffffffffL) << 4); r[i+3] = (sp_digit)(n & 0x1ffffffffffffffL);
+        n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 4); r[i+3] = (sp_digit)(n & 0x1ffffffffffffffL);
         n >>= 57; s = a[i+40] + (s >> 57);
-        n += (sp_digit)((s & 0x1ffffffffffffffL) << 4); r[i+4] = (sp_digit)(n & 0x1ffffffffffffffL);
+        n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 4); r[i+4] = (sp_digit)(n & 0x1ffffffffffffffL);
         n >>= 57; s = a[i+41] + (s >> 57);
-        n += (sp_digit)((s & 0x1ffffffffffffffL) << 4); r[i+5] = (sp_digit)(n & 0x1ffffffffffffffL);
+        n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 4); r[i+5] = (sp_digit)(n & 0x1ffffffffffffffL);
         n >>= 57; s = a[i+42] + (s >> 57);
-        n += (sp_digit)((s & 0x1ffffffffffffffL) << 4); r[i+6] = (sp_digit)(n & 0x1ffffffffffffffL);
+        n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 4); r[i+6] = (sp_digit)(n & 0x1ffffffffffffffL);
         n >>= 57; s = a[i+43] + (s >> 57);
-        n += (sp_digit)((s & 0x1ffffffffffffffL) << 4); r[i+7] = (sp_digit)(n & 0x1ffffffffffffffL);
+        n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 4); r[i+7] = (sp_digit)(n & 0x1ffffffffffffffL);
         n >>= 57; s = a[i+44] + (s >> 57);
     }
-    n += (sp_digit)((s & 0x1ffffffffffffffL) << 4); r[32] = (sp_digit)(n & 0x1ffffffffffffffL);
+    n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 4); r[32] = (sp_digit)(n & 0x1ffffffffffffffL);
     n >>= 57; s = a[69] + (s >> 57);
-    n += (sp_digit)((s & 0x1ffffffffffffffL) << 4); r[33] = (sp_digit)(n & 0x1ffffffffffffffL);
+    n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 4); r[33] = (sp_digit)(n & 0x1ffffffffffffffL);
     n >>= 57; s = a[70] + (s >> 57);
-    n += (sp_digit)((s & 0x1ffffffffffffffL) << 4); r[34] = (sp_digit)(n & 0x1ffffffffffffffL);
+    n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 4); r[34] = (sp_digit)(n & 0x1ffffffffffffffL);
     n >>= 57; s = a[71] + (s >> 57);
-    n += s << 4;              r[35] = n;
+    n += (sp_uint64)s << 4;              r[35] = n;
     XMEMSET(&r[36], 0, sizeof(*r) * 36U);
 }
 
@@ -5414,18 +5425,18 @@ SP_NOINLINE static void sp_2048_rshift_36(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<32; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint64)a[i+1] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint64)a[i+2] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint64)a[i+3] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint64)a[i+4] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint64)a[i+5] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint64)a[i+6] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint64)a[i+7] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint64)a[i+8] << (57 - n)) & 0x1ffffffffffffffL);
     }
-    r[32] = (a[32] >> n) | (sp_digit)((a[33] << (57 - n)) & 0x1ffffffffffffffL);
-    r[33] = (a[33] >> n) | (sp_digit)((a[34] << (57 - n)) & 0x1ffffffffffffffL);
-    r[34] = (a[34] >> n) | (sp_digit)((a[35] << (57 - n)) & 0x1ffffffffffffffL);
+    r[32] = (a[32] >> n) | (sp_digit)(((sp_uint64)a[33] << (57 - n)) & 0x1ffffffffffffffL);
+    r[33] = (a[33] >> n) | (sp_digit)(((sp_uint64)a[34] << (57 - n)) & 0x1ffffffffffffffL);
+    r[34] = (a[34] >> n) | (sp_digit)(((sp_uint64)a[35] << (57 - n)) & 0x1ffffffffffffffL);
     r[35] = a[35] >> n;
 }
 
@@ -5464,7 +5475,7 @@ static WC_INLINE sp_digit sp_2048_div_word_36(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 55; i >= 1; i--) {
         t1 += t1 + (((sp_uint64)t0 >> 56) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint64)t0 << 1;
         t2 = (sp_digit)(((sp_uint64)(dv - t1)) >> 63);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -5659,9 +5670,9 @@ static int sp_2048_mod_exp_36(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 57;
-        c = bits % 57;
-        n = e[i--] << (57 - c);
+        i = (bits - 1) / 57;
+        c = ((bits - 1) % 57) + 1;
+        n = (sp_uint64)e[i--] << (57 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -5673,7 +5684,7 @@ static int sp_2048_mod_exp_36(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 56) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_2048_mont_mul_36(t[y^1], t[0], t[1], m, mp);
 
@@ -5735,9 +5746,9 @@ static int sp_2048_mod_exp_36(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 57;
-        c = bits % 57;
-        n = e[i--] << (57 - c);
+        i = (bits - 1) / 57;
+        c = ((bits - 1) % 57) + 1;
+        n = (sp_uint64)e[i--] << (57 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -5749,7 +5760,7 @@ static int sp_2048_mod_exp_36(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 56) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_2048_mont_mul_36(t[y^1], t[0], t[1], m, mp);
 
@@ -5834,38 +5845,38 @@ static int sp_2048_mod_exp_36(sp_digit* r, const sp_digit* a, const sp_digit* e,
             c = 57;
         }
         if (i < 36) {
-            n = e[i--] << (64 - c);
+            n = (sp_uint64)e[i--] << (64 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 4) {
-            n |= e[i--] << (7 - c);
+            n |= (sp_uint64)e[i--] << (7 - c);
             c += 57;
         }
         y = (int)((n >> 60) & 0xf);
-        n <<= 4;
+        n = (sp_uint64)n << 4;
         c -= 4;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 72);
         while ((i >= 0) || (c >= 4)) {
             if (c >= 4) {
                 y = (byte)((n >> 60) & 0xf);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c -= 4;
             }
             else if (c == 0) {
-                n = e[i--] << 7;
+                n = (sp_uint64)e[i--] << 7;
                 y = (byte)((n >> 60) & 0xf);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c = 53;
             }
             else {
                 y = (byte)((n >> 60) & 0xf);
-                n = e[i--] << 7;
+                n = (sp_uint64)e[i--] << 7;
                 c = 4 - c;
                 y |= (byte)((n >> (64 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 57 - c;
             }
 
@@ -5961,10 +5972,10 @@ static int sp_2048_mod_exp_36_nb(sp_2048_mod_exp_36_ctx* ctx,
     case 4: /* BIT_INIT: index the most-significant exponent limb without
               reading off the end when bits is an exact multiple of the
               limb width. (bits-1) keeps i within the populated range and
-              c = (bits-1) % @bits + 1 keeps the shift in [1, @bits]. */
+              c = (bits-1) % 57 + 1 keeps the shift in [1, 57]. */
         ctx->i = (ctx->bits - 1) / 57;
         ctx->c = ((ctx->bits - 1) % 57) + 1;
-        ctx->n = e[ctx->i--] << (57 - ctx->c);
+        ctx->n = (sp_uint64)e[ctx->i--] << (57 - ctx->c);
         ctx->state = 5;
         break;
     case 5: /* BIT_NEXT: refill on word boundary, peel one exponent bit */
@@ -5977,7 +5988,7 @@ static int sp_2048_mod_exp_36_nb(sp_2048_mod_exp_36_ctx* ctx,
             ctx->c = 57;
         }
         ctx->y = (byte)((ctx->n >> 56) & 1);
-        ctx->n <<= 1;
+        ctx->n = (sp_uint64)ctx->n << 1;
         ctx->state = 6;
         break;
     case 6: /* MUL: t[y^1] = t[0] * t[1] in Montgomery form */
@@ -6548,7 +6559,7 @@ static int sp_2048_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 36; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -6573,7 +6584,7 @@ static int sp_2048_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 36; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 57 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -6774,7 +6785,7 @@ SP_NOINLINE static void sp_2048_lshift_36(sp_digit* r, const sp_digit* a,
     r[2] = (sp_digit)(((s << n) | (t >> (57U - n))) & 0x1ffffffffffffffUL);
     s = (sp_int_digit)(a[1]); t = (sp_int_digit)(a[0]);
     r[1] = (sp_digit)(((s << n) | (t >> (57U - n))) & 0x1ffffffffffffffUL);
-    r[0] = (sp_digit)((a[0] << n) & 0x1ffffffffffffffL);
+    r[0] = (sp_digit)(((sp_uint64)a[0] << n) & 0x1ffffffffffffffL);
 }
 
 /* Modular exponentiate 2 to the e mod m. (r = 2^e mod m)
@@ -6820,38 +6831,38 @@ static int sp_2048_mod_exp_2_36(sp_digit* r, const sp_digit* e, int bits, const
             c = 57;
         }
         if (i < 36) {
-            n = e[i--] << (64 - c);
+            n = (sp_uint64)e[i--] << (64 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 5) {
-            n |= e[i--] << (7 - c);
+            n |= (sp_uint64)e[i--] << (7 - c);
             c += 57;
         }
         y = (int)((n >> 59) & 0x1f);
-        n <<= 5;
+        n = (sp_uint64)n << 5;
         c -= 5;
         sp_2048_lshift_36(r, norm, (byte)y);
         while ((i >= 0) || (c >= 5)) {
             if (c >= 5) {
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c -= 5;
             }
             else if (c == 0) {
-                n = e[i--] << 7;
+                n = (sp_uint64)e[i--] << 7;
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c = 52;
             }
             else {
                 y = (byte)((n >> 59) & 0x1f);
-                n = e[i--] << 7;
+                n = (sp_uint64)e[i--] << 7;
                 c = 5 - c;
                 y |= (byte)((n >> (64 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 57 - c;
             }
 
@@ -7081,7 +7092,7 @@ static void sp_3072_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     r[0] = 0;
     for (i = n-1; i >= 0; i--) {
-        r[j] |= (((sp_digit)a[i]) << s);
+        r[j] |= (((sp_uint64)a[i]) << s);
         if (s >= 52U) {
             r[j] &= 0xfffffffffffffffL;
             s = 60U - s;
@@ -7127,7 +7138,7 @@ static void sp_3072_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0xfffffffffffffffL;
         s = 60U - s;
         if (j + 1 >= size) {
@@ -7162,7 +7173,7 @@ static void sp_3072_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 60) {
             r[j] &= 0xfffffffffffffffL;
             if (j + 1 >= size) {
@@ -7211,7 +7222,7 @@ static void sp_3072_to_bin_52(sp_digit* r, byte* a)
     for (i=0; i<52 && j>=0; i++) {
         b = 0;
         /* lint allow cast of mismatch sp_digit and int */
-        a[j--] |= (byte)(r[i] << s); /*lint !e9033*/
+        a[j--] |= (byte)((sp_uint64)r[i] << s); /*lint !e9033*/
         b += 8 - s;
         if (j < 0) {
             break;
@@ -7348,10 +7359,10 @@ SP_NOINLINE static void sp_3072_sqr_52(sp_digit* r, const sp_digit* a)
  */
 static void sp_3072_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint64 x;
+    sp_uint64 b;
 
-    b = a[0];
+    b = (sp_uint64)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
@@ -7360,7 +7371,7 @@ static void sp_3072_mont_setup(const sp_digit* a, sp_digit* rho)
     x &= 0xfffffffffffffffL;
 
     /* rho = -1/m mod b */
-    *rho = ((sp_digit)1 << 60) - x;
+    *rho = (sp_digit)(((sp_uint64)1 << 60) - (sp_digit)x);
 }
 
 /* Multiply a by scalar b into r. (r = a * b)
@@ -7723,7 +7734,7 @@ SP_NOINLINE static void sp_3072_rshift_26(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<25; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (60 - n))) & 0xfffffffffffffffL);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint64)a[i + 1] << (60 - n))) & 0xfffffffffffffffL);
     }
     r[25] = a[25] >> n;
 }
@@ -7763,7 +7774,7 @@ static WC_INLINE sp_digit sp_3072_div_word_26(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 58; i >= 1; i--) {
         t1 += t1 + (((sp_uint64)t0 >> 59) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint64)t0 << 1;
         t2 = (sp_digit)(((sp_uint64)(dv - t1)) >> 63);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -7955,9 +7966,9 @@ static int sp_3072_mod_exp_26(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 60;
-        c = bits % 60;
-        n = e[i--] << (60 - c);
+        i = (bits - 1) / 60;
+        c = ((bits - 1) % 60) + 1;
+        n = (sp_uint64)e[i--] << (60 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -7969,7 +7980,7 @@ static int sp_3072_mod_exp_26(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 59) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_3072_mont_mul_26(t[y^1], t[0], t[1], m, mp);
 
@@ -8031,9 +8042,9 @@ static int sp_3072_mod_exp_26(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 60;
-        c = bits % 60;
-        n = e[i--] << (60 - c);
+        i = (bits - 1) / 60;
+        c = ((bits - 1) % 60) + 1;
+        n = (sp_uint64)e[i--] << (60 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -8045,7 +8056,7 @@ static int sp_3072_mod_exp_26(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 59) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_3072_mont_mul_26(t[y^1], t[0], t[1], m, mp);
 
@@ -8146,38 +8157,38 @@ static int sp_3072_mod_exp_26(sp_digit* r, const sp_digit* a, const sp_digit* e,
             c = 60;
         }
         if (i < 26) {
-            n = e[i--] << (64 - c);
+            n = (sp_uint64)e[i--] << (64 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 5) {
-            n |= e[i--] << (4 - c);
+            n |= (sp_uint64)e[i--] << (4 - c);
             c += 60;
         }
         y = (int)((n >> 59) & 0x1f);
-        n <<= 5;
+        n = (sp_uint64)n << 5;
         c -= 5;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 52);
         while ((i >= 0) || (c >= 5)) {
             if (c >= 5) {
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c -= 5;
             }
             else if (c == 0) {
-                n = e[i--] << 4;
+                n = (sp_uint64)e[i--] << 4;
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c = 55;
             }
             else {
                 y = (byte)((n >> 59) & 0x1f);
-                n = e[i--] << 4;
+                n = (sp_uint64)e[i--] << 4;
                 c = 5 - c;
                 y |= (byte)((n >> (64 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 60 - c;
             }
 
@@ -8493,7 +8504,7 @@ SP_NOINLINE static void sp_3072_rshift_52(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<51; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (60 - n))) & 0xfffffffffffffffL);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint64)a[i + 1] << (60 - n))) & 0xfffffffffffffffL);
     }
     r[51] = a[51] >> n;
 }
@@ -8533,7 +8544,7 @@ static WC_INLINE sp_digit sp_3072_div_word_52(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 58; i >= 1; i--) {
         t1 += t1 + (((sp_uint64)t0 >> 59) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint64)t0 << 1;
         t2 = (sp_digit)(((sp_uint64)(dv - t1)) >> 63);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -8726,9 +8737,9 @@ static int sp_3072_mod_exp_52(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 60;
-        c = bits % 60;
-        n = e[i--] << (60 - c);
+        i = (bits - 1) / 60;
+        c = ((bits - 1) % 60) + 1;
+        n = (sp_uint64)e[i--] << (60 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -8740,7 +8751,7 @@ static int sp_3072_mod_exp_52(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 59) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_3072_mont_mul_52(t[y^1], t[0], t[1], m, mp);
 
@@ -8802,9 +8813,9 @@ static int sp_3072_mod_exp_52(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 60;
-        c = bits % 60;
-        n = e[i--] << (60 - c);
+        i = (bits - 1) / 60;
+        c = ((bits - 1) % 60) + 1;
+        n = (sp_uint64)e[i--] << (60 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -8816,7 +8827,7 @@ static int sp_3072_mod_exp_52(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 59) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_3072_mont_mul_52(t[y^1], t[0], t[1], m, mp);
 
@@ -8901,38 +8912,38 @@ static int sp_3072_mod_exp_52(sp_digit* r, const sp_digit* a, const sp_digit* e,
             c = 60;
         }
         if (i < 52) {
-            n = e[i--] << (64 - c);
+            n = (sp_uint64)e[i--] << (64 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 4) {
-            n |= e[i--] << (4 - c);
+            n |= (sp_uint64)e[i--] << (4 - c);
             c += 60;
         }
         y = (int)((n >> 60) & 0xf);
-        n <<= 4;
+        n = (sp_uint64)n << 4;
         c -= 4;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 104);
         while ((i >= 0) || (c >= 4)) {
             if (c >= 4) {
                 y = (byte)((n >> 60) & 0xf);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c -= 4;
             }
             else if (c == 0) {
-                n = e[i--] << 4;
+                n = (sp_uint64)e[i--] << 4;
                 y = (byte)((n >> 60) & 0xf);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c = 56;
             }
             else {
                 y = (byte)((n >> 60) & 0xf);
-                n = e[i--] << 4;
+                n = (sp_uint64)e[i--] << 4;
                 c = 4 - c;
                 y |= (byte)((n >> (64 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 60 - c;
             }
 
@@ -9028,10 +9039,10 @@ static int sp_3072_mod_exp_52_nb(sp_3072_mod_exp_52_ctx* ctx,
     case 4: /* BIT_INIT: index the most-significant exponent limb without
               reading off the end when bits is an exact multiple of the
               limb width. (bits-1) keeps i within the populated range and
-              c = (bits-1) % @bits + 1 keeps the shift in [1, @bits]. */
+              c = (bits-1) % 60 + 1 keeps the shift in [1, 60]. */
         ctx->i = (ctx->bits - 1) / 60;
         ctx->c = ((ctx->bits - 1) % 60) + 1;
-        ctx->n = e[ctx->i--] << (60 - ctx->c);
+        ctx->n = (sp_uint64)e[ctx->i--] << (60 - ctx->c);
         ctx->state = 5;
         break;
     case 5: /* BIT_NEXT: refill on word boundary, peel one exponent bit */
@@ -9044,7 +9055,7 @@ static int sp_3072_mod_exp_52_nb(sp_3072_mod_exp_52_ctx* ctx,
             ctx->c = 60;
         }
         ctx->y = (byte)((ctx->n >> 59) & 1);
-        ctx->n <<= 1;
+        ctx->n = (sp_uint64)ctx->n << 1;
         ctx->state = 6;
         break;
     case 6: /* MUL: t[y^1] = t[0] * t[1] in Montgomery form */
@@ -9815,7 +9826,7 @@ static int sp_3072_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 52; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -9840,7 +9851,7 @@ static int sp_3072_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 52; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 60 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -10056,9 +10067,9 @@ SP_NOINLINE static void sp_3072_lshift_52(sp_digit* r, const sp_digit* a,
 
     r[52] = a[51] >> (60 - n);
     for (i=51; i>0; i--) {
-        r[i] = (sp_digit)(((a[i] << n) | (a[i-1] >> (60 - n))) & 0xfffffffffffffffL);
+        r[i] = (sp_digit)((((sp_uint64)a[i] << n) | (a[i-1] >> (60 - n))) & 0xfffffffffffffffL);
     }
-    r[0] = (sp_digit)((a[0] << n) & 0xfffffffffffffffL);
+    r[0] = (sp_digit)(((sp_uint64)a[0] << n) & 0xfffffffffffffffL);
 }
 
 /* Modular exponentiate 2 to the e mod m. (r = 2^e mod m)
@@ -10104,38 +10115,38 @@ static int sp_3072_mod_exp_2_52(sp_digit* r, const sp_digit* e, int bits, const
             c = 60;
         }
         if (i < 52) {
-            n = e[i--] << (64 - c);
+            n = (sp_uint64)e[i--] << (64 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 5) {
-            n |= e[i--] << (4 - c);
+            n |= (sp_uint64)e[i--] << (4 - c);
             c += 60;
         }
         y = (int)((n >> 59) & 0x1f);
-        n <<= 5;
+        n = (sp_uint64)n << 5;
         c -= 5;
         sp_3072_lshift_52(r, norm, (byte)y);
         while ((i >= 0) || (c >= 5)) {
             if (c >= 5) {
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c -= 5;
             }
             else if (c == 0) {
-                n = e[i--] << 4;
+                n = (sp_uint64)e[i--] << 4;
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c = 55;
             }
             else {
                 y = (byte)((n >> 59) & 0x1f);
-                n = e[i--] << 4;
+                n = (sp_uint64)e[i--] << 4;
                 c = 5 - c;
                 y |= (byte)((n >> (64 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 60 - c;
             }
 
@@ -10463,7 +10474,7 @@ static void sp_3072_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     r[0] = 0;
     for (i = n-1; i >= 0; i--) {
-        r[j] |= (((sp_digit)a[i]) << s);
+        r[j] |= (((sp_uint64)a[i]) << s);
         if (s >= 49U) {
             r[j] &= 0x1ffffffffffffffL;
             s = 57U - s;
@@ -10509,7 +10520,7 @@ static void sp_3072_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0x1ffffffffffffffL;
         s = 57U - s;
         if (j + 1 >= size) {
@@ -10544,7 +10555,7 @@ static void sp_3072_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 57) {
             r[j] &= 0x1ffffffffffffffL;
             if (j + 1 >= size) {
@@ -10593,7 +10604,7 @@ static void sp_3072_to_bin_54(sp_digit* r, byte* a)
     for (i=0; i<54 && j>=0; i++) {
         b = 0;
         /* lint allow cast of mismatch sp_digit and int */
-        a[j--] |= (byte)(r[i] << s); /*lint !e9033*/
+        a[j--] |= (byte)((sp_uint64)r[i] << s); /*lint !e9033*/
         b += 8 - s;
         if (j < 0) {
             break;
@@ -11165,10 +11176,10 @@ SP_NOINLINE static void sp_3072_sqr_54(sp_digit* r, const sp_digit* a)
  */
 static void sp_3072_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint64 x;
+    sp_uint64 b;
 
-    b = a[0];
+    b = (sp_uint64)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
@@ -11177,7 +11188,7 @@ static void sp_3072_mont_setup(const sp_digit* a, sp_digit* rho)
     x &= 0x1ffffffffffffffL;
 
     /* rho = -1/m mod b */
-    *rho = ((sp_digit)1 << 57) - x;
+    *rho = (sp_digit)(((sp_uint64)1 << 57) - (sp_digit)x);
 }
 
 /* Multiply a by scalar b into r. (r = a * b)
@@ -11396,28 +11407,28 @@ static void sp_3072_mont_shift_27(sp_digit* r, const sp_digit* a)
 
     s = a[27]; n = a[26] >> 54;
     for (i = 0; i < 24; i += 8) {
-        n += (sp_digit)((s & 0x1ffffffffffffffL) << 3); r[i+0] = (sp_digit)(n & 0x1ffffffffffffffL);
+        n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 3); r[i+0] = (sp_digit)(n & 0x1ffffffffffffffL);
         n >>= 57; s = a[i+28] + (s >> 57);
-        n += (sp_digit)((s & 0x1ffffffffffffffL) << 3); r[i+1] = (sp_digit)(n & 0x1ffffffffffffffL);
+        n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 3); r[i+1] = (sp_digit)(n & 0x1ffffffffffffffL);
         n >>= 57; s = a[i+29] + (s >> 57);
-        n += (sp_digit)((s & 0x1ffffffffffffffL) << 3); r[i+2] = (sp_digit)(n & 0x1ffffffffffffffL);
+        n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 3); r[i+2] = (sp_digit)(n & 0x1ffffffffffffffL);
         n >>= 57; s = a[i+30] + (s >> 57);
-        n += (sp_digit)((s & 0x1ffffffffffffffL) << 3); r[i+3] = (sp_digit)(n & 0x1ffffffffffffffL);
+        n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 3); r[i+3] = (sp_digit)(n & 0x1ffffffffffffffL);
         n >>= 57; s = a[i+31] + (s >> 57);
-        n += (sp_digit)((s & 0x1ffffffffffffffL) << 3); r[i+4] = (sp_digit)(n & 0x1ffffffffffffffL);
+        n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 3); r[i+4] = (sp_digit)(n & 0x1ffffffffffffffL);
         n >>= 57; s = a[i+32] + (s >> 57);
-        n += (sp_digit)((s & 0x1ffffffffffffffL) << 3); r[i+5] = (sp_digit)(n & 0x1ffffffffffffffL);
+        n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 3); r[i+5] = (sp_digit)(n & 0x1ffffffffffffffL);
         n >>= 57; s = a[i+33] + (s >> 57);
-        n += (sp_digit)((s & 0x1ffffffffffffffL) << 3); r[i+6] = (sp_digit)(n & 0x1ffffffffffffffL);
+        n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 3); r[i+6] = (sp_digit)(n & 0x1ffffffffffffffL);
         n >>= 57; s = a[i+34] + (s >> 57);
-        n += (sp_digit)((s & 0x1ffffffffffffffL) << 3); r[i+7] = (sp_digit)(n & 0x1ffffffffffffffL);
+        n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 3); r[i+7] = (sp_digit)(n & 0x1ffffffffffffffL);
         n >>= 57; s = a[i+35] + (s >> 57);
     }
-    n += (sp_digit)((s & 0x1ffffffffffffffL) << 3); r[24] = (sp_digit)(n & 0x1ffffffffffffffL);
+    n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 3); r[24] = (sp_digit)(n & 0x1ffffffffffffffL);
     n >>= 57; s = a[52] + (s >> 57);
-    n += (sp_digit)((s & 0x1ffffffffffffffL) << 3); r[25] = (sp_digit)(n & 0x1ffffffffffffffL);
+    n += (sp_digit)((sp_uint64)(s & 0x1ffffffffffffffL) << 3); r[25] = (sp_digit)(n & 0x1ffffffffffffffL);
     n >>= 57; s = a[53] + (s >> 57);
-    n += s << 3;              r[26] = n;
+    n += (sp_uint64)s << 3;              r[26] = n;
     XMEMSET(&r[27], 0, sizeof(*r) * 27U);
 }
 
@@ -11565,17 +11576,17 @@ SP_NOINLINE static void sp_3072_rshift_27(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<24; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint64)a[i+1] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint64)a[i+2] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint64)a[i+3] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint64)a[i+4] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint64)a[i+5] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint64)a[i+6] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint64)a[i+7] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint64)a[i+8] << (57 - n)) & 0x1ffffffffffffffL);
     }
-    r[24] = (a[24] >> n) | (sp_digit)((a[25] << (57 - n)) & 0x1ffffffffffffffL);
-    r[25] = (a[25] >> n) | (sp_digit)((a[26] << (57 - n)) & 0x1ffffffffffffffL);
+    r[24] = (a[24] >> n) | (sp_digit)(((sp_uint64)a[25] << (57 - n)) & 0x1ffffffffffffffL);
+    r[25] = (a[25] >> n) | (sp_digit)(((sp_uint64)a[26] << (57 - n)) & 0x1ffffffffffffffL);
     r[26] = a[26] >> n;
 }
 
@@ -11614,7 +11625,7 @@ static WC_INLINE sp_digit sp_3072_div_word_27(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 55; i >= 1; i--) {
         t1 += t1 + (((sp_uint64)t0 >> 56) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint64)t0 << 1;
         t2 = (sp_digit)(((sp_uint64)(dv - t1)) >> 63);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -11806,9 +11817,9 @@ static int sp_3072_mod_exp_27(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 57;
-        c = bits % 57;
-        n = e[i--] << (57 - c);
+        i = (bits - 1) / 57;
+        c = ((bits - 1) % 57) + 1;
+        n = (sp_uint64)e[i--] << (57 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -11820,7 +11831,7 @@ static int sp_3072_mod_exp_27(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 56) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_3072_mont_mul_27(t[y^1], t[0], t[1], m, mp);
 
@@ -11882,9 +11893,9 @@ static int sp_3072_mod_exp_27(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 57;
-        c = bits % 57;
-        n = e[i--] << (57 - c);
+        i = (bits - 1) / 57;
+        c = ((bits - 1) % 57) + 1;
+        n = (sp_uint64)e[i--] << (57 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -11896,7 +11907,7 @@ static int sp_3072_mod_exp_27(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 56) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_3072_mont_mul_27(t[y^1], t[0], t[1], m, mp);
 
@@ -11997,38 +12008,38 @@ static int sp_3072_mod_exp_27(sp_digit* r, const sp_digit* a, const sp_digit* e,
             c = 57;
         }
         if (i < 27) {
-            n = e[i--] << (64 - c);
+            n = (sp_uint64)e[i--] << (64 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 5) {
-            n |= e[i--] << (7 - c);
+            n |= (sp_uint64)e[i--] << (7 - c);
             c += 57;
         }
         y = (int)((n >> 59) & 0x1f);
-        n <<= 5;
+        n = (sp_uint64)n << 5;
         c -= 5;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 54);
         while ((i >= 0) || (c >= 5)) {
             if (c >= 5) {
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c -= 5;
             }
             else if (c == 0) {
-                n = e[i--] << 7;
+                n = (sp_uint64)e[i--] << 7;
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c = 52;
             }
             else {
                 y = (byte)((n >> 59) & 0x1f);
-                n = e[i--] << 7;
+                n = (sp_uint64)e[i--] << 7;
                 c = 5 - c;
                 y |= (byte)((n >> (64 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 57 - c;
             }
 
@@ -12398,20 +12409,20 @@ SP_NOINLINE static void sp_3072_rshift_54(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<48; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint64)a[i+1] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint64)a[i+2] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint64)a[i+3] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint64)a[i+4] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint64)a[i+5] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint64)a[i+6] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint64)a[i+7] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint64)a[i+8] << (57 - n)) & 0x1ffffffffffffffL);
     }
-    r[48] = (a[48] >> n) | (sp_digit)((a[49] << (57 - n)) & 0x1ffffffffffffffL);
-    r[49] = (a[49] >> n) | (sp_digit)((a[50] << (57 - n)) & 0x1ffffffffffffffL);
-    r[50] = (a[50] >> n) | (sp_digit)((a[51] << (57 - n)) & 0x1ffffffffffffffL);
-    r[51] = (a[51] >> n) | (sp_digit)((a[52] << (57 - n)) & 0x1ffffffffffffffL);
-    r[52] = (a[52] >> n) | (sp_digit)((a[53] << (57 - n)) & 0x1ffffffffffffffL);
+    r[48] = (a[48] >> n) | (sp_digit)(((sp_uint64)a[49] << (57 - n)) & 0x1ffffffffffffffL);
+    r[49] = (a[49] >> n) | (sp_digit)(((sp_uint64)a[50] << (57 - n)) & 0x1ffffffffffffffL);
+    r[50] = (a[50] >> n) | (sp_digit)(((sp_uint64)a[51] << (57 - n)) & 0x1ffffffffffffffL);
+    r[51] = (a[51] >> n) | (sp_digit)(((sp_uint64)a[52] << (57 - n)) & 0x1ffffffffffffffL);
+    r[52] = (a[52] >> n) | (sp_digit)(((sp_uint64)a[53] << (57 - n)) & 0x1ffffffffffffffL);
     r[53] = a[53] >> n;
 }
 
@@ -12450,7 +12461,7 @@ static WC_INLINE sp_digit sp_3072_div_word_54(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 55; i >= 1; i--) {
         t1 += t1 + (((sp_uint64)t0 >> 56) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint64)t0 << 1;
         t2 = (sp_digit)(((sp_uint64)(dv - t1)) >> 63);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -12645,9 +12656,9 @@ static int sp_3072_mod_exp_54(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 57;
-        c = bits % 57;
-        n = e[i--] << (57 - c);
+        i = (bits - 1) / 57;
+        c = ((bits - 1) % 57) + 1;
+        n = (sp_uint64)e[i--] << (57 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -12659,7 +12670,7 @@ static int sp_3072_mod_exp_54(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 56) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_3072_mont_mul_54(t[y^1], t[0], t[1], m, mp);
 
@@ -12721,9 +12732,9 @@ static int sp_3072_mod_exp_54(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 57;
-        c = bits % 57;
-        n = e[i--] << (57 - c);
+        i = (bits - 1) / 57;
+        c = ((bits - 1) % 57) + 1;
+        n = (sp_uint64)e[i--] << (57 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -12735,7 +12746,7 @@ static int sp_3072_mod_exp_54(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 56) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_3072_mont_mul_54(t[y^1], t[0], t[1], m, mp);
 
@@ -12820,38 +12831,38 @@ static int sp_3072_mod_exp_54(sp_digit* r, const sp_digit* a, const sp_digit* e,
             c = 57;
         }
         if (i < 54) {
-            n = e[i--] << (64 - c);
+            n = (sp_uint64)e[i--] << (64 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 4) {
-            n |= e[i--] << (7 - c);
+            n |= (sp_uint64)e[i--] << (7 - c);
             c += 57;
         }
         y = (int)((n >> 60) & 0xf);
-        n <<= 4;
+        n = (sp_uint64)n << 4;
         c -= 4;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 108);
         while ((i >= 0) || (c >= 4)) {
             if (c >= 4) {
                 y = (byte)((n >> 60) & 0xf);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c -= 4;
             }
             else if (c == 0) {
-                n = e[i--] << 7;
+                n = (sp_uint64)e[i--] << 7;
                 y = (byte)((n >> 60) & 0xf);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c = 53;
             }
             else {
                 y = (byte)((n >> 60) & 0xf);
-                n = e[i--] << 7;
+                n = (sp_uint64)e[i--] << 7;
                 c = 4 - c;
                 y |= (byte)((n >> (64 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 57 - c;
             }
 
@@ -12947,10 +12958,10 @@ static int sp_3072_mod_exp_54_nb(sp_3072_mod_exp_54_ctx* ctx,
     case 4: /* BIT_INIT: index the most-significant exponent limb without
               reading off the end when bits is an exact multiple of the
               limb width. (bits-1) keeps i within the populated range and
-              c = (bits-1) % @bits + 1 keeps the shift in [1, @bits]. */
+              c = (bits-1) % 57 + 1 keeps the shift in [1, 57]. */
         ctx->i = (ctx->bits - 1) / 57;
         ctx->c = ((ctx->bits - 1) % 57) + 1;
-        ctx->n = e[ctx->i--] << (57 - ctx->c);
+        ctx->n = (sp_uint64)e[ctx->i--] << (57 - ctx->c);
         ctx->state = 5;
         break;
     case 5: /* BIT_NEXT: refill on word boundary, peel one exponent bit */
@@ -12963,7 +12974,7 @@ static int sp_3072_mod_exp_54_nb(sp_3072_mod_exp_54_ctx* ctx,
             ctx->c = 57;
         }
         ctx->y = (byte)((ctx->n >> 56) & 1);
-        ctx->n <<= 1;
+        ctx->n = (sp_uint64)ctx->n << 1;
         ctx->state = 6;
         break;
     case 6: /* MUL: t[y^1] = t[0] * t[1] in Montgomery form */
@@ -13534,7 +13545,7 @@ static int sp_3072_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 54; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -13559,7 +13570,7 @@ static int sp_3072_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 54; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 57 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -13796,7 +13807,7 @@ SP_NOINLINE static void sp_3072_lshift_54(sp_digit* r, const sp_digit* a,
     r[2] = (sp_digit)(((s << n) | (t >> (57U - n))) & 0x1ffffffffffffffUL);
     s = (sp_int_digit)(a[1]); t = (sp_int_digit)(a[0]);
     r[1] = (sp_digit)(((s << n) | (t >> (57U - n))) & 0x1ffffffffffffffUL);
-    r[0] = (sp_digit)((a[0] << n) & 0x1ffffffffffffffL);
+    r[0] = (sp_digit)(((sp_uint64)a[0] << n) & 0x1ffffffffffffffL);
 }
 
 /* Modular exponentiate 2 to the e mod m. (r = 2^e mod m)
@@ -13842,38 +13853,38 @@ static int sp_3072_mod_exp_2_54(sp_digit* r, const sp_digit* e, int bits, const
             c = 57;
         }
         if (i < 54) {
-            n = e[i--] << (64 - c);
+            n = (sp_uint64)e[i--] << (64 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 5) {
-            n |= e[i--] << (7 - c);
+            n |= (sp_uint64)e[i--] << (7 - c);
             c += 57;
         }
         y = (int)((n >> 59) & 0x1f);
-        n <<= 5;
+        n = (sp_uint64)n << 5;
         c -= 5;
         sp_3072_lshift_54(r, norm, (byte)y);
         while ((i >= 0) || (c >= 5)) {
             if (c >= 5) {
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c -= 5;
             }
             else if (c == 0) {
-                n = e[i--] << 7;
+                n = (sp_uint64)e[i--] << 7;
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c = 52;
             }
             else {
                 y = (byte)((n >> 59) & 0x1f);
-                n = e[i--] << 7;
+                n = (sp_uint64)e[i--] << 7;
                 c = 5 - c;
                 y |= (byte)((n >> (64 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 57 - c;
             }
 
@@ -14103,7 +14114,7 @@ static void sp_4096_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     r[0] = 0;
     for (i = n-1; i >= 0; i--) {
-        r[j] |= (((sp_digit)a[i]) << s);
+        r[j] |= (((sp_uint64)a[i]) << s);
         if (s >= 51U) {
             r[j] &= 0x7ffffffffffffffL;
             s = 59U - s;
@@ -14149,7 +14160,7 @@ static void sp_4096_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0x7ffffffffffffffL;
         s = 59U - s;
         if (j + 1 >= size) {
@@ -14184,7 +14195,7 @@ static void sp_4096_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 59) {
             r[j] &= 0x7ffffffffffffffL;
             if (j + 1 >= size) {
@@ -14233,7 +14244,7 @@ static void sp_4096_to_bin_70(sp_digit* r, byte* a)
     for (i=0; i<70 && j>=0; i++) {
         b = 0;
         /* lint allow cast of mismatch sp_digit and int */
-        a[j--] |= (byte)(r[i] << s); /*lint !e9033*/
+        a[j--] |= (byte)((sp_uint64)r[i] << s); /*lint !e9033*/
         b += 8 - s;
         if (j < 0) {
             break;
@@ -14372,10 +14383,10 @@ SP_NOINLINE static void sp_4096_sqr_70(sp_digit* r, const sp_digit* a)
  */
 static void sp_4096_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint64 x;
+    sp_uint64 b;
 
-    b = a[0];
+    b = (sp_uint64)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
@@ -14384,7 +14395,7 @@ static void sp_4096_mont_setup(const sp_digit* a, sp_digit* rho)
     x &= 0x7ffffffffffffffL;
 
     /* rho = -1/m mod b */
-    *rho = ((sp_digit)1 << 59) - x;
+    *rho = (sp_digit)(((sp_uint64)1 << 59) - (sp_digit)x);
 }
 
 /* Multiply a by scalar b into r. (r = a * b)
@@ -14751,7 +14762,7 @@ SP_NOINLINE static void sp_4096_rshift_35(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<34; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (59 - n))) & 0x7ffffffffffffffL);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint64)a[i + 1] << (59 - n))) & 0x7ffffffffffffffL);
     }
     r[34] = a[34] >> n;
 }
@@ -14791,7 +14802,7 @@ static WC_INLINE sp_digit sp_4096_div_word_35(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 57; i >= 1; i--) {
         t1 += t1 + (((sp_uint64)t0 >> 58) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint64)t0 << 1;
         t2 = (sp_digit)(((sp_uint64)(dv - t1)) >> 63);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -14983,9 +14994,9 @@ static int sp_4096_mod_exp_35(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 59;
-        c = bits % 59;
-        n = e[i--] << (59 - c);
+        i = (bits - 1) / 59;
+        c = ((bits - 1) % 59) + 1;
+        n = (sp_uint64)e[i--] << (59 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -14997,7 +15008,7 @@ static int sp_4096_mod_exp_35(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 58) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_4096_mont_mul_35(t[y^1], t[0], t[1], m, mp);
 
@@ -15059,9 +15070,9 @@ static int sp_4096_mod_exp_35(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 59;
-        c = bits % 59;
-        n = e[i--] << (59 - c);
+        i = (bits - 1) / 59;
+        c = ((bits - 1) % 59) + 1;
+        n = (sp_uint64)e[i--] << (59 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -15073,7 +15084,7 @@ static int sp_4096_mod_exp_35(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 58) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_4096_mont_mul_35(t[y^1], t[0], t[1], m, mp);
 
@@ -15174,38 +15185,38 @@ static int sp_4096_mod_exp_35(sp_digit* r, const sp_digit* a, const sp_digit* e,
             c = 59;
         }
         if (i < 35) {
-            n = e[i--] << (64 - c);
+            n = (sp_uint64)e[i--] << (64 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 5) {
-            n |= e[i--] << (5 - c);
+            n |= (sp_uint64)e[i--] << (5 - c);
             c += 59;
         }
         y = (int)((n >> 59) & 0x1f);
-        n <<= 5;
+        n = (sp_uint64)n << 5;
         c -= 5;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 70);
         while ((i >= 0) || (c >= 5)) {
             if (c >= 5) {
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c -= 5;
             }
             else if (c == 0) {
-                n = e[i--] << 5;
+                n = (sp_uint64)e[i--] << 5;
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c = 54;
             }
             else {
                 y = (byte)((n >> 59) & 0x1f);
-                n = e[i--] << 5;
+                n = (sp_uint64)e[i--] << 5;
                 c = 5 - c;
                 y |= (byte)((n >> (64 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 59 - c;
             }
 
@@ -15516,7 +15527,7 @@ SP_NOINLINE static void sp_4096_rshift_70(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<69; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (59 - n))) & 0x7ffffffffffffffL);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint64)a[i + 1] << (59 - n))) & 0x7ffffffffffffffL);
     }
     r[69] = a[69] >> n;
 }
@@ -15556,7 +15567,7 @@ static WC_INLINE sp_digit sp_4096_div_word_70(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 57; i >= 1; i--) {
         t1 += t1 + (((sp_uint64)t0 >> 58) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint64)t0 << 1;
         t2 = (sp_digit)(((sp_uint64)(dv - t1)) >> 63);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -15749,9 +15760,9 @@ static int sp_4096_mod_exp_70(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 59;
-        c = bits % 59;
-        n = e[i--] << (59 - c);
+        i = (bits - 1) / 59;
+        c = ((bits - 1) % 59) + 1;
+        n = (sp_uint64)e[i--] << (59 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -15763,7 +15774,7 @@ static int sp_4096_mod_exp_70(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 58) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_4096_mont_mul_70(t[y^1], t[0], t[1], m, mp);
 
@@ -15825,9 +15836,9 @@ static int sp_4096_mod_exp_70(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 59;
-        c = bits % 59;
-        n = e[i--] << (59 - c);
+        i = (bits - 1) / 59;
+        c = ((bits - 1) % 59) + 1;
+        n = (sp_uint64)e[i--] << (59 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -15839,7 +15850,7 @@ static int sp_4096_mod_exp_70(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 58) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_4096_mont_mul_70(t[y^1], t[0], t[1], m, mp);
 
@@ -15924,38 +15935,38 @@ static int sp_4096_mod_exp_70(sp_digit* r, const sp_digit* a, const sp_digit* e,
             c = 59;
         }
         if (i < 70) {
-            n = e[i--] << (64 - c);
+            n = (sp_uint64)e[i--] << (64 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 4) {
-            n |= e[i--] << (5 - c);
+            n |= (sp_uint64)e[i--] << (5 - c);
             c += 59;
         }
         y = (int)((n >> 60) & 0xf);
-        n <<= 4;
+        n = (sp_uint64)n << 4;
         c -= 4;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 140);
         while ((i >= 0) || (c >= 4)) {
             if (c >= 4) {
                 y = (byte)((n >> 60) & 0xf);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c -= 4;
             }
             else if (c == 0) {
-                n = e[i--] << 5;
+                n = (sp_uint64)e[i--] << 5;
                 y = (byte)((n >> 60) & 0xf);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c = 55;
             }
             else {
                 y = (byte)((n >> 60) & 0xf);
-                n = e[i--] << 5;
+                n = (sp_uint64)e[i--] << 5;
                 c = 4 - c;
                 y |= (byte)((n >> (64 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 59 - c;
             }
 
@@ -16051,10 +16062,10 @@ static int sp_4096_mod_exp_70_nb(sp_4096_mod_exp_70_ctx* ctx,
     case 4: /* BIT_INIT: index the most-significant exponent limb without
               reading off the end when bits is an exact multiple of the
               limb width. (bits-1) keeps i within the populated range and
-              c = (bits-1) % @bits + 1 keeps the shift in [1, @bits]. */
+              c = (bits-1) % 59 + 1 keeps the shift in [1, 59]. */
         ctx->i = (ctx->bits - 1) / 59;
         ctx->c = ((ctx->bits - 1) % 59) + 1;
-        ctx->n = e[ctx->i--] << (59 - ctx->c);
+        ctx->n = (sp_uint64)e[ctx->i--] << (59 - ctx->c);
         ctx->state = 5;
         break;
     case 5: /* BIT_NEXT: refill on word boundary, peel one exponent bit */
@@ -16067,7 +16078,7 @@ static int sp_4096_mod_exp_70_nb(sp_4096_mod_exp_70_ctx* ctx,
             ctx->c = 59;
         }
         ctx->y = (byte)((ctx->n >> 58) & 1);
-        ctx->n <<= 1;
+        ctx->n = (sp_uint64)ctx->n << 1;
         ctx->state = 6;
         break;
     case 6: /* MUL: t[y^1] = t[0] * t[1] in Montgomery form */
@@ -16838,7 +16849,7 @@ static int sp_4096_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 70; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -16863,7 +16874,7 @@ static int sp_4096_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 70; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 59 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -17079,9 +17090,9 @@ SP_NOINLINE static void sp_4096_lshift_70(sp_digit* r, const sp_digit* a,
 
     r[70] = a[69] >> (59 - n);
     for (i=69; i>0; i--) {
-        r[i] = (sp_digit)(((a[i] << n) | (a[i-1] >> (59 - n))) & 0x7ffffffffffffffL);
+        r[i] = (sp_digit)((((sp_uint64)a[i] << n) | (a[i-1] >> (59 - n))) & 0x7ffffffffffffffL);
     }
-    r[0] = (sp_digit)((a[0] << n) & 0x7ffffffffffffffL);
+    r[0] = (sp_digit)(((sp_uint64)a[0] << n) & 0x7ffffffffffffffL);
 }
 
 /* Modular exponentiate 2 to the e mod m. (r = 2^e mod m)
@@ -17127,38 +17138,38 @@ static int sp_4096_mod_exp_2_70(sp_digit* r, const sp_digit* e, int bits, const
             c = 59;
         }
         if (i < 70) {
-            n = e[i--] << (64 - c);
+            n = (sp_uint64)e[i--] << (64 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 5) {
-            n |= e[i--] << (5 - c);
+            n |= (sp_uint64)e[i--] << (5 - c);
             c += 59;
         }
         y = (int)((n >> 59) & 0x1f);
-        n <<= 5;
+        n = (sp_uint64)n << 5;
         c -= 5;
         sp_4096_lshift_70(r, norm, (byte)y);
         while ((i >= 0) || (c >= 5)) {
             if (c >= 5) {
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c -= 5;
             }
             else if (c == 0) {
-                n = e[i--] << 5;
+                n = (sp_uint64)e[i--] << 5;
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c = 54;
             }
             else {
                 y = (byte)((n >> 59) & 0x1f);
-                n = e[i--] << 5;
+                n = (sp_uint64)e[i--] << 5;
                 c = 5 - c;
                 y |= (byte)((n >> (64 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 59 - c;
             }
 
@@ -17385,7 +17396,7 @@ static void sp_4096_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     r[0] = 0;
     for (i = n-1; i >= 0; i--) {
-        r[j] |= (((sp_digit)a[i]) << s);
+        r[j] |= (((sp_uint64)a[i]) << s);
         if (s >= 45U) {
             r[j] &= 0x1fffffffffffffL;
             s = 53U - s;
@@ -17431,7 +17442,7 @@ static void sp_4096_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0x1fffffffffffffL;
         s = 53U - s;
         if (j + 1 >= size) {
@@ -17466,7 +17477,7 @@ static void sp_4096_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 53) {
             r[j] &= 0x1fffffffffffffL;
             if (j + 1 >= size) {
@@ -17515,7 +17526,7 @@ static void sp_4096_to_bin_78(sp_digit* r, byte* a)
     for (i=0; i<78 && j>=0; i++) {
         b = 0;
         /* lint allow cast of mismatch sp_digit and int */
-        a[j--] |= (byte)(r[i] << s); /*lint !e9033*/
+        a[j--] |= (byte)((sp_uint64)r[i] << s); /*lint !e9033*/
         b += 8 - s;
         if (j < 0) {
             break;
@@ -18255,10 +18266,10 @@ SP_NOINLINE static void sp_4096_sqr_78(sp_digit* r, const sp_digit* a)
  */
 static void sp_4096_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint64 x;
+    sp_uint64 b;
 
-    b = a[0];
+    b = (sp_uint64)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
@@ -18267,7 +18278,7 @@ static void sp_4096_mont_setup(const sp_digit* a, sp_digit* rho)
     x &= 0x1fffffffffffffL;
 
     /* rho = -1/m mod b */
-    *rho = ((sp_digit)1 << 53) - x;
+    *rho = (sp_digit)(((sp_uint64)1 << 53) - (sp_digit)x);
 }
 
 /* Multiply a by scalar b into r. (r = a * b)
@@ -18684,21 +18695,21 @@ SP_NOINLINE static void sp_4096_rshift_39(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<32; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (53 - n)) & 0x1fffffffffffffL);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (53 - n)) & 0x1fffffffffffffL);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (53 - n)) & 0x1fffffffffffffL);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (53 - n)) & 0x1fffffffffffffL);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (53 - n)) & 0x1fffffffffffffL);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (53 - n)) & 0x1fffffffffffffL);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (53 - n)) & 0x1fffffffffffffL);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (53 - n)) & 0x1fffffffffffffL);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint64)a[i+1] << (53 - n)) & 0x1fffffffffffffL);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint64)a[i+2] << (53 - n)) & 0x1fffffffffffffL);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint64)a[i+3] << (53 - n)) & 0x1fffffffffffffL);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint64)a[i+4] << (53 - n)) & 0x1fffffffffffffL);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint64)a[i+5] << (53 - n)) & 0x1fffffffffffffL);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint64)a[i+6] << (53 - n)) & 0x1fffffffffffffL);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint64)a[i+7] << (53 - n)) & 0x1fffffffffffffL);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint64)a[i+8] << (53 - n)) & 0x1fffffffffffffL);
     }
-    r[32] = (a[32] >> n) | (sp_digit)((a[33] << (53 - n)) & 0x1fffffffffffffL);
-    r[33] = (a[33] >> n) | (sp_digit)((a[34] << (53 - n)) & 0x1fffffffffffffL);
-    r[34] = (a[34] >> n) | (sp_digit)((a[35] << (53 - n)) & 0x1fffffffffffffL);
-    r[35] = (a[35] >> n) | (sp_digit)((a[36] << (53 - n)) & 0x1fffffffffffffL);
-    r[36] = (a[36] >> n) | (sp_digit)((a[37] << (53 - n)) & 0x1fffffffffffffL);
-    r[37] = (a[37] >> n) | (sp_digit)((a[38] << (53 - n)) & 0x1fffffffffffffL);
+    r[32] = (a[32] >> n) | (sp_digit)(((sp_uint64)a[33] << (53 - n)) & 0x1fffffffffffffL);
+    r[33] = (a[33] >> n) | (sp_digit)(((sp_uint64)a[34] << (53 - n)) & 0x1fffffffffffffL);
+    r[34] = (a[34] >> n) | (sp_digit)(((sp_uint64)a[35] << (53 - n)) & 0x1fffffffffffffL);
+    r[35] = (a[35] >> n) | (sp_digit)(((sp_uint64)a[36] << (53 - n)) & 0x1fffffffffffffL);
+    r[36] = (a[36] >> n) | (sp_digit)(((sp_uint64)a[37] << (53 - n)) & 0x1fffffffffffffL);
+    r[37] = (a[37] >> n) | (sp_digit)(((sp_uint64)a[38] << (53 - n)) & 0x1fffffffffffffL);
     r[38] = a[38] >> n;
 }
 
@@ -18737,7 +18748,7 @@ static WC_INLINE sp_digit sp_4096_div_word_39(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 51; i >= 1; i--) {
         t1 += t1 + (((sp_uint64)t0 >> 52) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint64)t0 << 1;
         t2 = (sp_digit)(((sp_uint64)(dv - t1)) >> 63);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -18929,9 +18940,9 @@ static int sp_4096_mod_exp_39(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 53;
-        c = bits % 53;
-        n = e[i--] << (53 - c);
+        i = (bits - 1) / 53;
+        c = ((bits - 1) % 53) + 1;
+        n = (sp_uint64)e[i--] << (53 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -18943,7 +18954,7 @@ static int sp_4096_mod_exp_39(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 52) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_4096_mont_mul_39(t[y^1], t[0], t[1], m, mp);
 
@@ -19005,9 +19016,9 @@ static int sp_4096_mod_exp_39(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 53;
-        c = bits % 53;
-        n = e[i--] << (53 - c);
+        i = (bits - 1) / 53;
+        c = ((bits - 1) % 53) + 1;
+        n = (sp_uint64)e[i--] << (53 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -19019,7 +19030,7 @@ static int sp_4096_mod_exp_39(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 52) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_4096_mont_mul_39(t[y^1], t[0], t[1], m, mp);
 
@@ -19120,38 +19131,38 @@ static int sp_4096_mod_exp_39(sp_digit* r, const sp_digit* a, const sp_digit* e,
             c = 53;
         }
         if (i < 39) {
-            n = e[i--] << (64 - c);
+            n = (sp_uint64)e[i--] << (64 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 5) {
-            n |= e[i--] << (11 - c);
+            n |= (sp_uint64)e[i--] << (11 - c);
             c += 53;
         }
         y = (int)((n >> 59) & 0x1f);
-        n <<= 5;
+        n = (sp_uint64)n << 5;
         c -= 5;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 78);
         while ((i >= 0) || (c >= 5)) {
             if (c >= 5) {
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c -= 5;
             }
             else if (c == 0) {
-                n = e[i--] << 11;
+                n = (sp_uint64)e[i--] << 11;
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c = 48;
             }
             else {
                 y = (byte)((n >> 59) & 0x1f);
-                n = e[i--] << 11;
+                n = (sp_uint64)e[i--] << 11;
                 c = 5 - c;
                 y |= (byte)((n >> (64 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 53 - c;
             }
 
@@ -19522,20 +19533,20 @@ SP_NOINLINE static void sp_4096_rshift_78(sp_digit* r, const sp_digit* a,
     int i;
 
     for (i=0; i<72; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (53 - n)) & 0x1fffffffffffffL);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (53 - n)) & 0x1fffffffffffffL);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (53 - n)) & 0x1fffffffffffffL);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (53 - n)) & 0x1fffffffffffffL);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (53 - n)) & 0x1fffffffffffffL);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (53 - n)) & 0x1fffffffffffffL);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (53 - n)) & 0x1fffffffffffffL);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (53 - n)) & 0x1fffffffffffffL);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint64)a[i+1] << (53 - n)) & 0x1fffffffffffffL);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint64)a[i+2] << (53 - n)) & 0x1fffffffffffffL);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint64)a[i+3] << (53 - n)) & 0x1fffffffffffffL);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint64)a[i+4] << (53 - n)) & 0x1fffffffffffffL);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint64)a[i+5] << (53 - n)) & 0x1fffffffffffffL);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint64)a[i+6] << (53 - n)) & 0x1fffffffffffffL);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint64)a[i+7] << (53 - n)) & 0x1fffffffffffffL);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint64)a[i+8] << (53 - n)) & 0x1fffffffffffffL);
     }
-    r[72] = (a[72] >> n) | (sp_digit)((a[73] << (53 - n)) & 0x1fffffffffffffL);
-    r[73] = (a[73] >> n) | (sp_digit)((a[74] << (53 - n)) & 0x1fffffffffffffL);
-    r[74] = (a[74] >> n) | (sp_digit)((a[75] << (53 - n)) & 0x1fffffffffffffL);
-    r[75] = (a[75] >> n) | (sp_digit)((a[76] << (53 - n)) & 0x1fffffffffffffL);
-    r[76] = (a[76] >> n) | (sp_digit)((a[77] << (53 - n)) & 0x1fffffffffffffL);
+    r[72] = (a[72] >> n) | (sp_digit)(((sp_uint64)a[73] << (53 - n)) & 0x1fffffffffffffL);
+    r[73] = (a[73] >> n) | (sp_digit)(((sp_uint64)a[74] << (53 - n)) & 0x1fffffffffffffL);
+    r[74] = (a[74] >> n) | (sp_digit)(((sp_uint64)a[75] << (53 - n)) & 0x1fffffffffffffL);
+    r[75] = (a[75] >> n) | (sp_digit)(((sp_uint64)a[76] << (53 - n)) & 0x1fffffffffffffL);
+    r[76] = (a[76] >> n) | (sp_digit)(((sp_uint64)a[77] << (53 - n)) & 0x1fffffffffffffL);
     r[77] = a[77] >> n;
 }
 
@@ -19574,7 +19585,7 @@ static WC_INLINE sp_digit sp_4096_div_word_78(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 51; i >= 1; i--) {
         t1 += t1 + (((sp_uint64)t0 >> 52) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint64)t0 << 1;
         t2 = (sp_digit)(((sp_uint64)(dv - t1)) >> 63);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -19769,9 +19780,9 @@ static int sp_4096_mod_exp_78(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 53;
-        c = bits % 53;
-        n = e[i--] << (53 - c);
+        i = (bits - 1) / 53;
+        c = ((bits - 1) % 53) + 1;
+        n = (sp_uint64)e[i--] << (53 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -19783,7 +19794,7 @@ static int sp_4096_mod_exp_78(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 52) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_4096_mont_mul_78(t[y^1], t[0], t[1], m, mp);
 
@@ -19845,9 +19856,9 @@ static int sp_4096_mod_exp_78(sp_digit* r, const sp_digit* a, const sp_digit* e,
     }
 
     if (err == MP_OKAY) {
-        i = bits / 53;
-        c = bits % 53;
-        n = e[i--] << (53 - c);
+        i = (bits - 1) / 53;
+        c = ((bits - 1) % 53) + 1;
+        n = (sp_uint64)e[i--] << (53 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1) {
@@ -19859,7 +19870,7 @@ static int sp_4096_mod_exp_78(sp_digit* r, const sp_digit* a, const sp_digit* e,
             }
 
             y = (int)((n >> 52) & 1);
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_4096_mont_mul_78(t[y^1], t[0], t[1], m, mp);
 
@@ -19944,38 +19955,38 @@ static int sp_4096_mod_exp_78(sp_digit* r, const sp_digit* a, const sp_digit* e,
             c = 53;
         }
         if (i < 78) {
-            n = e[i--] << (64 - c);
+            n = (sp_uint64)e[i--] << (64 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 4) {
-            n |= e[i--] << (11 - c);
+            n |= (sp_uint64)e[i--] << (11 - c);
             c += 53;
         }
         y = (int)((n >> 60) & 0xf);
-        n <<= 4;
+        n = (sp_uint64)n << 4;
         c -= 4;
         XMEMCPY(rt, t[y], sizeof(sp_digit) * 156);
         while ((i >= 0) || (c >= 4)) {
             if (c >= 4) {
                 y = (byte)((n >> 60) & 0xf);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c -= 4;
             }
             else if (c == 0) {
-                n = e[i--] << 11;
+                n = (sp_uint64)e[i--] << 11;
                 y = (byte)((n >> 60) & 0xf);
-                n <<= 4;
+                n = (sp_uint64)n << 4;
                 c = 49;
             }
             else {
                 y = (byte)((n >> 60) & 0xf);
-                n = e[i--] << 11;
+                n = (sp_uint64)e[i--] << 11;
                 c = 4 - c;
                 y |= (byte)((n >> (64 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 53 - c;
             }
 
@@ -20071,10 +20082,10 @@ static int sp_4096_mod_exp_78_nb(sp_4096_mod_exp_78_ctx* ctx,
     case 4: /* BIT_INIT: index the most-significant exponent limb without
               reading off the end when bits is an exact multiple of the
               limb width. (bits-1) keeps i within the populated range and
-              c = (bits-1) % @bits + 1 keeps the shift in [1, @bits]. */
+              c = (bits-1) % 53 + 1 keeps the shift in [1, 53]. */
         ctx->i = (ctx->bits - 1) / 53;
         ctx->c = ((ctx->bits - 1) % 53) + 1;
-        ctx->n = e[ctx->i--] << (53 - ctx->c);
+        ctx->n = (sp_uint64)e[ctx->i--] << (53 - ctx->c);
         ctx->state = 5;
         break;
     case 5: /* BIT_NEXT: refill on word boundary, peel one exponent bit */
@@ -20087,7 +20098,7 @@ static int sp_4096_mod_exp_78_nb(sp_4096_mod_exp_78_ctx* ctx,
             ctx->c = 53;
         }
         ctx->y = (byte)((ctx->n >> 52) & 1);
-        ctx->n <<= 1;
+        ctx->n = (sp_uint64)ctx->n << 1;
         ctx->state = 6;
         break;
     case 6: /* MUL: t[y^1] = t[0] * t[1] in Montgomery form */
@@ -20658,7 +20669,7 @@ static int sp_4096_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 78; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -20683,7 +20694,7 @@ static int sp_4096_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 78; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 53 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -20968,7 +20979,7 @@ SP_NOINLINE static void sp_4096_lshift_78(sp_digit* r, const sp_digit* a,
     r[2] = (sp_digit)(((s << n) | (t >> (53U - n))) & 0x1fffffffffffffUL);
     s = (sp_int_digit)(a[1]); t = (sp_int_digit)(a[0]);
     r[1] = (sp_digit)(((s << n) | (t >> (53U - n))) & 0x1fffffffffffffUL);
-    r[0] = (sp_digit)((a[0] << n) & 0x1fffffffffffffL);
+    r[0] = (sp_digit)(((sp_uint64)a[0] << n) & 0x1fffffffffffffL);
 }
 
 /* Modular exponentiate 2 to the e mod m. (r = 2^e mod m)
@@ -21014,38 +21025,38 @@ static int sp_4096_mod_exp_2_78(sp_digit* r, const sp_digit* e, int bits, const
             c = 53;
         }
         if (i < 78) {
-            n = e[i--] << (64 - c);
+            n = (sp_uint64)e[i--] << (64 - c);
         }
         else {
             n = 0;
             i--;
         }
         if (c < 5) {
-            n |= e[i--] << (11 - c);
+            n |= (sp_uint64)e[i--] << (11 - c);
             c += 53;
         }
         y = (int)((n >> 59) & 0x1f);
-        n <<= 5;
+        n = (sp_uint64)n << 5;
         c -= 5;
         sp_4096_lshift_78(r, norm, (byte)y);
         while ((i >= 0) || (c >= 5)) {
             if (c >= 5) {
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c -= 5;
             }
             else if (c == 0) {
-                n = e[i--] << 11;
+                n = (sp_uint64)e[i--] << 11;
                 y = (byte)((n >> 59) & 0x1f);
-                n <<= 5;
+                n = (sp_uint64)n << 5;
                 c = 48;
             }
             else {
                 y = (byte)((n >> 59) & 0x1f);
-                n = e[i--] << 11;
+                n = (sp_uint64)e[i--] << 11;
                 c = 5 - c;
                 y |= (byte)((n >> (64 - c)) & ((1 << c) - 1));
-                n <<= c;
+                n = (sp_uint64)n << c;
                 c = 53 - c;
             }
 
@@ -21513,7 +21524,7 @@ static void sp_256_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0xfffffffffffffL;
         s = 52U - s;
         if (j + 1 >= size) {
@@ -21548,7 +21559,7 @@ static void sp_256_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 52) {
             r[j] &= 0xfffffffffffffL;
             if (j + 1 >= size) {
@@ -21614,7 +21625,7 @@ static int sp_256_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 5; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -21639,7 +21650,7 @@ static int sp_256_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 5; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 52 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -21886,9 +21897,9 @@ static void sp_256_mont_reduce_5(sp_digit* a, const sp_digit* m, sp_digit mp)
         /* Fifth word of modulus word */
         t = am; t *= 0x0ffffffff0000L;
 
-        a[i + 1] += (sp_digit)((am << 44) & 0xfffffffffffffL);
+        a[i + 1] += (sp_digit)(((sp_uint64)am << 44) & 0xfffffffffffffL);
         a[i + 2] += am >> 8;
-        a[i + 3] += (sp_digit)((am << 36) & 0xfffffffffffffL);
+        a[i + 3] += (sp_digit)(((sp_uint64)am << 36) & 0xfffffffffffffL);
         a[i + 4] += (am >> 16) + (sp_digit)(t & 0xfffffffffffffL);
         a[i + 5] +=  t >> 52;
 
@@ -21898,17 +21909,17 @@ static void sp_256_mont_reduce_5(sp_digit* a, const sp_digit* m, sp_digit mp)
     /* Fifth word of modulus word */
     t = am; t *= 0x0ffffffff0000L;
 
-    a[4 + 1] += (sp_digit)((am << 44) & 0xfffffffffffffL);
+    a[4 + 1] += (sp_digit)(((sp_uint64)am << 44) & 0xfffffffffffffL);
     a[4 + 2] += am >> 8;
-    a[4 + 3] += (sp_digit)((am << 36) & 0xfffffffffffffL);
+    a[4 + 3] += (sp_digit)(((sp_uint64)am << 36) & 0xfffffffffffffL);
     a[4 + 4] += (am >> 16) + (sp_digit)(t & 0xfffffffffffffL);
     a[4 + 5] +=  t >> 52;
 
-    a[0] = (a[4] >> 48) + (sp_digit)((a[5] << 4) & 0xfffffffffffffL);
-    a[1] = (a[5] >> 48) + (sp_digit)((a[6] << 4) & 0xfffffffffffffL);
-    a[2] = (a[6] >> 48) + (sp_digit)((a[7] << 4) & 0xfffffffffffffL);
-    a[3] = (a[7] >> 48) + (sp_digit)((a[8] << 4) & 0xfffffffffffffL);
-    a[4] = (a[8] >> 48) +  (a[9] << 4);
+    a[0] = (a[4] >> 48) + (sp_digit)(((sp_uint64)a[5] << 4) & 0xfffffffffffffL);
+    a[1] = (a[5] >> 48) + (sp_digit)(((sp_uint64)a[6] << 4) & 0xfffffffffffffL);
+    a[2] = (a[6] >> 48) + (sp_digit)(((sp_uint64)a[7] << 4) & 0xfffffffffffffL);
+    a[3] = (a[7] >> 48) + (sp_digit)(((sp_uint64)a[8] << 4) & 0xfffffffffffffL);
+    a[4] = (a[8] >> 48) +  (sp_digit)((sp_uint64)a[9] << 4);
 
     a[1] += a[0] >> 52; a[0] &= 0xfffffffffffffL;
     a[2] += a[1] >> 52; a[1] &= 0xfffffffffffffL;
@@ -22005,7 +22016,7 @@ static void sp_256_mont_inv_5(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 5);
     for (i=254; i>=0; i--) {
         sp_256_mont_sqr_5(t, t, p256_mod, p256_mp_mod);
-        if (p256_mod_minus_2[i / 64] & ((sp_digit)1 << (i % 64)))
+        if (p256_mod_minus_2[i / 64] & ((sp_uint64)1 << (i % 64)))
             sp_256_mont_mul_5(t, t, a, p256_mod, p256_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 5);
@@ -22221,13 +22232,13 @@ SP_NOINLINE static void sp_256_rshift1_5(sp_digit* r, const sp_digit* a)
     int i;
 
     for (i=0; i<4; i++) {
-        r[i] = (a[i] >> 1) + (sp_digit)((a[i + 1] << 51) & 0xfffffffffffffL);
+        r[i] = (a[i] >> 1) + (sp_digit)(((sp_uint64)a[i + 1] << 51) & 0xfffffffffffffL);
     }
 #else
-    r[0] = (a[0] >> 1) + (sp_digit)((a[1] << 51) & 0xfffffffffffffL);
-    r[1] = (a[1] >> 1) + (sp_digit)((a[2] << 51) & 0xfffffffffffffL);
-    r[2] = (a[2] >> 1) + (sp_digit)((a[3] << 51) & 0xfffffffffffffL);
-    r[3] = (a[3] >> 1) + (sp_digit)((a[4] << 51) & 0xfffffffffffffL);
+    r[0] = (a[0] >> 1) + (sp_digit)(((sp_uint64)a[1] << 51) & 0xfffffffffffffL);
+    r[1] = (a[1] >> 1) + (sp_digit)(((sp_uint64)a[2] << 51) & 0xfffffffffffffL);
+    r[2] = (a[2] >> 1) + (sp_digit)(((sp_uint64)a[3] << 51) & 0xfffffffffffffL);
+    r[3] = (a[3] >> 1) + (sp_digit)(((sp_uint64)a[4] << 51) & 0xfffffffffffffL);
 #endif
     r[4] = a[4] >> 1;
 }
@@ -22785,18 +22796,18 @@ static int sp_256_mod_mul_norm_5(sp_digit* r, const sp_digit* a, const sp_digit*
 
         a32[0] = (sp_digit)(a[0]) & 0xffffffffL;
         a32[1] = (sp_digit)(a[0] >> 32U);
-        a32[1] |= (sp_digit)(a[1] << 20U);
+        a32[1] |= (sp_digit)((sp_uint64)a[1] << 20U);
         a32[1] &= 0xffffffffL;
         a32[2] = (sp_digit)(a[1] >> 12U) & 0xffffffffL;
         a32[3] = (sp_digit)(a[1] >> 44U);
-        a32[3] |= (sp_digit)(a[2] << 8U);
+        a32[3] |= (sp_digit)((sp_uint64)a[2] << 8U);
         a32[3] &= 0xffffffffL;
         a32[4] = (sp_digit)(a[2] >> 24U);
-        a32[4] |= (sp_digit)(a[3] << 28U);
+        a32[4] |= (sp_digit)((sp_uint64)a[3] << 28U);
         a32[4] &= 0xffffffffL;
         a32[5] = (sp_digit)(a[3] >> 4U) & 0xffffffffL;
         a32[6] = (sp_digit)(a[3] >> 36U);
-        a32[6] |= (sp_digit)(a[4] << 16U);
+        a32[6] |= (sp_digit)((sp_uint64)a[4] << 16U);
         a32[6] &= 0xffffffffL;
         a32[7] = (sp_digit)(a[4] >> 16U) & 0xffffffffL;
 
@@ -22910,7 +22921,7 @@ static int sp_256_ecc_mulmod_5(sp_point_256* r, const sp_point_256* g,
     if (err == MP_OKAY) {
         i = 4;
         c = 48;
-        n = k[i--] << (52 - c);
+        n = (sp_uint64)k[i--] << (52 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1)
@@ -22921,7 +22932,7 @@ static int sp_256_ecc_mulmod_5(sp_point_256* r, const sp_point_256* g,
             }
 
             y = (n >> 51) & 1;
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_256_proj_point_add_5(&t[y^1], &t[0], &t[1], tmp);
 
@@ -22980,7 +22991,7 @@ static int sp_256_ecc_mulmod_5_nb(sp_ecc_ctx_t* sp_ctx, sp_point_256* r,
         XMEMSET(ctx->t, 0, sizeof(sp_point_256) * 3);
         ctx->i = 4;
         ctx->c = 48;
-        ctx->n = k[ctx->i--] << (52 - ctx->c);
+        ctx->n = (sp_uint64)k[ctx->i--] << (52 - ctx->c);
 
         /* t[0] = {0, 0, 1} * norm */
         ctx->t[0].infinity = 1;
@@ -23010,7 +23021,7 @@ static int sp_256_ecc_mulmod_5_nb(sp_ecc_ctx_t* sp_ctx, sp_point_256* r,
             ctx->c = 52;
         }
         ctx->y = (ctx->n >> 51) & 1;
-        ctx->n <<= 1;
+        ctx->n = (sp_uint64)ctx->n << 1;
         XMEMSET(&ctx->add_ctx, 0, sizeof(ctx->add_ctx));
         ctx->state = 5;
         break;
@@ -23405,7 +23416,7 @@ static void sp_256_ecc_recode_6_5(const sp_digit* k, ecc_recode_256* v)
         }
         else if (++j < 5) {
             n = k[j];
-            y |= (word8)((n << (52 - o)) & 0x3f);
+            y |= (word8)(((sp_uint64)n << (52 - o)) & 0x3f);
             o -= 46;
             n >>= o;
         }
@@ -25615,7 +25626,7 @@ static void sp_256_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     r[0] = 0;
     for (i = n-1; i >= 0; i--) {
-        r[j] |= (((sp_digit)a[i]) << s);
+        r[j] |= (((sp_uint64)a[i]) << s);
         if (s >= 44U) {
             r[j] &= 0xfffffffffffffL;
             s = 52U - s;
@@ -25836,7 +25847,7 @@ static void sp_256_to_bin_5(sp_digit* r, byte* a)
     for (i=0; i<5 && j>=0; i++) {
         b = 0;
         /* lint allow cast of mismatch sp_digit and int */
-        a[j--] |= (byte)(r[i] << s); /*lint !e9033*/
+        a[j--] |= (byte)((sp_uint64)r[i] << s); /*lint !e9033*/
         b += 8 - s;
         if (j < 0) {
             break;
@@ -25960,23 +25971,23 @@ SP_NOINLINE static void sp_256_rshift_5(sp_digit* r, const sp_digit* a,
 
