diff --git a/wolfcrypt/benchmark/benchmark.c b/wolfcrypt/benchmark/benchmark.c
index fbb79c3e7..76cd20f71 100644
--- a/wolfcrypt/benchmark/benchmark.c
+++ b/wolfcrypt/benchmark/benchmark.c
@@ -525,7 +525,41 @@ void bench_aes(int show)
 #endif
 
     if (show) {
-        printf("AES      %d %s took %5.3f seconds, %8.3f MB/s", numBlocks,
+        printf("AES enc  %d %s took %5.3f seconds, %8.3f MB/s", numBlocks,
+                                                  blockType, total, persec);
+        SHOW_INTEL_CYCLES
+        printf("\n");
+    }
+#ifdef HAVE_CAVIUM
+    wc_AesFreeCavium(&enc);
+    if (wc_AesInitCavium(&enc, CAVIUM_DEV_ID) != 0) {
+        printf("aes init cavium failed\n");
+        return;
+    }
+#endif
+
+    ret = wc_AesSetKey(&enc, key, 16, iv, AES_DECRYPTION);
+    if (ret != 0) {
+        printf("AesSetKey failed, ret = %d\n", ret);
+        return;
+    }
+    start = current_time(1);
+    BEGIN_INTEL_CYCLES
+
+    for(i = 0; i < numBlocks; i++)
+        wc_AesCbcDecrypt(&enc, plain, cipher, sizeof(plain));
+
+    END_INTEL_CYCLES
+    total = current_time(0) - start;
+
+    persec = 1 / total * numBlocks;
+#ifdef BENCH_EMBEDDED
+    /* since using kB, convert to MB/s */
+    persec = persec / 1024;
+#endif
+
+    if (show) {
+        printf("AES dec  %d %s took %5.3f seconds, %8.3f MB/s", numBlocks,
                                                   blockType, total, persec);
         SHOW_INTEL_CYCLES
         printf("\n");
diff --git a/wolfcrypt/src/aes.c b/wolfcrypt/src/aes.c
index 6d6eae21e..6d5d9fa96 100644
--- a/wolfcrypt/src/aes.c
+++ b/wolfcrypt/src/aes.c
@@ -1094,10 +1094,22 @@ void AES_CBC_encrypt(const unsigned char* in, unsigned char* out,
                      XASM_LINK("AES_CBC_encrypt");
 
 #ifdef HAVE_AES_DECRYPT
-void AES_CBC_decrypt(const unsigned char* in, unsigned char* out,
-                     unsigned char* ivec, unsigned long length,
-                     const unsigned char* KS, int nr)
-                     XASM_LINK("AES_CBC_decrypt");
+    #if defined(WOLFSSL_AESNI_BY4)
+    void AES_CBC_decrypt_by4(const unsigned char* in, unsigned char* out,
+                             unsigned char* ivec, unsigned long length,
+                             const unsigned char* KS, int nr)
+                             XASM_LINK("AES_CBC_decrypt_by4");
+    #elif defined(WOLFSSL_AESNI_BY6)
+    void AES_CBC_decrypt_by6(const unsigned char* in, unsigned char* out,
+                             unsigned char* ivec, unsigned long length,
+                             const unsigned char* KS, int nr)
+                             XASM_LINK("AES_CBC_decrypt_by6");
+    #else /* WOLFSSL_AESNI_BYx */
+    void AES_CBC_decrypt_by8(const unsigned char* in, unsigned char* out,
+                             unsigned char* ivec, unsigned long length,
+                             const unsigned char* KS, int nr)
+                             XASM_LINK("AES_CBC_decrypt_by8");
+    #endif /* WOLFSSL_AESNI_BYx */
 #endif /* HAVE_AES_DECRYPT */
 #endif /* HAVE_AES_CBC */
 
@@ -2549,8 +2561,16 @@ int wc_AesSetIV(Aes* aes, const byte* iv)
 
             /* if input and output same will overwrite input iv */
             XMEMCPY(aes->tmp, in + sz - AES_BLOCK_SIZE, AES_BLOCK_SIZE);
-            AES_CBC_decrypt(in, out, (byte*)aes->reg, sz, (byte*)aes->key,
+            #if defined(WOLFSSL_AESNI_BY4)
+            AES_CBC_decrypt_by4(in, out, (byte*)aes->reg, sz, (byte*)aes->key,
                             aes->rounds);
+            #elif defined(WOLFSSL_AESNI_BY6)
+            AES_CBC_decrypt_by6(in, out, (byte*)aes->reg, sz, (byte*)aes->key,
+                            aes->rounds);
+            #else /* WOLFSSL_AESNI_BYx */
+            AES_CBC_decrypt_by8(in, out, (byte*)aes->reg, sz, (byte*)aes->key,
+                            aes->rounds);
+            #endif /* WOLFSSL_AESNI_BYx */
             /* store iv for next call */
             XMEMCPY(aes->reg, aes->tmp, AES_BLOCK_SIZE);
             return 0;
diff --git a/wolfcrypt/src/aes_asm.asm b/wolfcrypt/src/aes_asm.asm
index 5453d2e45..6fe026d5c 100644
--- a/wolfcrypt/src/aes_asm.asm
+++ b/wolfcrypt/src/aes_asm.asm
@@ -101,220 +101,753 @@ LAST:
 AES_CBC_encrypt ENDP
 
 
+; void AES_CBC_decrypt_by4(const unsigned char* in,
+;                          unsigned char* out,
+;                          unsigned char ivec[16],
+;                          unsigned long length,
+;                          const unsigned char* KS,
+;                          int nr)
+AES_CBC_decrypt_by4 PROC
+; parameter 1: rdi
+; parameter 2: rsi
+; parameter 3: rdx
+; parameter 4: rcx
+; parameter 5: r8
+; parameter 6: r9d
 
-;	/*
-;	AES_CBC_decrypt[const	,unsigned	char*in
-;	unsigned	,char*out
-;	unsigned	,char	ivec+16
-;	unsigned	,long	length
-;	const	,unsigned	char*KS
-;	int	nr]
-;	*/
-;	.	globl	AES_CBC_decrypt
-AES_CBC_decrypt PROC
-;#	parameter	1:	rdi
-;#	parameter	2:	rsi
-;#	parameter	3:	rdx
