wolfSSL/wolfssl
1
0
Fork 1
mirror of https://github.com/wolfSSL/wolfssl.git synced 2026-07-05 07:40:54 +02:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #10756 from SparkiDev/aes_asm_ymm_zmm

Browse Source 
Intel x64 ASM: Add new assembly for AES
This commit is contained in:
David Garske
2026-06-25 21:41:17 -07:00
committed by GitHub
parent 39c0336cb1 a342eba578
commit 0cecccdf6e
28 changed files with 47480 additions and 3449 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.github/workflows/symbol-prefixes.yml
+1 -1
View File
@@ -49,7 +49,7 @@ jobs:
{
if (($7 !~ /^[0-9]+$/) ||
($8 ~ /^(wc_|wolf|WOLF|__pfx|fe_|sp_[a-zA-Z090-0_]*[0-9])/) ||
($8 ~ /(_avx[12]|_AVX[12]|_sse[12]|_SSE[12]|_aesni|_AESNI|_bmi2|_x64$)/))
($8 ~ /(_avx[12]|_AVX[12]|_sse[12]|_SSE[12]|_aesni|_AESNI|_vaes|_VAES|_avx512|_AVX512|_bmi2|_x64$)/))
{
next;
}
.wolfssl_known_macro_extras
+2
View File
@@ -381,6 +381,7 @@ NO_AES_DECRYPT
NO_ARDUINO_DEFAULT
NO_ASM
NO_ASN_OLD_TYPE_NAMES
NO_AVX512_SUPPORT
NO_CAMELLIA_CBC
NO_CERT
NO_CERT_IN_TICKET
@@ -459,6 +460,7 @@ NO_STDIO_FGETS_REMAP
NO_STM32_HMAC
NO_TKERNEL_MEM_POOL
NO_TLSX_PSKKEM_PLAIN_ANNOUNCE
NO_VAES_SUPPORT
NO_VERIFY_OID
NO_WC_DHGENERATEPUBLIC
NO_WC_SHE_GETUID
linuxkm/Kbuild
+2
View File
@@ -200,6 +200,8 @@ $(obj)/wolfcrypt/src/aes_gcm_asm.o: asflags-y := $(WOLFSSL_ASFLAGS) $(ASFLAGS_FP
$(obj)/wolfcrypt/src/aes_gcm_asm.o: OBJECT_FILES_NON_STANDARD := y
$(obj)/wolfcrypt/src/aes_xts_asm.o: asflags-y := $(WOLFSSL_ASFLAGS) $(ASFLAGS_FPU_DISABLE_SIMD_ENABLE)
$(obj)/wolfcrypt/src/aes_xts_asm.o: OBJECT_FILES_NON_STANDARD := y
$(obj)/wolfcrypt/src/aes_x86_64_asm.o: asflags-y := $(WOLFSSL_ASFLAGS) $(ASFLAGS_FPU_DISABLE_SIMD_ENABLE)
$(obj)/wolfcrypt/src/aes_x86_64_asm.o: OBJECT_FILES_NON_STANDARD := y
$(obj)/wolfcrypt/src/sp_x86_64_asm.o: asflags-y := $(WOLFSSL_ASFLAGS) $(ASFLAGS_FPU_DISABLE_SIMD_ENABLE)
$(obj)/wolfcrypt/src/sp_x86_64_asm.o: OBJECT_FILES_NON_STANDARD := y
$(obj)/wolfcrypt/src/sha256_asm.o: asflags-y := $(WOLFSSL_ASFLAGS) $(ASFLAGS_FPU_DISABLE_SIMD_ENABLE)
src/include.am
+5
View File
@@ -109,6 +109,7 @@ endif
if BUILD_AESNI
src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/aes_asm.S
src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/aes_x86_64_asm.S
if BUILD_X86_ASM
src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/aes_gcm_x86_asm.S
else
@@ -259,6 +260,7 @@ endif BUILD_PPC64_ASM
if BUILD_AESNI
src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/aes_asm.S
src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/aes_x86_64_asm.S
if BUILD_X86_ASM
src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/aes_gcm_x86_asm.S
else
@@ -532,6 +534,7 @@ endif BUILD_PPC64_ASM
if BUILD_AESNI
src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/aes_asm.S
src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/aes_x86_64_asm.S
if BUILD_X86_ASM