 #ifdef WOLFSSL_SP_SMALL
     for (i=0; i<4; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (52 - n))) & 0xfffffffffffffL);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint64)a[i + 1] << (52 - n))) & 0xfffffffffffffL);
     }
 #else
     for (i=0; i<0; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (52 - n)) & 0xfffffffffffffL);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (52 - n)) & 0xfffffffffffffL);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (52 - n)) & 0xfffffffffffffL);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (52 - n)) & 0xfffffffffffffL);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (52 - n)) & 0xfffffffffffffL);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (52 - n)) & 0xfffffffffffffL);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (52 - n)) & 0xfffffffffffffL);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (52 - n)) & 0xfffffffffffffL);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint64)a[i+1] << (52 - n)) & 0xfffffffffffffL);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint64)a[i+2] << (52 - n)) & 0xfffffffffffffL);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint64)a[i+3] << (52 - n)) & 0xfffffffffffffL);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint64)a[i+4] << (52 - n)) & 0xfffffffffffffL);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint64)a[i+5] << (52 - n)) & 0xfffffffffffffL);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint64)a[i+6] << (52 - n)) & 0xfffffffffffffL);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint64)a[i+7] << (52 - n)) & 0xfffffffffffffL);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint64)a[i+8] << (52 - n)) & 0xfffffffffffffL);
     }
-    r[0] = (a[0] >> n) | (sp_digit)((a[1] << (52 - n)) & 0xfffffffffffffL);
-    r[1] = (a[1] >> n) | (sp_digit)((a[2] << (52 - n)) & 0xfffffffffffffL);
-    r[2] = (a[2] >> n) | (sp_digit)((a[3] << (52 - n)) & 0xfffffffffffffL);
-    r[3] = (a[3] >> n) | (sp_digit)((a[4] << (52 - n)) & 0xfffffffffffffL);
+    r[0] = (a[0] >> n) | (sp_digit)(((sp_uint64)a[1] << (52 - n)) & 0xfffffffffffffL);
+    r[1] = (a[1] >> n) | (sp_digit)(((sp_uint64)a[2] << (52 - n)) & 0xfffffffffffffL);
+    r[2] = (a[2] >> n) | (sp_digit)(((sp_uint64)a[3] << (52 - n)) & 0xfffffffffffffL);
+    r[3] = (a[3] >> n) | (sp_digit)(((sp_uint64)a[4] << (52 - n)) & 0xfffffffffffffL);
 #endif /* WOLFSSL_SP_SMALL */
     r[4] = a[4] >> n;
 }
@@ -26027,7 +26038,7 @@ SP_NOINLINE static void sp_256_lshift_10(sp_digit* r, const sp_digit* a,
 
     r[10] = a[9] >> (52 - n);
     for (i=9; i>0; i--) {
-        r[i] = (sp_digit)(((a[i] << n) | (a[i-1] >> (52 - n))) & 0xfffffffffffffL);
+        r[i] = (sp_digit)((((sp_uint64)a[i] << n) | (a[i-1] >> (52 - n))) & 0xfffffffffffffL);
     }
 #else
     sp_int_digit s;
@@ -26054,7 +26065,7 @@ SP_NOINLINE static void sp_256_lshift_10(sp_digit* r, const sp_digit* a,
     s = (sp_int_digit)(a[1]); t = (sp_int_digit)(a[0]);
     r[1] = (sp_digit)(((s << n) | (t >> (52U - n))) & 0xfffffffffffffUL);
 #endif /* WOLFSSL_SP_SMALL */
-    r[0] = (sp_digit)((a[0] << n) & 0xfffffffffffffL);
+    r[0] = (sp_digit)(((sp_uint64)a[0] << n) & 0xfffffffffffffL);
 }
 
 /* Divide d in a and put remainder into r (m*d + r = a)
@@ -27960,7 +27971,7 @@ static void sp_384_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0x7fffffffffffffL;
         s = 55U - s;
         if (j + 1 >= size) {
@@ -27995,7 +28006,7 @@ static void sp_384_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 55) {
             r[j] &= 0x7fffffffffffffL;
             if (j + 1 >= size) {
@@ -28061,7 +28072,7 @@ static int sp_384_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 7; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -28086,7 +28097,7 @@ static int sp_384_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 7; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 55 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -28346,31 +28357,31 @@ static void sp_384_mont_reduce_7(sp_digit* a, const sp_digit* m, sp_digit mp)
     (void)mp;
 
     for (i = 0; i < 6; i++) {
-        am = (sp_digit)((a[i] * 0x100000001) & 0x7fffffffffffffL);
-        a[i + 0] += (sp_digit)((am << 32) & 0x7fffffffffffffL);
-        a[i + 1] += (am >> 23) - (sp_digit)((am << 41) & 0x7fffffffffffffL);
-        a[i + 2] += -(am >> 14) - ((sp_digit)(am << 18) & 0x7fffffffffffffL);
+        am = (sp_digit)((((sp_uint64)a[i]) * 0x100000001) & 0x7fffffffffffffL);
+        a[i + 0] += (sp_digit)(((sp_uint64)am << 32) & 0x7fffffffffffffL);
+        a[i + 1] += (am >> 23) - (sp_digit)(((sp_uint64)am << 41) & 0x7fffffffffffffL);
+        a[i + 2] += -(am >> 14) - (sp_digit)(((sp_uint64)am << 18) & 0x7fffffffffffffL);
         a[i + 3] += -(am >> 37);
-        a[i + 6] += ((sp_digit)(am << 54) & 0x7fffffffffffffL);
+        a[i + 6] += (sp_digit)(((sp_uint64)am << 54) & 0x7fffffffffffffL);
         a[i + 7] += am >> 1;
 
         a[i + 1] += a[i] >> 55;
     }
-    am = (sp_digit)((a[6] * 0x100000001) & 0x3fffffffffffff);
-    a[6 + 0] += (sp_digit)((am << 32) & 0x7fffffffffffffL);
-    a[6 + 1] += (am >> 23) - (sp_digit)((am << 41) & 0x7fffffffffffffL);
-    a[6 + 2] += -(am >> 14) - (sp_digit)((am << 18) & 0x7fffffffffffffL);
+    am = (sp_digit)((((sp_uint64)a[6]) * 0x100000001) & 0x3fffffffffffff);
+    a[6 + 0] += (sp_digit)(((sp_uint64)am << 32) & 0x7fffffffffffffL);
+    a[6 + 1] += (am >> 23) - (sp_digit)(((sp_uint64)am << 41) & 0x7fffffffffffffL);
+    a[6 + 2] += -(am >> 14) - (sp_digit)(((sp_uint64)am << 18) & 0x7fffffffffffffL);
     a[6 + 3] += -(am >> 37);
-    a[6 + 6] += (sp_digit)((am << 54) & 0x7fffffffffffffL);
+    a[6 + 6] += (sp_digit)(((sp_uint64)am << 54) & 0x7fffffffffffffL);
     a[6 + 7] += am >> 1;
 
-    a[0] = (a[6] >> 54) + (sp_digit)((a[7] << 1) & 0x7fffffffffffffL);
-    a[1] = (a[7] >> 54) + (sp_digit)((a[8] << 1) & 0x7fffffffffffffL);
-    a[2] = (a[8] >> 54) + (sp_digit)((a[9] << 1) & 0x7fffffffffffffL);
-    a[3] = (a[9] >> 54) + (sp_digit)((a[10] << 1) & 0x7fffffffffffffL);
-    a[4] = (a[10] >> 54) + (sp_digit)((a[11] << 1) & 0x7fffffffffffffL);
-    a[5] = (a[11] >> 54) + (sp_digit)((a[12] << 1) & 0x7fffffffffffffL);
-    a[6] = (a[12] >> 54) +  (a[13] << 1);
+    a[0] = (a[6] >> 54) + (sp_digit)(((sp_uint64)a[7] << 1) & 0x7fffffffffffffL);
+    a[1] = (a[7] >> 54) + (sp_digit)(((sp_uint64)a[8] << 1) & 0x7fffffffffffffL);
+    a[2] = (a[8] >> 54) + (sp_digit)(((sp_uint64)a[9] << 1) & 0x7fffffffffffffL);
+    a[3] = (a[9] >> 54) + (sp_digit)(((sp_uint64)a[10] << 1) & 0x7fffffffffffffL);
+    a[4] = (a[10] >> 54) + (sp_digit)(((sp_uint64)a[11] << 1) & 0x7fffffffffffffL);
+    a[5] = (a[11] >> 54) + (sp_digit)(((sp_uint64)a[12] << 1) & 0x7fffffffffffffL);
+    a[6] = (a[12] >> 54) +  (sp_digit)((sp_uint64)a[13] << 1);
 
     a[1] += a[0] >> 55; a[0] &= 0x7fffffffffffffL;
     a[2] += a[1] >> 55; a[1] &= 0x7fffffffffffffL;
@@ -28473,7 +28484,7 @@ static void sp_384_mont_inv_7(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 7);
     for (i=382; i>=0; i--) {
         sp_384_mont_sqr_7(t, t, p384_mod, p384_mp_mod);
-        if (p384_mod_minus_2[i / 64] & ((sp_digit)1 << (i % 64)))
+        if (p384_mod_minus_2[i / 64] & ((sp_uint64)1 << (i % 64)))
             sp_384_mont_mul_7(t, t, a, p384_mod, p384_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 7);
@@ -28707,15 +28718,15 @@ SP_NOINLINE static void sp_384_rshift1_7(sp_digit* r, const sp_digit* a)
     int i;
 
     for (i=0; i<6; i++) {
-        r[i] = (a[i] >> 1) + (sp_digit)((a[i + 1] << 54) & 0x7fffffffffffffL);
+        r[i] = (a[i] >> 1) + (sp_digit)(((sp_uint64)a[i + 1] << 54) & 0x7fffffffffffffL);
     }
 #else
-    r[0] = (a[0] >> 1) + (sp_digit)((a[1] << 54) & 0x7fffffffffffffL);
-    r[1] = (a[1] >> 1) + (sp_digit)((a[2] << 54) & 0x7fffffffffffffL);
-    r[2] = (a[2] >> 1) + (sp_digit)((a[3] << 54) & 0x7fffffffffffffL);
-    r[3] = (a[3] >> 1) + (sp_digit)((a[4] << 54) & 0x7fffffffffffffL);
-    r[4] = (a[4] >> 1) + (sp_digit)((a[5] << 54) & 0x7fffffffffffffL);
-    r[5] = (a[5] >> 1) + (sp_digit)((a[6] << 54) & 0x7fffffffffffffL);
+    r[0] = (a[0] >> 1) + (sp_digit)(((sp_uint64)a[1] << 54) & 0x7fffffffffffffL);
+    r[1] = (a[1] >> 1) + (sp_digit)(((sp_uint64)a[2] << 54) & 0x7fffffffffffffL);
+    r[2] = (a[2] >> 1) + (sp_digit)(((sp_uint64)a[3] << 54) & 0x7fffffffffffffL);
+    r[3] = (a[3] >> 1) + (sp_digit)(((sp_uint64)a[4] << 54) & 0x7fffffffffffffL);
+    r[4] = (a[4] >> 1) + (sp_digit)(((sp_uint64)a[5] << 54) & 0x7fffffffffffffL);
+    r[5] = (a[5] >> 1) + (sp_digit)(((sp_uint64)a[6] << 54) & 0x7fffffffffffffL);
 #endif
     r[6] = a[6] >> 1;
 }
@@ -29274,26 +29285,26 @@ static int sp_384_mod_mul_norm_7(sp_digit* r, const sp_digit* a, const sp_digit*
 
         a32[0] = (sp_digit)(a[0]) & 0xffffffffL;
         a32[1] = (sp_digit)(a[0] >> 32U);
-        a32[1] |= (sp_digit)(a[1] << 23U);
+        a32[1] |= (sp_digit)((sp_uint64)a[1] << 23U);
         a32[1] &= 0xffffffffL;
         a32[2] = (sp_digit)(a[1] >> 9U) & 0xffffffffL;
         a32[3] = (sp_digit)(a[1] >> 41U);
-        a32[3] |= (sp_digit)(a[2] << 14U);
+        a32[3] |= (sp_digit)((sp_uint64)a[2] << 14U);
         a32[3] &= 0xffffffffL;
         a32[4] = (sp_digit)(a[2] >> 18U) & 0xffffffffL;
         a32[5] = (sp_digit)(a[2] >> 50U);
-        a32[5] |= (sp_digit)(a[3] << 5U);
+        a32[5] |= (sp_digit)((sp_uint64)a[3] << 5U);
         a32[5] &= 0xffffffffL;
         a32[6] = (sp_digit)(a[3] >> 27U);
-        a32[6] |= (sp_digit)(a[4] << 28U);
+        a32[6] |= (sp_digit)((sp_uint64)a[4] << 28U);
         a32[6] &= 0xffffffffL;
         a32[7] = (sp_digit)(a[4] >> 4U) & 0xffffffffL;
         a32[8] = (sp_digit)(a[4] >> 36U);
-        a32[8] |= (sp_digit)(a[5] << 19U);
+        a32[8] |= (sp_digit)((sp_uint64)a[5] << 19U);
         a32[8] &= 0xffffffffL;
         a32[9] = (sp_digit)(a[5] >> 13U) & 0xffffffffL;
         a32[10] = (sp_digit)(a[5] >> 45U);
-        a32[10] |= (sp_digit)(a[6] << 10U);
+        a32[10] |= (sp_digit)((sp_uint64)a[6] << 10U);
         a32[10] &= 0xffffffffL;
         a32[11] = (sp_digit)(a[6] >> 22U) & 0xffffffffL;
 
@@ -29431,7 +29442,7 @@ static int sp_384_ecc_mulmod_7(sp_point_384* r, const sp_point_384* g,
     if (err == MP_OKAY) {
         i = 6;
         c = 54;
-        n = k[i--] << (55 - c);
+        n = (sp_uint64)k[i--] << (55 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1)
@@ -29442,7 +29453,7 @@ static int sp_384_ecc_mulmod_7(sp_point_384* r, const sp_point_384* g,
             }
 
             y = (n >> 54) & 1;
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_384_proj_point_add_7(&t[y^1], &t[0], &t[1], tmp);
 
@@ -29501,7 +29512,7 @@ static int sp_384_ecc_mulmod_7_nb(sp_ecc_ctx_t* sp_ctx, sp_point_384* r,
         XMEMSET(ctx->t, 0, sizeof(sp_point_384) * 3);
         ctx->i = 6;
         ctx->c = 54;
-        ctx->n = k[ctx->i--] << (55 - ctx->c);
+        ctx->n = (sp_uint64)k[ctx->i--] << (55 - ctx->c);
 
         /* t[0] = {0, 0, 1} * norm */
         ctx->t[0].infinity = 1;
@@ -29531,7 +29542,7 @@ static int sp_384_ecc_mulmod_7_nb(sp_ecc_ctx_t* sp_ctx, sp_point_384* r,
             ctx->c = 55;
         }
         ctx->y = (ctx->n >> 54) & 1;
-        ctx->n <<= 1;
+        ctx->n = (sp_uint64)ctx->n << 1;
         XMEMSET(&ctx->add_ctx, 0, sizeof(ctx->add_ctx));
         ctx->state = 5;
         break;
@@ -29930,7 +29941,7 @@ static void sp_384_ecc_recode_6_7(const sp_digit* k, ecc_recode_384* v)
         }
         else if (++j < 7) {
             n = k[j];
-            y |= (word8)((n << (55 - o)) & 0x3f);
+            y |= (word8)(((sp_uint64)n << (55 - o)) & 0x3f);
             o -= 49;
             n >>= o;
         }
@@ -32670,7 +32681,7 @@ static void sp_384_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     r[0] = 0;
     for (i = n-1; i >= 0; i--) {
-        r[j] |= (((sp_digit)a[i]) << s);
+        r[j] |= (((sp_uint64)a[i]) << s);
         if (s >= 47U) {
             r[j] &= 0x7fffffffffffffL;
             s = 55U - s;
@@ -32891,7 +32902,7 @@ static void sp_384_to_bin_7(sp_digit* r, byte* a)
     for (i=0; i<7 && j>=0; i++) {
         b = 0;
         /* lint allow cast of mismatch sp_digit and int */
-        a[j--] |= (byte)(r[i] << s); /*lint !e9033*/
+        a[j--] |= (byte)((sp_uint64)r[i] << s); /*lint !e9033*/
         b += 8 - s;
         if (j < 0) {
             break;
@@ -33015,25 +33026,25 @@ SP_NOINLINE static void sp_384_rshift_7(sp_digit* r, const sp_digit* a,
 
 #ifdef WOLFSSL_SP_SMALL
     for (i=0; i<6; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (55 - n))) & 0x7fffffffffffffL);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint64)a[i + 1] << (55 - n))) & 0x7fffffffffffffL);
     }
 #else
     for (i=0; i<0; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (55 - n)) & 0x7fffffffffffffL);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (55 - n)) & 0x7fffffffffffffL);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (55 - n)) & 0x7fffffffffffffL);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (55 - n)) & 0x7fffffffffffffL);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (55 - n)) & 0x7fffffffffffffL);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (55 - n)) & 0x7fffffffffffffL);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (55 - n)) & 0x7fffffffffffffL);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (55 - n)) & 0x7fffffffffffffL);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint64)a[i+1] << (55 - n)) & 0x7fffffffffffffL);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint64)a[i+2] << (55 - n)) & 0x7fffffffffffffL);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint64)a[i+3] << (55 - n)) & 0x7fffffffffffffL);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint64)a[i+4] << (55 - n)) & 0x7fffffffffffffL);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint64)a[i+5] << (55 - n)) & 0x7fffffffffffffL);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint64)a[i+6] << (55 - n)) & 0x7fffffffffffffL);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint64)a[i+7] << (55 - n)) & 0x7fffffffffffffL);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint64)a[i+8] << (55 - n)) & 0x7fffffffffffffL);
     }
-    r[0] = (a[0] >> n) | (sp_digit)((a[1] << (55 - n)) & 0x7fffffffffffffL);
-    r[1] = (a[1] >> n) | (sp_digit)((a[2] << (55 - n)) & 0x7fffffffffffffL);
-    r[2] = (a[2] >> n) | (sp_digit)((a[3] << (55 - n)) & 0x7fffffffffffffL);
-    r[3] = (a[3] >> n) | (sp_digit)((a[4] << (55 - n)) & 0x7fffffffffffffL);
-    r[4] = (a[4] >> n) | (sp_digit)((a[5] << (55 - n)) & 0x7fffffffffffffL);
-    r[5] = (a[5] >> n) | (sp_digit)((a[6] << (55 - n)) & 0x7fffffffffffffL);
+    r[0] = (a[0] >> n) | (sp_digit)(((sp_uint64)a[1] << (55 - n)) & 0x7fffffffffffffL);
+    r[1] = (a[1] >> n) | (sp_digit)(((sp_uint64)a[2] << (55 - n)) & 0x7fffffffffffffL);
+    r[2] = (a[2] >> n) | (sp_digit)(((sp_uint64)a[3] << (55 - n)) & 0x7fffffffffffffL);
+    r[3] = (a[3] >> n) | (sp_digit)(((sp_uint64)a[4] << (55 - n)) & 0x7fffffffffffffL);
+    r[4] = (a[4] >> n) | (sp_digit)(((sp_uint64)a[5] << (55 - n)) & 0x7fffffffffffffL);
+    r[5] = (a[5] >> n) | (sp_digit)(((sp_uint64)a[6] << (55 - n)) & 0x7fffffffffffffL);
 #endif /* WOLFSSL_SP_SMALL */
     r[6] = a[6] >> n;
 }
@@ -33088,7 +33099,7 @@ SP_NOINLINE static void sp_384_lshift_14(sp_digit* r, const sp_digit* a,
 
     r[14] = a[13] >> (55 - n);
     for (i=13; i>0; i--) {
-        r[i] = (sp_digit)(((a[i] << n) | (a[i-1] >> (55 - n))) & 0x7fffffffffffffL);
+        r[i] = (sp_digit)((((sp_uint64)a[i] << n) | (a[i-1] >> (55 - n))) & 0x7fffffffffffffL);
     }
 #else
     sp_int_digit s;
@@ -33123,7 +33134,7 @@ SP_NOINLINE static void sp_384_lshift_14(sp_digit* r, const sp_digit* a,
     s = (sp_int_digit)(a[1]); t = (sp_int_digit)(a[0]);
     r[1] = (sp_digit)(((s << n) | (t >> (55U - n))) & 0x7fffffffffffffUL);
 #endif /* WOLFSSL_SP_SMALL */
-    r[0] = (sp_digit)((a[0] << n) & 0x7fffffffffffffL);
+    r[0] = (sp_digit)(((sp_uint64)a[0] << n) & 0x7fffffffffffffL);
 }
 
 /* Divide d in a and put remainder into r (m*d + r = a)
@@ -35101,7 +35112,7 @@ static void sp_521_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0x3ffffffffffffffL;
         s = 58U - s;
         if (j + 1 >= size) {
@@ -35136,7 +35147,7 @@ static void sp_521_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 58) {
             r[j] &= 0x3ffffffffffffffL;
             if (j + 1 >= size) {
@@ -35202,7 +35213,7 @@ static int sp_521_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 9; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -35227,7 +35238,7 @@ static int sp_521_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 9; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 58 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -35308,10 +35319,10 @@ static void sp_521_mont_reduce_9(sp_digit* a, const sp_digit* m, sp_digit mp)
     (void)mp;
 
     for (i = 0; i < 8; i++) {
-        a[i] += (sp_digit)(((a[8 + i] >> 57) + (a[8 + i + 1] << 1)) & 0x3ffffffffffffffL);
+        a[i] += (sp_digit)(((a[8 + i] >> 57) + ((sp_uint64)a[8 + i + 1] << 1)) & 0x3ffffffffffffffL);
     }
     a[8] &= 0x1ffffffffffffff;
-    a[8] += (sp_digit)(((a[16] >> 57) + (a[17] << 1)) & 0x3ffffffffffffffL);
+    a[8] += (sp_digit)(((a[16] >> 57) + ((sp_uint64)a[17] << 1)) & 0x3ffffffffffffffL);
 
     sp_521_norm_9(a);
 
@@ -35577,7 +35588,7 @@ static void sp_521_mont_inv_9(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 9);
     for (i=519; i>=0; i--) {
         sp_521_mont_sqr_9(t, t, p521_mod, p521_mp_mod);
-        if (p521_mod_minus_2[i / 64] & ((sp_digit)1 << (i % 64)))
+        if (p521_mod_minus_2[i / 64] & ((sp_uint64)1 << (i % 64)))
             sp_521_mont_mul_9(t, t, a, p521_mod, p521_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 9);
@@ -35809,17 +35820,17 @@ SP_NOINLINE static void sp_521_rshift1_9(sp_digit* r, const sp_digit* a)
     int i;
 
     for (i=0; i<8; i++) {
-        r[i] = (a[i] >> 1) + (sp_digit)((a[i + 1] << 57) & 0x3ffffffffffffffL);
+        r[i] = (a[i] >> 1) + (sp_digit)(((sp_uint64)a[i + 1] << 57) & 0x3ffffffffffffffL);
     }
 #else
-    r[0] = (a[0] >> 1) + (sp_digit)((a[1] << 57) & 0x3ffffffffffffffL);
-    r[1] = (a[1] >> 1) + (sp_digit)((a[2] << 57) & 0x3ffffffffffffffL);
-    r[2] = (a[2] >> 1) + (sp_digit)((a[3] << 57) & 0x3ffffffffffffffL);
-    r[3] = (a[3] >> 1) + (sp_digit)((a[4] << 57) & 0x3ffffffffffffffL);
-    r[4] = (a[4] >> 1) + (sp_digit)((a[5] << 57) & 0x3ffffffffffffffL);
-    r[5] = (a[5] >> 1) + (sp_digit)((a[6] << 57) & 0x3ffffffffffffffL);
-    r[6] = (a[6] >> 1) + (sp_digit)((a[7] << 57) & 0x3ffffffffffffffL);
-    r[7] = (a[7] >> 1) + (sp_digit)((a[8] << 57) & 0x3ffffffffffffffL);
+    r[0] = (a[0] >> 1) + (sp_digit)(((sp_uint64)a[1] << 57) & 0x3ffffffffffffffL);
+    r[1] = (a[1] >> 1) + (sp_digit)(((sp_uint64)a[2] << 57) & 0x3ffffffffffffffL);
+    r[2] = (a[2] >> 1) + (sp_digit)(((sp_uint64)a[3] << 57) & 0x3ffffffffffffffL);
+    r[3] = (a[3] >> 1) + (sp_digit)(((sp_uint64)a[4] << 57) & 0x3ffffffffffffffL);
+    r[4] = (a[4] >> 1) + (sp_digit)(((sp_uint64)a[5] << 57) & 0x3ffffffffffffffL);
+    r[5] = (a[5] >> 1) + (sp_digit)(((sp_uint64)a[6] << 57) & 0x3ffffffffffffffL);
+    r[6] = (a[6] >> 1) + (sp_digit)(((sp_uint64)a[7] << 57) & 0x3ffffffffffffffL);
+    r[7] = (a[7] >> 1) + (sp_digit)(((sp_uint64)a[8] << 57) & 0x3ffffffffffffffL);
 #endif
     r[8] = a[8] >> 1;
 }
@@ -36425,7 +36436,7 @@ static int sp_521_ecc_mulmod_9(sp_point_521* r, const sp_point_521* g,
     if (err == MP_OKAY) {
         i = 8;
         c = 57;
-        n = k[i--] << (58 - c);
+        n = (sp_uint64)k[i--] << (58 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1)
@@ -36436,7 +36447,7 @@ static int sp_521_ecc_mulmod_9(sp_point_521* r, const sp_point_521* g,
             }
 
             y = (n >> 57) & 1;
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_521_proj_point_add_9(&t[y^1], &t[0], &t[1], tmp);
 
@@ -36495,7 +36506,7 @@ static int sp_521_ecc_mulmod_9_nb(sp_ecc_ctx_t* sp_ctx, sp_point_521* r,
         XMEMSET(ctx->t, 0, sizeof(sp_point_521) * 3);
         ctx->i = 8;
         ctx->c = 57;
-        ctx->n = k[ctx->i--] << (58 - ctx->c);
+        ctx->n = (sp_uint64)k[ctx->i--] << (58 - ctx->c);
 
         /* t[0] = {0, 0, 1} * norm */
         ctx->t[0].infinity = 1;
@@ -36525,7 +36536,7 @@ static int sp_521_ecc_mulmod_9_nb(sp_ecc_ctx_t* sp_ctx, sp_point_521* r,
             ctx->c = 58;
         }
         ctx->y = (ctx->n >> 57) & 1;
-        ctx->n <<= 1;
+        ctx->n = (sp_uint64)ctx->n << 1;
         XMEMSET(&ctx->add_ctx, 0, sizeof(ctx->add_ctx));
         ctx->state = 5;
         break;
@@ -36928,7 +36939,7 @@ static void sp_521_ecc_recode_6_9(const sp_digit* k, ecc_recode_521* v)
         }
         else if (++j < 9) {
             n = k[j];
-            y |= (word8)((n << (58 - o)) & 0x3f);
+            y |= (word8)(((sp_uint64)n << (58 - o)) & 0x3f);
             o -= 52;
             n >>= o;
         }
@@ -39688,7 +39699,7 @@ static void sp_521_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     r[0] = 0;
     for (i = n-1; i >= 0; i--) {
-        r[j] |= (((sp_digit)a[i]) << s);
+        r[j] |= (((sp_uint64)a[i]) << s);
         if (s >= 50U) {
             r[j] &= 0x3ffffffffffffffL;
             s = 58U - s;
@@ -39910,7 +39921,7 @@ static void sp_521_to_bin_9(sp_digit* r, byte* a)
     for (i=0; i<9 && j>=0; i++) {
         b = 0;
         /* lint allow cast of mismatch sp_digit and int */
-        a[j--] |= (byte)(r[i] << s); /*lint !e9033*/
+        a[j--] |= (byte)((sp_uint64)r[i] << s); /*lint !e9033*/
         b += 8 - s;
         if (j < 0) {
             break;
@@ -40032,18 +40043,18 @@ SP_NOINLINE static void sp_521_rshift_9(sp_digit* r, const sp_digit* a,
 
 #ifdef WOLFSSL_SP_SMALL
     for (i=0; i<8; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (58 - n))) & 0x3ffffffffffffffL);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint64)a[i + 1] << (58 - n))) & 0x3ffffffffffffffL);
     }
 #else
     for (i=0; i<8; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (58 - n)) & 0x3ffffffffffffffL);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (58 - n)) & 0x3ffffffffffffffL);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (58 - n)) & 0x3ffffffffffffffL);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (58 - n)) & 0x3ffffffffffffffL);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (58 - n)) & 0x3ffffffffffffffL);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (58 - n)) & 0x3ffffffffffffffL);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (58 - n)) & 0x3ffffffffffffffL);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (58 - n)) & 0x3ffffffffffffffL);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint64)a[i+1] << (58 - n)) & 0x3ffffffffffffffL);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint64)a[i+2] << (58 - n)) & 0x3ffffffffffffffL);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint64)a[i+3] << (58 - n)) & 0x3ffffffffffffffL);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint64)a[i+4] << (58 - n)) & 0x3ffffffffffffffL);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint64)a[i+5] << (58 - n)) & 0x3ffffffffffffffL);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint64)a[i+6] << (58 - n)) & 0x3ffffffffffffffL);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint64)a[i+7] << (58 - n)) & 0x3ffffffffffffffL);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint64)a[i+8] << (58 - n)) & 0x3ffffffffffffffL);
     }
 #endif /* WOLFSSL_SP_SMALL */
     r[8] = a[8] >> n;
@@ -40105,7 +40116,7 @@ SP_NOINLINE static void sp_521_lshift_18(sp_digit* r, const sp_digit* a,
 
     r[18] = a[17] >> (58 - n);
     for (i=17; i>0; i--) {
-        r[i] = (sp_digit)(((a[i] << n) | (a[i-1] >> (58 - n))) & 0x3ffffffffffffffL);
+        r[i] = (sp_digit)((((sp_uint64)a[i] << n) | (a[i-1] >> (58 - n))) & 0x3ffffffffffffffL);
     }
 #else
     sp_int_digit s;
@@ -40148,7 +40159,7 @@ SP_NOINLINE static void sp_521_lshift_18(sp_digit* r, const sp_digit* a,
     s = (sp_int_digit)(a[1]); t = (sp_int_digit)(a[0]);
     r[1] = (sp_digit)(((s << n) | (t >> (58U - n))) & 0x3ffffffffffffffUL);
 #endif /* WOLFSSL_SP_SMALL */
-    r[0] = (sp_digit)((a[0] << n) & 0x3ffffffffffffffL);
+    r[0] = (sp_digit)(((sp_uint64)a[0] << n) & 0x3ffffffffffffffL);
 }
 
 /* Divide d in a and put remainder into r (m*d + r = a)
@@ -41565,7 +41576,7 @@ static int sp_521_mont_sqrt_9(sp_digit* y)
             XMEMCPY(t, y, sizeof(sp_digit) * 9);
             for (i=518; i>=0; i--) {
                 sp_521_mont_sqr_9(t, t, p521_mod, p521_mp_mod);
-                if (p521_sqrt_power[i / 64] & ((sp_digit)1 << (i % 64)))
+                if (p521_sqrt_power[i / 64] & ((sp_uint64)1 << (i % 64)))
                     sp_521_mont_mul_9(t, t, y, p521_mod, p521_mp_mod);
             }
             XMEMCPY(y, t, sizeof(sp_digit) * 9);
@@ -42351,20 +42362,20 @@ SP_NOINLINE static void sp_1024_rshift_18(sp_digit* r, const sp_digit* a,
 
 #ifdef WOLFSSL_SP_SMALL
     for (i=0; i<17; i++) {
-        r[i] = (sp_digit)(((a[i] >> n) | (a[i + 1] << (57 - n))) & 0x1ffffffffffffffL);
+        r[i] = (sp_digit)(((a[i] >> n) | ((sp_uint64)a[i + 1] << (57 - n))) & 0x1ffffffffffffffL);
     }
 #else
     for (i=0; i<16; i += 8) {
-        r[i+0] = (a[i+0] >> n) | (sp_digit)((a[i+1] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+1] = (a[i+1] >> n) | (sp_digit)((a[i+2] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+2] = (a[i+2] >> n) | (sp_digit)((a[i+3] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+3] = (a[i+3] >> n) | (sp_digit)((a[i+4] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+4] = (a[i+4] >> n) | (sp_digit)((a[i+5] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+5] = (a[i+5] >> n) | (sp_digit)((a[i+6] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+6] = (a[i+6] >> n) | (sp_digit)((a[i+7] << (57 - n)) & 0x1ffffffffffffffL);
-        r[i+7] = (a[i+7] >> n) | (sp_digit)((a[i+8] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+0] = (a[i+0] >> n) | (sp_digit)(((sp_uint64)a[i+1] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+1] = (a[i+1] >> n) | (sp_digit)(((sp_uint64)a[i+2] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+2] = (a[i+2] >> n) | (sp_digit)(((sp_uint64)a[i+3] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+3] = (a[i+3] >> n) | (sp_digit)(((sp_uint64)a[i+4] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+4] = (a[i+4] >> n) | (sp_digit)(((sp_uint64)a[i+5] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+5] = (a[i+5] >> n) | (sp_digit)(((sp_uint64)a[i+6] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+6] = (a[i+6] >> n) | (sp_digit)(((sp_uint64)a[i+7] << (57 - n)) & 0x1ffffffffffffffL);
+        r[i+7] = (a[i+7] >> n) | (sp_digit)(((sp_uint64)a[i+8] << (57 - n)) & 0x1ffffffffffffffL);
     }
-    r[16] = (a[16] >> n) | (sp_digit)((a[17] << (57 - n)) & 0x1ffffffffffffffL);
+    r[16] = (a[16] >> n) | (sp_digit)(((sp_uint64)a[17] << (57 - n)) & 0x1ffffffffffffffL);
 #endif /* WOLFSSL_SP_SMALL */
     r[17] = a[17] >> n;
 }
@@ -42404,7 +42415,7 @@ static WC_INLINE sp_digit sp_1024_div_word_18(sp_digit d1, sp_digit d0,
     t1 -= dv & (0 - r);
     for (i = 55; i >= 1; i--) {
         t1 += t1 + (((sp_uint64)t0 >> 56) & 1);
-        t0 <<= 1;
+        t0 = (sp_uint64)t0 << 1;
         t2 = (sp_digit)(((sp_uint64)(dv - t1)) >> 63);
         r += r + t2;
         t1 -= dv & (0 - t2);
@@ -42647,7 +42658,7 @@ static void sp_1024_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0x1ffffffffffffffL;
         s = 57U - s;
         if (j + 1 >= size) {
@@ -42682,7 +42693,7 @@ static void sp_1024_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 57) {
             r[j] &= 0x1ffffffffffffffL;
             if (j + 1 >= size) {
@@ -42748,7 +42759,7 @@ static int sp_1024_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 18; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -42773,7 +42784,7 @@ static int sp_1024_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 18; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 57 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -43009,11 +43020,11 @@ static void sp_1024_mont_reduce_18(sp_digit* a, const sp_digit* m, sp_digit mp)
 
     if (mp != 1) {
         for (i=0; i<17; i++) {
-            mu = (sp_digit)((a[i] * mp) & 0x1ffffffffffffffL);
+            mu = (sp_digit)(((sp_uint64)a[i] * (sp_uint64)mp) & 0x1ffffffffffffffL);
             sp_1024_mul_add_18(a+i, m, mu);
             a[i+1] += a[i] >> 57;
         }
-        mu = (sp_digit)((a[i] * mp) & 0x7fffffffffffffL);
+        mu = (sp_digit)(((sp_uint64)a[i] * (sp_uint64)mp) & 0x7fffffffffffffL);
         sp_1024_mul_add_18(a+i, m, mu);
         a[i+1] += a[i] >> 57;
         a[i] &= 0x1ffffffffffffffL;
@@ -43249,26 +43260,26 @@ SP_NOINLINE static void sp_1024_rshift1_18(sp_digit* r, const sp_digit* a)
     int i;
 
     for (i=0; i<17; i++) {
-        r[i] = (a[i] >> 1) + (sp_digit)((a[i + 1] << 56) & 0x1ffffffffffffffL);
+        r[i] = (a[i] >> 1) + (sp_digit)(((sp_uint64)a[i + 1] << 56) & 0x1ffffffffffffffL);
     }
 #else
-    r[0] = (a[0] >> 1) + (sp_digit)((a[1] << 56) & 0x1ffffffffffffffL);
-    r[1] = (a[1] >> 1) + (sp_digit)((a[2] << 56) & 0x1ffffffffffffffL);
-    r[2] = (a[2] >> 1) + (sp_digit)((a[3] << 56) & 0x1ffffffffffffffL);
-    r[3] = (a[3] >> 1) + (sp_digit)((a[4] << 56) & 0x1ffffffffffffffL);
-    r[4] = (a[4] >> 1) + (sp_digit)((a[5] << 56) & 0x1ffffffffffffffL);
-    r[5] = (a[5] >> 1) + (sp_digit)((a[6] << 56) & 0x1ffffffffffffffL);
-    r[6] = (a[6] >> 1) + (sp_digit)((a[7] << 56) & 0x1ffffffffffffffL);
-    r[7] = (a[7] >> 1) + (sp_digit)((a[8] << 56) & 0x1ffffffffffffffL);
-    r[8] = (a[8] >> 1) + (sp_digit)((a[9] << 56) & 0x1ffffffffffffffL);
-    r[9] = (a[9] >> 1) + (sp_digit)((a[10] << 56) & 0x1ffffffffffffffL);
-    r[10] = (a[10] >> 1) + (sp_digit)((a[11] << 56) & 0x1ffffffffffffffL);
-    r[11] = (a[11] >> 1) + (sp_digit)((a[12] << 56) & 0x1ffffffffffffffL);
-    r[12] = (a[12] >> 1) + (sp_digit)((a[13] << 56) & 0x1ffffffffffffffL);
-    r[13] = (a[13] >> 1) + (sp_digit)((a[14] << 56) & 0x1ffffffffffffffL);
-    r[14] = (a[14] >> 1) + (sp_digit)((a[15] << 56) & 0x1ffffffffffffffL);
-    r[15] = (a[15] >> 1) + (sp_digit)((a[16] << 56) & 0x1ffffffffffffffL);
-    r[16] = (a[16] >> 1) + (sp_digit)((a[17] << 56) & 0x1ffffffffffffffL);
+    r[0] = (a[0] >> 1) + (sp_digit)(((sp_uint64)a[1] << 56) & 0x1ffffffffffffffL);
+    r[1] = (a[1] >> 1) + (sp_digit)(((sp_uint64)a[2] << 56) & 0x1ffffffffffffffL);
+    r[2] = (a[2] >> 1) + (sp_digit)(((sp_uint64)a[3] << 56) & 0x1ffffffffffffffL);
+    r[3] = (a[3] >> 1) + (sp_digit)(((sp_uint64)a[4] << 56) & 0x1ffffffffffffffL);
+    r[4] = (a[4] >> 1) + (sp_digit)(((sp_uint64)a[5] << 56) & 0x1ffffffffffffffL);
+    r[5] = (a[5] >> 1) + (sp_digit)(((sp_uint64)a[6] << 56) & 0x1ffffffffffffffL);
+    r[6] = (a[6] >> 1) + (sp_digit)(((sp_uint64)a[7] << 56) & 0x1ffffffffffffffL);
+    r[7] = (a[7] >> 1) + (sp_digit)(((sp_uint64)a[8] << 56) & 0x1ffffffffffffffL);
+    r[8] = (a[8] >> 1) + (sp_digit)(((sp_uint64)a[9] << 56) & 0x1ffffffffffffffL);
+    r[9] = (a[9] >> 1) + (sp_digit)(((sp_uint64)a[10] << 56) & 0x1ffffffffffffffL);
+    r[10] = (a[10] >> 1) + (sp_digit)(((sp_uint64)a[11] << 56) & 0x1ffffffffffffffL);
+    r[11] = (a[11] >> 1) + (sp_digit)(((sp_uint64)a[12] << 56) & 0x1ffffffffffffffL);
+    r[12] = (a[12] >> 1) + (sp_digit)(((sp_uint64)a[13] << 56) & 0x1ffffffffffffffL);
+    r[13] = (a[13] >> 1) + (sp_digit)(((sp_uint64)a[14] << 56) & 0x1ffffffffffffffL);
+    r[14] = (a[14] >> 1) + (sp_digit)(((sp_uint64)a[15] << 56) & 0x1ffffffffffffffL);
+    r[15] = (a[15] >> 1) + (sp_digit)(((sp_uint64)a[16] << 56) & 0x1ffffffffffffffL);
+    r[16] = (a[16] >> 1) + (sp_digit)(((sp_uint64)a[17] << 56) & 0x1ffffffffffffffL);
 #endif
     r[17] = a[17] >> 1;
 }
@@ -43860,7 +43871,7 @@ static int sp_1024_ecc_mulmod_18(sp_point_1024* r, const sp_point_1024* g,
     if (err == MP_OKAY) {
         i = 17;
         c = 55;
-        n = k[i--] << (57 - c);
+        n = (sp_uint64)k[i--] << (57 - c);
         for (; ; c--) {
             if (c == 0) {
                 if (i == -1)
@@ -43871,7 +43882,7 @@ static int sp_1024_ecc_mulmod_18(sp_point_1024* r, const sp_point_1024* g,
             }
 
             y = (n >> 56) & 1;
-            n <<= 1;
+            n = (sp_uint64)n << 1;
 
             sp_1024_proj_point_add_18(&t[y^1], &t[0], &t[1], tmp);
 
@@ -43930,7 +43941,7 @@ static int sp_1024_ecc_mulmod_18_nb(sp_ecc_ctx_t* sp_ctx, sp_point_1024* r,
         XMEMSET(ctx->t, 0, sizeof(sp_point_1024) * 3);
         ctx->i = 17;
         ctx->c = 55;
-        ctx->n = k[ctx->i--] << (57 - ctx->c);
+        ctx->n = (sp_uint64)k[ctx->i--] << (57 - ctx->c);
 
         /* t[0] = {0, 0, 1} * norm */
         ctx->t[0].infinity = 1;
@@ -43960,7 +43971,7 @@ static int sp_1024_ecc_mulmod_18_nb(sp_ecc_ctx_t* sp_ctx, sp_point_1024* r,
             ctx->c = 57;
         }
         ctx->y = (ctx->n >> 56) & 1;
-        ctx->n <<= 1;
+        ctx->n = (sp_uint64)ctx->n << 1;
         XMEMSET(&ctx->add_ctx, 0, sizeof(ctx->add_ctx));
         ctx->state = 5;
         break;
@@ -44389,7 +44400,7 @@ static void sp_1024_ecc_recode_7_18(const sp_digit* k, ecc_recode_1024* v)
         }
         else if (++j < 18) {
             n = k[j];
-            y |= (word8)((n << (57 - o)) & 0x7f);
+            y |= (word8)(((sp_uint64)n << (57 - o)) & 0x7f);
             o -= 50;
             n >>= o;
         }
@@ -51732,7 +51743,7 @@ static void sp_1024_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     r[0] = 0;
     for (i = n-1; i >= 0; i--) {
-        r[j] |= (((sp_digit)a[i]) << s);
+        r[j] |= (((sp_uint64)a[i]) << s);
         if (s >= 49U) {
             r[j] &= 0x1ffffffffffffffL;
             s = 57U - s;
diff --git a/wolfcrypt/src/sp_cortexm.c b/wolfcrypt/src/sp_cortexm.c
index 9a334f4d4a..f9d22092d9 100644
--- a/wolfcrypt/src/sp_cortexm.c
+++ b/wolfcrypt/src/sp_cortexm.c
@@ -47,6 +47,17 @@
 
 #include <wolfssl/wolfcrypt/sp.h>
 
+#if defined(WOLFSSL_USE_SAVE_VECTOR_REGISTERS) && !defined(WOLFSSL_SP_ASM) && \
+        !defined(DEBUG_VECTOR_REGISTER_ACCESS)
+    /* force off unneeded vector register save/restore. */
+    #undef SAVE_VECTOR_REGISTERS
+    #define SAVE_VECTOR_REGISTERS(fail_clause) SAVE_NO_VECTOR_REGISTERS(fail_clause)
+    #undef SAVE_VECTOR_REGISTERS2
+    #define SAVE_VECTOR_REGISTERS2() SAVE_NO_VECTOR_REGISTERS2()
+    #undef RESTORE_VECTOR_REGISTERS
+    #define RESTORE_VECTOR_REGISTERS() RESTORE_NO_VECTOR_REGISTERS()
+#endif
+
 #ifdef __IAR_SYSTEMS_ICC__
 #define __asm__        asm
 #define __volatile__   volatile
@@ -141,10 +152,10 @@ static void sp_2048_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
@@ -199,7 +210,7 @@ static void sp_2048_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -234,7 +245,7 @@ static void sp_2048_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -3136,17 +3147,17 @@ WC_OMIT_FRAME_POINTER static void sp_2048_sqr_32(sp_digit* r, const sp_digit* a)
  */
 static void sp_2048_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint32 x;
+    sp_uint32 b;
 
-    b = a[0];
+    b = (sp_uint32)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**32 */
 
     /* rho = -1/m mod b */
-    *rho = (sp_digit)0 - x;
+    *rho = (sp_digit)((sp_int32)0 - (sp_int32)x);
 }
 
 #ifdef WOLFSSL_SP_SMALL
@@ -5600,10 +5611,10 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -5612,14 +5623,14 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 32);
         for (; i>=0 || c>=4; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 28);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 28;
             }
             else if (c < 4) {
@@ -5627,12 +5638,12 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 4 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
 
@@ -5753,10 +5764,10 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -5765,14 +5776,14 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 32);
         for (; i>=0 || c>=5; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 27);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 27;
             }
             else if (c < 5) {
@@ -5780,12 +5791,12 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 5 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
 
@@ -8772,10 +8783,10 @@ static int sp_2048_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -8784,14 +8795,14 @@ static int sp_2048_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 64);
         for (; i>=0 || c>=3; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 29);
-                n <<= 3;
+                n = (sp_uint32)n << 3;
                 c = 29;
             }
             else if (c < 3) {
@@ -8799,12 +8810,12 @@ static int sp_2048_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 3 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 29) & 0x7);
-                n <<= 3;
+                n = (sp_uint32)n << 3;
                 c -= 3;
             }
 
@@ -8908,10 +8919,10 @@ static int sp_2048_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -8920,14 +8931,14 @@ static int sp_2048_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 64);
         for (; i>=0 || c>=4; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 28);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 28;
             }
             else if (c < 4) {
@@ -8935,12 +8946,12 @@ static int sp_2048_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 4 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
 
@@ -9013,7 +9024,7 @@ int sp_RsaPublic_2048(const byte* in, word32 inLen, const mp_int* em,
 #else
         e[0] = em->dp[0];
         if (em->used > 1) {
-            e[0] |= ((sp_digit)em->dp[1]) << DIGIT_BIT;
+            e[0] |= ((sp_uint32)em->dp[1]) << DIGIT_BIT;
         }
 #endif
         if (e[0] == 0) {
@@ -9505,7 +9516,7 @@ static int sp_2048_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 64; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -9530,7 +9541,7 @@ static int sp_2048_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 64; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -10064,10 +10075,10 @@ static int sp_2048_mod_exp_2_64(sp_digit* r, const sp_digit* e, int bits,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -10076,14 +10087,14 @@ static int sp_2048_mod_exp_2_64(sp_digit* r, const sp_digit* e, int bits,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         sp_2048_lshift_64(r, norm, y);
         for (; i>=0 || c>=5; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 27);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 27;
             }
             else if (c < 5) {
@@ -10091,12 +10102,12 @@ static int sp_2048_mod_exp_2_64(sp_digit* r, const sp_digit* e, int bits,
                 n = e[i--];
                 c = 5 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
 
@@ -10264,10 +10275,10 @@ static void sp_3072_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
@@ -10322,7 +10333,7 @@ static void sp_3072_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -10357,7 +10368,7 @@ static void sp_3072_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -14361,17 +14372,17 @@ WC_OMIT_FRAME_POINTER static void sp_3072_sqr_48(sp_digit* r, const sp_digit* a)
  */
 static void sp_3072_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint32 x;
+    sp_uint32 b;
 
-    b = a[0];
+    b = (sp_uint32)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**32 */
 
     /* rho = -1/m mod b */
-    *rho = (sp_digit)0 - x;
+    *rho = (sp_digit)((sp_int32)0 - (sp_int32)x);
 }
 
 #ifdef WOLFSSL_SP_SMALL
@@ -17505,10 +17516,10 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -17517,14 +17528,14 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 48);
         for (; i>=0 || c>=4; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 28);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 28;
             }
             else if (c < 4) {
@@ -17532,12 +17543,12 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 4 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
 
@@ -17658,10 +17669,10 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -17670,14 +17681,14 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 48);
         for (; i>=0 || c>=5; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 27);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 27;
             }
             else if (c < 5) {
@@ -17685,12 +17696,12 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 5 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
 
@@ -21613,10 +21624,10 @@ static int sp_3072_mod_exp_96(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -21625,14 +21636,14 @@ static int sp_3072_mod_exp_96(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 96);
         for (; i>=0 || c>=3; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 29);
-                n <<= 3;
+                n = (sp_uint32)n << 3;
                 c = 29;
             }
             else if (c < 3) {
@@ -21640,12 +21651,12 @@ static int sp_3072_mod_exp_96(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 3 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 29) & 0x7);
-                n <<= 3;
+                n = (sp_uint32)n << 3;
                 c -= 3;
             }
 
@@ -21749,10 +21760,10 @@ static int sp_3072_mod_exp_96(sp_digit* r, const sp_digit* a, const sp_digit* e,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -21761,14 +21772,14 @@ static int sp_3072_mod_exp_96(sp_digit* r, const sp_digit* a, const sp_digit* e,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 96);
         for (; i>=0 || c>=4; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 28);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 28;
             }
             else if (c < 4) {
@@ -21776,12 +21787,12 @@ static int sp_3072_mod_exp_96(sp_digit* r, const sp_digit* a, const sp_digit* e,
                 n = e[i--];
                 c = 4 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
 
@@ -21854,7 +21865,7 @@ int sp_RsaPublic_3072(const byte* in, word32 inLen, const mp_int* em,
 #else
         e[0] = em->dp[0];
         if (em->used > 1) {
-            e[0] |= ((sp_digit)em->dp[1]) << DIGIT_BIT;
+            e[0] |= ((sp_uint32)em->dp[1]) << DIGIT_BIT;
         }
 #endif
         if (e[0] == 0) {
@@ -22402,7 +22413,7 @@ static int sp_3072_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 96; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -22427,7 +22438,7 @@ static int sp_3072_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 96; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -23153,10 +23164,10 @@ static int sp_3072_mod_exp_2_96(sp_digit* r, const sp_digit* e, int bits,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -23165,14 +23176,14 @@ static int sp_3072_mod_exp_2_96(sp_digit* r, const sp_digit* e, int bits,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         sp_3072_lshift_96(r, norm, y);
         for (; i>=0 || c>=5; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 27);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 27;
             }
             else if (c < 5) {
@@ -23180,12 +23191,12 @@ static int sp_3072_mod_exp_2_96(sp_digit* r, const sp_digit* e, int bits,
                 n = e[i--];
                 c = 5 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
 
@@ -23353,10 +23364,10 @@ static void sp_4096_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
@@ -23411,7 +23422,7 @@ static void sp_4096_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -23446,7 +23457,7 @@ static void sp_4096_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -24488,17 +24499,17 @@ WC_OMIT_FRAME_POINTER static void sp_4096_sqr_128(sp_digit* r,
  */
 static void sp_4096_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint32 x;
+    sp_uint32 b;
 
-    b = a[0];
+    b = (sp_uint32)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**32 */
 
     /* rho = -1/m mod b */
-    *rho = (sp_digit)0 - x;
+    *rho = (sp_digit)((sp_int32)0 - (sp_int32)x);
 }
 
 #ifdef WOLFSSL_SP_SMALL
@@ -30072,10 +30083,10 @@ static int sp_4096_mod_exp_128(sp_digit* r, const sp_digit* a, const sp_digit* e
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -30084,14 +30095,14 @@ static int sp_4096_mod_exp_128(sp_digit* r, const sp_digit* a, const sp_digit* e
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 128);
         for (; i>=0 || c>=3; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 29);
-                n <<= 3;
+                n = (sp_uint32)n << 3;
                 c = 29;
             }
             else if (c < 3) {
@@ -30099,12 +30110,12 @@ static int sp_4096_mod_exp_128(sp_digit* r, const sp_digit* a, const sp_digit* e
                 n = e[i--];
                 c = 3 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 29) & 0x7);
-                n <<= 3;
+                n = (sp_uint32)n << 3;
                 c -= 3;
             }
 
@@ -30208,10 +30219,10 @@ static int sp_4096_mod_exp_128(sp_digit* r, const sp_digit* a, const sp_digit* e
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -30220,14 +30231,14 @@ static int sp_4096_mod_exp_128(sp_digit* r, const sp_digit* a, const sp_digit* e
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         XMEMCPY(r, t[y], sizeof(sp_digit) * 128);
         for (; i>=0 || c>=4; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 28);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c = 28;
             }
             else if (c < 4) {
@@ -30235,12 +30246,12 @@ static int sp_4096_mod_exp_128(sp_digit* r, const sp_digit* a, const sp_digit* e
                 n = e[i--];
                 c = 4 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 28) & 0xf);
-                n <<= 4;
+                n = (sp_uint32)n << 4;
                 c -= 4;
             }
 
@@ -30313,7 +30324,7 @@ int sp_RsaPublic_4096(const byte* in, word32 inLen, const mp_int* em,
 #else
         e[0] = em->dp[0];
         if (em->used > 1) {
-            e[0] |= ((sp_digit)em->dp[1]) << DIGIT_BIT;
+            e[0] |= ((sp_uint32)em->dp[1]) << DIGIT_BIT;
         }
 #endif
         if (e[0] == 0) {
@@ -30917,7 +30928,7 @@ static int sp_4096_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 128; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -30942,7 +30953,7 @@ static int sp_4096_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 128; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -31860,10 +31871,10 @@ static int sp_4096_mod_exp_2_128(sp_digit* r, const sp_digit* e, int bits,
         if (c < 0) {
             /* Number of bits in top word is less than number needed. */
             c = -c;
-            y = (byte)(n << c);
+            y = (byte)((sp_uint32)n << c);
             n = e[i--];
             y |= (byte)(n >> (32 - c));
-            n <<= c;
+            n = (sp_uint32)n << c;
             c = 32 - c;
         }
         else if (c == 0) {
@@ -31872,14 +31883,14 @@ static int sp_4096_mod_exp_2_128(sp_digit* r, const sp_digit* e, int bits,
         }
         else {
             y = (byte)(n >> c);
-            n <<= 32 - c;
+            n = (sp_uint32)n << (32 - c);
         }
         sp_4096_lshift_128(r, norm, y);
         for (; i>=0 || c>=5; ) {
             if (c == 0) {
                 n = e[i--];
                 y = (byte)(n >> 27);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c = 27;
             }
             else if (c < 5) {
@@ -31887,12 +31898,12 @@ static int sp_4096_mod_exp_2_128(sp_digit* r, const sp_digit* e, int bits,
                 n = e[i--];
                 c = 5 - c;
                 y |= (byte)(n >> (32 - c));
-                n <<= c;
+                n = (sp_uint32)n << c;
                 c = 32 - c;
             }
             else {
                 y = (byte)((n >> 27) & 0x1f);
-                n <<= 5;
+                n = (sp_uint32)n << 5;
                 c -= 5;
             }
 
@@ -33632,7 +33643,7 @@ static void sp_256_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -33667,7 +33678,7 @@ static void sp_256_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -33733,7 +33744,7 @@ static int sp_256_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 8; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -33758,7 +33769,7 @@ static int sp_256_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 8; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -35269,7 +35280,7 @@ static void sp_256_mont_inv_8(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 8);
     for (i=254; i>=0; i--) {
         sp_256_mont_sqr_8(t, t, p256_mod, p256_mp_mod);
-        if (p256_mod_minus_2[i / 32] & ((sp_digit)1 << (i % 32)))
+        if (p256_mod_minus_2[i / 32] & ((sp_uint32)1 << (i % 32)))
             sp_256_mont_mul_8(t, t, a, p256_mod, p256_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 8);
@@ -37352,7 +37363,7 @@ static int sp_256_ecc_mulmod_fast_8(sp_point_256* r, const sp_point_256* g, cons
         t[15].infinity = 0;
 
         i = 6;
-        n = k[i+1] << 0;
+        n = (sp_uint32)k[i+1] << 0;
         c = 28;
         y = (int)(n >> 28);
     #ifndef WC_NO_CACHE_RESISTANT
@@ -37365,14 +37376,14 @@ static int sp_256_ecc_mulmod_fast_8(sp_point_256* r, const sp_point_256* g, cons
         {
             XMEMCPY(rt, &t[y], sizeof(sp_point_256));
         }
-        n <<= 4;
+        n = (sp_uint32)n << (4);
         for (; i>=0 || c>=4; ) {
             if (c < 4) {
                 n |= k[i--];
                 c += 32;
             }
             y = (n >> 28) & 0xf;
-            n <<= 4;
+            n = (sp_uint32)n << 4;
             c -= 4;
 
             sp_256_proj_point_dbl_8(rt, rt, tmp);
@@ -40038,10 +40049,10 @@ static void sp_256_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
@@ -45025,7 +45036,7 @@ static void sp_384_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -45060,7 +45071,7 @@ static void sp_384_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -45126,7 +45137,7 @@ static int sp_384_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 12; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -45151,7 +45162,7 @@ static int sp_384_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 12; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -45703,7 +45714,7 @@ static void sp_384_mont_inv_12(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 12);
     for (i=382; i>=0; i--) {
         sp_384_mont_sqr_12(t, t, p384_mod, p384_mp_mod);
-        if (p384_mod_minus_2[i / 32] & ((sp_digit)1 << (i % 32)))
+        if (p384_mod_minus_2[i / 32] & ((sp_uint32)1 << (i % 32)))
             sp_384_mont_mul_12(t, t, a, p384_mod, p384_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 12);
@@ -47444,7 +47455,7 @@ static int sp_384_ecc_mulmod_fast_12(sp_point_384* r, const sp_point_384* g, con
         t[15].infinity = 0;
 
         i = 10;
-        n = k[i+1] << 0;
+        n = (sp_uint32)k[i+1] << 0;
         c = 28;
         y = (int)(n >> 28);
     #ifndef WC_NO_CACHE_RESISTANT
@@ -47457,14 +47468,14 @@ static int sp_384_ecc_mulmod_fast_12(sp_point_384* r, const sp_point_384* g, con
         {
             XMEMCPY(rt, &t[y], sizeof(sp_point_384));
         }
-        n <<= 4;
+        n = (sp_uint32)n << (4);
         for (; i>=0 || c>=4; ) {
             if (c < 4) {
                 n |= k[i--];
                 c += 32;
             }
             y = (n >> 28) & 0xf;
-            n <<= 4;
+            n = (sp_uint32)n << 4;
             c -= 4;
 
             sp_384_proj_point_dbl_12(rt, rt, tmp);
@@ -50168,10 +50179,10 @@ static void sp_384_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
@@ -56725,7 +56736,7 @@ static void sp_521_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -56760,7 +56771,7 @@ static void sp_521_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -56826,7 +56837,7 @@ static int sp_521_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 17; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -56851,7 +56862,7 @@ static int sp_521_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 17; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -57804,7 +57815,7 @@ static void sp_521_mont_inv_17(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 17);
     for (i=519; i>=0; i--) {
         sp_521_mont_sqr_17(t, t, p521_mod, p521_mp_mod);
-        if (p521_mod_minus_2[i / 32] & ((sp_digit)1 << (i % 32)))
+        if (p521_mod_minus_2[i / 32] & ((sp_uint32)1 << (i % 32)))
             sp_521_mont_mul_17(t, t, a, p521_mod, p521_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 17);
@@ -58677,7 +58688,7 @@ static void sp_521_mont_div2_17(sp_digit* r, const sp_digit* a, const sp_digit*
 
     (void)m;
 
-    sp_521_rshift1_17(r, r);
+    sp_521_rshift1_17(r, a);
     r[16] |= o << 8;
 }
 
@@ -59407,7 +59418,7 @@ static int sp_521_ecc_mulmod_fast_17(sp_point_521* r, const sp_point_521* g, con
         t[15].infinity = 0;
 
         i = 15;
-        n = k[i+1] << 0;
+        n = (sp_uint32)k[i+1] << 0;
         c = 5;
         y = (int)(n >> 5);
     #ifndef WC_NO_CACHE_RESISTANT
@@ -59420,15 +59431,15 @@ static int sp_521_ecc_mulmod_fast_17(sp_point_521* r, const sp_point_521* g, con
         {
             XMEMCPY(rt, &t[y], sizeof(sp_point_521));
         }
-        n <<= 27;
+        n = (sp_uint32)n << (27);
         for (; i>=0 || c>=4; ) {
             if (c < 4) {
-                n = (k[i+1] << 31) | (k[i] >> 1);
+                n = ((sp_uint32)k[i+1] << 31) | (k[i] >> 1);
                 i--;
                 c += 32;
             }
             y = (n >> 28) & 0xf;
-            n <<= 4;
+            n = (sp_uint32)n << 4;
             c -= 4;
 
             sp_521_proj_point_dbl_17(rt, rt, tmp);
@@ -62738,10 +62749,10 @@ static void sp_521_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
@@ -64063,8 +64074,6 @@ static WC_INLINE int sp_521_div_17(const sp_digit* a, const sp_digit* d,
     sp_digit r1;
     int i;
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)m;
     div = (d[16] << 23) | (d[15] >> 9);
     XMEMCPY(t1, a, sizeof(*t1) * 2 * 17);
@@ -66139,7 +66148,7 @@ static int sp_521_mont_sqrt_17(sp_digit* y)
             XMEMCPY(t, y, sizeof(sp_digit) * 17);
             for (i=518; i>=0; i--) {
                 sp_521_mont_sqr_17(t, t, p521_mod, p521_mp_mod);
-                if (p521_sqrt_power[i / 32] & ((sp_digit)1 << (i % 32)))
+                if (p521_sqrt_power[i / 32] & ((sp_uint32)1 << (i % 32)))
                     sp_521_mont_mul_17(t, t, y, p521_mod, p521_mp_mod);
             }
             XMEMCPY(y, t, sizeof(sp_digit) * 17);
@@ -71319,7 +71328,7 @@ static void sp_1024_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint32)a->dp[i] << s);
         r[j] &= 0xffffffff;
         s = 32U - s;
         if (j + 1 >= size) {
@@ -71354,7 +71363,7 @@ static void sp_1024_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint32)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 32) {
             r[j] &= 0xffffffff;
             if (j + 1 >= size) {
@@ -71420,7 +71429,7 @@ static int sp_1024_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 32; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint32)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -71445,7 +71454,7 @@ static int sp_1024_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 32; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint32)a[i]) << s;
             if (s + 32 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -74032,18 +74041,18 @@ static int sp_1024_ecc_mulmod_fast_32(sp_point_1024* r, const sp_point_1024* g,
         t[15].infinity = 0;
 
         i = 30;
-        n = k[i+1] << 0;
+        n = (sp_uint32)k[i+1] << 0;
         c = 28;
         y = (int)(n >> 28);
         XMEMCPY(rt, &t[y], sizeof(sp_point_1024));
-        n <<= 4;
+        n = (sp_uint32)n << (4);
         for (; i>=0 || c>=4; ) {
             if (c < 4) {
                 n |= k[i--];
                 c += 32;
             }
             y = (n >> 28) & 0xf;
-            n <<= 4;
+            n = (sp_uint32)n << 4;
             c -= 4;
 
             sp_1024_proj_point_dbl_32(rt, rt, tmp);
@@ -82148,10 +82157,10 @@ static void sp_1024_from_bin(sp_digit* r, int size, const byte* a, int n)
 
     j = 0;
     for (i = n - 1; i >= 3; i -= 4) {
-        r[j]  = ((sp_digit)a[i - 0] <<  0) |
-                ((sp_digit)a[i - 1] <<  8) |
-                ((sp_digit)a[i - 2] << 16) |
-                ((sp_digit)a[i - 3] << 24);
+        r[j]  = ((sp_uint32)a[i - 0] <<  0) |
+                ((sp_uint32)a[i - 1] <<  8) |
+                ((sp_uint32)a[i - 2] << 16) |
+                ((sp_uint32)a[i - 3] << 24);
         j++;
     }
 
diff --git a/wolfcrypt/src/sp_int.c b/wolfcrypt/src/sp_int.c
index 71ad6c25be..ff5a33e2e6 100644
--- a/wolfcrypt/src/sp_int.c
+++ b/wolfcrypt/src/sp_int.c
@@ -125,8 +125,6 @@ This library provides single precision (SP) integer math functions.
  * WOLFSSL_NO_ASM:              Disable all assembly implementations
  * WOLFSSL_KEIL:                Keil compiler in use, affects inline assembly
  *      syntax
- * WOLFSSL_USE_SAVE_VECTOR_REGISTERS: Save/restore vector registers around
- *      SP ASM calls
  * WOLFSSL_SP_INT_LARGE_COMBA:  Enable large Comba multiplication and
  *      squaring
  * WOLFSSL_SP_INT_SQR_VOLATILE: Declare squaring intermediate variables as
@@ -198,15 +196,6 @@ This library provides single precision (SP) integer math functions.
     PRAGMA_GCC("GCC diagnostic ignored \"-Warray-bounds\"")
 #endif
 
-#if defined(WOLFSSL_USE_SAVE_VECTOR_REGISTERS) && !defined(WOLFSSL_SP_ASM)
-    /* force off unneeded vector register save/restore. */
-    #undef SAVE_VECTOR_REGISTERS
-    #define SAVE_VECTOR_REGISTERS(fail_clause) \
-        SAVE_NO_VECTOR_REGISTERS(fail_clause)
-    #undef RESTORE_VECTOR_REGISTERS
-    #define RESTORE_VECTOR_REGISTERS() RESTORE_NO_VECTOR_REGISTERS()
-#endif
-
 /* DECL_SP_INT: Declare one variable of type 'sp_int'. */
 #if (defined(WOLFSSL_SMALL_STACK) || defined(SP_ALLOC)) && \
     !defined(WOLFSSL_SP_NO_MALLOC)
@@ -5487,6 +5476,22 @@ static void _sp_copy_2_ct(const sp_int* a1, const sp_int* a2, sp_int* r1,
     sp_int* r2, int y, unsigned int used)
 {
     unsigned int i;
+#ifdef WC_NO_GLOBAL_OBJECT_POINTERS
+    static const wc_ptr_t wc_off_on_addr[2] =
+    {
+    #if defined(WC_64BIT_CPU)
+        W64LIT(0x0000000000000000),
+        W64LIT(0xffffffffffffffff)
+    #elif defined(WC_16BIT_CPU)
+        0x0000U,
+        0xffffU
+    #else
+        /* 32 bit */
+        0x00000000U,
+        0xffffffffU
+    #endif
+    };
+#endif
 
     /* Copy data - constant time. */
     for (i = 0; i < used; i++) {
@@ -14329,11 +14334,9 @@ int sp_exptmod(const sp_int* b, const sp_int* e, const sp_int* m, sp_int* r)
     if ((b == NULL) || (e == NULL) || (m == NULL) || (r == NULL)) {
         err = MP_VAL;
     }
-    SAVE_VECTOR_REGISTERS(err = _svr_ret;);
     if (err == MP_OKAY) {
         err = sp_exptmod_ex(b, e, (int)e->used, m, r);
     }
-    RESTORE_VECTOR_REGISTERS();
     return err;
 }
 #endif
@@ -19554,8 +19557,6 @@ int sp_prime_is_prime(const sp_int* a, int trials, int* result)
         haveRes = 1;
     }
 
-    SAVE_VECTOR_REGISTERS(err = _svr_ret;);
-
     /* Check against known small primes when a has 1 digit. */
     if ((err == MP_OKAY) && (!haveRes) && (a->used == 1) &&
             (a->dp[0] <= sp_primes[SP_PRIME_SIZE - 1])) {
@@ -19572,8 +19573,6 @@ int sp_prime_is_prime(const sp_int* a, int trials, int* result)
         err = _sp_prime_trials(a, trials, result);
     }
 
-    RESTORE_VECTOR_REGISTERS();
-
     return err;
 }
 
@@ -19714,8 +19713,6 @@ int sp_prime_is_prime_ex(const sp_int* a, int trials, int* result, WC_RNG* rng)
         haveRes = 1;
     }
 
-    SAVE_VECTOR_REGISTERS(err = _svr_ret;);
-
     /* Check against known small primes when a has 1 digit. */
     if ((err == MP_OKAY) && (!haveRes) && (a->used == 1) &&
             (a->dp[0] <= (sp_int_digit)sp_primes[SP_PRIME_SIZE - 1])) {
@@ -19740,8 +19737,6 @@ int sp_prime_is_prime_ex(const sp_int* a, int trials, int* result, WC_RNG* rng)
         *result = ret;
     }
 
-    RESTORE_VECTOR_REGISTERS();
-
     return err;
 }
 #endif /* WOLFSSL_SP_PRIME_GEN */
@@ -19783,8 +19778,6 @@ static WC_INLINE int _sp_gcd(const sp_int* a, const sp_int* b, sp_int* r)
     unsigned int used = (a->used >= b->used) ? a->used + 1U : b->used + 1U;
     DECL_SP_INT_ARRAY(d, used, 3);
 
-    SAVE_VECTOR_REGISTERS(err = _svr_ret;);
-
     ALLOC_SP_INT_ARRAY(d, used, 3, err, NULL);
     if (err == MP_OKAY) {
         u = d[0];
@@ -19849,8 +19842,6 @@ static WC_INLINE int _sp_gcd(const sp_int* a, const sp_int* b, sp_int* r)
 
     FREE_SP_INT_ARRAY(d, NULL);
 
-    RESTORE_VECTOR_REGISTERS();
-
     return err;
 }
 
@@ -19955,8 +19946,6 @@ static int _sp_lcm(const sp_int* a, const sp_int* b, sp_int* r)
         _sp_init_size(t[0], used);
         _sp_init_size(t[1], used);
 
-        SAVE_VECTOR_REGISTERS(err = _svr_ret;);
-
         if (err == MP_OKAY) {
             /* 1. t0 = gcd(a, b) */
             err = sp_gcd(a, b, t[0]);
@@ -19985,8 +19974,6 @@ static int _sp_lcm(const sp_int* a, const sp_int* b, sp_int* r)
                 }
             }
         }
-
-        RESTORE_VECTOR_REGISTERS();
     }
 
     FREE_SP_INT_ARRAY(t, NULL);
diff --git a/wolfcrypt/src/sp_x86_64.c b/wolfcrypt/src/sp_x86_64.c
index 1f7cd9e396..7ce9b9ab6a 100644
--- a/wolfcrypt/src/sp_x86_64.c
+++ b/wolfcrypt/src/sp_x86_64.c
@@ -47,6 +47,17 @@
 
 #include <wolfssl/wolfcrypt/sp.h>
 
+#if defined(WOLFSSL_USE_SAVE_VECTOR_REGISTERS) && !defined(WOLFSSL_SP_ASM) && \
+        !defined(DEBUG_VECTOR_REGISTER_ACCESS)
+    /* force off unneeded vector register save/restore. */
+    #undef SAVE_VECTOR_REGISTERS
+    #define SAVE_VECTOR_REGISTERS(fail_clause) SAVE_NO_VECTOR_REGISTERS(fail_clause)
+    #undef SAVE_VECTOR_REGISTERS2
+    #define SAVE_VECTOR_REGISTERS2() SAVE_NO_VECTOR_REGISTERS2()
+    #undef RESTORE_VECTOR_REGISTERS
+    #define RESTORE_VECTOR_REGISTERS() RESTORE_NO_VECTOR_REGISTERS()
+#endif
+
 #ifdef __IAR_SYSTEMS_ICC__
 #define __asm__        asm
 #define __volatile__   volatile
@@ -188,7 +199,7 @@ static void sp_2048_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0xffffffffffffffffl;
         s = 64U - s;
         if (j + 1 >= size) {
@@ -223,7 +234,7 @@ static void sp_2048_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 64) {
             r[j] &= 0xffffffffffffffffl;
             if (j + 1 >= size) {
@@ -397,10 +408,10 @@ extern sp_digit sp_2048_sub_in_place_16(sp_digit* a, const sp_digit* b);
  */
 static void sp_2048_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint64 x;
+    sp_uint64 b;
 
-    b = a[0];
+    b = (sp_uint64)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
@@ -408,7 +419,7 @@ static void sp_2048_mont_setup(const sp_digit* a, sp_digit* rho)
     x *= 2 - b * x;               /* here x*a==1 mod 2**64 */
 
     /* rho = -1/m mod b */
-    *rho = (sp_digit)0 - x;
+    *rho = (sp_digit)((sp_int64)0 - (sp_int64)x);
 }
 