-;#	parameter	4:	rcx
-;#	parameter	5:	r8
-;#	parameter	6:	r9d
-
-; save rdi and rsi to rax and r11, restore before ret
-	mov rax,rdi
-	mov r11,rsi
-
-; convert to what we had for att&t convention
-	mov rdi,rcx
-	mov rsi,rdx
-	mov rdx,r8
-	mov rcx,r9
-	mov r8,[rsp+40]
-	mov r9d,[rsp+48]
-
-; on microsoft xmm6-xmm15 are non volaitle, let's save on stack and restore at end
-	sub rsp,8+8*16  ; 8 = align stack , 8 xmm6-12,15 16 bytes each
-	movdqa [rsp+0], xmm6
-	movdqa [rsp+16], xmm7
-	movdqa [rsp+32], xmm8
-	movdqa [rsp+48], xmm9
-	movdqa [rsp+64], xmm10
-	movdqa [rsp+80], xmm11
-	movdqa [rsp+96], xmm12
-	movdqa [rsp+112], xmm15
-
-	mov	r10,rcx
-	shr	rcx,4
-	shl	r10,60
-	je	DNO_PARTS_4
-	add	rcx,1
+        ; save rdi and rsi to rax and r11, restore before ret
+        mov         rax, rdi
+        mov         r11, rsi
+        ; convert to what we had for att&t convention
+        mov         rdi, rcx
+        mov         rsi, rdx
+        mov         rdx, r8
+        mov         rcx,r9
+        mov         r8, [rsp+40]
+        mov         r9d, [rsp+48]
+        ; on microsoft xmm6-xmm15 are non volatile,
+        ; let's save on stack and restore at end
+        sub         rsp, 8+8*16  ; 8 = align stack , 8 xmm6-12,15 16 bytes each
+        movdqa      [rsp+0], xmm6
+        movdqa      [rsp+16], xmm7
+        movdqa      [rsp+32], xmm8
+        movdqa      [rsp+48], xmm9
+        movdqa      [rsp+64], xmm10
+        movdqa      [rsp+80], xmm11
+        movdqa      [rsp+96], xmm12
+        movdqa      [rsp+112], xmm15
+        ; back to our original code, more or less
+        mov         r10, rcx
+        shr         rcx, 4
+        shl         r10, 60
+        je          DNO_PARTS_4
+        add         rcx, 1
 DNO_PARTS_4:
-	mov	r10,rcx
-	shl	r10,62
-	shr	r10,62
-	shr	rcx,2
-	movdqu xmm5,[rdx]
-	je	DREMAINDER_4
-	sub	rsi,64
+        mov         r10, rcx
+        shl         r10, 62
+        shr         r10, 62
+        shr         rcx, 2
+        movdqu      xmm5, [rdx]
+        je          DREMAINDER_4
+        sub         rsi, 64
 DLOOP_4:
-	movdqu  xmm1,[rdi]
-	movdqu	xmm2,16[rdi]
-	movdqu	xmm3,32[rdi]
-	movdqu	xmm4,48[rdi]
-	movdqa	xmm6,xmm1
-	movdqa	xmm7,xmm2
-	movdqa	xmm8,xmm3
-	movdqa	xmm15,xmm4
-	movdqa  xmm9,[r8]
-	movdqa	xmm10,16[r8]
-	movdqa	xmm11,32[r8]
-	movdqa	xmm12,48[r8]
-	pxor	xmm1,xmm9
-	pxor	xmm2,xmm9
-	pxor	xmm3,xmm9
-
-	pxor	xmm4,xmm9
-	aesdec	xmm1,xmm10
-	aesdec	xmm2,xmm10
-	aesdec	xmm3,xmm10
-	aesdec	xmm4,xmm10
-	aesdec	xmm1,xmm11
-	aesdec	xmm2,xmm11
-	aesdec	xmm3,xmm11
-	aesdec	xmm4,xmm11
-	aesdec	xmm1,xmm12
-	aesdec	xmm2,xmm12
-	aesdec	xmm3,xmm12
-	aesdec	xmm4,xmm12
-	movdqa	xmm9,64[r8]
-	movdqa	xmm10,80[r8]
-	movdqa	xmm11,96[r8]
-	movdqa	xmm12,112[r8]
-	aesdec	xmm1,xmm9
-	aesdec	xmm2,xmm9
-	aesdec	xmm3,xmm9
-	aesdec	xmm4,xmm9
-	aesdec	xmm1,xmm10
-	aesdec	xmm2,xmm10
-	aesdec	xmm3,xmm10
-	aesdec	xmm4,xmm10
-	aesdec	xmm1,xmm11
-	aesdec	xmm2,xmm11
-	aesdec	xmm3,xmm11
-	aesdec	xmm4,xmm11
-	aesdec	xmm1,xmm12
-	aesdec	xmm2,xmm12
-	aesdec	xmm3,xmm12
-	aesdec	xmm4,xmm12
-	movdqa	xmm9,128[r8]
-	movdqa	xmm10,144[r8]
-	movdqa	xmm11,160[r8]
-	cmp	r9d,12
-	aesdec	xmm1,xmm9
-	aesdec	xmm2,xmm9
-	aesdec	xmm3,xmm9
-	aesdec	xmm4,xmm9
-	aesdec	xmm1,xmm10
-	aesdec	xmm2,xmm10
-	aesdec	xmm3,xmm10
-	aesdec	xmm4,xmm10
-	jb	DLAST_4
-	movdqa	xmm9,160[r8]
-	movdqa	xmm10,176[r8]
-	movdqa	xmm11,192[r8]
-	cmp	r9d,14
-	aesdec	xmm1,xmm9
-	aesdec	xmm2,xmm9
-	aesdec	xmm3,xmm9
-	aesdec	xmm4,xmm9
-	aesdec	xmm1,xmm10
-	aesdec	xmm2,xmm10
-	aesdec	xmm3,xmm10
-	aesdec	xmm4,xmm10
-	jb	DLAST_4
-
-	movdqa	xmm9,192[r8]
-	movdqa	xmm10,208[r8]
-	movdqa	xmm11,224[r8]
-	aesdec	xmm1,xmm9
-	aesdec	xmm2,xmm9
-	aesdec	xmm3,xmm9
-	aesdec	xmm4,xmm9
-	aesdec	xmm1,xmm10
-	aesdec	xmm2,xmm10
-	aesdec	xmm3,xmm10
-	aesdec	xmm4,xmm10
+        movdqu      xmm1, [rdi]
+        movdqu      xmm2, 16[rdi]
+        movdqu      xmm3, 32[rdi]
+        movdqu      xmm4, 48[rdi]
+        movdqa      xmm6, xmm1
+        movdqa      xmm7, xmm2
+        movdqa      xmm8, xmm3
+        movdqa      xmm15, xmm4
+        movdqa      xmm9, [r8]
+        movdqa      xmm10, 16[r8]
+        movdqa      xmm11, 32[r8]
+        movdqa      xmm12, 48[r8]
+        pxor        xmm1, xmm9
+        pxor        xmm2, xmm9
+        pxor        xmm3, xmm9
+        pxor        xmm4, xmm9
+        aesdec      xmm1, xmm10
+        aesdec      xmm2, xmm10
+        aesdec      xmm3, xmm10
+        aesdec      xmm4, xmm10
+        aesdec      xmm1, xmm11
+        aesdec      xmm2, xmm11
+        aesdec      xmm3, xmm11
+        aesdec      xmm4, xmm11
+        aesdec      xmm1, xmm12
+        aesdec      xmm2, xmm12
+        aesdec      xmm3, xmm12
+        aesdec      xmm4, xmm12
+        movdqa      xmm9, 64[r8]
+        movdqa      xmm10, 80[r8]
+        movdqa      xmm11, 96[r8]
+        movdqa      xmm12, 112[r8]
+        aesdec      xmm1, xmm9
+        aesdec      xmm2, xmm9
+        aesdec      xmm3, xmm9