src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/aes_gcm_x86_asm.S
else
@@ -867,6 +870,7 @@ endif BUILD_AES
if BUILD_AESNI
src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/aes_asm.S
src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/aes_x86_64_asm.S
if BUILD_X86_ASM
src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/aes_gcm_x86_asm.S
else
@@ -1708,6 +1712,7 @@ endif
if !BUILD_FIPS_V2_PLUS
if BUILD_AESNI
src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/aes_asm.S
src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/aes_x86_64_asm.S
if BUILD_X86_ASM
src_libwolfssl@LIBSUFFIX@_la_SOURCES += wolfcrypt/src/aes_gcm_x86_asm.S
else
wolfcrypt/benchmark/benchmark.c
+2
View File
@@ -4794,6 +4794,8 @@ static void print_cpu_features(void)
if (IS_INTEL_MOVBE(cpuid_flags)) printf(" movbe");
if (IS_INTEL_BMI1(cpuid_flags)) printf(" bmi1");
if (IS_INTEL_SHA(cpuid_flags)) printf(" sha");
if (IS_INTEL_VAES(cpuid_flags)) printf(" vaes");
if (IS_INTEL_AVX512(cpuid_flags)) printf(" avx512");
#endif
#ifdef __aarch64__
printf("Aarch64 -");
wolfcrypt/src/aes.c
+773 -2
View File
File diff suppressed because it is too large Load Diff
wolfcrypt/src/aes_asm.S
+2 -1309
View File
File diff suppressed because it is too large Load Diff
wolfcrypt/src/aes_asm.asm
+51 -1528
View File
File diff suppressed because it is too large Load Diff
wolfcrypt/src/aes_gcm_asm.S
+14233 -16
View File
File diff suppressed because it is too large Load Diff
wolfcrypt/src/aes_gcm_asm.asm
+14165 -16
View File
File diff suppressed because it is too large Load Diff
wolfcrypt/src/aes_x86_64_asm.S
+4375
View File
File diff suppressed because it is too large Load Diff
wolfcrypt/src/aes_x86_64_asm.asm
+4283
View File
File diff suppressed because it is too large Load Diff
wolfcrypt/src/aes_xts_asm.S
+4412
View File
File diff suppressed because it is too large Load Diff
wolfcrypt/src/aes_xts_asm.asm
+4468
View File
File diff suppressed because it is too large Load Diff
wolfcrypt/src/chacha_asm.S
+10
View File
@@ -46,6 +46,16 @@
#define HAVE_INTEL_AVX2
#endif /* HAVE_INTEL_AVX2 */
#endif /* NO_AVX2_SUPPORT */
#ifndef NO_VAES_SUPPORT
#ifndef HAVE_INTEL_VAES
#define HAVE_INTEL_VAES
#endif /* HAVE_INTEL_VAES */
#endif /* NO_VAES_SUPPORT */
#ifndef NO_AVX512_SUPPORT
#ifndef HAVE_INTEL_AVX512
#define HAVE_INTEL_AVX512
#endif /* HAVE_INTEL_AVX512 */
#endif /* NO_AVX512_SUPPORT */
#ifdef WOLFSSL_X86_64_BUILD
#ifndef __APPLE__
wolfcrypt/src/cpuid.c
+2
View File
@@ -130,6 +130,8 @@
if (cpuid_flag(1, 0, ECX, 22)) { new_cpuid_flags |= CPUID_MOVBE ; }
if (cpuid_flag(7, 0, EBX, 3)) { new_cpuid_flags |= CPUID_BMI1 ; }
if (cpuid_flag(7, 0, EBX, 29)) { new_cpuid_flags |= CPUID_SHA ; }
if (cpuid_flag(7, 0, ECX, 9)) { new_cpuid_flags |= CPUID_VAES ; }
if (cpuid_flag(7, 0, EBX, 16)) { new_cpuid_flags |= CPUID_AVX512; }
(void)wolfSSL_Atomic_Uint_CompareExchange