 #ifdef __cplusplus
@@ -429,8 +440,6 @@ static void sp_2048_mont_norm_16(sp_digit* r, const sp_digit* m)
 {
     XMEMSET(r, 0, sizeof(sp_digit) * 16);
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     /* r = 2^n mod m */
     sp_2048_sub_in_place_16(r, m);
 }
@@ -520,7 +529,6 @@ extern sp_digit div_2048_word_asm_16(sp_digit d1, sp_digit d0, sp_digit div);
 static WC_INLINE sp_digit div_2048_word_16(sp_digit d1, sp_digit d0,
         sp_digit div)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
 #if _MSC_VER >= 1920
     return _udiv128(d1, d0, div, NULL);
 #else
@@ -539,7 +547,6 @@ static WC_INLINE sp_digit div_2048_word_16(sp_digit d1, sp_digit d0,
         sp_digit div)
 {
     register sp_digit r asm("rax");
-    ASSERT_SAVED_VECTOR_REGISTERS();
     __asm__ __volatile__ (
         "divq %3"
         : "=a" (r)
@@ -607,8 +614,6 @@ static WC_INLINE int sp_2048_div_16(const sp_digit* a, const sp_digit* d, sp_dig
     word32 cpuid_flags = cpuid_get_flags();
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)m;
 
     div = d[15];
@@ -661,18 +666,38 @@ static WC_INLINE int sp_2048_div_16(const sp_digit* a, const sp_digit* d, sp_dig
 static WC_INLINE int sp_2048_mod_16(sp_digit* r, const sp_digit* a,
         const sp_digit* m)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
     return sp_2048_div_16(a, m, NULL, r);
 }
 
-#ifdef __cplusplus
-extern "C" {
-#endif
-extern void sp_2048_get_from_table_16(sp_digit* r, sp_digit** table, int idx);
-#ifdef __cplusplus
-}
-#endif
+#ifndef WC_NO_CACHE_RESISTANT
+static void sp_2048_get_from_table_16(sp_digit* r,
+        sp_digit** table, int idx)
+{
+    int e, j;
+    sp_digit mask;
+    sp_digit diff;
 
+    for (j = 0; j < 16; j++) {
+        r[j] = 0;
+    }
+
+    for (e = 0; e < 32; e++) {
+        /* Constant-time: mask = (e == idx) ? ~0 : 0
+         * diff = e ^ idx is 0 iff equal.
+         * (diff | -diff) has its sign bit set iff diff != 0.
+         * Shift to get 1/0, subtract from 1 to invert, negate to mask.
+         */
+        diff = (sp_digit)(e ^ idx);
+        diff = (diff | ((sp_digit)0 - diff))
+               >> (sizeof(sp_digit) * 8 - 1);  /* 1 if !=, 0 if == */
+        mask = (sp_digit)0 - ((sp_digit)1 - diff);  /* all-1 if ==, else 0 */
+
+        for (j = 0; j < 16; j++) {
+            r[j] |= table[e][j] & mask;
+        }
+    }
+}
+#endif /* !WC_NO_CACHE_RESISTANT */
 /* Modular exponentiate a to the e mod m. (r = a^e mod m)
  *
  * r     A single precision number that is the result of the operation.
@@ -699,8 +724,6 @@ static int sp_2048_mod_exp_16(sp_digit* r, const sp_digit* a, const sp_digit* e,
     byte y;
     int err = MP_OKAY;
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (bits == 0) {
         err = MP_VAL;
     }
@@ -869,6 +892,7 @@ extern void sp_2048_mont_reduce_avx2_16(sp_digit* a, const sp_digit* m, sp_digit
 SP_NOINLINE static void sp_2048_mont_mul_avx2_16(sp_digit* r, const sp_digit* a,
         const sp_digit* b, const sp_digit* m, sp_digit mp)
 {
+    ASSERT_SAVED_VECTOR_REGISTERS();
     sp_2048_mul_avx2_16(r, a, b);
     sp_2048_mont_reduce_avx2_16(r, m, mp);
 }
@@ -885,6 +909,7 @@ SP_NOINLINE static void sp_2048_mont_mul_avx2_16(sp_digit* r, const sp_digit* a,
 SP_NOINLINE static void sp_2048_mont_sqr_avx2_16(sp_digit* r, const sp_digit* a,
         const sp_digit* m, sp_digit mp)
 {
+    ASSERT_SAVED_VECTOR_REGISTERS();
     sp_2048_sqr_avx2_16(r, a);
     sp_2048_mont_reduce_avx2_16(r, m, mp);
 }
@@ -926,7 +951,6 @@ static int sp_2048_mod_exp_avx2_16(sp_digit* r, const sp_digit* a, const sp_digi
     int err = MP_OKAY;
 
     ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (bits == 0) {
         err = MP_VAL;
     }
@@ -1089,8 +1113,6 @@ static void sp_2048_mont_norm_32(sp_digit* r, const sp_digit* m)
 {
     XMEMSET(r, 0, sizeof(sp_digit) * 32);
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     /* r = 2^n mod m */
     sp_2048_sub_in_place_32(r, m);
 }
@@ -1174,7 +1196,6 @@ extern sp_digit div_2048_word_asm_32(sp_digit d1, sp_digit d0, sp_digit div);
 static WC_INLINE sp_digit div_2048_word_32(sp_digit d1, sp_digit d0,
         sp_digit div)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
 #if _MSC_VER >= 1920
     return _udiv128(d1, d0, div, NULL);
 #else
@@ -1193,7 +1214,6 @@ static WC_INLINE sp_digit div_2048_word_32(sp_digit d1, sp_digit d0,
         sp_digit div)
 {
     register sp_digit r asm("rax");
-    ASSERT_SAVED_VECTOR_REGISTERS();
     __asm__ __volatile__ (
         "divq %3"
         : "=a" (r)
@@ -1224,8 +1244,6 @@ static WC_INLINE int sp_2048_div_32_cond(const sp_digit* a, const sp_digit* d, s
     word32 cpuid_flags = cpuid_get_flags();
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)m;
 
     div = d[31];
@@ -1284,7 +1302,6 @@ static WC_INLINE int sp_2048_div_32_cond(const sp_digit* a, const sp_digit* d, s
 static WC_INLINE int sp_2048_mod_32_cond(sp_digit* r, const sp_digit* a,
         const sp_digit* m)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
     return sp_2048_div_32_cond(a, m, NULL, r);
 }
 
@@ -1354,8 +1371,6 @@ static WC_INLINE int sp_2048_div_32(const sp_digit* a, const sp_digit* d, sp_dig
     word32 cpuid_flags = cpuid_get_flags();
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)m;
 
     div = d[31];
@@ -1409,19 +1424,39 @@ static WC_INLINE int sp_2048_div_32(const sp_digit* a, const sp_digit* d, sp_dig
 static WC_INLINE int sp_2048_mod_32(sp_digit* r, const sp_digit* a,
         const sp_digit* m)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
     return sp_2048_div_32(a, m, NULL, r);
 }
 
 #endif /* WOLFSSL_HAVE_SP_DH || !WOLFSSL_RSA_PUBLIC_ONLY */
-#ifdef __cplusplus
-extern "C" {
-#endif
-extern void sp_2048_get_from_table_32(sp_digit* r, sp_digit** table, int idx);
-#ifdef __cplusplus
-}
-#endif
+#ifndef WC_NO_CACHE_RESISTANT
+static void sp_2048_get_from_table_32(sp_digit* r,
+        sp_digit** table, int idx)
+{
+    int e, j;
+    sp_digit mask;
+    sp_digit diff;
 
+    for (j = 0; j < 32; j++) {
+        r[j] = 0;
+    }
+
+    for (e = 0; e < 64; e++) {
+        /* Constant-time: mask = (e == idx) ? ~0 : 0
+         * diff = e ^ idx is 0 iff equal.
+         * (diff | -diff) has its sign bit set iff diff != 0.
+         * Shift to get 1/0, subtract from 1 to invert, negate to mask.
+         */
+        diff = (sp_digit)(e ^ idx);
+        diff = (diff | ((sp_digit)0 - diff))
+               >> (sizeof(sp_digit) * 8 - 1);  /* 1 if !=, 0 if == */
+        mask = (sp_digit)0 - ((sp_digit)1 - diff);  /* all-1 if ==, else 0 */
+
+        for (j = 0; j < 32; j++) {
+            r[j] |= table[e][j] & mask;
+        }
+    }
+}
+#endif /* !WC_NO_CACHE_RESISTANT */
 /* Modular exponentiate a to the e mod m. (r = a^e mod m)
  *
  * r     A single precision number that is the result of the operation.
@@ -1448,8 +1483,6 @@ static int sp_2048_mod_exp_32(sp_digit* r, const sp_digit* a, const sp_digit* e,
     byte y;
     int err = MP_OKAY;
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (bits == 0) {
         err = MP_VAL;
     }
@@ -1652,6 +1685,7 @@ extern void sp_2048_mont_reduce_avx2_32(sp_digit* a, const sp_digit* m, sp_digit
 SP_NOINLINE static void sp_2048_mont_mul_avx2_32(sp_digit* r, const sp_digit* a,
         const sp_digit* b, const sp_digit* m, sp_digit mp)
 {
+    ASSERT_SAVED_VECTOR_REGISTERS();
     sp_2048_mul_avx2_32(r, a, b);
     sp_2048_mont_reduce_avx2_32(r, m, mp);
 }
@@ -1668,6 +1702,7 @@ SP_NOINLINE static void sp_2048_mont_mul_avx2_32(sp_digit* r, const sp_digit* a,
 SP_NOINLINE static void sp_2048_mont_sqr_avx2_32(sp_digit* r, const sp_digit* a,
         const sp_digit* m, sp_digit mp)
 {
+    ASSERT_SAVED_VECTOR_REGISTERS();
     sp_2048_sqr_avx2_32(r, a);
     sp_2048_mont_reduce_avx2_32(r, m, mp);
 }
@@ -1710,7 +1745,6 @@ static int sp_2048_mod_exp_avx2_32(sp_digit* r, const sp_digit* a, const sp_digi
     int err = MP_OKAY;
 
     ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (bits == 0) {
         err = MP_VAL;
     }
@@ -1918,10 +1952,9 @@ int sp_RsaPublic_2048(const byte* in, word32 inLen, const mp_int* em,
     int err = MP_OKAY;
 #ifdef HAVE_INTEL_AVX2
     word32 cpuid_flags = cpuid_get_flags();
+    int saved_vector_registers = 0;
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (*outLen < 256) {
         err = MP_TO_E;
     }
@@ -1953,6 +1986,13 @@ int sp_RsaPublic_2048(const byte* in, word32 inLen, const mp_int* em,
     if (err == MP_OKAY) {
         sp_2048_from_mp(m, 32, mm);
 
+#ifdef HAVE_INTEL_AVX2
+        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
+            IS_INTEL_AVX2(cpuid_flags) &&
+            (SAVE_VECTOR_REGISTERS2() == 0))
+            saved_vector_registers = 1;
+#endif
+
         if (e == 0x10001) {
             int i;
             sp_digit mp;
@@ -1967,8 +2007,7 @@ int sp_RsaPublic_2048(const byte* in, word32 inLen, const mp_int* em,
             if (err == MP_OKAY) {
                 /* r = a ^ 0x10000 => r = a squared 16 times */
 #ifdef HAVE_INTEL_AVX2
-                if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                        IS_INTEL_AVX2(cpuid_flags)) {
+                if (saved_vector_registers) {
                     for (i = 15; i >= 0; i--) {
                         sp_2048_mont_sqr_avx2_32(r, r, m, mp);
                     }
@@ -1999,8 +2038,7 @@ int sp_RsaPublic_2048(const byte* in, word32 inLen, const mp_int* em,
         }
         else if (e == 0x3) {
 #ifdef HAVE_INTEL_AVX2
-            if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+            if (saved_vector_registers) {
                 if (err == MP_OKAY) {
                     sp_2048_sqr_avx2_32(r, ah);
                     err = sp_2048_mod_32_cond(r, r, m);
@@ -2042,8 +2080,7 @@ int sp_RsaPublic_2048(const byte* in, word32 inLen, const mp_int* em,
 
                 XMEMCPY(r, a, sizeof(sp_digit) * 32);
 #ifdef HAVE_INTEL_AVX2
-                if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                        IS_INTEL_AVX2(cpuid_flags)) {
+                if (saved_vector_registers) {
                     for (i--; i>=0; i--) {
                         sp_2048_mont_sqr_avx2_32(r, r, m, mp);
                         if (((e >> i) & 1) == 1) {
@@ -2081,6 +2118,11 @@ int sp_RsaPublic_2048(const byte* in, word32 inLen, const mp_int* em,
         *outLen = 256;
     }
 
+#ifdef HAVE_INTEL_AVX2
+    if (saved_vector_registers)
+        RESTORE_VECTOR_REGISTERS();
+#endif
+
     SP_FREE_VAR(a, NULL, DYNAMIC_TYPE_RSA);
 
     return err;
@@ -2115,8 +2157,6 @@ int sp_RsaPrivate_2048(const byte* in, word32 inLen, const mp_int* dm,
     sp_digit* r = NULL;
     int err = MP_OKAY;
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)pm;
     (void)qm;
     (void)dpm;
@@ -2214,10 +2254,9 @@ int sp_RsaPrivate_2048(const byte* in, word32 inLen, const mp_int* dm,
     int err = MP_OKAY;
 #ifdef HAVE_INTEL_AVX2
     word32 cpuid_flags = cpuid_get_flags();
+    int saved_vector_registers = 0;
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)dm;
     (void)mm;
 
@@ -2238,6 +2277,7 @@ int sp_RsaPrivate_2048(const byte* in, word32 inLen, const mp_int* dm,
     }
 
     SP_ALLOC_VAR(sp_digit, a, 16 * 11, NULL, DYNAMIC_TYPE_RSA);
+
     if (err == MP_OKAY) {
         p = a + 32 * 2;
         q = p + 16;
@@ -2253,9 +2293,11 @@ int sp_RsaPrivate_2048(const byte* in, word32 inLen, const mp_int* dm,
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+            IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0))
+            saved_vector_registers = 1;
+
+        if (saved_vector_registers)
             err = sp_2048_mod_exp_avx2_16(tmpa, a, dp, 1024, p, 1);
-        }
         else
 #endif
             err = sp_2048_mod_exp_16(tmpa, a, dp, 1024, p, 1);
@@ -2263,10 +2305,8 @@ int sp_RsaPrivate_2048(const byte* in, word32 inLen, const mp_int* dm,
     if (err == MP_OKAY) {
         sp_2048_from_mp(dq, 16, dqm);
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             err = sp_2048_mod_exp_avx2_16(tmpb, a, dq, 1024, q, 1);
-       }
        else
 #endif
             err = sp_2048_mod_exp_16(tmpb, a, dq, 1024, q, 1);
@@ -2275,8 +2315,7 @@ int sp_RsaPrivate_2048(const byte* in, word32 inLen, const mp_int* dm,
     if (err == MP_OKAY) {
         c = sp_2048_sub_in_place_16(tmpa, tmpb);
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers) {
             c += sp_2048_cond_add_avx2_16(tmpa, tmpa, p, c);
             sp_2048_cond_add_avx2_16(tmpa, tmpa, p, c);
         }
@@ -2289,29 +2328,21 @@ int sp_RsaPrivate_2048(const byte* in, word32 inLen, const mp_int* dm,
 
         sp_2048_from_mp(qi, 16, qim);
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_2048_mul_avx2_16(tmpa, tmpa, qi);
-        }
         else
 #endif
-        {
             sp_2048_mul_16(tmpa, tmpa, qi);
-        }
         err = sp_2048_mod_16(tmpa, tmpa, p);
     }
 
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_2048_mul_avx2_16(tmpa, q, tmpa);
-        }
         else
 #endif
-        {
             sp_2048_mul_16(tmpa, q, tmpa);
-        }
         XMEMSET(&tmpb[16], 0, sizeof(sp_digit) * 16);
         sp_2048_add_32(r, tmpb, tmpa);
 
@@ -2319,6 +2350,11 @@ int sp_RsaPrivate_2048(const byte* in, word32 inLen, const mp_int* dm,
         *outLen = 256;
     }
 
+#ifdef HAVE_INTEL_AVX2
+    if (saved_vector_registers)
+        RESTORE_VECTOR_REGISTERS();
+#endif
+
     SP_ZEROFREE_VAR(sp_digit, a, 16 * 11, NULL, DYNAMIC_TYPE_RSA);
 
     return err;
@@ -2350,7 +2386,7 @@ static int sp_2048_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 32; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -2375,7 +2411,7 @@ static int sp_2048_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 32; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 64 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -2418,8 +2454,6 @@ int sp_ModExp_2048(const mp_int* base, const mp_int* exp, const mp_int* mod,
 #endif
     int expBits = mp_count_bits(exp);
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (mp_count_bits(base) > 2048) {
         err = MP_READ_E;
     }
@@ -2445,8 +2479,9 @@ int sp_ModExp_2048(const mp_int* base, const mp_int* exp, const mp_int* mod,
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_2048_mod_exp_avx2_32(r, b, e, expBits, m, 0);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -2500,7 +2535,6 @@ static int sp_2048_mod_exp_2_avx2_32(sp_digit* r, const sp_digit* e, int bits,
     int err = MP_OKAY;
 
     ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (bits == 0) {
         err = MP_VAL;
     }
@@ -2618,8 +2652,6 @@ static int sp_2048_mod_exp_2_32(sp_digit* r, const sp_digit* e, int bits,
     byte y;
     int err = MP_OKAY;
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (bits == 0) {
         err = MP_VAL;
     }
@@ -2738,8 +2770,6 @@ int sp_DhExp_2048(const mp_int* base, const byte* exp, word32 expLen,
     word32 cpuid_flags = cpuid_get_flags();
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (mp_count_bits(base) > 2048) {
         err = MP_READ_E;
     }
@@ -2767,8 +2797,10 @@ int sp_DhExp_2048(const mp_int* base, const byte* exp, word32 expLen,
         if (base->used == 1 && base->dp[0] == 2 && m[31] == (sp_digit)-1) {
 #ifdef HAVE_INTEL_AVX2
             if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+                    IS_INTEL_AVX2(cpuid_flags) &&
+                    (SAVE_VECTOR_REGISTERS2() == 0)) {
                 err = sp_2048_mod_exp_2_avx2_32(r, e, (int)expLen * 8, m);
+                RESTORE_VECTOR_REGISTERS();
             }
             else
 #endif
@@ -2779,8 +2811,10 @@ int sp_DhExp_2048(const mp_int* base, const byte* exp, word32 expLen,
         {
 #ifdef HAVE_INTEL_AVX2
             if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+                    IS_INTEL_AVX2(cpuid_flags) &&
+                    (SAVE_VECTOR_REGISTERS2() == 0)) {
                 err = sp_2048_mod_exp_avx2_32(r, b, e, (int)expLen * 8, m, 0);
+                RESTORE_VECTOR_REGISTERS();
             }
             else
 #endif
@@ -2827,8 +2861,6 @@ int sp_ModExp_1024(const mp_int* base, const mp_int* exp, const mp_int* mod,
 #endif
     int expBits = mp_count_bits(exp);
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (mp_count_bits(base) > 1024) {
         err = MP_READ_E;
     }
@@ -2854,8 +2886,9 @@ int sp_ModExp_1024(const mp_int* base, const mp_int* exp, const mp_int* mod,
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_2048_mod_exp_avx2_16(r, b, e, expBits, m, 0);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -2941,7 +2974,7 @@ static void sp_3072_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0xffffffffffffffffl;
         s = 64U - s;
         if (j + 1 >= size) {
@@ -2976,7 +3009,7 @@ static void sp_3072_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 64) {
             r[j] &= 0xffffffffffffffffl;
             if (j + 1 >= size) {
@@ -3189,10 +3222,10 @@ extern void sp_3072_sqr_avx2_48(sp_digit* r, const sp_digit* a);
  */
 static void sp_3072_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint64 x;
+    sp_uint64 b;
 
-    b = a[0];
+    b = (sp_uint64)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
@@ -3200,7 +3233,7 @@ static void sp_3072_mont_setup(const sp_digit* a, sp_digit* rho)
     x *= 2 - b * x;               /* here x*a==1 mod 2**64 */
 
     /* rho = -1/m mod b */
-    *rho = (sp_digit)0 - x;
+    *rho = (sp_digit)((sp_int64)0 - (sp_int64)x);
 }
 
 #ifdef __cplusplus
@@ -3221,8 +3254,6 @@ static void sp_3072_mont_norm_24(sp_digit* r, const sp_digit* m)
 {
     XMEMSET(r, 0, sizeof(sp_digit) * 24);
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     /* r = 2^n mod m */
     sp_3072_sub_in_place_24(r, m);
 }
@@ -3312,7 +3343,6 @@ extern sp_digit div_3072_word_asm_24(sp_digit d1, sp_digit d0, sp_digit div);
 static WC_INLINE sp_digit div_3072_word_24(sp_digit d1, sp_digit d0,
         sp_digit div)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
 #if _MSC_VER >= 1920
     return _udiv128(d1, d0, div, NULL);
 #else
@@ -3331,7 +3361,6 @@ static WC_INLINE sp_digit div_3072_word_24(sp_digit d1, sp_digit d0,
         sp_digit div)
 {
     register sp_digit r asm("rax");
-    ASSERT_SAVED_VECTOR_REGISTERS();
     __asm__ __volatile__ (
         "divq %3"
         : "=a" (r)
@@ -3399,8 +3428,6 @@ static WC_INLINE int sp_3072_div_24(const sp_digit* a, const sp_digit* d, sp_dig
     word32 cpuid_flags = cpuid_get_flags();
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)m;
 
     div = d[23];
@@ -3453,18 +3480,38 @@ static WC_INLINE int sp_3072_div_24(const sp_digit* a, const sp_digit* d, sp_dig
 static WC_INLINE int sp_3072_mod_24(sp_digit* r, const sp_digit* a,
         const sp_digit* m)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
     return sp_3072_div_24(a, m, NULL, r);
 }
 
-#ifdef __cplusplus
-extern "C" {
-#endif
-extern void sp_3072_get_from_table_24(sp_digit* r, sp_digit** table, int idx);
-#ifdef __cplusplus
-}
-#endif
+#ifndef WC_NO_CACHE_RESISTANT
+static void sp_3072_get_from_table_24(sp_digit* r,
+        sp_digit** table, int idx)
+{
+    int e, j;
+    sp_digit mask;
+    sp_digit diff;
 
+    for (j = 0; j < 24; j++) {
+        r[j] = 0;
+    }
+
+    for (e = 0; e < 32; e++) {
+        /* Constant-time: mask = (e == idx) ? ~0 : 0
+         * diff = e ^ idx is 0 iff equal.
+         * (diff | -diff) has its sign bit set iff diff != 0.
+         * Shift to get 1/0, subtract from 1 to invert, negate to mask.
+         */
+        diff = (sp_digit)(e ^ idx);
+        diff = (diff | ((sp_digit)0 - diff))
+               >> (sizeof(sp_digit) * 8 - 1);  /* 1 if !=, 0 if == */
+        mask = (sp_digit)0 - ((sp_digit)1 - diff);  /* all-1 if ==, else 0 */
+
+        for (j = 0; j < 24; j++) {
+            r[j] |= table[e][j] & mask;
+        }
+    }
+}
+#endif /* !WC_NO_CACHE_RESISTANT */
 /* Modular exponentiate a to the e mod m. (r = a^e mod m)
  *
  * r     A single precision number that is the result of the operation.
@@ -3491,8 +3538,6 @@ static int sp_3072_mod_exp_24(sp_digit* r, const sp_digit* a, const sp_digit* e,
     byte y;
     int err = MP_OKAY;
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (bits == 0) {
         err = MP_VAL;
     }
@@ -3661,6 +3706,7 @@ extern void sp_3072_mont_reduce_avx2_24(sp_digit* a, const sp_digit* m, sp_digit
 SP_NOINLINE static void sp_3072_mont_mul_avx2_24(sp_digit* r, const sp_digit* a,
         const sp_digit* b, const sp_digit* m, sp_digit mp)
 {
+    ASSERT_SAVED_VECTOR_REGISTERS();
     sp_3072_mul_avx2_24(r, a, b);
     sp_3072_mont_reduce_avx2_24(r, m, mp);
 }
@@ -3677,6 +3723,7 @@ SP_NOINLINE static void sp_3072_mont_mul_avx2_24(sp_digit* r, const sp_digit* a,
 SP_NOINLINE static void sp_3072_mont_sqr_avx2_24(sp_digit* r, const sp_digit* a,
         const sp_digit* m, sp_digit mp)
 {
+    ASSERT_SAVED_VECTOR_REGISTERS();
     sp_3072_sqr_avx2_24(r, a);
     sp_3072_mont_reduce_avx2_24(r, m, mp);
 }
@@ -3718,7 +3765,6 @@ static int sp_3072_mod_exp_avx2_24(sp_digit* r, const sp_digit* a, const sp_digi
     int err = MP_OKAY;
 
     ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (bits == 0) {
         err = MP_VAL;
     }
@@ -3881,8 +3927,6 @@ static void sp_3072_mont_norm_48(sp_digit* r, const sp_digit* m)
 {
     XMEMSET(r, 0, sizeof(sp_digit) * 48);
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     /* r = 2^n mod m */
     sp_3072_sub_in_place_48(r, m);
 }
@@ -3966,7 +4010,6 @@ extern sp_digit div_3072_word_asm_48(sp_digit d1, sp_digit d0, sp_digit div);
 static WC_INLINE sp_digit div_3072_word_48(sp_digit d1, sp_digit d0,
         sp_digit div)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
 #if _MSC_VER >= 1920
     return _udiv128(d1, d0, div, NULL);
 #else
@@ -3985,7 +4028,6 @@ static WC_INLINE sp_digit div_3072_word_48(sp_digit d1, sp_digit d0,
         sp_digit div)
 {
     register sp_digit r asm("rax");
-    ASSERT_SAVED_VECTOR_REGISTERS();
     __asm__ __volatile__ (
         "divq %3"
         : "=a" (r)
@@ -4016,8 +4058,6 @@ static WC_INLINE int sp_3072_div_48_cond(const sp_digit* a, const sp_digit* d, s
     word32 cpuid_flags = cpuid_get_flags();
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)m;
 
     div = d[47];
@@ -4076,7 +4116,6 @@ static WC_INLINE int sp_3072_div_48_cond(const sp_digit* a, const sp_digit* d, s
 static WC_INLINE int sp_3072_mod_48_cond(sp_digit* r, const sp_digit* a,
         const sp_digit* m)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
     return sp_3072_div_48_cond(a, m, NULL, r);
 }
 
@@ -4146,8 +4185,6 @@ static WC_INLINE int sp_3072_div_48(const sp_digit* a, const sp_digit* d, sp_dig
     word32 cpuid_flags = cpuid_get_flags();
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)m;
 
     div = d[47];
@@ -4201,19 +4238,39 @@ static WC_INLINE int sp_3072_div_48(const sp_digit* a, const sp_digit* d, sp_dig
 static WC_INLINE int sp_3072_mod_48(sp_digit* r, const sp_digit* a,
         const sp_digit* m)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
     return sp_3072_div_48(a, m, NULL, r);
 }
 
 #endif /* WOLFSSL_HAVE_SP_DH || !WOLFSSL_RSA_PUBLIC_ONLY */
-#ifdef __cplusplus
-extern "C" {
-#endif
-extern void sp_3072_get_from_table_48(sp_digit* r, sp_digit** table, int idx);
-#ifdef __cplusplus
-}
-#endif
+#ifndef WC_NO_CACHE_RESISTANT
+static void sp_3072_get_from_table_48(sp_digit* r,
+        sp_digit** table, int idx)
+{
+    int e, j;
+    sp_digit mask;
+    sp_digit diff;
 
+    for (j = 0; j < 48; j++) {
+        r[j] = 0;
+    }
+
+    for (e = 0; e < 16; e++) {
+        /* Constant-time: mask = (e == idx) ? ~0 : 0
+         * diff = e ^ idx is 0 iff equal.
+         * (diff | -diff) has its sign bit set iff diff != 0.
+         * Shift to get 1/0, subtract from 1 to invert, negate to mask.
+         */
+        diff = (sp_digit)(e ^ idx);
+        diff = (diff | ((sp_digit)0 - diff))
+               >> (sizeof(sp_digit) * 8 - 1);  /* 1 if !=, 0 if == */
+        mask = (sp_digit)0 - ((sp_digit)1 - diff);  /* all-1 if ==, else 0 */
+
+        for (j = 0; j < 48; j++) {
+            r[j] |= table[e][j] & mask;
+        }
+    }
+}
+#endif /* !WC_NO_CACHE_RESISTANT */
 /* Modular exponentiate a to the e mod m. (r = a^e mod m)
  *
  * r     A single precision number that is the result of the operation.
@@ -4240,8 +4297,6 @@ static int sp_3072_mod_exp_48(sp_digit* r, const sp_digit* a, const sp_digit* e,
     byte y;
     int err = MP_OKAY;
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (bits == 0) {
         err = MP_VAL;
     }
@@ -4392,6 +4447,7 @@ extern void sp_3072_mont_reduce_avx2_48(sp_digit* a, const sp_digit* m, sp_digit
 SP_NOINLINE static void sp_3072_mont_mul_avx2_48(sp_digit* r, const sp_digit* a,
         const sp_digit* b, const sp_digit* m, sp_digit mp)
 {
+    ASSERT_SAVED_VECTOR_REGISTERS();
     sp_3072_mul_avx2_48(r, a, b);
     sp_3072_mont_reduce_avx2_48(r, m, mp);
 }
@@ -4408,6 +4464,7 @@ SP_NOINLINE static void sp_3072_mont_mul_avx2_48(sp_digit* r, const sp_digit* a,
 SP_NOINLINE static void sp_3072_mont_sqr_avx2_48(sp_digit* r, const sp_digit* a,
         const sp_digit* m, sp_digit mp)
 {
+    ASSERT_SAVED_VECTOR_REGISTERS();
     sp_3072_sqr_avx2_48(r, a);
     sp_3072_mont_reduce_avx2_48(r, m, mp);
 }
@@ -4450,7 +4507,6 @@ static int sp_3072_mod_exp_avx2_48(sp_digit* r, const sp_digit* a, const sp_digi
     int err = MP_OKAY;
 
     ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (bits == 0) {
         err = MP_VAL;
     }
@@ -4606,10 +4662,9 @@ int sp_RsaPublic_3072(const byte* in, word32 inLen, const mp_int* em,
     int err = MP_OKAY;
 #ifdef HAVE_INTEL_AVX2
     word32 cpuid_flags = cpuid_get_flags();
+    int saved_vector_registers = 0;
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (*outLen < 384) {
         err = MP_TO_E;
     }
@@ -4641,6 +4696,13 @@ int sp_RsaPublic_3072(const byte* in, word32 inLen, const mp_int* em,
     if (err == MP_OKAY) {
         sp_3072_from_mp(m, 48, mm);
 
+#ifdef HAVE_INTEL_AVX2
+        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
+            IS_INTEL_AVX2(cpuid_flags) &&
+            (SAVE_VECTOR_REGISTERS2() == 0))
+            saved_vector_registers = 1;
+#endif
+
         if (e == 0x10001) {
             int i;
             sp_digit mp;
@@ -4655,8 +4717,7 @@ int sp_RsaPublic_3072(const byte* in, word32 inLen, const mp_int* em,
             if (err == MP_OKAY) {
                 /* r = a ^ 0x10000 => r = a squared 16 times */
 #ifdef HAVE_INTEL_AVX2
-                if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                        IS_INTEL_AVX2(cpuid_flags)) {
+                if (saved_vector_registers) {
                     for (i = 15; i >= 0; i--) {
                         sp_3072_mont_sqr_avx2_48(r, r, m, mp);
                     }
@@ -4687,8 +4748,7 @@ int sp_RsaPublic_3072(const byte* in, word32 inLen, const mp_int* em,
         }
         else if (e == 0x3) {
 #ifdef HAVE_INTEL_AVX2
-            if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+            if (saved_vector_registers) {
                 if (err == MP_OKAY) {
                     sp_3072_sqr_avx2_48(r, ah);
                     err = sp_3072_mod_48_cond(r, r, m);
@@ -4730,8 +4790,7 @@ int sp_RsaPublic_3072(const byte* in, word32 inLen, const mp_int* em,
 
                 XMEMCPY(r, a, sizeof(sp_digit) * 48);
 #ifdef HAVE_INTEL_AVX2
-                if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                        IS_INTEL_AVX2(cpuid_flags)) {
+                if (saved_vector_registers) {
                     for (i--; i>=0; i--) {
                         sp_3072_mont_sqr_avx2_48(r, r, m, mp);
                         if (((e >> i) & 1) == 1) {
@@ -4769,6 +4828,11 @@ int sp_RsaPublic_3072(const byte* in, word32 inLen, const mp_int* em,
         *outLen = 384;
     }
 
+#ifdef HAVE_INTEL_AVX2
+    if (saved_vector_registers)
+        RESTORE_VECTOR_REGISTERS();
+#endif
+
     SP_FREE_VAR(a, NULL, DYNAMIC_TYPE_RSA);
 
     return err;
@@ -4803,8 +4867,6 @@ int sp_RsaPrivate_3072(const byte* in, word32 inLen, const mp_int* dm,
     sp_digit* r = NULL;
     int err = MP_OKAY;
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)pm;
     (void)qm;
     (void)dpm;
@@ -4902,10 +4964,9 @@ int sp_RsaPrivate_3072(const byte* in, word32 inLen, const mp_int* dm,
     int err = MP_OKAY;
 #ifdef HAVE_INTEL_AVX2
     word32 cpuid_flags = cpuid_get_flags();
+    int saved_vector_registers = 0;
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)dm;
     (void)mm;
 
@@ -4926,6 +4987,7 @@ int sp_RsaPrivate_3072(const byte* in, word32 inLen, const mp_int* dm,
     }
 
     SP_ALLOC_VAR(sp_digit, a, 24 * 11, NULL, DYNAMIC_TYPE_RSA);
+
     if (err == MP_OKAY) {
         p = a + 48 * 2;
         q = p + 24;
@@ -4941,9 +5003,11 @@ int sp_RsaPrivate_3072(const byte* in, word32 inLen, const mp_int* dm,
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+            IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0))
+            saved_vector_registers = 1;
+
+        if (saved_vector_registers)
             err = sp_3072_mod_exp_avx2_24(tmpa, a, dp, 1536, p, 1);
-        }
         else
 #endif
             err = sp_3072_mod_exp_24(tmpa, a, dp, 1536, p, 1);
@@ -4951,10 +5015,8 @@ int sp_RsaPrivate_3072(const byte* in, word32 inLen, const mp_int* dm,
     if (err == MP_OKAY) {
         sp_3072_from_mp(dq, 24, dqm);
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             err = sp_3072_mod_exp_avx2_24(tmpb, a, dq, 1536, q, 1);
-       }
        else
 #endif
             err = sp_3072_mod_exp_24(tmpb, a, dq, 1536, q, 1);
@@ -4963,8 +5025,7 @@ int sp_RsaPrivate_3072(const byte* in, word32 inLen, const mp_int* dm,
     if (err == MP_OKAY) {
         c = sp_3072_sub_in_place_24(tmpa, tmpb);
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers) {
             c += sp_3072_cond_add_avx2_24(tmpa, tmpa, p, c);
             sp_3072_cond_add_avx2_24(tmpa, tmpa, p, c);
         }
@@ -4977,29 +5038,21 @@ int sp_RsaPrivate_3072(const byte* in, word32 inLen, const mp_int* dm,
 
         sp_3072_from_mp(qi, 24, qim);
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_3072_mul_avx2_24(tmpa, tmpa, qi);
-        }
         else
 #endif
-        {
             sp_3072_mul_24(tmpa, tmpa, qi);
-        }
         err = sp_3072_mod_24(tmpa, tmpa, p);
     }
 
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_3072_mul_avx2_24(tmpa, q, tmpa);
-        }
         else
 #endif
-        {
             sp_3072_mul_24(tmpa, q, tmpa);
-        }
         XMEMSET(&tmpb[24], 0, sizeof(sp_digit) * 24);
         sp_3072_add_48(r, tmpb, tmpa);
 
@@ -5007,6 +5060,11 @@ int sp_RsaPrivate_3072(const byte* in, word32 inLen, const mp_int* dm,
         *outLen = 384;
     }
 
+#ifdef HAVE_INTEL_AVX2
+    if (saved_vector_registers)
+        RESTORE_VECTOR_REGISTERS();
+#endif
+
     SP_ZEROFREE_VAR(sp_digit, a, 24 * 11, NULL, DYNAMIC_TYPE_RSA);
 
     return err;
@@ -5038,7 +5096,7 @@ static int sp_3072_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 48; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -5063,7 +5121,7 @@ static int sp_3072_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 48; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 64 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -5106,8 +5164,6 @@ int sp_ModExp_3072(const mp_int* base, const mp_int* exp, const mp_int* mod,
 #endif
     int expBits = mp_count_bits(exp);
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (mp_count_bits(base) > 3072) {
         err = MP_READ_E;
     }
@@ -5133,8 +5189,9 @@ int sp_ModExp_3072(const mp_int* base, const mp_int* exp, const mp_int* mod,
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_3072_mod_exp_avx2_48(r, b, e, expBits, m, 0);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -5188,7 +5245,6 @@ static int sp_3072_mod_exp_2_avx2_48(sp_digit* r, const sp_digit* e, int bits,
     int err = MP_OKAY;
 
     ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (bits == 0) {
         err = MP_VAL;
     }
@@ -5306,8 +5362,6 @@ static int sp_3072_mod_exp_2_48(sp_digit* r, const sp_digit* e, int bits,
     byte y;
     int err = MP_OKAY;
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (bits == 0) {
         err = MP_VAL;
     }
@@ -5426,8 +5480,6 @@ int sp_DhExp_3072(const mp_int* base, const byte* exp, word32 expLen,
     word32 cpuid_flags = cpuid_get_flags();
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (mp_count_bits(base) > 3072) {
         err = MP_READ_E;
     }
@@ -5455,8 +5507,10 @@ int sp_DhExp_3072(const mp_int* base, const byte* exp, word32 expLen,
         if (base->used == 1 && base->dp[0] == 2 && m[47] == (sp_digit)-1) {
 #ifdef HAVE_INTEL_AVX2
             if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+                    IS_INTEL_AVX2(cpuid_flags) &&
+                    (SAVE_VECTOR_REGISTERS2() == 0)) {
                 err = sp_3072_mod_exp_2_avx2_48(r, e, (int)expLen * 8, m);
+                RESTORE_VECTOR_REGISTERS();
             }
             else
 #endif
@@ -5467,8 +5521,10 @@ int sp_DhExp_3072(const mp_int* base, const byte* exp, word32 expLen,
         {
 #ifdef HAVE_INTEL_AVX2
             if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+                    IS_INTEL_AVX2(cpuid_flags) &&
+                    (SAVE_VECTOR_REGISTERS2() == 0)) {
                 err = sp_3072_mod_exp_avx2_48(r, b, e, (int)expLen * 8, m, 0);
+                RESTORE_VECTOR_REGISTERS();
             }
             else
 #endif
@@ -5515,8 +5571,6 @@ int sp_ModExp_1536(const mp_int* base, const mp_int* exp, const mp_int* mod,
 #endif
     int expBits = mp_count_bits(exp);
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (mp_count_bits(base) > 1536) {
         err = MP_READ_E;
     }
@@ -5542,8 +5596,9 @@ int sp_ModExp_1536(const mp_int* base, const mp_int* exp, const mp_int* mod,
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_3072_mod_exp_avx2_24(r, b, e, expBits, m, 0);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -5629,7 +5684,7 @@ static void sp_4096_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0xffffffffffffffffl;
         s = 64U - s;
         if (j + 1 >= size) {
@@ -5664,7 +5719,7 @@ static void sp_4096_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 64) {
             r[j] &= 0xffffffffffffffffl;
             if (j + 1 >= size) {
@@ -5793,10 +5848,10 @@ extern void sp_4096_sqr_avx2_64(sp_digit* r, const sp_digit* a);
  */
 static void sp_4096_mont_setup(const sp_digit* a, sp_digit* rho)
 {
-    sp_digit x;
-    sp_digit b;
+    sp_uint64 x;
+    sp_uint64 b;
 
-    b = a[0];
+    b = (sp_uint64)a[0];
     x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
     x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
@@ -5804,7 +5859,7 @@ static void sp_4096_mont_setup(const sp_digit* a, sp_digit* rho)
     x *= 2 - b * x;               /* here x*a==1 mod 2**64 */
 
     /* rho = -1/m mod b */
-    *rho = (sp_digit)0 - x;
+    *rho = (sp_digit)((sp_int64)0 - (sp_int64)x);
 }
 
 #ifdef __cplusplus
@@ -5825,8 +5880,6 @@ static void sp_4096_mont_norm_64(sp_digit* r, const sp_digit* m)
 {
     XMEMSET(r, 0, sizeof(sp_digit) * 64);
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     /* r = 2^n mod m */
     sp_4096_sub_in_place_64(r, m);
 }
@@ -5910,7 +5963,6 @@ extern sp_digit div_4096_word_asm_64(sp_digit d1, sp_digit d0, sp_digit div);
 static WC_INLINE sp_digit div_4096_word_64(sp_digit d1, sp_digit d0,
         sp_digit div)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
 #if _MSC_VER >= 1920
     return _udiv128(d1, d0, div, NULL);
 #else
@@ -5929,7 +5981,6 @@ static WC_INLINE sp_digit div_4096_word_64(sp_digit d1, sp_digit d0,
         sp_digit div)
 {
     register sp_digit r asm("rax");
-    ASSERT_SAVED_VECTOR_REGISTERS();
     __asm__ __volatile__ (
         "divq %3"
         : "=a" (r)
@@ -5960,8 +6011,6 @@ static WC_INLINE int sp_4096_div_64_cond(const sp_digit* a, const sp_digit* d, s
     word32 cpuid_flags = cpuid_get_flags();
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)m;
 
     div = d[63];
@@ -6020,7 +6069,6 @@ static WC_INLINE int sp_4096_div_64_cond(const sp_digit* a, const sp_digit* d, s
 static WC_INLINE int sp_4096_mod_64_cond(sp_digit* r, const sp_digit* a,
         const sp_digit* m)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
     return sp_4096_div_64_cond(a, m, NULL, r);
 }
 
@@ -6090,8 +6138,6 @@ static WC_INLINE int sp_4096_div_64(const sp_digit* a, const sp_digit* d, sp_dig
     word32 cpuid_flags = cpuid_get_flags();
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)m;
 
     div = d[63];
@@ -6145,19 +6191,39 @@ static WC_INLINE int sp_4096_div_64(const sp_digit* a, const sp_digit* d, sp_dig
 static WC_INLINE int sp_4096_mod_64(sp_digit* r, const sp_digit* a,
         const sp_digit* m)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
     return sp_4096_div_64(a, m, NULL, r);
 }
 
 #endif /* WOLFSSL_HAVE_SP_DH || !WOLFSSL_RSA_PUBLIC_ONLY */
-#ifdef __cplusplus
-extern "C" {
-#endif
-extern void sp_4096_get_from_table_64(sp_digit* r, sp_digit** table, int idx);
-#ifdef __cplusplus
-}
-#endif
+#ifndef WC_NO_CACHE_RESISTANT
+static void sp_4096_get_from_table_64(sp_digit* r,
+        sp_digit** table, int idx)
+{
+    int e, j;
+    sp_digit mask;
+    sp_digit diff;
 
+    for (j = 0; j < 64; j++) {
+        r[j] = 0;
+    }
+
+    for (e = 0; e < 16; e++) {
+        /* Constant-time: mask = (e == idx) ? ~0 : 0
+         * diff = e ^ idx is 0 iff equal.
+         * (diff | -diff) has its sign bit set iff diff != 0.
+         * Shift to get 1/0, subtract from 1 to invert, negate to mask.
+         */
+        diff = (sp_digit)(e ^ idx);
+        diff = (diff | ((sp_digit)0 - diff))
+               >> (sizeof(sp_digit) * 8 - 1);  /* 1 if !=, 0 if == */
+        mask = (sp_digit)0 - ((sp_digit)1 - diff);  /* all-1 if ==, else 0 */
+
+        for (j = 0; j < 64; j++) {
+            r[j] |= table[e][j] & mask;
+        }
+    }
+}
+#endif /* !WC_NO_CACHE_RESISTANT */
 /* Modular exponentiate a to the e mod m. (r = a^e mod m)
  *
  * r     A single precision number that is the result of the operation.
@@ -6184,8 +6250,6 @@ static int sp_4096_mod_exp_64(sp_digit* r, const sp_digit* a, const sp_digit* e,
     byte y;
     int err = MP_OKAY;
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (bits == 0) {
         err = MP_VAL;
     }
@@ -6336,6 +6400,7 @@ extern void sp_4096_mont_reduce_avx2_64(sp_digit* a, const sp_digit* m, sp_digit
 SP_NOINLINE static void sp_4096_mont_mul_avx2_64(sp_digit* r, const sp_digit* a,
         const sp_digit* b, const sp_digit* m, sp_digit mp)
 {
+    ASSERT_SAVED_VECTOR_REGISTERS();
     sp_4096_mul_avx2_64(r, a, b);
     sp_4096_mont_reduce_avx2_64(r, m, mp);
 }
@@ -6352,6 +6417,7 @@ SP_NOINLINE static void sp_4096_mont_mul_avx2_64(sp_digit* r, const sp_digit* a,
 SP_NOINLINE static void sp_4096_mont_sqr_avx2_64(sp_digit* r, const sp_digit* a,
         const sp_digit* m, sp_digit mp)
 {
+    ASSERT_SAVED_VECTOR_REGISTERS();
     sp_4096_sqr_avx2_64(r, a);
     sp_4096_mont_reduce_avx2_64(r, m, mp);
 }
@@ -6394,7 +6460,6 @@ static int sp_4096_mod_exp_avx2_64(sp_digit* r, const sp_digit* a, const sp_digi
     int err = MP_OKAY;
 
     ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (bits == 0) {
         err = MP_VAL;
     }
@@ -6550,10 +6615,9 @@ int sp_RsaPublic_4096(const byte* in, word32 inLen, const mp_int* em,
     int err = MP_OKAY;
 #ifdef HAVE_INTEL_AVX2
     word32 cpuid_flags = cpuid_get_flags();
+    int saved_vector_registers = 0;
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (*outLen < 512) {
         err = MP_TO_E;
     }
@@ -6585,6 +6649,13 @@ int sp_RsaPublic_4096(const byte* in, word32 inLen, const mp_int* em,
     if (err == MP_OKAY) {
         sp_4096_from_mp(m, 64, mm);
 
+#ifdef HAVE_INTEL_AVX2
+        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
+            IS_INTEL_AVX2(cpuid_flags) &&
+            (SAVE_VECTOR_REGISTERS2() == 0))
+            saved_vector_registers = 1;
+#endif
+
         if (e == 0x10001) {
             int i;
             sp_digit mp;
@@ -6599,8 +6670,7 @@ int sp_RsaPublic_4096(const byte* in, word32 inLen, const mp_int* em,
             if (err == MP_OKAY) {
                 /* r = a ^ 0x10000 => r = a squared 16 times */
 #ifdef HAVE_INTEL_AVX2
-                if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                        IS_INTEL_AVX2(cpuid_flags)) {
+                if (saved_vector_registers) {
                     for (i = 15; i >= 0; i--) {
                         sp_4096_mont_sqr_avx2_64(r, r, m, mp);
                     }
@@ -6631,8 +6701,7 @@ int sp_RsaPublic_4096(const byte* in, word32 inLen, const mp_int* em,
         }
         else if (e == 0x3) {
 #ifdef HAVE_INTEL_AVX2
-            if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+            if (saved_vector_registers) {
                 if (err == MP_OKAY) {
                     sp_4096_sqr_avx2_64(r, ah);
                     err = sp_4096_mod_64_cond(r, r, m);
@@ -6674,8 +6743,7 @@ int sp_RsaPublic_4096(const byte* in, word32 inLen, const mp_int* em,
 
                 XMEMCPY(r, a, sizeof(sp_digit) * 64);
 #ifdef HAVE_INTEL_AVX2
-                if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                        IS_INTEL_AVX2(cpuid_flags)) {
+                if (saved_vector_registers) {
                     for (i--; i>=0; i--) {
                         sp_4096_mont_sqr_avx2_64(r, r, m, mp);
                         if (((e >> i) & 1) == 1) {
@@ -6713,6 +6781,11 @@ int sp_RsaPublic_4096(const byte* in, word32 inLen, const mp_int* em,
         *outLen = 512;
     }
 
+#ifdef HAVE_INTEL_AVX2
+    if (saved_vector_registers)
+        RESTORE_VECTOR_REGISTERS();
+#endif
+
     SP_FREE_VAR(a, NULL, DYNAMIC_TYPE_RSA);
 
     return err;
@@ -6747,8 +6820,6 @@ int sp_RsaPrivate_4096(const byte* in, word32 inLen, const mp_int* dm,
     sp_digit* r = NULL;
     int err = MP_OKAY;
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)pm;
     (void)qm;
     (void)dpm;
@@ -6846,10 +6917,9 @@ int sp_RsaPrivate_4096(const byte* in, word32 inLen, const mp_int* dm,
     int err = MP_OKAY;
 #ifdef HAVE_INTEL_AVX2
     word32 cpuid_flags = cpuid_get_flags();
+    int saved_vector_registers = 0;
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)dm;
     (void)mm;
 
@@ -6870,6 +6940,7 @@ int sp_RsaPrivate_4096(const byte* in, word32 inLen, const mp_int* dm,
     }
 
     SP_ALLOC_VAR(sp_digit, a, 32 * 11, NULL, DYNAMIC_TYPE_RSA);
+
     if (err == MP_OKAY) {
         p = a + 64 * 2;
         q = p + 32;
@@ -6885,9 +6956,11 @@ int sp_RsaPrivate_4096(const byte* in, word32 inLen, const mp_int* dm,
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+            IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0))
+            saved_vector_registers = 1;
+
+        if (saved_vector_registers)
             err = sp_2048_mod_exp_avx2_32(tmpa, a, dp, 2048, p, 1);
-        }
         else
 #endif
             err = sp_2048_mod_exp_32(tmpa, a, dp, 2048, p, 1);
@@ -6895,10 +6968,8 @@ int sp_RsaPrivate_4096(const byte* in, word32 inLen, const mp_int* dm,
     if (err == MP_OKAY) {
         sp_4096_from_mp(dq, 32, dqm);
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             err = sp_2048_mod_exp_avx2_32(tmpb, a, dq, 2048, q, 1);
-       }
        else
 #endif
             err = sp_2048_mod_exp_32(tmpb, a, dq, 2048, q, 1);
@@ -6907,8 +6978,7 @@ int sp_RsaPrivate_4096(const byte* in, word32 inLen, const mp_int* dm,
     if (err == MP_OKAY) {
         c = sp_2048_sub_in_place_32(tmpa, tmpb);
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers) {
             c += sp_4096_cond_add_avx2_32(tmpa, tmpa, p, c);
             sp_4096_cond_add_avx2_32(tmpa, tmpa, p, c);
         }
@@ -6921,29 +6991,21 @@ int sp_RsaPrivate_4096(const byte* in, word32 inLen, const mp_int* dm,
 
         sp_2048_from_mp(qi, 32, qim);
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_2048_mul_avx2_32(tmpa, tmpa, qi);
-        }
         else
 #endif
-        {
             sp_2048_mul_32(tmpa, tmpa, qi);
-        }
         err = sp_2048_mod_32(tmpa, tmpa, p);
     }
 
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_2048_mul_avx2_32(tmpa, q, tmpa);
-        }
         else
 #endif
-        {
             sp_2048_mul_32(tmpa, q, tmpa);
-        }
         XMEMSET(&tmpb[32], 0, sizeof(sp_digit) * 32);
         sp_4096_add_64(r, tmpb, tmpa);
 
@@ -6951,6 +7013,11 @@ int sp_RsaPrivate_4096(const byte* in, word32 inLen, const mp_int* dm,
         *outLen = 512;
     }
 
+#ifdef HAVE_INTEL_AVX2
+    if (saved_vector_registers)
+        RESTORE_VECTOR_REGISTERS();
+#endif
+
     SP_ZEROFREE_VAR(sp_digit, a, 32 * 11, NULL, DYNAMIC_TYPE_RSA);
 
     return err;
@@ -6982,7 +7049,7 @@ static int sp_4096_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 64; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -7007,7 +7074,7 @@ static int sp_4096_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 64; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 64 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -7050,8 +7117,6 @@ int sp_ModExp_4096(const mp_int* base, const mp_int* exp, const mp_int* mod,
 #endif
     int expBits = mp_count_bits(exp);
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (mp_count_bits(base) > 4096) {
         err = MP_READ_E;
     }
@@ -7077,8 +7142,9 @@ int sp_ModExp_4096(const mp_int* base, const mp_int* exp, const mp_int* mod,
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_4096_mod_exp_avx2_64(r, b, e, expBits, m, 0);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -7132,7 +7198,6 @@ static int sp_4096_mod_exp_2_avx2_64(sp_digit* r, const sp_digit* e, int bits,
     int err = MP_OKAY;
 
     ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (bits == 0) {
         err = MP_VAL;
     }
@@ -7250,8 +7315,6 @@ static int sp_4096_mod_exp_2_64(sp_digit* r, const sp_digit* e, int bits,
     byte y;
     int err = MP_OKAY;
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (bits == 0) {
         err = MP_VAL;
     }
@@ -7370,8 +7433,6 @@ int sp_DhExp_4096(const mp_int* base, const byte* exp, word32 expLen,
     word32 cpuid_flags = cpuid_get_flags();
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     if (mp_count_bits(base) > 4096) {
         err = MP_READ_E;
     }
@@ -7399,8 +7460,10 @@ int sp_DhExp_4096(const mp_int* base, const byte* exp, word32 expLen,
         if (base->used == 1 && base->dp[0] == 2 && m[63] == (sp_digit)-1) {
 #ifdef HAVE_INTEL_AVX2
             if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+                    IS_INTEL_AVX2(cpuid_flags) &&
+                    (SAVE_VECTOR_REGISTERS2() == 0)) {
                 err = sp_4096_mod_exp_2_avx2_64(r, e, (int)expLen * 8, m);
+                RESTORE_VECTOR_REGISTERS();
             }
             else
 #endif
@@ -7411,8 +7474,10 @@ int sp_DhExp_4096(const mp_int* base, const byte* exp, word32 expLen,
         {
 #ifdef HAVE_INTEL_AVX2
             if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+                    IS_INTEL_AVX2(cpuid_flags) &&
+                    (SAVE_VECTOR_REGISTERS2() == 0)) {
                 err = sp_4096_mod_exp_avx2_64(r, b, e, (int)expLen * 8, m, 0);
+                RESTORE_VECTOR_REGISTERS();
             }
             else
 #endif
@@ -7586,8 +7651,6 @@ static int sp_256_mod_mul_norm_4(sp_digit* r, const sp_digit* a, const sp_digit*
     int64_t a32[8];
     int64_t o;
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)m;
 
     a32[0] = (int64_t)(a[0] & 0xffffffff);
@@ -7647,10 +7710,10 @@ static int sp_256_mod_mul_norm_4(sp_digit* r, const sp_digit* a, const sp_digit*
     t[5] += t[4] >> 32; t[4] &= 0xffffffff;
     t[6] += t[5] >> 32; t[5] &= 0xffffffff;
     t[7] += t[6] >> 32; t[6] &= 0xffffffff;
-    r[0] = (sp_digit)((t[1] << 32) | t[0]);
-    r[1] = (sp_digit)((t[3] << 32) | t[2]);
-    r[2] = (sp_digit)((t[5] << 32) | t[4]);
-    r[3] = (sp_digit)((t[7] << 32) | t[6]);
+    r[0] = (sp_digit)(((sp_uint64)t[1] << 32) | (sp_uint64)t[0]);
+    r[1] = (sp_digit)(((sp_uint64)t[3] << 32) | (sp_uint64)t[2]);
+    r[2] = (sp_digit)(((sp_uint64)t[5] << 32) | (sp_uint64)t[4]);
+    r[3] = (sp_digit)(((sp_uint64)t[7] << 32) | (sp_uint64)t[6]);
 
     return MP_OKAY;
 }
@@ -7681,7 +7744,7 @@ static void sp_256_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0xffffffffffffffffl;
         s = 64U - s;
         if (j + 1 >= size) {
@@ -7716,7 +7779,7 @@ static void sp_256_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 64) {
             r[j] &= 0xffffffffffffffffl;
             if (j + 1 >= size) {
@@ -7782,7 +7845,7 @@ static int sp_256_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 4; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -7807,7 +7870,7 @@ static int sp_256_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 4; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 64 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -7914,7 +7977,7 @@ static void sp_256_mont_inv_4(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 4);
     for (i=254; i>=0; i--) {
         sp_256_mont_sqr_4(t, t, p256_mod, p256_mp_mod);
-        if (p256_mod_minus_2[i / 64] & ((sp_digit)1 << (i % 64)))
+        if (p256_mod_minus_2[i / 64] & ((sp_uint64)1 << (i % 64)))
             sp_256_mont_mul_4(t, t, a, p256_mod, p256_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 4);
@@ -8885,7 +8948,7 @@ static void sp_256_ecc_recode_6_4(const sp_digit* k, ecc_recode_256* v)
         }
         else if (++j < 4) {
             n = k[j];
-            y |= (word8)((n << (64 - o)) & 0x3f);
+            y |= (word8)(((sp_uint64)n << (64 - o)) & 0x3f);
             o -= 58;
             n >>= o;
         }
@@ -9088,7 +9151,7 @@ static void sp_256_mont_inv_avx2_4(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 4);
     for (i=254; i>=0; i--) {
         sp_256_mont_sqr_avx2_4(t, t, p256_mod, p256_mp_mod);
-        if (p256_mod_minus_2[i / 64] & ((sp_digit)1 << (i % 64)))
+        if (p256_mod_minus_2[i / 64] & ((sp_uint64)1 << (i % 64)))
             sp_256_mont_mul_avx2_4(t, t, a, p256_mod, p256_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 4);
@@ -10872,8 +10935,9 @@ int sp_ecc_mulmod_256(const mp_int* km, const ecc_point* gm, ecc_point* r,
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_256_ecc_mulmod_avx2_4(point, point, k, map, 1, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -10911,6 +10975,7 @@ int sp_ecc_mulmod_add_256(const mp_int* km, const ecc_point* gm,
     int err = MP_OKAY;
 #ifdef HAVE_INTEL_AVX2
     word32 cpuid_flags = cpuid_get_flags();
+    int saved_vector_registers = 0;
 #endif
 
     SP_ALLOC_VAR(sp_point_256, point, 2, heap, DYNAMIC_TYPE_ECC);
@@ -10935,29 +11000,26 @@ int sp_ecc_mulmod_add_256(const mp_int* km, const ecc_point* gm,
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0))
+            saved_vector_registers = 1;
+        if (saved_vector_registers)
             err = sp_256_ecc_mulmod_avx2_4(point, point, k, 0, 0, heap);
-        }
         else
 #endif
             err = sp_256_ecc_mulmod_4(point, point, k, 0, 0, heap);
     }
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_256_proj_point_add_avx2_4(point, point, addP, tmp);
-        }
         else
 #endif
             sp_256_proj_point_add_4(point, point, addP, tmp);
 
         if (map) {
 #ifdef HAVE_INTEL_AVX2
-            if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+            if (saved_vector_registers)
                 sp_256_map_avx2_4(point, point, tmp);
-            }
             else
 #endif
                 sp_256_map_4(point, point, tmp);
@@ -10966,6 +11028,11 @@ int sp_ecc_mulmod_add_256(const mp_int* km, const ecc_point* gm,
         err = sp_256_point_to_ecc_point_4(point, r);
     }
 
+#ifdef HAVE_INTEL_AVX2
+    if (saved_vector_registers)
+        RESTORE_VECTOR_REGISTERS();
+#endif
+
     SP_FREE_VAR(k, heap, DYNAMIC_TYPE_ECC);
     SP_FREE_VAR(point, heap, DYNAMIC_TYPE_ECC);
 
@@ -11404,7 +11471,7 @@ static void sp_256_ecc_recode_7_4(const sp_digit* k, ecc_recode_256* v)
         }
         else if (++j < 4) {
             n = k[j];
-            y |= (word8)((n << (64 - o)) & 0x7f);
+            y |= (word8)(((sp_uint64)n << (64 - o)) & 0x7f);
             o -= 57;
             n >>= o;
         }
@@ -23624,8 +23691,9 @@ int sp_ecc_mulmod_base_256(const mp_int* km, ecc_point* r, int map, void* heap)
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_256_ecc_mulmod_base_avx2_4(point, k, map, 1, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -23662,6 +23730,7 @@ int sp_ecc_mulmod_base_add_256(const mp_int* km, const ecc_point* am,
     int err = MP_OKAY;
 #ifdef HAVE_INTEL_AVX2
     word32 cpuid_flags = cpuid_get_flags();
+    int saved_vector_registers = 0;
 #endif
 
     SP_ALLOC_VAR(sp_point_256, point, 2, NULL, DYNAMIC_TYPE_ECC);
@@ -23685,29 +23754,26 @@ int sp_ecc_mulmod_base_add_256(const mp_int* km, const ecc_point* am,
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0))
+            saved_vector_registers = 1;
+        if (saved_vector_registers)
             err = sp_256_ecc_mulmod_base_avx2_4(point, k, 0, 0, heap);
-        }
         else
 #endif
             err = sp_256_ecc_mulmod_base_4(point, k, 0, 0, heap);
     }
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_256_proj_point_add_avx2_4(point, point, addP, tmp);
-        }
         else
 #endif
             sp_256_proj_point_add_4(point, point, addP, tmp);
 
         if (map) {
 #ifdef HAVE_INTEL_AVX2
-            if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+            if (saved_vector_registers)
                 sp_256_map_avx2_4(point, point, tmp);
-            }
             else
 #endif
                 sp_256_map_4(point, point, tmp);
@@ -23716,6 +23782,11 @@ int sp_ecc_mulmod_base_add_256(const mp_int* km, const ecc_point* am,
         err = sp_256_point_to_ecc_point_4(point, r);
     }
 
+#ifdef HAVE_INTEL_AVX2
+    if (saved_vector_registers)
+        RESTORE_VECTOR_REGISTERS();
+#endif
+
     SP_FREE_VAR(k, NULL, DYNAMIC_TYPE_ECC);
     SP_FREE_VAR(point, NULL, DYNAMIC_TYPE_ECC);
 
@@ -23827,6 +23898,7 @@ int sp_ecc_make_key_256(WC_RNG* rng, mp_int* priv, ecc_point* pub, void* heap)
 
 #ifdef HAVE_INTEL_AVX2
     word32 cpuid_flags = cpuid_get_flags();
+    int saved_vector_registers = 0;
 #endif
 
     (void)heap;
@@ -23847,9 +23919,11 @@ int sp_ecc_make_key_256(WC_RNG* rng, mp_int* priv, ecc_point* pub, void* heap)
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0))
+            saved_vector_registers = 1;
+
+        if (saved_vector_registers)
             err = sp_256_ecc_mulmod_base_avx2_4(point, k, 1, 1, NULL);
-       }
         else
 #endif
             err = sp_256_ecc_mulmod_base_4(point, k, 1, 1, NULL);
@@ -23858,8 +23932,7 @@ int sp_ecc_make_key_256(WC_RNG* rng, mp_int* priv, ecc_point* pub, void* heap)
 #ifdef WOLFSSL_VALIDATE_ECC_KEYGEN
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers) {
             err = sp_256_ecc_mulmod_avx2_4(infinity, point, p256_order, 1, 1,
                                                                           NULL);
         }
@@ -23874,6 +23947,11 @@ int sp_ecc_make_key_256(WC_RNG* rng, mp_int* priv, ecc_point* pub, void* heap)
     }
 #endif
 
+#ifdef HAVE_INTEL_AVX2
+    if (saved_vector_registers)
+        RESTORE_VECTOR_REGISTERS();
+#endif
+
     if (err == MP_OKAY) {
         err = sp_256_to_mp(k, priv);
     }
@@ -24035,8 +24113,9 @@ int sp_ecc_secret_gen_256(const mp_int* priv, const ecc_point* pub, byte* out,
         sp_256_point_from_ecc_point_4(point, pub);
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_256_ecc_mulmod_avx2_4(point, point, k, 1, 1, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -24148,7 +24227,6 @@ extern sp_digit div_256_word_asm_4(sp_digit d1, sp_digit d0, sp_digit div);
 static WC_INLINE sp_digit div_256_word_4(sp_digit d1, sp_digit d0,
         sp_digit div)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
 #if _MSC_VER >= 1920
     return _udiv128(d1, d0, div, NULL);
 #else
@@ -24167,7 +24245,6 @@ static WC_INLINE sp_digit div_256_word_4(sp_digit d1, sp_digit d0,
         sp_digit div)
 {
     register sp_digit r asm("rax");
-    ASSERT_SAVED_VECTOR_REGISTERS();
     __asm__ __volatile__ (
         "divq %3"
         : "=a" (r)
@@ -24220,8 +24297,6 @@ static WC_INLINE int sp_256_div_4(const sp_digit* a, const sp_digit* d, sp_digit
     word32 cpuid_flags = cpuid_get_flags();
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)m;
 
     div = d[3];
@@ -24274,7 +24349,6 @@ static WC_INLINE int sp_256_div_4(const sp_digit* a, const sp_digit* d, sp_digit
 static WC_INLINE int sp_256_mod_4(sp_digit* r, const sp_digit* a,
         const sp_digit* m)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
     return sp_256_div_4(a, m, NULL, r);
 }
 
@@ -24288,7 +24362,6 @@ static WC_INLINE int sp_256_mod_4(sp_digit* r, const sp_digit* a,
  */
 static void sp_256_mont_mul_order_4(sp_digit* r, const sp_digit* a, const sp_digit* b)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
     sp_256_mul_4(r, a, b);
     sp_256_mont_reduce_order_4(r, p256_order, p256_mp_order);
 }
@@ -24314,7 +24387,6 @@ static const word64 p256_order_low[2] = {
  */
 static void sp_256_mont_sqr_order_4(sp_digit* r, const sp_digit* a)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
     sp_256_sqr_4(r, a);
     sp_256_mont_reduce_order_4(r, p256_order, p256_mp_order);
 }
@@ -24330,8 +24402,6 @@ static void sp_256_mont_sqr_n_order_4(sp_digit* r, const sp_digit* a, int n)
 {
     int i;
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     sp_256_mont_sqr_order_4(r, a);
     for (i=1; i<n; i++) {
         sp_256_mont_sqr_order_4(r, r);
@@ -24358,8 +24428,6 @@ static int sp_256_mont_inv_order_4_nb(sp_ecc_ctx_t* sp_ctx, sp_digit* r, const s
 {
     int err = FP_WOULDBLOCK;
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     sp_256_mont_inv_order_4_ctx* ctx = (sp_256_mont_inv_order_4_ctx*)sp_ctx;
 
     typedef char ctx_size_test[sizeof(sp_256_mont_inv_order_4_ctx) >= sizeof(*sp_ctx) ? -1 : 1];
@@ -24420,8 +24488,6 @@ static void sp_256_mont_inv_order_4(sp_digit* r, const sp_digit* a,
     sp_digit* t4 = td + 6 * 4;
     int i;
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     /* t = a^2 */
     sp_256_mont_sqr_order_4(t, a);
     /* t = a^3 = t * a */
@@ -24554,7 +24620,6 @@ static void sp_256_mont_sqr_n_order_avx2_4(sp_digit* r, const sp_digit* a, int n
     int i;
 
     ASSERT_SAVED_VECTOR_REGISTERS();
-
     sp_256_mont_sqr_order_avx2_4(r, a);
     for (i=1; i<n; i++) {
         sp_256_mont_sqr_order_avx2_4(r, r);
@@ -24582,7 +24647,6 @@ static int sp_256_mont_inv_order_avx2_4_nb(sp_ecc_ctx_t* sp_ctx, sp_digit* r, co
     int err = FP_WOULDBLOCK;
 
     ASSERT_SAVED_VECTOR_REGISTERS();
-
     sp_256_mont_inv_order_avx2_4_ctx* ctx = (sp_256_mont_inv_order_avx2_4_ctx*)sp_ctx;
 
     typedef char ctx_size_test[sizeof(sp_256_mont_inv_order_avx2_4_ctx) >= sizeof(*sp_ctx) ? -1 : 1];
@@ -24644,7 +24708,6 @@ static void sp_256_mont_inv_order_avx2_4(sp_digit* r, const sp_digit* a,
     int i;
 
     ASSERT_SAVED_VECTOR_REGISTERS();
-
     /* t = a^2 */
     sp_256_mont_sqr_order_avx2_4(t, a);
     /* t = a^3 = t * a */
@@ -24776,14 +24839,16 @@ static int sp_256_calc_s_4(sp_digit* s, const sp_digit* r, sp_digit* k,
     sp_digit* kInv = k;
 #ifdef HAVE_INTEL_AVX2
     word32 cpuid_flags = cpuid_get_flags();
+    int saved_vector_registers = 0;
 #endif
 
     /* Conv k to Montgomery form (mod order) */
 #ifdef HAVE_INTEL_AVX2
     if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-            IS_INTEL_AVX2(cpuid_flags)) {
+            IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0))
+        saved_vector_registers = 1;
+    if (saved_vector_registers)
         sp_256_mul_avx2_4(k, k, p256_norm_order);
-    }
     else
 #endif
         sp_256_mul_4(k, k, p256_norm_order);
@@ -24793,10 +24858,8 @@ static int sp_256_calc_s_4(sp_digit* s, const sp_digit* r, sp_digit* k,
 
         /* kInv = 1/k mod order */
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_256_mont_inv_order_avx2_4(kInv, k, tmp);
-        }
         else
 #endif
             sp_256_mont_inv_order_4(kInv, k, tmp);
@@ -24804,10 +24867,8 @@ static int sp_256_calc_s_4(sp_digit* s, const sp_digit* r, sp_digit* k,
 
         /* s = r * x + e */
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_256_mul_avx2_4(x, x, r);
-        }
         else
 #endif
             sp_256_mul_4(x, x, r);
@@ -24825,16 +24886,19 @@ static int sp_256_calc_s_4(sp_digit* s, const sp_digit* r, sp_digit* k,
 
         /* s = s * k^-1 mod order */
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_256_mont_mul_order_avx2_4(s, s, kInv);
-        }
         else
 #endif
             sp_256_mont_mul_order_4(s, s, kInv);
         sp_256_norm_4(s);
     }
 
+#ifdef HAVE_INTEL_AVX2
+    if (saved_vector_registers)
+        RESTORE_VECTOR_REGISTERS();
+#endif
+
     return err;
 }
 
@@ -24899,8 +24963,10 @@ int sp_ecc_sign_256(const byte* hash, word32 hashLen, WC_RNG* rng,
         if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
             if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+                    IS_INTEL_AVX2(cpuid_flags) &&
+                    (SAVE_VECTOR_REGISTERS2() == 0)) {
                 err = sp_256_ecc_mulmod_base_avx2_4(point, k, 1, 1, heap);
+                RESTORE_VECTOR_REGISTERS();
             }
             else
 #endif
@@ -25148,8 +25214,9 @@ static void sp_256_add_points_4(sp_point_256* p1, const sp_point_256* p2,
 