+        aesdec      xmm4, xmm9
+        aesdec      xmm1, xmm10
+        aesdec      xmm2, xmm10
+        aesdec      xmm3, xmm10
+        aesdec      xmm4, xmm10
+        aesdec      xmm1, xmm11
+        aesdec      xmm2, xmm11
+        aesdec      xmm3, xmm11
+        aesdec      xmm4, xmm11
+        aesdec      xmm1, xmm12
+        aesdec      xmm2, xmm12
+        aesdec      xmm3, xmm12
+        aesdec      xmm4, xmm12
+        movdqa      xmm9, 128[r8]
+        movdqa      xmm10, 144[r8]
+        movdqa      xmm11, 160[r8]
+        cmp         r9d, 12
+        aesdec      xmm1, xmm9
+        aesdec      xmm2, xmm9
+        aesdec      xmm3, xmm9
+        aesdec      xmm4, xmm9
+        aesdec      xmm1, xmm10
+        aesdec      xmm2, xmm10
+        aesdec      xmm3, xmm10
+        aesdec      xmm4, xmm10
+        jb          DLAST_4
+        movdqa      xmm9, 160[r8]
+        movdqa      xmm10, 176[r8]
+        movdqa      xmm11, 192[r8]
+        cmp         r9d, 14
+        aesdec      xmm1, xmm9
+        aesdec      xmm2, xmm9
+        aesdec      xmm3, xmm9
+        aesdec      xmm4, xmm9
+        aesdec      xmm1, xmm10
+        aesdec      xmm2, xmm10
+        aesdec      xmm3, xmm10
+        aesdec      xmm4, xmm10
+        jb          DLAST_4
+        movdqa      xmm9, 192[r8]
+        movdqa      xmm10, 208[r8]
+        movdqa      xmm11, 224[r8]
+        aesdec      xmm1, xmm9
+        aesdec      xmm2, xmm9
+        aesdec      xmm3, xmm9
+        aesdec      xmm4, xmm9
+        aesdec      xmm1, xmm10
+        aesdec      xmm2, xmm10
+        aesdec      xmm3, xmm10
+        aesdec      xmm4, xmm10
 DLAST_4:
-	add	rdi,64
-	add	rsi,64
-	dec	rcx
-	aesdeclast	xmm1,xmm11
-	aesdeclast	xmm2,xmm11
-	aesdeclast	xmm3,xmm11
-	aesdeclast	xmm4,xmm11
-	pxor	xmm1,xmm5
-	pxor	xmm2,xmm6
-	pxor	xmm3,xmm7
-	pxor	xmm4,xmm8
-	movdqu	[rsi],xmm1
-	movdqu	16[rsi],xmm2
-	movdqu	32[rsi],xmm3
-	movdqu	48[rsi],xmm4
-	movdqa	xmm5,xmm15
-	jne	DLOOP_4
-	add	rsi,64
+        add         rdi, 64
+        add         rsi, 64
+        dec         rcx
+        aesdeclast  xmm1, xmm11
+        aesdeclast  xmm2, xmm11
+        aesdeclast  xmm3, xmm11
+        aesdeclast  xmm4, xmm11
+        pxor        xmm1, xmm5
+        pxor        xmm2, xmm6
+        pxor        xmm3, xmm7
+        pxor        xmm4, xmm8
+        movdqu      [rsi], xmm1
+        movdqu      16[rsi], xmm2
+        movdqu      32[rsi], xmm3
+        movdqu      48[rsi], xmm4
+        movdqa      xmm5, xmm15
+        jne         DLOOP_4
+        add         rsi, 64
 DREMAINDER_4:
-	cmp	r10,0
-	je	DEND_4
+        cmp         r10, 0
+        je          DEND_4
 DLOOP_4_2:
-	movdqu  xmm1,[rdi]
-	movdqa	xmm15,xmm1
-	add	rdi,16
-	pxor	xmm1,[r8]
-	movdqu	xmm2,160[r8]
-	cmp	r9d,12
-	aesdec	xmm1,16[r8]
-	aesdec	xmm1,32[r8]
-	aesdec	xmm1,48[r8]
-	aesdec	xmm1,64[r8]
-	aesdec	xmm1,80[r8]
-	aesdec	xmm1,96[r8]
-	aesdec	xmm1,112[r8]
-	aesdec	xmm1,128[r8]
-	aesdec	xmm1,144[r8]
-	jb	DLAST_4_2
-	movdqu	xmm2,192[r8]
-	cmp	r9d,14
-	aesdec	xmm1,160[r8]
-	aesdec	xmm1,176[r8]
-	jb	DLAST_4_2
-	movdqu	xmm2,224[r8]
-	aesdec	xmm1,192[r8]
-	aesdec	xmm1,208[r8]
+        movdqu      xmm1, [rdi]
+        movdqa      xmm15, xmm1
+        add         rdi, 16
+        pxor        xmm1, [r8]
+        movdqu      xmm2, 160[r8]
+        cmp         r9d, 12
+        aesdec      xmm1, 16[r8]
+        aesdec      xmm1, 32[r8]
+        aesdec      xmm1, 48[r8]
+        aesdec      xmm1, 64[r8]
+        aesdec      xmm1, 80[r8]
+        aesdec      xmm1, 96[r8]
+        aesdec      xmm1, 112[r8]
+        aesdec      xmm1, 128[r8]
+        aesdec      xmm1, 144[r8]
+        jb          DLAST_4_2
+        movdqu      xmm2, 192[r8]
+        cmp         r9d, 14
+        aesdec      xmm1, 160[r8]
+        aesdec      xmm1, 176[r8]
+        jb          DLAST_4_2
+        movdqu      xmm2, 224[r8]
+        aesdec      xmm1, 192[r8]
+        aesdec      xmm1, 208[r8]
 DLAST_4_2:
-	aesdeclast	xmm1,xmm2
-	pxor	xmm1,xmm5
-	movdqa	xmm5,xmm15
-	movdqu	[rsi],xmm1
-
-	add	rsi,16
-	dec	r10
-	jne	DLOOP_4_2
+        aesdeclast  xmm1, xmm2
+        pxor        xmm1, xmm5
+        movdqa      xmm5, xmm15
+        movdqu      [rsi], xmm1
+        add         rsi, 16
+        dec         r10
+        jne         DLOOP_4_2
 DEND_4:
-	; restore non volatile rdi,rsi
-	mov rdi,rax
-	mov rsi,r11
-	; restore non volatile xmms from stack
-	movdqa xmm6, [rsp+0]
-	movdqa xmm7, [rsp+16]
-	movdqa xmm8, [rsp+32]
-	movdqa xmm9, [rsp+48]
-	movdqa xmm10, [rsp+64]
-	movdqa xmm11, [rsp+80]
-	movdqa xmm12, [rsp+96]
-	movdqa xmm15, [rsp+112]
-	add rsp,8+8*16 ; 8 = align stack , 8 xmm6-12,15 16 bytes each
-	ret
-AES_CBC_decrypt ENDP
+        ; restore non volatile rdi,rsi
+        mov         rdi, rax
+        mov         rsi, r11
+        ; restore non volatile xmms from stack
+        movdqa      xmm6, [rsp+0]
+        movdqa      xmm7, [rsp+16]
+        movdqa      xmm8, [rsp+32]