(&cpuid_flags, &old_cpuid_flags, new_cpuid_flags);
}
wolfcrypt/src/fe_x25519_asm.S
+10
View File
@@ -46,6 +46,16 @@
#define HAVE_INTEL_AVX2
#endif /* HAVE_INTEL_AVX2 */
#endif /* NO_AVX2_SUPPORT */
#ifndef NO_VAES_SUPPORT
#ifndef HAVE_INTEL_VAES
#define HAVE_INTEL_VAES
#endif /* HAVE_INTEL_VAES */
#endif /* NO_VAES_SUPPORT */
#ifndef NO_AVX512_SUPPORT
#ifndef HAVE_INTEL_AVX512
#define HAVE_INTEL_AVX512
#endif /* HAVE_INTEL_AVX512 */
#endif /* NO_AVX512_SUPPORT */
#ifndef __APPLE__
.text
wolfcrypt/src/include.am
+1
View File
@@ -16,6 +16,7 @@ EXTRA_DIST += wolfcrypt/src/evp.c
EXTRA_DIST += wolfcrypt/src/evp_pk.c
EXTRA_DIST += wolfcrypt/src/asm.c
EXTRA_DIST += wolfcrypt/src/aes_asm.asm
EXTRA_DIST += wolfcrypt/src/aes_x86_64_asm.asm
EXTRA_DIST += wolfcrypt/src/aes_gcm_asm.asm
EXTRA_DIST += wolfcrypt/src/aes_xts_asm.asm
EXTRA_DIST += wolfcrypt/src/chacha_asm.asm
wolfcrypt/src/poly1305_asm.S
+10
View File
@@ -46,6 +46,16 @@
#define HAVE_INTEL_AVX2
#endif /* HAVE_INTEL_AVX2 */
#endif /* NO_AVX2_SUPPORT */
#ifndef NO_VAES_SUPPORT
#ifndef HAVE_INTEL_VAES
#define HAVE_INTEL_VAES
#endif /* HAVE_INTEL_VAES */
#endif /* NO_VAES_SUPPORT */
#ifndef NO_AVX512_SUPPORT
#ifndef HAVE_INTEL_AVX512
#define HAVE_INTEL_AVX512
#endif /* HAVE_INTEL_AVX512 */
#endif /* NO_AVX512_SUPPORT */
#ifdef WOLFSSL_X86_64_BUILD
#ifdef HAVE_INTEL_AVX1
wolfcrypt/src/sha256_asm.S
+10
View File
@@ -46,6 +46,16 @@
#define HAVE_INTEL_AVX2
#endif /* HAVE_INTEL_AVX2 */
#endif /* NO_AVX2_SUPPORT */
#ifndef NO_VAES_SUPPORT
#ifndef HAVE_INTEL_VAES
#define HAVE_INTEL_VAES
#endif /* HAVE_INTEL_VAES */
#endif /* NO_VAES_SUPPORT */
#ifndef NO_AVX512_SUPPORT
#ifndef HAVE_INTEL_AVX512
#define HAVE_INTEL_AVX512
#endif /* HAVE_INTEL_AVX512 */
#endif /* NO_AVX512_SUPPORT */
#ifdef WOLFSSL_X86_64_BUILD
#ifndef __APPLE__
wolfcrypt/src/sha3_asm.S
+10
View File
@@ -46,6 +46,16 @@
#define HAVE_INTEL_AVX2
#endif /* HAVE_INTEL_AVX2 */
#endif /* NO_AVX2_SUPPORT */
#ifndef NO_VAES_SUPPORT
#ifndef HAVE_INTEL_VAES
#define HAVE_INTEL_VAES
#endif /* HAVE_INTEL_VAES */
#endif /* NO_VAES_SUPPORT */
#ifndef NO_AVX512_SUPPORT
#ifndef HAVE_INTEL_AVX512
#define HAVE_INTEL_AVX512
#endif /* HAVE_INTEL_AVX512 */
#endif /* NO_AVX512_SUPPORT */
#ifndef __APPLE__
.data
wolfcrypt/src/sha512_asm.S
+10
View File
@@ -46,6 +46,16 @@
#define HAVE_INTEL_AVX2
#endif /* HAVE_INTEL_AVX2 */
#endif /* NO_AVX2_SUPPORT */
#ifndef NO_VAES_SUPPORT
#ifndef HAVE_INTEL_VAES
#define HAVE_INTEL_VAES
#endif /* HAVE_INTEL_VAES */
#endif /* NO_VAES_SUPPORT */
#ifndef NO_AVX512_SUPPORT
#ifndef HAVE_INTEL_AVX512
#define HAVE_INTEL_AVX512
#endif /* HAVE_INTEL_AVX512 */
#endif /* NO_AVX512_SUPPORT */
#ifdef HAVE_INTEL_AVX1
#ifndef __APPLE__
wolfcrypt/src/wc_mldsa_asm.S
+10
View File
@@ -46,6 +46,16 @@
#define HAVE_INTEL_AVX2
#endif /* HAVE_INTEL_AVX2 */
#endif /* NO_AVX2_SUPPORT */
#ifndef NO_VAES_SUPPORT