 #ifdef HAVE_INTEL_AVX2
     if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-            IS_INTEL_AVX2(cpuid_flags)) {
+            IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
         sp_256_proj_point_add_avx2_4(p1, p1, p2, tmp);
+        RESTORE_VECTOR_REGISTERS();
     }
     else
 #endif
@@ -25158,8 +25225,10 @@ static void sp_256_add_points_4(sp_point_256* p1, const sp_point_256* p2,
         if (sp_256_iszero_4(p1->x) && sp_256_iszero_4(p1->y)) {
 #ifdef HAVE_INTEL_AVX2
             if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+                    IS_INTEL_AVX2(cpuid_flags) &&
+                    (SAVE_VECTOR_REGISTERS2() == 0)) {
                 sp_256_proj_point_dbl_avx2_4(p1, p2, tmp);
+                RESTORE_VECTOR_REGISTERS();
             }
             else
 #endif
@@ -25197,8 +25266,9 @@ static int sp_256_calc_vfy_point_4(sp_point_256* p1, sp_point_256* p2,
 #ifndef WOLFSSL_SP_SMALL
 #ifdef HAVE_INTEL_AVX2
     if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-            IS_INTEL_AVX2(cpuid_flags)) {
+            IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
         sp_256_mod_inv_avx2_4(s, s, p256_order);
+        RESTORE_VECTOR_REGISTERS();
     }
     else
 #endif
@@ -25209,8 +25279,9 @@ static int sp_256_calc_vfy_point_4(sp_point_256* p1, sp_point_256* p2,
     {
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_256_mul_avx2_4(s, s, p256_norm_order);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -25224,10 +25295,11 @@ static int sp_256_calc_vfy_point_4(sp_point_256* p1, sp_point_256* p2,
 #ifdef WOLFSSL_SP_SMALL
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_256_mont_inv_order_avx2_4(s, s, tmp);
             sp_256_mont_mul_order_avx2_4(u1, u1, s);
             sp_256_mont_mul_order_avx2_4(u2, u2, s);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -25239,9 +25311,10 @@ static int sp_256_calc_vfy_point_4(sp_point_256* p1, sp_point_256* p2,
 #else
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_256_mont_mul_order_avx2_4(u1, u1, s);
             sp_256_mont_mul_order_avx2_4(u2, u2, s);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -25252,8 +25325,9 @@ static int sp_256_calc_vfy_point_4(sp_point_256* p1, sp_point_256* p2,
 #endif /* WOLFSSL_SP_SMALL */
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_256_ecc_mulmod_base_avx2_4(p1, u1, 0, 0, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -25267,8 +25341,9 @@ static int sp_256_calc_vfy_point_4(sp_point_256* p1, sp_point_256* p2,
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_256_ecc_mulmod_avx2_4(p2, p2, u2, 0, 0, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -25354,16 +25429,18 @@ int sp_ecc_verify_256(const byte* hash, word32 hashLen, const mp_int* pX,
         /* u1 = r.z'.z' mod prime */
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_256_mont_sqr_avx2_4(p1->z, p1->z, p256_mod, p256_mp_mod);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
             sp_256_mont_sqr_4(p1->z, p1->z, p256_mod, p256_mp_mod);
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_256_mont_mul_avx2_4(u1, u2, p1->z, p256_mod, p256_mp_mod);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -25388,9 +25465,11 @@ int sp_ecc_verify_256(const byte* hash, word32 hashLen, const mp_int* pX,
                 /* u1 = (r + 1*order).z'.z' mod prime */
 #ifdef HAVE_INTEL_AVX2
                 if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                        IS_INTEL_AVX2(cpuid_flags)) {
+                        IS_INTEL_AVX2(cpuid_flags) &&
+                        (SAVE_VECTOR_REGISTERS2() == 0)) {
                     sp_256_mont_mul_avx2_4(u1, u2, p1->z, p256_mod,
                         p256_mp_mod);
+                    RESTORE_VECTOR_REGISTERS();
                 }
                 else
 #endif
@@ -25698,8 +25777,9 @@ int sp_ecc_check_key_256(const mp_int* pX, const mp_int* pY,
         /* Point * order = infinity */
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_256_ecc_mulmod_avx2_4(p, pub, p256_order, 1, 1, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -25716,8 +25796,10 @@ int sp_ecc_check_key_256(const mp_int* pX, const mp_int* pY,
             /* Base * private = point */
 #ifdef HAVE_INTEL_AVX2
             if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+                    IS_INTEL_AVX2(cpuid_flags) &&
+                    (SAVE_VECTOR_REGISTERS2() == 0)) {
                 err = sp_256_ecc_mulmod_base_avx2_4(p, priv, 1, 1, heap);
+                RESTORE_VECTOR_REGISTERS();
             }
             else
 #endif
@@ -25782,8 +25864,9 @@ int sp_ecc_proj_add_point_256(mp_int* pX, mp_int* pY, mp_int* pZ,
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_256_proj_point_add_avx2_4(p, p, q, tmp);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -25838,8 +25921,9 @@ int sp_ecc_proj_dbl_point_256(mp_int* pX, mp_int* pY, mp_int* pZ,
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_256_proj_point_dbl_avx2_4(p, p, tmp);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -25891,8 +25975,9 @@ int sp_ecc_map_256(mp_int* pX, mp_int* pY, mp_int* pZ)
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_256_map_avx2_4(p, p, tmp);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -25936,7 +26021,7 @@ static int sp_256_mont_sqrt_4(sp_digit* y)
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             /* t2 = y ^ 0x2 */
             sp_256_mont_sqr_avx2_4(t2, y, p256_mod, p256_mp_mod);
             /* t1 = y ^ 0x3 */
@@ -25966,6 +26051,7 @@ static int sp_256_mont_sqrt_4(sp_digit* y)
             /* t1 = y ^ 0xffffffff00000001000000000000000000000001 */
             sp_256_mont_mul_avx2_4(t1, t1, y, p256_mod, p256_mp_mod);
             sp_256_mont_sqr_n_avx2_4(y, t1, 94, p256_mod, p256_mp_mod);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -26035,9 +26121,10 @@ int sp_ecc_uncompress_256(mp_int* xm, int odd, mp_int* ym)
         /* y = x^3 */
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_256_mont_sqr_avx2_4(y, x, p256_mod, p256_mp_mod);
             sp_256_mont_mul_avx2_4(y, y, x, p256_mod, p256_mp_mod);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -26293,12 +26380,12 @@ static int sp_384_mod_mul_norm_6(sp_digit* r, const sp_digit* a, const sp_digit*
         t[10] += t[9] >> 32; t[9] &= 0xffffffff;
         t[11] += t[10] >> 32; t[10] &= 0xffffffff;
 
-        r[0] = (sp_digit)((t[1] << 32) | t[0]);
-        r[1] = (sp_digit)((t[3] << 32) | t[2]);
-        r[2] = (sp_digit)((t[5] << 32) | t[4]);
-        r[3] = (sp_digit)((t[7] << 32) | t[6]);
-        r[4] = (sp_digit)((t[9] << 32) | t[8]);
-        r[5] = (sp_digit)((t[11] << 32) | t[10]);
+        r[0] = (sp_digit)(((sp_uint64)t[1] << 32) | (sp_uint64)t[0]);
+        r[1] = (sp_digit)(((sp_uint64)t[3] << 32) | (sp_uint64)t[2]);
+        r[2] = (sp_digit)(((sp_uint64)t[5] << 32) | (sp_uint64)t[4]);
+        r[3] = (sp_digit)(((sp_uint64)t[7] << 32) | (sp_uint64)t[6]);
+        r[4] = (sp_digit)(((sp_uint64)t[9] << 32) | (sp_uint64)t[8]);
+        r[5] = (sp_digit)(((sp_uint64)t[11] << 32) | (sp_uint64)t[10]);
     }
 
     SP_FREE_VAR(t, NULL, DYNAMIC_TYPE_ECC);
@@ -26332,7 +26419,7 @@ static void sp_384_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0xffffffffffffffffl;
         s = 64U - s;
         if (j + 1 >= size) {
@@ -26367,7 +26454,7 @@ static void sp_384_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 64) {
             r[j] &= 0xffffffffffffffffl;
             if (j + 1 >= size) {
@@ -26433,7 +26520,7 @@ static int sp_384_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 6; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -26458,7 +26545,7 @@ static int sp_384_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 6; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 64 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -26602,7 +26689,7 @@ static void sp_384_mont_inv_6(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 6);
     for (i=382; i>=0; i--) {
         sp_384_mont_sqr_6(t, t, p384_mod, p384_mp_mod);
-        if (p384_mod_minus_2[i / 64] & ((sp_digit)1 << (i % 64)))
+        if (p384_mod_minus_2[i / 64] & ((sp_uint64)1 << (i % 64)))
             sp_384_mont_mul_6(t, t, a, p384_mod, p384_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 6);
@@ -27579,7 +27666,7 @@ static void sp_384_ecc_recode_6_6(const sp_digit* k, ecc_recode_384* v)
         }
         else if (++j < 6) {
             n = k[j];
-            y |= (word8)((n << (64 - o)) & 0x3f);
+            y |= (word8)(((sp_uint64)n << (64 - o)) & 0x3f);
             o -= 58;
             n >>= o;
         }
@@ -27753,6 +27840,7 @@ extern void sp_384_mont_reduce_order_avx2_6(sp_digit* a, const sp_digit* m, sp_d
 SP_NOINLINE static void sp_384_mont_mul_avx2_6(sp_digit* r, const sp_digit* a,
         const sp_digit* b, const sp_digit* m, sp_digit mp)
 {
+    ASSERT_SAVED_VECTOR_REGISTERS();
     sp_384_mul_avx2_6(r, a, b);
     sp_384_mont_reduce_avx2_6(r, m, mp);
 }
@@ -27769,6 +27857,7 @@ SP_NOINLINE static void sp_384_mont_mul_avx2_6(sp_digit* r, const sp_digit* a,
 SP_NOINLINE static void sp_384_mont_sqr_avx2_6(sp_digit* r, const sp_digit* a,
         const sp_digit* m, sp_digit mp)
 {
+    ASSERT_SAVED_VECTOR_REGISTERS();
     sp_384_sqr_avx2_6(r, a);
     sp_384_mont_reduce_avx2_6(r, m, mp);
 }
@@ -27810,7 +27899,7 @@ static void sp_384_mont_inv_avx2_6(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 6);
     for (i=382; i>=0; i--) {
         sp_384_mont_sqr_avx2_6(t, t, p384_mod, p384_mp_mod);
-        if (p384_mod_minus_2[i / 64] & ((sp_digit)1 << (i % 64)))
+        if (p384_mod_minus_2[i / 64] & ((sp_uint64)1 << (i % 64)))
             sp_384_mont_mul_avx2_6(t, t, a, p384_mod, p384_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 6);
@@ -29625,8 +29714,9 @@ int sp_ecc_mulmod_384(const mp_int* km, const ecc_point* gm, ecc_point* r,
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_384_ecc_mulmod_avx2_6(point, point, k, map, 1, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -29664,6 +29754,7 @@ int sp_ecc_mulmod_add_384(const mp_int* km, const ecc_point* gm,
     int err = MP_OKAY;
 #ifdef HAVE_INTEL_AVX2
     word32 cpuid_flags = cpuid_get_flags();
+    int saved_vector_registers = 0;
 #endif
 
     SP_ALLOC_VAR(sp_point_384, point, 2, heap, DYNAMIC_TYPE_ECC);
@@ -29688,29 +29779,26 @@ int sp_ecc_mulmod_add_384(const mp_int* km, const ecc_point* gm,
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0))
+            saved_vector_registers = 1;
+        if (saved_vector_registers)
             err = sp_384_ecc_mulmod_avx2_6(point, point, k, 0, 0, heap);
-        }
         else
 #endif
             err = sp_384_ecc_mulmod_6(point, point, k, 0, 0, heap);
     }
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_384_proj_point_add_avx2_6(point, point, addP, tmp);
-        }
         else
 #endif
             sp_384_proj_point_add_6(point, point, addP, tmp);
 
         if (map) {
 #ifdef HAVE_INTEL_AVX2
-            if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+            if (saved_vector_registers)
                 sp_384_map_avx2_6(point, point, tmp);
-            }
             else
 #endif
                 sp_384_map_6(point, point, tmp);
@@ -29719,6 +29807,11 @@ int sp_ecc_mulmod_add_384(const mp_int* km, const ecc_point* gm,
         err = sp_384_point_to_ecc_point_6(point, r);
     }
 
+#ifdef HAVE_INTEL_AVX2
+    if (saved_vector_registers)
+        RESTORE_VECTOR_REGISTERS();
+#endif
+
     SP_FREE_VAR(k, heap, DYNAMIC_TYPE_ECC);
     SP_FREE_VAR(point, heap, DYNAMIC_TYPE_ECC);
 
@@ -30157,7 +30250,7 @@ static void sp_384_ecc_recode_7_6(const sp_digit* k, ecc_recode_384* v)
         }
         else if (++j < 6) {
             n = k[j];
-            y |= (word8)((n << (64 - o)) & 0x7f);
+            y |= (word8)(((sp_uint64)n << (64 - o)) & 0x7f);
             o -= 57;
             n >>= o;
         }
@@ -48191,8 +48284,9 @@ int sp_ecc_mulmod_base_384(const mp_int* km, ecc_point* r, int map, void* heap)
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_384_ecc_mulmod_base_avx2_6(point, k, map, 1, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -48229,6 +48323,7 @@ int sp_ecc_mulmod_base_add_384(const mp_int* km, const ecc_point* am,
     int err = MP_OKAY;
 #ifdef HAVE_INTEL_AVX2
     word32 cpuid_flags = cpuid_get_flags();
+    int saved_vector_registers = 0;
 #endif
 
     SP_ALLOC_VAR(sp_point_384, point, 2, NULL, DYNAMIC_TYPE_ECC);
@@ -48252,29 +48347,26 @@ int sp_ecc_mulmod_base_add_384(const mp_int* km, const ecc_point* am,
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0))
+            saved_vector_registers = 1;
+        if (saved_vector_registers)
             err = sp_384_ecc_mulmod_base_avx2_6(point, k, 0, 0, heap);
-        }
         else
 #endif
             err = sp_384_ecc_mulmod_base_6(point, k, 0, 0, heap);
     }
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_384_proj_point_add_avx2_6(point, point, addP, tmp);
-        }
         else
 #endif
             sp_384_proj_point_add_6(point, point, addP, tmp);
 
         if (map) {
 #ifdef HAVE_INTEL_AVX2
-            if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+            if (saved_vector_registers)
                 sp_384_map_avx2_6(point, point, tmp);
-            }
             else
 #endif
                 sp_384_map_6(point, point, tmp);
@@ -48283,6 +48375,11 @@ int sp_ecc_mulmod_base_add_384(const mp_int* km, const ecc_point* am,
         err = sp_384_point_to_ecc_point_6(point, r);
     }
 
+#ifdef HAVE_INTEL_AVX2
+    if (saved_vector_registers)
+        RESTORE_VECTOR_REGISTERS();
+#endif
+
     SP_FREE_VAR(k, NULL, DYNAMIC_TYPE_ECC);
     SP_FREE_VAR(point, NULL, DYNAMIC_TYPE_ECC);
 
@@ -48394,6 +48491,7 @@ int sp_ecc_make_key_384(WC_RNG* rng, mp_int* priv, ecc_point* pub, void* heap)
 
 #ifdef HAVE_INTEL_AVX2
     word32 cpuid_flags = cpuid_get_flags();
+    int saved_vector_registers = 0;
 #endif
 
     (void)heap;
@@ -48414,9 +48512,11 @@ int sp_ecc_make_key_384(WC_RNG* rng, mp_int* priv, ecc_point* pub, void* heap)
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0))
+            saved_vector_registers = 1;
+
+        if (saved_vector_registers)
             err = sp_384_ecc_mulmod_base_avx2_6(point, k, 1, 1, NULL);
-       }
         else
 #endif
             err = sp_384_ecc_mulmod_base_6(point, k, 1, 1, NULL);
@@ -48425,8 +48525,7 @@ int sp_ecc_make_key_384(WC_RNG* rng, mp_int* priv, ecc_point* pub, void* heap)
 #ifdef WOLFSSL_VALIDATE_ECC_KEYGEN
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers) {
             err = sp_384_ecc_mulmod_avx2_6(infinity, point, p384_order, 1, 1,
                                                                           NULL);
         }
@@ -48441,6 +48540,11 @@ int sp_ecc_make_key_384(WC_RNG* rng, mp_int* priv, ecc_point* pub, void* heap)
     }
 #endif
 
+#ifdef HAVE_INTEL_AVX2
+    if (saved_vector_registers)
+        RESTORE_VECTOR_REGISTERS();
+#endif
+
     if (err == MP_OKAY) {
         err = sp_384_to_mp(k, priv);
     }
@@ -48602,8 +48706,9 @@ int sp_ecc_secret_gen_384(const mp_int* priv, const ecc_point* pub, byte* out,
         sp_384_point_from_ecc_point_6(point, pub);
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_384_ecc_mulmod_avx2_6(point, point, k, 1, 1, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -48715,7 +48820,6 @@ extern sp_digit div_384_word_asm_6(sp_digit d1, sp_digit d0, sp_digit div);
 static WC_INLINE sp_digit div_384_word_6(sp_digit d1, sp_digit d0,
         sp_digit div)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
 #if _MSC_VER >= 1920
     return _udiv128(d1, d0, div, NULL);
 #else
@@ -48734,7 +48838,6 @@ static WC_INLINE sp_digit div_384_word_6(sp_digit d1, sp_digit d0,
         sp_digit div)
 {
     register sp_digit r asm("rax");
-    ASSERT_SAVED_VECTOR_REGISTERS();
     __asm__ __volatile__ (
         "divq %3"
         : "=a" (r)
@@ -48789,8 +48892,6 @@ static WC_INLINE int sp_384_div_6(const sp_digit* a, const sp_digit* d, sp_digit
     word32 cpuid_flags = cpuid_get_flags();
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)m;
 
     div = d[5];
@@ -48843,7 +48944,6 @@ static WC_INLINE int sp_384_div_6(const sp_digit* a, const sp_digit* d, sp_digit
 static WC_INLINE int sp_384_mod_6(sp_digit* r, const sp_digit* a,
         const sp_digit* m)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
     return sp_384_div_6(a, m, NULL, r);
 }
 
@@ -49201,14 +49301,16 @@ static int sp_384_calc_s_6(sp_digit* s, const sp_digit* r, sp_digit* k,
     sp_digit* kInv = k;
 #ifdef HAVE_INTEL_AVX2
     word32 cpuid_flags = cpuid_get_flags();
+    int saved_vector_registers = 0;
 #endif
 
     /* Conv k to Montgomery form (mod order) */
 #ifdef HAVE_INTEL_AVX2
     if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-            IS_INTEL_AVX2(cpuid_flags)) {
+            IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0))
+        saved_vector_registers = 1;
+    if (saved_vector_registers)
         sp_384_mul_avx2_6(k, k, p384_norm_order);
-    }
     else
 #endif
         sp_384_mul_6(k, k, p384_norm_order);
@@ -49218,10 +49320,8 @@ static int sp_384_calc_s_6(sp_digit* s, const sp_digit* r, sp_digit* k,
 
         /* kInv = 1/k mod order */
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_384_mont_inv_order_avx2_6(kInv, k, tmp);
-        }
         else
 #endif
             sp_384_mont_inv_order_6(kInv, k, tmp);
@@ -49229,10 +49329,8 @@ static int sp_384_calc_s_6(sp_digit* s, const sp_digit* r, sp_digit* k,
 
         /* s = r * x + e */
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_384_mul_avx2_6(x, x, r);
-        }
         else
 #endif
             sp_384_mul_6(x, x, r);
@@ -49250,16 +49348,19 @@ static int sp_384_calc_s_6(sp_digit* s, const sp_digit* r, sp_digit* k,
 
         /* s = s * k^-1 mod order */
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_384_mont_mul_order_avx2_6(s, s, kInv);
-        }
         else
 #endif
             sp_384_mont_mul_order_6(s, s, kInv);
         sp_384_norm_6(s);
     }
 
+#ifdef HAVE_INTEL_AVX2
+    if (saved_vector_registers)
+        RESTORE_VECTOR_REGISTERS();
+#endif
+
     return err;
 }
 
@@ -49324,8 +49425,10 @@ int sp_ecc_sign_384(const byte* hash, word32 hashLen, WC_RNG* rng,
         if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
             if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+                    IS_INTEL_AVX2(cpuid_flags) &&
+                    (SAVE_VECTOR_REGISTERS2() == 0)) {
                 err = sp_384_ecc_mulmod_base_avx2_6(point, k, 1, 1, heap);
+                RESTORE_VECTOR_REGISTERS();
             }
             else
 #endif
@@ -49662,8 +49765,9 @@ static void sp_384_add_points_6(sp_point_384* p1, const sp_point_384* p2,
 
 #ifdef HAVE_INTEL_AVX2
     if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-            IS_INTEL_AVX2(cpuid_flags)) {
+            IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
         sp_384_proj_point_add_avx2_6(p1, p1, p2, tmp);
+        RESTORE_VECTOR_REGISTERS();
     }
     else
 #endif
@@ -49672,8 +49776,10 @@ static void sp_384_add_points_6(sp_point_384* p1, const sp_point_384* p2,
         if (sp_384_iszero_6(p1->x) && sp_384_iszero_6(p1->y)) {
 #ifdef HAVE_INTEL_AVX2
             if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+                    IS_INTEL_AVX2(cpuid_flags) &&
+                    (SAVE_VECTOR_REGISTERS2() == 0)) {
                 sp_384_proj_point_dbl_avx2_6(p1, p2, tmp);
+                RESTORE_VECTOR_REGISTERS();
             }
             else
 #endif
@@ -49717,8 +49823,9 @@ static int sp_384_calc_vfy_point_6(sp_point_384* p1, sp_point_384* p2,
     {
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_384_mul_avx2_6(s, s, p384_norm_order);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -49732,10 +49839,11 @@ static int sp_384_calc_vfy_point_6(sp_point_384* p1, sp_point_384* p2,
 #ifdef WOLFSSL_SP_SMALL
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_384_mont_inv_order_avx2_6(s, s, tmp);
             sp_384_mont_mul_order_avx2_6(u1, u1, s);
             sp_384_mont_mul_order_avx2_6(u2, u2, s);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -49747,9 +49855,10 @@ static int sp_384_calc_vfy_point_6(sp_point_384* p1, sp_point_384* p2,
 #else
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_384_mont_mul_order_avx2_6(u1, u1, s);
             sp_384_mont_mul_order_avx2_6(u2, u2, s);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -49760,8 +49869,9 @@ static int sp_384_calc_vfy_point_6(sp_point_384* p1, sp_point_384* p2,
 #endif /* WOLFSSL_SP_SMALL */
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_384_ecc_mulmod_base_avx2_6(p1, u1, 0, 0, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -49775,8 +49885,9 @@ static int sp_384_calc_vfy_point_6(sp_point_384* p1, sp_point_384* p2,
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_384_ecc_mulmod_avx2_6(p2, p2, u2, 0, 0, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -49862,16 +49973,18 @@ int sp_ecc_verify_384(const byte* hash, word32 hashLen, const mp_int* pX,
         /* u1 = r.z'.z' mod prime */
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_384_mont_sqr_avx2_6(p1->z, p1->z, p384_mod, p384_mp_mod);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
             sp_384_mont_sqr_6(p1->z, p1->z, p384_mod, p384_mp_mod);
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_384_mont_mul_avx2_6(u1, u2, p1->z, p384_mod, p384_mp_mod);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -49896,9 +50009,11 @@ int sp_ecc_verify_384(const byte* hash, word32 hashLen, const mp_int* pX,
                 /* u1 = (r + 1*order).z'.z' mod prime */
 #ifdef HAVE_INTEL_AVX2
                 if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                        IS_INTEL_AVX2(cpuid_flags)) {
+                        IS_INTEL_AVX2(cpuid_flags) &&
+                        (SAVE_VECTOR_REGISTERS2() == 0)) {
                     sp_384_mont_mul_avx2_6(u1, u2, p1->z, p384_mod,
                         p384_mp_mod);
+                    RESTORE_VECTOR_REGISTERS();
                 }
                 else
 #endif
@@ -50206,8 +50321,9 @@ int sp_ecc_check_key_384(const mp_int* pX, const mp_int* pY,
         /* Point * order = infinity */
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_384_ecc_mulmod_avx2_6(p, pub, p384_order, 1, 1, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -50224,8 +50340,10 @@ int sp_ecc_check_key_384(const mp_int* pX, const mp_int* pY,
             /* Base * private = point */
 #ifdef HAVE_INTEL_AVX2
             if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+                    IS_INTEL_AVX2(cpuid_flags) &&
+                    (SAVE_VECTOR_REGISTERS2() == 0)) {
                 err = sp_384_ecc_mulmod_base_avx2_6(p, priv, 1, 1, heap);
+                RESTORE_VECTOR_REGISTERS();
             }
             else
 #endif
@@ -50290,8 +50408,9 @@ int sp_ecc_proj_add_point_384(mp_int* pX, mp_int* pY, mp_int* pZ,
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_384_proj_point_add_avx2_6(p, p, q, tmp);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -50346,8 +50465,9 @@ int sp_ecc_proj_dbl_point_384(mp_int* pX, mp_int* pY, mp_int* pZ,
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_384_proj_point_dbl_avx2_6(p, p, tmp);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -50399,8 +50519,9 @@ int sp_ecc_map_384(mp_int* pX, mp_int* pY, mp_int* pZ)
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_384_map_avx2_6(p, p, tmp);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -50450,7 +50571,7 @@ static int sp_384_mont_sqrt_6(sp_digit* y)
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             /* t2 = y ^ 0x2 */
             sp_384_mont_sqr_avx2_6(t2, y, p384_mod, p384_mp_mod);
             /* t1 = y ^ 0x3 */
@@ -50505,6 +50626,7 @@ static int sp_384_mont_sqrt_6(sp_digit* y)
             sp_384_mont_mul_avx2_6(t1, y, t2, p384_mod, p384_mp_mod);
             /* t2 = y ^ 0x3fffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffbfffffffc00000000000000040000000 */
             sp_384_mont_sqr_n_avx2_6(y, t1, 30, p384_mod, p384_mp_mod);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -50599,9 +50721,10 @@ int sp_ecc_uncompress_384(mp_int* xm, int odd, mp_int* ym)
         /* y = x^3 */
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_384_mont_sqr_avx2_6(y, x, p384_mod, p384_mp_mod);
             sp_384_mont_mul_avx2_6(y, y, x, p384_mod, p384_mp_mod);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -50824,7 +50947,7 @@ static void sp_521_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0xffffffffffffffffl;
         s = 64U - s;
         if (j + 1 >= size) {
@@ -50859,7 +50982,7 @@ static void sp_521_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 64) {
             r[j] &= 0xffffffffffffffffl;
             if (j + 1 >= size) {
@@ -50925,7 +51048,7 @@ static int sp_521_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 9; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -50950,7 +51073,7 @@ static int sp_521_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 9; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 64 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -51058,7 +51181,7 @@ static void sp_521_mont_inv_9(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 9);
     for (i=519; i>=0; i--) {
         sp_521_mont_sqr_9(t, t, p521_mod, p521_mp_mod);
-        if (p521_mod_minus_2[i / 64] & ((sp_digit)1 << (i % 64)))
+        if (p521_mod_minus_2[i / 64] & ((sp_uint64)1 << (i % 64)))
             sp_521_mont_mul_9(t, t, a, p521_mod, p521_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 9);
@@ -52054,7 +52177,7 @@ static void sp_521_ecc_recode_6_9(const sp_digit* k, ecc_recode_521* v)
         }
         else if (++j < 9) {
             n = k[j];
-            y |= (word8)((n << (64 - o)) & 0x3f);
+            y |= (word8)(((sp_uint64)n << (64 - o)) & 0x3f);
             o -= 58;
             n >>= o;
         }
@@ -52257,7 +52380,7 @@ static void sp_521_mont_inv_avx2_9(sp_digit* r, const sp_digit* a, sp_digit* td)
     XMEMCPY(t, a, sizeof(sp_digit) * 9);
     for (i=519; i>=0; i--) {
         sp_521_mont_sqr_avx2_9(t, t, p521_mod, p521_mp_mod);
-        if (p521_mod_minus_2[i / 64] & ((sp_digit)1 << (i % 64)))
+        if (p521_mod_minus_2[i / 64] & ((sp_uint64)1 << (i % 64)))
             sp_521_mont_mul_avx2_9(t, t, a, p521_mod, p521_mp_mod);
     }
     XMEMCPY(r, t, sizeof(sp_digit) * 9);
@@ -54077,8 +54200,9 @@ int sp_ecc_mulmod_521(const mp_int* km, const ecc_point* gm, ecc_point* r,
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_521_ecc_mulmod_avx2_9(point, point, k, map, 1, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -54116,6 +54240,7 @@ int sp_ecc_mulmod_add_521(const mp_int* km, const ecc_point* gm,
     int err = MP_OKAY;
 #ifdef HAVE_INTEL_AVX2
     word32 cpuid_flags = cpuid_get_flags();
+    int saved_vector_registers = 0;
 #endif
 
     SP_ALLOC_VAR(sp_point_521, point, 2, heap, DYNAMIC_TYPE_ECC);
@@ -54140,29 +54265,26 @@ int sp_ecc_mulmod_add_521(const mp_int* km, const ecc_point* gm,
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0))
+            saved_vector_registers = 1;
+        if (saved_vector_registers)
             err = sp_521_ecc_mulmod_avx2_9(point, point, k, 0, 0, heap);
-        }
         else
 #endif
             err = sp_521_ecc_mulmod_9(point, point, k, 0, 0, heap);
     }
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_521_proj_point_add_avx2_9(point, point, addP, tmp);
-        }
         else
 #endif
             sp_521_proj_point_add_9(point, point, addP, tmp);
 
         if (map) {
 #ifdef HAVE_INTEL_AVX2
-            if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+            if (saved_vector_registers)
                 sp_521_map_avx2_9(point, point, tmp);
-            }
             else
 #endif
                 sp_521_map_9(point, point, tmp);
@@ -54171,6 +54293,11 @@ int sp_ecc_mulmod_add_521(const mp_int* km, const ecc_point* gm,
         err = sp_521_point_to_ecc_point_9(point, r);
     }
 
+#ifdef HAVE_INTEL_AVX2
+    if (saved_vector_registers)
+        RESTORE_VECTOR_REGISTERS();
+#endif
+
     SP_FREE_VAR(k, heap, DYNAMIC_TYPE_ECC);
     SP_FREE_VAR(point, heap, DYNAMIC_TYPE_ECC);
 
@@ -54735,7 +54862,7 @@ static void sp_521_ecc_recode_7_9(const sp_digit* k, ecc_recode_521* v)
         }
         else if (++j < 9) {
             n = k[j];
-            y |= (word8)((n << (64 - o)) & 0x7f);
+            y |= (word8)(((sp_uint64)n << (64 - o)) & 0x7f);
             o -= 57;
             n >>= o;
         }
@@ -88829,8 +88956,9 @@ int sp_ecc_mulmod_base_521(const mp_int* km, ecc_point* r, int map, void* heap)
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_521_ecc_mulmod_base_avx2_9(point, k, map, 1, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -88867,6 +88995,7 @@ int sp_ecc_mulmod_base_add_521(const mp_int* km, const ecc_point* am,
     int err = MP_OKAY;
 #ifdef HAVE_INTEL_AVX2
     word32 cpuid_flags = cpuid_get_flags();
+    int saved_vector_registers = 0;
 #endif
 
     SP_ALLOC_VAR(sp_point_521, point, 2, NULL, DYNAMIC_TYPE_ECC);
@@ -88890,29 +89019,26 @@ int sp_ecc_mulmod_base_add_521(const mp_int* km, const ecc_point* am,
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0))
+            saved_vector_registers = 1;
+        if (saved_vector_registers)
             err = sp_521_ecc_mulmod_base_avx2_9(point, k, 0, 0, heap);
-        }
         else
 #endif
             err = sp_521_ecc_mulmod_base_9(point, k, 0, 0, heap);
     }
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_521_proj_point_add_avx2_9(point, point, addP, tmp);
-        }
         else
 #endif
             sp_521_proj_point_add_9(point, point, addP, tmp);
 
         if (map) {
 #ifdef HAVE_INTEL_AVX2
-            if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+            if (saved_vector_registers)
                 sp_521_map_avx2_9(point, point, tmp);
-            }
             else
 #endif
                 sp_521_map_9(point, point, tmp);
@@ -88921,6 +89047,11 @@ int sp_ecc_mulmod_base_add_521(const mp_int* km, const ecc_point* am,
         err = sp_521_point_to_ecc_point_9(point, r);
     }
 
+#ifdef HAVE_INTEL_AVX2
+    if (saved_vector_registers)
+        RESTORE_VECTOR_REGISTERS();
+#endif
+
     SP_FREE_VAR(k, NULL, DYNAMIC_TYPE_ECC);
     SP_FREE_VAR(point, NULL, DYNAMIC_TYPE_ECC);
 
@@ -89033,6 +89164,7 @@ int sp_ecc_make_key_521(WC_RNG* rng, mp_int* priv, ecc_point* pub, void* heap)
 
 #ifdef HAVE_INTEL_AVX2
     word32 cpuid_flags = cpuid_get_flags();
+    int saved_vector_registers = 0;
 #endif
 
     (void)heap;
@@ -89053,9 +89185,11 @@ int sp_ecc_make_key_521(WC_RNG* rng, mp_int* priv, ecc_point* pub, void* heap)
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0))
+            saved_vector_registers = 1;
+
+        if (saved_vector_registers)
             err = sp_521_ecc_mulmod_base_avx2_9(point, k, 1, 1, NULL);
-       }
         else
 #endif
             err = sp_521_ecc_mulmod_base_9(point, k, 1, 1, NULL);
@@ -89064,8 +89198,7 @@ int sp_ecc_make_key_521(WC_RNG* rng, mp_int* priv, ecc_point* pub, void* heap)
 #ifdef WOLFSSL_VALIDATE_ECC_KEYGEN
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers) {
             err = sp_521_ecc_mulmod_avx2_9(infinity, point, p521_order, 1, 1,
                                                                           NULL);
         }
@@ -89080,6 +89213,11 @@ int sp_ecc_make_key_521(WC_RNG* rng, mp_int* priv, ecc_point* pub, void* heap)
     }
 #endif
 
+#ifdef HAVE_INTEL_AVX2
+    if (saved_vector_registers)
+        RESTORE_VECTOR_REGISTERS();
+#endif
+
     if (err == MP_OKAY) {
         err = sp_521_to_mp(k, priv);
     }
@@ -89241,8 +89379,9 @@ int sp_ecc_secret_gen_521(const mp_int* priv, const ecc_point* pub, byte* out,
         sp_521_point_from_ecc_point_9(point, pub);
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_521_ecc_mulmod_avx2_9(point, point, k, 1, 1, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -89375,7 +89514,6 @@ extern sp_digit div_521_word_asm_9(sp_digit d1, sp_digit d0, sp_digit div);
 static WC_INLINE sp_digit div_521_word_9(sp_digit d1, sp_digit d0,
         sp_digit div)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
 #if _MSC_VER >= 1920
     return _udiv128(d1, d0, div, NULL);
 #else
@@ -89394,7 +89532,6 @@ static WC_INLINE sp_digit div_521_word_9(sp_digit d1, sp_digit d0,
         sp_digit div)
 {
     register sp_digit r asm("rax");
-    ASSERT_SAVED_VECTOR_REGISTERS();
     __asm__ __volatile__ (
         "divq %3"
         : "=a" (r)
@@ -89453,8 +89590,6 @@ static WC_INLINE int sp_521_div_9(const sp_digit* a, const sp_digit* d, sp_digit
     word32 cpuid_flags = cpuid_get_flags();
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)m;
 
     div = (d[8] << 55) | (d[7] >> 9);
@@ -89509,7 +89644,6 @@ static WC_INLINE int sp_521_div_9(const sp_digit* a, const sp_digit* d, sp_digit
 static WC_INLINE int sp_521_mod_9(sp_digit* r, const sp_digit* a,
         const sp_digit* m)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
     return sp_521_div_9(a, m, NULL, r);
 }
 
@@ -89895,14 +90029,16 @@ static int sp_521_calc_s_9(sp_digit* s, const sp_digit* r, sp_digit* k,
     sp_digit* kInv = k;
 #ifdef HAVE_INTEL_AVX2
     word32 cpuid_flags = cpuid_get_flags();
+    int saved_vector_registers = 0;
 #endif
 
     /* Conv k to Montgomery form (mod order) */
 #ifdef HAVE_INTEL_AVX2
     if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-            IS_INTEL_AVX2(cpuid_flags)) {
+            IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0))
+        saved_vector_registers = 1;
+    if (saved_vector_registers)
         sp_521_mul_avx2_9(k, k, p521_norm_order);
-    }
     else
 #endif
         sp_521_mul_9(k, k, p521_norm_order);
@@ -89912,10 +90048,8 @@ static int sp_521_calc_s_9(sp_digit* s, const sp_digit* r, sp_digit* k,
 
         /* kInv = 1/k mod order */
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_521_mont_inv_order_avx2_9(kInv, k, tmp);
-        }
         else
 #endif
             sp_521_mont_inv_order_9(kInv, k, tmp);
@@ -89923,10 +90057,8 @@ static int sp_521_calc_s_9(sp_digit* s, const sp_digit* r, sp_digit* k,
 
         /* s = r * x + e */
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_521_mul_avx2_9(x, x, r);
-        }
         else
 #endif
             sp_521_mul_9(x, x, r);
@@ -89944,16 +90076,19 @@ static int sp_521_calc_s_9(sp_digit* s, const sp_digit* r, sp_digit* k,
 
         /* s = s * k^-1 mod order */
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_521_mont_mul_order_avx2_9(s, s, kInv);
-        }
         else
 #endif
             sp_521_mont_mul_order_9(s, s, kInv);
         sp_521_norm_9(s);
     }
 
+#ifdef HAVE_INTEL_AVX2
+    if (saved_vector_registers)
+        RESTORE_VECTOR_REGISTERS();
+#endif
+
     return err;
 }
 
@@ -90018,8 +90153,10 @@ int sp_ecc_sign_521(const byte* hash, word32 hashLen, WC_RNG* rng,
         if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
             if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+                    IS_INTEL_AVX2(cpuid_flags) &&
+                    (SAVE_VECTOR_REGISTERS2() == 0)) {
                 err = sp_521_ecc_mulmod_base_avx2_9(point, k, 1, 1, heap);
+                RESTORE_VECTOR_REGISTERS();
             }
             else
 #endif
@@ -90364,8 +90501,9 @@ static void sp_521_add_points_9(sp_point_521* p1, const sp_point_521* p2,
 
 #ifdef HAVE_INTEL_AVX2
     if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-            IS_INTEL_AVX2(cpuid_flags)) {
+            IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
         sp_521_proj_point_add_avx2_9(p1, p1, p2, tmp);
+        RESTORE_VECTOR_REGISTERS();
     }
     else
 #endif
@@ -90374,8 +90512,10 @@ static void sp_521_add_points_9(sp_point_521* p1, const sp_point_521* p2,
         if (sp_521_iszero_9(p1->x) && sp_521_iszero_9(p1->y)) {
 #ifdef HAVE_INTEL_AVX2
             if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+                    IS_INTEL_AVX2(cpuid_flags) &&
+                    (SAVE_VECTOR_REGISTERS2() == 0)) {
                 sp_521_proj_point_dbl_avx2_9(p1, p2, tmp);
+                RESTORE_VECTOR_REGISTERS();
             }
             else
 #endif
@@ -90422,8 +90562,9 @@ static int sp_521_calc_vfy_point_9(sp_point_521* p1, sp_point_521* p2,
     {
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_521_mul_avx2_9(s, s, p521_norm_order);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -90437,10 +90578,11 @@ static int sp_521_calc_vfy_point_9(sp_point_521* p1, sp_point_521* p2,
 #ifdef WOLFSSL_SP_SMALL
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_521_mont_inv_order_avx2_9(s, s, tmp);
             sp_521_mont_mul_order_avx2_9(u1, u1, s);
             sp_521_mont_mul_order_avx2_9(u2, u2, s);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -90452,9 +90594,10 @@ static int sp_521_calc_vfy_point_9(sp_point_521* p1, sp_point_521* p2,
 #else
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_521_mont_mul_order_avx2_9(u1, u1, s);
             sp_521_mont_mul_order_avx2_9(u2, u2, s);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -90465,8 +90608,9 @@ static int sp_521_calc_vfy_point_9(sp_point_521* p1, sp_point_521* p2,
 #endif /* WOLFSSL_SP_SMALL */
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_521_ecc_mulmod_base_avx2_9(p1, u1, 0, 0, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -90480,8 +90624,9 @@ static int sp_521_calc_vfy_point_9(sp_point_521* p1, sp_point_521* p2,
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_521_ecc_mulmod_avx2_9(p2, p2, u2, 0, 0, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -90571,16 +90716,18 @@ int sp_ecc_verify_521(const byte* hash, word32 hashLen, const mp_int* pX,
         /* u1 = r.z'.z' mod prime */
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_521_mont_sqr_avx2_9(p1->z, p1->z, p521_mod, p521_mp_mod);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
             sp_521_mont_sqr_9(p1->z, p1->z, p521_mod, p521_mp_mod);
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_521_mont_mul_avx2_9(u1, u2, p1->z, p521_mod, p521_mp_mod);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -90605,9 +90752,11 @@ int sp_ecc_verify_521(const byte* hash, word32 hashLen, const mp_int* pX,
                 /* u1 = (r + 1*order).z'.z' mod prime */
 #ifdef HAVE_INTEL_AVX2
                 if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                        IS_INTEL_AVX2(cpuid_flags)) {
+                        IS_INTEL_AVX2(cpuid_flags) &&
+                        (SAVE_VECTOR_REGISTERS2() == 0)) {
                     sp_521_mont_mul_avx2_9(u1, u2, p1->z, p521_mod,
                         p521_mp_mod);
+                    RESTORE_VECTOR_REGISTERS();
                 }
                 else
 #endif
@@ -90918,8 +91067,9 @@ int sp_ecc_check_key_521(const mp_int* pX, const mp_int* pY,
         /* Point * order = infinity */
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_521_ecc_mulmod_avx2_9(p, pub, p521_order, 1, 1, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -90936,8 +91086,10 @@ int sp_ecc_check_key_521(const mp_int* pX, const mp_int* pY,
             /* Base * private = point */
 #ifdef HAVE_INTEL_AVX2
             if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+                    IS_INTEL_AVX2(cpuid_flags) &&
+                    (SAVE_VECTOR_REGISTERS2() == 0)) {
                 err = sp_521_ecc_mulmod_base_avx2_9(p, priv, 1, 1, heap);
+                RESTORE_VECTOR_REGISTERS();
             }
             else
 #endif
@@ -91002,8 +91154,9 @@ int sp_ecc_proj_add_point_521(mp_int* pX, mp_int* pY, mp_int* pZ,
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_521_proj_point_add_avx2_9(p, p, q, tmp);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -91058,8 +91211,9 @@ int sp_ecc_proj_dbl_point_521(mp_int* pX, mp_int* pY, mp_int* pZ,
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_521_proj_point_dbl_avx2_9(p, p, tmp);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -91111,8 +91265,9 @@ int sp_ecc_map_521(mp_int* pX, mp_int* pY, mp_int* pZ)
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_521_map_avx2_9(p, p, tmp);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -91161,16 +91316,17 @@ static int sp_521_mont_sqrt_9(sp_digit* y)
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             int i;
 
             XMEMCPY(t, y, sizeof(sp_digit) * 9);
             for (i=518; i>=0; i--) {
                 sp_521_mont_sqr_avx2_9(t, t, p521_mod, p521_mp_mod);
-                if (p521_sqrt_power[i / 64] & ((sp_digit)1 << (i % 64)))
+                if (p521_sqrt_power[i / 64] & ((sp_uint64)1 << (i % 64)))
                     sp_521_mont_mul_avx2_9(t, t, y, p521_mod, p521_mp_mod);
             }
             XMEMCPY(y, t, sizeof(sp_digit) * 9);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -91180,7 +91336,7 @@ static int sp_521_mont_sqrt_9(sp_digit* y)
             XMEMCPY(t, y, sizeof(sp_digit) * 9);
             for (i=518; i>=0; i--) {
                 sp_521_mont_sqr_9(t, t, p521_mod, p521_mp_mod);
-                if (p521_sqrt_power[i / 64] & ((sp_digit)1 << (i % 64)))
+                if (p521_sqrt_power[i / 64] & ((sp_uint64)1 << (i % 64)))
                     sp_521_mont_mul_9(t, t, y, p521_mod, p521_mp_mod);
             }
             XMEMCPY(y, t, sizeof(sp_digit) * 9);
@@ -91220,9 +91376,10 @@ int sp_ecc_uncompress_521(mp_int* xm, int odd, mp_int* ym)
         /* y = x^3 */
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             sp_521_mont_sqr_avx2_9(y, x, p521_mod, p521_mp_mod);
             sp_521_mont_mul_avx2_9(y, y, x, p521_mod, p521_mp_mod);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -91439,7 +91596,6 @@ extern sp_digit div_1024_word_asm_16(sp_digit d1, sp_digit d0, sp_digit div);
 static WC_INLINE sp_digit div_1024_word_16(sp_digit d1, sp_digit d0,
         sp_digit div)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
 #if _MSC_VER >= 1920
     return _udiv128(d1, d0, div, NULL);
 #else
@@ -91458,7 +91614,6 @@ static WC_INLINE sp_digit div_1024_word_16(sp_digit d1, sp_digit d0,
         sp_digit div)
 {
     register sp_digit r asm("rax");
-    ASSERT_SAVED_VECTOR_REGISTERS();
     __asm__ __volatile__ (
         "divq %3"
         : "=a" (r)
@@ -91526,8 +91681,6 @@ static WC_INLINE int sp_1024_div_16(const sp_digit* a, const sp_digit* d, sp_dig
     word32 cpuid_flags = cpuid_get_flags();
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
     (void)m;
 
     div = d[15];
@@ -91580,7 +91733,6 @@ static WC_INLINE int sp_1024_div_16(const sp_digit* a, const sp_digit* d, sp_dig
 static WC_INLINE int sp_1024_mod_16(sp_digit* r, const sp_digit* a,
         const sp_digit* m)
 {
-    ASSERT_SAVED_VECTOR_REGISTERS();
     return sp_1024_div_16(a, m, NULL, r);
 }
 
@@ -91702,7 +91854,7 @@ static void sp_1024_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i] << s);
+        r[j] |= ((sp_uint64)a->dp[i] << s);
         r[j] &= 0xffffffffffffffffl;
         s = 64U - s;
         if (j + 1 >= size) {
@@ -91737,7 +91889,7 @@ static void sp_1024_from_mp(sp_digit* r, int size, const mp_int* a)
 
     r[0] = 0;
     for (i = 0; i < (unsigned int)a->used && j < size; i++) {
-        r[j] |= ((sp_digit)a->dp[i]) << s;
+        r[j] |= ((sp_uint64)a->dp[i]) << s;
         if (s + DIGIT_BIT >= 64) {
             r[j] &= 0xffffffffffffffffl;
             if (j + 1 >= size) {
@@ -91803,7 +91955,7 @@ static int sp_1024_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 16; i++) {
-            r->dp[j] |= (mp_digit)(a[i] << s);
+            r->dp[j] |= (mp_digit)((sp_uint64)a[i] << s);
             r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
             s = DIGIT_BIT - s;
             r->dp[++j] = (mp_digit)(a[i] >> s);
@@ -91828,7 +91980,7 @@ static int sp_1024_to_mp(const sp_digit* a, mp_int* r)
 
         r->dp[0] = 0;
         for (i = 0; i < 16; i++) {
-            r->dp[j] |= ((mp_digit)a[i]) << s;
+            r->dp[j] |= ((sp_uint64)a[i]) << s;
             if (s + 64 >= DIGIT_BIT) {
     #if DIGIT_BIT != 32 && DIGIT_BIT != 64
                 r->dp[j] &= ((sp_digit)1 << DIGIT_BIT) - 1;
@@ -92890,7 +93042,7 @@ static void sp_1024_ecc_recode_7_16(const sp_digit* k, ecc_recode_1024* v)
         }
         else if (++j < 16) {
             n = k[j];
-            y |= (word8)((n << (64 - o)) & 0x7f);
+            y |= (word8)(((sp_uint64)n << (64 - o)) & 0x7f);
             o -= 57;
             n >>= o;
         }
@@ -93052,6 +93204,7 @@ extern void sp_1024_mont_reduce_avx2_16(sp_digit* a, const sp_digit* m, sp_digit
 SP_NOINLINE static void sp_1024_mont_mul_avx2_16(sp_digit* r, const sp_digit* a,
         const sp_digit* b, const sp_digit* m, sp_digit mp)
 {
+    ASSERT_SAVED_VECTOR_REGISTERS();
     sp_1024_mul_avx2_16(r, a, b);
     sp_1024_mont_reduce_avx2_16(r, m, mp);
 }
@@ -93068,6 +93221,7 @@ SP_NOINLINE static void sp_1024_mont_mul_avx2_16(sp_digit* r, const sp_digit* a,
 SP_NOINLINE static void sp_1024_mont_sqr_avx2_16(sp_digit* r, const sp_digit* a,
         const sp_digit* m, sp_digit mp)
 {
+    ASSERT_SAVED_VECTOR_REGISTERS();
     sp_1024_sqr_avx2_16(r, a);
     sp_1024_mont_reduce_avx2_16(r, m, mp);
 }
@@ -94831,8 +94985,9 @@ int sp_ecc_mulmod_1024(const mp_int* km, const ecc_point* gm, ecc_point* r,
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_1024_ecc_mulmod_avx2_16(point, point, k, map, 1, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -98246,8 +98401,9 @@ int sp_ecc_mulmod_base_1024(const mp_int* km, ecc_point* r, int map, void* heap)
 
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_1024_ecc_mulmod_base_avx2_16(point, k, map, 1, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -98284,6 +98440,7 @@ int sp_ecc_mulmod_base_add_1024(const mp_int* km, const ecc_point* am,
     int err = MP_OKAY;
 #ifdef HAVE_INTEL_AVX2
     word32 cpuid_flags = cpuid_get_flags();
+    int saved_vector_registers = 0;
 #endif
 
     SP_ALLOC_VAR(sp_point_1024, point, 2, NULL, DYNAMIC_TYPE_ECC);
@@ -98307,29 +98464,26 @@ int sp_ecc_mulmod_base_add_1024(const mp_int* km, const ecc_point* am,
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0))
+            saved_vector_registers = 1;
+        if (saved_vector_registers)
             err = sp_1024_ecc_mulmod_base_avx2_16(point, k, 0, 0, heap);
-        }
         else
 #endif
             err = sp_1024_ecc_mulmod_base_16(point, k, 0, 0, heap);
     }
     if (err == MP_OKAY) {
 #ifdef HAVE_INTEL_AVX2
-        if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+        if (saved_vector_registers)
             sp_1024_proj_point_add_avx2_16(point, point, addP, tmp);
-        }
         else
 #endif
             sp_1024_proj_point_add_16(point, point, addP, tmp);
 
         if (map) {
 #ifdef HAVE_INTEL_AVX2
-            if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+            if (saved_vector_registers)
                 sp_1024_map_avx2_16(point, point, tmp);
-            }
             else
 #endif
                 sp_1024_map_16(point, point, tmp);
@@ -98338,6 +98492,11 @@ int sp_ecc_mulmod_base_add_1024(const mp_int* km, const ecc_point* am,
         err = sp_1024_point_to_ecc_point_16(point, r);
     }
 
+#ifdef HAVE_INTEL_AVX2
+    if (saved_vector_registers)
+        RESTORE_VECTOR_REGISTERS();
+#endif
+
     SP_FREE_VAR(k, NULL, DYNAMIC_TYPE_ECC);
     SP_FREE_VAR(point, NULL, DYNAMIC_TYPE_ECC);
 
@@ -98382,9 +98541,10 @@ int sp_ecc_gen_table_1024(const ecc_point* gm, byte* table, word32* len,
         sp_1024_point_from_ecc_point_16(point, gm);
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_1024_gen_stripe_table_avx2_16(point,
                 (sp_table_entry_1024*)table, t, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -98465,9 +98625,10 @@ int sp_ecc_mulmod_table_1024(const mp_int* km, const ecc_point* gm, byte* table,
 #ifndef WOLFSSL_SP_SMALL
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_1024_ecc_mulmod_stripe_avx2_16(point, point,
                 (const sp_table_entry_1024*)table, k, map, 0, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -100536,11 +100697,11 @@ int sp_ModExp_Fp_star_1024(const mp_int* base, mp_int* exp, mp_int* res)
     word32 cpuid_flags = cpuid_get_flags();
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
 #ifdef HAVE_INTEL_AVX2
-    if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags)) {
+    if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
+            IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
         err = sp_ModExp_Fp_star_avx2_1024(base, exp, res);
+        RESTORE_VECTOR_REGISTERS();
     }
     else
 #endif
@@ -102138,11 +102299,11 @@ int sp_Pairing_1024(const ecc_point* pm, const ecc_point* qm, mp_int* res)
     word32 cpuid_flags = cpuid_get_flags();
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
 #ifdef HAVE_INTEL_AVX2
-    if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags)) {
+    if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
+            IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
         err = sp_Pairing_avx2_1024(pm, qm, res);
+        RESTORE_VECTOR_REGISTERS();
     }
     else
 #endif
@@ -103263,11 +103424,11 @@ int sp_Pairing_gen_precomp_1024(const ecc_point* pm, byte* table, word32* len)
     word32 cpuid_flags = cpuid_get_flags();
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
 #ifdef HAVE_INTEL_AVX2
-    if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags)) {
+    if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
+            IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
         err = sp_Pairing_gen_precomp_avx2_1024(pm, table, len);
+        RESTORE_VECTOR_REGISTERS();
     }
     else
 #endif
@@ -103299,11 +103460,11 @@ int sp_Pairing_precomp_1024(const ecc_point* pm, const ecc_point* qm, mp_int* re
     word32 cpuid_flags = cpuid_get_flags();
 #endif
 