+        movdqa      xmm9, [rsp+48]
+        movdqa      xmm10, [rsp+64]
+        movdqa      xmm11, [rsp+80]
+        movdqa      xmm12, [rsp+96]
+        movdqa      xmm15, [rsp+112]
+        add         rsp, 8+8*16 ; 8 = align stack , 8 xmm6-12,15 16 bytes each
+        ret
+AES_CBC_decrypt_by4 ENDP
+
+
+; void AES_CBC_decrypt_by6(const unsigned char *in,
+;                          unsigned char *out,
+;                          unsigned char ivec[16],
+;                          unsigned long length,
+;                          const unsigned char *KS,
+;                          int nr)
+AES_CBC_decrypt_by6 PROC
+; parameter 1: rdi - in
+; parameter 2: rsi - out
+; parameter 3: rdx - ivec
+; parameter 4: rcx - length
+; parameter 5: r8  - KS
+; parameter 6: r9d - nr
+
+        ; save rdi and rsi to rax and r11, restore before ret
+        mov         rax, rdi
+        mov         r11, rsi
+        ; convert to what we had for att&t convention
+        mov         rdi, rcx
+        mov         rsi, rdx
+        mov         rdx, r8
+        mov         rcx, r9
+        mov         r8, [rsp+40]
+        mov         r9d, [rsp+48]
+        ; on microsoft xmm6-xmm15 are non volatile,
+        ; let's save on stack and restore at end
+        sub         rsp, 8+9*16  ; 8 = align stack , 9 xmm6-14 16 bytes each
+        movdqa      [rsp+0], xmm6
+        movdqa      [rsp+16], xmm7
+        movdqa      [rsp+32], xmm8
+        movdqa      [rsp+48], xmm9
+        movdqa      [rsp+64], xmm10
+        movdqa      [rsp+80], xmm11
+        movdqa      [rsp+96], xmm12
+        movdqa      [rsp+112], xmm13
+        movdqa      [rsp+128], xmm14
+        ; back to our original code, more or less
+        mov         r10, rcx
+        shr         rcx, 4
+        shl         r10, 60
+        je          DNO_PARTS_6
+        add         rcx, 1
+DNO_PARTS_6:
+        mov         r12, rax
+        mov         r13, rdx
+        mov         r14, rbx
+        mov         rdx, 0
+        mov         rax, rcx
+        mov         rbx, 6
+        div         rbx
+        mov         rcx, rax
+        mov         r10, rdx
+        mov         rax, r12
+        mov         rdx, r13
+        mov         rbx, r14
+        cmp         rcx, 0
+        movdqu      xmm7, [rdx]
+        je          DREMAINDER_6
+        sub         rsi, 96
+DLOOP_6:
+        movdqu      xmm1, [rdi]
+        movdqu      xmm2, 16[rdi]
+        movdqu      xmm3, 32[rdi]
+        movdqu      xmm4, 48[rdi]
+        movdqu      xmm5, 64[rdi]
+        movdqu      xmm6, 80[rdi]
+        movdqa      xmm8, [r8]
+        movdqa      xmm9, 16[r8]
+        movdqa      xmm10, 32[r8]
+        movdqa      xmm11, 48[r8]
+        pxor        xmm1, xmm8
+        pxor        xmm2, xmm8
+        pxor        xmm3, xmm8
+        pxor        xmm4, xmm8
+        pxor        xmm5, xmm8
+        pxor        xmm6, xmm8
+        aesdec      xmm1, xmm9
+        aesdec      xmm2, xmm9
+        aesdec      xmm3, xmm9
+        aesdec      xmm4, xmm9
+        aesdec      xmm5, xmm9
+        aesdec      xmm6, xmm9
+        aesdec      xmm1, xmm10
+        aesdec      xmm2, xmm10
+        aesdec      xmm3, xmm10
+        aesdec      xmm4, xmm10
+        aesdec      xmm5, xmm10
+        aesdec      xmm6, xmm10
+        aesdec      xmm1, xmm11
+        aesdec      xmm2, xmm11
+        aesdec      xmm3, xmm11
+        aesdec      xmm4, xmm11
+        aesdec      xmm5, xmm11
+        aesdec      xmm6, xmm11
+        movdqa      xmm8, 64[r8]
+        movdqa      xmm9, 80[r8]
+        movdqa      xmm10, 96[r8]
+        movdqa      xmm11, 112[r8]
+        aesdec      xmm1, xmm8
+        aesdec      xmm2, xmm8
+        aesdec      xmm3, xmm8
+        aesdec      xmm4, xmm8
+        aesdec      xmm5, xmm8
+        aesdec      xmm6, xmm8
+        aesdec      xmm1, xmm9
+        aesdec      xmm2, xmm9
+        aesdec      xmm3, xmm9
+        aesdec      xmm4, xmm9
+        aesdec      xmm5, xmm9
+        aesdec      xmm6, xmm9
+        aesdec      xmm1, xmm10
+        aesdec      xmm2, xmm10
+        aesdec      xmm3, xmm10
+        aesdec      xmm4, xmm10
+        aesdec      xmm5, xmm10
+        aesdec      xmm6, xmm10
+        aesdec      xmm1, xmm11
+        aesdec      xmm2, xmm11
+        aesdec      xmm3, xmm11
+        aesdec      xmm4, xmm11
+        aesdec      xmm5, xmm11
+        aesdec      xmm6, xmm11
+        movdqa      xmm8, 128[r8]
+        movdqa      xmm9, 144[r8]
+        movdqa      xmm10, 160[r8]
+        cmp         r9d, 12
+        aesdec      xmm1, xmm8
+        aesdec      xmm2, xmm8
+        aesdec      xmm3, xmm8
+        aesdec      xmm4, xmm8
+        aesdec      xmm5, xmm8
+        aesdec      xmm6, xmm8
+        aesdec      xmm1, xmm9
+        aesdec      xmm2, xmm9
+        aesdec      xmm3, xmm9
+        aesdec      xmm4, xmm9
+        aesdec      xmm5, xmm9
+        aesdec      xmm6, xmm9
+        jb          DLAST_6