#ifndef HAVE_INTEL_VAES
#define HAVE_INTEL_VAES
#endif /* HAVE_INTEL_VAES */
#endif /* NO_VAES_SUPPORT */
#ifndef NO_AVX512_SUPPORT
#ifndef HAVE_INTEL_AVX512
#define HAVE_INTEL_AVX512
#endif /* HAVE_INTEL_AVX512 */
#endif /* NO_AVX512_SUPPORT */
#ifdef WOLFSSL_HAVE_MLDSA
#ifdef HAVE_INTEL_AVX2
wolfcrypt/src/wc_mlkem_asm.S
+10
View File
@@ -46,6 +46,16 @@
#define HAVE_INTEL_AVX2
#endif /* HAVE_INTEL_AVX2 */
#endif /* NO_AVX2_SUPPORT */
#ifndef NO_VAES_SUPPORT
#ifndef HAVE_INTEL_VAES
#define HAVE_INTEL_VAES
#endif /* HAVE_INTEL_VAES */
#endif /* NO_VAES_SUPPORT */
#ifndef NO_AVX512_SUPPORT
#ifndef HAVE_INTEL_AVX512
#define HAVE_INTEL_AVX512
#endif /* HAVE_INTEL_AVX512 */
#endif /* NO_AVX512_SUPPORT */
#ifdef WOLFSSL_HAVE_MLKEM
#ifdef HAVE_INTEL_AVX2
wolfssl-VS2022.vcxproj
+591 -577
View File
File diff suppressed because it is too large Load Diff
wolfssl.vcxproj
+14
View File
@@ -489,6 +489,20 @@
<Outputs Condition="'$(Configuration)|$(Platform)'=='Release|x64'">$(OutDir)%(Filename).obj</Outputs>
<Outputs Condition="'$(Configuration)|$(Platform)'=='DLL Release|x64'">$(IntDir)%(Filename).obj</Outputs>
</CustomBuild>
<CustomBuild Include="wolfcrypt\src\aes_x86_64_asm.asm">
<ExcludedFromBuild Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">false</ExcludedFromBuild>
<ExcludedFromBuild Condition="'$(Configuration)|$(Platform)'=='DLL Debug|x64'">false</ExcludedFromBuild>
<Command Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">ml64.exe /c /Zi /Fo"$(OutDir)%(Filename).obj" %(Identity)</Command>
<Command Condition="'$(Configuration)|$(Platform)'=='DLL Debug|x64'">ml64.exe /c /Zi /Fo"$(IntDir)%(Filename).obj" %(Identity)</Command>
<Outputs Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">$(OutDir)%(Filename).obj</Outputs>
<Outputs Condition="'$(Configuration)|$(Platform)'=='DLL Debug|x64'">$(IntDir)%(Filename).obj</Outputs>
<ExcludedFromBuild Condition="'$(Configuration)|$(Platform)'=='Release|x64'">false</ExcludedFromBuild>
<ExcludedFromBuild Condition="'$(Configuration)|$(Platform)'=='DLL Release|x64'">false</ExcludedFromBuild>
<Command Condition="'$(Configuration)|$(Platform)'=='Release|x64'">ml64.exe /c /Zi /Fo"$(OutDir)%(Filename).obj" %(Identity)</Command>
<Command Condition="'$(Configuration)|$(Platform)'=='DLL Release|x64'">ml64.exe /c /Zi /Fo"$(IntDir)%(Filename).obj" %(Identity)</Command>
<Outputs Condition="'$(Configuration)|$(Platform)'=='Release|x64'">$(OutDir)%(Filename).obj</Outputs>
<Outputs Condition="'$(Configuration)|$(Platform)'=='DLL Release|x64'">$(IntDir)%(Filename).obj</Outputs>
</CustomBuild>
<CustomBuild Include="wolfcrypt\src\aes_gcm_asm.asm">
<ExcludedFromBuild Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">false</ExcludedFromBuild>
<ExcludedFromBuild Condition="'$(Configuration)|$(Platform)'=='DLL Debug|x64'">false</ExcludedFromBuild>
wolfssl/wolfcrypt/cpuid.h