-    ASSERT_SAVED_VECTOR_REGISTERS();
-
 #ifdef HAVE_INTEL_AVX2
-    if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags)) {
+    if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
+            IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
         err = sp_Pairing_precomp_avx2_1024(pm, qm, res, table, len);
+        RESTORE_VECTOR_REGISTERS();
     }
     else
 #endif
@@ -103494,8 +103655,9 @@ int sp_ecc_check_key_1024(const mp_int* pX, const mp_int* pY,
         /* Point * order = infinity */
 #ifdef HAVE_INTEL_AVX2
         if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                IS_INTEL_AVX2(cpuid_flags)) {
+                IS_INTEL_AVX2(cpuid_flags) && (SAVE_VECTOR_REGISTERS2() == 0)) {
             err = sp_1024_ecc_mulmod_avx2_16(p, pub, p1024_order, 1, 1, heap);
+            RESTORE_VECTOR_REGISTERS();
         }
         else
 #endif
@@ -103512,8 +103674,10 @@ int sp_ecc_check_key_1024(const mp_int* pX, const mp_int* pY,
             /* Base * private = point */
 #ifdef HAVE_INTEL_AVX2
             if (IS_INTEL_BMI2(cpuid_flags) && IS_INTEL_ADX(cpuid_flags) &&
-                    IS_INTEL_AVX2(cpuid_flags)) {
+                    IS_INTEL_AVX2(cpuid_flags) &&
+                    (SAVE_VECTOR_REGISTERS2() == 0)) {
                 err = sp_1024_ecc_mulmod_base_avx2_16(p, priv, 1, 1, heap);
+                RESTORE_VECTOR_REGISTERS();
             }
             else
 #endif
diff --git a/wolfcrypt/src/sp_x86_64_asm.S b/wolfcrypt/src/sp_x86_64_asm.S
index 7aaae4bb60..fbd2de64a8 100644
--- a/wolfcrypt/src/sp_x86_64_asm.S
+++ b/wolfcrypt/src/sp_x86_64_asm.S
@@ -9983,1173 +9983,6 @@ _sp_2048_cmp_16:
 #ifndef __APPLE__
 .size	sp_2048_cmp_16,.-sp_2048_cmp_16
 #endif /* __APPLE__ */
-#ifndef WC_NO_CACHE_RESISTANT
-#ifndef __APPLE__
-.text
-.globl	sp_2048_get_from_table_16
-.type	sp_2048_get_from_table_16,@function
-.align	16
-sp_2048_get_from_table_16:
-#else
-.section	__TEXT,__text
-.globl	_sp_2048_get_from_table_16
-.p2align	4
-_sp_2048_get_from_table_16:
-#endif /* __APPLE__ */
-        movq	$0x01, %rax
-        movd	%rdx, %xmm10
-        movd	%rax, %xmm11
-        pxor	%xmm13, %xmm13
-        pshufd	$0x00, %xmm11, %xmm11
-        pshufd	$0x00, %xmm10, %xmm10
-        # START: 0-7
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 16
-        movq	128(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 17
-        movq	136(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 18
-        movq	144(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 19
-        movq	152(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 20
-        movq	160(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 21
-        movq	168(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 22
-        movq	176(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 23
-        movq	184(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 24
-        movq	192(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 25
-        movq	200(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 26
-        movq	208(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 27
-        movq	216(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 28
-        movq	224(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 29
-        movq	232(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 30
-        movq	240(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 31
-        movq	248(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        addq	$0x40, %rdi
-        # END: 0-7
-        # START: 8-15
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 16
-        movq	128(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 17
-        movq	136(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 18
-        movq	144(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 19
-        movq	152(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 20
-        movq	160(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 21
-        movq	168(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 22
-        movq	176(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 23
-        movq	184(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 24
-        movq	192(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 25
-        movq	200(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 26
-        movq	208(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 27
-        movq	216(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 28
-        movq	224(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 29
-        movq	232(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 30
-        movq	240(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 31
-        movq	248(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        # END: 8-15
-        repz retq
-#ifndef __APPLE__
-.size	sp_2048_get_from_table_16,.-sp_2048_get_from_table_16
-#endif /* __APPLE__ */
-#endif /* !WC_NO_CACHE_RESISTANT */
 #ifdef HAVE_INTEL_AVX2
 /* Reduce the number back to 2048 bits using Montgomery reduction.
  *
@@ -13501,4621 +12334,6 @@ _sp_2048_cmp_32:
 #ifndef __APPLE__
 .size	sp_2048_cmp_32,.-sp_2048_cmp_32
 #endif /* __APPLE__ */
-#ifndef WC_NO_CACHE_RESISTANT
-#ifndef __APPLE__
-.text
-.globl	sp_2048_get_from_table_32
-.type	sp_2048_get_from_table_32,@function
-.align	16
-sp_2048_get_from_table_32:
-#else
-.section	__TEXT,__text
-.globl	_sp_2048_get_from_table_32
-.p2align	4
-_sp_2048_get_from_table_32:
-#endif /* __APPLE__ */
-        movq	$0x01, %rax
-        movd	%rdx, %xmm10
-        movd	%rax, %xmm11
-        pxor	%xmm13, %xmm13
-        pshufd	$0x00, %xmm11, %xmm11
-        pshufd	$0x00, %xmm10, %xmm10
-        # START: 0-7
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 16
-        movq	128(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 17
-        movq	136(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 18
-        movq	144(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 19
-        movq	152(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 20
-        movq	160(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 21
-        movq	168(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 22
-        movq	176(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 23
-        movq	184(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 24
-        movq	192(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 25
-        movq	200(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 26
-        movq	208(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 27
-        movq	216(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 28
-        movq	224(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 29
-        movq	232(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 30
-        movq	240(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 31
-        movq	248(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 32
-        movq	256(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 33
-        movq	264(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 34
-        movq	272(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 35
-        movq	280(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 36
-        movq	288(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 37
-        movq	296(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 38
-        movq	304(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 39
-        movq	312(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 40
-        movq	320(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 41
-        movq	328(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 42
-        movq	336(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 43
-        movq	344(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 44
-        movq	352(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 45
-        movq	360(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 46
-        movq	368(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 47
-        movq	376(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 48
-        movq	384(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 49
-        movq	392(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 50
-        movq	400(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 51
-        movq	408(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 52
-        movq	416(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 53
-        movq	424(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 54
-        movq	432(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 55
-        movq	440(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 56
-        movq	448(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 57
-        movq	456(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 58
-        movq	464(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 59
-        movq	472(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 60
-        movq	480(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 61
-        movq	488(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 62
-        movq	496(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 63
-        movq	504(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        addq	$0x40, %rdi
-        # END: 0-7
-        # START: 8-15
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 16
-        movq	128(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 17
-        movq	136(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 18
-        movq	144(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 19
-        movq	152(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 20
-        movq	160(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 21
-        movq	168(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 22
-        movq	176(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 23
-        movq	184(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 24
-        movq	192(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 25
-        movq	200(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 26
-        movq	208(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 27
-        movq	216(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 28
-        movq	224(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 29
-        movq	232(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 30
-        movq	240(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 31
-        movq	248(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 32
-        movq	256(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 33
-        movq	264(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 34
-        movq	272(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 35
-        movq	280(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 36
-        movq	288(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 37
-        movq	296(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 38
-        movq	304(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 39
-        movq	312(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 40
-        movq	320(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 41
-        movq	328(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 42
-        movq	336(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 43
-        movq	344(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 44
-        movq	352(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 45
-        movq	360(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 46
-        movq	368(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 47
-        movq	376(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 48
-        movq	384(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 49
-        movq	392(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 50
-        movq	400(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 51
-        movq	408(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 52
-        movq	416(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 53
-        movq	424(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 54
-        movq	432(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 55
-        movq	440(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 56
-        movq	448(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 57
-        movq	456(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 58
-        movq	464(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 59
-        movq	472(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 60
-        movq	480(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 61
-        movq	488(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 62
-        movq	496(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 63
-        movq	504(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        addq	$0x40, %rdi
-        # END: 8-15
-        # START: 16-23
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 16
-        movq	128(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 17
-        movq	136(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 18
-        movq	144(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 19
-        movq	152(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 20
-        movq	160(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 21
-        movq	168(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 22
-        movq	176(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 23
-        movq	184(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 24
-        movq	192(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 25
-        movq	200(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 26
-        movq	208(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 27
-        movq	216(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 28
-        movq	224(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 29
-        movq	232(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 30
-        movq	240(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 31
-        movq	248(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 32
-        movq	256(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 33
-        movq	264(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 34
-        movq	272(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 35
-        movq	280(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 36
-        movq	288(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 37
-        movq	296(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 38
-        movq	304(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 39
-        movq	312(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 40
-        movq	320(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 41
-        movq	328(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 42
-        movq	336(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 43
-        movq	344(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 44
-        movq	352(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 45
-        movq	360(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 46
-        movq	368(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 47
-        movq	376(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 48
-        movq	384(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 49
-        movq	392(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 50
-        movq	400(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 51
-        movq	408(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 52
-        movq	416(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 53
-        movq	424(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 54
-        movq	432(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 55
-        movq	440(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 56
-        movq	448(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 57
-        movq	456(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 58
-        movq	464(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 59
-        movq	472(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 60
-        movq	480(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 61
-        movq	488(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 62
-        movq	496(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 63
-        movq	504(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        addq	$0x40, %rdi
-        # END: 16-23
-        # START: 24-31
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 16
-        movq	128(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 17
-        movq	136(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 18
-        movq	144(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 19
-        movq	152(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 20
-        movq	160(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 21
-        movq	168(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 22
-        movq	176(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 23
-        movq	184(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 24
-        movq	192(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 25
-        movq	200(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 26
-        movq	208(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 27
-        movq	216(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 28
-        movq	224(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 29
-        movq	232(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 30
-        movq	240(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 31
-        movq	248(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 32
-        movq	256(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 33
-        movq	264(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 34
-        movq	272(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 35
-        movq	280(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 36
-        movq	288(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 37
-        movq	296(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 38
-        movq	304(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 39
-        movq	312(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 40
-        movq	320(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 41
-        movq	328(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 42
-        movq	336(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 43
-        movq	344(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 44
-        movq	352(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 45
-        movq	360(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 46
-        movq	368(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 47
-        movq	376(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 48
-        movq	384(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 49
-        movq	392(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 50
-        movq	400(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 51
-        movq	408(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 52
-        movq	416(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 53
-        movq	424(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 54
-        movq	432(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 55
-        movq	440(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 56
-        movq	448(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 57
-        movq	456(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 58
-        movq	464(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 59
-        movq	472(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 60
-        movq	480(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 61
-        movq	488(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 62
-        movq	496(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 63
-        movq	504(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        # END: 24-31
-        repz retq
-#ifndef __APPLE__
-.size	sp_2048_get_from_table_32,.-sp_2048_get_from_table_32
-#endif /* __APPLE__ */
-#endif /* !WC_NO_CACHE_RESISTANT */
 #ifdef HAVE_INTEL_AVX2
 /* Reduce the number back to 2048 bits using Montgomery reduction.
  *
@@ -32549,1761 +26767,6 @@ _sp_3072_cmp_24:
 #ifndef __APPLE__
 .size	sp_3072_cmp_24,.-sp_3072_cmp_24
 #endif /* __APPLE__ */
-#ifndef WC_NO_CACHE_RESISTANT
-#ifndef __APPLE__
-.text
-.globl	sp_3072_get_from_table_24
-.type	sp_3072_get_from_table_24,@function
-.align	16
-sp_3072_get_from_table_24:
-#else
-.section	__TEXT,__text
-.globl	_sp_3072_get_from_table_24
-.p2align	4
-_sp_3072_get_from_table_24:
-#endif /* __APPLE__ */
-        movq	$0x01, %rax
-        movd	%rdx, %xmm10
-        movd	%rax, %xmm11
-        pxor	%xmm13, %xmm13
-        pshufd	$0x00, %xmm11, %xmm11
-        pshufd	$0x00, %xmm10, %xmm10
-        # START: 0-7
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 16
-        movq	128(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 17
-        movq	136(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 18
-        movq	144(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 19
-        movq	152(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 20
-        movq	160(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 21
-        movq	168(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 22
-        movq	176(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 23
-        movq	184(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 24
-        movq	192(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 25
-        movq	200(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 26
-        movq	208(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 27
-        movq	216(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 28
-        movq	224(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 29
-        movq	232(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 30
-        movq	240(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 31
-        movq	248(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        addq	$0x40, %rdi
-        # END: 0-7
-        # START: 8-15
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 16
-        movq	128(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 17
-        movq	136(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 18
-        movq	144(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 19
-        movq	152(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 20
-        movq	160(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 21
-        movq	168(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 22
-        movq	176(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 23
-        movq	184(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 24
-        movq	192(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 25
-        movq	200(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 26
-        movq	208(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 27
-        movq	216(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 28
-        movq	224(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 29
-        movq	232(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 30
-        movq	240(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 31
-        movq	248(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        addq	$0x40, %rdi
-        # END: 8-15
-        # START: 16-23
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 16
-        movq	128(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 17
-        movq	136(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 18
-        movq	144(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 19
-        movq	152(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 20
-        movq	160(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 21
-        movq	168(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 22
-        movq	176(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 23
-        movq	184(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 24
-        movq	192(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 25
-        movq	200(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 26
-        movq	208(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 27
-        movq	216(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 28
-        movq	224(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 29
-        movq	232(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 30
-        movq	240(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 31
-        movq	248(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        # END: 16-23
-        repz retq
-#ifndef __APPLE__
-.size	sp_3072_get_from_table_24,.-sp_3072_get_from_table_24
-#endif /* __APPLE__ */
-#endif /* !WC_NO_CACHE_RESISTANT */
 #ifdef HAVE_INTEL_AVX2
 /* Reduce the number back to 3072 bits using Montgomery reduction.
  *
@@ -37626,1813 +30089,6 @@ _sp_3072_cmp_48:
 #ifndef __APPLE__
 .size	sp_3072_cmp_48,.-sp_3072_cmp_48
 #endif /* __APPLE__ */
-#ifndef WC_NO_CACHE_RESISTANT
-#ifndef __APPLE__
-.text
-.globl	sp_3072_get_from_table_48
-.type	sp_3072_get_from_table_48,@function
-.align	16
-sp_3072_get_from_table_48:
-#else
-.section	__TEXT,__text
-.globl	_sp_3072_get_from_table_48
-.p2align	4
-_sp_3072_get_from_table_48:
-#endif /* __APPLE__ */
-        movq	$0x01, %rax
-        movd	%rdx, %xmm10
-        movd	%rax, %xmm11
-        pxor	%xmm13, %xmm13
-        pshufd	$0x00, %xmm11, %xmm11
-        pshufd	$0x00, %xmm10, %xmm10
-        # START: 0-7
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        addq	$0x40, %rdi
-        # END: 0-7
-        # START: 8-15
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        addq	$0x40, %rdi
-        # END: 8-15
-        # START: 16-23
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        addq	$0x40, %rdi
-        # END: 16-23
-        # START: 24-31
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        addq	$0x40, %rdi
-        # END: 24-31
-        # START: 32-39
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        addq	$0x40, %rdi
-        # END: 32-39
-        # START: 40-47
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        # END: 40-47
-        repz retq
-#ifndef __APPLE__
-.size	sp_3072_get_from_table_48,.-sp_3072_get_from_table_48
-#endif /* __APPLE__ */
-#endif /* !WC_NO_CACHE_RESISTANT */
 #ifdef HAVE_INTEL_AVX2
 /* Reduce the number back to 3072 bits using Montgomery reduction.
  *
@@ -50237,2413 +40893,6 @@ _sp_4096_cmp_64:
 #ifndef __APPLE__
 .size	sp_4096_cmp_64,.-sp_4096_cmp_64
 #endif /* __APPLE__ */
-#ifndef WC_NO_CACHE_RESISTANT
-#ifndef __APPLE__
-.text
-.globl	sp_4096_get_from_table_64
-.type	sp_4096_get_from_table_64,@function
-.align	16
-sp_4096_get_from_table_64:
-#else
-.section	__TEXT,__text
-.globl	_sp_4096_get_from_table_64
-.p2align	4
-_sp_4096_get_from_table_64:
-#endif /* __APPLE__ */
-        movq	$0x01, %rax
-        movd	%rdx, %xmm10
-        movd	%rax, %xmm11
-        pxor	%xmm13, %xmm13
-        pshufd	$0x00, %xmm11, %xmm11
-        pshufd	$0x00, %xmm10, %xmm10
-        # START: 0-7
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        addq	$0x40, %rdi
-        # END: 0-7
-        # START: 8-15
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        addq	$0x40, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        addq	$0x40, %rdi
-        # END: 8-15
-        # START: 16-23
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        addq	$0x80, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        addq	$0x40, %rdi
-        # END: 16-23
-        # START: 24-31
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        addq	$0xc0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        addq	$0x40, %rdi
-        # END: 24-31
-        # START: 32-39
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        addq	$0x100, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        addq	$0x40, %rdi
-        # END: 32-39
-        # START: 40-47
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        addq	$0x140, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        addq	$0x40, %rdi
-        # END: 40-47
-        # START: 48-55
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        addq	$0x180, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        addq	$0x180, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        addq	$0x180, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        addq	$0x180, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        addq	$0x180, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        addq	$0x180, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        addq	$0x180, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        addq	$0x180, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        addq	$0x180, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        addq	$0x180, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        addq	$0x180, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        addq	$0x180, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        addq	$0x180, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        addq	$0x180, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        addq	$0x180, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        addq	$0x180, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        addq	$0x40, %rdi
-        # END: 48-55
-        # START: 56-63
-        pxor	%xmm13, %xmm13
-        pxor	%xmm4, %xmm4
-        pxor	%xmm5, %xmm5
-        pxor	%xmm6, %xmm6
-        pxor	%xmm7, %xmm7
-        # ENTRY: 0
-        movq	(%rsi), %rcx
-        addq	$0x1c0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 1
-        movq	8(%rsi), %rcx
-        addq	$0x1c0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 2
-        movq	16(%rsi), %rcx
-        addq	$0x1c0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 3
-        movq	24(%rsi), %rcx
-        addq	$0x1c0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 4
-        movq	32(%rsi), %rcx
-        addq	$0x1c0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 5
-        movq	40(%rsi), %rcx
-        addq	$0x1c0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 6
-        movq	48(%rsi), %rcx
-        addq	$0x1c0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 7
-        movq	56(%rsi), %rcx
-        addq	$0x1c0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 8
-        movq	64(%rsi), %rcx
-        addq	$0x1c0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 9
-        movq	72(%rsi), %rcx
-        addq	$0x1c0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 10
-        movq	80(%rsi), %rcx
-        addq	$0x1c0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 11
-        movq	88(%rsi), %rcx
-        addq	$0x1c0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 12
-        movq	96(%rsi), %rcx
-        addq	$0x1c0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 13
-        movq	104(%rsi), %rcx
-        addq	$0x1c0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 14
-        movq	112(%rsi), %rcx
-        addq	$0x1c0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        # ENTRY: 15
-        movq	120(%rsi), %rcx
-        addq	$0x1c0, %rcx
-        movdqu	%xmm13, %xmm12
-        pcmpeqd	%xmm10, %xmm12
-        movdqu	(%rcx), %xmm0
-        movdqu	16(%rcx), %xmm1
-        movdqu	32(%rcx), %xmm2
-        movdqu	48(%rcx), %xmm3
-        pand	%xmm12, %xmm0
-        pand	%xmm12, %xmm1
-        pand	%xmm12, %xmm2
-        pand	%xmm12, %xmm3
-        por	%xmm0, %xmm4
-        por	%xmm1, %xmm5
-        por	%xmm2, %xmm6
-        por	%xmm3, %xmm7
-        paddd	%xmm11, %xmm13
-        movdqu	%xmm4, (%rdi)
-        movdqu	%xmm5, 16(%rdi)
-        movdqu	%xmm6, 32(%rdi)
-        movdqu	%xmm7, 48(%rdi)
-        # END: 56-63
-        repz retq
-#ifndef __APPLE__
-.size	sp_4096_get_from_table_64,.-sp_4096_get_from_table_64
-#endif /* __APPLE__ */
-#endif /* !WC_NO_CACHE_RESISTANT */
 #ifdef HAVE_INTEL_AVX2
 /* Reduce the number back to 4096 bits using Montgomery reduction.
  *
diff --git a/wolfcrypt/src/sp_x86_64_asm.asm b/wolfcrypt/src/sp_x86_64_asm.asm
index cdf5789407..2ad38bb43b 100644
--- a/wolfcrypt/src/sp_x86_64_asm.asm
+++ b/wolfcrypt/src/sp_x86_64_asm.asm
@@ -9709,1180 +9709,6 @@ sp_2048_cmp_16 PROC
         ret
 sp_2048_cmp_16 ENDP
 _TEXT ENDS
-IFNDEF WC_NO_CACHE_RESISTANT
-_TEXT SEGMENT READONLY PARA
-sp_2048_get_from_table_16 PROC
-        sub	rsp, 128
-        movdqu	OWORD PTR [rsp], xmm6
-        movdqu	OWORD PTR [rsp+16], xmm7
-        movdqu	OWORD PTR [rsp+32], xmm8
-        movdqu	OWORD PTR [rsp+48], xmm9
-        movdqu	OWORD PTR [rsp+64], xmm10
-        movdqu	OWORD PTR [rsp+80], xmm11
-        movdqu	OWORD PTR [rsp+96], xmm12
-        movdqu	OWORD PTR [rsp+112], xmm13
-        mov	rax, 1
-        movd	xmm10, r8
-        movd	xmm11, rax
-        pxor	xmm13, xmm13
-        pshufd	xmm11, xmm11, 0
-        pshufd	xmm10, xmm10, 0
-        ; START: 0-7
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 16
-        mov	r9, QWORD PTR [rdx+128]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 17
-        mov	r9, QWORD PTR [rdx+136]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 18
-        mov	r9, QWORD PTR [rdx+144]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 19
-        mov	r9, QWORD PTR [rdx+152]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 20
-        mov	r9, QWORD PTR [rdx+160]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 21
-        mov	r9, QWORD PTR [rdx+168]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 22
-        mov	r9, QWORD PTR [rdx+176]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 23
-        mov	r9, QWORD PTR [rdx+184]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 24
-        mov	r9, QWORD PTR [rdx+192]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 25
-        mov	r9, QWORD PTR [rdx+200]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 26
-        mov	r9, QWORD PTR [rdx+208]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 27
-        mov	r9, QWORD PTR [rdx+216]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 28
-        mov	r9, QWORD PTR [rdx+224]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 29
-        mov	r9, QWORD PTR [rdx+232]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 30
-        mov	r9, QWORD PTR [rdx+240]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 31
-        mov	r9, QWORD PTR [rdx+248]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        add	rcx, 64
-        ; END: 0-7
-        ; START: 8-15
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 16
-        mov	r9, QWORD PTR [rdx+128]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 17
-        mov	r9, QWORD PTR [rdx+136]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 18
-        mov	r9, QWORD PTR [rdx+144]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 19
-        mov	r9, QWORD PTR [rdx+152]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 20
-        mov	r9, QWORD PTR [rdx+160]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 21
-        mov	r9, QWORD PTR [rdx+168]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 22
-        mov	r9, QWORD PTR [rdx+176]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 23
-        mov	r9, QWORD PTR [rdx+184]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 24
-        mov	r9, QWORD PTR [rdx+192]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 25
-        mov	r9, QWORD PTR [rdx+200]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 26
-        mov	r9, QWORD PTR [rdx+208]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 27
-        mov	r9, QWORD PTR [rdx+216]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 28
-        mov	r9, QWORD PTR [rdx+224]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 29
-        mov	r9, QWORD PTR [rdx+232]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 30
-        mov	r9, QWORD PTR [rdx+240]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 31
-        mov	r9, QWORD PTR [rdx+248]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        ; END: 8-15
-        movdqu	xmm6, OWORD PTR [rsp]
-        movdqu	xmm7, OWORD PTR [rsp+16]
-        movdqu	xmm8, OWORD PTR [rsp+32]
-        movdqu	xmm9, OWORD PTR [rsp+48]
-        movdqu	xmm10, OWORD PTR [rsp+64]
-        movdqu	xmm11, OWORD PTR [rsp+80]
-        movdqu	xmm12, OWORD PTR [rsp+96]
-        movdqu	xmm13, OWORD PTR [rsp+112]
-        add	rsp, 128
-        ret
-sp_2048_get_from_table_16 ENDP
-_TEXT ENDS
-ENDIF
 IFDEF HAVE_INTEL_AVX2
 ; /* Reduce the number back to 2048 bits using Montgomery reduction.
 ;  *
@@ -13166,4628 +11992,6 @@ sp_2048_cmp_32 PROC
         ret
 sp_2048_cmp_32 ENDP
 _TEXT ENDS
-IFNDEF WC_NO_CACHE_RESISTANT
-_TEXT SEGMENT READONLY PARA
-sp_2048_get_from_table_32 PROC
-        sub	rsp, 128
-        movdqu	OWORD PTR [rsp], xmm6
-        movdqu	OWORD PTR [rsp+16], xmm7
-        movdqu	OWORD PTR [rsp+32], xmm8
-        movdqu	OWORD PTR [rsp+48], xmm9
-        movdqu	OWORD PTR [rsp+64], xmm10
-        movdqu	OWORD PTR [rsp+80], xmm11
-        movdqu	OWORD PTR [rsp+96], xmm12
-        movdqu	OWORD PTR [rsp+112], xmm13
-        mov	rax, 1
-        movd	xmm10, r8
-        movd	xmm11, rax
-        pxor	xmm13, xmm13
-        pshufd	xmm11, xmm11, 0
-        pshufd	xmm10, xmm10, 0
-        ; START: 0-7
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 16
-        mov	r9, QWORD PTR [rdx+128]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 17
-        mov	r9, QWORD PTR [rdx+136]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 18
-        mov	r9, QWORD PTR [rdx+144]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 19
-        mov	r9, QWORD PTR [rdx+152]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 20
-        mov	r9, QWORD PTR [rdx+160]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 21
-        mov	r9, QWORD PTR [rdx+168]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 22
-        mov	r9, QWORD PTR [rdx+176]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 23
-        mov	r9, QWORD PTR [rdx+184]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 24
-        mov	r9, QWORD PTR [rdx+192]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 25
-        mov	r9, QWORD PTR [rdx+200]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 26
-        mov	r9, QWORD PTR [rdx+208]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 27
-        mov	r9, QWORD PTR [rdx+216]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 28
-        mov	r9, QWORD PTR [rdx+224]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 29
-        mov	r9, QWORD PTR [rdx+232]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 30
-        mov	r9, QWORD PTR [rdx+240]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 31
-        mov	r9, QWORD PTR [rdx+248]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 32
-        mov	r9, QWORD PTR [rdx+256]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 33
-        mov	r9, QWORD PTR [rdx+264]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 34
-        mov	r9, QWORD PTR [rdx+272]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 35
-        mov	r9, QWORD PTR [rdx+280]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 36
-        mov	r9, QWORD PTR [rdx+288]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 37
-        mov	r9, QWORD PTR [rdx+296]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 38
-        mov	r9, QWORD PTR [rdx+304]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 39
-        mov	r9, QWORD PTR [rdx+312]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 40
-        mov	r9, QWORD PTR [rdx+320]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 41
-        mov	r9, QWORD PTR [rdx+328]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 42
-        mov	r9, QWORD PTR [rdx+336]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 43
-        mov	r9, QWORD PTR [rdx+344]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 44
-        mov	r9, QWORD PTR [rdx+352]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 45
-        mov	r9, QWORD PTR [rdx+360]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 46
-        mov	r9, QWORD PTR [rdx+368]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 47
-        mov	r9, QWORD PTR [rdx+376]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 48
-        mov	r9, QWORD PTR [rdx+384]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 49
-        mov	r9, QWORD PTR [rdx+392]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 50
-        mov	r9, QWORD PTR [rdx+400]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 51
-        mov	r9, QWORD PTR [rdx+408]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 52
-        mov	r9, QWORD PTR [rdx+416]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 53
-        mov	r9, QWORD PTR [rdx+424]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 54
-        mov	r9, QWORD PTR [rdx+432]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 55
-        mov	r9, QWORD PTR [rdx+440]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 56
-        mov	r9, QWORD PTR [rdx+448]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 57
-        mov	r9, QWORD PTR [rdx+456]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 58
-        mov	r9, QWORD PTR [rdx+464]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 59
-        mov	r9, QWORD PTR [rdx+472]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 60
-        mov	r9, QWORD PTR [rdx+480]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 61
-        mov	r9, QWORD PTR [rdx+488]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 62
-        mov	r9, QWORD PTR [rdx+496]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 63
-        mov	r9, QWORD PTR [rdx+504]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        add	rcx, 64
-        ; END: 0-7
-        ; START: 8-15
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 16
-        mov	r9, QWORD PTR [rdx+128]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 17
-        mov	r9, QWORD PTR [rdx+136]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 18
-        mov	r9, QWORD PTR [rdx+144]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 19
-        mov	r9, QWORD PTR [rdx+152]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 20
-        mov	r9, QWORD PTR [rdx+160]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 21
-        mov	r9, QWORD PTR [rdx+168]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 22
-        mov	r9, QWORD PTR [rdx+176]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 23
-        mov	r9, QWORD PTR [rdx+184]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 24
-        mov	r9, QWORD PTR [rdx+192]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 25
-        mov	r9, QWORD PTR [rdx+200]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 26
-        mov	r9, QWORD PTR [rdx+208]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 27
-        mov	r9, QWORD PTR [rdx+216]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 28
-        mov	r9, QWORD PTR [rdx+224]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 29
-        mov	r9, QWORD PTR [rdx+232]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 30
-        mov	r9, QWORD PTR [rdx+240]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 31
-        mov	r9, QWORD PTR [rdx+248]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 32
-        mov	r9, QWORD PTR [rdx+256]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 33
-        mov	r9, QWORD PTR [rdx+264]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 34
-        mov	r9, QWORD PTR [rdx+272]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 35
-        mov	r9, QWORD PTR [rdx+280]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 36
-        mov	r9, QWORD PTR [rdx+288]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 37
-        mov	r9, QWORD PTR [rdx+296]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 38
-        mov	r9, QWORD PTR [rdx+304]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 39
-        mov	r9, QWORD PTR [rdx+312]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 40
-        mov	r9, QWORD PTR [rdx+320]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 41
-        mov	r9, QWORD PTR [rdx+328]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 42
-        mov	r9, QWORD PTR [rdx+336]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 43
-        mov	r9, QWORD PTR [rdx+344]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 44
-        mov	r9, QWORD PTR [rdx+352]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 45
-        mov	r9, QWORD PTR [rdx+360]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 46
-        mov	r9, QWORD PTR [rdx+368]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 47
-        mov	r9, QWORD PTR [rdx+376]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 48
-        mov	r9, QWORD PTR [rdx+384]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 49
-        mov	r9, QWORD PTR [rdx+392]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 50
-        mov	r9, QWORD PTR [rdx+400]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 51
-        mov	r9, QWORD PTR [rdx+408]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 52
-        mov	r9, QWORD PTR [rdx+416]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 53
-        mov	r9, QWORD PTR [rdx+424]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 54
-        mov	r9, QWORD PTR [rdx+432]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 55
-        mov	r9, QWORD PTR [rdx+440]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 56
-        mov	r9, QWORD PTR [rdx+448]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 57
-        mov	r9, QWORD PTR [rdx+456]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 58
-        mov	r9, QWORD PTR [rdx+464]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 59
-        mov	r9, QWORD PTR [rdx+472]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 60
-        mov	r9, QWORD PTR [rdx+480]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 61
-        mov	r9, QWORD PTR [rdx+488]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 62
-        mov	r9, QWORD PTR [rdx+496]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 63
-        mov	r9, QWORD PTR [rdx+504]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        add	rcx, 64
-        ; END: 8-15
-        ; START: 16-23
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 16
-        mov	r9, QWORD PTR [rdx+128]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 17
-        mov	r9, QWORD PTR [rdx+136]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 18
-        mov	r9, QWORD PTR [rdx+144]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 19
-        mov	r9, QWORD PTR [rdx+152]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 20
-        mov	r9, QWORD PTR [rdx+160]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 21
-        mov	r9, QWORD PTR [rdx+168]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 22
-        mov	r9, QWORD PTR [rdx+176]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 23
-        mov	r9, QWORD PTR [rdx+184]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 24
-        mov	r9, QWORD PTR [rdx+192]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 25
-        mov	r9, QWORD PTR [rdx+200]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 26
-        mov	r9, QWORD PTR [rdx+208]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 27
-        mov	r9, QWORD PTR [rdx+216]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 28
-        mov	r9, QWORD PTR [rdx+224]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 29
-        mov	r9, QWORD PTR [rdx+232]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 30
-        mov	r9, QWORD PTR [rdx+240]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 31
-        mov	r9, QWORD PTR [rdx+248]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 32
-        mov	r9, QWORD PTR [rdx+256]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 33
-        mov	r9, QWORD PTR [rdx+264]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 34
-        mov	r9, QWORD PTR [rdx+272]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 35
-        mov	r9, QWORD PTR [rdx+280]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 36
-        mov	r9, QWORD PTR [rdx+288]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 37
-        mov	r9, QWORD PTR [rdx+296]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 38
-        mov	r9, QWORD PTR [rdx+304]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 39
-        mov	r9, QWORD PTR [rdx+312]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 40
-        mov	r9, QWORD PTR [rdx+320]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 41
-        mov	r9, QWORD PTR [rdx+328]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 42
-        mov	r9, QWORD PTR [rdx+336]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 43
-        mov	r9, QWORD PTR [rdx+344]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 44
-        mov	r9, QWORD PTR [rdx+352]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 45
-        mov	r9, QWORD PTR [rdx+360]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 46
-        mov	r9, QWORD PTR [rdx+368]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 47
-        mov	r9, QWORD PTR [rdx+376]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 48
-        mov	r9, QWORD PTR [rdx+384]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 49
-        mov	r9, QWORD PTR [rdx+392]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 50
-        mov	r9, QWORD PTR [rdx+400]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 51
-        mov	r9, QWORD PTR [rdx+408]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 52
-        mov	r9, QWORD PTR [rdx+416]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 53
-        mov	r9, QWORD PTR [rdx+424]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 54
-        mov	r9, QWORD PTR [rdx+432]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 55
-        mov	r9, QWORD PTR [rdx+440]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 56
-        mov	r9, QWORD PTR [rdx+448]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 57
-        mov	r9, QWORD PTR [rdx+456]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 58
-        mov	r9, QWORD PTR [rdx+464]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 59
-        mov	r9, QWORD PTR [rdx+472]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 60
-        mov	r9, QWORD PTR [rdx+480]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 61
-        mov	r9, QWORD PTR [rdx+488]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 62
-        mov	r9, QWORD PTR [rdx+496]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 63
-        mov	r9, QWORD PTR [rdx+504]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        add	rcx, 64
-        ; END: 16-23
-        ; START: 24-31
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 16
-        mov	r9, QWORD PTR [rdx+128]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 17
-        mov	r9, QWORD PTR [rdx+136]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 18
-        mov	r9, QWORD PTR [rdx+144]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 19
-        mov	r9, QWORD PTR [rdx+152]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 20
-        mov	r9, QWORD PTR [rdx+160]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 21
-        mov	r9, QWORD PTR [rdx+168]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 22
-        mov	r9, QWORD PTR [rdx+176]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 23
-        mov	r9, QWORD PTR [rdx+184]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 24
-        mov	r9, QWORD PTR [rdx+192]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 25
-        mov	r9, QWORD PTR [rdx+200]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 26
-        mov	r9, QWORD PTR [rdx+208]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 27
-        mov	r9, QWORD PTR [rdx+216]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 28
-        mov	r9, QWORD PTR [rdx+224]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 29
-        mov	r9, QWORD PTR [rdx+232]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 30
-        mov	r9, QWORD PTR [rdx+240]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 31
-        mov	r9, QWORD PTR [rdx+248]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 32
-        mov	r9, QWORD PTR [rdx+256]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 33
-        mov	r9, QWORD PTR [rdx+264]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 34
-        mov	r9, QWORD PTR [rdx+272]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 35
-        mov	r9, QWORD PTR [rdx+280]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 36
-        mov	r9, QWORD PTR [rdx+288]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 37
-        mov	r9, QWORD PTR [rdx+296]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 38
-        mov	r9, QWORD PTR [rdx+304]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 39
-        mov	r9, QWORD PTR [rdx+312]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 40
-        mov	r9, QWORD PTR [rdx+320]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 41
-        mov	r9, QWORD PTR [rdx+328]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 42
-        mov	r9, QWORD PTR [rdx+336]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 43
-        mov	r9, QWORD PTR [rdx+344]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 44
-        mov	r9, QWORD PTR [rdx+352]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 45
-        mov	r9, QWORD PTR [rdx+360]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 46
-        mov	r9, QWORD PTR [rdx+368]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 47
-        mov	r9, QWORD PTR [rdx+376]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 48
-        mov	r9, QWORD PTR [rdx+384]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 49
-        mov	r9, QWORD PTR [rdx+392]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 50
-        mov	r9, QWORD PTR [rdx+400]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 51
-        mov	r9, QWORD PTR [rdx+408]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 52
-        mov	r9, QWORD PTR [rdx+416]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 53
-        mov	r9, QWORD PTR [rdx+424]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 54
-        mov	r9, QWORD PTR [rdx+432]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 55
-        mov	r9, QWORD PTR [rdx+440]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 56
-        mov	r9, QWORD PTR [rdx+448]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 57
-        mov	r9, QWORD PTR [rdx+456]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 58
-        mov	r9, QWORD PTR [rdx+464]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 59
-        mov	r9, QWORD PTR [rdx+472]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 60
-        mov	r9, QWORD PTR [rdx+480]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 61
-        mov	r9, QWORD PTR [rdx+488]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 62
-        mov	r9, QWORD PTR [rdx+496]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 63
-        mov	r9, QWORD PTR [rdx+504]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        ; END: 24-31
-        movdqu	xmm6, OWORD PTR [rsp]
-        movdqu	xmm7, OWORD PTR [rsp+16]
-        movdqu	xmm8, OWORD PTR [rsp+32]
-        movdqu	xmm9, OWORD PTR [rsp+48]
-        movdqu	xmm10, OWORD PTR [rsp+64]
-        movdqu	xmm11, OWORD PTR [rsp+80]
-        movdqu	xmm12, OWORD PTR [rsp+96]
-        movdqu	xmm13, OWORD PTR [rsp+112]
-        add	rsp, 128
-        ret
-sp_2048_get_from_table_32 ENDP
-_TEXT ENDS
-ENDIF
 IFDEF HAVE_INTEL_AVX2
 ; /* Reduce the number back to 2048 bits using Montgomery reduction.
 ;  *
@@ -31834,1768 +26038,6 @@ sp_3072_cmp_24 PROC
         ret
 sp_3072_cmp_24 ENDP
 _TEXT ENDS
-IFNDEF WC_NO_CACHE_RESISTANT
-_TEXT SEGMENT READONLY PARA
-sp_3072_get_from_table_24 PROC
-        sub	rsp, 128
-        movdqu	OWORD PTR [rsp], xmm6
-        movdqu	OWORD PTR [rsp+16], xmm7
-        movdqu	OWORD PTR [rsp+32], xmm8
-        movdqu	OWORD PTR [rsp+48], xmm9
-        movdqu	OWORD PTR [rsp+64], xmm10
-        movdqu	OWORD PTR [rsp+80], xmm11
-        movdqu	OWORD PTR [rsp+96], xmm12
-        movdqu	OWORD PTR [rsp+112], xmm13
-        mov	rax, 1
-        movd	xmm10, r8
-        movd	xmm11, rax
-        pxor	xmm13, xmm13
-        pshufd	xmm11, xmm11, 0
-        pshufd	xmm10, xmm10, 0
-        ; START: 0-7
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 16
-        mov	r9, QWORD PTR [rdx+128]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 17
-        mov	r9, QWORD PTR [rdx+136]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 18
-        mov	r9, QWORD PTR [rdx+144]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 19
-        mov	r9, QWORD PTR [rdx+152]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 20
-        mov	r9, QWORD PTR [rdx+160]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 21
-        mov	r9, QWORD PTR [rdx+168]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 22
-        mov	r9, QWORD PTR [rdx+176]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 23
-        mov	r9, QWORD PTR [rdx+184]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 24
-        mov	r9, QWORD PTR [rdx+192]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 25
-        mov	r9, QWORD PTR [rdx+200]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 26
-        mov	r9, QWORD PTR [rdx+208]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 27
-        mov	r9, QWORD PTR [rdx+216]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 28
-        mov	r9, QWORD PTR [rdx+224]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 29
-        mov	r9, QWORD PTR [rdx+232]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 30
-        mov	r9, QWORD PTR [rdx+240]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 31
-        mov	r9, QWORD PTR [rdx+248]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        add	rcx, 64
-        ; END: 0-7
-        ; START: 8-15
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 16
-        mov	r9, QWORD PTR [rdx+128]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 17
-        mov	r9, QWORD PTR [rdx+136]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 18
-        mov	r9, QWORD PTR [rdx+144]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 19
-        mov	r9, QWORD PTR [rdx+152]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 20
-        mov	r9, QWORD PTR [rdx+160]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 21
-        mov	r9, QWORD PTR [rdx+168]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 22
-        mov	r9, QWORD PTR [rdx+176]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 23
-        mov	r9, QWORD PTR [rdx+184]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 24
-        mov	r9, QWORD PTR [rdx+192]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 25
-        mov	r9, QWORD PTR [rdx+200]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 26
-        mov	r9, QWORD PTR [rdx+208]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 27
-        mov	r9, QWORD PTR [rdx+216]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 28
-        mov	r9, QWORD PTR [rdx+224]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 29
-        mov	r9, QWORD PTR [rdx+232]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 30
-        mov	r9, QWORD PTR [rdx+240]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 31
-        mov	r9, QWORD PTR [rdx+248]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        add	rcx, 64
-        ; END: 8-15
-        ; START: 16-23
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 16
-        mov	r9, QWORD PTR [rdx+128]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 17
-        mov	r9, QWORD PTR [rdx+136]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 18
-        mov	r9, QWORD PTR [rdx+144]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 19
-        mov	r9, QWORD PTR [rdx+152]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 20
-        mov	r9, QWORD PTR [rdx+160]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 21
-        mov	r9, QWORD PTR [rdx+168]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 22
-        mov	r9, QWORD PTR [rdx+176]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 23
-        mov	r9, QWORD PTR [rdx+184]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 24
-        mov	r9, QWORD PTR [rdx+192]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 25
-        mov	r9, QWORD PTR [rdx+200]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 26
-        mov	r9, QWORD PTR [rdx+208]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 27
-        mov	r9, QWORD PTR [rdx+216]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 28
-        mov	r9, QWORD PTR [rdx+224]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 29
-        mov	r9, QWORD PTR [rdx+232]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 30
-        mov	r9, QWORD PTR [rdx+240]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 31
-        mov	r9, QWORD PTR [rdx+248]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        ; END: 16-23
-        movdqu	xmm6, OWORD PTR [rsp]
-        movdqu	xmm7, OWORD PTR [rsp+16]
-        movdqu	xmm8, OWORD PTR [rsp+32]
-        movdqu	xmm9, OWORD PTR [rsp+48]
-        movdqu	xmm10, OWORD PTR [rsp+64]
-        movdqu	xmm11, OWORD PTR [rsp+80]
-        movdqu	xmm12, OWORD PTR [rsp+96]
-        movdqu	xmm13, OWORD PTR [rsp+112]
-        add	rsp, 128
-        ret
-sp_3072_get_from_table_24 ENDP
-_TEXT ENDS
-ENDIF
 IFDEF HAVE_INTEL_AVX2
 ; /* Reduce the number back to 3072 bits using Montgomery reduction.
 ;  *
@@ -36850,1820 +29292,6 @@ sp_3072_cmp_48 PROC
         ret
 sp_3072_cmp_48 ENDP
 _TEXT ENDS
-IFNDEF WC_NO_CACHE_RESISTANT
-_TEXT SEGMENT READONLY PARA
-sp_3072_get_from_table_48 PROC
-        sub	rsp, 128
-        movdqu	OWORD PTR [rsp], xmm6
-        movdqu	OWORD PTR [rsp+16], xmm7
-        movdqu	OWORD PTR [rsp+32], xmm8
-        movdqu	OWORD PTR [rsp+48], xmm9
-        movdqu	OWORD PTR [rsp+64], xmm10
-        movdqu	OWORD PTR [rsp+80], xmm11
-        movdqu	OWORD PTR [rsp+96], xmm12
-        movdqu	OWORD PTR [rsp+112], xmm13
-        mov	rax, 1
-        movd	xmm10, r8
-        movd	xmm11, rax
-        pxor	xmm13, xmm13
-        pshufd	xmm11, xmm11, 0
-        pshufd	xmm10, xmm10, 0
-        ; START: 0-7
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        add	rcx, 64
-        ; END: 0-7
-        ; START: 8-15
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        add	rcx, 64
-        ; END: 8-15
-        ; START: 16-23
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        add	rcx, 64
-        ; END: 16-23
-        ; START: 24-31
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        add	rcx, 64
-        ; END: 24-31
-        ; START: 32-39
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        add	rcx, 64
-        ; END: 32-39
-        ; START: 40-47
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        ; END: 40-47
-        movdqu	xmm6, OWORD PTR [rsp]
-        movdqu	xmm7, OWORD PTR [rsp+16]
-        movdqu	xmm8, OWORD PTR [rsp+32]
-        movdqu	xmm9, OWORD PTR [rsp+48]
-        movdqu	xmm10, OWORD PTR [rsp+64]
-        movdqu	xmm11, OWORD PTR [rsp+80]
-        movdqu	xmm12, OWORD PTR [rsp+96]
-        movdqu	xmm13, OWORD PTR [rsp+112]
-        add	rsp, 128
-        ret
-sp_3072_get_from_table_48 ENDP
-_TEXT ENDS
-ENDIF
 IFDEF HAVE_INTEL_AVX2
 ; /* Reduce the number back to 3072 bits using Montgomery reduction.
 ;  *
@@ -49221,2420 +39849,6 @@ sp_4096_cmp_64 PROC
         ret
 sp_4096_cmp_64 ENDP
 _TEXT ENDS
-IFNDEF WC_NO_CACHE_RESISTANT
-_TEXT SEGMENT READONLY PARA
-sp_4096_get_from_table_64 PROC
-        sub	rsp, 128
-        movdqu	OWORD PTR [rsp], xmm6
-        movdqu	OWORD PTR [rsp+16], xmm7
-        movdqu	OWORD PTR [rsp+32], xmm8
-        movdqu	OWORD PTR [rsp+48], xmm9
-        movdqu	OWORD PTR [rsp+64], xmm10
-        movdqu	OWORD PTR [rsp+80], xmm11
-        movdqu	OWORD PTR [rsp+96], xmm12
-        movdqu	OWORD PTR [rsp+112], xmm13
-        mov	rax, 1
-        movd	xmm10, r8
-        movd	xmm11, rax
-        pxor	xmm13, xmm13
-        pshufd	xmm11, xmm11, 0
-        pshufd	xmm10, xmm10, 0
-        ; START: 0-7
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        add	rcx, 64
-        ; END: 0-7
-        ; START: 8-15
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        add	r9, 64
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        add	rcx, 64
-        ; END: 8-15
-        ; START: 16-23
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        add	r9, 128
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        add	rcx, 64
-        ; END: 16-23
-        ; START: 24-31
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        add	r9, 192
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        add	rcx, 64
-        ; END: 24-31
-        ; START: 32-39
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        add	r9, 256
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        add	rcx, 64
-        ; END: 32-39
-        ; START: 40-47
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        add	r9, 320
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        add	rcx, 64
-        ; END: 40-47
-        ; START: 48-55
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        add	r9, 384
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        add	r9, 384
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        add	r9, 384
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        add	r9, 384
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        add	r9, 384
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        add	r9, 384
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        add	r9, 384
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        add	r9, 384
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        add	r9, 384
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        add	r9, 384
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        add	r9, 384
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        add	r9, 384
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        add	r9, 384
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        add	r9, 384
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        add	r9, 384
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        add	r9, 384
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        add	rcx, 64
-        ; END: 48-55
-        ; START: 56-63
-        pxor	xmm13, xmm13
-        pxor	xmm4, xmm4
-        pxor	xmm5, xmm5
-        pxor	xmm6, xmm6
-        pxor	xmm7, xmm7
-        ; ENTRY: 0
-        mov	r9, QWORD PTR [rdx]
-        add	r9, 448
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 1
-        mov	r9, QWORD PTR [rdx+8]
-        add	r9, 448
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 2
-        mov	r9, QWORD PTR [rdx+16]
-        add	r9, 448
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 3
-        mov	r9, QWORD PTR [rdx+24]
-        add	r9, 448
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 4
-        mov	r9, QWORD PTR [rdx+32]
-        add	r9, 448
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 5
-        mov	r9, QWORD PTR [rdx+40]
-        add	r9, 448
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 6
-        mov	r9, QWORD PTR [rdx+48]
-        add	r9, 448
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 7
-        mov	r9, QWORD PTR [rdx+56]
-        add	r9, 448
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 8
-        mov	r9, QWORD PTR [rdx+64]
-        add	r9, 448
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 9
-        mov	r9, QWORD PTR [rdx+72]
-        add	r9, 448
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 10
-        mov	r9, QWORD PTR [rdx+80]
-        add	r9, 448
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 11
-        mov	r9, QWORD PTR [rdx+88]
-        add	r9, 448
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 12
-        mov	r9, QWORD PTR [rdx+96]
-        add	r9, 448
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 13
-        mov	r9, QWORD PTR [rdx+104]
-        add	r9, 448
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 14
-        mov	r9, QWORD PTR [rdx+112]
-        add	r9, 448
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        ; ENTRY: 15
-        mov	r9, QWORD PTR [rdx+120]
-        add	r9, 448
-        movdqu	xmm12, xmm13
-        pcmpeqd	xmm12, xmm10
-        movdqu	xmm0, OWORD PTR [r9]
-        movdqu	xmm1, OWORD PTR [r9+16]
-        movdqu	xmm2, OWORD PTR [r9+32]
-        movdqu	xmm3, OWORD PTR [r9+48]
-        pand	xmm0, xmm12
-        pand	xmm1, xmm12
-        pand	xmm2, xmm12
-        pand	xmm3, xmm12
-        por	xmm4, xmm0
-        por	xmm5, xmm1
-        por	xmm6, xmm2
-        por	xmm7, xmm3
-        paddd	xmm13, xmm11
-        movdqu	OWORD PTR [rcx], xmm4
-        movdqu	OWORD PTR [rcx+16], xmm5
-        movdqu	OWORD PTR [rcx+32], xmm6
-        movdqu	OWORD PTR [rcx+48], xmm7
-        ; END: 56-63
-        movdqu	xmm6, OWORD PTR [rsp]
-        movdqu	xmm7, OWORD PTR [rsp+16]
-        movdqu	xmm8, OWORD PTR [rsp+32]
-        movdqu	xmm9, OWORD PTR [rsp+48]
-        movdqu	xmm10, OWORD PTR [rsp+64]
-        movdqu	xmm11, OWORD PTR [rsp+80]
-        movdqu	xmm12, OWORD PTR [rsp+96]
-        movdqu	xmm13, OWORD PTR [rsp+112]
-        add	rsp, 128
-        ret
-sp_4096_get_from_table_64 ENDP
-_TEXT ENDS
-ENDIF
 IFDEF HAVE_INTEL_AVX2
 ; /* Reduce the number back to 4096 bits using Montgomery reduction.
 ;  *
diff --git a/wolfcrypt/src/wolfmath.c b/wolfcrypt/src/wolfmath.c
index 06fb8ed017..f073be133e 100644
--- a/wolfcrypt/src/wolfmath.c
+++ b/wolfcrypt/src/wolfmath.c
@@ -43,12 +43,13 @@
 
 #if !defined(NO_BIG_INT) || defined(WOLFSSL_SP_MATH)
 
-#if (!defined(WC_NO_CACHE_RESISTANT) && \
-     ((defined(HAVE_ECC) && defined(ECC_TIMING_RESISTANT)) || \
-      (defined(USE_FAST_MATH) && defined(TFM_TIMING_RESISTANT)))) || \
-    ((defined(WOLFSSL_SP_MATH_ALL) && !defined(WOLFSSL_RSA_VERIFY_ONLY) && \
-      !defined(WOLFSSL_RSA_PUBLIC_ONLY)) || !defined(NO_DH) || \
-     defined(OPENSSL_ALL) && defined(WC_PROTECT_ENCRYPTED_MEM))
+#if !defined(WC_NO_GLOBAL_OBJECT_POINTERS) && \
+    ((!defined(WC_NO_CACHE_RESISTANT) && \
+      ((defined(HAVE_ECC) && defined(ECC_TIMING_RESISTANT)) || \
+       (defined(USE_FAST_MATH) && defined(TFM_TIMING_RESISTANT)))) || \
+     ((defined(WOLFSSL_SP_MATH_ALL) && !defined(WOLFSSL_RSA_VERIFY_ONLY) && \
+       !defined(WOLFSSL_RSA_PUBLIC_ONLY)) || !defined(NO_DH) || \
+      defined(OPENSSL_ALL) && defined(WC_PROTECT_ENCRYPTED_MEM)))
 
     /* all off / all on pointer addresses for constant calculations */
     /* ecc.c uses same table */
diff --git a/wolfcrypt/test/test.c b/wolfcrypt/test/test.c
index 329c88cffd..cd3275650b 100644
--- a/wolfcrypt/test/test.c
+++ b/wolfcrypt/test/test.c
@@ -44767,7 +44767,9 @@ static wc_test_ret_t ed25519_test_check_key(void)
         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
     };
-    /* Y-ordinate value equal to prime - 1. */
+    /* Y-ordinate value equal to prime - 1. Older FIPS modules accept
+     * this as a valid key; the current source rejects it as an order-2
+     * point. */
     WOLFSSL_SMALL_STACK_STATIC const byte key_y_is_p_minus_1[] = {
         0x40,
         0xec,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
@@ -44775,6 +44777,15 @@ static wc_test_ret_t ed25519_test_check_key(void)
         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x7f,
     };
+    /* RFC 8032 section 7.1 test-vector public key: a genuinely valid
+     * Ed25519 point used as a positive control. */
+    WOLFSSL_SMALL_STACK_STATIC const byte key_good[] = {
+        0x40,
+        0xd7,0x5a,0x98,0x01,0x82,0xb1,0x0a,0xb7,
+        0xd5,0x4b,0xfe,0xd3,0xc9,0x64,0x07,0x3a,
+        0x0e,0xe1,0x72,0xf3,0xda,0xa6,0x23,0x25,
+        0xaf,0x02,0x1a,0x68,0xf7,0x07,0x51,0x1a,
+    };
     ed25519_key key;
     int ret;
     int res = 0;
@@ -44807,9 +44818,26 @@ static wc_test_ret_t ed25519_test_check_key(void)
         }
     }
     if (res == 0) {
-        /* Load good public key only and perform checks. */
         ret = wc_ed25519_import_public(key_y_is_p_minus_1,
             ED25519_PUB_KEY_SIZE + 1, &key);
+#if !defined(HAVE_FIPS) || FIPS_VERSION3_GE(7,0,0)
+        /* y = p - 1 is an order-2 point; check_key rejects it because
+         * h*A is the neutral element for small-order public keys and
+         * forged signatures would otherwise verify. */
+        if (ret != WC_NO_ERR_TRACE(PUBLIC_KEY_E)) {
+            res = WC_TEST_RET_ENC_NC;
+        }
+#else
+        /* Older FIPS modules accept this order-2 point. */
+        if (ret != 0) {
+            res = WC_TEST_RET_ENC_NC;
+        }
+#endif
+    }
+    if (res == 0) {
+        /* Positive control: a real Ed25519 public key must be accepted. */
+        ret = wc_ed25519_import_public(key_good, ED25519_PUB_KEY_SIZE + 1,
+            &key);
         if (ret != 0) {
             res = WC_TEST_RET_ENC_NC;
         }
@@ -46484,7 +46512,9 @@ static wc_test_ret_t ed448_test_check_key(void)
         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
         0xff
     };
-    /* Y-ordinate value equal to prime - 1. */
+    /* Y-ordinate value equal to prime - 1. Older FIPS modules accept
+     * this as a valid key; the current source rejects it as an order-2
+     * point. */
     WOLFSSL_SMALL_STACK_STATIC const byte key_y_is_p_minus_1[] = {
         0x40,
         0xfe,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
@@ -46496,6 +46526,19 @@ static wc_test_ret_t ed448_test_check_key(void)
         0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,
         0xff
     };
+    /* RFC 8032 section 7.4 test-vector public key: a genuinely valid
+     * Ed448 point used as a positive control. */
+    WOLFSSL_SMALL_STACK_STATIC const byte key_good[] = {
+        0x40,
+        0x5f,0xd7,0x44,0x9b,0x59,0xb4,0x61,0xfd,
+        0x2c,0xe7,0x87,0xec,0x61,0x6a,0xd4,0x6a,
+        0x1d,0xa1,0x34,0x24,0x85,0xa7,0x0e,0x1f,
+        0x8a,0x0e,0xa7,0x5d,0x80,0xe9,0x67,0x78,
+        0xed,0xf1,0x24,0x76,0x9b,0x46,0xc7,0x06,
+        0x1b,0xd6,0x78,0x3d,0xf1,0xe5,0x0f,0x6c,
+        0xd1,0xfa,0x1a,0xbe,0xaf,0xe8,0x25,0x61,
+        0x80
+    };
     ed448_key key;
     int ret;
     int res = 0;
@@ -46528,9 +46571,25 @@ static wc_test_ret_t ed448_test_check_key(void)
         }
     }
     if (res == 0) {
-        /* Load good public key only and perform checks. */
         ret = wc_ed448_import_public(key_y_is_p_minus_1, ED448_PUB_KEY_SIZE + 1,
             &key);
+#if !defined(HAVE_FIPS) || FIPS_VERSION3_GE(7,0,0)
+        /* y = p - 1 is an order-2 point; check_key rejects it because
+         * h*A is the neutral element for small-order public keys and
+         * forged signatures would otherwise verify. */
+        if (ret != WC_NO_ERR_TRACE(PUBLIC_KEY_E)) {
+            res = WC_TEST_RET_ENC_NC;
+        }
+#else
+        /* Older FIPS modules accept this order-2 point. */
+        if (ret != 0) {
+            res = WC_TEST_RET_ENC_NC;
+        }
+#endif
+    }
+    if (res == 0) {
+        /* Positive control: a real Ed448 public key must be accepted. */
+        ret = wc_ed448_import_public(key_good, ED448_PUB_KEY_SIZE + 1, &key);
         if (ret != 0) {
             res = WC_TEST_RET_ENC_NC;
         }
@@ -60460,38 +60519,25 @@ WOLFSSL_TEST_SUBROUTINE wc_test_ret_t eccsi_test(void)
     wc_test_ret_t ret = 0;
     WC_RNG rng;
     int rng_inited = 0;
-    EccsiKey* priv = NULL;
-    EccsiKey* pub  = NULL;
-    mp_int* ssk    = NULL;
+    WC_DECLARE_VAR(priv, EccsiKey, 1, HEAP_HINT);
+    WC_DECLARE_VAR(pub, EccsiKey, 1, HEAP_HINT);
+    WC_DECLARE_VAR(ssk, mp_int, 1, HEAP_HINT);
     ecc_point* pvt = NULL;
     WOLFSSL_ENTER("eccsi_test");
 
-    priv = (EccsiKey*)XMALLOC(sizeof(EccsiKey), HEAP_HINT,
-            DYNAMIC_TYPE_TMP_BUFFER);
-    if (priv == NULL)
-        ret = WC_TEST_RET_ENC_NC;
-    else
+    WC_ALLOC_VAR_EX(priv, EccsiKey, 1, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER, ret = WC_TEST_RET_ENC_EC(MEMORY_E));
+    if (ret == 0) {
         XMEMSET(priv, 0, sizeof(*priv));
-
+        WC_ALLOC_VAR_EX(pub, EccsiKey, 1, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER, ret = WC_TEST_RET_ENC_EC(MEMORY_E));
+    }
     if (ret == 0) {
-        pub = (EccsiKey*)XMALLOC(sizeof(EccsiKey), HEAP_HINT,
-            DYNAMIC_TYPE_TMP_BUFFER);
-        if (pub == NULL)
-            ret = WC_TEST_RET_ENC_NC;
-        else
-            XMEMSET(pub, 0, sizeof(*pub));
+        XMEMSET(pub, 0, sizeof(*pub));
+        WC_ALLOC_VAR_EX(ssk, mp_int, 1, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER, ret = WC_TEST_RET_ENC_EC(MEMORY_E));
     }
 
     if (ret == 0) {
-        ssk = (mp_int*)XMALLOC(sizeof(mp_int), HEAP_HINT,
-                DYNAMIC_TYPE_TMP_BUFFER);
-        if (ssk == NULL)
-            ret = WC_TEST_RET_ENC_NC;
-        else
-            XMEMSET(ssk, 0, sizeof(*ssk));
-    }
+        XMEMSET(ssk, 0, sizeof(*ssk));
 
-    if (ret == 0) {
     #ifndef HAVE_FIPS
         ret = wc_InitRng_ex(&rng, HEAP_HINT, devId);
     #else
@@ -60547,18 +60593,12 @@ WOLFSSL_TEST_SUBROUTINE wc_test_ret_t eccsi_test(void)
         wc_ecc_del_point(pvt);
     if (rng_inited)
         wc_FreeRng(&rng);
-    if (ssk != NULL) {
-        mp_free(ssk);
-        XFREE(ssk, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
-    }
-    if (pub != NULL) {
-        wc_FreeEccsiKey(pub);
-        XFREE(pub, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
-    }
-    if (priv != NULL) {
-        wc_FreeEccsiKey(priv);
-        XFREE(priv, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
-    }
+    mp_free(ssk);
+    WC_FREE_VAR_EX(ssk, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
+    wc_FreeEccsiKey(pub);
+    WC_FREE_VAR_EX(pub, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
+    wc_FreeEccsiKey(priv);
+    WC_FREE_VAR_EX(priv, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
 
     return ret;
 }
@@ -61704,38 +61744,25 @@ WOLFSSL_TEST_SUBROUTINE wc_test_ret_t sakke_test(void)
     wc_test_ret_t ret = 0;
     WC_RNG rng;
     int rng_inited = 0;
-    SakkeKey* priv = NULL;
-    SakkeKey* pub  = NULL;
-    SakkeKey* key  = NULL;
+    WC_DECLARE_VAR(priv, SakkeKey, 1, HEAP_HINT);
+    WC_DECLARE_VAR(pub, SakkeKey, 1, HEAP_HINT);
+    WC_DECLARE_VAR(key, SakkeKey, 1, HEAP_HINT);
     ecc_point* rsk = NULL;
     WOLFSSL_ENTER("sakke_test");
 
-    priv = (SakkeKey*)XMALLOC(sizeof(SakkeKey), HEAP_HINT,
-            DYNAMIC_TYPE_TMP_BUFFER);
-    if (priv == NULL)
-        ret = WC_TEST_RET_ENC_NC;
-    else
+    WC_ALLOC_VAR_EX(priv, SakkeKey, 1, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER, ret = WC_TEST_RET_ENC_EC(MEMORY_E));
+    if (ret == 0) {
         XMEMSET(priv, 0, sizeof(*priv));
-
+        WC_ALLOC_VAR_EX(pub, SakkeKey, 1, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER, ret = WC_TEST_RET_ENC_EC(MEMORY_E));
+    }
     if (ret == 0) {
-        pub = (SakkeKey*)XMALLOC(sizeof(SakkeKey), HEAP_HINT,
-            DYNAMIC_TYPE_TMP_BUFFER);
-        if (pub == NULL)
-            ret = WC_TEST_RET_ENC_NC;
-        else
-            XMEMSET(pub, 0, sizeof(*pub));
+        XMEMSET(pub, 0, sizeof(*pub));
+        WC_ALLOC_VAR_EX(key, SakkeKey, 1, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER, ret = WC_TEST_RET_ENC_EC(MEMORY_E));
     }
 
     if (ret == 0) {
-        key = (SakkeKey*)XMALLOC(sizeof(SakkeKey), HEAP_HINT,
-            DYNAMIC_TYPE_TMP_BUFFER);
-        if (key == NULL)
-            ret = WC_TEST_RET_ENC_NC;
-        else
-            XMEMSET(key, 0, sizeof(*key));
-    }
+        XMEMSET(key, 0, sizeof(*key));
 
-    if (ret == 0) {
     #ifndef HAVE_FIPS
         ret = wc_InitRng_ex(&rng, HEAP_HINT, devId);
     #else
@@ -61791,15 +61818,12 @@ WOLFSSL_TEST_SUBROUTINE wc_test_ret_t sakke_test(void)
     }
     if (rng_inited)
         wc_FreeRng(&rng);
-    XFREE(key, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
-    if (pub != NULL) {
-        wc_FreeSakkeKey(pub);
-        XFREE(pub, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
-    }
-    if (priv != NULL) {
-        wc_FreeSakkeKey(priv);
-        XFREE(priv, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
-    }
+
+    WC_FREE_VAR_EX(key, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
+    wc_FreeSakkeKey(pub);
+    WC_FREE_VAR_EX(pub, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
+    wc_FreeSakkeKey(priv);
+    WC_FREE_VAR_EX(priv, HEAP_HINT, DYNAMIC_TYPE_TMP_BUFFER);
 
     return ret;
 }
diff --git a/wolfssl/internal.h b/wolfssl/internal.h
index 025878a8d3..794808f38b 100644
--- a/wolfssl/internal.h
+++ b/wolfssl/internal.h
@@ -1419,7 +1419,7 @@ enum {
 /* Limit is 2^24.5
  * https://www.rfc-editor.org/rfc/rfc8446#section-5.5
  * Without the fraction is 23726566 (0x016A09E6) */
-#define AEAD_AES_LIMIT                           w64From32(0x016A, 0x09E6)
+#define AEAD_AES_LIMIT                           w64From32(0, 0x016A09E6)
 /* Limit is 2^23
  * https://www.rfc-editor.org/rfc/rfc9147.html#name-integrity-limits */
 #define DTLS_AEAD_AES_CCM_LIMIT                  w64From32(0, 1 << 22)
@@ -1436,8 +1436,8 @@ enum {
  * https://www.rfc-editor.org/rfc/rfc9147.html#name-integrity-limits
  * Without the fraction is 11863283 (0x00B504F3)
  * Half of this value is    5931641 (0x005A8279) */
-#define DTLS_AEAD_AES_CCM_FAIL_LIMIT             w64From32(0x00B5, 0x04F3)
-#define DTLS_AEAD_AES_CCM_FAIL_KU_LIMIT          w64From32(0x005A, 0x8279)
+#define DTLS_AEAD_AES_CCM_FAIL_LIMIT             w64From32(0, 0x00B504F3)
+#define DTLS_AEAD_AES_CCM_FAIL_KU_LIMIT          w64From32(0, 0x005A8279)
 
 /* Limit is (2^22 - 1) full messages [2^36 - 31 octets]
  * https://www.rfc-editor.org/rfc/rfc8998.html#name-aead_sm4_gcm
@@ -1632,25 +1632,28 @@ enum Misc {
     DTLS_EXPORT_PRO          = 165,/* wolfSSL protocol for serialized session */
     DTLS_EXPORT_STATE_PRO    = 166,/* wolfSSL protocol for serialized state */
     TLS_EXPORT_PRO           = 167,/* wolfSSL protocol for serialized TLS */
-    DTLS_EXPORT_OPT_SZ       = 62, /* amount of bytes used from Options */
-    DTLS_EXPORT_OPT_SZ_4     = 61, /* amount of bytes used from Options */
-    TLS_EXPORT_OPT_SZ        = 66, /* amount of bytes used from Options */
-    TLS_EXPORT_OPT_SZ_4      = 65, /* amount of bytes used from Options */
-    DTLS_EXPORT_OPT_SZ_3     = 60, /* amount of bytes used from Options */
+    DTLS_EXPORT_OPT_SZ       = 66, /* number of bytes used from Options */
+    DTLS_EXPORT_OPT_SZ_5     = 62, /* number of bytes used from Options */
+    DTLS_EXPORT_OPT_SZ_4     = 61, /* number of bytes used from Options */
+    TLS_EXPORT_OPT_SZ        = 66, /* number of bytes used from Options */
+    TLS_EXPORT_OPT_SZ_5      = 66, /* number of bytes used from Options */
+    TLS_EXPORT_OPT_SZ_4      = 65, /* number of bytes used from Options */
+    DTLS_EXPORT_OPT_SZ_3     = 60, /* number of bytes used from Options */
     DTLS_EXPORT_KEY_SZ       = 325 + (DTLS_SEQ_SZ * 2),
-                                   /* max amount of bytes used from Keys */
+                                   /* max number of bytes used from Keys */
     DTLS_EXPORT_MIN_KEY_SZ   = 85 + (DTLS_SEQ_SZ * 2),
-                                   /* min amount of bytes used from Keys */
+                                   /* min number of bytes used from Keys */
     WOLFSSL_EXPORT_TLS       = 1,
     WOLFSSL_EXPORT_DTLS      = 0,
 #ifndef WOLFSSL_EXPORT_SPC_SZ
-    WOLFSSL_EXPORT_SPC_SZ    = 16, /* amount of bytes used from CipherSpecs */
+    WOLFSSL_EXPORT_SPC_SZ    = 16, /* number of bytes used from CipherSpecs */
 #endif
     WOLFSSL_EXPORT_LEN       = 2,  /* 2 bytes for length and protocol */
-    WOLFSSL_EXPORT_VERSION   = 5,  /* wolfSSL version for serialized session */
+    WOLFSSL_EXPORT_VERSION   = 6,  /* wolfSSL version for serialized session */
 
-    WOLFSSL_EXPORT_VERSION_4 = 4,  /* 5.6.4 release and before */
     /* older export versions supported */
+    WOLFSSL_EXPORT_VERSION_5 = 5,  /* version before DTLS Encrypt-Then-MAC */
+    WOLFSSL_EXPORT_VERSION_4 = 4,  /* 5.6.4 release and before */
     WOLFSSL_EXPORT_VERSION_3 = 3,  /* wolfSSL version before TLS 1.3 addition */
 
     MAX_EXPORT_IP            = 46, /* max ip size IPv4 mapped IPv6 */
diff --git a/wolfssl/openssl/evp.h b/wolfssl/openssl/evp.h
index 9b90563cf8..bc873695bf 100644
--- a/wolfssl/openssl/evp.h
+++ b/wolfssl/openssl/evp.h
@@ -455,6 +455,12 @@ enum {
 #endif
 #ifdef HAVE_ED448
     WC_EVP_PKEY_ED448         = WC_NID_ED448,
+#endif
+#ifdef HAVE_CURVE25519
+    WC_EVP_PKEY_X25519        = WC_NID_X25519,
+#endif
+#ifdef HAVE_CURVE448
+    WC_EVP_PKEY_X448          = WC_NID_X448,
 #endif
     WC_AES_128_CFB1_TYPE      = 24,
     WC_AES_192_CFB1_TYPE      = 25,
@@ -529,6 +535,12 @@ enum {
 #ifdef HAVE_ED448
 #define EVP_PKEY_ED448 WC_EVP_PKEY_ED448
 #endif
+#ifdef HAVE_CURVE25519
+#define EVP_PKEY_X25519 WC_EVP_PKEY_X25519
+#endif
+#ifdef HAVE_CURVE448
+#define EVP_PKEY_X448 WC_EVP_PKEY_X448
+#endif
 #define AES_128_CFB1_TYPE WC_AES_128_CFB1_TYPE
 #define AES_192_CFB1_TYPE WC_AES_192_CFB1_TYPE
 #define AES_256_CFB1_TYPE WC_AES_256_CFB1_TYPE
diff --git a/wolfssl/openssl/pem.h b/wolfssl/openssl/pem.h
index 9ccc4731de..6ad42052b2 100644
--- a/wolfssl/openssl/pem.h
+++ b/wolfssl/openssl/pem.h
@@ -231,6 +231,13 @@ WOLFSSL_API
 int wolfSSL_PEM_write_X509(XFILE fp, WOLFSSL_X509 *x);
 WOLFSSL_API
 int wolfSSL_PEM_write_DHparams(XFILE fp, WOLFSSL_DH* dh);
+WOLFSSL_API
+int wolfSSL_PEM_write_PrivateKey(XFILE fp, WOLFSSL_EVP_PKEY* key,
+                                 const WOLFSSL_EVP_CIPHER* cipher,
+                                 unsigned char* passwd, int len,
+                                 wc_pem_password_cb* cb, void* arg);
+WOLFSSL_API
+int wolfSSL_PEM_write_PUBKEY(XFILE fp, WOLFSSL_EVP_PKEY* key);
 #endif /* NO_FILESYSTEM */
 
 #ifndef OPENSSL_COEXIST
@@ -244,6 +251,7 @@ int wolfSSL_PEM_write_DHparams(XFILE fp, WOLFSSL_DH* dh);
 
 #define PEM_read_X509                   wolfSSL_PEM_read_X509
 #define PEM_read_PrivateKey             wolfSSL_PEM_read_PrivateKey
+#define PEM_write_PrivateKey            wolfSSL_PEM_write_PrivateKey
 #define PEM_write_X509                  wolfSSL_PEM_write_X509
 #define PEM_write_bio_PrivateKey        wolfSSL_PEM_write_bio_PrivateKey
 #define PEM_write_bio_PKCS8PrivateKey   wolfSSL_PEM_write_bio_PKCS8PrivateKey
@@ -287,6 +295,7 @@ int wolfSSL_PEM_write_DHparams(XFILE fp, WOLFSSL_DH* dh);
 #define PEM_read_PUBKEY                 wolfSSL_PEM_read_PUBKEY
 #define PEM_read_bio_PUBKEY             wolfSSL_PEM_read_bio_PUBKEY
 #define PEM_write_bio_PUBKEY            wolfSSL_PEM_write_bio_PUBKEY
+#define PEM_write_PUBKEY                wolfSSL_PEM_write_PUBKEY
 
 #define PEM_write_bio_PKCS8_PRIV_KEY_INFO wolfSSL_PEM_write_bio_PKCS8_PRIV_KEY_INFO
 #define PEM_read_bio_PKCS8_PRIV_KEY_INFO wolfSSL_PEM_read_bio_PKCS8_PRIV_KEY_INFO
diff --git a/wolfssl/ssl.h b/wolfssl/ssl.h
index 0081cfc494..2281bb2f26 100644
--- a/wolfssl/ssl.h
+++ b/wolfssl/ssl.h
@@ -609,6 +609,12 @@ struct WOLFSSL_EVP_PKEY {
     #ifdef HAVE_ED448
     struct ed448_key* ed448;
     #endif
+    #ifdef HAVE_CURVE25519
+    struct curve25519_key* curve25519;
+    #endif
+    #ifdef HAVE_CURVE448
+    struct curve448_key* curve448;
+    #endif
     WC_RNG rng;
     #ifdef HAVE_HKDF
     const WOLFSSL_EVP_MD* hkdfMd;
@@ -640,6 +646,12 @@ struct WOLFSSL_EVP_PKEY {
 #ifdef HAVE_ED448
     WC_BITFIELD ownEd448:1;   /* if struct owns Ed448 and should free it */
 #endif
+#ifdef HAVE_CURVE25519
+    WC_BITFIELD ownCurve25519:1; /* if struct owns X25519 and should free it */
+#endif
+#ifdef HAVE_CURVE448
+    WC_BITFIELD ownCurve448:1;   /* if struct owns X448 and should free it */
+#endif
 };
 
 
diff --git a/wolfssl/wolfcrypt/aes.h b/wolfssl/wolfcrypt/aes.h
index 0afc129009..e3d7637470 100644
--- a/wolfssl/wolfcrypt/aes.h
+++ b/wolfssl/wolfcrypt/aes.h
@@ -1112,6 +1112,53 @@ WOLFSSL_LOCAL void AES_XTS_decrypt_AARCH32(const byte* in, byte* out,
 #endif /* !__aarch64__ && !WOLFSSL_ARMASM_NO_HW_CRYPTO */
 #endif /* WOLFSSL_ARMASM */
 
+#if defined(WOLFSSL_PPC64_ASM)
+WOLFSSL_LOCAL void AES_set_encrypt_key(const unsigned char* key, word32 len,
+    unsigned char* ks);
+WOLFSSL_LOCAL void AES_invert_key(unsigned char* ks, word32 rounds);
+WOLFSSL_LOCAL void AES_ECB_encrypt(const unsigned char* in, unsigned char* out,
+    unsigned long len, const unsigned char* ks, int nr);
+WOLFSSL_LOCAL void AES_ECB_decrypt(const unsigned char* in, unsigned char* out,
+    unsigned long len, const unsigned char* ks, int nr);
+WOLFSSL_LOCAL void AES_CBC_encrypt(const unsigned char* in, unsigned char* out,
+    unsigned long len, const unsigned char* ks, int nr, unsigned char* iv);
+WOLFSSL_LOCAL void AES_CBC_decrypt(const unsigned char* in, unsigned char* out,
+    unsigned long len, const unsigned char* ks, int nr, unsigned char* iv);
+WOLFSSL_LOCAL void AES_CTR_encrypt(const unsigned char* in, unsigned char* out,
+    unsigned long len, const unsigned char* ks, int nr, unsigned char* ctr);
+#if defined(GCM_TABLE) || defined(GCM_TABLE_4BIT)
+/* in pre-C2x C, constness conflicts for dimensioned arrays can't be resolved.
+ */
+WOLFSSL_LOCAL void GCM_gmult_len(byte* x, const byte** m,
+    const unsigned char* data, unsigned long len);
+#endif
+WOLFSSL_LOCAL void AES_GCM_encrypt(const unsigned char* in, unsigned char* out,
+    unsigned long len, const unsigned char* ks, int nr, unsigned char* ctr);
+
+#if defined(BUILDING_WOLFSSL)
+ WOLFSSL_API WARN_UNUSED_RESULT int wc_AesEncryptDirect(Aes* aes, byte* out,
+                                                        const byte* in);
+ WOLFSSL_API WARN_UNUSED_RESULT int wc_AesDecryptDirect(Aes* aes, byte* out,
+                                                        const byte* in);
+ WOLFSSL_API WARN_UNUSED_RESULT int wc_AesSetKeyDirect(Aes* aes,
+                                                       const byte* key,
+                                                       word32 len,
+                                const byte* iv, int dir);
+#else
+ WOLFSSL_API int wc_AesEncryptDirect(Aes* aes, byte* out, const byte* in);
+ WOLFSSL_API int wc_AesDecryptDirect(Aes* aes, byte* out, const byte* in);
+ WOLFSSL_API int wc_AesSetKeyDirect(Aes* aes, const byte* key, word32 len,
+                                const byte* iv, int dir);
+#endif
+
+#if defined(WOLFSSL_AES_XTS)
+WOLFSSL_LOCAL void AES_XTS_encrypt(const byte* in, byte* out, word32 sz,
+    const byte* i, byte* key, byte* key2, byte* tmp, int nr);
+WOLFSSL_LOCAL void AES_XTS_decrypt(const byte* in, byte* out, word32 sz,
+    const byte* i, byte* key, byte* key2, byte* tmp, int nr);
+#endif
+#endif /* WOLFSSL_PPC64_ASM */
+
 #ifdef __cplusplus
     } /* extern "C" */
 #endif
diff --git a/wolfssl/wolfcrypt/asn.h b/wolfssl/wolfcrypt/asn.h
index 5db17afab1..70f823fc80 100644
--- a/wolfssl/wolfcrypt/asn.h
+++ b/wolfssl/wolfcrypt/asn.h
@@ -3213,6 +3213,13 @@ WOLFSSL_TEST_VIS int  wolfssl_local_MatchBaseName(int type, const char* name,
 WOLFSSL_TEST_VIS int  wolfssl_local_MatchIpSubnet(const byte* ip, int ipSz,
                                                   const byte* constraint,
                                                   int constraintSz);
+WOLFSSL_TEST_VIS int  wolfssl_local_MatchUriNameConstraint(const char* uri,
+                                                  int uriSz, const char* base,
+                                                  int baseSz);
+WOLFSSL_TEST_VIS int  wolfssl_local_MatchDnsConstraintWildcard(
+                                                  const char* name, int nameSz,
+                                                  const char* base, int baseSz,
+                                                  int permitted);
 #endif
 
 #if ((defined(HAVE_ED25519) && defined(HAVE_ED25519_KEY_IMPORT)) \
diff --git a/wolfssl/wolfcrypt/chacha.h b/wolfssl/wolfcrypt/chacha.h
index 5ae021fa43..a2467ba1f9 100644
--- a/wolfssl/wolfcrypt/chacha.h
+++ b/wolfssl/wolfcrypt/chacha.h
@@ -86,9 +86,10 @@ typedef struct ChaCha {
     byte extra[12];
 #endif
     word32 left;                            /* number of bytes leftover */
-#if defined(USE_INTEL_CHACHA_SPEEDUP) || defined(USE_ARM_CHACHA_SPEEDUP) || \
-    defined(WOLFSSL_RISCV_ASM)
+#if defined(USE_INTEL_CHACHA_SPEEDUP) || defined(USE_ARM_CHACHA_SPEEDUP)
     word32 over[CHACHA_CHUNK_WORDS];
+#elif defined(WOLFSSL_RISCV_ASM)
+    ALIGN8 word32 over[CHACHA_CHUNK_WORDS];
 #endif
 } ChaCha;
 
diff --git a/wolfssl/wolfcrypt/cryptocb.h b/wolfssl/wolfcrypt/cryptocb.h
index 1d86bf1169..b0aaad2f37 100644
--- a/wolfssl/wolfcrypt/cryptocb.h
+++ b/wolfssl/wolfcrypt/cryptocb.h
@@ -871,7 +871,11 @@ WOLFSSL_LOCAL int wc_CryptoCb_Sha384Hash(wc_Sha384* sha384, const byte* in,
 #endif
 #ifdef WOLFSSL_SHA512
 WOLFSSL_LOCAL int wc_CryptoCb_Sha512Hash(wc_Sha512* sha512, const byte* in,
-    word32 inSz, byte* digest, size_t digestSz);
+    word32 inSz, byte* digest
+#if !(defined(HAVE_FIPS) && FIPS_VERSION_LT(7,0))
+    , size_t digestSz
+#endif
+    );
 #endif
 
 #ifdef WOLFSSL_SHA3
diff --git a/wolfssl/wolfcrypt/error-crypt.h b/wolfssl/wolfcrypt/error-crypt.h
index 8d49ffc18e..736954a0a7 100644
--- a/wolfssl/wolfcrypt/error-crypt.h
+++ b/wolfssl/wolfcrypt/error-crypt.h
@@ -356,10 +356,13 @@ WOLFSSL_API void wc_ErrorString(int err, char* buff);
 WOLFSSL_ABI WOLFSSL_API const char* wc_GetErrorString(int error);
 #endif
 
+#ifdef WOLFSSL_DEBUG_BACKTRACE_ERROR_CODES
+    WOLFSSL_API extern int wc_backtrace_render(void);
+#endif
+
 #if defined(WOLFSSL_DEBUG_TRACE_ERROR_CODES) && \
         (defined(BUILDING_WOLFSSL) || \
          defined(WOLFSSL_DEBUG_TRACE_ERROR_CODES_ALWAYS))
-    WOLFSSL_API extern int wc_backtrace_render(void);
     #define WC_NO_ERR_TRACE(label) (CONST_NUM_ERR_ ## label)
     #ifndef WOLFSSL_DEBUG_BACKTRACE_RENDER_CLAUSE
         #ifdef WOLFSSL_DEBUG_BACKTRACE_ERROR_CODES
diff --git a/wolfssl/wolfcrypt/logging.h b/wolfssl/wolfcrypt/logging.h
index 414b41c233..efed186fdb 100644
--- a/wolfssl/wolfcrypt/logging.h
+++ b/wolfssl/wolfcrypt/logging.h
@@ -579,12 +579,13 @@ WOLFSSL_API void wolfSSL_SetLoggingPrefix(const char* prefix);
     #error "Failed: Cannot WOLFSSL_DEBUG_CERTS with WOLFSSL_DEBUG_ERRORS_ONLY"
 #endif
 
+#if defined(WOLFSSL_DEBUG_BACKTRACE_ERROR_CODES) && defined(XFILE)
+    WOLFSSL_API XFILE wc_backtrace_set_fp(XFILE new_fp);
+#endif
+
 #ifdef WOLFSSL_DEBUG_TRACE_ERROR_CODES
     WOLFSSL_API int wc_debug_trace_error_codes_enabled(void);
     WOLFSSL_API int wc_debug_trace_error_codes_set(int state);
-    #ifdef XFILE
-        WOLFSSL_API XFILE wc_backtrace_set_fp(XFILE new_fp);
-    #endif
 #endif
 
 #ifdef __cplusplus
diff --git a/wolfssl/wolfcrypt/memory.h b/wolfssl/wolfcrypt/memory.h
index 0651869060..cdc6a04045 100644
--- a/wolfssl/wolfcrypt/memory.h
+++ b/wolfssl/wolfcrypt/memory.h
@@ -370,6 +370,8 @@ WOLFSSL_LOCAL int wc_debug_CipherLifecycleFree(void **CipherLifecycleTag,
         #define DEBUG_VECTOR_REGISTERS_EXTRA_FAIL_CLAUSE abort();
     #elif defined(DEBUG_VECTOR_REGISTERS_EXIT_ON_FAIL)
         #define DEBUG_VECTOR_REGISTERS_EXTRA_FAIL_CLAUSE exit(1);
+    #elif defined(DEBUG_VECTOR_REGISTERS_BACKTRACE_ON_FAIL)
+        #define DEBUG_VECTOR_REGISTERS_EXTRA_FAIL_CLAUSE wc_backtrace_render();
     #elif !defined(DEBUG_VECTOR_REGISTERS_EXTRA_FAIL_CLAUSE)
         #define DEBUG_VECTOR_REGISTERS_EXTRA_FAIL_CLAUSE
     #endif
@@ -381,8 +383,9 @@ WOLFSSL_LOCAL int wc_debug_CipherLifecycleFree(void **CipherLifecycleTag,
           ++wc_svr_count;                                           \
           if (wc_svr_count > 5) {                                   \
               fprintf(stderr,                                       \
-                      ("%s @ L%d : incr : "                         \
-                       "wc_svr_count %d (last op %s L%d)\n"),       \
+                      ("%s() %s @ L %d : incr : "                   \
+                       "wc_svr_count %d (last op %s L %d)\n"),      \
+                        __FUNCTION__,                               \
                       __FILE__,                                     \
                       __LINE__,                                     \
                       wc_svr_count,                                 \
@@ -403,8 +406,9 @@ WOLFSSL_LOCAL int wc_debug_CipherLifecycleFree(void **CipherLifecycleTag,
 #define WC_DEBUG_SET_VECTOR_REGISTERS_RETVAL(x) do { \
             if (((x) != 0) && (wc_svr_count > 0)) {                 \
                 fprintf(stderr,                                     \
-                        ("%s @ L%d : incr : "                       \
-                         "wc_svr_count %d (last op %s L%d)\n"),     \
+                        ("%s() %s @ L %d : incr : "                 \
+                         "wc_svr_count %d (last op %s L %d)\n"),    \
+                        __FUNCTION__,                               \
                         __FILE__,                                   \
                         __LINE__,                                   \
                         wc_svr_count,                               \
@@ -422,8 +426,9 @@ WOLFSSL_LOCAL int wc_debug_CipherLifecycleFree(void **CipherLifecycleTag,
             ++wc_svr_count;                                         \
             if (wc_svr_count > 5) {                                 \
                 fprintf(stderr,                                     \
-                        ("%s @ L%d : incr : "                       \
-                         "wc_svr_count %d (last op %s L%d)\n"),     \
+                        ("%s() %s @ L %d : incr : "                 \
+                         "wc_svr_count %d (last op %s L %d)\n"),    \
+                        __FUNCTION__,                               \
                         __FILE__,                                   \
                         __LINE__,                                   \
                         wc_svr_count,                               \
@@ -445,8 +450,9 @@ WOLFSSL_LOCAL int wc_debug_CipherLifecycleFree(void **CipherLifecycleTag,
         if (wc_debug_vector_registers_retval != 0) {                \
             if (wc_svr_count > 0) {                                 \
                 fprintf(stderr,                                     \
-                        ("%s @ L%d : incr : "                       \
-                        "wc_svr_count %d (last op %s L%d)\n"),      \
+                        ("%s() %s @ L %d : incr : "                 \
+                        "wc_svr_count %d (last op %s L %d)\n"),     \
+                        __FUNCTION__,                               \
                         __FILE__,                                   \
                         __LINE__,                                   \
                         wc_svr_count,                               \
@@ -459,8 +465,9 @@ WOLFSSL_LOCAL int wc_debug_CipherLifecycleFree(void **CipherLifecycleTag,
             ++wc_svr_count;                                         \
             if (wc_svr_count > 5) {                                 \
                 fprintf(stderr,                                     \
-                        ("%s @ L%d : incr : "                       \
-                         "wc_svr_count %d (last op %s L%d)\n"),     \
+                        ("%s() %s @ L %d : incr : "                 \
+                         "wc_svr_count %d (last op %s L %d)\n"),    \
+                        __FUNCTION__,                               \
                         __FILE__,                                   \
                         __LINE__,                                   \
                         wc_svr_count,                               \
@@ -480,8 +487,9 @@ WOLFSSL_LOCAL int wc_debug_CipherLifecycleFree(void **CipherLifecycleTag,
     #define ASSERT_SAVED_VECTOR_REGISTERS() do {                    \
         if (wc_svr_count <= 0) {                                    \
             fprintf(stderr,                                         \
-                    ("ASSERT_SAVED_VECTOR_REGISTERS : %s @ L%d : "  \
-                    "wc_svr_count %d (last op %s L%d)\n"),          \
+                    ("ASSERT_SAVED_VECTOR_REGISTERS : %s() %s @ L %d : "  \
+                    "wc_svr_count %d (last op %s L %d)\n"),         \
+                        __FUNCTION__,                               \
                     __FILE__,                                       \
                     __LINE__,                                       \
                     wc_svr_count,                                   \
@@ -493,8 +501,9 @@ WOLFSSL_LOCAL int wc_debug_CipherLifecycleFree(void **CipherLifecycleTag,
     #define ASSERT_RESTORED_VECTOR_REGISTERS(fail_clause) do {      \
         if (wc_svr_count != 0) {                                    \
             fprintf(stderr,                                         \
-                    ("ASSERT_RESTORED_VECTOR_REGISTERS : %s @ L%d"  \
-                     " : wc_svr_count %d (last op %s L%d)\n"),      \
+                    ("ASSERT_RESTORED_VECTOR_REGISTERS : %s() %s @ L %d"  \
+                     " : wc_svr_count %d (last op %s L %d)\n"),     \
+                        __FUNCTION__,                               \
                     __FILE__,                                       \
                     __LINE__,                                       \
                     wc_svr_count,                                   \
@@ -508,8 +517,9 @@ WOLFSSL_LOCAL int wc_debug_CipherLifecycleFree(void **CipherLifecycleTag,
         --wc_svr_count;                                             \
         if ((wc_svr_count > 4) || (wc_svr_count < 0)) {             \
             fprintf(stderr,                                         \
-                    ("%s @ L%d : decr : "                           \
-                     "wc_svr_count %d (last op %s L%d)\n"),         \
+                    ("%s() %s @ L %d : decr : "                     \
+                     "wc_svr_count %d (last op %s L %d)\n"),        \
+                        __FUNCTION__,                               \
                     __FILE__,                                       \
                     __LINE__,                                       \
                     wc_svr_count,                                   \
diff --git a/wolfssl/wolfcrypt/port/riscv/riscv-64-asm.h b/wolfssl/wolfcrypt/port/riscv/riscv-64-asm.h
index 70cabff507..11090c65e5 100644
--- a/wolfssl/wolfcrypt/port/riscv/riscv-64-asm.h
+++ b/wolfssl/wolfcrypt/port/riscv/riscv-64-asm.h
@@ -181,6 +181,633 @@
 /* 32-bit width when loading. */
 #define WIDTH_32  0b110
 