+        movdqa      xmm8, 160[r8]
+        movdqa      xmm9, 176[r8]
+        movdqa      xmm10, 192[r8]
+        cmp         r9d, 14
+        aesdec      xmm1, xmm8
+        aesdec      xmm2, xmm8
+        aesdec      xmm3, xmm8
+        aesdec      xmm4, xmm8
+        aesdec      xmm5, xmm8
+        aesdec      xmm6, xmm8
+        aesdec      xmm1, xmm9
+        aesdec      xmm2, xmm9
+        aesdec      xmm3, xmm9
+        aesdec      xmm4, xmm9
+        aesdec      xmm5, xmm9
+        aesdec      xmm6, xmm9
+        jb          DLAST_6
+        movdqa      xmm8, 192[r8]
+        movdqa      xmm9, 208[r8]
+        movdqa      xmm10, 224[r8]
+        aesdec      xmm1, xmm8
+        aesdec      xmm2, xmm8
+        aesdec      xmm3, xmm8
+        aesdec      xmm4, xmm8
+        aesdec      xmm5, xmm8
+        aesdec      xmm6, xmm8
+        aesdec      xmm1, xmm9
+        aesdec      xmm2, xmm9
+        aesdec      xmm3, xmm9
+        aesdec      xmm4, xmm9
+        aesdec      xmm5, xmm9
+        aesdec      xmm6, xmm9
+DLAST_6:
+        add         rsi, 96
+        aesdeclast  xmm1, xmm10
+        aesdeclast  xmm2, xmm10
+        aesdeclast  xmm3, xmm10
+        aesdeclast  xmm4, xmm10
+        aesdeclast  xmm5, xmm10
+        aesdeclast  xmm6, xmm10
+        movdqu      xmm8, [rdi]
+        movdqu      xmm9, 16[rdi]
+        movdqu      xmm10, 32[rdi]
+        movdqu      xmm11, 48[rdi]
+        movdqu      xmm12, 64[rdi]
+        movdqu      xmm13, 80[rdi]
+        pxor        xmm1, xmm7
+        pxor        xmm2, xmm8
+        pxor        xmm3, xmm9
+        pxor        xmm4, xmm10
+        pxor        xmm5, xmm11
+        pxor        xmm6, xmm12
+        movdqu      xmm7, xmm13
+        movdqu      [rsi], xmm1
+        movdqu      16[rsi], xmm2
+        movdqu      32[rsi], xmm3
+        movdqu      48[rsi], xmm4
+        movdqu      64[rsi], xmm5
+        movdqu      80[rsi], xmm6
+        add         rdi, 96
+        dec         rcx
+        jne         DLOOP_6
+        add         rsi, 96
+DREMAINDER_6:
+        cmp         r10, 0
+        je          DEND_6
+DLOOP_6_2:
+        movdqu      xmm1, [rdi]
+        movdqa      xmm10, xmm1
+        add         rdi, 16
+        pxor        xmm1, [r8]
+        movdqu      xmm2, 160[r8]
+        cmp         r9d, 12
+        aesdec      xmm1, 16[r8]
+        aesdec      xmm1, 32[r8]
+        aesdec      xmm1, 48[r8]
+        aesdec      xmm1, 64[r8]
+        aesdec      xmm1, 80[r8]
+        aesdec      xmm1, 96[r8]
+        aesdec      xmm1, 112[r8]
+        aesdec      xmm1, 128[r8]
+        aesdec      xmm1, 144[r8]
+        jb          DLAST_6_2
+        movdqu      xmm2, 192[r8]
+        cmp         r9d, 14
+        aesdec      xmm1, 160[r8]
+        aesdec      xmm1, 176[r8]
+        jb          DLAST_6_2
+        movdqu      xmm2, 224[r8]
+        aesdec      xmm1, 192[r8]
+        aesdec      xmm1, 208[r8]
+DLAST_6_2:
+        aesdeclast  xmm1, xmm2
+        pxor        xmm1, xmm7
+        movdqa      xmm7, xmm10
+        movdqu      [rsi], xmm1
+        add         rsi, 16
+        dec         r10
+        jne         DLOOP_6_2
+DEND_6:
+        ; restore non volatile rdi,rsi
+        mov         rdi, rax
+        mov         rsi, r11
+        ; restore non volatile xmms from stack
+        movdqa      xmm6, [rsp+0]
+        movdqa      xmm7, [rsp+16]
+        movdqa      xmm8, [rsp+32]
+        movdqa      xmm9, [rsp+48]
+        movdqa      xmm10, [rsp+64]
+        movdqa      xmm11, [rsp+80]
+        movdqa      xmm12, [rsp+96]
+        movdqa      xmm13, [rsp+112]
+        movdqa      xmm14, [rsp+128]
+        add         rsp, 8+9*16 ; 8 = align stack , 9 xmm6-14 16 bytes each
+        ret
+AES_CBC_decrypt_by6 ENDP
+
+
+; void AES_CBC_decrypt_by8(const unsigned char *in,
+;                          unsigned char *out,
+;                          unsigned char ivec[16],
+;                          unsigned long length,
+;                          const unsigned char *KS,
+;                          int nr)
+AES_CBC_decrypt_by8 PROC
+; parameter 1: rdi - in
+; parameter 2: rsi - out
+; parameter 3: rdx - ivec
+; parameter 4: rcx - length
+; parameter 5: r8  - KS
+; parameter 6: r9d - nr
+
+        ; save rdi and rsi to rax and r11, restore before ret
+        mov         rax, rdi
+        mov         r11, rsi
+        ; convert to what we had for att&t convention
+        mov         rdi, rcx
+        mov         rsi, rdx
+        mov         rdx, r8
+        mov         rcx,r9
+        mov         r8, [rsp+40]
+        mov         r9d, [rsp+48]
+        ; on microsoft xmm6-xmm15 are non volatile,
+        ; let's save on stack and restore at end
+        sub         rsp, 8+8*16  ; 8 = align stack , 8 xmm6-13 16 bytes each
+        movdqa      [rsp+0], xmm6
+        movdqa      [rsp+16], xmm7
+        movdqa      [rsp+32], xmm8
+        movdqa      [rsp+48], xmm9
+        movdqa      [rsp+64], xmm10
