Fixup assembly to compile with ARMv7a

configure.ac

@@ -2081,11 +2081,11 @@ then
             AC_MSG_NOTICE([64bit ARMv8 found, setting mcpu to generic+crypto])
             ;;
         armv7a)
-            AM_CPPFLAGS="$AM_CPPFLAGS -march=armv7-a -mfpu=neon-vfpv3 -DWOLFSSL_ARMASM_NO_CRYPTO"
+            AM_CPPFLAGS="$AM_CPPFLAGS -march=armv7-a -mfpu=neon -DWOLFSSL_ARMASM_NO_CRYPTO -DWOLFSSL_ARM_ARCH=7"
             # Include options.h
             AM_CCASFLAGS="$AM_CCASFLAGS -DEXTERNAL_OPTS_OPENVPN"
             ENABLED_ARMASM_CRYPTO=no
-            AC_MSG_NOTICE([32bit ARMv7-a found, setting mfpu to neon-vfpv3])
+            AC_MSG_NOTICE([32bit ARMv7-a found, setting mfpu to neon])
             ;;
         *)
             AM_CPPFLAGS="$AM_CPPFLAGS -mfpu=crypto-neon-fp-armv8"

wolfcrypt/src/port/arm/armv8-chacha.c

@@ -987,7 +987,12 @@ static WC_INLINE int wc_Chacha_encrypt_256(const word32 input[CHACHA_CHUNK_WORDS
         "VMOV d4, r8, r9 \n\t"
         "STRD r10, r11, %[x_10] \n\t"
         "VMOV d5, r10, r11 \n\t"
+#if defined(WOLFSSL_ARM_ARCH) && (WOLFSSL_ARM_ARCH < 8)
+        "LDR r11, [r14, #4*14] \n\t"
+        "LDR r10, [r14, #4*15] \n\t"
+#else
         "LDRD r11, r10, [r14, #4*14] \n\t"
+#endif
         "VMOV q4, q0 \n\t"
         "VMOV q5, q1 \n\t"
         "VMOV q6, q2 \n\t"
@@ -2754,11 +2759,9 @@ static WC_INLINE void wc_Chacha_encrypt_64(const word32* input, const byte* m,
         /* XOR 8 bytes */
         "CMP        %[bytes], #8         \n\t"
         "BLT        L_chacha20_arm32_64_lt_8_%= \n\t"
-        "VLDR       d8, [%[m], #0]       \n\t"
+        "VLD1.64    { d8 }, [%[m]]!      \n\t"
-        "ADD        %[m], %[m], #8       \n\t"
         "VEOR       d8, d8, d0           \n\t"
-        "VSTR       d8, [%[c], #0]       \n\t"
+        "VST1.64    { d8 }, [%[c]]!      \n\t"
-        "ADD        %[c], %[c], #8       \n\t"
         "SUBS       %[bytes], %[bytes], #8 \n\t"
         "VMOV       d0, d1               \n\t"
         "BEQ        L_chacha20_arm32_64_done_%= \n\t"
@@ -2772,7 +2775,7 @@ static WC_INLINE void wc_Chacha_encrypt_64(const word32* input, const byte* m,
         "EOR        r12, r12, r14        \n\t"
         "STR        r12, [%[c]], #4      \n\t"
         "SUBS       %[bytes], %[bytes], #4 \n\t"
-        "VTRN.32    d0, d0               \n\t"
+        "VSHR.U64   d0, d0, #32          \n\t"
         "BEQ        L_chacha20_arm32_64_done_%= \n\t"
         "\n"
     "L_chacha20_arm32_64_lt_4_%=: \n\t"

wolfcrypt/src/port/arm/armv8-poly1305.c

@@ -29,6 +29,7 @@
 #endif
 
 #include <wolfssl/wolfcrypt/settings.h>
+#include <wolfssl/wolfcrypt/types.h>
 
 #ifdef WOLFSSL_ARMASM
 #ifdef __aarch64__

wolfcrypt/src/port/arm/armv8-sha256.c

@@ -1537,7 +1537,11 @@ int wc_Sha256Transform(wc_Sha256* sha256, const unsigned char* data)
 #else
     XMEMCPY(sha256->buffer, data, WC_SHA256_BLOCK_SIZE);
 #endif
+#ifndef WOLFSSL_ARMASM_NO_CRYPTO
     Sha256Transform(sha256, data, 1);
+#else
+    Transform_Sha256_Len(sha256, data, WC_SHA256_BLOCK_SIZE);
+#endif
     return 0;
 }
 #endif

wolfcrypt/src/sha256.c

@@ -43,6 +43,7 @@ on the specific device platform.
 #endif
 
 #include <wolfssl/wolfcrypt/settings.h>
+#include <wolfssl/wolfcrypt/types.h>
 
 /*
  * SHA256 Build Options:

wolfcrypt/test/test.c

@@ -5921,8 +5921,10 @@ WOLFSSL_TEST_SUBROUTINE int chacha_test(void)
         return -4722;
 
     for (i = 0; i < 18; ++i) {
-        /* this will test all paths */
+        /* this will test all paths
-        /* block sizes: 1 2 3 4 7 8 15 16 31 32 63 64 127 128 255 256 511 512 */
+         * block sizes: 1 3 7 15 31 63 127 255 511 (i = 0- 8)
+         *              2 4 8 16 32 64 128 256 512 (i = 9-17)
+         */
         block_size = (2 << (i%9)) - (i<9?1:0);
         keySz = 32;
 
@@ -5936,16 +5938,16 @@ WOLFSSL_TEST_SUBROUTINE int chacha_test(void)
         if (ret != 0)
             return ret;
 
-        ret |= wc_Chacha_Process(&enc, cipher_big, plain_big,  block_size);
+        ret |= wc_Chacha_Process(&enc, cipher_big, plain_big , block_size);
-        ret |= wc_Chacha_Process(&dec, plain_big,  cipher_big, block_size);
+        ret |= wc_Chacha_Process(&dec, plain_big , cipher_big, block_size);
         if (ret != 0)
             return ret;
 
         if (XMEMCMP(plain_big, input_big, block_size))
-            return -4723-i;
+            return -4740-i*2;
 
         if (XMEMCMP(cipher_big, cipher_big_result, block_size))
-            return -4724-i;
+            return -4741-i*2;
     }
 
     /* Streaming test */