+/*
+ * Scalar load/store helpers.
+ *
+ * Each macro performs the same operation as the ld/lwu/lw/lh/sd/sw/sh
+ * instruction it is named after. By default it expands to that native
+ * instruction. When built with WOLFSSL_RISCV_ASM_NO_UNALIGNED - for cores that
+ * don't support misaligned access - it checks the effective address alignment
+ * at run time and dispatches to the widest supported sequence: word-wise
+ * (lwu/sw) when 4-byte aligned, half-wise (lhu/sh) when 2-byte aligned,
+ * otherwise byte-wise (lbu/sb). The narrower _BY_BYTE / _BY_HALF / _BY_WORD
+ * forms are also exposed for sites that already know the alignment. Values
+ * are little-endian, matching the native instructions either way.
+ *
+ * Bulk variants UNALIGNED_<LD|SD|LWU|LW|SW><N> (N = 2, 4, 8) issue N
+ * consecutive accesses starting at o(p) - stride 8 bytes for LD/SD, 4 bytes
+ * for LWU/LW/SW. Under WOLFSSL_RISCV_ASM_NO_UNALIGNED they share a single
+ * alignment check across all N elements; otherwise they are just N native
+ * instructions back-to-back.
+ *
+ *   r = data register: destination (loads) or source (stores)
+ *   o = constant byte offset
+ *   p = base address register
+ *   t = scratch register - must differ from r and p; clobbered. For stores the
+ *       data register r is preserved. Only used when
+ *       WOLFSSL_RISCV_ASM_NO_UNALIGNED is defined; ignored otherwise.
+ */
+
+/* Apply X to N doublewords at 8-byte stride starting at o(p). */
+#define UNALIGNED_DW_REP2(X, r0, r1, o, p, t)                           \
+    X(r0, o,    p, t)                                                   \
+    X(r1, o+8,  p, t)
+#define UNALIGNED_DW_REP4(X, r0, r1, r2, r3, o, p, t)                   \
+    X(r0, o,    p, t)                                                   \
+    X(r1, o+8,  p, t)                                                   \
+    X(r2, o+16, p, t)                                                   \
+    X(r3, o+24, p, t)
+#define UNALIGNED_DW_REP8(X, r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)   \
+    X(r0, o,    p, t)                                                   \
+    X(r1, o+8,  p, t)                                                   \
+    X(r2, o+16, p, t)                                                   \
+    X(r3, o+24, p, t)                                                   \
+    X(r4, o+32, p, t)                                                   \
+    X(r5, o+40, p, t)                                                   \
+    X(r6, o+48, p, t)                                                   \
+    X(r7, o+56, p, t)
+
+/* Apply X to N words at 4-byte stride starting at o(p). */
+#define UNALIGNED_W_REP2(X, r0, r1, o, p, t)                            \
+    X(r0, o,    p, t)                                                   \
+    X(r1, o+4,  p, t)
+#define UNALIGNED_W_REP4(X, r0, r1, r2, r3, o, p, t)                    \
+    X(r0, o,    p, t)                                                   \
+    X(r1, o+4,  p, t)                                                   \
+    X(r2, o+8,  p, t)                                                   \
+    X(r3, o+12, p, t)
+#define UNALIGNED_W_REP8(X, r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)    \
+    X(r0, o,    p, t)                                                   \
+    X(r1, o+4,  p, t)                                                   \
+    X(r2, o+8,  p, t)                                                   \
+    X(r3, o+12, p, t)                                                   \
+    X(r4, o+16, p, t)                                                   \
+    X(r5, o+20, p, t)                                                   \
+    X(r6, o+24, p, t)                                                   \
+    X(r7, o+28, p, t)
+
+#ifndef WOLFSSL_RISCV_ASM_NO_UNALIGNED
+
+/* Load 64-bits. */
+#define UNALIGNED_LD(r, o, p, t)                \
+    "ld     " #r ", " #o "(" #p ")\n\t"
+
+/* Load 32-bits, zero extended. */
+#define UNALIGNED_LWU(r, o, p, t)               \
+    "lwu    " #r ", " #o "(" #p ")\n\t"
+
+/* Load 32-bits, sign extended. */
+#define UNALIGNED_LW(r, o, p, t)                \
+    "lw     " #r ", " #o "(" #p ")\n\t"
+
+/* Load 16-bits, sign extended. */
+#define UNALIGNED_LH(r, o, p, t)                \
+    "lh     " #r ", " #o "(" #p ")\n\t"
+
+/* Store 64-bits. */
+#define UNALIGNED_SD(r, o, p, t)                \
+    "sd     " #r ", " #o "(" #p ")\n\t"
+
+/* Store 32-bits. */
+#define UNALIGNED_SW(r, o, p, t)                \
+    "sw     " #r ", " #o "(" #p ")\n\t"
+
+/* Store 16-bits. */
+#define UNALIGNED_SH(r, o, p, t)                \
+    "sh     " #r ", " #o "(" #p ")\n\t"
+
+/* Bulk variants - hardware handles unaligned access, so just emit N native
+ * instructions. */
+#define UNALIGNED_LD2(r0, r1, o, p, t)                              \
+    UNALIGNED_DW_REP2(UNALIGNED_LD, r0, r1, o, p, t)
+#define UNALIGNED_LD4(r0, r1, r2, r3, o, p, t)                      \
+    UNALIGNED_DW_REP4(UNALIGNED_LD, r0, r1, r2, r3, o, p, t)
+#define UNALIGNED_LD8(r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)      \
+    UNALIGNED_DW_REP8(UNALIGNED_LD,                                 \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)
+#define UNALIGNED_SD2(r0, r1, o, p, t)                              \
+    UNALIGNED_DW_REP2(UNALIGNED_SD, r0, r1, o, p, t)
+#define UNALIGNED_SD4(r0, r1, r2, r3, o, p, t)                      \
+    UNALIGNED_DW_REP4(UNALIGNED_SD, r0, r1, r2, r3, o, p, t)
+#define UNALIGNED_SD8(r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)      \
+    UNALIGNED_DW_REP8(UNALIGNED_SD,                                 \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)
+#define UNALIGNED_LWU2(r0, r1, o, p, t)                             \
+    UNALIGNED_W_REP2(UNALIGNED_LWU, r0, r1, o, p, t)
+#define UNALIGNED_LWU4(r0, r1, r2, r3, o, p, t)                     \
+    UNALIGNED_W_REP4(UNALIGNED_LWU, r0, r1, r2, r3, o, p, t)
+#define UNALIGNED_LWU8(r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)     \
+    UNALIGNED_W_REP8(UNALIGNED_LWU,                                 \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)
+#define UNALIGNED_LW2(r0, r1, o, p, t)                              \
+    UNALIGNED_W_REP2(UNALIGNED_LW, r0, r1, o, p, t)
+#define UNALIGNED_LW4(r0, r1, r2, r3, o, p, t)                      \
+    UNALIGNED_W_REP4(UNALIGNED_LW, r0, r1, r2, r3, o, p, t)
+#define UNALIGNED_LW8(r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)      \
+    UNALIGNED_W_REP8(UNALIGNED_LW,                                  \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)
+#define UNALIGNED_SW2(r0, r1, o, p, t)                              \
+    UNALIGNED_W_REP2(UNALIGNED_SW, r0, r1, o, p, t)
+#define UNALIGNED_SW4(r0, r1, r2, r3, o, p, t)                      \
+    UNALIGNED_W_REP4(UNALIGNED_SW, r0, r1, r2, r3, o, p, t)
+#define UNALIGNED_SW8(r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)      \
+    UNALIGNED_W_REP8(UNALIGNED_SW,                                  \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)
+
+#else
+
+/* Load 64-bits. */
+#define UNALIGNED_LD_BY_BYTE(r, o, p, t)        \
+    "lbu    " #r ", " #o "+0(" #p ")\n\t"       \
+    "lbu    " #t ", " #o "+1(" #p ")\n\t"       \
+    "slli   " #t ", " #t ", 8\n\t"              \
+    "or     " #r ", " #r ", " #t "\n\t"         \
+    "lbu    " #t ", " #o "+2(" #p ")\n\t"       \
+    "slli   " #t ", " #t ", 16\n\t"             \
+    "or     " #r ", " #r ", " #t "\n\t"         \
+    "lbu    " #t ", " #o "+3(" #p ")\n\t"       \
+    "slli   " #t ", " #t ", 24\n\t"             \
+    "or     " #r ", " #r ", " #t "\n\t"         \
+    "lbu    " #t ", " #o "+4(" #p ")\n\t"       \
+    "slli   " #t ", " #t ", 32\n\t"             \
+    "or     " #r ", " #r ", " #t "\n\t"         \
+    "lbu    " #t ", " #o "+5(" #p ")\n\t"       \
+    "slli   " #t ", " #t ", 40\n\t"             \
+    "or     " #r ", " #r ", " #t "\n\t"         \
+    "lbu    " #t ", " #o "+6(" #p ")\n\t"       \
+    "slli   " #t ", " #t ", 48\n\t"             \
+    "or     " #r ", " #r ", " #t "\n\t"         \
+    "lbu    " #t ", " #o "+7(" #p ")\n\t"       \
+    "slli   " #t ", " #t ", 56\n\t"             \
+    "or     " #r ", " #r ", " #t "\n\t"
+#define UNALIGNED_LD_BY_HALF(r, o, p, t)        \
+    "lhu    " #r ", " #o "+0(" #p ")\n\t"       \
+    "lhu    " #t ", " #o "+2(" #p ")\n\t"       \
+    "slli   " #t ", " #t ", 16\n\t"             \
+    "or     " #r ", " #r ", " #t "\n\t"         \
+    "lhu    " #t ", " #o "+4(" #p ")\n\t"       \
+    "slli   " #t ", " #t ", 32\n\t"             \
+    "or     " #r ", " #r ", " #t "\n\t"         \
+    "lhu    " #t ", " #o "+6(" #p ")\n\t"       \
+    "slli   " #t ", " #t ", 48\n\t"             \
+    "or     " #r ", " #r ", " #t "\n\t"
+#define UNALIGNED_LD_BY_WORD(r, o, p, t)        \
+    "lwu    " #r ", " #o "+0(" #p ")\n\t"       \
+    "lwu    " #t ", " #o "+4(" #p ")\n\t"       \
+    "slli   " #t ", " #t ", 32\n\t"             \
+    "or     " #r ", " #r ", " #t "\n\t"
+#define UNALIGNED_LD_BY_DWORD(r, o, p, t)       \
+    "ld     " #r ", " #o "(" #p ")\n\t"
+/* Assumes o is a multiple of 8. */
+#define UNALIGNED_LD(r, o, p, t)                \
+    "andi   " #t ", " #p ", 7\n\t"              \
+    "bnez   " #t ", 1f\n\t"                     \
+    UNALIGNED_LD_BY_DWORD(r, o, p, t)           \
+    "j      4f\n\t"                             \
+    "1:\n\t"                                    \
+    "andi   " #t ", " #t ", 3\n\t"              \
+    "bnez   " #t ", 2f\n\t"                     \
+    UNALIGNED_LD_BY_WORD(r, o, p, t)            \
+    "j      4f\n\t"                             \
+    "2:\n\t"                                    \
+    "andi   " #t ", " #t ", 1\n\t"              \
+    "bnez   " #t ", 3f\n\t"                     \
+    UNALIGNED_LD_BY_HALF(r, o, p, t)            \
+    "j      4f\n\t"                             \
+    "3:\n\t"                                    \
+    UNALIGNED_LD_BY_BYTE(r, o, p, t)            \
+    "4:\n\t"
+
+/* Load 32-bits, zero extended. */
+#define UNALIGNED_LWU_BY_BYTE(r, o, p, t)       \
+    "lbu    " #r ", " #o "+0(" #p ")\n\t"       \
+    "lbu    " #t ", " #o "+1(" #p ")\n\t"       \
+    "slli   " #t ", " #t ", 8\n\t"              \
+    "or     " #r ", " #r ", " #t "\n\t"         \
+    "lbu    " #t ", " #o "+2(" #p ")\n\t"       \
+    "slli   " #t ", " #t ", 16\n\t"             \
+    "or     " #r ", " #r ", " #t "\n\t"         \
+    "lbu    " #t ", " #o "+3(" #p ")\n\t"       \
+    "slli   " #t ", " #t ", 24\n\t"             \
+    "or     " #r ", " #r ", " #t "\n\t"
+#define UNALIGNED_LWU_BY_HALF(r, o, p, t)       \
+    "lhu    " #r ", " #o "+0(" #p ")\n\t"       \
+    "lhu    " #t ", " #o "+2(" #p ")\n\t"       \
+    "slli   " #t ", " #t ", 16\n\t"             \
+    "or     " #r ", " #r ", " #t "\n\t"
+#define UNALIGNED_LWU_BY_WORD(r, o, p, t)       \
+    "lwu    " #r ", " #o "(" #p ")\n\t"
+/* Assumes o is a multiple of 4. */
+#define UNALIGNED_LWU(r, o, p, t)               \
+    "andi   " #t ", " #p ", 3\n\t"              \
+    "bnez   " #t ", 1f\n\t"                     \
+    UNALIGNED_LWU_BY_WORD(r, o, p, t)           \
+    "j      3f\n\t"                             \
+    "1:\n\t"                                    \
+    "andi   " #t ", " #t ", 1\n\t"              \
+    "bnez   " #t ", 2f\n\t"                     \
+    UNALIGNED_LWU_BY_HALF(r, o, p, t)           \
+    "j      3f\n\t"                             \
+    "2:\n\t"                                    \
+    UNALIGNED_LWU_BY_BYTE(r, o, p, t)           \
+    "3:\n\t"
+
+/* Load 32-bits, sign extended. */
+#define UNALIGNED_LW_BY_BYTE(r, o, p, t)        \
+    UNALIGNED_LWU_BY_BYTE(r, o, p, t)           \
+    "sext.w " #r ", " #r "\n\t"
+#define UNALIGNED_LW_BY_HALF(r, o, p, t)        \
+    UNALIGNED_LWU_BY_HALF(r, o, p, t)           \
+    "sext.w " #r ", " #r "\n\t"
+#define UNALIGNED_LW_BY_WORD(r, o, p, t)        \
+    "lw     " #r ", " #o "(" #p ")\n\t"
+/* Assumes o is a multiple of 4. */
+#define UNALIGNED_LW(r, o, p, t)                \
+    "andi   " #t ", " #p ", 3\n\t"              \
+    "bnez   " #t ", 1f\n\t"                     \
+    UNALIGNED_LW_BY_WORD(r, o, p, t)            \
+    "j      3f\n\t"                             \
+    "1:\n\t"                                    \
+    "andi   " #t ", " #t ", 1\n\t"              \
+    "bnez   " #t ", 2f\n\t"                     \
+    UNALIGNED_LW_BY_HALF(r, o, p, t)            \
+    "j      3f\n\t"                             \
+    "2:\n\t"                                    \
+    UNALIGNED_LW_BY_BYTE(r, o, p, t)            \
+    "3:\n\t"
+
+/* Load 16-bits, sign extended. */
+#define UNALIGNED_LH_BY_BYTE(r, o, p, t)        \
+    "lbu    " #r ", " #o "+0(" #p ")\n\t"       \
+    "lb     " #t ", " #o "+1(" #p ")\n\t"       \
+    "slli   " #t ", " #t ", 8\n\t"              \
+    "or     " #r ", " #r ", " #t "\n\t"
+#define UNALIGNED_LH(r, o, p, t)                \
+    UNALIGNED_LH_BY_BYTE(r, o, p, t)
+
+/* Store 64-bits. */
+#define UNALIGNED_SD_BY_BYTE(r, o, p, t)        \
+    "sb     " #r ", " #o "+0(" #p ")\n\t"       \
+    "srli   " #t ", " #r ", 8\n\t"              \
+    "sb     " #t ", " #o "+1(" #p ")\n\t"       \
+    "srli   " #t ", " #r ", 16\n\t"             \
+    "sb     " #t ", " #o "+2(" #p ")\n\t"       \
+    "srli   " #t ", " #r ", 24\n\t"             \
+    "sb     " #t ", " #o "+3(" #p ")\n\t"       \
+    "srli   " #t ", " #r ", 32\n\t"             \
+    "sb     " #t ", " #o "+4(" #p ")\n\t"       \
+    "srli   " #t ", " #r ", 40\n\t"             \
+    "sb     " #t ", " #o "+5(" #p ")\n\t"       \
+    "srli   " #t ", " #r ", 48\n\t"             \
+    "sb     " #t ", " #o "+6(" #p ")\n\t"       \
+    "srli   " #t ", " #r ", 56\n\t"             \
+    "sb     " #t ", " #o "+7(" #p ")\n\t"
+#define UNALIGNED_SD_BY_HALF(r, o, p, t)        \
+    "sh     " #r ", " #o "+0(" #p ")\n\t"       \
+    "srli   " #t ", " #r ", 16\n\t"             \
+    "sh     " #t ", " #o "+2(" #p ")\n\t"       \
+    "srli   " #t ", " #r ", 32\n\t"             \
+    "sh     " #t ", " #o "+4(" #p ")\n\t"       \
+    "srli   " #t ", " #r ", 48\n\t"             \
+    "sh     " #t ", " #o "+6(" #p ")\n\t"
+#define UNALIGNED_SD_BY_WORD(r, o, p, t)        \
+    "sw     " #r ", " #o "+0(" #p ")\n\t"       \
+    "srli   " #t ", " #r ", 32\n\t"             \
+    "sw     " #t ", " #o "+4(" #p ")\n\t"
+#define UNALIGNED_SD_BY_DWORD(r, o, p, t)       \
+    "sd     " #r ", " #o "(" #p ")\n\t"
+/* Assumes o is a multiple of 8. */
+#define UNALIGNED_SD(r, o, p, t)                \
+    "andi   " #t ", " #p ", 7\n\t"              \
+    "bnez   " #t ", 1f\n\t"                     \
+    UNALIGNED_SD_BY_DWORD(r, o, p, t)           \
+    "j      4f\n\t"                             \
+    "1:\n\t"                                    \
+    "andi   " #t ", " #t ", 3\n\t"              \
+    "bnez   " #t ", 2f\n\t"                     \
+    UNALIGNED_SD_BY_WORD(r, o, p, t)            \
+    "j      4f\n\t"                             \
+    "2:\n\t"                                    \
+    "andi   " #t ", " #t ", 1\n\t"              \
+    "bnez   " #t ", 3f\n\t"                     \
+    UNALIGNED_SD_BY_HALF(r, o, p, t)            \
+    "j      4f\n\t"                             \
+    "3:\n\t"                                    \
+    UNALIGNED_SD_BY_BYTE(r, o, p, t)            \
+    "4:\n\t"
+
+/* Store 32-bits. */
+#define UNALIGNED_SW_BY_BYTE(r, o, p, t)        \
+    "sb     " #r ", " #o "+0(" #p ")\n\t"       \
+    "srli   " #t ", " #r ", 8\n\t"              \
+    "sb     " #t ", " #o "+1(" #p ")\n\t"       \
+    "srli   " #t ", " #r ", 16\n\t"             \
+    "sb     " #t ", " #o "+2(" #p ")\n\t"       \
+    "srli   " #t ", " #r ", 24\n\t"             \
+    "sb     " #t ", " #o "+3(" #p ")\n\t"
+#define UNALIGNED_SW_BY_HALF(r, o, p, t)        \
+    "sh     " #r ", " #o "+0(" #p ")\n\t"       \
+    "srli   " #t ", " #r ", 16\n\t"             \
+    "sh     " #t ", " #o "+2(" #p ")\n\t"
+#define UNALIGNED_SW_BY_WORD(r, o, p, t)        \
+    "sw     " #r ", " #o "(" #p ")\n\t"
+/* Assumes o is a multiple of 4. */
+#define UNALIGNED_SW(r, o, p, t)                \
+    "andi   " #t ", " #p ", 3\n\t"              \
+    "bnez   " #t ", 1f\n\t"                     \
+    UNALIGNED_SW_BY_WORD(r, o, p, t)            \
+    "j      3f\n\t"                             \
+    "1:\n\t"                                    \
+    "andi   " #t ", " #t ", 1\n\t"              \
+    "bnez   " #t ", 2f\n\t"                     \
+    UNALIGNED_SW_BY_HALF(r, o, p, t)            \
+    "j      3f\n\t"                             \
+    "2:\n\t"                                    \
+    UNALIGNED_SW_BY_BYTE(r, o, p, t)            \
+    "3:\n\t"
+
+/* Store 16-bits. */
+#define UNALIGNED_SH_BY_BYTE(r, o, p, t)        \
+    "sb     " #r ", " #o "+0(" #p ")\n\t"       \
+    "srli   " #t ", " #r ", 8\n\t"              \
+    "sb     " #t ", " #o "+1(" #p ")\n\t"
+#define UNALIGNED_SH(r, o, p, t)                \
+    UNALIGNED_SH_BY_BYTE(r, o, p, t)
+
+/* Load 2 64-bits. Assumes o is a multiple of 8. */
+#define UNALIGNED_LD2(r0, r1, o, p, t)                                  \
+    "andi   " #t ", " #p ", 7\n\t"                                      \
+    "bnez   " #t ", 1f\n\t"                                             \
+    UNALIGNED_DW_REP2(UNALIGNED_LD_BY_DWORD, r0, r1, o, p, t)           \
+    "j      4f\n\t"                                                     \
+    "1:\n\t"                                                            \
+    "andi   " #t ", " #t ", 3\n\t"                                      \
+    "bnez   " #t ", 2f\n\t"                                             \
+    UNALIGNED_DW_REP2(UNALIGNED_LD_BY_WORD, r0, r1, o, p, t)            \
+    "j      4f\n\t"                                                     \
+    "2:\n\t"                                                            \
+    "andi   " #t ", " #t ", 1\n\t"                                      \
+    "bnez   " #t ", 3f\n\t"                                             \
+    UNALIGNED_DW_REP2(UNALIGNED_LD_BY_HALF, r0, r1, o, p, t)            \
+    "j      4f\n\t"                                                     \
+    "3:\n\t"                                                            \
+    UNALIGNED_DW_REP2(UNALIGNED_LD_BY_BYTE, r0, r1, o, p, t)            \
+    "4:\n\t"
+
+/* Load 4 64-bits. Assumes o is a multiple of 8. */
+#define UNALIGNED_LD4(r0, r1, r2, r3, o, p, t)                          \
+    "andi   " #t ", " #p ", 7\n\t"                                      \
+    "bnez   " #t ", 1f\n\t"                                             \
+    UNALIGNED_DW_REP4(UNALIGNED_LD_BY_DWORD, r0, r1, r2, r3, o, p, t)   \
+    "j      4f\n\t"                                                     \
+    "1:\n\t"                                                            \
+    "andi   " #t ", " #t ", 3\n\t"                                      \
+    "bnez   " #t ", 2f\n\t"                                             \
+    UNALIGNED_DW_REP4(UNALIGNED_LD_BY_WORD, r0, r1, r2, r3, o, p, t)    \
+    "j      4f\n\t"                                                     \
+    "2:\n\t"                                                            \
+    "andi   " #t ", " #t ", 1\n\t"                                      \
+    "bnez   " #t ", 3f\n\t"                                             \
+    UNALIGNED_DW_REP4(UNALIGNED_LD_BY_HALF, r0, r1, r2, r3, o, p, t)    \
+    "j      4f\n\t"                                                     \
+    "3:\n\t"                                                            \
+    UNALIGNED_DW_REP4(UNALIGNED_LD_BY_BYTE, r0, r1, r2, r3, o, p, t)    \
+    "4:\n\t"
+
+/* Load 8 64-bits. Assumes o is a multiple of 8. */
+#define UNALIGNED_LD8(r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)          \
+    "andi   " #t ", " #p ", 7\n\t"                                      \
+    "bnez   " #t ", 1f\n\t"                                             \
+    UNALIGNED_DW_REP8(UNALIGNED_LD_BY_DWORD,                            \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)                        \
+    "j      4f\n\t"                                                     \
+    "1:\n\t"                                                            \
+    "andi   " #t ", " #t ", 3\n\t"                                      \
+    "bnez   " #t ", 2f\n\t"                                             \
+    UNALIGNED_DW_REP8(UNALIGNED_LD_BY_WORD,                             \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)                        \
+    "j      4f\n\t"                                                     \
+    "2:\n\t"                                                            \
+    "andi   " #t ", " #t ", 1\n\t"                                      \
+    "bnez   " #t ", 3f\n\t"                                             \
+    UNALIGNED_DW_REP8(UNALIGNED_LD_BY_HALF,                             \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)                        \
+    "j      4f\n\t"                                                     \
+    "3:\n\t"                                                            \
+    UNALIGNED_DW_REP8(UNALIGNED_LD_BY_BYTE,                             \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)                        \
+    "4:\n\t"
+
+/* Store 2 64-bits. Assumes o is a multiple of 8. */
+#define UNALIGNED_SD2(r0, r1, o, p, t)                                  \
+    "andi   " #t ", " #p ", 7\n\t"                                      \
+    "bnez   " #t ", 1f\n\t"                                             \
+    UNALIGNED_DW_REP2(UNALIGNED_SD_BY_DWORD, r0, r1, o, p, t)           \
+    "j      4f\n\t"                                                     \
+    "1:\n\t"                                                            \
+    "andi   " #t ", " #t ", 3\n\t"                                      \
+    "bnez   " #t ", 2f\n\t"                                             \
+    UNALIGNED_DW_REP2(UNALIGNED_SD_BY_WORD, r0, r1, o, p, t)            \
+    "j      4f\n\t"                                                     \
+    "2:\n\t"                                                            \
+    "andi   " #t ", " #t ", 1\n\t"                                      \
+    "bnez   " #t ", 3f\n\t"                                             \
+    UNALIGNED_DW_REP2(UNALIGNED_SD_BY_HALF, r0, r1, o, p, t)            \
+    "j      4f\n\t"                                                     \
+    "3:\n\t"                                                            \
+    UNALIGNED_DW_REP2(UNALIGNED_SD_BY_BYTE, r0, r1, o, p, t)            \
+    "4:\n\t"
+
+/* Store 4 64-bits. Assumes o is a multiple of 8. */
+#define UNALIGNED_SD4(r0, r1, r2, r3, o, p, t)                          \
+    "andi   " #t ", " #p ", 7\n\t"                                      \
+    "bnez   " #t ", 1f\n\t"                                             \
+    UNALIGNED_DW_REP4(UNALIGNED_SD_BY_DWORD, r0, r1, r2, r3, o, p, t)   \
+    "j      4f\n\t"                                                     \
+    "1:\n\t"                                                            \
+    "andi   " #t ", " #t ", 3\n\t"                                      \
+    "bnez   " #t ", 2f\n\t"                                             \
+    UNALIGNED_DW_REP4(UNALIGNED_SD_BY_WORD, r0, r1, r2, r3, o, p, t)    \
+    "j      4f\n\t"                                                     \
+    "2:\n\t"                                                            \
+    "andi   " #t ", " #t ", 1\n\t"                                      \
+    "bnez   " #t ", 3f\n\t"                                             \
+    UNALIGNED_DW_REP4(UNALIGNED_SD_BY_HALF, r0, r1, r2, r3, o, p, t)    \
+    "j      4f\n\t"                                                     \
+    "3:\n\t"                                                            \
+    UNALIGNED_DW_REP4(UNALIGNED_SD_BY_BYTE, r0, r1, r2, r3, o, p, t)    \
+    "4:\n\t"
+
+/* Store 8 64-bits. Assumes o is a multiple of 8. */
+#define UNALIGNED_SD8(r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)          \
+    "andi   " #t ", " #p ", 7\n\t"                                      \
+    "bnez   " #t ", 1f\n\t"                                             \
+    UNALIGNED_DW_REP8(UNALIGNED_SD_BY_DWORD,                            \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)                        \
+    "j      4f\n\t"                                                     \
+    "1:\n\t"                                                            \
+    "andi   " #t ", " #t ", 3\n\t"                                      \
+    "bnez   " #t ", 2f\n\t"                                             \
+    UNALIGNED_DW_REP8(UNALIGNED_SD_BY_WORD,                             \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)                        \
+    "j      4f\n\t"                                                     \
+    "2:\n\t"                                                            \
+    "andi   " #t ", " #t ", 1\n\t"                                      \
+    "bnez   " #t ", 3f\n\t"                                             \
+    UNALIGNED_DW_REP8(UNALIGNED_SD_BY_HALF,                             \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)                        \
+    "j      4f\n\t"                                                     \
+    "3:\n\t"                                                            \
+    UNALIGNED_DW_REP8(UNALIGNED_SD_BY_BYTE,                             \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)                        \
+    "4:\n\t"
+
+/* Load 2 32-bits, zero extended. Assumes o is a multiple of 4. */
+#define UNALIGNED_LWU2(r0, r1, o, p, t)                                 \
+    "andi   " #t ", " #p ", 3\n\t"                                      \
+    "bnez   " #t ", 1f\n\t"                                             \
+    UNALIGNED_W_REP2(UNALIGNED_LWU_BY_WORD, r0, r1, o, p, t)            \
+    "j      3f\n\t"                                                     \
+    "1:\n\t"                                                            \
+    "andi   " #t ", " #t ", 1\n\t"                                      \
+    "bnez   " #t ", 2f\n\t"                                             \
+    UNALIGNED_W_REP2(UNALIGNED_LWU_BY_HALF, r0, r1, o, p, t)            \
+    "j      3f\n\t"                                                     \
+    "2:\n\t"                                                            \
+    UNALIGNED_W_REP2(UNALIGNED_LWU_BY_BYTE, r0, r1, o, p, t)            \
+    "3:\n\t"
+
+/* Load 4 32-bits, zero extended. Assumes o is a multiple of 4. */
+#define UNALIGNED_LWU4(r0, r1, r2, r3, o, p, t)                         \
+    "andi   " #t ", " #p ", 3\n\t"                                      \
+    "bnez   " #t ", 1f\n\t"                                             \
+    UNALIGNED_W_REP4(UNALIGNED_LWU_BY_WORD, r0, r1, r2, r3, o, p, t)    \
+    "j      3f\n\t"                                                     \
+    "1:\n\t"                                                            \
+    "andi   " #t ", " #t ", 1\n\t"                                      \
+    "bnez   " #t ", 2f\n\t"                                             \
+    UNALIGNED_W_REP4(UNALIGNED_LWU_BY_HALF, r0, r1, r2, r3, o, p, t)    \
+    "j      3f\n\t"                                                     \
+    "2:\n\t"                                                            \
+    UNALIGNED_W_REP4(UNALIGNED_LWU_BY_BYTE, r0, r1, r2, r3, o, p, t)    \
+    "3:\n\t"
+
+/* Load 8 32-bits, zero extended. Assumes o is a multiple of 4. */
+#define UNALIGNED_LWU8(r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)         \
+    "andi   " #t ", " #p ", 3\n\t"                                      \
+    "bnez   " #t ", 1f\n\t"                                             \
+    UNALIGNED_W_REP8(UNALIGNED_LWU_BY_WORD,                             \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)                        \
+    "j      3f\n\t"                                                     \
+    "1:\n\t"                                                            \
+    "andi   " #t ", " #t ", 1\n\t"                                      \
+    "bnez   " #t ", 2f\n\t"                                             \
+    UNALIGNED_W_REP8(UNALIGNED_LWU_BY_HALF,                             \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)                        \
+    "j      3f\n\t"                                                     \
+    "2:\n\t"                                                            \
+    UNALIGNED_W_REP8(UNALIGNED_LWU_BY_BYTE,                             \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)                        \
+    "3:\n\t"
+
+/* Load 2 32-bits, sign extended. Assumes o is a multiple of 4. */
+#define UNALIGNED_LW2(r0, r1, o, p, t)                                  \
+    "andi   " #t ", " #p ", 3\n\t"                                      \
+    "bnez   " #t ", 1f\n\t"                                             \
+    UNALIGNED_W_REP2(UNALIGNED_LW_BY_WORD, r0, r1, o, p, t)             \
+    "j      3f\n\t"                                                     \
+    "1:\n\t"                                                            \
+    "andi   " #t ", " #t ", 1\n\t"                                      \
+    "bnez   " #t ", 2f\n\t"                                             \
+    UNALIGNED_W_REP2(UNALIGNED_LW_BY_HALF, r0, r1, o, p, t)             \
+    "j      3f\n\t"                                                     \
+    "2:\n\t"                                                            \
+    UNALIGNED_W_REP2(UNALIGNED_LW_BY_BYTE, r0, r1, o, p, t)             \
+    "3:\n\t"
+
+/* Load 4 32-bits, sign extended. Assumes o is a multiple of 4. */
+#define UNALIGNED_LW4(r0, r1, r2, r3, o, p, t)                          \
+    "andi   " #t ", " #p ", 3\n\t"                                      \
+    "bnez   " #t ", 1f\n\t"                                             \
+    UNALIGNED_W_REP4(UNALIGNED_LW_BY_WORD, r0, r1, r2, r3, o, p, t)     \
+    "j      3f\n\t"                                                     \
+    "1:\n\t"                                                            \
+    "andi   " #t ", " #t ", 1\n\t"                                      \
+    "bnez   " #t ", 2f\n\t"                                             \
+    UNALIGNED_W_REP4(UNALIGNED_LW_BY_HALF, r0, r1, r2, r3, o, p, t)     \
+    "j      3f\n\t"                                                     \
+    "2:\n\t"                                                            \
+    UNALIGNED_W_REP4(UNALIGNED_LW_BY_BYTE, r0, r1, r2, r3, o, p, t)     \
+    "3:\n\t"
+
+/* Load 8 32-bits, sign extended. Assumes o is a multiple of 4. */
+#define UNALIGNED_LW8(r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)          \
+    "andi   " #t ", " #p ", 3\n\t"                                      \
+    "bnez   " #t ", 1f\n\t"                                             \
+    UNALIGNED_W_REP8(UNALIGNED_LW_BY_WORD,                              \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)                        \
+    "j      3f\n\t"                                                     \
+    "1:\n\t"                                                            \
+    "andi   " #t ", " #t ", 1\n\t"                                      \
+    "bnez   " #t ", 2f\n\t"                                             \
+    UNALIGNED_W_REP8(UNALIGNED_LW_BY_HALF,                              \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)                        \
+    "j      3f\n\t"                                                     \
+    "2:\n\t"                                                            \
+    UNALIGNED_W_REP8(UNALIGNED_LW_BY_BYTE,                              \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)                        \
+    "3:\n\t"
+
+/* Store 2 32-bits. Assumes o is a multiple of 4. */
+#define UNALIGNED_SW2(r0, r1, o, p, t)                                  \
+    "andi   " #t ", " #p ", 3\n\t"                                      \
+    "bnez   " #t ", 1f\n\t"                                             \
+    UNALIGNED_W_REP2(UNALIGNED_SW_BY_WORD, r0, r1, o, p, t)             \
+    "j      3f\n\t"                                                     \
+    "1:\n\t"                                                            \
+    "andi   " #t ", " #t ", 1\n\t"                                      \
+    "bnez   " #t ", 2f\n\t"                                             \
+    UNALIGNED_W_REP2(UNALIGNED_SW_BY_HALF, r0, r1, o, p, t)             \
+    "j      3f\n\t"                                                     \
+    "2:\n\t"                                                            \
+    UNALIGNED_W_REP2(UNALIGNED_SW_BY_BYTE, r0, r1, o, p, t)             \
+    "3:\n\t"
+
+/* Store 4 32-bits. Assumes o is a multiple of 4. */
+#define UNALIGNED_SW4(r0, r1, r2, r3, o, p, t)                          \
+    "andi   " #t ", " #p ", 3\n\t"                                      \
+    "bnez   " #t ", 1f\n\t"                                             \
+    UNALIGNED_W_REP4(UNALIGNED_SW_BY_WORD, r0, r1, r2, r3, o, p, t)     \
+    "j      3f\n\t"                                                     \
+    "1:\n\t"                                                            \
+    "andi   " #t ", " #t ", 1\n\t"                                      \
+    "bnez   " #t ", 2f\n\t"                                             \
+    UNALIGNED_W_REP4(UNALIGNED_SW_BY_HALF, r0, r1, r2, r3, o, p, t)     \
+    "j      3f\n\t"                                                     \
+    "2:\n\t"                                                            \
+    UNALIGNED_W_REP4(UNALIGNED_SW_BY_BYTE, r0, r1, r2, r3, o, p, t)     \
+    "3:\n\t"
+
+/* Store 8 32-bits. Assumes o is a multiple of 4. */
+#define UNALIGNED_SW8(r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)          \
+    "andi   " #t ", " #p ", 3\n\t"                                      \
+    "bnez   " #t ", 1f\n\t"                                             \
+    UNALIGNED_W_REP8(UNALIGNED_SW_BY_WORD,                              \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)                        \
+    "j      3f\n\t"                                                     \
+    "1:\n\t"                                                            \
+    "andi   " #t ", " #t ", 1\n\t"                                      \
+    "bnez   " #t ", 2f\n\t"                                             \
+    UNALIGNED_W_REP8(UNALIGNED_SW_BY_HALF,                              \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)                        \
+    "j      3f\n\t"                                                     \
+    "2:\n\t"                                                            \
+    UNALIGNED_W_REP8(UNALIGNED_SW_BY_BYTE,                              \
+        r0, r1, r2, r3, r4, r5, r6, r7, o, p, t)                        \
+    "3:\n\t"
+
+#endif /* !WOLFSSL_RISCV_ASM_NO_UNALIGNED */
+
 
 #define VLSEG_V(vd, rs1, cnt, width) \
     ASM_WORD(0b0000111 | (width << 12) | (0b10101000 << 20) |   \
diff --git a/wolfssl/wolfcrypt/settings.h b/wolfssl/wolfcrypt/settings.h
index c3ced88da7..e7d9b568f6 100644
--- a/wolfssl/wolfcrypt/settings.h
+++ b/wolfssl/wolfcrypt/settings.h
@@ -327,7 +327,7 @@
         #define WC_16BIT_CPU
         #define WOLFSSL_OLD_OID_SUM
     #elif defined(__SAM3X8E__)
-        #define WOLFSSL_NO_ATOMIC
+        #define WOLFSSL_NO_STDATOMIC_FENCE
         #define WOLFSSL_NO_SOCK
         #define WOLFSSL_USER_IO
         #define NO_WRITEV
@@ -3963,6 +3963,9 @@ extern void uITRON4_free(void *p) ;
         #ifndef WC_NO_INTERNAL_FUNCTION_POINTERS
             #define WC_NO_INTERNAL_FUNCTION_POINTERS
         #endif
+        #ifndef WC_NO_GLOBAL_OBJECT_POINTERS
+            #define WC_NO_GLOBAL_OBJECT_POINTERS
+        #endif
         #ifndef WOLFSSL_ECC_CURVE_STATIC
             #define WOLFSSL_ECC_CURVE_STATIC
         #endif
diff --git a/wolfssl/wolfcrypt/wc_port.h b/wolfssl/wolfcrypt/wc_port.h
index 042551bfc0..775794ae1f 100644
--- a/wolfssl/wolfcrypt/wc_port.h
+++ b/wolfssl/wolfcrypt/wc_port.h
@@ -1786,7 +1786,7 @@ WOLFSSL_ABI WOLFSSL_API int wolfCrypt_Cleanup(void);
         /* use user-supplied XFENCE definition. */
     #elif defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L) && \
           !defined(__STDC_NO_ATOMICS__) && !defined(NO_STDATOMIC_H)
-        #ifdef WOLFSSL_NO_ATOMIC
+        #ifdef WOLFSSL_NO_STDATOMIC_FENCE
             #define XFENCE() WC_DO_NOTHING
         #else
             #include <stdatomic.h>
diff --git a/wolfssl/wolfcrypt/wolfmath.h b/wolfssl/wolfcrypt/wolfmath.h
index f656329979..1353f40d17 100644
--- a/wolfssl/wolfcrypt/wolfmath.h
+++ b/wolfssl/wolfcrypt/wolfmath.h
@@ -74,12 +74,13 @@ This library provides big integer math functions.
 #endif
 
 /* timing resistance array */
-#if (!defined(WC_NO_CACHE_RESISTANT) && \
-     ((defined(HAVE_ECC) && defined(ECC_TIMING_RESISTANT)) || \
-      (defined(USE_FAST_MATH) && defined(TFM_TIMING_RESISTANT)))) || \
-    ((defined(WOLFSSL_SP_MATH_ALL) && !defined(WOLFSSL_RSA_VERIFY_ONLY) && \
-      !defined(WOLFSSL_RSA_PUBLIC_ONLY)) || !defined(NO_DH) || \
-     defined(OPENSSL_ALL) && defined(WC_PROTECT_ENCRYPTED_MEM))
+#if !defined(WC_NO_GLOBAL_OBJECT_POINTERS) && \
+    ((!defined(WC_NO_CACHE_RESISTANT) && \
+      ((defined(HAVE_ECC) && defined(ECC_TIMING_RESISTANT)) || \
+       (defined(USE_FAST_MATH) && defined(TFM_TIMING_RESISTANT)))) || \
+     ((defined(WOLFSSL_SP_MATH_ALL) && !defined(WOLFSSL_RSA_VERIFY_ONLY) && \
+       !defined(WOLFSSL_RSA_PUBLIC_ONLY)) || !defined(NO_DH) || \
+      defined(OPENSSL_ALL) && defined(WC_PROTECT_ENCRYPTED_MEM)))
 
     extern const wc_ptr_t wc_off_on_addr[2];
 #endif
diff --git a/wrapper/rust/wolfssl-wolfcrypt/Cargo.toml b/wrapper/rust/wolfssl-wolfcrypt/Cargo.toml
index 0e3ceed443..6a38933619 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/Cargo.toml
+++ b/wrapper/rust/wolfssl-wolfcrypt/Cargo.toml
@@ -11,7 +11,7 @@ categories = ["cryptography", "security", "api-bindings"]
 readme = "README.md"
 
 [features]
-std = []
+alloc = []
 rand_core = ["dep:rand_core"]
 aead = ["dep:aead"]
 cipher = ["dep:cipher"]
diff --git a/wrapper/rust/wolfssl-wolfcrypt/Makefile b/wrapper/rust/wolfssl-wolfcrypt/Makefile
index f705c4f8c0..0b7eac1d47 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/Makefile
+++ b/wrapper/rust/wolfssl-wolfcrypt/Makefile
@@ -1,4 +1,4 @@
-FEATURES := rand_core,aead,cipher,digest,mac,signature,password-hash,kem
+FEATURES := alloc,rand_core,aead,cipher,digest,mac,signature,password-hash,kem
 CARGO_FEATURE_FLAGS := --features $(FEATURES)
 
 .PHONY: all
diff --git a/wrapper/rust/wolfssl-wolfcrypt/README.md b/wrapper/rust/wolfssl-wolfcrypt/README.md
index d2509b508e..ec4fea9221 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/README.md
+++ b/wrapper/rust/wolfssl-wolfcrypt/README.md
@@ -3,6 +3,8 @@
 The `wolfssl-wolfcrypt` crate is a Rust wrapper for the wolfCrypt cryptographic
 algorithms portion of the wolfSSL C library.
 
+This crate requires wolfSSL version 5.9.0 or newer.
+
 ## Installation
 
 The `wolfssl` C library must be installed to be used by the Rust crate.
diff --git a/wrapper/rust/wolfssl-wolfcrypt/build.rs b/wrapper/rust/wolfssl-wolfcrypt/build.rs
index 6495d6bdb8..3d0c7ea6b1 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/build.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/build.rs
@@ -371,6 +371,9 @@ fn scan_cfg() -> Result<()> {
     check_cfg(&binding, "wc_Dh_ffdhe6144_Get", "dh_ffdhe_6144");
     check_cfg(&binding, "wc_Dh_ffdhe8192_Get", "dh_ffdhe_8192");
 
+    /* crypto callback */
+    check_cfg(&binding, "wc_CryptoCb_RegisterDevice", "wolf_crypto_cb");
+
     /* ecc */
     check_cfg(&binding, "wc_ecc_init", "ecc");
     check_cfg(&binding, "wc_ecc_export_point_der_compressed", "ecc_comp_key");
diff --git a/wrapper/rust/wolfssl-wolfcrypt/src/aes.rs b/wrapper/rust/wolfssl-wolfcrypt/src/aes.rs
index 9aede17418..7d954d2164 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/src/aes.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/src/aes.rs
@@ -33,13 +33,10 @@ use zeroize::{Zeroize, ZeroizeOnDrop};
 use aead::{AeadCore, AeadInPlace, KeyInit, KeySizeUser};
 
 #[cfg(feature = "aead")]
-use aead::generic_array::typenum::{U0, U12, U16, U32};
+use aead::generic_array::typenum::{U0, U12, U16, U24, U32};
 
 #[cfg(all(feature = "cipher", not(feature = "aead")))]
-use cipher::typenum::consts::{U16, U32};
-
-#[cfg(feature = "cipher")]
-use cipher::typenum::consts::U24;
+use cipher::typenum::consts::{U16, U24, U32};
 
 #[cfg(feature = "cipher")]
 use cipher::{
@@ -557,6 +554,58 @@ impl AeadInPlace for Aes128Ccm {
     }
 }
 
+/// AES-192-CCM authenticated encryption (12-byte nonce, 16-byte tag).
+#[cfg(all(aes_ccm, feature = "aead"))]
+#[derive(Zeroize, ZeroizeOnDrop)]
+pub struct Aes192Ccm {
+    key: [u8; 24],
+}
+
+#[cfg(all(aes_ccm, feature = "aead"))]
+impl KeySizeUser for Aes192Ccm {
+    type KeySize = U24;
+}
+
+#[cfg(all(aes_ccm, feature = "aead"))]
+impl AeadCore for Aes192Ccm {
+    type NonceSize = U12;
+    type TagSize = U16;
+    type CiphertextOverhead = U0;
+}
+
+#[cfg(all(aes_ccm, feature = "aead"))]
+impl KeyInit for Aes192Ccm {
+    fn new(key: &aead::Key<Self>) -> Self {
+        let mut k = [0u8; 24];
+        k.copy_from_slice(key.as_ref());
+        Aes192Ccm { key: k }
+    }
+}
+
+#[cfg(all(aes_ccm, feature = "aead"))]
+impl AeadInPlace for Aes192Ccm {
+    fn encrypt_in_place_detached(
+        &self,
+        nonce: &aead::Nonce<Self>,
+        associated_data: &[u8],
+        buffer: &mut [u8],
+    ) -> Result<aead::Tag<Self>, aead::Error> {
+        let mut tag = aead::Tag::<Self>::default();
+        ccm_encrypt_in_place(&self.key, nonce.as_ref(), associated_data, buffer, tag.as_mut())?;
+        Ok(tag)
+    }
+
+    fn decrypt_in_place_detached(
+        &self,
+        nonce: &aead::Nonce<Self>,
+        associated_data: &[u8],
+        buffer: &mut [u8],
+        tag: &aead::Tag<Self>,
+    ) -> Result<(), aead::Error> {
+        ccm_decrypt_in_place(&self.key, nonce.as_ref(), associated_data, buffer, tag.as_ref())
+    }
+}
+
 /// AES-256-CCM authenticated encryption (12-byte nonce, 16-byte tag).
 #[cfg(all(aes_ccm, feature = "aead"))]
 #[derive(Zeroize, ZeroizeOnDrop)]
@@ -1713,6 +1762,58 @@ impl AeadInPlace for Aes128Gcm {
     }
 }
 
+/// AES-192-GCM authenticated encryption (12-byte nonce, 16-byte tag).
+#[cfg(all(aes_gcm, feature = "aead"))]
+#[derive(Zeroize, ZeroizeOnDrop)]
+pub struct Aes192Gcm {
+    key: [u8; 24],
+}
+
+#[cfg(all(aes_gcm, feature = "aead"))]
+impl KeySizeUser for Aes192Gcm {
+    type KeySize = U24;
+}
+
+#[cfg(all(aes_gcm, feature = "aead"))]
+impl AeadCore for Aes192Gcm {
+    type NonceSize = U12;
+    type TagSize = U16;
+    type CiphertextOverhead = U0;
+}
+
+#[cfg(all(aes_gcm, feature = "aead"))]
+impl KeyInit for Aes192Gcm {
+    fn new(key: &aead::Key<Self>) -> Self {
+        let mut k = [0u8; 24];
+        k.copy_from_slice(key.as_ref());
+        Aes192Gcm { key: k }
+    }
+}
+
+#[cfg(all(aes_gcm, feature = "aead"))]
+impl AeadInPlace for Aes192Gcm {
+    fn encrypt_in_place_detached(
+        &self,
+        nonce: &aead::Nonce<Self>,
+        associated_data: &[u8],
+        buffer: &mut [u8],
+    ) -> Result<aead::Tag<Self>, aead::Error> {
+        let mut tag = aead::Tag::<Self>::default();
+        gcm_encrypt_in_place(&self.key, nonce.as_ref(), associated_data, buffer, tag.as_mut())?;
+        Ok(tag)
+    }
+
+    fn decrypt_in_place_detached(
+        &self,
+        nonce: &aead::Nonce<Self>,
+        associated_data: &[u8],
+        buffer: &mut [u8],
+        tag: &aead::Tag<Self>,
+    ) -> Result<(), aead::Error> {
+        gcm_decrypt_in_place(&self.key, nonce.as_ref(), associated_data, buffer, tag.as_ref())
+    }
+}
+
 /// AES-256-GCM authenticated encryption (12-byte nonce, 16-byte tag).
 #[cfg(all(aes_gcm, feature = "aead"))]
 #[derive(Zeroize, ZeroizeOnDrop)]
diff --git a/wrapper/rust/wolfssl-wolfcrypt/src/blake2_digest.rs b/wrapper/rust/wolfssl-wolfcrypt/src/blake2_digest.rs
new file mode 100644
index 0000000000..133b66d97f
--- /dev/null
+++ b/wrapper/rust/wolfssl-wolfcrypt/src/blake2_digest.rs
@@ -0,0 +1,169 @@
+/*
+ * Copyright (C) 2006-2026 wolfSSL Inc.
+ *
+ * This file is part of wolfSSL.
+ *
+ * wolfSSL is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * wolfSSL is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1335, USA
+ */
+
+/*!
+RustCrypto `digest` trait implementations for the wolfCrypt BLAKE2 hash
+types.
+
+Because BLAKE2b and BLAKE2s have variable digest sizes, this module exposes
+typed wrappers that pin the digest size of an underlying [`crate::blake2`]
+instance to a specific value, matching the typed aliases provided by the
+RustCrypto `blake2` crate (`Blake2b512`, `Blake2b256`, `Blake2s256`, etc.).
+
+Each typed wrapper implements the traits from the `digest` crate
+(`HashMarker`, `OutputSizeUser`, `BlockSizeUser`, `Update`, `Reset`,
+`FixedOutput`, and `FixedOutputReset`). With these implementations the
+`digest::Digest` trait becomes available via its blanket implementation,
+allowing these hashers to be used anywhere a RustCrypto `Digest` is
+accepted.
+
+Any failure returned by the underlying wolfCrypt call in a trait method
+will result in a panic, matching the infallible signatures required by the
+RustCrypto traits.
+*/
+
+use digest::consts::{U16, U24, U32, U48, U64, U128};
+
+macro_rules! impl_blake2_digest {
+    (
+        $(#[$attr:meta])*
+        $name:ident,
+        wc_ty = $wc_ty:path,
+        digest_size = $digest_size:literal,
+        out = $out_size:ty,
+        block = $block_size:ty
+    ) => {
+        $(#[$attr])*
+        pub struct $name {
+            blake2: $wc_ty,
+        }
+
+        $(#[$attr])*
+        impl Default for $name {
+            fn default() -> Self {
+                Self {
+                    blake2: <$wc_ty>::new($digest_size)
+                        .expect("wolfCrypt BLAKE2 init failed"),
+                }
+            }
+        }
+
+        $(#[$attr])*
+        impl digest::HashMarker for $name {}
+
+        $(#[$attr])*
+        impl digest::OutputSizeUser for $name {
+            type OutputSize = $out_size;
+        }
+
+        $(#[$attr])*
+        impl digest::block_api::BlockSizeUser for $name {
+            type BlockSize = $block_size;
+        }
+
+        $(#[$attr])*
+        impl digest::Update for $name {
+            fn update(&mut self, data: &[u8]) {
+                <$wc_ty>::update(&mut self.blake2, data)
+                    .expect("wolfCrypt BLAKE2 update failed");
+            }
+        }
+
+        $(#[$attr])*
+        impl digest::Reset for $name {
+            fn reset(&mut self) {
+                self.blake2 = <$wc_ty>::new($digest_size)
+                    .expect("wolfCrypt BLAKE2 init failed");
+            }
+        }
+
+        $(#[$attr])*
+        impl digest::FixedOutput for $name {
+            fn finalize_into(mut self, out: &mut digest::Output<Self>) {
+                <$wc_ty>::finalize(&mut self.blake2, out.as_mut_slice())
+                    .expect("wolfCrypt BLAKE2 finalize failed");
+            }
+        }
+
+        $(#[$attr])*
+        impl digest::FixedOutputReset for $name {
+            fn finalize_into_reset(&mut self, out: &mut digest::Output<Self>) {
+                <$wc_ty>::finalize(&mut self.blake2, out.as_mut_slice())
+                    .expect("wolfCrypt BLAKE2 finalize failed");
+                self.blake2 = <$wc_ty>::new($digest_size)
+                    .expect("wolfCrypt BLAKE2 init failed");
+            }
+        }
+    };
+}
+
+impl_blake2_digest! {
+    #[cfg(blake2b)]
+    Blake2b256,
+    wc_ty = crate::blake2::BLAKE2b,
+    digest_size = 32,
+    out = U32,
+    block = U128
+}
+
+impl_blake2_digest! {
+    #[cfg(blake2b)]
+    Blake2b384,
+    wc_ty = crate::blake2::BLAKE2b,
+    digest_size = 48,
+    out = U48,
+    block = U128
+}
+
+impl_blake2_digest! {
+    #[cfg(blake2b)]
+    Blake2b512,
+    wc_ty = crate::blake2::BLAKE2b,
+    digest_size = 64,
+    out = U64,
+    block = U128
+}
+
+impl_blake2_digest! {
+    #[cfg(blake2s)]
+    Blake2s128,
+    wc_ty = crate::blake2::BLAKE2s,
+    digest_size = 16,
+    out = U16,
+    block = U64
+}
+
+impl_blake2_digest! {
+    #[cfg(blake2s)]
+    Blake2s192,
+    wc_ty = crate::blake2::BLAKE2s,
+    digest_size = 24,
+    out = U24,
+    block = U64
+}
+
+impl_blake2_digest! {
+    #[cfg(blake2s)]
+    Blake2s256,
+    wc_ty = crate::blake2::BLAKE2s,
+    digest_size = 32,
+    out = U32,
+    block = U64
+}
diff --git a/wrapper/rust/wolfssl-wolfcrypt/src/blake2_mac.rs b/wrapper/rust/wolfssl-wolfcrypt/src/blake2_mac.rs
new file mode 100644
index 0000000000..561cba9498
--- /dev/null
+++ b/wrapper/rust/wolfssl-wolfcrypt/src/blake2_mac.rs
@@ -0,0 +1,163 @@
+/*
+ * Copyright (C) 2006-2026 wolfSSL Inc.
+ *
+ * This file is part of wolfSSL.
+ *
+ * wolfSSL is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * wolfSSL is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1335, USA
+ */
+
+/*!
+RustCrypto `digest::Mac` trait implementations for the wolfCrypt keyed
+BLAKE2 types.
+
+These wrappers cover keyed BLAKE2 (the construction exposed by RustCrypto's
+`blake2` crate as `Blake2bMac` / `Blake2sMac`), not the wolfCrypt-specific
+HMAC-BLAKE2 construction. With these implementations the `digest::Mac`
+trait becomes available via its blanket implementation, allowing these
+types to be used anywhere a RustCrypto `Mac` is accepted.
+
+Each typed wrapper pins the digest size and the maximum key size to a
+specific algorithm parameterization. The `Blake2sMac128`, `Blake2sMac192`,
+and `Blake2sMac256` wrappers all accept keys of any length up to 32 bytes
+(the BLAKE2s maximum), and the `Blake2bMac512` wrapper accepts keys of any
+length up to 64 bytes (the BLAKE2b maximum), matching the variable-length
+key behavior of the RustCrypto `blake2` crate's `Blake2sMac` / `Blake2bMac`
+types.
+
+Any failure returned by the underlying wolfCrypt call in a trait method
+will result in a panic, matching the infallible signatures required by the
+RustCrypto traits.
+*/
+
+use digest::consts::{U16, U24, U32, U48, U64};
+
+macro_rules! impl_blake2_mac {
+    (
+        $(#[$attr:meta])*
+        $name:ident,
+        wc_ty = $wc_ty:path,
+        digest_size = $digest_size:literal,
+        out = $out_size:ty,
+        key = $key_size:ty
+    ) => {
+        $(#[$attr])*
+        pub struct $name {
+            blake2: $wc_ty,
+        }
+
+        $(#[$attr])*
+        impl digest::MacMarker for $name {}
+
+        $(#[$attr])*
+        impl digest::OutputSizeUser for $name {
+            type OutputSize = $out_size;
+        }
+
+        $(#[$attr])*
+        impl digest::common::KeySizeUser for $name {
+            type KeySize = $key_size;
+        }
+
+        $(#[$attr])*
+        impl digest::KeyInit for $name {
+            fn new(key: &digest::Key<Self>) -> Self {
+                Self {
+                    blake2: <$wc_ty>::new_with_key($digest_size, key.as_slice())
+                        .expect("wolfCrypt BLAKE2 init failed"),
+                }
+            }
+
+            fn new_from_slice(key: &[u8]) -> Result<Self, digest::InvalidLength> {
+                if key.len() > <Self as digest::common::KeySizeUser>::key_size() {
+                    return Err(digest::InvalidLength);
+                }
+                Ok(Self {
+                    blake2: <$wc_ty>::new_with_key($digest_size, key)
+                        .map_err(|_| digest::InvalidLength)?,
+                })
+            }
+        }
+
+        $(#[$attr])*
+        impl digest::Update for $name {
+            fn update(&mut self, data: &[u8]) {
+                <$wc_ty>::update(&mut self.blake2, data)
+                    .expect("wolfCrypt BLAKE2 update failed");
+            }
+        }
+
+        $(#[$attr])*
+        impl digest::FixedOutput for $name {
+            fn finalize_into(mut self, out: &mut digest::Output<Self>) {
+                <$wc_ty>::finalize(&mut self.blake2, out.as_mut_slice())
+                    .expect("wolfCrypt BLAKE2 finalize failed");
+            }
+        }
+    };
+}
+
+impl_blake2_mac! {
+    #[cfg(blake2b)]
+    Blake2bMac256,
+    wc_ty = crate::blake2::BLAKE2b,
+    digest_size = 32,
+    out = U32,
+    key = U64
+}
+
+impl_blake2_mac! {
+    #[cfg(blake2b)]
+    Blake2bMac384,
+    wc_ty = crate::blake2::BLAKE2b,
+    digest_size = 48,
+    out = U48,
+    key = U64
+}
+
+impl_blake2_mac! {
+    #[cfg(blake2b)]
+    Blake2bMac512,
+    wc_ty = crate::blake2::BLAKE2b,
+    digest_size = 64,
+    out = U64,
+    key = U64
+}
+
+impl_blake2_mac! {
+    #[cfg(blake2s)]
+    Blake2sMac128,
+    wc_ty = crate::blake2::BLAKE2s,
+    digest_size = 16,
+    out = U16,
+    key = U32
+}
+
+impl_blake2_mac! {
+    #[cfg(blake2s)]
+    Blake2sMac192,
+    wc_ty = crate::blake2::BLAKE2s,
+    digest_size = 24,
+    out = U24,
+    key = U32
+}
+
+impl_blake2_mac! {
+    #[cfg(blake2s)]
+    Blake2sMac256,
+    wc_ty = crate::blake2::BLAKE2s,
+    digest_size = 32,
+    out = U32,
+    key = U32
+}
diff --git a/wrapper/rust/wolfssl-wolfcrypt/src/curve25519.rs b/wrapper/rust/wolfssl-wolfcrypt/src/curve25519.rs
index ef3c3ca8bd..7be171ff2d 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/src/curve25519.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/src/curve25519.rs
@@ -26,12 +26,15 @@ functionality.
 #![cfg(curve25519)]
 
 #[cfg(random)]
-use crate::random::RNG;
+use crate::random::{RNG, RngHandle};
 use crate::sys;
 use core::mem::MaybeUninit;
 
 pub struct Curve25519Key {
     wc_key: sys::curve25519_key,
+    /// RNG bound via `set_rng`, kept alive while the C struct holds its pointer.
+    #[cfg(random)]
+    rng: Option<RngHandle>,
 }
 
 impl Curve25519Key {
@@ -73,7 +76,7 @@ impl Curve25519Key {
     /// Returns either Ok(curve25519key) on success or Err(e) containing the
     /// wolfSSL library error code value.
     #[cfg(random)]
-    pub fn generate(rng: &mut RNG) -> Result<Self, i32> {
+    pub fn generate(rng: &RNG) -> Result<Self, i32> {
         let mut wc_key: MaybeUninit<sys::curve25519_key> = MaybeUninit::uninit();
         let rc = unsafe {
             sys::wc_curve25519_init(wc_key.as_mut_ptr())
@@ -82,9 +85,13 @@ impl Curve25519Key {
             return Err(rc);
         }
         let wc_key = unsafe { wc_key.assume_init() };
-        let mut curve25519key = Curve25519Key { wc_key };
+        let mut curve25519key = Curve25519Key {
+            wc_key,
+            #[cfg(random)]
+            rng: None,
+        };
         let rc = unsafe {
-            sys::wc_curve25519_make_key(&mut rng.wc_rng, Self::KEYSIZE as i32,
+            sys::wc_curve25519_make_key(rng.wc_rng, Self::KEYSIZE as i32,
                 &mut curve25519key.wc_key)
         };
         if rc != 0 {
@@ -104,12 +111,12 @@ impl Curve25519Key {
     /// Returns either Ok(()) on success or Err(e) containing the wolfSSL
     /// library error code value.
     #[cfg(random)]
-    pub fn generate_priv(rng: &mut RNG, out: &mut [u8]) -> Result<(), i32> {
+    pub fn generate_priv(rng: &RNG, out: &mut [u8]) -> Result<(), i32> {
         if out.len() != Self::KEYSIZE {
             return Err(sys::wolfCrypt_ErrorCodes_BUFFER_E);
         }
         let rc = unsafe {
-            sys::wc_curve25519_make_priv(&mut rng.wc_rng, Self::KEYSIZE as i32, out.as_mut_ptr())
+            sys::wc_curve25519_make_priv(rng.wc_rng, Self::KEYSIZE as i32, out.as_mut_ptr())
         };
         if rc != 0 {
             return Err(rc);
@@ -137,7 +144,11 @@ impl Curve25519Key {
             return Err(rc);
         }
         let wc_key = unsafe { wc_key.assume_init() };
-        let mut curve25519key = Curve25519Key { wc_key };
+        let mut curve25519key = Curve25519Key {
+            wc_key,
+            #[cfg(random)]
+            rng: None,
+        };
         let rc = unsafe {
             sys::wc_curve25519_import_private(private.as_ptr(), private_size,
                 &mut curve25519key.wc_key)
@@ -169,7 +180,11 @@ impl Curve25519Key {
             return Err(rc);
         }
         let wc_key = unsafe { wc_key.assume_init() };
-        let mut curve25519key = Curve25519Key { wc_key };
+        let mut curve25519key = Curve25519Key {
+            wc_key,
+            #[cfg(random)]
+            rng: None,
+        };
         let endian = if big_endian {sys::EC25519_BIG_ENDIAN} else {sys::EC25519_LITTLE_ENDIAN};
         let rc = unsafe {
             sys::wc_curve25519_import_private_ex(private.as_ptr(),
@@ -203,7 +218,11 @@ impl Curve25519Key {
             return Err(rc);
         }
         let wc_key = unsafe { wc_key.assume_init() };
-        let mut curve25519key = Curve25519Key { wc_key };
+        let mut curve25519key = Curve25519Key {
+            wc_key,
+            #[cfg(random)]
+            rng: None,
+        };
         let rc = unsafe {
             sys::wc_curve25519_import_private_raw(private.as_ptr(),
                 private_size, public.as_ptr(), public_size,
@@ -238,7 +257,11 @@ impl Curve25519Key {
             return Err(rc);
         }
         let wc_key = unsafe { wc_key.assume_init() };
-        let mut curve25519key = Curve25519Key { wc_key };
+        let mut curve25519key = Curve25519Key {
+            wc_key,
+            #[cfg(random)]
+            rng: None,
+        };
         let endian = if big_endian {sys::EC25519_BIG_ENDIAN} else {sys::EC25519_LITTLE_ENDIAN};
         let rc = unsafe {
             sys::wc_curve25519_import_private_raw_ex(private.as_ptr(),
@@ -271,7 +294,11 @@ impl Curve25519Key {
             return Err(rc);
         }
         let wc_key = unsafe { wc_key.assume_init() };
-        let mut curve25519key = Curve25519Key { wc_key };
+        let mut curve25519key = Curve25519Key {
+            wc_key,
+            #[cfg(random)]
+            rng: None,
+        };
         let rc = unsafe {
             sys::wc_curve25519_import_public(public.as_ptr(), public_size,
                 &mut curve25519key.wc_key)
@@ -303,7 +330,11 @@ impl Curve25519Key {
             return Err(rc);
         }
         let wc_key = unsafe { wc_key.assume_init() };
-        let mut curve25519key = Curve25519Key { wc_key };
+        let mut curve25519key = Curve25519Key {
+            wc_key,
+            #[cfg(random)]
+            rng: None,
+        };
         let endian = if big_endian {sys::EC25519_BIG_ENDIAN} else {sys::EC25519_LITTLE_ENDIAN};
         let rc = unsafe {
             sys::wc_curve25519_import_public_ex(public.as_ptr(), public_size,
@@ -353,12 +384,12 @@ impl Curve25519Key {
     /// Returns either Ok(()) on success or Err(e) containing the wolfSSL
     /// library error code value.
     #[cfg(all(curve25519_blinding, random))]
-    pub fn make_pub_blind(private: &[u8], public: &mut [u8], rng: &mut RNG) -> Result<(), i32> {
+    pub fn make_pub_blind(private: &[u8], public: &mut [u8], rng: &RNG) -> Result<(), i32> {
         let private_size = crate::buffer_len_to_i32(private.len())?;
         let public_size = crate::buffer_len_to_i32(public.len())?;
         let rc = unsafe {
             sys::wc_curve25519_make_pub_blind(public_size, public.as_mut_ptr(),
-                private_size, private.as_ptr(), &mut rng.wc_rng)
+                private_size, private.as_ptr(), rng.wc_rng)
         };
         if rc != 0 {
             return Err(rc);
@@ -408,14 +439,14 @@ impl Curve25519Key {
     /// Returns either Ok(()) on success or Err(e) containing the wolfSSL
     /// library error code value.
     #[cfg(all(curve25519_blinding, random))]
-    pub fn make_pub_generic_blind(private: &[u8], public: &mut [u8], basepoint: &[u8], rng: &mut RNG) -> Result<(), i32> {
+    pub fn make_pub_generic_blind(private: &[u8], public: &mut [u8], basepoint: &[u8], rng: &RNG) -> Result<(), i32> {
         let private_size = crate::buffer_len_to_i32(private.len())?;
         let public_size = crate::buffer_len_to_i32(public.len())?;
         let basepoint_size = crate::buffer_len_to_i32(basepoint.len())?;
         let rc = unsafe {
             sys::wc_curve25519_generic_blind(public_size, public.as_mut_ptr(),
                 private_size, private.as_ptr(), basepoint_size, basepoint.as_ptr(),
-                &mut rng.wc_rng)
+                rng.wc_rng)
         };
         if rc != 0 {
             return Err(rc);
@@ -454,27 +485,57 @@ impl Curve25519Key {
     /// This is necessary when generating a shared secret if wolfSSL is built
     /// with the `WOLFSSL_CURVE25519_BLINDING` build option enabled.
     ///
+    /// The key takes ownership of the RNG, so the underlying `WC_RNG` is
+    /// guaranteed to outlive this key.
+    ///
     /// # Parameters
     ///
     /// * `rng`: The `RNG` struct instance to associate with this
-    ///   `Curve25519Key` instance. The `RNG` struct should not be moved in
-    ///   memory after calling this method.
+    ///   `Curve25519Key` instance.
     ///
     /// # Returns
     ///
     /// Returns Ok(()) on success or Err(e) containing the wolfSSL library
     /// error code value.
     #[cfg(all(curve25519_blinding, random))]
-    pub fn set_rng(&mut self, rng: &mut RNG) -> Result<(), i32> {
+    pub fn set_rng(&mut self, rng: RNG) -> Result<(), i32> {
+        let wc_rng = rng.wc_rng;
         let rc = unsafe {
-            sys::wc_curve25519_set_rng(&mut self.wc_key, &mut rng.wc_rng)
+            sys::wc_curve25519_set_rng(&mut self.wc_key, wc_rng)
         };
         if rc != 0 {
             return Err(rc);
         }
+        self.rng = Some(RngHandle::Owned(rng));
         Ok(())
     }
 
+    /// Bind a shared `RNG` to this key. Available when the `alloc` feature
+    /// is enabled.
+    #[cfg(all(curve25519_blinding, random, feature = "alloc"))]
+    pub fn set_shared_rng(&mut self, rng: alloc::rc::Rc<RNG>) -> Result<(), i32> {
+        let wc_rng = rng.wc_rng;
+        let rc = unsafe {
+            sys::wc_curve25519_set_rng(&mut self.wc_key, wc_rng)
+        };
+        if rc != 0 {
+            return Err(rc);
+        }
+        self.rng = Some(RngHandle::Shared(rng));
+        Ok(())
+    }
+
+    /// Borrow the RNG previously bound via `set_rng` or `set_shared_rng`.
+    #[cfg(random)]
+    pub fn rng(&self) -> Option<&RNG> {
+        match &self.rng {
+            Some(RngHandle::Owned(rng)) => Some(rng),
+            #[cfg(feature = "alloc")]
+            Some(RngHandle::Shared(rng)) => Some(rng),
+            None => None,
+        }
+    }
+
     /// Compute a shared secret key given a secret private key and a received
     /// public key. It stores the generated secret key in the buffer out and
     /// returns the generated key size. Supports big or little endian.
diff --git a/wrapper/rust/wolfssl-wolfcrypt/src/dh.rs b/wrapper/rust/wolfssl-wolfcrypt/src/dh.rs
index 36c4fca3fa..c6f2140f37 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/src/dh.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/src/dh.rs
@@ -194,7 +194,7 @@ impl DH {
     /// }
     /// ```
     #[cfg(all(dh_keygen, random))]
-    pub fn generate(rng: &mut RNG, modulus_size: i32) -> Result<Self, i32> {
+    pub fn generate(rng: &RNG, modulus_size: i32) -> Result<Self, i32> {
         Self::generate_ex(rng, modulus_size, None, None)
     }
 
@@ -225,7 +225,7 @@ impl DH {
     /// }
     /// ```
     #[cfg(all(dh_keygen, random))]
-    pub fn generate_ex(rng: &mut RNG, modulus_size: i32, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
+    pub fn generate_ex(rng: &RNG, modulus_size: i32, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
         let mut wc_dhkey: MaybeUninit<sys::DhKey> = MaybeUninit::uninit();
         let heap = match heap {
             Some(heap) => heap,
@@ -242,7 +242,7 @@ impl DH {
         let wc_dhkey = unsafe { wc_dhkey.assume_init() };
         let mut dh = DH { wc_dhkey };
         let rc = unsafe {
-            sys::wc_DhGenerateParams(&mut rng.wc_rng, modulus_size, &mut dh.wc_dhkey)
+            sys::wc_DhGenerateParams(rng.wc_rng, modulus_size, &mut dh.wc_dhkey)
         };
         if rc != 0 {
             return Err(rc);
@@ -921,7 +921,7 @@ impl DH {
     /// }
     /// ```
     #[cfg(random)]
-    pub fn new_from_pgq_with_check(p: &[u8], g: &[u8], q: &[u8], trusted: i32, rng: &mut RNG) -> Result<Self, i32> {
+    pub fn new_from_pgq_with_check(p: &[u8], g: &[u8], q: &[u8], trusted: i32, rng: &RNG) -> Result<Self, i32> {
         Self::new_from_pgq_with_check_ex(p, g, q, trusted, rng, None, None)
     }
 
@@ -1030,7 +1030,7 @@ impl DH {
     /// }
     /// ```
     #[cfg(random)]
-    pub fn new_from_pgq_with_check_ex(p: &[u8], g: &[u8], q: &[u8], trusted: i32, rng: &mut RNG, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
+    pub fn new_from_pgq_with_check_ex(p: &[u8], g: &[u8], q: &[u8], trusted: i32, rng: &RNG, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
         let p_size = crate::buffer_len_to_u32(p.len())?;
         let g_size = crate::buffer_len_to_u32(g.len())?;
         let q_size = crate::buffer_len_to_u32(q.len())?;
@@ -1050,7 +1050,7 @@ impl DH {
         let wc_dhkey = unsafe { wc_dhkey.assume_init() };
         let mut dh = DH { wc_dhkey };
         let rc = unsafe {
-            sys::wc_DhSetCheckKey(&mut dh.wc_dhkey, p.as_ptr(), p_size, g.as_ptr(), g_size, q.as_ptr(), q_size, trusted, &mut rng.wc_rng)
+            sys::wc_DhSetCheckKey(&mut dh.wc_dhkey, p.as_ptr(), p_size, g.as_ptr(), g_size, q.as_ptr(), q_size, trusted, rng.wc_rng)
         };
         if rc != 0 {
             return Err(rc);
@@ -1509,13 +1509,13 @@ impl DH {
     /// }
     /// ```
     #[cfg(random)]
-    pub fn generate_key_pair(&mut self, rng: &mut RNG,
+    pub fn generate_key_pair(&mut self, rng: &RNG,
             private: &mut [u8], private_size: &mut u32,
             public: &mut [u8], public_size: &mut u32) -> Result<(), i32> {
         *private_size = crate::buffer_len_to_u32(private.len())?;
         *public_size = crate::buffer_len_to_u32(public.len())?;
         let rc = unsafe {
-            sys::wc_DhGenerateKeyPair(&mut self.wc_dhkey, &mut rng.wc_rng,
+            sys::wc_DhGenerateKeyPair(&mut self.wc_dhkey, rng.wc_rng,
                 private.as_mut_ptr(), private_size,
                 public.as_mut_ptr(), public_size)
         };
diff --git a/wrapper/rust/wolfssl-wolfcrypt/src/dilithium.rs b/wrapper/rust/wolfssl-wolfcrypt/src/dilithium.rs
index 0005147859..6e1ac0564c 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/src/dilithium.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/src/dilithium.rs
@@ -159,7 +159,7 @@ impl Dilithium {
     /// }
     /// ```
     #[cfg(all(dilithium_make_key, random))]
-    pub fn generate(level: u8, rng: &mut RNG) -> Result<Self, i32> {
+    pub fn generate(level: u8, rng: &RNG) -> Result<Self, i32> {
         Self::generate_ex(level, rng, None, None)
     }
 
@@ -193,7 +193,7 @@ impl Dilithium {
     #[cfg(all(dilithium_make_key, random))]
     pub fn generate_ex(
         level: u8,
-        rng: &mut RNG,
+        rng: &RNG,
         heap: Option<*mut core::ffi::c_void>,
         dev_id: Option<i32>,
     ) -> Result<Self, i32> {
@@ -202,7 +202,7 @@ impl Dilithium {
         if rc != 0 {
             return Err(rc);
         }
-        let rc = unsafe { sys::wc_dilithium_make_key(&mut key.ws_key, &mut rng.wc_rng) };
+        let rc = unsafe { sys::wc_dilithium_make_key(&mut key.ws_key, rng.wc_rng) };
         if rc != 0 {
             return Err(rc);
         }
@@ -859,7 +859,7 @@ impl Dilithium {
         &mut self,
         msg: &[u8],
         sig: &mut [u8],
-        rng: &mut RNG,
+        rng: &RNG,
     ) -> Result<usize, i32> {
         let msg_len = crate::buffer_len_to_u32(msg.len())?;
         let mut sig_len = crate::buffer_len_to_u32(sig.len())?;
@@ -869,7 +869,7 @@ impl Dilithium {
                 msg.as_ptr(), msg_len,
                 sig.as_mut_ptr(), &mut sig_len,
                 &mut self.ws_key,
-                &mut rng.wc_rng,
+                rng.wc_rng,
             )
         };
         if rc != 0 {
@@ -917,7 +917,7 @@ impl Dilithium {
         ctx: &[u8],
         msg: &[u8],
         sig: &mut [u8],
-        rng: &mut RNG,
+        rng: &RNG,
     ) -> Result<usize, i32> {
         if ctx.len() > 255 {
             return Err(sys::wolfCrypt_ErrorCodes_BUFFER_E);
@@ -931,7 +931,7 @@ impl Dilithium {
                 msg.as_ptr(), msg_len,
                 sig.as_mut_ptr(), &mut sig_len,
                 &mut self.ws_key,
-                &mut rng.wc_rng,
+                rng.wc_rng,
             )
         };
         if rc != 0 {
@@ -966,7 +966,7 @@ impl Dilithium {
         hash_alg: i32,
         hash: &[u8],
         sig: &mut [u8],
-        rng: &mut RNG,
+        rng: &RNG,
     ) -> Result<usize, i32> {
         if ctx.len() > 255 {
             return Err(sys::wolfCrypt_ErrorCodes_BUFFER_E);
@@ -981,7 +981,7 @@ impl Dilithium {
                 hash.as_ptr(), hash_len,
                 sig.as_mut_ptr(), &mut sig_len,
                 &mut self.ws_key,
-                &mut rng.wc_rng,
+                rng.wc_rng,
             )
         };
         if rc != 0 {
diff --git a/wrapper/rust/wolfssl-wolfcrypt/src/ecc.rs b/wrapper/rust/wolfssl-wolfcrypt/src/ecc.rs
index d154654a96..d0e3fefc79 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/src/ecc.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/src/ecc.rs
@@ -30,8 +30,7 @@ wolfSSL `ecc_key` object. It ensures proper initialization and deallocation.
 