+        movdqa      [rsp+80], xmm11
+        movdqa      [rsp+96], xmm12
+        movdqa      [rsp+112], xmm13
+        ; back to our original code, more or less
+        mov         r10, rcx
+        shr         rcx, 4
+        shl         r10, 60
+        je          DNO_PARTS_8
+        add         rcx, 1
+DNO_PARTS_8:
+        mov         r10, rcx
+        shl         r10, 61
+        shr         r10, 61
+        shr         rcx, 3
+        movdqu      xmm9, [rdx]
+        je          DREMAINDER_8
+        sub         rsi, 128
+DLOOP_8:
+        movdqu      xmm1, [rdi]
+        movdqu      xmm2, 16[rdi]
+        movdqu      xmm3, 32[rdi]
+        movdqu      xmm4, 48[rdi]
+        movdqu      xmm5, 64[rdi]
+        movdqu      xmm6, 80[rdi]
+        movdqu      xmm7, 96[rdi]
+        movdqu      xmm8, 112[rdi]
+        movdqa      xmm10, [r8]
+        movdqa      xmm11, 16[r8]
+        movdqa      xmm12, 32[r8]
+        movdqa      xmm13, 48[r8]
+        pxor        xmm1, xmm10
+        pxor        xmm2, xmm10
+        pxor        xmm3, xmm10
+        pxor        xmm4, xmm10
+        pxor        xmm5, xmm10
+        pxor        xmm6, xmm10
+        pxor        xmm7, xmm10
+        pxor        xmm8, xmm10
+        aesdec      xmm1, xmm11
+        aesdec      xmm2, xmm11
+        aesdec      xmm3, xmm11
+        aesdec      xmm4, xmm11
+        aesdec      xmm5, xmm11
+        aesdec      xmm6, xmm11
+        aesdec      xmm7, xmm11
+        aesdec      xmm8, xmm11
+        aesdec      xmm1, xmm12
+        aesdec      xmm2, xmm12
+        aesdec      xmm3, xmm12
+        aesdec      xmm4, xmm12
+        aesdec      xmm5, xmm12
+        aesdec      xmm6, xmm12
+        aesdec      xmm7, xmm12
+        aesdec      xmm8, xmm12
+        aesdec      xmm1, xmm13
+        aesdec      xmm2, xmm13
+        aesdec      xmm3, xmm13
+        aesdec      xmm4, xmm13
+        aesdec      xmm5, xmm13
+        aesdec      xmm6, xmm13
+        aesdec      xmm7, xmm13
+        aesdec      xmm8, xmm13
+        movdqa      xmm10, 64[r8]
+        movdqa      xmm11, 80[r8]
+        movdqa      xmm12, 96[r8]
+        movdqa      xmm13, 112[r8]
+        aesdec      xmm1, xmm10
+        aesdec      xmm2, xmm10
+        aesdec      xmm3, xmm10
+        aesdec      xmm4, xmm10
+        aesdec      xmm5, xmm10
+        aesdec      xmm6, xmm10
+        aesdec      xmm7, xmm10
+        aesdec      xmm8, xmm10
+        aesdec      xmm1, xmm11
+        aesdec      xmm2, xmm11
+        aesdec      xmm3, xmm11
+        aesdec      xmm4, xmm11
+        aesdec      xmm5, xmm11
+        aesdec      xmm6, xmm11
+        aesdec      xmm7, xmm11
+        aesdec      xmm8, xmm11
+        aesdec      xmm1, xmm12
+        aesdec      xmm2, xmm12
+        aesdec      xmm3, xmm12
+        aesdec      xmm4, xmm12
+        aesdec      xmm5, xmm12
+        aesdec      xmm6, xmm12
+        aesdec      xmm7, xmm12
+        aesdec      xmm8, xmm12
+        aesdec      xmm1, xmm13
+        aesdec      xmm2, xmm13
+        aesdec      xmm3, xmm13
+        aesdec      xmm4, xmm13
+        aesdec      xmm5, xmm13
+        aesdec      xmm6, xmm13
+        aesdec      xmm7, xmm13
+        aesdec      xmm8, xmm13
+        movdqa      xmm10, 128[r8]
+        movdqa      xmm11, 144[r8]
+        movdqa      xmm12, 160[r8]
+        cmp         r9d, 12
+        aesdec      xmm1, xmm10
+        aesdec      xmm2, xmm10
+        aesdec      xmm3, xmm10
+        aesdec      xmm4, xmm10
+        aesdec      xmm5, xmm10
+        aesdec      xmm6, xmm10
+        aesdec      xmm7, xmm10
+        aesdec      xmm8, xmm10
+        aesdec      xmm1, xmm11
+        aesdec      xmm2, xmm11
+        aesdec      xmm3, xmm11
+        aesdec      xmm4, xmm11
+        aesdec      xmm5, xmm11
+        aesdec      xmm6, xmm11
+        aesdec      xmm7, xmm11
+        aesdec      xmm8, xmm11
+        jb          DLAST_8
+        movdqa      xmm10, 160[r8]
+        movdqa      xmm11, 176[r8]
+        movdqa      xmm12, 192[r8]
+        cmp         r9d, 14
+        aesdec      xmm1, xmm10
+        aesdec      xmm2, xmm10
+        aesdec      xmm3, xmm10
+        aesdec      xmm4, xmm10
+        aesdec      xmm5, xmm10
+        aesdec      xmm6, xmm10
+        aesdec      xmm7, xmm10
+        aesdec      xmm8, xmm10
+        aesdec      xmm1, xmm11
+        aesdec      xmm2, xmm11
+        aesdec      xmm3, xmm11
+        aesdec      xmm4, xmm11
+        aesdec      xmm5, xmm11
+        aesdec      xmm6, xmm11
+        aesdec      xmm7, xmm11
+        aesdec      xmm8, xmm11
+        jb          DLAST_8
+        movdqa      xmm10, 192[r8]
+        movdqa      xmm11, 208[r8]
+        movdqa      xmm12, 224[r8]
+        aesdec      xmm1, xmm10
+        aesdec      xmm2, xmm10
+        aesdec      xmm3, xmm10
+        aesdec      xmm4, xmm10
+        aesdec      xmm5, xmm10
+        aesdec      xmm6, xmm10
+        aesdec      xmm7, xmm10