+4
View File
@@ -67,6 +67,8 @@ typedef word32 cpuid_flags_t;
#define CPUID_MOVBE 0x0080 /* Move and byte swap */
#define CPUID_BMI1 0x0100 /* ANDN */
#define CPUID_SHA 0x0200 /* SHA-1 and SHA-256 instructions */
#define CPUID_VAES 0x0400
#define CPUID_AVX512 0x0800
#define IS_INTEL_AVX1(f) (WOLFSSL_ATOMIC_COERCE_UINT(f) & CPUID_AVX1)
#define IS_INTEL_AVX2(f) (WOLFSSL_ATOMIC_COERCE_UINT(f) & CPUID_AVX2)
@@ -78,6 +80,8 @@ typedef word32 cpuid_flags_t;
#define IS_INTEL_MOVBE(f) (WOLFSSL_ATOMIC_COERCE_UINT(f) & CPUID_MOVBE)
#define IS_INTEL_BMI1(f) (WOLFSSL_ATOMIC_COERCE_UINT(f) & CPUID_BMI1)
#define IS_INTEL_SHA(f) (WOLFSSL_ATOMIC_COERCE_UINT(f) & CPUID_SHA)
#define IS_INTEL_VAES(f) (WOLFSSL_ATOMIC_COERCE_UINT(f) & CPUID_VAES)
#define IS_INTEL_AVX512(f) (WOLFSSL_ATOMIC_COERCE_UINT(f) & CPUID_AVX512)
#elif defined(HAVE_CPUID_AARCH64)
wrapper/CSharp/wolfssl.vcxproj
+14
View File
@@ -371,6 +371,20 @@
<Outputs Condition="'$(Configuration)|$(Platform)'=='Release|x64'">$(OutDir)%(Filename).obj</Outputs>
<Outputs Condition="'$(Configuration)|$(Platform)'=='DLL Release|x64'">$(IntDir)%(Filename).obj</Outputs>
</CustomBuild>
<CustomBuild Include="..\..\wolfcrypt\src\aes_x86_64_asm.asm">
<ExcludedFromBuild Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">false</ExcludedFromBuild>
<ExcludedFromBuild Condition="'$(Configuration)|$(Platform)'=='DLL Debug|x64'">false</ExcludedFromBuild>
<Command Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">ml64.exe /c /Zi /Fo"$(OutDir)%(Filename).obj" %(Identity)</Command>
<Command Condition="'$(Configuration)|$(Platform)'=='DLL Debug|x64'">ml64.exe /c /Zi /Fo"$(IntDir)%(Filename).obj" %(Identity)</Command>
<Outputs Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">$(OutDir)%(Filename).obj</Outputs>
<Outputs Condition="'$(Configuration)|$(Platform)'=='DLL Debug|x64'">$(IntDir)%(Filename).obj</Outputs>
<ExcludedFromBuild Condition="'$(Configuration)|$(Platform)'=='Release|x64'">false</ExcludedFromBuild>
<ExcludedFromBuild Condition="'$(Configuration)|$(Platform)'=='DLL Release|x64'">false</ExcludedFromBuild>
<Command Condition="'$(Configuration)|$(Platform)'=='Release|x64'">ml64.exe /c /Zi /Fo"$(OutDir)%(Filename).obj" %(Identity)</Command>
<Command Condition="'$(Configuration)|$(Platform)'=='DLL Release|x64'">ml64.exe /c /Zi /Fo"$(IntDir)%(Filename).obj" %(Identity)</Command>
<Outputs Condition="'$(Configuration)|$(Platform)'=='Release|x64'">$(OutDir)%(Filename).obj</Outputs>
<Outputs Condition="'$(Configuration)|$(Platform)'=='DLL Release|x64'">$(IntDir)%(Filename).obj</Outputs>
</CustomBuild>
<CustomBuild Include="..\..\wolfcrypt\src\aes_gcm_asm.asm">
<ExcludedFromBuild Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">false</ExcludedFromBuild>
<ExcludedFromBuild Condition="'$(Configuration)|$(Platform)'=='DLL Debug|x64'">false</ExcludedFromBuild>