 use crate::sys;
 #[cfg(random)]
-use crate::random::RNG;
-use core::mem::{MaybeUninit};
+use crate::random::{RNG, RngHandle};
 
 /// Rust wrapper for wolfSSL `ecc_point` object.
 pub struct ECCPoint {
@@ -60,11 +59,11 @@ impl ECCPoint {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::{ECC,ECCPoint};
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
+    /// let rng = RNG::new().expect("Failed to create RNG");
     /// let curve_id = ECC::SECP256R1;
     /// let curve_size = ECC::get_curve_size_from_id(curve_id).expect("Error with get_curve_size_from_id()");
-    /// let mut ecc = ECC::generate_ex(curve_size, &mut rng, curve_id, None, None).expect("Error with generate()");
-    /// let ecc_point = ecc.make_pub_to_point(Some(&mut rng), None).expect("Error with make_pub_to_point()");
+    /// let mut ecc = ECC::generate_ex(curve_size, &rng, curve_id, None, None).expect("Error with generate()");
+    /// let ecc_point = ecc.make_pub_to_point(Some(&rng), None).expect("Error with make_pub_to_point()");
     /// let mut der = [0u8; 128];
     /// let size = ecc_point.export_der(&mut der, curve_id).expect("Error with export_der()");
     /// ECCPoint::import_der(&der[0..size], curve_id, None).expect("Error with import_der()");
@@ -118,11 +117,11 @@ impl ECCPoint {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::{ECC,ECCPoint};
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
+    /// let rng = RNG::new().expect("Failed to create RNG");
     /// let curve_id = ECC::SECP256R1;
     /// let curve_size = ECC::get_curve_size_from_id(curve_id).expect("Error with get_curve_size_from_id()");
-    /// let mut ecc = ECC::generate_ex(curve_size, &mut rng, curve_id, None, None).expect("Error with generate()");
-    /// let ecc_point = ecc.make_pub_to_point(Some(&mut rng), None).expect("Error with make_pub_to_point()");
+    /// let mut ecc = ECC::generate_ex(curve_size, &rng, curve_id, None, None).expect("Error with generate()");
+    /// let ecc_point = ecc.make_pub_to_point(Some(&rng), None).expect("Error with make_pub_to_point()");
     /// let mut der = [0u8; 128];
     /// let size = ecc_point.export_der(&mut der, curve_id).expect("Error with export_der()");
     /// ECCPoint::import_der_ex(&der[0..size], curve_id, 0, None).expect("Error with import_der_ex()");
@@ -173,11 +172,11 @@ impl ECCPoint {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::{ECC,ECCPoint};
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
+    /// let rng = RNG::new().expect("Failed to create RNG");
     /// let curve_id = ECC::SECP256R1;
     /// let curve_size = ECC::get_curve_size_from_id(curve_id).expect("Error with get_curve_size_from_id()");
-    /// let mut ecc = ECC::generate_ex(curve_size, &mut rng, curve_id, None, None).expect("Error with generate()");
-    /// let ecc_point = ecc.make_pub_to_point(Some(&mut rng), None).expect("Error with make_pub_to_point()");
+    /// let mut ecc = ECC::generate_ex(curve_size, &rng, curve_id, None, None).expect("Error with generate()");
+    /// let ecc_point = ecc.make_pub_to_point(Some(&rng), None).expect("Error with make_pub_to_point()");
     /// let mut der = [0u8; 128];
     /// let size = ecc_point.export_der(&mut der, curve_id).expect("Error with export_der()");
     /// assert!(size > 0 && size <= der.len());
@@ -220,11 +219,11 @@ impl ECCPoint {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::{ECC,ECCPoint};
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
+    /// let rng = RNG::new().expect("Failed to create RNG");
     /// let curve_id = ECC::SECP256R1;
     /// let curve_size = ECC::get_curve_size_from_id(curve_id).expect("Error with get_curve_size_from_id()");
-    /// let mut ecc = ECC::generate_ex(curve_size, &mut rng, curve_id, None, None).expect("Error with generate()");
-    /// let ecc_point = ecc.make_pub_to_point(Some(&mut rng), None).expect("Error with make_pub_to_point()");
+    /// let mut ecc = ECC::generate_ex(curve_size, &rng, curve_id, None, None).expect("Error with generate()");
+    /// let ecc_point = ecc.make_pub_to_point(Some(&rng), None).expect("Error with make_pub_to_point()");
     /// let mut der = [0u8; 128];
     /// let size = ecc_point.export_der_compressed(&mut der, curve_id).expect("Error with export_der_compressed()");
     /// }
@@ -256,9 +255,9 @@ impl ECCPoint {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
-    /// let mut ecc = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
-    /// let mut ecc_point = ecc.make_pub_to_point(Some(&mut rng), None).expect("Error with make_pub_to_point()");
+    /// let rng = RNG::new().expect("Failed to create RNG");
+    /// let mut ecc = ECC::generate(32, &rng, None, None).expect("Error with generate()");
+    /// let mut ecc_point = ecc.make_pub_to_point(Some(&rng), None).expect("Error with make_pub_to_point()");
     /// ecc_point.forcezero();
     /// }
     /// ```
@@ -290,13 +289,18 @@ impl Drop for ECCPoint {
 
 /// The `ECC` struct manages the lifecycle of a wolfSSL `ecc_key` object.
 ///
-/// It ensures proper initialization and deallocation.
+/// It ensures proper initialization and deallocation. The struct holds a
+/// pointer to a `sys::ecc_key` allocated on the C heap.
 ///
 /// An instance can be created with `generate()`, `import_x963()`,
 /// `import_x963_ex()`, `import_private_key()`, `import_private_key_ex()`,
 /// `import_raw()`, or `import_raw_ex()`.
 pub struct ECC {
-    pub(crate) wc_ecc_key: sys::ecc_key,
+    pub(crate) wc_ecc_key: *mut sys::ecc_key,
+    /// RNG bound to this key via `set_rng`, kept alive for as long as the C
+    /// struct holds its pointer.
+    #[cfg(random)]
+    rng: Option<RngHandle>,
 }
 
 #[cfg(ecc_curve_ids)]
@@ -392,6 +396,25 @@ impl ECC {
     pub const FLAG_COFACTOR: i32 = sys::WC_ECC_FLAG_COFACTOR as i32;
     pub const FLAG_DEC_SIGN: i32 = sys::WC_ECC_FLAG_DEC_SIGN as i32;
 
+    /// Allocate and initialize a new `sys::ecc_key` on the C heap.
+    fn new_ecc_key(heap: *mut core::ffi::c_void, dev_id: i32) -> Result<*mut sys::ecc_key, i32> {
+        let key = unsafe { sys::wc_ecc_key_new(heap) };
+        if key.is_null() {
+            return Err(sys::wolfCrypt_ErrorCodes_MEMORY_E);
+        }
+        // wc_ecc_key_new() always initializes the key with INVALID_DEVID.
+        // Calling wc_ecc_init_ex() a second time to install the user's dev_id
+        // would re-run every initialization step (including allocations under
+        // some build configurations such as async crypto without WOLF_CRYPTO_CB)
+        // and orphan resources from the first init. Instead, just patch devId
+        // in place for WOLF_CRYPTO_CB builds.
+        #[cfg(wolf_crypto_cb)]
+        unsafe { (*key).devId = dev_id; }
+        #[cfg(not(wolf_crypto_cb))]
+        let _ = dev_id;
+        Ok(key)
+    }
+
     /// Generate a new ECC key with the given size.
     ///
     /// # Parameters
@@ -415,30 +438,23 @@ impl ECC {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
-    /// let mut ecc = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
+    /// let rng = RNG::new().expect("Failed to create RNG");
+    /// let mut ecc = ECC::generate(32, &rng, None, None).expect("Error with generate()");
     /// ecc.check().expect("Error with check()");
     /// }
     /// ```
     #[cfg(random)]
-    pub fn generate(size: i32, rng: &mut RNG, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
-        let mut wc_ecc_key: MaybeUninit<sys::ecc_key> = MaybeUninit::uninit();
-        let heap = match heap {
-            Some(heap) => heap,
-            None => core::ptr::null_mut(),
+    pub fn generate(size: i32, rng: &RNG, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
+        let heap = heap.unwrap_or(core::ptr::null_mut());
+        let dev_id = dev_id.unwrap_or(sys::INVALID_DEVID);
+        let wc_ecc_key = Self::new_ecc_key(heap, dev_id)?;
+        let ecc = ECC {
+            wc_ecc_key,
+            #[cfg(random)]
+            rng: None,
         };
-        let dev_id = match dev_id {
-            Some(dev_id) => dev_id,
-            None => sys::INVALID_DEVID,
-        };
-        let rc = unsafe { sys::wc_ecc_init_ex(wc_ecc_key.as_mut_ptr(), heap, dev_id) };
-        if rc != 0 {
-            return Err(rc);
-        }
-        let wc_ecc_key = unsafe { wc_ecc_key.assume_init() };
-        let mut ecc = ECC { wc_ecc_key };
         let rc = unsafe {
-            sys::wc_ecc_make_key(&mut rng.wc_rng, size, &mut ecc.wc_ecc_key)
+            sys::wc_ecc_make_key(rng.wc_rng, size, ecc.wc_ecc_key)
         };
         if rc != 0 {
             return Err(rc);
@@ -470,32 +486,25 @@ impl ECC {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
+    /// let rng = RNG::new().expect("Failed to create RNG");
     /// let curve_id = ECC::SECP256R1;
     /// let curve_size = ECC::get_curve_size_from_id(curve_id).expect("Error with get_curve_size_from_id()");
-    /// let mut ecc = ECC::generate_ex(curve_size, &mut rng, curve_id, None, None).expect("Error with generate_ex()");
+    /// let mut ecc = ECC::generate_ex(curve_size, &rng, curve_id, None, None).expect("Error with generate_ex()");
     /// ecc.check().expect("Error with check()");
     /// }
     /// ```
     #[cfg(random)]
-    pub fn generate_ex(size: i32, rng: &mut RNG, curve_id: i32, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
-        let mut wc_ecc_key: MaybeUninit<sys::ecc_key> = MaybeUninit::uninit();
-        let heap = match heap {
-            Some(heap) => heap,
-            None => core::ptr::null_mut(),
+    pub fn generate_ex(size: i32, rng: &RNG, curve_id: i32, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
+        let heap = heap.unwrap_or(core::ptr::null_mut());
+        let dev_id = dev_id.unwrap_or(sys::INVALID_DEVID);
+        let wc_ecc_key = Self::new_ecc_key(heap, dev_id)?;
+        let ecc = ECC {
+            wc_ecc_key,
+            #[cfg(random)]
+            rng: None,
         };
-        let dev_id = match dev_id {
-            Some(dev_id) => dev_id,
-            None => sys::INVALID_DEVID,
-        };
-        let rc = unsafe { sys::wc_ecc_init_ex(wc_ecc_key.as_mut_ptr(), heap, dev_id) };
-        if rc != 0 {
-            return Err(rc);
-        }
-        let wc_ecc_key = unsafe { wc_ecc_key.assume_init() };
-        let mut ecc = ECC { wc_ecc_key };
         let rc = unsafe {
-            sys::wc_ecc_make_key_ex(&mut rng.wc_rng, size, &mut ecc.wc_ecc_key, curve_id)
+            sys::wc_ecc_make_key_ex(rng.wc_rng, size, ecc.wc_ecc_key, curve_id)
         };
         if rc != 0 {
             return Err(rc);
@@ -528,32 +537,25 @@ impl ECC {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
+    /// let rng = RNG::new().expect("Failed to create RNG");
     /// let curve_id = ECC::SECP256R1;
     /// let curve_size = ECC::get_curve_size_from_id(curve_id).expect("Error with get_curve_size_from_id()");
-    /// let mut ecc = ECC::generate_ex2(curve_size, &mut rng, curve_id, ECC::FLAG_COFACTOR, None, None).expect("Error with generate_ex2()");
+    /// let mut ecc = ECC::generate_ex2(curve_size, &rng, curve_id, ECC::FLAG_COFACTOR, None, None).expect("Error with generate_ex2()");
     /// ecc.check().expect("Error with check()");
     /// }
     /// ```
     #[cfg(random)]
-    pub fn generate_ex2(size: i32, rng: &mut RNG, curve_id: i32, flags: i32, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
-        let mut wc_ecc_key: MaybeUninit<sys::ecc_key> = MaybeUninit::uninit();
-        let heap = match heap {
-            Some(heap) => heap,
-            None => core::ptr::null_mut(),
+    pub fn generate_ex2(size: i32, rng: &RNG, curve_id: i32, flags: i32, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
+        let heap = heap.unwrap_or(core::ptr::null_mut());
+        let dev_id = dev_id.unwrap_or(sys::INVALID_DEVID);
+        let wc_ecc_key = Self::new_ecc_key(heap, dev_id)?;
+        let ecc = ECC {
+            wc_ecc_key,
+            #[cfg(random)]
+            rng: None,
         };
-        let dev_id = match dev_id {
-            Some(dev_id) => dev_id,
-            None => sys::INVALID_DEVID,
-        };
-        let rc = unsafe { sys::wc_ecc_init_ex(wc_ecc_key.as_mut_ptr(), heap, dev_id) };
-        if rc != 0 {
-            return Err(rc);
-        }
-        let wc_ecc_key = unsafe { wc_ecc_key.assume_init() };
-        let mut ecc = ECC { wc_ecc_key };
         let rc = unsafe {
-            sys::wc_ecc_make_key_ex2(&mut rng.wc_rng, size, &mut ecc.wc_ecc_key, curve_id, flags)
+            sys::wc_ecc_make_key_ex2(rng.wc_rng, size, ecc.wc_ecc_key, curve_id, flags)
         };
         if rc != 0 {
             return Err(rc);
@@ -580,10 +582,10 @@ impl ECC {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
+    /// let rng = RNG::new().expect("Failed to create RNG");
     /// let curve_id = ECC::SECP256R1;
     /// let curve_size = ECC::get_curve_size_from_id(curve_id).expect("Error with get_curve_size_from_id()");
-    /// let mut ecc = ECC::generate_ex(curve_size, &mut rng, curve_id, None, None).expect("Error with generate()");
+    /// let mut ecc = ECC::generate_ex(curve_size, &rng, curve_id, None, None).expect("Error with generate()");
     /// ecc.check().expect("Error with check()");
     /// }
     /// ```
@@ -617,32 +619,25 @@ impl ECC {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
     /// use std::fs;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
+    /// let rng = RNG::new().expect("Failed to create RNG");
     /// let key_path = "../../../certs/ecc-client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut ecc = ECC::import_der(&der, None, None).expect("Error with import_der()");
     /// }
     /// ```
     pub fn import_der(der: &[u8], heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
-        let mut wc_ecc_key: MaybeUninit<sys::ecc_key> = MaybeUninit::uninit();
-        let heap = match heap {
-            Some(heap) => heap,
-            None => core::ptr::null_mut(),
+        let heap = heap.unwrap_or(core::ptr::null_mut());
+        let dev_id = dev_id.unwrap_or(sys::INVALID_DEVID);
+        let wc_ecc_key = Self::new_ecc_key(heap, dev_id)?;
+        let ecc = ECC {
+            wc_ecc_key,
+            #[cfg(random)]
+            rng: None,
         };
-        let dev_id = match dev_id {
-            Some(dev_id) => dev_id,
-            None => sys::INVALID_DEVID,
-        };
-        let rc = unsafe { sys::wc_ecc_init_ex(wc_ecc_key.as_mut_ptr(), heap, dev_id) };
-        if rc != 0 {
-            return Err(rc);
-        }
-        let wc_ecc_key = unsafe { wc_ecc_key.assume_init() };
-        let mut ecc = ECC { wc_ecc_key };
         let mut idx = 0u32;
         let der_size = crate::buffer_len_to_u32(der.len())?;
         let rc = unsafe {
-            sys::wc_EccPrivateKeyDecode(der.as_ptr(), &mut idx, &mut ecc.wc_ecc_key, der_size)
+            sys::wc_EccPrivateKeyDecode(der.as_ptr(), &mut idx, ecc.wc_ecc_key, der_size)
         };
         if rc != 0 {
             return Err(rc);
@@ -672,13 +667,13 @@ impl ECC {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
     /// use std::fs;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
+    /// let rng = RNG::new().expect("Failed to create RNG");
     /// let key_path = "../../../certs/ecc-client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut ecc = ECC::import_der(&der, None, None).expect("Error with import_der()");
     /// let hash = [0x42u8; 32];
     /// let mut signature = [0u8; 128];
-    /// let signature_length = ecc.sign_hash(&hash, &mut signature, &mut rng).expect("Error with sign_hash()");
+    /// let signature_length = ecc.sign_hash(&hash, &mut signature, &rng).expect("Error with sign_hash()");
     /// assert!(signature_length > 0 && signature_length <= signature.len());
     /// let signature = &mut signature[0..signature_length];
     /// let key_path = "../../../certs/ecc-client-keyPub.der";
@@ -687,25 +682,18 @@ impl ECC {
     /// }
     /// ```
     pub fn import_public_der(der: &[u8], heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
-        let mut wc_ecc_key: MaybeUninit<sys::ecc_key> = MaybeUninit::uninit();
-        let heap = match heap {
-            Some(heap) => heap,
-            None => core::ptr::null_mut(),
+        let heap = heap.unwrap_or(core::ptr::null_mut());
+        let dev_id = dev_id.unwrap_or(sys::INVALID_DEVID);
+        let wc_ecc_key = Self::new_ecc_key(heap, dev_id)?;
+        let ecc = ECC {
+            wc_ecc_key,
+            #[cfg(random)]
+            rng: None,
         };
-        let dev_id = match dev_id {
-            Some(dev_id) => dev_id,
-            None => sys::INVALID_DEVID,
-        };
-        let rc = unsafe { sys::wc_ecc_init_ex(wc_ecc_key.as_mut_ptr(), heap, dev_id) };
-        if rc != 0 {
-            return Err(rc);
-        }
-        let wc_ecc_key = unsafe { wc_ecc_key.assume_init() };
-        let mut ecc = ECC { wc_ecc_key };
         let mut idx = 0u32;
         let der_size = crate::buffer_len_to_u32(der.len())?;
         let rc = unsafe {
-            sys::wc_EccPublicKeyDecode(der.as_ptr(), &mut idx, &mut ecc.wc_ecc_key, der_size)
+            sys::wc_EccPublicKeyDecode(der.as_ptr(), &mut idx, ecc.wc_ecc_key, der_size)
         };
         if rc != 0 {
             return Err(rc);
@@ -738,11 +726,11 @@ impl ECC {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
-    /// let mut ecc = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
+    /// let rng = RNG::new().expect("Failed to create RNG");
+    /// let mut ecc = ECC::generate(32, &rng, None, None).expect("Error with generate()");
     /// let hash = [0x42u8; 32];
     /// let mut signature = [0u8; 128];
-    /// let signature_length = ecc.sign_hash(&hash, &mut signature, &mut rng).expect("Error with sign_hash()");
+    /// let signature_length = ecc.sign_hash(&hash, &mut signature, &rng).expect("Error with sign_hash()");
     /// let signature = &signature[0..signature_length];
     /// let mut d = [0u8; 32];
     /// let d_size = ecc.export_private(&mut d).expect("Error with export_private()");
@@ -756,27 +744,20 @@ impl ECC {
     /// ```
     #[cfg(ecc_import)]
     pub fn import_private_key(priv_buf: &[u8], pub_buf: &[u8], heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
-        let mut wc_ecc_key: MaybeUninit<sys::ecc_key> = MaybeUninit::uninit();
-        let heap = match heap {
-            Some(heap) => heap,
-            None => core::ptr::null_mut(),
+        let heap = heap.unwrap_or(core::ptr::null_mut());
+        let dev_id = dev_id.unwrap_or(sys::INVALID_DEVID);
+        let wc_ecc_key = Self::new_ecc_key(heap, dev_id)?;
+        let ecc = ECC {
+            wc_ecc_key,
+            #[cfg(random)]
+            rng: None,
         };
-        let dev_id = match dev_id {
-            Some(dev_id) => dev_id,
-            None => sys::INVALID_DEVID,
-        };
-        let rc = unsafe { sys::wc_ecc_init_ex(wc_ecc_key.as_mut_ptr(), heap, dev_id) };
-        if rc != 0 {
-            return Err(rc);
-        }
-        let wc_ecc_key = unsafe { wc_ecc_key.assume_init() };
-        let mut ecc = ECC { wc_ecc_key };
         let priv_size = crate::buffer_len_to_u32(priv_buf.len())?;
         let pub_ptr = if pub_buf.is_empty() {core::ptr::null()} else {pub_buf.as_ptr()};
         let pub_size = crate::buffer_len_to_u32(pub_buf.len())?;
         let rc = unsafe {
             sys::wc_ecc_import_private_key(priv_buf.as_ptr(), priv_size,
-                pub_ptr, pub_size, &mut ecc.wc_ecc_key)
+                pub_ptr, pub_size, ecc.wc_ecc_key)
         };
         if rc != 0 {
             return Err(rc);
@@ -810,13 +791,13 @@ impl ECC {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
+    /// let rng = RNG::new().expect("Failed to create RNG");
     /// let curve_id = ECC::SECP256R1;
     /// let curve_size = ECC::get_curve_size_from_id(curve_id).expect("Error with get_curve_size_from_id()");
-    /// let mut ecc = ECC::generate_ex(curve_size, &mut rng, curve_id, None, None).expect("Error with generate_ex()");
+    /// let mut ecc = ECC::generate_ex(curve_size, &rng, curve_id, None, None).expect("Error with generate_ex()");
     /// let hash = [0x42u8; 32];
     /// let mut signature = [0u8; 128];
-    /// let signature_length = ecc.sign_hash(&hash, &mut signature, &mut rng).expect("Error with sign_hash()");
+    /// let signature_length = ecc.sign_hash(&hash, &mut signature, &rng).expect("Error with sign_hash()");
     /// let signature = &signature[0..signature_length];
     /// let mut d = [0u8; 32];
     /// let d_size = ecc.export_private(&mut d).expect("Error with export_private()");
@@ -830,27 +811,20 @@ impl ECC {
     /// ```
     #[cfg(ecc_import)]
     pub fn import_private_key_ex(priv_buf: &[u8], pub_buf: &[u8], curve_id: i32, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
-        let mut wc_ecc_key: MaybeUninit<sys::ecc_key> = MaybeUninit::uninit();
-        let heap = match heap {
-            Some(heap) => heap,
-            None => core::ptr::null_mut(),
+        let heap = heap.unwrap_or(core::ptr::null_mut());
+        let dev_id = dev_id.unwrap_or(sys::INVALID_DEVID);
+        let wc_ecc_key = Self::new_ecc_key(heap, dev_id)?;
+        let ecc = ECC {
+            wc_ecc_key,
+            #[cfg(random)]
+            rng: None,
         };
-        let dev_id = match dev_id {
-            Some(dev_id) => dev_id,
-            None => sys::INVALID_DEVID,
-        };
-        let rc = unsafe { sys::wc_ecc_init_ex(wc_ecc_key.as_mut_ptr(), heap, dev_id) };
-        if rc != 0 {
-            return Err(rc);
-        }
-        let wc_ecc_key = unsafe { wc_ecc_key.assume_init() };
-        let mut ecc = ECC { wc_ecc_key };
         let priv_size = crate::buffer_len_to_u32(priv_buf.len())?;
         let pub_ptr = if pub_buf.is_empty() {core::ptr::null()} else {pub_buf.as_ptr()};
         let pub_size = crate::buffer_len_to_u32(pub_buf.len())?;
         let rc = unsafe {
             sys::wc_ecc_import_private_key_ex(priv_buf.as_ptr(), priv_size,
-                pub_ptr, pub_size, &mut ecc.wc_ecc_key, curve_id)
+                pub_ptr, pub_size, ecc.wc_ecc_key, curve_id)
         };
         if rc != 0 {
             return Err(rc);
@@ -889,27 +863,24 @@ impl ECC {
     /// ```
     #[cfg(ecc_import)]
     pub fn import_raw(qx: &[u8], qy: &[u8], d: &[u8], curve_name: &[u8], heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
-        let mut wc_ecc_key: MaybeUninit<sys::ecc_key> = MaybeUninit::uninit();
-        let heap = match heap {
-            Some(heap) => heap,
-            None => core::ptr::null_mut(),
-        };
-        let dev_id = match dev_id {
-            Some(dev_id) => dev_id,
-            None => sys::INVALID_DEVID,
-        };
-        let rc = unsafe { sys::wc_ecc_init_ex(wc_ecc_key.as_mut_ptr(), heap, dev_id) };
-        if rc != 0 {
-            return Err(rc);
+        if qx.is_empty() || qy.is_empty() || d.is_empty() || curve_name.is_empty() ||
+            qx[qx.len() - 1] != 0 || qy[qy.len() - 1] != 0 || d[d.len() - 1] != 0 || curve_name[curve_name.len() - 1] != 0 {
+            return Err(sys::wolfCrypt_ErrorCodes_BAD_FUNC_ARG);
         }
-        let wc_ecc_key = unsafe { wc_ecc_key.assume_init() };
-        let mut ecc = ECC { wc_ecc_key };
+        let heap = heap.unwrap_or(core::ptr::null_mut());
+        let dev_id = dev_id.unwrap_or(sys::INVALID_DEVID);
+        let wc_ecc_key = Self::new_ecc_key(heap, dev_id)?;
+        let ecc = ECC {
+            wc_ecc_key,
+            #[cfg(random)]
+            rng: None,
+        };
         let qx_ptr = qx.as_ptr() as *const core::ffi::c_char;
         let qy_ptr = qy.as_ptr() as *const core::ffi::c_char;
         let d_ptr = d.as_ptr() as *const core::ffi::c_char;
         let curve_name_ptr = curve_name.as_ptr() as *const core::ffi::c_char;
         let rc = unsafe {
-            sys::wc_ecc_import_raw(&mut ecc.wc_ecc_key, qx_ptr, qy_ptr, d_ptr,
+            sys::wc_ecc_import_raw(ecc.wc_ecc_key, qx_ptr, qy_ptr, d_ptr,
                 curve_name_ptr)
         };
         if rc != 0 {
@@ -949,26 +920,23 @@ impl ECC {
     /// ```
     #[cfg(ecc_import)]
     pub fn import_raw_ex(qx: &[u8], qy: &[u8], d: &[u8], curve_id: i32, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
-        let mut wc_ecc_key: MaybeUninit<sys::ecc_key> = MaybeUninit::uninit();
-        let heap = match heap {
-            Some(heap) => heap,
-            None => core::ptr::null_mut(),
-        };
-        let dev_id = match dev_id {
-            Some(dev_id) => dev_id,
-            None => sys::INVALID_DEVID,
-        };
-        let rc = unsafe { sys::wc_ecc_init_ex(wc_ecc_key.as_mut_ptr(), heap, dev_id) };
-        if rc != 0 {
-            return Err(rc);
+        if qx.is_empty() || qy.is_empty() || d.is_empty() ||
+            qx[qx.len() - 1] != 0 || qy[qy.len() - 1] != 0 || d[d.len() - 1] != 0 {
+            return Err(sys::wolfCrypt_ErrorCodes_BAD_FUNC_ARG);
         }
-        let wc_ecc_key = unsafe { wc_ecc_key.assume_init() };
-        let mut ecc = ECC { wc_ecc_key };
+        let heap = heap.unwrap_or(core::ptr::null_mut());
+        let dev_id = dev_id.unwrap_or(sys::INVALID_DEVID);
+        let wc_ecc_key = Self::new_ecc_key(heap, dev_id)?;
+        let ecc = ECC {
+            wc_ecc_key,
+            #[cfg(random)]
+            rng: None,
+        };
         let qx_ptr = qx.as_ptr() as *const core::ffi::c_char;
         let qy_ptr = qy.as_ptr() as *const core::ffi::c_char;
         let d_ptr = d.as_ptr() as *const core::ffi::c_char;
         let rc = unsafe {
-            sys::wc_ecc_import_raw_ex(&mut ecc.wc_ecc_key, qx_ptr, qy_ptr,
+            sys::wc_ecc_import_raw_ex(ecc.wc_ecc_key, qx_ptr, qy_ptr,
                 d_ptr, curve_id)
         };
         if rc != 0 {
@@ -1001,10 +969,10 @@ impl ECC {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
+    /// let rng = RNG::new().expect("Failed to create RNG");
     /// let curve_id = ECC::SECP256R1;
     /// let curve_size = ECC::get_curve_size_from_id(curve_id).expect("Error with get_curve_size_from_id()");
-    /// let mut ecc = ECC::generate_ex(curve_size, &mut rng, curve_id, None, None).expect("Error with generate()");
+    /// let mut ecc = ECC::generate_ex(curve_size, &rng, curve_id, None, None).expect("Error with generate()");
     /// let mut qx = [0u8; 32];
     /// let mut qx_len = 0u32;
     /// let mut qy = [0u8; 32];
@@ -1017,23 +985,20 @@ impl ECC {
     /// ```
     #[cfg(ecc_import)]
     pub fn import_unsigned(qx: &[u8], qy: &[u8], d: &[u8], curve_id: i32, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
-        let mut wc_ecc_key: MaybeUninit<sys::ecc_key> = MaybeUninit::uninit();
-        let heap = match heap {
-            Some(heap) => heap,
-            None => core::ptr::null_mut(),
-        };
-        let dev_id = match dev_id {
-            Some(dev_id) => dev_id,
-            None => sys::INVALID_DEVID,
-        };
-        let rc = unsafe { sys::wc_ecc_init_ex(wc_ecc_key.as_mut_ptr(), heap, dev_id) };
-        if rc != 0 {
-            return Err(rc);
+        let curve_size = Self::get_curve_size_from_id(curve_id)? as usize;
+        if qx.len() < curve_size || qy.len() < curve_size || d.len() < curve_size {
+            return Err(sys::wolfCrypt_ErrorCodes_BAD_FUNC_ARG);
         }
-        let wc_ecc_key = unsafe { wc_ecc_key.assume_init() };
-        let mut ecc = ECC { wc_ecc_key };
+        let heap = heap.unwrap_or(core::ptr::null_mut());
+        let dev_id = dev_id.unwrap_or(sys::INVALID_DEVID);
+        let wc_ecc_key = Self::new_ecc_key(heap, dev_id)?;
+        let ecc = ECC {
+            wc_ecc_key,
+            #[cfg(random)]
+            rng: None,
+        };
         let rc = unsafe {
-            sys::wc_ecc_import_unsigned(&mut ecc.wc_ecc_key, qx.as_ptr(),
+            sys::wc_ecc_import_unsigned(ecc.wc_ecc_key, qx.as_ptr(),
                 qy.as_ptr(), d.as_ptr(), curve_id)
         };
         if rc != 0 {
@@ -1065,8 +1030,8 @@ impl ECC {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
-    /// let mut ecc = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
+    /// let rng = RNG::new().expect("Failed to create RNG");
+    /// let mut ecc = ECC::generate(32, &rng, None, None).expect("Error with generate()");
     /// let mut x963 = [0u8; 128];
     /// let x963_size = ecc.export_x963(&mut x963).expect("Error with export_x963()");
     /// let x963 = &x963[0..x963_size];
@@ -1076,23 +1041,16 @@ impl ECC {
     #[cfg(ecc_import)]
     pub fn import_x963(din: &[u8], heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<ECC, i32> {
         let din_size = crate::buffer_len_to_u32(din.len())?;
-        let mut wc_ecc_key: MaybeUninit<sys::ecc_key> = MaybeUninit::uninit();
-        let heap = match heap {
-            Some(heap) => heap,
-            None => core::ptr::null_mut(),
+        let heap = heap.unwrap_or(core::ptr::null_mut());
+        let dev_id = dev_id.unwrap_or(sys::INVALID_DEVID);
+        let wc_ecc_key = Self::new_ecc_key(heap, dev_id)?;
+        let ecc = ECC {
+            wc_ecc_key,
+            #[cfg(random)]
+            rng: None,
         };
-        let dev_id = match dev_id {
-            Some(dev_id) => dev_id,
-            None => sys::INVALID_DEVID,
-        };
-        let rc = unsafe { sys::wc_ecc_init_ex(wc_ecc_key.as_mut_ptr(), heap, dev_id) };
-        if rc != 0 {
-            return Err(rc);
-        }
-        let wc_ecc_key = unsafe { wc_ecc_key.assume_init() };
-        let mut ecc = ECC { wc_ecc_key };
         let rc = unsafe {
-            sys::wc_ecc_import_x963(din.as_ptr(), din_size, &mut ecc.wc_ecc_key)
+            sys::wc_ecc_import_x963(din.as_ptr(), din_size, ecc.wc_ecc_key)
         };
         if rc != 0 {
             return Err(rc);
@@ -1126,10 +1084,10 @@ impl ECC {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
+    /// let rng = RNG::new().expect("Failed to create RNG");
     /// let curve_id = ECC::SECP256R1;
     /// let curve_size = ECC::get_curve_size_from_id(curve_id).expect("Error with get_curve_size_from_id()");
-    /// let mut ecc = ECC::generate_ex(curve_size, &mut rng, curve_id, None, None).expect("Error with generate_ex()");
+    /// let mut ecc = ECC::generate_ex(curve_size, &rng, curve_id, None, None).expect("Error with generate_ex()");
     /// let mut x963 = [0u8; 128];
     /// let x963_size = ecc.export_x963(&mut x963).expect("Error with export_x963()");
     /// let x963 = &x963[0..x963_size];
@@ -1139,23 +1097,16 @@ impl ECC {
     #[cfg(ecc_import)]
     pub fn import_x963_ex(din: &[u8], curve_id: i32, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<ECC, i32> {
         let din_size = crate::buffer_len_to_u32(din.len())?;
-        let mut wc_ecc_key: MaybeUninit<sys::ecc_key> = MaybeUninit::uninit();
-        let heap = match heap {
-            Some(heap) => heap,
-            None => core::ptr::null_mut(),
+        let heap = heap.unwrap_or(core::ptr::null_mut());
+        let dev_id = dev_id.unwrap_or(sys::INVALID_DEVID);
+        let wc_ecc_key = Self::new_ecc_key(heap, dev_id)?;
+        let ecc = ECC {
+            wc_ecc_key,
+            #[cfg(random)]
+            rng: None,
         };
-        let dev_id = match dev_id {
-            Some(dev_id) => dev_id,
-            None => sys::INVALID_DEVID,
-        };
-        let rc = unsafe { sys::wc_ecc_init_ex(wc_ecc_key.as_mut_ptr(), heap, dev_id) };
-        if rc != 0 {
-            return Err(rc);
-        }
-        let wc_ecc_key = unsafe { wc_ecc_key.assume_init() };
-        let mut ecc = ECC { wc_ecc_key };
         let rc = unsafe {
-            sys::wc_ecc_import_x963_ex(din.as_ptr(), din_size, &mut ecc.wc_ecc_key, curve_id)
+            sys::wc_ecc_import_x963_ex(din.as_ptr(), din_size, ecc.wc_ecc_key, curve_id)
         };
         if rc != 0 {
             return Err(rc);
@@ -1196,13 +1147,13 @@ impl ECC {
     ///     hex_string.push('\0');
     ///     hex_string
     /// }
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
+    /// let rng = RNG::new().expect("Failed to create RNG");
     /// let key_path = "../../../certs/ecc-client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut ecc = ECC::import_der(&der, None, None).expect("Error with import_der()");
     /// let hash = [0x42u8; 32];
     /// let mut signature = [0u8; 128];
-    /// let signature_length = ecc.sign_hash(&hash, &mut signature, &mut rng).expect("Error with sign_hash()");
+    /// let signature_length = ecc.sign_hash(&hash, &mut signature, &rng).expect("Error with sign_hash()");
     /// let signature = &mut signature[0..signature_length];
     /// let mut r = [0u8; 32];
     /// let mut r_size = 0u32;
@@ -1258,13 +1209,13 @@ impl ECC {
     /// use std::fs;
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
+    /// let rng = RNG::new().expect("Failed to create RNG");
     /// let key_path = "../../../certs/ecc-client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut ecc = ECC::import_der(&der, None, None).expect("Error with import_der()");
     /// let hash = [0x42u8; 32];
     /// let mut signature = [0u8; 128];
-    /// let signature_length = ecc.sign_hash(&hash, &mut signature, &mut rng).expect("Error with sign_hash()");
+    /// let signature_length = ecc.sign_hash(&hash, &mut signature, &rng).expect("Error with sign_hash()");
     /// let signature = &mut signature[0..signature_length];
     /// let mut r = [0u8; 32];
     /// let mut r_size = 0u32;
@@ -1311,13 +1262,13 @@ impl ECC {
     /// use std::fs;
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
+    /// let rng = RNG::new().expect("Failed to create RNG");
     /// let key_path = "../../../certs/ecc-client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut ecc = ECC::import_der(&der, None, None).expect("Error with import_der()");
     /// let hash = [0x42u8; 32];
     /// let mut signature = [0u8; 128];
-    /// let signature_length = ecc.sign_hash(&hash, &mut signature, &mut rng).expect("Error with sign_hash()");
+    /// let signature_length = ecc.sign_hash(&hash, &mut signature, &rng).expect("Error with sign_hash()");
     /// let signature = &mut signature[0..signature_length];
     /// let mut r = [0u8; 32];
     /// let mut r_size = 0u32;
@@ -1360,13 +1311,13 @@ impl ECC {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
-    /// let mut ecc = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
+    /// let rng = RNG::new().expect("Failed to create RNG");
+    /// let mut ecc = ECC::generate(32, &rng, None, None).expect("Error with generate()");
     /// ecc.check().expect("Error with check()");
     /// }
     /// ```
     pub fn check(&mut self) -> Result<(), i32> {
-        let rc = unsafe { sys::wc_ecc_check_key(&mut self.wc_ecc_key) };
+        let rc = unsafe { sys::wc_ecc_check_key(self.wc_ecc_key) };
         if rc != 0 {
             return Err(rc);
         }
@@ -1396,8 +1347,8 @@ impl ECC {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
-    /// let mut ecc = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
+    /// let rng = RNG::new().expect("Failed to create RNG");
+    /// let mut ecc = ECC::generate(32, &rng, None, None).expect("Error with generate()");
     /// let mut qx = [0u8; 32];
     /// let mut qx_len = 0u32;
     /// let mut qy = [0u8; 32];
@@ -1414,7 +1365,7 @@ impl ECC {
         *qy_len = crate::buffer_len_to_u32(qy.len())?;
         *d_len = crate::buffer_len_to_u32(d.len())?;
         let rc = unsafe {
-            sys::wc_ecc_export_private_raw(&mut self.wc_ecc_key,
+            sys::wc_ecc_export_private_raw(self.wc_ecc_key,
                 qx.as_mut_ptr(), qx_len,
                 qy.as_mut_ptr(), qy_len,
                 d.as_mut_ptr(), d_len)
@@ -1451,8 +1402,8 @@ impl ECC {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
-    /// let mut ecc = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
+    /// let rng = RNG::new().expect("Failed to create RNG");
+    /// let mut ecc = ECC::generate(32, &rng, None, None).expect("Error with generate()");
     /// let mut qx = [0u8; 32];
     /// let mut qx_len = 0u32;
     /// let mut qy = [0u8; 32];
@@ -1477,7 +1428,7 @@ impl ECC {
                 sys::WC_TYPE_UNSIGNED_BIN as i32
             };
         let rc = unsafe {
-            sys::wc_ecc_export_ex(&mut self.wc_ecc_key,
+            sys::wc_ecc_export_ex(self.wc_ecc_key,
                 qx.as_mut_ptr(), qx_len,
                 qy.as_mut_ptr(), qy_len,
                 d.as_mut_ptr(), d_len,
@@ -1507,8 +1458,8 @@ impl ECC {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
-    /// let mut ecc = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
+    /// let rng = RNG::new().expect("Failed to create RNG");
+    /// let mut ecc = ECC::generate(32, &rng, None, None).expect("Error with generate()");
     /// let mut d = [0u8; 32];
     /// let d_size = ecc.export_private(&mut d).expect("Error with export_private()");
     /// assert_eq!(d_size, 32);
@@ -1518,7 +1469,7 @@ impl ECC {
     pub fn export_private(&mut self, d: &mut [u8]) -> Result<usize, i32> {
         let mut d_size = crate::buffer_len_to_u32(d.len())?;
         let rc = unsafe {
-            sys::wc_ecc_export_private_only(&mut self.wc_ecc_key,
+            sys::wc_ecc_export_private_only(self.wc_ecc_key,
                 d.as_mut_ptr(), &mut d_size)
         };
         if rc != 0 {
@@ -1548,8 +1499,8 @@ impl ECC {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
-    /// let mut ecc = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
+    /// let rng = RNG::new().expect("Failed to create RNG");
+    /// let mut ecc = ECC::generate(32, &rng, None, None).expect("Error with generate()");
     /// let mut qx = [0u8; 32];
     /// let mut qx_len = 0u32;
     /// let mut qy = [0u8; 32];
@@ -1563,7 +1514,7 @@ impl ECC {
         *qx_len = crate::buffer_len_to_u32(qx.len())?;
         *qy_len = crate::buffer_len_to_u32(qy.len())?;
         let rc = unsafe {
-            sys::wc_ecc_export_public_raw(&mut self.wc_ecc_key,
+            sys::wc_ecc_export_public_raw(self.wc_ecc_key,
                 qx.as_mut_ptr(), qx_len,
                 qy.as_mut_ptr(), qy_len)
         };
@@ -1591,8 +1542,8 @@ impl ECC {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
-    /// let mut ecc = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
+    /// let rng = RNG::new().expect("Failed to create RNG");
+    /// let mut ecc = ECC::generate(32, &rng, None, None).expect("Error with generate()");
     /// let mut x963 = [0u8; 128];
     /// let _x963_size = ecc.export_x963(&mut x963).expect("Error with export_x963()");
     /// }
@@ -1601,7 +1552,7 @@ impl ECC {
     pub fn export_x963(&mut self, dout: &mut [u8]) -> Result<usize, i32> {
         let mut out_len = crate::buffer_len_to_u32(dout.len())?;
         let rc = unsafe {
-            sys::wc_ecc_export_x963(&mut self.wc_ecc_key, dout.as_mut_ptr(), &mut out_len)
+            sys::wc_ecc_export_x963(self.wc_ecc_key, dout.as_mut_ptr(), &mut out_len)
         };
         if rc != 0 {
             return Err(rc);
@@ -1627,8 +1578,8 @@ impl ECC {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
-    /// let mut ecc = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
+    /// let rng = RNG::new().expect("Failed to create RNG");
+    /// let mut ecc = ECC::generate(32, &rng, None, None).expect("Error with generate()");
     /// let mut x963 = [0u8; 128];
     /// let _x963_size = ecc.export_x963_compressed(&mut x963).expect("Error with export_x963_compressed()");
     /// }
@@ -1637,7 +1588,7 @@ impl ECC {
     pub fn export_x963_compressed(&mut self, dout: &mut [u8]) -> Result<usize, i32> {
         let mut out_len = crate::buffer_len_to_u32(dout.len())?;
         let rc = unsafe {
-            sys::wc_ecc_export_x963_ex(&mut self.wc_ecc_key, dout.as_mut_ptr(), &mut out_len, 1)
+            sys::wc_ecc_export_x963_ex(self.wc_ecc_key, dout.as_mut_ptr(), &mut out_len, 1)
         };
         if rc != 0 {
             return Err(rc);
@@ -1666,21 +1617,21 @@ impl ECC {
     /// use std::fs;
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
+    /// let rng = RNG::new().expect("Failed to create RNG");
     /// let key_path = "../../../certs/ecc-client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut ecc = ECC::import_der(&der, None, None).expect("Error with import_der()");
-    /// ecc.make_pub(Some(&mut rng)).expect("Error with make_pub()");
+    /// ecc.make_pub(Some(&rng)).expect("Error with make_pub()");
     /// }
     /// ```
     #[cfg(random)]
-    pub fn make_pub(&mut self, rng: Option<&mut RNG>) -> Result<(), i32> {
+    pub fn make_pub(&mut self, rng: Option<&RNG>) -> Result<(), i32> {
         let rng_ptr = match rng {
-            Some(rng) => &mut rng.wc_rng,
+            Some(rng) => rng.wc_rng,
             None => core::ptr::null_mut(),
         };
         let rc = unsafe {
-            sys::wc_ecc_make_pub_ex(&mut self.wc_ecc_key, core::ptr::null_mut(), rng_ptr)
+            sys::wc_ecc_make_pub_ex(self.wc_ecc_key, core::ptr::null_mut(), rng_ptr)
         };
         if rc != 0 {
             return Err(rc);
@@ -1710,17 +1661,17 @@ impl ECC {
     /// use std::fs;
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
+    /// let rng = RNG::new().expect("Failed to create RNG");
     /// let key_path = "../../../certs/ecc-client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut ecc = ECC::import_der(&der, None, None).expect("Error with import_der()");
-    /// ecc.make_pub_to_point(Some(&mut rng), None).expect("Error with make_pub_to_point()");
+    /// ecc.make_pub_to_point(Some(&rng), None).expect("Error with make_pub_to_point()");
     /// }
     /// ```
     #[cfg(random)]
-    pub fn make_pub_to_point(&mut self, rng: Option<&mut RNG>, heap: Option<*mut core::ffi::c_void>) -> Result<ECCPoint, i32> {
+    pub fn make_pub_to_point(&mut self, rng: Option<&RNG>, heap: Option<*mut core::ffi::c_void>) -> Result<ECCPoint, i32> {
         let rng_ptr = match rng {
-            Some(rng) => &mut rng.wc_rng,
+            Some(rng) => rng.wc_rng,
             None => core::ptr::null_mut(),
         };
         let heap = match heap {
@@ -1733,7 +1684,7 @@ impl ECC {
         }
         let ecc_point = ECCPoint { wc_ecc_point, heap };
         let rc = unsafe {
-            sys::wc_ecc_make_pub_ex(&mut self.wc_ecc_key, wc_ecc_point, rng_ptr)
+            sys::wc_ecc_make_pub_ex(self.wc_ecc_key, wc_ecc_point, rng_ptr)
         };
         if rc != 0 {
             return Err(rc);
@@ -1749,8 +1700,7 @@ impl ECC {
     /// # Parameters
     ///
     /// * `rng`: The `RNG` struct instance to associate with this `ECC`
-    ///   instance. The `RNG` struct should not be moved in memory after
-    ///   calling this method.
+    ///   instance.
     ///
     /// # Returns
     ///
@@ -1765,22 +1715,51 @@ impl ECC {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
-    /// let mut ecc = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
-    /// ecc.set_rng(&mut rng).expect("Error with set_rng()");
+    /// let blinding_rng = RNG::new().expect("Failed to create RNG");
+    /// let key_gen_rng = RNG::new().expect("Failed to create RNG");
+    /// let mut ecc = ECC::generate(32, &key_gen_rng, None, None).expect("Error with generate()");
+    /// ecc.set_rng(blinding_rng).expect("Error with set_rng()");
     /// }
     /// ```
     #[cfg(random)]
-    pub fn set_rng(&mut self, rng: &mut RNG) -> Result<(), i32> {
+    pub fn set_rng(&mut self, rng: RNG) -> Result<(), i32> {
+        let wc_rng = rng.wc_rng;
         let rc = unsafe {
-            sys::wc_ecc_set_rng(&mut self.wc_ecc_key, &mut rng.wc_rng)
+            sys::wc_ecc_set_rng(self.wc_ecc_key, wc_rng)
         };
         if rc != 0 {
             return Err(rc);
         }
+        self.rng = Some(RngHandle::Owned(rng));
         Ok(())
     }
 
+    /// Bind a shared `RNG` to this key. Available when the `alloc` feature
+    /// is enabled.
+    #[cfg(all(random, feature = "alloc"))]
+    pub fn set_shared_rng(&mut self, rng: alloc::rc::Rc<RNG>) -> Result<(), i32> {
+        let wc_rng = rng.wc_rng;
+        let rc = unsafe {
+            sys::wc_ecc_set_rng(self.wc_ecc_key, wc_rng)
+        };
+        if rc != 0 {
+            return Err(rc);
+        }
+        self.rng = Some(RngHandle::Shared(rng));
+        Ok(())
+    }
+
+    /// Borrow the RNG previously bound via `set_rng` or `set_shared_rng`.
+    #[cfg(random)]
+    pub fn rng(&self) -> Option<&RNG> {
+        match &self.rng {
+            Some(RngHandle::Owned(rng)) => Some(rng),
+            #[cfg(feature = "alloc")]
+            Some(RngHandle::Shared(rng)) => Some(rng),
+            None => None,
+        }
+    }
+
     /// Compute the ECDH shared secret using this key's private component
     /// and the peer public key.
     ///
@@ -1797,17 +1776,18 @@ impl ECC {
     /// # Example
     ///
     /// ```rust
-    /// #[cfg(all(ecc_dh, random))]
+    /// #[cfg(all(ecc_dh, random, feature = "alloc"))]
     /// {
+    /// use std::rc::Rc;
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
-    /// let mut ecc0 = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
-    /// let mut ecc1 = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
+    /// let rng = Rc::new(RNG::new().expect("Failed to create RNG"));
+    /// let mut ecc0 = ECC::generate(32, &rng, None, None).expect("Error with generate()");
+    /// let mut ecc1 = ECC::generate(32, &rng, None, None).expect("Error with generate()");
     /// let mut ss0 = [0u8; 128];
     /// let mut ss1 = [0u8; 128];
-    /// ecc0.set_rng(&mut rng).expect("Error with set_rng()");
-    /// ecc1.set_rng(&mut rng).expect("Error with set_rng()");
+    /// ecc0.set_shared_rng(Rc::clone(&rng)).expect("Error with set_shared_rng()");
+    /// ecc1.set_shared_rng(Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let ss0_size = ecc0.shared_secret(&mut ecc1, &mut ss0).expect("Error with shared_secret()");
     /// let ss1_size = ecc1.shared_secret(&mut ecc0, &mut ss1).expect("Error with shared_secret()");
     /// assert_eq!(ss0_size, ss1_size);
@@ -1820,8 +1800,8 @@ impl ECC {
     pub fn shared_secret(&mut self, peer_key: &mut ECC, dout: &mut [u8]) -> Result<usize, i32> {
         let mut out_len = crate::buffer_len_to_u32(dout.len())?;
         let rc = unsafe {
-            sys::wc_ecc_shared_secret(&mut self.wc_ecc_key,
-                &mut peer_key.wc_ecc_key, dout.as_mut_ptr(), &mut out_len)
+            sys::wc_ecc_shared_secret(self.wc_ecc_key,
+                peer_key.wc_ecc_key, dout.as_mut_ptr(), &mut out_len)
         };
         if rc != 0 {
             return Err(rc);
@@ -1846,18 +1826,19 @@ impl ECC {
     /// # Example
     ///
     /// ```rust
-    /// #[cfg(all(ecc_dh, random))]
+    /// #[cfg(all(ecc_dh, random, feature = "alloc"))]
     /// {
+    /// use std::rc::Rc;
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
-    /// let mut ecc0 = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
-    /// let mut ecc1 = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
+    /// let rng = Rc::new(RNG::new().expect("Failed to create RNG"));
+    /// let mut ecc0 = ECC::generate(32, &rng, None, None).expect("Error with generate()");
+    /// let mut ecc1 = ECC::generate(32, &rng, None, None).expect("Error with generate()");
     /// let ecc1_point = ecc1.make_pub_to_point(None, None).expect("Error with make_pub_to_point()");
     /// let mut ss0 = [0u8; 128];
     /// let mut ss1 = [0u8; 128];
-    /// ecc0.set_rng(&mut rng).expect("Error with set_rng()");
-    /// ecc1.set_rng(&mut rng).expect("Error with set_rng()");
+    /// ecc0.set_shared_rng(Rc::clone(&rng)).expect("Error with set_shared_rng()");
+    /// ecc1.set_shared_rng(Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let ss0_size = ecc0.shared_secret_ex(&ecc1_point, &mut ss0).expect("Error with shared_secret_ex()");
     /// let ss1_size = ecc1.shared_secret(&mut ecc0, &mut ss1).expect("Error with shared_secret()");
     /// assert_eq!(ss0_size, ss1_size);
@@ -1870,7 +1851,7 @@ impl ECC {
     pub fn shared_secret_ex(&mut self, peer: &ECCPoint, dout: &mut [u8]) -> Result<usize, i32> {
         let mut out_len = crate::buffer_len_to_u32(dout.len())?;
         let rc = unsafe {
-            sys::wc_ecc_shared_secret_ex(&mut self.wc_ecc_key,
+            sys::wc_ecc_shared_secret_ex(self.wc_ecc_key,
                 peer.wc_ecc_point, dout.as_mut_ptr(), &mut out_len)
         };
         if rc != 0 {
@@ -1899,23 +1880,23 @@ impl ECC {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
-    /// let mut ecc = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
+    /// let rng = RNG::new().expect("Failed to create RNG");
+    /// let mut ecc = ECC::generate(32, &rng, None, None).expect("Error with generate()");
     /// let hash = [0x42u8; 32];
     /// let mut signature = [0u8; 128];
-    /// let signature_length = ecc.sign_hash(&hash, &mut signature, &mut rng).expect("Error with sign_hash()");
+    /// let signature_length = ecc.sign_hash(&hash, &mut signature, &rng).expect("Error with sign_hash()");
     /// let signature = &mut signature[0..signature_length];
     /// let valid = ecc.verify_hash(&signature, &hash).expect("Error with verify_hash()");
     /// assert_eq!(valid, true);
     /// }
     /// ```
     #[cfg(all(ecc_sign, random))]
-    pub fn sign_hash(&mut self, din: &[u8], dout: &mut [u8], rng: &mut RNG) -> Result<usize, i32> {
+    pub fn sign_hash(&mut self, din: &[u8], dout: &mut [u8], rng: &RNG) -> Result<usize, i32> {
         let din_size = crate::buffer_len_to_u32(din.len())?;
         let mut dout_size = crate::buffer_len_to_u32(dout.len())?;
         let rc = unsafe {
             sys::wc_ecc_sign_hash(din.as_ptr(), din_size, dout.as_mut_ptr(),
-                &mut dout_size, &mut rng.wc_rng, &mut self.wc_ecc_key)
+                &mut dout_size, rng.wc_rng, self.wc_ecc_key)
         };
         if rc != 0 {
             return Err(rc);
@@ -1942,11 +1923,11 @@ impl ECC {
     /// {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::ecc::ECC;
-    /// let mut rng = RNG::new().expect("Failed to create RNG");
-    /// let mut ecc = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
+    /// let rng = RNG::new().expect("Failed to create RNG");
+    /// let mut ecc = ECC::generate(32, &rng, None, None).expect("Error with generate()");
     /// let hash = [0x42u8; 32];
     /// let mut signature = [0u8; 128];
-    /// let signature_length = ecc.sign_hash(&hash, &mut signature, &mut rng).expect("Error with sign_hash()");
+    /// let signature_length = ecc.sign_hash(&hash, &mut signature, &rng).expect("Error with sign_hash()");
     /// let signature = &mut signature[0..signature_length];
     /// let valid = ecc.verify_hash(&signature, &hash).expect("Error with verify_hash()");
     /// assert_eq!(valid, true);
@@ -1959,7 +1940,7 @@ impl ECC {
         let hash_len = crate::buffer_len_to_u32(hash.len())?;
         let rc = unsafe {
             sys::wc_ecc_verify_hash(sig.as_ptr(), sig_len,
-                hash.as_ptr(), hash_len, &mut res, &mut self.wc_ecc_key)
+                hash.as_ptr(), hash_len, &mut res, self.wc_ecc_key)
         };
         if rc != 0 {
             return Err(rc);
@@ -1968,22 +1949,16 @@ impl ECC {
     }
 }
 
-impl ECC {
-    fn zeroize(&mut self) {
-        unsafe { crate::zeroize_raw(&mut self.wc_ecc_key); }
-    }
-}
-
 impl Drop for ECC {
     /// Safely free the underlying wolfSSL ECC context.
     ///
-    /// This calls the `wc_ecc_key_free()` wolfssl library function.
+    /// This calls the `wc_ecc_key_free()` wolfssl library function, which
+    /// frees the C-heap-allocated `ecc_key` object.
     ///
     /// The Rust Drop trait guarantees that this method is called when the ECC
     /// struct goes out of scope, automatically cleaning up resources and
     /// preventing memory leaks.
     fn drop(&mut self) {
-        unsafe { sys::wc_ecc_free(&mut self.wc_ecc_key); }
-        self.zeroize();
+        unsafe { sys::wc_ecc_key_free(self.wc_ecc_key); }
     }
 }
diff --git a/wrapper/rust/wolfssl-wolfcrypt/src/ecdsa.rs b/wrapper/rust/wolfssl-wolfcrypt/src/ecdsa.rs
index b771694b05..f3c06d451f 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/src/ecdsa.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/src/ecdsa.rs
@@ -252,9 +252,9 @@ macro_rules! define_ecdsa_curve {
                         sys::wc_SignatureType_WC_SIGNATURE_TYPE_ECC,
                         msg.as_ptr(), msg_len,
                         der.as_mut_ptr(), &mut der_len,
-                        &mut self.inner.wc_ecc_key as *mut _ as *mut c_void,
+                        self.inner.wc_ecc_key as *mut c_void,
                         size_of::<sys::ecc_key>() as u32,
-                        &mut self.rng.wc_rng,
+                        self.rng.wc_rng,
                     )
                 };
                 if rc != 0 {
@@ -313,7 +313,7 @@ macro_rules! define_ecdsa_curve {
             fn verify(&self, msg: &[u8], sig: &$signature) -> Result<(), Error> {
                 let mut der = [0u8; $der_max];
                 let der_len = rs_to_der::<$field_size>(&sig.0, &mut der)?;
-                let mut key = ECC::import_x963_ex(&self.pub_bytes, $curve_id, None, None)
+                let key = ECC::import_x963_ex(&self.pub_bytes, $curve_id, None, None)
                     .map_err(|_| Error::new())?;
                 let msg_len: u32 = msg.len().try_into().map_err(|_| Error::new())?;
                 let rc = unsafe {
@@ -322,7 +322,7 @@ macro_rules! define_ecdsa_curve {
                         sys::wc_SignatureType_WC_SIGNATURE_TYPE_ECC,
                         msg.as_ptr(), msg_len,
                         der.as_ptr(), der_len as u32,
-                        &mut key.wc_ecc_key as *mut _ as *mut c_void,
+                        key.wc_ecc_key as *mut c_void,
                         size_of::<sys::ecc_key>() as u32,
                     )
                 };
diff --git a/wrapper/rust/wolfssl-wolfcrypt/src/ed25519.rs b/wrapper/rust/wolfssl-wolfcrypt/src/ed25519.rs
index a0b91f1e27..5a950e729c 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/src/ed25519.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/src/ed25519.rs
@@ -72,7 +72,7 @@ impl Ed25519 {
     /// let mut rng = RNG::new().expect("Error creating RNG");
     /// let ed = Ed25519::generate(&mut rng).expect("Error with generate()");
     /// ```
-    pub fn generate(rng: &mut RNG) -> Result<Self, i32> {
+    pub fn generate(rng: &RNG) -> Result<Self, i32> {
         Self::generate_ex(rng, None, None)
     }
 
@@ -97,7 +97,7 @@ impl Ed25519 {
     /// let mut rng = RNG::new().expect("Error creating RNG");
     /// let ed = Ed25519::generate_ex(&mut rng, None, None).expect("Error with generate_ex()");
     /// ```
-    pub fn generate_ex(rng: &mut RNG, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
+    pub fn generate_ex(rng: &RNG, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
         let mut ws_key: MaybeUninit<sys::ed25519_key> = MaybeUninit::uninit();
         let heap = match heap {
             Some(heap) => heap,
@@ -114,7 +114,7 @@ impl Ed25519 {
         let ws_key = unsafe { ws_key.assume_init() };
         let mut ed25519 = Ed25519 { ws_key };
         let rc = unsafe {
-            sys::wc_ed25519_make_key(&mut rng.wc_rng,
+            sys::wc_ed25519_make_key(rng.wc_rng,
                 sys::ED25519_KEY_SIZE as i32, &mut ed25519.ws_key)
         };
         if rc != 0 {
diff --git a/wrapper/rust/wolfssl-wolfcrypt/src/ed448.rs b/wrapper/rust/wolfssl-wolfcrypt/src/ed448.rs
index 0a77aa071e..bb134ce796 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/src/ed448.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/src/ed448.rs
@@ -71,7 +71,7 @@ impl Ed448 {
     /// let mut rng = RNG::new().expect("Error creating RNG");
     /// let ed = Ed448::generate(&mut rng).expect("Error with generate()");
     /// ```
-    pub fn generate(rng: &mut RNG) -> Result<Self, i32> {
+    pub fn generate(rng: &RNG) -> Result<Self, i32> {
         Self::generate_ex(rng, None, None)
     }
 
@@ -96,7 +96,7 @@ impl Ed448 {
     /// let mut rng = RNG::new().expect("Error creating RNG");
     /// let ed = Ed448::generate_ex(&mut rng, None, None).expect("Error with generate_ex()");
     /// ```
-    pub fn generate_ex(rng: &mut RNG, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
+    pub fn generate_ex(rng: &RNG, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
         let mut ws_key: MaybeUninit<sys::ed448_key> = MaybeUninit::uninit();
         let heap = match heap {
             Some(heap) => heap,
@@ -113,7 +113,7 @@ impl Ed448 {
         let ws_key = unsafe { ws_key.assume_init() };
         let mut ed448 = Ed448 { ws_key };
         let rc = unsafe {
-            sys::wc_ed448_make_key(&mut rng.wc_rng,
+            sys::wc_ed448_make_key(rng.wc_rng,
                 sys::ED448_KEY_SIZE as i32, &mut ed448.ws_key)
         };
         if rc != 0 {
diff --git a/wrapper/rust/wolfssl-wolfcrypt/src/hmac.rs b/wrapper/rust/wolfssl-wolfcrypt/src/hmac.rs
index b1d004006c..11a05adbbd 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/src/hmac.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/src/hmac.rs
@@ -330,6 +330,24 @@ impl HMAC {
     }
 }
 
+impl Clone for HMAC {
+    /// Deep-copy the HMAC state via `wc_HmacCopy()`.
+    ///
+    /// Panics if the underlying wolfSSL copy fails.
+    fn clone(&self) -> Self {
+        let mut wc_hmac: MaybeUninit<sys::Hmac> = MaybeUninit::uninit();
+        let rc = unsafe {
+            sys::wc_HmacCopy(&self.wc_hmac as *const _ as *mut _,
+                wc_hmac.as_mut_ptr())
+        };
+        if rc != 0 {
+            panic!("wc_HmacCopy() failed: {}", rc);
+        }
+        let wc_hmac = unsafe { wc_hmac.assume_init() };
+        HMAC { wc_hmac }
+    }
+}
+
 impl Drop for HMAC {
     /// Safely free the underlying wolfSSL Hmac context.
     ///
diff --git a/wrapper/rust/wolfssl-wolfcrypt/src/hmac_mac.rs b/wrapper/rust/wolfssl-wolfcrypt/src/hmac_mac.rs
index 452debaa22..e5060d6f8d 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/src/hmac_mac.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/src/hmac_mac.rs
@@ -42,6 +42,7 @@ macro_rules! impl_hmac_mac {
         $name:ident, hmac_type = $hmac_type:expr, key = $key_size:ty, out = $out_size:ty
     ) => {
         $(#[$attr])*
+        #[derive(Clone)]
         pub struct $name {
             hmac: crate::hmac::HMAC,
         }
diff --git a/wrapper/rust/wolfssl-wolfcrypt/src/lib.rs b/wrapper/rust/wolfssl-wolfcrypt/src/lib.rs
index 1c6a23dcdd..36dc42bd94 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/src/lib.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/src/lib.rs
@@ -20,6 +20,9 @@
 
 #![no_std]
 
+#[cfg(feature = "alloc")]
+extern crate alloc;
+
 /* bindgen-generated bindings to the C library */
 pub mod sys;
 
@@ -42,6 +45,10 @@ pub(crate) unsafe fn zeroize_raw<T>(val: &mut T) {
 
 pub mod aes;
 pub mod blake2;
+#[cfg(all(any(blake2b, blake2s), feature = "digest"))]
+pub mod blake2_digest;
+#[cfg(all(any(blake2b, blake2s), feature = "mac"))]
+pub mod blake2_mac;
 pub mod chacha20_poly1305;
 pub mod cmac;
 #[cfg(all(cmac, feature = "mac"))]
diff --git a/wrapper/rust/wolfssl-wolfcrypt/src/lms.rs b/wrapper/rust/wolfssl-wolfcrypt/src/lms.rs
index 90d4f1127f..4eca40433c 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/src/lms.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/src/lms.rs
@@ -442,8 +442,8 @@ impl Lms {
     /// }
     /// ```
     #[cfg(all(lms_make_key, random))]
-    pub fn make_key(&mut self, rng: &mut RNG) -> Result<(), i32> {
-        let rc = unsafe { sys::wc_LmsKey_MakeKey(&mut self.ws_key, &mut rng.wc_rng) };
+    pub fn make_key(&mut self, rng: &RNG) -> Result<(), i32> {
+        let rc = unsafe { sys::wc_LmsKey_MakeKey(&mut self.ws_key, rng.wc_rng) };
         if rc != 0 {
             return Err(rc);
         }
diff --git a/wrapper/rust/wolfssl-wolfcrypt/src/mlkem.rs b/wrapper/rust/wolfssl-wolfcrypt/src/mlkem.rs
index e63709dea5..b2711fc996 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/src/mlkem.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/src/mlkem.rs
@@ -123,7 +123,7 @@ impl MlKem {
     /// }
     /// ```
     #[cfg(random)]
-    pub fn generate(key_type: i32, rng: &mut RNG) -> Result<Self, i32> {
+    pub fn generate(key_type: i32, rng: &RNG) -> Result<Self, i32> {
         Self::generate_ex(key_type, rng, None, None)
     }
 
@@ -157,12 +157,12 @@ impl MlKem {
     #[cfg(random)]
     pub fn generate_ex(
         key_type: i32,
-        rng: &mut RNG,
+        rng: &RNG,
         heap: Option<*mut core::ffi::c_void>,
         dev_id: Option<i32>,
     ) -> Result<Self, i32> {
         let key = Self::new_ex(key_type, heap, dev_id)?;
-        let rc = unsafe { sys::wc_MlKemKey_MakeKey(key.ws_key, &mut rng.wc_rng) };
+        let rc = unsafe { sys::wc_MlKemKey_MakeKey(key.ws_key, rng.wc_rng) };
         if rc != 0 {
             return Err(rc);
         }
@@ -472,7 +472,7 @@ impl MlKem {
         &mut self,
         ct: &mut [u8],
         ss: &mut [u8],
-        rng: &mut RNG,
+        rng: &RNG,
     ) -> Result<(), i32> {
         // Verify the cipher text length is as expected based on the parameter
         // set (key type) in use.
@@ -489,7 +489,7 @@ impl MlKem {
                 self.ws_key,
                 ct.as_mut_ptr(),
                 ss.as_mut_ptr(),
-                &mut rng.wc_rng,
+                rng.wc_rng,
             )
         };
         if rc != 0 {
diff --git a/wrapper/rust/wolfssl-wolfcrypt/src/random.rs b/wrapper/rust/wolfssl-wolfcrypt/src/random.rs
index a5904daa40..5e3a1833ac 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/src/random.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/src/random.rs
@@ -31,7 +31,7 @@ wolfSSL `WC_RNG` object. It ensures proper initialization and deallocation.
 use wolfssl_wolfcrypt::random::RNG;
 
 // Create a RNG instance.
-let mut rng = RNG::new().expect("Failed to create RNG");
+let rng = RNG::new().expect("Failed to create RNG");
 
 // Generate a single random byte value.
 let byte = rng.generate_byte().expect("Failed to generate a single byte");
@@ -45,18 +45,42 @@ rng.generate_block(&mut buffer).expect("Failed to generate a block");
 #![cfg(random)]
 
 use crate::sys;
-use core::mem::{size_of_val, MaybeUninit};
+use core::mem::size_of_val;
 use num_traits::PrimInt;
 
 /// A cryptographically secure random number generator based on the wolfSSL
 /// library.
 ///
-/// This struct wraps the wolfssl `WC_RNG` type, providing a high-level API
-/// for generating random bytes and blocks of data. The `Drop` implementation
-/// ensures that the underlying wolfSSL RNG context is correctly freed when the
-/// `RNG` struct goes out of scope, preventing memory leaks.
+/// This struct wraps a pointer to a wolfssl `WC_RNG` allocated on the C heap,
+/// providing a high-level API for generating random bytes and blocks of data.
+/// The `Drop` implementation ensures that the underlying wolfSSL RNG context is
+/// correctly freed when the `RNG` struct goes out of scope, preventing memory
+/// leaks.
+///
+/// All generation methods take `&self`. The actual mutation of the DRBG state
+/// happens through the raw pointer in the C library; the `RNG` struct itself
+/// is logically immutable after construction.
 pub struct RNG {
-    pub(crate) wc_rng: sys::WC_RNG,
+    pub(crate) wc_rng: *mut sys::WC_RNG,
+}
+
+// Safety: the only field of `RNG` is a non-null pointer to a `WC_RNG` that
+// lives on the C heap and is never reassigned after construction. Moving the
+// struct between threads is sound.
+unsafe impl Send for RNG {}
+
+// Note: `RNG` is intentionally not `Sync`. The underlying C `WC_RNG` state is
+// mutated by every call to a generation routine, with no internal locking.
+// Callers that need cross-thread sharing of a single RNG struct must implement
+// their own locking.
+
+/// Storage for an RNG that a consumer (e.g. `RSA`, `ECC`) has been bound to
+/// via `set_rng`. The consumer keeps the `RngHandle` alive for as long as the
+/// C struct holds its pointer, ensuring the `WC_RNG` outlives the consumer.
+pub(crate) enum RngHandle {
+    Owned(RNG),
+    #[cfg(feature = "alloc")]
+    Shared(alloc::rc::Rc<RNG>),
 }
 
 impl RNG {
@@ -97,7 +121,7 @@ impl RNG {
                 return Err(rc);
             }
         }
-        let mut wc_rng: MaybeUninit<sys::WC_RNG> = MaybeUninit::uninit();
+        let mut wc_rng: *mut sys::WC_RNG = core::ptr::null_mut();
         let heap = match heap {
             Some(heap) => heap,
             None => core::ptr::null_mut(),
@@ -107,12 +131,10 @@ impl RNG {
             None => sys::INVALID_DEVID,
         };
         let rc = unsafe {
-            sys::wc_InitRng_ex(wc_rng.as_mut_ptr(), heap, dev_id)
+            sys::wc_rng_new_ex(&mut wc_rng, core::ptr::null_mut(), 0, heap, dev_id)
         };
         if rc == 0 {
-            let wc_rng = unsafe { wc_rng.assume_init() };
-            let rng = RNG {wc_rng};
-            Ok(rng)
+            Ok(RNG {wc_rng})
         } else {
             Err(rc)
         }
@@ -159,7 +181,7 @@ impl RNG {
         }
         let ptr = nonce.as_mut_ptr() as *mut u8;
         let size = crate::buffer_len_to_u32(size_of_val(nonce))?;
-        let mut wc_rng: MaybeUninit<sys::WC_RNG> = MaybeUninit::uninit();
+        let mut wc_rng: *mut sys::WC_RNG = core::ptr::null_mut();
         let heap = match heap {
             Some(heap) => heap,
             None => core::ptr::null_mut(),
@@ -169,12 +191,10 @@ impl RNG {
             None => sys::INVALID_DEVID,
         };
         let rc = unsafe {
-            sys::wc_InitRngNonce_ex(wc_rng.as_mut_ptr(), ptr, size, heap, dev_id)
+            sys::wc_rng_new_ex(&mut wc_rng, ptr, size, heap, dev_id)
         };
         if rc == 0 {
-            let wc_rng = unsafe { wc_rng.assume_init() };
-            let rng = RNG {wc_rng};
-            Ok(rng)
+            Ok(RNG {wc_rng})
         } else {
             Err(rc)
         }
@@ -315,9 +335,9 @@ impl RNG {
     ///
     /// A `Result` which is `Ok(u8)` containing the random byte on success or
     /// an `Err` with the wolfssl library return code on failure.
-    pub fn generate_byte(&mut self) -> Result<u8, i32> {
+    pub fn generate_byte(&self) -> Result<u8, i32> {
         let mut b: u8 = 0;
-        let rc = unsafe { sys::wc_RNG_GenerateByte(&mut self.wc_rng, &mut b) };
+        let rc = unsafe { sys::wc_RNG_GenerateByte(self.wc_rng, &mut b) };
         if rc == 0 {
             Ok(b)
         } else {
@@ -339,10 +359,10 @@ impl RNG {
     ///
     /// A `Result` which is `Ok(())` on success or an `Err` with the wolfssl
     /// library return code on failure.
-    pub fn generate_block<T: PrimInt>(&mut self, buf: &mut [T]) -> Result<(), i32> {
+    pub fn generate_block<T: PrimInt>(&self, buf: &mut [T]) -> Result<(), i32> {
         let ptr = buf.as_mut_ptr() as *mut u8;
         let size = crate::buffer_len_to_u32(size_of_val(buf))?;
-        let rc = unsafe { sys::wc_RNG_GenerateBlock(&mut self.wc_rng, ptr, size) };
+        let rc = unsafe { sys::wc_RNG_GenerateBlock(self.wc_rng, ptr, size) };
         if rc == 0 {
             Ok(())
         } else {
@@ -371,10 +391,10 @@ impl RNG {
     /// rng.reseed(&seed).expect("Error with reseed()");
     /// ```
     #[cfg(random_hashdrbg)]
-    pub fn reseed(&mut self, seed: &[u8]) -> Result<(), i32> {
+    pub fn reseed(&self, seed: &[u8]) -> Result<(), i32> {
         let seed_size = crate::buffer_len_to_u32(seed.len())?;
         let rc = unsafe {
-            sys::wc_RNG_DRBG_Reseed(&mut self.wc_rng, seed.as_ptr(), seed_size)
+            sys::wc_RNG_DRBG_Reseed(self.wc_rng, seed.as_ptr(), seed_size)
         };
         if rc != 0 {
             return Err(rc);
@@ -411,22 +431,16 @@ impl rand_core::TryRng for RNG {
 #[cfg(feature = "rand_core")]
 impl rand_core::TryCryptoRng for RNG {}
 