+        aesdec      xmm8, xmm10
+        aesdec      xmm1, xmm11
+        aesdec      xmm2, xmm11
+        aesdec      xmm3, xmm11
+        aesdec      xmm4, xmm11
+        aesdec      xmm5, xmm11
+        aesdec      xmm6, xmm11
+        aesdec      xmm7, xmm11
+        aesdec      xmm8, xmm11
+DLAST_8:
+        add         rsi, 128
+        aesdeclast  xmm1, xmm12
+        aesdeclast  xmm2, xmm12
+        aesdeclast  xmm3, xmm12
+        aesdeclast  xmm4, xmm12
+        aesdeclast  xmm5, xmm12
+        aesdeclast  xmm6, xmm12
+        aesdeclast  xmm7, xmm12
+        aesdeclast  xmm8, xmm12
+        movdqu      xmm10, [rdi]
+        movdqu      xmm11, 16[rdi]
+        movdqu      xmm12, 32[rdi]
+        movdqu      xmm13, 48[rdi]
+        pxor        xmm1, xmm9
+        pxor        xmm2, xmm10
+        pxor        xmm3, xmm11
+        pxor        xmm4, xmm12
+        pxor        xmm5, xmm13
+        movdqu      xmm10, 64[rdi]
+        movdqu      xmm11, 80[rdi]
+        movdqu      xmm12, 96[rdi]
+        movdqu      xmm9, 112[rdi]
+        pxor        xmm6, xmm10
+        pxor        xmm7, xmm11
+        pxor        xmm8, xmm12
+        movdqu      [rsi], xmm1
+        movdqu      16[rsi], xmm2
+        movdqu      32[rsi], xmm3
+        movdqu      48[rsi], xmm4
+        movdqu      64[rsi], xmm5
+        movdqu      80[rsi], xmm6
+        movdqu      96[rsi], xmm7
+        movdqu      112[rsi], xmm8
+        add         rdi, 128
+        dec         rcx
+        jne         DLOOP_8
+        add         rsi, 128
+DREMAINDER_8:
+        cmp         r10, 0 
+        je          DEND_8
+DLOOP_8_2:
+        movdqu      xmm1, [rdi]
+        movdqa      xmm10, xmm1
+        add         rdi, 16
+        pxor        xmm1, [r8]
+        movdqu      xmm2, 160[r8]
+        cmp         r9d, 12
+        aesdec      xmm1, 16[r8]
+        aesdec      xmm1, 32[r8]
+        aesdec      xmm1, 48[r8]
+        aesdec      xmm1, 64[r8]
+        aesdec      xmm1, 80[r8]
+        aesdec      xmm1, 96[r8]
+        aesdec      xmm1, 112[r8]
+        aesdec      xmm1, 128[r8]
+        aesdec      xmm1, 144[r8]
+        jb          DLAST_8_2
+        movdqu      xmm2, 192[r8]
+        cmp         r9d, 14
+        aesdec      xmm1, 160[r8]
+        aesdec      xmm1, 176[r8]
+        jb          DLAST_8_2
+        movdqu      xmm2, 224[r8]
+        aesdec      xmm1, 192[r8]
+        aesdec      xmm1, 208[r8]
+DLAST_8_2:
+        aesdeclast  xmm1, xmm2
+        pxor        xmm1, xmm9
+        movdqa      xmm9, xmm10
+        movdqu      [rsi], xmm1
+        add         rsi, 16
+        dec         r10
+        jne         DLOOP_8_2
+DEND_8:
+        ; restore non volatile rdi,rsi
+        mov         rdi, rax
+        mov         rsi, r11
+        ; restore non volatile xmms from stack
+        movdqa      xmm6, [rsp+0]
+        movdqa      xmm7, [rsp+16]
+        movdqa      xmm8, [rsp+32]
+        movdqa      xmm9, [rsp+48]
+        movdqa      xmm10, [rsp+64]
+        movdqa      xmm11, [rsp+80]
+        movdqa      xmm12, [rsp+96]
+        movdqa      xmm13, [rsp+112]
+        add         rsp, 8+8*16 ; 8 = align stack , 8 xmm6-13 16 bytes each
+        ret
+AES_CBC_decrypt_by8 ENDP
+
 
 ;	/*
 ;	AES_ECB_encrypt[const	,unsigned	char*in
diff --git a/wolfcrypt/src/aes_asm.s b/wolfcrypt/src/aes_asm.s
index 385a2d49a..ac67a09ee 100644
--- a/wolfcrypt/src/aes_asm.s
+++ b/wolfcrypt/src/aes_asm.s
@@ -86,18 +86,18 @@ jne	LOOP
 ret
 
 
-
+#if defined(WOLFSSL_AESNI_BY4)
 
 /*
-AES_CBC_decrypt (const unsigned char *in,
+AES_CBC_decrypt_by4 (const unsigned char *in,
   unsigned char *out,
   unsigned char ivec[16],
   unsigned long length,
   const unsigned char *KS,
   int nr)
 */
-.globl AES_CBC_decrypt
-AES_CBC_decrypt:
+.globl AES_CBC_decrypt_by4
+AES_CBC_decrypt_by4:
 # parameter 1: %rdi
 # parameter 2: %rsi
 # parameter 3: %rdx
@@ -105,165 +105,638 @@ AES_CBC_decrypt:
 # parameter 5: %r8
 # parameter 6: %r9d
 
-movq    %rcx, %r10
-shrq $4, %rcx
-shlq   $60, %r10
-je    DNO_PARTS_4
-addq    $1, %rcx
+        movq        %rcx, %r10
+        shrq        $4, %rcx
+        shlq        $60, %r10
+        je          DNO_PARTS_4
+        addq        $1, %rcx
 DNO_PARTS_4:
-movq   %rcx, %r10
-shlq    $62, %r10
-shrq  $62, %r10
-shrq  $2, %rcx
-movdqu (%rdx),%xmm5
-je DREMAINDER_4
-subq   $64, %rsi
+        movq        %rcx, %r10
+        shlq        $62, %r10
+        shrq        $62, %r10
+        shrq        $2, %rcx
+        movdqu      (%rdx),%xmm5
+        je          DREMAINDER_4
+        subq        $64, %rsi
 DLOOP_4:
-movdqu (%rdi), %xmm1
-movdqu  16(%rdi), %xmm2
-movdqu  32(%rdi), %xmm3
-movdqu  48(%rdi), %xmm4
-movdqa  %xmm1, %xmm6
-movdqa %xmm2, %xmm7
-movdqa %xmm3, %xmm8
-movdqa %xmm4, %xmm15
-movdqa    (%r8), %xmm9
-movdqa 16(%r8), %xmm10
-movdqa  32(%r8), %xmm11
-movdqa  48(%r8), %xmm12
-pxor    %xmm9, %xmm1
-pxor   %xmm9, %xmm2
-pxor   %xmm9, %xmm3
-
-pxor    %xmm9, %xmm4