-impl RNG {
-    fn zeroize(&mut self) {
-        unsafe { crate::zeroize_raw(&mut self.wc_rng); }
-    }
-}
-
 impl Drop for RNG {
     /// Safely free the underlying wolfSSL RNG context.
     ///
-    /// This calls the `wc_FreeRng` wolfssl library function.
+    /// This calls the `wc_rng_free` wolfssl library function, which frees the
+    /// C-heap-allocated `WC_RNG` object.
     ///
     /// The Rust Drop trait guarantees that this method is called when the RNG
     /// struct goes out of scope, automatically cleaning up resources and
     /// preventing memory leaks.
     fn drop(&mut self) {
-        unsafe { sys::wc_FreeRng(&mut self.wc_rng); }
-        self.zeroize();
+        unsafe { sys::wc_rng_free(self.wc_rng); }
     }
 }
diff --git a/wrapper/rust/wolfssl-wolfcrypt/src/rsa.rs b/wrapper/rust/wolfssl-wolfcrypt/src/rsa.rs
index 3e89b79142..da9dc765d1 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/src/rsa.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/src/rsa.rs
@@ -29,26 +29,26 @@ wolfSSL `RsaKey` object. It ensures proper initialization and deallocation.
 
 ```rust
 # extern crate std;
-#[cfg(random)]
+#[cfg(all(random, feature = "alloc"))]
 {
 use std::fs;
 use wolfssl_wolfcrypt::random::RNG;
 use wolfssl_wolfcrypt::rsa::RSA;
 
-let mut rng = RNG::new().expect("Error creating RNG");
+let rng = std::rc::Rc::new(RNG::new().expect("Error creating RNG"));
 let key_path = "../../../certs/client-keyPub.der";
 let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
 let mut rsa = RSA::new_public_from_der(&der).expect("Error with new_public_from_der()");
-rsa.set_rng(&mut rng).expect("Error with set_rng()");
+rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
 let plain: &[u8] = b"Test message";
 let mut enc: [u8; 512] = [0; 512];
-let enc_len = rsa.public_encrypt(plain, &mut enc, &mut rng).expect("Error with public_encrypt()");
+let enc_len = rsa.public_encrypt(plain, &mut enc, &rng).expect("Error with public_encrypt()");
 assert!(enc_len > 0 && enc_len <= 512);
 
 let key_path = "../../../certs/client-key.der";
 let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
 let mut rsa = RSA::new_from_der(&der).expect("Error with new_from_der()");
-rsa.set_rng(&mut rng).expect("Error with set_rng()");
+rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
 let mut plain_out: [u8; 512] = [0; 512];
 let dec_len = rsa.private_decrypt(&enc[0..enc_len], &mut plain_out).expect("Error with private_decrypt()");
 assert!(dec_len as usize == plain.len());
@@ -61,7 +61,7 @@ assert_eq!(plain_out[0..dec_len], *plain);
 
 use crate::sys;
 #[cfg(random)]
-use crate::random::RNG;
+use crate::random::{RNG, RngHandle};
 use core::mem::{MaybeUninit};
 
 /// The `RSA` struct manages the lifecycle of a wolfSSL `RsaKey` object.
@@ -72,6 +72,10 @@ use core::mem::{MaybeUninit};
 /// or `generate()`.
 pub struct RSA {
     pub(crate) wc_rsakey: sys::RsaKey,
+    /// RNG bound to this key via `set_rng`. Kept alive here so the C struct's
+    /// internal `WC_RNG` pointer remains valid for as long as the key exists.
+    #[cfg(random)]
+    rng: Option<RngHandle>,
 }
 
 impl RSA {
@@ -143,26 +147,26 @@ impl RSA {
     ///
     /// ```rust
     /// # extern crate std;
-    /// #[cfg(random)]
+    /// #[cfg(all(random, feature = "alloc"))]
     /// {
     /// use std::fs;
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
+    /// let rng = std::rc::Rc::new(RNG::new().expect("Error creating RNG"));
     /// let key_path = "../../../certs/client-keyPub.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_public_from_der(&der).expect("Error with new_public_from_der()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let plain: &[u8] = b"Test message";
     /// let mut enc: [u8; 512] = [0; 512];
-    /// let enc_len = rsa.public_encrypt(plain, &mut enc, &mut rng).expect("Error with public_encrypt()");
+    /// let enc_len = rsa.public_encrypt(plain, &mut enc, &rng).expect("Error with public_encrypt()");
     /// assert!(enc_len > 0 && enc_len <= 512);
     ///
     /// let key_path = "../../../certs/client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_from_der(&der).expect("Error with new_from_der()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let mut plain_out: [u8; 512] = [0; 512];
     /// let dec_len = rsa.private_decrypt(&enc[0..enc_len], &mut plain_out).expect("Error with private_decrypt()");
     /// assert!(dec_len as usize == plain.len());
@@ -191,26 +195,26 @@ impl RSA {
     ///
     /// ```rust
     /// # extern crate std;
-    /// #[cfg(random)]
+    /// #[cfg(all(random, feature = "alloc"))]
     /// {
     /// use std::fs;
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
+    /// let rng = std::rc::Rc::new(RNG::new().expect("Error creating RNG"));
     /// let key_path = "../../../certs/client-keyPub.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_public_from_der(&der).expect("Error with new_public_from_der()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let plain: &[u8] = b"Test message";
     /// let mut enc: [u8; 512] = [0; 512];
-    /// let enc_len = rsa.public_encrypt(plain, &mut enc, &mut rng).expect("Error with public_encrypt()");
+    /// let enc_len = rsa.public_encrypt(plain, &mut enc, &rng).expect("Error with public_encrypt()");
     /// assert!(enc_len > 0 && enc_len <= 512);
     ///
     /// let key_path = "../../../certs/client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_from_der_ex(&der, None, None).expect("Error with new_from_der_ex()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let mut plain_out: [u8; 512] = [0; 512];
     /// let dec_len = rsa.private_decrypt(&enc[0..enc_len], &mut plain_out).expect("Error with private_decrypt()");
     /// assert!(dec_len as usize == plain.len());
@@ -241,7 +245,11 @@ impl RSA {
             unsafe { sys::wc_FreeRsaKey(&mut wc_rsakey); }
             return Err(rc);
         }
-        let rsa = RSA { wc_rsakey };
+        let rsa = RSA {
+            wc_rsakey,
+            #[cfg(random)]
+            rng: None,
+        };
         Ok(rsa)
     }
 
@@ -260,26 +268,26 @@ impl RSA {
     ///
     /// ```rust
     /// # extern crate std;
-    /// #[cfg(random)]
+    /// #[cfg(all(random, feature = "alloc"))]
     /// {
     /// use std::fs;
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
+    /// let rng = std::rc::Rc::new(RNG::new().expect("Error creating RNG"));
     /// let key_path = "../../../certs/client-keyPub.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_public_from_der(&der).expect("Error with new_public_from_der()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let plain: &[u8] = b"Test message";
     /// let mut enc: [u8; 512] = [0; 512];
-    /// let enc_len = rsa.public_encrypt(plain, &mut enc, &mut rng).expect("Error with public_encrypt()");
+    /// let enc_len = rsa.public_encrypt(plain, &mut enc, &rng).expect("Error with public_encrypt()");
     /// assert!(enc_len > 0 && enc_len <= 512);
     ///
     /// let key_path = "../../../certs/client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_from_der(&der).expect("Error with new_from_der()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let mut plain_out: [u8; 512] = [0; 512];
     /// let dec_len = rsa.private_decrypt(&enc[0..enc_len], &mut plain_out).expect("Error with private_decrypt()");
     /// assert!(dec_len as usize == plain.len());
@@ -308,26 +316,26 @@ impl RSA {
     ///
     /// ```rust
     /// # extern crate std;
-    /// #[cfg(random)]
+    /// #[cfg(all(random, feature = "alloc"))]
     /// {
     /// use std::fs;
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
+    /// let rng = std::rc::Rc::new(RNG::new().expect("Error creating RNG"));
     /// let key_path = "../../../certs/client-keyPub.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_public_from_der_ex(&der, None, None).expect("Error with new_public_from_der_ex()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let plain: &[u8] = b"Test message";
     /// let mut enc: [u8; 512] = [0; 512];
-    /// let enc_len = rsa.public_encrypt(plain, &mut enc, &mut rng).expect("Error with public_encrypt()");
+    /// let enc_len = rsa.public_encrypt(plain, &mut enc, &rng).expect("Error with public_encrypt()");
     /// assert!(enc_len > 0 && enc_len <= 512);
     ///
     /// let key_path = "../../../certs/client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_from_der(&der).expect("Error with new_from_der()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let mut plain_out: [u8; 512] = [0; 512];
     /// let dec_len = rsa.private_decrypt(&enc[0..enc_len], &mut plain_out).expect("Error with private_decrypt()");
     /// assert!(dec_len as usize == plain.len());
@@ -358,7 +366,11 @@ impl RSA {
             unsafe { sys::wc_FreeRsaKey(&mut wc_rsakey); }
             return Err(rc);
         }
-        let rsa = RSA { wc_rsakey };
+        let rsa = RSA {
+            wc_rsakey,
+            #[cfg(random)]
+            rng: None,
+        };
         Ok(rsa)
     }
 
@@ -411,7 +423,11 @@ impl RSA {
             unsafe { sys::wc_FreeRsaKey(&mut wc_rsakey); }
             return Err(rc);
         }
-        Ok(RSA { wc_rsakey })
+        Ok(RSA {
+            wc_rsakey,
+            #[cfg(random)]
+            rng: None,
+        })
     }
 
     /// Generate a new RSA key using the given size and exponent.
@@ -446,15 +462,15 @@ impl RSA {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
-    /// let mut rsa = RSA::generate(2048, 65537, &mut rng).expect("Error with generate()");
+    /// let rng = RNG::new().expect("Error creating RNG");
+    /// let mut rsa = RSA::generate(2048, 65537, &rng).expect("Error with generate()");
     /// rsa.check().expect("Error with check()");
     /// let encrypt_size = rsa.get_encrypt_size().expect("Error with get_encrypt_size()");
     /// assert_eq!(encrypt_size, 256);
     /// }
     /// ```
     #[cfg(all(random, rsa_keygen))]
-    pub fn generate(size: i32, e: i32, rng: &mut RNG) -> Result<Self, i32> {
+    pub fn generate(size: i32, e: i32, rng: &RNG) -> Result<Self, i32> {
         Self::generate_ex(size, e, rng, None, None)
     }
 
@@ -493,15 +509,15 @@ impl RSA {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
-    /// let mut rsa = RSA::generate_ex(2048, 65537, &mut rng, None, None).expect("Error with generate_ex()");
+    /// let rng = RNG::new().expect("Error creating RNG");
+    /// let mut rsa = RSA::generate_ex(2048, 65537, &rng, None, None).expect("Error with generate_ex()");
     /// rsa.check().expect("Error with check()");
     /// let encrypt_size = rsa.get_encrypt_size().expect("Error with get_encrypt_size()");
     /// assert_eq!(encrypt_size, 256);
     /// }
     /// ```
     #[cfg(all(random, rsa_keygen))]
-    pub fn generate_ex(size: i32, e: i32, rng: &mut RNG, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
+    pub fn generate_ex(size: i32, e: i32, rng: &RNG, heap: Option<*mut core::ffi::c_void>, dev_id: Option<i32>) -> Result<Self, i32> {
         let mut wc_rsakey: MaybeUninit<sys::RsaKey> = MaybeUninit::uninit();
         let heap = match heap {
             Some(heap) => heap,
@@ -518,13 +534,17 @@ impl RSA {
         let mut wc_rsakey = unsafe { wc_rsakey.assume_init() };
         let e = e as core::ffi::c_long;
         let rc = unsafe {
-            sys::wc_MakeRsaKey(&mut wc_rsakey, size, e, &mut rng.wc_rng)
+            sys::wc_MakeRsaKey(&mut wc_rsakey, size, e, rng.wc_rng)
         };
         if rc != 0 {
             unsafe { sys::wc_FreeRsaKey(&mut wc_rsakey); }
             return Err(rc);
         }
-        let rsa = RSA { wc_rsakey };
+        let rsa = RSA {
+            wc_rsakey,
+            #[cfg(random)]
+            rng: None,
+        };
         Ok(rsa)
     }
 
@@ -556,8 +576,8 @@ impl RSA {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
-    /// let mut rsa = RSA::generate(2048, 65537, &mut rng).expect("Error with generate()");
+    /// let rng = RNG::new().expect("Error creating RNG");
+    /// let mut rsa = RSA::generate(2048, 65537, &rng).expect("Error with generate()");
     /// let mut e: [u8; 256] = [0; 256];
     /// let mut e_size: u32 = 0;
     /// let mut n: [u8; 256] = [0; 256];
@@ -624,8 +644,8 @@ impl RSA {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
-    /// let mut rsa = RSA::generate(2048, 65537, &mut rng).expect("Error with generate()");
+    /// let rng = RNG::new().expect("Error creating RNG");
+    /// let mut rsa = RSA::generate(2048, 65537, &rng).expect("Error with generate()");
     /// let mut e: [u8; 256] = [0; 256];
     /// let mut e_size: u32 = 0;
     /// let mut n: [u8; 256] = [0; 256];
@@ -667,8 +687,8 @@ impl RSA {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
-    /// let mut rsa = RSA::generate(2048, 65537, &mut rng).expect("Error with generate()");
+    /// let rng = RNG::new().expect("Error creating RNG");
+    /// let mut rsa = RSA::generate(2048, 65537, &rng).expect("Error with generate()");
     /// let encrypt_size = rsa.get_encrypt_size().expect("Error with get_encrypt_size()");
     /// assert_eq!(encrypt_size, 256);
     /// }
@@ -696,8 +716,8 @@ impl RSA {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
-    /// let mut rsa = RSA::generate(2048, 65537, &mut rng).expect("Error with generate()");
+    /// let rng = RNG::new().expect("Error creating RNG");
+    /// let mut rsa = RSA::generate(2048, 65537, &rng).expect("Error with generate()");
     /// rsa.check().expect("Error with check()");
     /// }
     /// ```
@@ -729,26 +749,26 @@ impl RSA {
     ///
     /// ```rust
     /// # extern crate std;
-    /// #[cfg(random)]
+    /// #[cfg(all(random, feature = "alloc"))]
     /// {
     /// use std::fs;
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
+    /// let rng = std::rc::Rc::new(RNG::new().expect("Error creating RNG"));
     /// let key_path = "../../../certs/client-keyPub.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_public_from_der(&der).expect("Error with new_public_from_der()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let plain: &[u8] = b"Test message";
     /// let mut enc: [u8; 512] = [0; 512];
-    /// let enc_len = rsa.public_encrypt(plain, &mut enc, &mut rng).expect("Error with public_encrypt()");
+    /// let enc_len = rsa.public_encrypt(plain, &mut enc, &rng).expect("Error with public_encrypt()");
     /// assert!(enc_len > 0 && enc_len <= 512);
     ///
     /// let key_path = "../../../certs/client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_from_der(&der).expect("Error with new_from_der()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let mut plain_out: [u8; 512] = [0; 512];
     /// let dec_len = rsa.private_decrypt(&enc[0..enc_len], &mut plain_out).expect("Error with private_decrypt()");
     /// assert!(dec_len as usize == plain.len());
@@ -756,13 +776,13 @@ impl RSA {
     /// }
     /// ```
     #[cfg(random)]
-    pub fn public_encrypt(&mut self, din: &[u8], dout: &mut [u8], rng: &mut RNG) -> Result<usize, i32> {
+    pub fn public_encrypt(&mut self, din: &[u8], dout: &mut [u8], rng: &RNG) -> Result<usize, i32> {
         let din_size = crate::buffer_len_to_u32(din.len())?;
         let dout_size = crate::buffer_len_to_u32(dout.len())?;
         let rc = unsafe {
             sys::wc_RsaPublicEncrypt(din.as_ptr(), din_size,
                 dout.as_mut_ptr(), dout_size, &mut self.wc_rsakey,
-                &mut rng.wc_rng)
+                rng.wc_rng)
         };
         if rc < 0 {
             return Err(rc);
@@ -788,26 +808,26 @@ impl RSA {
     ///
     /// ```rust
     /// # extern crate std;
-    /// #[cfg(random)]
+    /// #[cfg(all(random, feature = "alloc"))]
     /// {
     /// use std::fs;
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
+    /// let rng = std::rc::Rc::new(RNG::new().expect("Error creating RNG"));
     /// let key_path = "../../../certs/client-keyPub.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_public_from_der(&der).expect("Error with new_public_from_der()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let plain: &[u8] = b"Test message";
     /// let mut enc: [u8; 512] = [0; 512];
-    /// let enc_len = rsa.public_encrypt(plain, &mut enc, &mut rng).expect("Error with public_encrypt()");
+    /// let enc_len = rsa.public_encrypt(plain, &mut enc, &rng).expect("Error with public_encrypt()");
     /// assert!(enc_len > 0 && enc_len <= 512);
     ///
     /// let key_path = "../../../certs/client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_from_der(&der).expect("Error with new_from_der()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let mut plain_out: [u8; 512] = [0; 512];
     /// let dec_len = rsa.private_decrypt(&enc[0..enc_len], &mut plain_out).expect("Error with private_decrypt()");
     /// assert!(dec_len as usize == plain.len());
@@ -849,26 +869,26 @@ impl RSA {
     /// # Example
     ///
     /// ```rust
-    /// #[cfg(all(random, rsa_pss))]
+    /// #[cfg(all(random, rsa_pss, feature = "alloc"))]
     /// {
     /// use std::fs;
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
+    /// let rng = std::rc::Rc::new(RNG::new().expect("Error creating RNG"));
     ///
     /// let key_path = "../../../certs/client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_from_der(&der).expect("Error with new_from_der()");
     /// let msg: &[u8] = b"This is the string to be signed!";
     /// let mut signature: [u8; 512] = [0; 512];
-    /// let sig_len = rsa.pss_sign(msg, &mut signature, RSA::HASH_TYPE_SHA256, RSA::MGF1SHA256, &mut rng).expect("Error with pss_sign()");
+    /// let sig_len = rsa.pss_sign(msg, &mut signature, RSA::HASH_TYPE_SHA256, RSA::MGF1SHA256, &rng).expect("Error with pss_sign()");
     /// assert!(sig_len > 0 && sig_len <= 512);
     ///
     /// let key_path = "../../../certs/client-keyPub.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_public_from_der(&der).expect("Error with new_public_from_der()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let signature = &signature[0..sig_len];
     /// let mut verify_out: [u8; 512] = [0; 512];
     /// let verify_out_size = rsa.pss_verify(signature, &mut verify_out, RSA::HASH_TYPE_SHA256, RSA::MGF1SHA256).expect("Error with pss_verify()");
@@ -880,12 +900,12 @@ impl RSA {
     /// }
     /// ```
     #[cfg(all(random, rsa_pss))]
-    pub fn pss_sign(&mut self, din: &[u8], dout: &mut [u8], hash_algo: u32, mgf: i32, rng: &mut RNG) -> Result<usize, i32> {
+    pub fn pss_sign(&mut self, din: &[u8], dout: &mut [u8], hash_algo: u32, mgf: i32, rng: &RNG) -> Result<usize, i32> {
         let din_size = crate::buffer_len_to_u32(din.len())?;
         let dout_size = crate::buffer_len_to_u32(dout.len())?;
         let rc = unsafe {
             sys::wc_RsaPSS_Sign(din.as_ptr(), din_size, dout.as_mut_ptr(), dout_size,
-                hash_algo, mgf, &mut self.wc_rsakey, &mut rng.wc_rng)
+                hash_algo, mgf, &mut self.wc_rsakey, rng.wc_rng)
         };
         if rc < 0 {
             return Err(rc);
@@ -912,26 +932,26 @@ impl RSA {
     /// # Example
     ///
     /// ```rust
-    /// #[cfg(all(random, rsa_pss, rsa_const_api))]
+    /// #[cfg(all(random, rsa_pss, rsa_const_api, feature = "alloc"))]
     /// {
     /// use std::fs;
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
+    /// let rng = std::rc::Rc::new(RNG::new().expect("Error creating RNG"));
     ///
     /// let key_path = "../../../certs/client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_from_der(&der).expect("Error with new_from_der()");
     /// let msg: &[u8] = b"This is the string to be signed!";
     /// let mut signature: [u8; 512] = [0; 512];
-    /// let sig_len = rsa.pss_sign(msg, &mut signature, RSA::HASH_TYPE_SHA256, RSA::MGF1SHA256, &mut rng).expect("Error with pss_sign()");
+    /// let sig_len = rsa.pss_sign(msg, &mut signature, RSA::HASH_TYPE_SHA256, RSA::MGF1SHA256, &rng).expect("Error with pss_sign()");
     /// assert!(sig_len > 0 && sig_len <= 512);
     ///
     /// let key_path = "../../../certs/client-keyPub.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_public_from_der(&der).expect("Error with new_public_from_der()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let signature = &signature[0..sig_len];
     /// let mut verify_out: [u8; 512] = [0; 512];
     /// let verify_out_size = rsa.pss_verify(signature, &mut verify_out, RSA::HASH_TYPE_SHA256, RSA::MGF1SHA256).expect("Error with pss_verify()");
@@ -978,26 +998,26 @@ impl RSA {
     /// # Example
     ///
     /// ```rust
-    /// #[cfg(all(random, rsa_pss, rsa_const_api))]
+    /// #[cfg(all(random, rsa_pss, rsa_const_api, feature = "alloc"))]
     /// {
     /// use std::fs;
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
+    /// let rng = std::rc::Rc::new(RNG::new().expect("Error creating RNG"));
     ///
     /// let key_path = "../../../certs/client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_from_der(&der).expect("Error with new_from_der()");
     /// let msg: &[u8] = b"This is the string to be signed!";
     /// let mut signature: [u8; 512] = [0; 512];
-    /// let sig_len = rsa.pss_sign(msg, &mut signature, RSA::HASH_TYPE_SHA256, RSA::MGF1SHA256, &mut rng).expect("Error with pss_sign()");
+    /// let sig_len = rsa.pss_sign(msg, &mut signature, RSA::HASH_TYPE_SHA256, RSA::MGF1SHA256, &rng).expect("Error with pss_sign()");
     /// assert!(sig_len > 0 && sig_len <= 512);
     ///
     /// let key_path = "../../../certs/client-keyPub.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_public_from_der(&der).expect("Error with new_public_from_der()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let signature = &signature[0..sig_len];
     /// let mut verify_out: [u8; 512] = [0; 512];
     /// let verify_out_size = rsa.pss_verify(signature, &mut verify_out, RSA::HASH_TYPE_SHA256, RSA::MGF1SHA256).expect("Error with pss_verify()");
@@ -1049,26 +1069,26 @@ impl RSA {
     /// # Example
     ///
     /// ```rust
-    /// #[cfg(all(random, rsa_pss, rsa_const_api))]
+    /// #[cfg(all(random, rsa_pss, rsa_const_api, feature = "alloc"))]
     /// {
     /// use std::fs;
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
+    /// let rng = std::rc::Rc::new(RNG::new().expect("Error creating RNG"));
     ///
     /// let key_path = "../../../certs/client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_from_der(&der).expect("Error with new_from_der()");
     /// let msg: &[u8] = b"This is the string to be signed!";
     /// let mut signature: [u8; 512] = [0; 512];
-    /// let sig_len = rsa.pss_sign(msg, &mut signature, RSA::HASH_TYPE_SHA256, RSA::MGF1SHA256, &mut rng).expect("Error with pss_sign()");
+    /// let sig_len = rsa.pss_sign(msg, &mut signature, RSA::HASH_TYPE_SHA256, RSA::MGF1SHA256, &rng).expect("Error with pss_sign()");
     /// assert!(sig_len > 0 && sig_len <= 512);
     ///
     /// let key_path = "../../../certs/client-keyPub.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_public_from_der(&der).expect("Error with new_public_from_der()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let signature = &signature[0..sig_len];
     /// let mut verify_out: [u8; 512] = [0; 512];
     /// let verify_out_size = rsa.pss_verify(signature, &mut verify_out, RSA::HASH_TYPE_SHA256, RSA::MGF1SHA256).expect("Error with pss_verify()");
@@ -1125,7 +1145,7 @@ impl RSA {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
+    /// let rng = RNG::new().expect("Error creating RNG");
     ///
     /// let key_path = "../../../certs/client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
@@ -1134,22 +1154,22 @@ impl RSA {
     /// let mut plain = [0u8; 256];
     /// plain[..msg.len()].copy_from_slice(msg);
     /// let mut enc = [0u8; 256];
-    /// let enc_len = rsa.rsa_direct(&plain, &mut enc, RSA::PRIVATE_ENCRYPT, &mut rng).expect("Error with rsa_direct()");
+    /// let enc_len = rsa.rsa_direct(&plain, &mut enc, RSA::PRIVATE_ENCRYPT, &rng).expect("Error with rsa_direct()");
     /// assert_eq!(enc_len, 256);
     /// let mut plain_out = [0u8; 256];
-    /// let dec_len = rsa.rsa_direct(&enc, &mut plain_out, RSA::PUBLIC_DECRYPT, &mut rng).expect("Error with rsa_direct()");
+    /// let dec_len = rsa.rsa_direct(&enc, &mut plain_out, RSA::PUBLIC_DECRYPT, &rng).expect("Error with rsa_direct()");
     /// assert_eq!(dec_len, 256);
     /// assert_eq!(plain_out, plain);
     /// }
     /// ```
     #[cfg(all(rsa_direct, rsa_const_api))]
-    pub fn rsa_direct(&mut self, din: &[u8], dout: &mut [u8], typ: i32, rng: &mut RNG) -> Result<usize, i32> {
+    pub fn rsa_direct(&mut self, din: &[u8], dout: &mut [u8], typ: i32, rng: &RNG) -> Result<usize, i32> {
         let din_size = crate::buffer_len_to_u32(din.len())?;
         let mut dout_size = crate::buffer_len_to_u32(dout.len())?;
         let rc = unsafe {
             sys::wc_RsaDirect(din.as_ptr(), din_size,
                 dout.as_mut_ptr(), &mut dout_size,
-                &mut self.wc_rsakey, typ, &mut rng.wc_rng)
+                &mut self.wc_rsakey, typ, rng.wc_rng)
         };
         if rc < 0 {
             return Err(rc);
@@ -1165,8 +1185,7 @@ impl RSA {
     /// # Parameters
     ///
     /// * `rng`: The `RNG` struct instance to associate with this `RSA`
-    ///   instance. The `RNG` struct should not be moved in memory after
-    ///   calling this method.
+    ///   instance.
     ///
     /// # Returns
     ///
@@ -1183,19 +1202,19 @@ impl RSA {
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
     /// let key_path = "../../../certs/client-keyPub.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_public_from_der(&der).expect("Error with new_public_from_der()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_rng(RNG::new().expect("Error creating RNG")).expect("Error with set_rng()");
     /// let plain: &[u8] = b"Test message";
     /// let mut enc: [u8; 512] = [0; 512];
-    /// let enc_len = rsa.public_encrypt(plain, &mut enc, &mut rng).expect("Error with public_encrypt()");
+    /// let enc_rng = RNG::new().expect("Error creating RNG");
+    /// let enc_len = rsa.public_encrypt(plain, &mut enc, &enc_rng).expect("Error with public_encrypt()");
     /// assert!(enc_len > 0 && enc_len <= 512);
     /// let key_path = "../../../certs/client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_from_der(&der).expect("Error with new_from_der()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_rng(RNG::new().expect("Error creating RNG")).expect("Error with set_rng()");
     /// let mut plain_out: [u8; 512] = [0; 512];
     /// let dec_len = rsa.private_decrypt(&enc[0..enc_len], &mut plain_out).expect("Error with private_decrypt()");
     /// assert!(dec_len as usize == plain.len());
@@ -1203,16 +1222,47 @@ impl RSA {
     /// }
     /// ```
     #[cfg(random)]
-    pub fn set_rng(&mut self, rng: &mut RNG) -> Result<(), i32> {
+    pub fn set_rng(&mut self, rng: RNG) -> Result<(), i32> {
+        let wc_rng = rng.wc_rng;
         let rc = unsafe {
-            sys::wc_RsaSetRNG(&mut self.wc_rsakey, &mut rng.wc_rng)
+            sys::wc_RsaSetRNG(&mut self.wc_rsakey, wc_rng)
         };
         if rc != 0 {
             return Err(rc);
         }
+        self.rng = Some(RngHandle::Owned(rng));
         Ok(())
     }
 
+    /// Bind a shared `RNG` to this key for blinding during private operations.
+    ///
+    /// Like `set_rng`, but takes an `Rc<RNG>` so the same RNG can be shared
+    /// among multiple consumers and used directly by the caller. Available
+    /// when the `alloc` feature is enabled.
+    #[cfg(all(random, feature = "alloc"))]
+    pub fn set_shared_rng(&mut self, rng: alloc::rc::Rc<RNG>) -> Result<(), i32> {
+        let wc_rng = rng.wc_rng;
+        let rc = unsafe {
+            sys::wc_RsaSetRNG(&mut self.wc_rsakey, wc_rng)
+        };
+        if rc != 0 {
+            return Err(rc);
+        }
+        self.rng = Some(RngHandle::Shared(rng));
+        Ok(())
+    }
+
+    /// Borrow the RNG previously bound via `set_rng` or `set_shared_rng`.
+    #[cfg(random)]
+    pub fn rng(&self) -> Option<&RNG> {
+        match &self.rng {
+            Some(RngHandle::Owned(rng)) => Some(rng),
+            #[cfg(feature = "alloc")]
+            Some(RngHandle::Shared(rng)) => Some(rng),
+            None => None,
+        }
+    }
+
     /// Sign the provided data with the private key.
     ///
     /// # Parameters
@@ -1233,26 +1283,26 @@ impl RSA {
     ///
     /// ```rust
     /// # extern crate std;
-    /// #[cfg(random)]
+    /// #[cfg(all(random, feature = "alloc"))]
     /// {
     /// use std::fs;
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
+    /// let rng = std::rc::Rc::new(RNG::new().expect("Error creating RNG"));
     ///
     /// let key_path = "../../../certs/client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_from_der(&der).expect("Error with new_from_der()");
     /// let msg: &[u8] = b"This is the string to be signed!";
     /// let mut signature: [u8; 512] = [0; 512];
-    /// let sig_len = rsa.ssl_sign(msg, &mut signature, &mut rng).expect("Error with ssl_sign()");
+    /// let sig_len = rsa.ssl_sign(msg, &mut signature, &rng).expect("Error with ssl_sign()");
     /// assert!(sig_len > 0 && sig_len <= 512);
     ///
     /// let key_path = "../../../certs/client-keyPub.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_public_from_der(&der).expect("Error with new_public_from_der()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let signature = &signature[0..sig_len];
     /// let mut verify_out: [u8; 512] = [0; 512];
     /// let verify_out_size = rsa.ssl_verify(signature, &mut verify_out).expect("Error with ssl_verify()");
@@ -1260,13 +1310,13 @@ impl RSA {
     /// }
     /// ```
     #[cfg(random)]
-    pub fn ssl_sign(&mut self, din: &[u8], dout: &mut [u8], rng: &mut RNG) -> Result<usize, i32> {
+    pub fn ssl_sign(&mut self, din: &[u8], dout: &mut [u8], rng: &RNG) -> Result<usize, i32> {
         let din_size = crate::buffer_len_to_u32(din.len())?;
         let dout_size = crate::buffer_len_to_u32(dout.len())?;
         let rc = unsafe {
             sys::wc_RsaSSL_Sign(din.as_ptr(), din_size,
                 dout.as_mut_ptr(), dout_size,
-                &mut self.wc_rsakey, &mut rng.wc_rng)
+                &mut self.wc_rsakey, rng.wc_rng)
         };
         if rc < 0 {
             return Err(rc);
@@ -1295,26 +1345,26 @@ impl RSA {
     ///
     /// ```rust
     /// # extern crate std;
-    /// #[cfg(random)]
+    /// #[cfg(all(random, feature = "alloc"))]
     /// {
     /// use std::fs;
     /// use wolfssl_wolfcrypt::random::RNG;
     /// use wolfssl_wolfcrypt::rsa::RSA;
     ///
-    /// let mut rng = RNG::new().expect("Error creating RNG");
+    /// let rng = std::rc::Rc::new(RNG::new().expect("Error creating RNG"));
     ///
     /// let key_path = "../../../certs/client-key.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_from_der(&der).expect("Error with new_from_der()");
     /// let msg: &[u8] = b"This is the string to be signed!";
     /// let mut signature: [u8; 512] = [0; 512];
-    /// let sig_len = rsa.ssl_sign(msg, &mut signature, &mut rng).expect("Error with ssl_sign()");
+    /// let sig_len = rsa.ssl_sign(msg, &mut signature, &rng).expect("Error with ssl_sign()");
     /// assert!(sig_len > 0 && sig_len <= 512);
     ///
     /// let key_path = "../../../certs/client-keyPub.der";
     /// let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     /// let mut rsa = RSA::new_public_from_der(&der).expect("Error with new_public_from_der()");
-    /// rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    /// rsa.set_shared_rng(std::rc::Rc::clone(&rng)).expect("Error with set_shared_rng()");
     /// let signature = &signature[0..sig_len];
     /// let mut verify_out: [u8; 512] = [0; 512];
     /// let verify_out_size = rsa.ssl_verify(signature, &mut verify_out).expect("Error with ssl_verify()");
diff --git a/wrapper/rust/wolfssl-wolfcrypt/src/rsa_pkcs1v15.rs b/wrapper/rust/wolfssl-wolfcrypt/src/rsa_pkcs1v15.rs
index 40314f9548..576f63e222 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/src/rsa_pkcs1v15.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/src/rsa_pkcs1v15.rs
@@ -149,9 +149,9 @@ pub struct SigningKey<H: Hash, const N: usize> {
 impl<H: Hash, const N: usize> SigningKey<H, N> {
     /// Generate a fresh `N * 8`-bit RSA key with public exponent 65537.
     #[cfg(rsa_keygen)]
-    pub fn generate(mut rng: RNG) -> Result<Self, i32> {
+    pub fn generate(rng: RNG) -> Result<Self, i32> {
         let bits: i32 = (N * 8).try_into().map_err(|_| sys::wolfCrypt_ErrorCodes_BAD_FUNC_ARG)?;
-        let rsa = RSA::generate(bits, 65537, &mut rng)?;
+        let rsa = RSA::generate(bits, 65537, &rng)?;
         Ok(Self { inner: rsa, rng, _hash: PhantomData })
     }
 
@@ -186,7 +186,7 @@ impl<H: Hash, const N: usize> SignerMut<Signature<N>> for SigningKey<H, N> {
                 sig.as_mut_ptr(), &mut sig_len,
                 &mut self.inner.wc_rsakey as *mut _ as *mut c_void,
                 size_of::<sys::RsaKey>() as u32,
-                &mut self.rng.wc_rng,
+                self.rng.wc_rng,
             )
         };
         if rc != 0 || sig_len as usize != N {
diff --git a/wrapper/rust/wolfssl-wolfcrypt/tests/test_aes.rs b/wrapper/rust/wolfssl-wolfcrypt/tests/test_aes.rs
index e215891691..d5a171fd9e 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/tests/test_aes.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/tests/test_aes.rs
@@ -1030,6 +1030,29 @@ fn test_aes256gcm_nist_tc14_encrypt() {
     assert_eq!(&tag[..], &expected_tag);
 }
 
+/// Roundtrip test for AES-192-GCM using `aead::Aead`.
+#[test]
+#[cfg(all(feature = "aead", aes_gcm))]
+fn test_aes192gcm_aead_roundtrip() {
+    let key = [0x77u8; 24];
+    let nonce_bytes = [0x88u8; 12];
+    let aad = b"aes-192-gcm test";
+    let plaintext = b"AES-192-GCM roundtrip test";
+
+    let cipher = Aes192Gcm::new_from_slice(&key).unwrap();
+    let nonce: aead::Nonce<Aes192Gcm> = nonce_bytes.into();
+
+    let ciphertext = cipher
+        .encrypt(&nonce, Payload { msg: plaintext, aad })
+        .expect("AES-192-GCM encrypt failed");
+
+    let recovered = cipher
+        .decrypt(&nonce, Payload { msg: &ciphertext, aad })
+        .expect("AES-192-GCM decrypt failed");
+
+    assert_eq!(recovered, plaintext);
+}
+
 /// Roundtrip test for AES-256-GCM using `aead::Aead`.
 #[test]
 #[cfg(all(feature = "aead", aes_gcm))]
@@ -1092,6 +1115,29 @@ fn test_aes128ccm_reject_tampered() {
     assert!(cipher.decrypt(&nonce, ct.as_slice()).is_err());
 }
 
+/// Roundtrip test for AES-192-CCM using `aead::Aead`.
+#[test]
+#[cfg(all(feature = "aead", aes_ccm))]
+fn test_aes192ccm_aead_roundtrip() {
+    let key = [0x33u8; 24];
+    let nonce_bytes = [0x44u8; 12];
+    let aad = b"aes-192-ccm test";
+    let plaintext = b"AES-192-CCM plaintext data";
+
+    let cipher = Aes192Ccm::new_from_slice(&key).unwrap();
+    let nonce: aead::Nonce<Aes192Ccm> = nonce_bytes.into();
+
+    let ciphertext = cipher
+        .encrypt(&nonce, Payload { msg: plaintext, aad })
+        .expect("AES-192-CCM encrypt failed");
+
+    let recovered = cipher
+        .decrypt(&nonce, Payload { msg: &ciphertext, aad })
+        .expect("AES-192-CCM decrypt failed");
+
+    assert_eq!(recovered, plaintext);
+}
+
 /// Roundtrip test for AES-256-CCM using `aead::Aead`.
 #[test]
 #[cfg(all(feature = "aead", aes_ccm))]
diff --git a/wrapper/rust/wolfssl-wolfcrypt/tests/test_blake2_digest.rs b/wrapper/rust/wolfssl-wolfcrypt/tests/test_blake2_digest.rs
new file mode 100644
index 0000000000..49d28c6fac
--- /dev/null
+++ b/wrapper/rust/wolfssl-wolfcrypt/tests/test_blake2_digest.rs
@@ -0,0 +1,126 @@
+#![cfg(all(any(blake2b, blake2s), feature = "digest"))]
+
+use digest::{Digest, FixedOutputReset};
+use digest::block_api::BlockSizeUser;
+
+mod common;
+
+fn check_digest<D: Digest + BlockSizeUser + FixedOutputReset + Default>(
+    input: &[u8],
+    expected: &[u8],
+    expected_block_size: usize,
+) {
+    assert_eq!(<D as Digest>::output_size(), expected.len());
+    assert_eq!(<D as BlockSizeUser>::block_size(), expected_block_size);
+
+    /* One-shot digest via associated function. */
+    let out = D::digest(input);
+    assert_eq!(out.as_slice(), expected);
+
+    /* Streaming via Digest::update and finalize. */
+    let mut hasher = D::new();
+    Digest::update(&mut hasher, input);
+    let out = hasher.finalize();
+    assert_eq!(out.as_slice(), expected);
+
+    /* Split update via Default + Update + FixedOutputReset::finalize_reset. */
+    let mut hasher = D::default();
+    if input.len() >= 2 {
+        let mid = input.len() / 2;
+        Digest::update(&mut hasher, &input[..mid]);
+        Digest::update(&mut hasher, &input[mid..]);
+    } else {
+        Digest::update(&mut hasher, input);
+    }
+    let out = hasher.finalize_reset();
+    assert_eq!(out.as_slice(), expected);
+
+    /* After reset, the same hasher should produce the same result. */
+    Digest::update(&mut hasher, input);
+    let out = hasher.finalize();
+    assert_eq!(out.as_slice(), expected);
+}
+
+#[test]
+#[cfg(blake2b)]
+fn test_digest_blake2b_512() {
+    use wolfssl_wolfcrypt::blake2_digest::Blake2b512;
+    common::setup();
+    check_digest::<Blake2b512>(
+        b"abc",
+        b"\xBA\x80\xA5\x3F\x98\x1C\x4D\x0D\x6A\x27\x97\xB6\x9F\x12\xF6\xE9\x4C\x21\x2F\x14\x68\x5A\xC4\xB7\x4B\x12\xBB\x6F\xDB\xFF\xA2\xD1\x7D\x87\xC5\x39\x2A\xAB\x79\x2D\xC2\x52\xD5\xDE\x45\x33\xCC\x95\x18\xD3\x8A\xA8\xDB\xF1\x92\x5A\xB9\x23\x86\xED\xD4\x00\x99\x23",
+        128,
+    );
+}
+
+#[test]
+#[cfg(blake2b)]
+fn test_digest_blake2b_384() {
+    use wolfssl_wolfcrypt::blake2::BLAKE2b;
+    use wolfssl_wolfcrypt::blake2_digest::Blake2b384;
+    common::setup();
+
+    let mut reference = BLAKE2b::new(48).expect("Error with new()");
+    reference.update(b"abc").expect("Error with update()");
+    let mut expected = [0u8; 48];
+    reference.finalize(&mut expected).expect("Error with finalize()");
+
+    check_digest::<Blake2b384>(b"abc", &expected, 128);
+}
+
+#[test]
+#[cfg(blake2b)]
+fn test_digest_blake2b_256() {
+    use wolfssl_wolfcrypt::blake2::BLAKE2b;
+    use wolfssl_wolfcrypt::blake2_digest::Blake2b256;
+    common::setup();
+
+    let mut reference = BLAKE2b::new(32).expect("Error with new()");
+    reference.update(b"abc").expect("Error with update()");
+    let mut expected = [0u8; 32];
+    reference.finalize(&mut expected).expect("Error with finalize()");
+
+    check_digest::<Blake2b256>(b"abc", &expected, 128);
+}
+
+#[test]
+#[cfg(blake2s)]
+fn test_digest_blake2s_256() {
+    use wolfssl_wolfcrypt::blake2_digest::Blake2s256;
+    common::setup();
+    check_digest::<Blake2s256>(
+        b"abc",
+        b"\x50\x8C\x5E\x8C\x32\x7C\x14\xE2\xE1\xA7\x2B\xA3\x4E\xEB\x45\x2F\x37\x45\x8B\x20\x9E\xD6\x3A\x29\x4D\x99\x9B\x4C\x86\x67\x59\x82",
+        64,
+    );
+}
+
+#[test]
+#[cfg(blake2s)]
+fn test_digest_blake2s_192() {
+    use wolfssl_wolfcrypt::blake2::BLAKE2s;
+    use wolfssl_wolfcrypt::blake2_digest::Blake2s192;
+    common::setup();
+
+    let mut reference = BLAKE2s::new(24).expect("Error with new()");
+    reference.update(b"abc").expect("Error with update()");
+    let mut expected = [0u8; 24];
+    reference.finalize(&mut expected).expect("Error with finalize()");
+
+    check_digest::<Blake2s192>(b"abc", &expected, 64);
+}
+
+#[test]
+#[cfg(blake2s)]
+fn test_digest_blake2s_128() {
+    use wolfssl_wolfcrypt::blake2::BLAKE2s;
+    use wolfssl_wolfcrypt::blake2_digest::Blake2s128;
+    common::setup();
+
+    let mut reference = BLAKE2s::new(16).expect("Error with new()");
+    reference.update(b"abc").expect("Error with update()");
+    let mut expected = [0u8; 16];
+    reference.finalize(&mut expected).expect("Error with finalize()");
+
+    check_digest::<Blake2s128>(b"abc", &expected, 64);
+}
diff --git a/wrapper/rust/wolfssl-wolfcrypt/tests/test_blake2_mac.rs b/wrapper/rust/wolfssl-wolfcrypt/tests/test_blake2_mac.rs
new file mode 100644
index 0000000000..a25761a2bd
--- /dev/null
+++ b/wrapper/rust/wolfssl-wolfcrypt/tests/test_blake2_mac.rs
@@ -0,0 +1,181 @@
+#![cfg(all(any(blake2b, blake2s), feature = "mac"))]
+
+use digest::{KeyInit, Mac};
+
+#[test]
+#[cfg(blake2b)]
+fn test_blake2b_mac_512() {
+    use wolfssl_wolfcrypt::blake2::BLAKE2b;
+    use wolfssl_wolfcrypt::blake2_mac::Blake2bMac512;
+
+    let key = [0x42u8; 64];
+    let input = b"The quick brown fox jumps over the lazy dog";
+
+    let mut reference = BLAKE2b::new_with_key(64, &key)
+        .expect("Error with new_with_key()");
+    reference.update(input).expect("Error with update()");
+    let mut expected = [0u8; 64];
+    reference.finalize(&mut expected)
+        .expect("Error with finalize()");
+
+    let mut mac = Blake2bMac512::new_from_slice(&key)
+        .expect("Blake2bMac512 init failed");
+    mac.update(input);
+    let tag = mac.finalize();
+    assert_eq!(tag.into_bytes().as_slice(), &expected);
+}
+
+#[test]
+#[cfg(blake2b)]
+fn test_blake2b_mac_256() {
+    use wolfssl_wolfcrypt::blake2::BLAKE2b;
+    use wolfssl_wolfcrypt::blake2_mac::Blake2bMac256;
+
+    let key = [0x33u8; 64];
+    let input = b"libsodium crypto_generichash analog";
+
+    let mut reference = BLAKE2b::new_with_key(32, &key)
+        .expect("Error with new_with_key()");
+    reference.update(input).expect("Error with update()");
+    let mut expected = [0u8; 32];
+    reference.finalize(&mut expected)
+        .expect("Error with finalize()");
+
+    let mut mac = Blake2bMac256::new_from_slice(&key)
+        .expect("Blake2bMac256 init failed");
+    mac.update(input);
+    let tag = mac.finalize();
+    assert_eq!(tag.into_bytes().as_slice(), &expected);
+}
+
+#[test]
+#[cfg(blake2b)]
+fn test_blake2b_mac_384() {
+    use wolfssl_wolfcrypt::blake2::BLAKE2b;
+    use wolfssl_wolfcrypt::blake2_mac::Blake2bMac384;
+
+    let key = [0x77u8; 64];
+    let input = b"sha-384 sized blake2b mac";
+
+    let mut reference = BLAKE2b::new_with_key(48, &key)
+        .expect("Error with new_with_key()");
+    reference.update(input).expect("Error with update()");
+    let mut expected = [0u8; 48];
+    reference.finalize(&mut expected)
+        .expect("Error with finalize()");
+
+    let mut mac = Blake2bMac384::new_from_slice(&key)
+        .expect("Blake2bMac384 init failed");
+    mac.update(input);
+    let tag = mac.finalize();
+    assert_eq!(tag.into_bytes().as_slice(), &expected);
+}
+
+#[test]
+#[cfg(blake2b)]
+fn test_blake2b_mac_512_chunked() {
+    use wolfssl_wolfcrypt::blake2::BLAKE2b;
+    use wolfssl_wolfcrypt::blake2_mac::Blake2bMac512;
+
+    let key = [0xA5u8; 64];
+    let input: Vec<u8> = (0u8..200).collect();
+
+    let mut reference = BLAKE2b::new_with_key(64, &key)
+        .expect("Error with new_with_key()");
+    reference.update(&input).expect("Error with update()");
+    let mut expected = [0u8; 64];
+    reference.finalize(&mut expected)
+        .expect("Error with finalize()");
+
+    let mut mac = Blake2bMac512::new_from_slice(&key)
+        .expect("Blake2bMac512 init failed");
+    for chunk in input.chunks(17) {
+        mac.update(chunk);
+    }
+    mac.verify_slice(&expected)
+        .expect("Blake2bMac512 verify failed");
+}
+
+#[test]
+#[cfg(blake2s)]
+fn test_blake2s_mac_128() {
+    use wolfssl_wolfcrypt::blake2::BLAKE2s;
+    use wolfssl_wolfcrypt::blake2_mac::Blake2sMac128;
+
+    let key = [0x55u8; 32];
+    let input = b"short blake2s mac";
+
+    let mut reference = BLAKE2s::new_with_key(16, &key)
+        .expect("Error with new_with_key()");
+    reference.update(input).expect("Error with update()");
+    let mut expected = [0u8; 16];
+    reference.finalize(&mut expected)
+        .expect("Error with finalize()");
+
+    let mut mac = Blake2sMac128::new_from_slice(&key)
+        .expect("Blake2sMac128 init failed");
+    mac.update(input);
+    let tag = mac.finalize();
+    assert_eq!(tag.into_bytes().as_slice(), &expected);
+}
+
+#[test]
+#[cfg(blake2s)]
+fn test_blake2s_mac_192() {
+    use wolfssl_wolfcrypt::blake2::BLAKE2s;
+    use wolfssl_wolfcrypt::blake2_mac::Blake2sMac192;
+
+    let key = [0x99u8; 32];
+    let input = b"medium blake2s mac";
+
+    let mut reference = BLAKE2s::new_with_key(24, &key)
+        .expect("Error with new_with_key()");
+    reference.update(input).expect("Error with update()");
+    let mut expected = [0u8; 24];
+    reference.finalize(&mut expected)
+        .expect("Error with finalize()");
+
+    let mut mac = Blake2sMac192::new_from_slice(&key)
+        .expect("Blake2sMac192 init failed");
+    mac.update(input);
+    let tag = mac.finalize();
+    assert_eq!(tag.into_bytes().as_slice(), &expected);
+}
+
+#[test]
+#[cfg(blake2s)]
+fn test_blake2s_mac_256() {
+    use wolfssl_wolfcrypt::blake2::BLAKE2s;
+    use wolfssl_wolfcrypt::blake2_mac::Blake2sMac256;
+
+    let key = [0x42u8; 32];
+    let input = b"The quick brown fox jumps over the lazy dog";
+
+    let mut reference = BLAKE2s::new_with_key(32, &key)
+        .expect("Error with new_with_key()");
+    reference.update(input).expect("Error with update()");
+    let mut expected = [0u8; 32];
+    reference.finalize(&mut expected)
+        .expect("Error with finalize()");
+
+    let mut mac = Blake2sMac256::new_from_slice(&key)
+        .expect("Blake2sMac256 init failed");
+    mac.update(input);
+    let tag = mac.finalize();
+    assert_eq!(tag.into_bytes().as_slice(), &expected);
+}
+
+#[test]
+#[cfg(blake2s)]
+fn test_blake2s_mac_256_verify_fail() {
+    use wolfssl_wolfcrypt::blake2_mac::Blake2sMac256;
+
+    let key = [0x0Bu8; 32];
+    let input = b"hello";
+    let wrong_tag = [0u8; 32];
+
+    let mut mac = Blake2sMac256::new_from_slice(&key)
+        .expect("Blake2sMac256 init failed");
+    mac.update(input);
+    assert!(mac.verify_slice(&wrong_tag).is_err());
+}
diff --git a/wrapper/rust/wolfssl-wolfcrypt/tests/test_curve25519.rs b/wrapper/rust/wolfssl-wolfcrypt/tests/test_curve25519.rs
index ce1e3cc39d..ae4881dd1a 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/tests/test_curve25519.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/tests/test_curve25519.rs
@@ -1,5 +1,7 @@
 #![cfg(all(curve25519, random))]
 
+#[cfg(curve25519_blinding)]
+use std::rc::Rc;
 use wolfssl_wolfcrypt::curve25519::*;
 use wolfssl_wolfcrypt::random::RNG;
 
@@ -97,14 +99,17 @@ fn test_make_pub_blind() {
 
 #[test]
 fn test_shared_secret() {
-    let mut rng = RNG::new().expect("Error with new()");
-    let mut key1 = Curve25519Key::generate(&mut rng).expect("Error with generate()");
-    let mut key2 = Curve25519Key::generate(&mut rng).expect("Error with generate()");
+    #[cfg(curve25519_blinding)]
+    let rng = Rc::new(RNG::new().expect("Error with new()"));
+    #[cfg(not(curve25519_blinding))]
+    let rng = RNG::new().expect("Error with new()");
+    let mut key1 = Curve25519Key::generate(&rng).expect("Error with generate()");
+    let mut key2 = Curve25519Key::generate(&rng).expect("Error with generate()");
 
     #[cfg(curve25519_blinding)]
-    key1.set_rng(&mut rng).expect("Error with set_rng()");
+    key1.set_shared_rng(Rc::clone(&rng)).expect("Error with set_shared_rng()");
     #[cfg(curve25519_blinding)]
-    key2.set_rng(&mut rng).expect("Error with set_rng()");
+    key2.set_shared_rng(Rc::clone(&rng)).expect("Error with set_shared_rng()");
 
     let mut public_buffer = [0u8; Curve25519Key::KEYSIZE];
     key1.export_public(&mut public_buffer).expect("Error with export_public()");
@@ -122,14 +127,17 @@ fn test_shared_secret() {
 
 #[test]
 fn test_shared_secret_ex() {
-    let mut rng = RNG::new().expect("Error with new()");
-    let mut key1 = Curve25519Key::generate(&mut rng).expect("Error with generate()");
-    let mut key2 = Curve25519Key::generate(&mut rng).expect("Error with generate()");
+    #[cfg(curve25519_blinding)]
+    let rng = Rc::new(RNG::new().expect("Error with new()"));
+    #[cfg(not(curve25519_blinding))]
+    let rng = RNG::new().expect("Error with new()");
+    let mut key1 = Curve25519Key::generate(&rng).expect("Error with generate()");
+    let mut key2 = Curve25519Key::generate(&rng).expect("Error with generate()");
 
     #[cfg(curve25519_blinding)]
-    key1.set_rng(&mut rng).expect("Error with set_rng()");
+    key1.set_shared_rng(Rc::clone(&rng)).expect("Error with set_shared_rng()");
     #[cfg(curve25519_blinding)]
-    key2.set_rng(&mut rng).expect("Error with set_rng()");
+    key2.set_shared_rng(Rc::clone(&rng)).expect("Error with set_shared_rng()");
 
     let mut public_buffer = [0u8; Curve25519Key::KEYSIZE];
     key1.export_public(&mut public_buffer).expect("Error with export_public()");
diff --git a/wrapper/rust/wolfssl-wolfcrypt/tests/test_ecc.rs b/wrapper/rust/wolfssl-wolfcrypt/tests/test_ecc.rs
index 30d2d5b3fc..8bc3d369c1 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/tests/test_ecc.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/tests/test_ecc.rs
@@ -4,9 +4,13 @@ mod common;
 
 #[cfg(any(all(ecc_import, ecc_export, ecc_sign, ecc_verify, random), random))]
 use std::fs;
+#[cfg(all(ecc_dh, random))]
+use std::rc::Rc;
 use wolfssl_wolfcrypt::ecc::*;
 #[cfg(random)]
 use wolfssl_wolfcrypt::random::RNG;
+#[cfg(ecc_import)]
+use wolfssl_wolfcrypt::sys;
 
 #[test]
 #[cfg(random)]
@@ -134,7 +138,7 @@ fn test_ecc_import_export_sign_verify() {
     let valid = ecc.verify_hash(&signature, &hash).expect("Error with verify_hash()");
     assert_eq!(valid, false);
 
-    ecc.set_rng(&mut rng).expect("Error with set_rng()");
+    ecc.set_rng(rng).expect("Error with set_rng()");
 }
 
 #[test]
@@ -142,13 +146,13 @@ fn test_ecc_import_export_sign_verify() {
 fn test_ecc_shared_secret() {
     common::setup();
 
-    let mut rng = RNG::new().expect("Failed to create RNG");
-    let mut ecc0 = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
-    let mut ecc1 = ECC::generate(32, &mut rng, None, None).expect("Error with generate()");
+    let rng = Rc::new(RNG::new().expect("Failed to create RNG"));
+    let mut ecc0 = ECC::generate(32, &rng, None, None).expect("Error with generate()");
+    let mut ecc1 = ECC::generate(32, &rng, None, None).expect("Error with generate()");
     let mut ss0 = [0u8; 128];
     let mut ss1 = [0u8; 128];
-    ecc0.set_rng(&mut rng).expect("Error with set_rng()");
-    ecc1.set_rng(&mut rng).expect("Error with set_rng()");
+    ecc0.set_shared_rng(Rc::clone(&rng)).expect("Error with set_shared_rng()");
+    ecc1.set_shared_rng(Rc::clone(&rng)).expect("Error with set_shared_rng()");
     let ss0_size = ecc0.shared_secret(&mut ecc1, &mut ss0).expect("Error with shared_secret()");
     let ss1_size = ecc1.shared_secret(&mut ecc0, &mut ss1).expect("Error with shared_secret()");
     assert_eq!(ss0_size, ss1_size);
@@ -290,6 +294,33 @@ fn test_ecc_import_unsigned() {
     assert_eq!(valid, true);
 }
 
+#[test]
+#[cfg(ecc_import)]
+fn test_ecc_import_unsigned_short_slices() {
+    common::setup();
+
+    let curve_id = ECC::SECP256R1;
+    let qx = [0u8; 32];
+    let qy = [0u8; 32];
+    let d = [0u8; 32];
+    let empty: [u8; 0] = [];
+
+    let cases: [(&[u8], &[u8], &[u8]); 6] = [
+        (&qx[..31], &qy,        &d       ),
+        (&qx,       &qy[..31],  &d       ),
+        (&qx,       &qy,        &d[..31] ),
+        (&empty,    &qy,        &d       ),
+        (&qx,       &empty,     &d       ),
+        (&qx,       &qy,        &empty   ),
+    ];
+    for (qx, qy, d) in cases {
+        match ECC::import_unsigned(qx, qy, d, curve_id, None, None) {
+            Ok(_) => panic!("import_unsigned() should fail with short slice"),
+            Err(rc) => assert_eq!(rc, sys::wolfCrypt_ErrorCodes_BAD_FUNC_ARG),
+        }
+    }
+}
+
 #[test]
 #[cfg(random)]
 fn test_ecc_make_pub() {
@@ -365,6 +396,38 @@ fn test_ecc_import() {
     ECC::import_raw_ex(qx, qy, d, ECC::SECP256R1, None, None).expect("Error with import_raw_ex()");
 }
 
+#[test]
+#[cfg(ecc_import)]
+fn test_ecc_import_raw_not_null_terminated() {
+    common::setup();
+
+    let qx = b"7a4e287890a1a47ad3457e52f2f76a83ce46cbc947616d0cbaa82323818a793d\0";
+    let qy = b"eec4084f5b29ebf29c44cce3b3059610922f8b30ea6e8811742ac7238fe87308\0";
+    let d  = b"8c14b793cb19137e323a6d2e2a870bca2e7a493ec1153b3a95feb8a4873f8d08\0";
+    let qx_no_nul: &[u8] = &qx[..qx.len() - 1];
+    let qy_no_nul: &[u8] = &qy[..qy.len() - 1];
+    let d_no_nul:  &[u8] = &d[..d.len() - 1];
+    let curve_name = b"SECP256R1\0";
+    let curve_name_no_nul: &[u8] = b"SECP256R1";
+    let empty: &[u8] = b"";
+
+    assert!(ECC::import_raw(qx_no_nul, qy, d, curve_name, None, None).is_err());
+    assert!(ECC::import_raw(qx, qy_no_nul, d, curve_name, None, None).is_err());
+    assert!(ECC::import_raw(qx, qy, d_no_nul, curve_name, None, None).is_err());
+    assert!(ECC::import_raw(qx, qy, d, curve_name_no_nul, None, None).is_err());
+    assert!(ECC::import_raw(empty, qy, d, curve_name, None, None).is_err());
+    assert!(ECC::import_raw(qx, empty, d, curve_name, None, None).is_err());
+    assert!(ECC::import_raw(qx, qy, empty, curve_name, None, None).is_err());
+    assert!(ECC::import_raw(qx, qy, d, empty, None, None).is_err());
+
+    assert!(ECC::import_raw_ex(qx_no_nul, qy, d, ECC::SECP256R1, None, None).is_err());
+    assert!(ECC::import_raw_ex(qx, qy_no_nul, d, ECC::SECP256R1, None, None).is_err());
+    assert!(ECC::import_raw_ex(qx, qy, d_no_nul, ECC::SECP256R1, None, None).is_err());
+    assert!(ECC::import_raw_ex(empty, qy, d, ECC::SECP256R1, None, None).is_err());
+    assert!(ECC::import_raw_ex(qx, empty, d, ECC::SECP256R1, None, None).is_err());
+    assert!(ECC::import_raw_ex(qx, qy, empty, ECC::SECP256R1, None, None).is_err());
+}
+
 #[test]
 fn test_ecc_rs_hex_to_sig_not_null_terminated() {
     common::setup();
diff --git a/wrapper/rust/wolfssl-wolfcrypt/tests/test_hmac_mac.rs b/wrapper/rust/wolfssl-wolfcrypt/tests/test_hmac_mac.rs
index 6a7067b0ee..fab799bae6 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/tests/test_hmac_mac.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/tests/test_hmac_mac.rs
@@ -40,6 +40,27 @@ fn test_hmac_sha256_mac_finalize() {
     assert_eq!(result.as_bytes().as_slice(), expected);
 }
 
+#[test]
+fn test_hmac_sha256_mac_clone() {
+    let key = b"\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b";
+    let prefix = b"Hi ";
+    let suffix = b"There";
+    let expected = b"\xb0\x34\x4c\x61\xd8\xdb\x38\x53\x5c\xa8\xaf\xce\xaf\x0b\xf1\x2b\x88\x1d\xc2\x00\xc9\x83\x3d\xa7\x26\xe9\x37\x6c\x2e\x32\xcf\xf7";
+
+    let mut mac = HmacSha256::new_from_slice(key)
+        .expect("HMAC init failed");
+    mac.update(prefix);
+
+    let mut forked = mac.clone();
+    forked.update(suffix);
+    let forked_tag = forked.finalize();
+    assert_eq!(forked_tag.as_bytes().as_slice(), expected);
+
+    mac.update(suffix);
+    let original_tag = mac.finalize();
+    assert_eq!(original_tag.as_bytes().as_slice(), expected);
+}
+
 #[test]
 fn test_hmac_sha256_mac_verify_fail() {
     let key = b"\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b";
diff --git a/wrapper/rust/wolfssl-wolfcrypt/tests/test_random.rs b/wrapper/rust/wolfssl-wolfcrypt/tests/test_random.rs
index bf7f1536a7..45977123ad 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/tests/test_random.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/tests/test_random.rs
@@ -52,7 +52,7 @@ fn test_test_seed() {
 fn test_rng_generate_byte() {
     // Since a single 0x00 or 0xFF could occur occasionally, we'll combine four
     // bytes into a u32 and make sure they aren't all 0x00 or all 0xFF.
-    let mut rng = RNG::new().expect("Failed to create RNG");
+    let rng = RNG::new().expect("Failed to create RNG");
     let mut v: u32 = 0;
     for _i in 0..4 {
         let byte = rng.generate_byte().expect("Failed to generate a single byte");
@@ -65,7 +65,7 @@ fn test_rng_generate_byte() {
 // Test that generate_block works for a slice of u8.
 #[test]
 fn test_rng_generate_block_u8() {
-    let mut rng = RNG::new().expect("Failed to create RNG");
+    let rng = RNG::new().expect("Failed to create RNG");
     let mut buffer = [0u8; 32];
     rng.generate_block(&mut buffer).expect("Failed to generate a block of bytes");
 
@@ -77,7 +77,7 @@ fn test_rng_generate_block_u8() {
 // Test that generate_block works for a slice of u32.
 #[test]
 fn test_rng_generate_block_u32() {
-    let mut rng = RNG::new().expect("Failed to create RNG");
+    let rng = RNG::new().expect("Failed to create RNG");
     let mut buffer = [0u32; 8];
     rng.generate_block(&mut buffer).expect("Failed to generate a block of u32");
 
@@ -93,7 +93,7 @@ fn test_rng_generate_block_u32() {
 #[test]
 #[cfg(random_hashdrbg)]
 fn test_rng_reseed() {
-    let mut rng = RNG::new().expect("Failed to create RNG");
+    let rng = RNG::new().expect("Failed to create RNG");
     let seed = [1u8, 2, 3, 4];
     rng.reseed(&seed).expect("Error with reseed()");
 }
diff --git a/wrapper/rust/wolfssl-wolfcrypt/tests/test_rsa.rs b/wrapper/rust/wolfssl-wolfcrypt/tests/test_rsa.rs
index b8d19cfaec..c935113433 100644
--- a/wrapper/rust/wolfssl-wolfcrypt/tests/test_rsa.rs
+++ b/wrapper/rust/wolfssl-wolfcrypt/tests/test_rsa.rs
@@ -5,6 +5,8 @@ mod common;
 #[cfg(any(all(sha256, random, rsa_pss), random, rsa_direct))]
 use std::fs;
 #[cfg(random)]
+use std::rc::Rc;
+#[cfg(random)]
 use wolfssl_wolfcrypt::random::RNG;
 #[cfg(any(random, rsa_direct, rsa_keygen))]
 use wolfssl_wolfcrypt::rsa::*;
@@ -14,8 +16,8 @@ use wolfssl_wolfcrypt::rsa::*;
 fn test_rsa_generate() {
     common::setup();
 
-    let mut rng = RNG::new().expect("Error creating RNG");
-    let mut rsa = RSA::generate(2048, 65537, &mut rng).expect("Error with generate()");
+    let rng = RNG::new().expect("Error creating RNG");
+    let mut rsa = RSA::generate(2048, 65537, &rng).expect("Error with generate()");
     rsa.check().expect("Error with check()");
 
     let encrypt_size = rsa.get_encrypt_size().expect("Error with get_encrypt_size()");
@@ -58,20 +60,20 @@ fn test_rsa_generate() {
 #[test]
 #[cfg(random)]
 fn test_rsa_encrypt_decrypt() {
-    let mut rng = RNG::new().expect("Error creating RNG");
+    let rng = Rc::new(RNG::new().expect("Error creating RNG"));
     let key_path = "../../../certs/client-keyPub.der";
     let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     let mut rsa = RSA::new_public_from_der(&der).expect("Error with new_public_from_der()");
-    rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    rsa.set_shared_rng(Rc::clone(&rng)).expect("Error with set_shared_rng()");
     let plain: &[u8] = b"Test message";
     let mut enc: [u8; 512] = [0; 512];
-    let enc_len = rsa.public_encrypt(plain, &mut enc, &mut rng).expect("Error with public_encrypt()");
+    let enc_len = rsa.public_encrypt(plain, &mut enc, &rng).expect("Error with public_encrypt()");
     assert!(enc_len > 0 && enc_len <= 512);
 
     let key_path = "../../../certs/client-key.der";
     let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     let mut rsa = RSA::new_from_der(&der).expect("Error with new_from_der()");
-    rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    rsa.set_shared_rng(Rc::clone(&rng)).expect("Error with set_shared_rng()");
     let mut plain_out: [u8; 512] = [0; 512];
     let dec_len = rsa.private_decrypt(&enc[0..enc_len], &mut plain_out).expect("Error with private_decrypt()");
     assert!(dec_len as usize == plain.len());
@@ -81,20 +83,20 @@ fn test_rsa_encrypt_decrypt() {
 #[test]
 #[cfg(all(sha256, random, rsa_pss))]
 fn test_rsa_pss() {
-    let mut rng = RNG::new().expect("Error creating RNG");
+    let rng = Rc::new(RNG::new().expect("Error creating RNG"));
 
     let key_path = "../../../certs/client-key.der";
     let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     let mut rsa = RSA::new_from_der(&der).expect("Error with new_from_der()");
     let msg: &[u8] = b"This is the string to be signed!";
     let mut signature: [u8; 512] = [0; 512];
-    let sig_len = rsa.pss_sign(msg, &mut signature, RSA::HASH_TYPE_SHA256, RSA::MGF1SHA256, &mut rng).expect("Error with pss_sign()");
+    let sig_len = rsa.pss_sign(msg, &mut signature, RSA::HASH_TYPE_SHA256, RSA::MGF1SHA256, &rng).expect("Error with pss_sign()");
     assert!(sig_len > 0 && sig_len <= 512);
 
     let key_path = "../../../certs/client-keyPub.der";
     let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     let mut rsa = RSA::new_public_from_der(&der).expect("Error with new_public_from_der()");
-    rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    rsa.set_shared_rng(Rc::clone(&rng)).expect("Error with set_shared_rng()");
     let signature = &signature[0..sig_len];
     let mut verify_out: [u8; 512] = [0; 512];
     let verify_out_size = rsa.pss_verify(signature, &mut verify_out, RSA::HASH_TYPE_SHA256, RSA::MGF1SHA256).expect("Error with pss_verify()");
@@ -108,7 +110,7 @@ fn test_rsa_pss() {
 #[test]
 #[cfg(rsa_direct)]
 fn test_rsa_direct() {
-    let mut rng = RNG::new().expect("Error creating RNG");
+    let rng = RNG::new().expect("Error creating RNG");
 
     let key_path = "../../../certs/client-key.der";
     let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
@@ -117,10 +119,10 @@ fn test_rsa_direct() {
     let mut plain = [0u8; 256];
     plain[..msg.len()].copy_from_slice(msg);
     let mut enc = [0u8; 256];
-    let enc_len = rsa.rsa_direct(&plain, &mut enc, RSA::PRIVATE_ENCRYPT, &mut rng).expect("Error with rsa_direct()");
+    let enc_len = rsa.rsa_direct(&plain, &mut enc, RSA::PRIVATE_ENCRYPT, &rng).expect("Error with rsa_direct()");
     assert_eq!(enc_len, 256);
     let mut plain_out = [0u8; 256];
-    let dec_len = rsa.rsa_direct(&enc, &mut plain_out, RSA::PUBLIC_DECRYPT, &mut rng).expect("Error with rsa_direct()");
+    let dec_len = rsa.rsa_direct(&enc, &mut plain_out, RSA::PUBLIC_DECRYPT, &rng).expect("Error with rsa_direct()");
     assert_eq!(dec_len, 256);
     assert_eq!(plain_out, plain);
 }
@@ -128,20 +130,20 @@ fn test_rsa_direct() {
 #[test]
 #[cfg(random)]
 fn test_rsa_ssl() {
-    let mut rng = RNG::new().expect("Error creating RNG");
+    let rng = Rc::new(RNG::new().expect("Error creating RNG"));
 
     let key_path = "../../../certs/client-key.der";
     let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     let mut rsa = RSA::new_from_der(&der).expect("Error with new_from_der()");
     let msg: &[u8] = b"This is the string to be signed!";
     let mut signature: [u8; 512] = [0; 512];
-    let sig_len = rsa.ssl_sign(msg, &mut signature, &mut rng).expect("Error with ssl_sign()");
+    let sig_len = rsa.ssl_sign(msg, &mut signature, &rng).expect("Error with ssl_sign()");
     assert!(sig_len > 0 && sig_len <= 512);
 
     let key_path = "../../../certs/client-keyPub.der";
     let der: Vec<u8> = fs::read(key_path).expect("Error reading key file");
     let mut rsa = RSA::new_public_from_der(&der).expect("Error with new_public_from_der()");
-    rsa.set_rng(&mut rng).expect("Error with set_rng()");
+    rsa.set_shared_rng(Rc::clone(&rng)).expect("Error with set_shared_rng()");
     let signature = &signature[0..sig_len];
     let mut verify_out: [u8; 512] = [0; 512];
     let verify_out_size = rsa.ssl_verify(signature, &mut verify_out).expect("Error with ssl_verify()");