-aesdec %xmm10, %xmm1
-aesdec    %xmm10, %xmm2
-aesdec    %xmm10, %xmm3
-aesdec    %xmm10, %xmm4
-aesdec    %xmm11, %xmm1
-aesdec    %xmm11, %xmm2
-aesdec    %xmm11, %xmm3
-aesdec    %xmm11, %xmm4
-aesdec    %xmm12, %xmm1
-aesdec    %xmm12, %xmm2
-aesdec    %xmm12, %xmm3
-aesdec    %xmm12, %xmm4
-movdqa    64(%r8), %xmm9
-movdqa   80(%r8), %xmm10
-movdqa  96(%r8), %xmm11
-movdqa  112(%r8), %xmm12
-aesdec %xmm9, %xmm1
-aesdec %xmm9, %xmm2
-aesdec %xmm9, %xmm3
-aesdec %xmm9, %xmm4
-aesdec %xmm10, %xmm1
-aesdec    %xmm10, %xmm2
-aesdec    %xmm10, %xmm3
-aesdec    %xmm10, %xmm4
-aesdec    %xmm11, %xmm1
-aesdec    %xmm11, %xmm2
-aesdec    %xmm11, %xmm3
-aesdec    %xmm11, %xmm4
-aesdec    %xmm12, %xmm1
-aesdec    %xmm12, %xmm2
-aesdec    %xmm12, %xmm3
-aesdec    %xmm12, %xmm4
-movdqa    128(%r8), %xmm9
-movdqa  144(%r8), %xmm10
-movdqa 160(%r8), %xmm11
-cmpl   $12, %r9d
-aesdec  %xmm9, %xmm1
-aesdec %xmm9, %xmm2
-aesdec %xmm9, %xmm3
-aesdec %xmm9, %xmm4
-aesdec %xmm10, %xmm1
-aesdec    %xmm10, %xmm2
-aesdec    %xmm10, %xmm3
-aesdec    %xmm10, %xmm4
-jb    DLAST_4
-movdqa  160(%r8), %xmm9
-movdqa  176(%r8), %xmm10
-movdqa 192(%r8), %xmm11
-cmpl   $14, %r9d
-aesdec  %xmm9, %xmm1
-aesdec %xmm9, %xmm2
-aesdec %xmm9, %xmm3
-aesdec %xmm9, %xmm4
-aesdec %xmm10, %xmm1
-aesdec    %xmm10, %xmm2
-aesdec    %xmm10, %xmm3
-aesdec    %xmm10, %xmm4
-jb    DLAST_4
-
-movdqa  192(%r8), %xmm9
-movdqa  208(%r8), %xmm10
-movdqa 224(%r8), %xmm11
-aesdec %xmm9, %xmm1
-aesdec %xmm9, %xmm2
-aesdec %xmm9, %xmm3
-aesdec %xmm9, %xmm4
-aesdec %xmm10, %xmm1
-aesdec    %xmm10, %xmm2
-aesdec    %xmm10, %xmm3
-aesdec    %xmm10, %xmm4
+        movdqu      (%rdi), %xmm1
+        movdqu      16(%rdi), %xmm2
+        movdqu      32(%rdi), %xmm3
+        movdqu      48(%rdi), %xmm4
+        movdqa      %xmm1, %xmm6
+        movdqa      %xmm2, %xmm7
+        movdqa      %xmm3, %xmm8
+        movdqa      %xmm4, %xmm15
+        movdqa      (%r8), %xmm9
+        movdqa      16(%r8), %xmm10
+        movdqa      32(%r8), %xmm11
+        movdqa      48(%r8), %xmm12
+        pxor        %xmm9, %xmm1
+        pxor        %xmm9, %xmm2
+        pxor        %xmm9, %xmm3
+        pxor        %xmm9, %xmm4
+        aesdec      %xmm10, %xmm1
+        aesdec      %xmm10, %xmm2
+        aesdec      %xmm10, %xmm3
+        aesdec      %xmm10, %xmm4
+        aesdec      %xmm11, %xmm1
+        aesdec      %xmm11, %xmm2
+        aesdec      %xmm11, %xmm3
+        aesdec      %xmm11, %xmm4
+        aesdec      %xmm12, %xmm1
+        aesdec      %xmm12, %xmm2
+        aesdec      %xmm12, %xmm3
+        aesdec      %xmm12, %xmm4
+        movdqa      64(%r8), %xmm9
+        movdqa      80(%r8), %xmm10
+        movdqa      96(%r8), %xmm11
+        movdqa      112(%r8), %xmm12
+        aesdec      %xmm9, %xmm1
+        aesdec      %xmm9, %xmm2
+        aesdec      %xmm9, %xmm3
+        aesdec      %xmm9, %xmm4
+        aesdec      %xmm10, %xmm1
+        aesdec      %xmm10, %xmm2
+        aesdec      %xmm10, %xmm3
+        aesdec      %xmm10, %xmm4
+        aesdec      %xmm11, %xmm1
+        aesdec      %xmm11, %xmm2
+        aesdec      %xmm11, %xmm3
+        aesdec      %xmm11, %xmm4
+        aesdec      %xmm12, %xmm1
+        aesdec      %xmm12, %xmm2
+        aesdec      %xmm12, %xmm3
+        aesdec      %xmm12, %xmm4
+        movdqa      128(%r8), %xmm9
+        movdqa      144(%r8), %xmm10
+        movdqa      160(%r8), %xmm11
+        cmpl        $12, %r9d
+        aesdec      %xmm9, %xmm1
+        aesdec      %xmm9, %xmm2
+        aesdec      %xmm9, %xmm3
+        aesdec      %xmm9, %xmm4
+        aesdec      %xmm10, %xmm1
+        aesdec      %xmm10, %xmm2
+        aesdec      %xmm10, %xmm3
+        aesdec      %xmm10, %xmm4
+        jb          DLAST_4
+        movdqa      160(%r8), %xmm9
+        movdqa      176(%r8), %xmm10
+        movdqa      192(%r8), %xmm11
+        cmpl        $14, %r9d
+        aesdec      %xmm9, %xmm1
+        aesdec      %xmm9, %xmm2
+        aesdec      %xmm9, %xmm3
+        aesdec      %xmm9, %xmm4
+        aesdec      %xmm10, %xmm1
+        aesdec      %xmm10, %xmm2
+        aesdec      %xmm10, %xmm3
+        aesdec      %xmm10, %xmm4
+        jb          DLAST_4
+        movdqa      192(%r8), %xmm9
+        movdqa      208(%r8), %xmm10
+        movdqa      224(%r8), %xmm11
+        aesdec      %xmm9, %xmm1
+        aesdec      %xmm9, %xmm2
+        aesdec      %xmm9, %xmm3
+        aesdec      %xmm9, %xmm4
+        aesdec      %xmm10, %xmm1
+        aesdec      %xmm10, %xmm2
+        aesdec      %xmm10, %xmm3
+        aesdec      %xmm10, %xmm4
 DLAST_4:
-addq   $64, %rdi
-addq    $64, %rsi
-decq  %rcx
-aesdeclast %xmm11, %xmm1
-aesdeclast %xmm11, %xmm2
-aesdeclast %xmm11, %xmm3
-aesdeclast %xmm11, %xmm4
-pxor   %xmm5 ,%xmm1
-pxor    %xmm6 ,%xmm2
-pxor   %xmm7 ,%xmm3
-pxor   %xmm8 ,%xmm4
-movdqu %xmm1, (%rsi)
-movdqu    %xmm2, 16(%rsi)
-movdqu  %xmm3, 32(%rsi)
-movdqu  %xmm4, 48(%rsi)
-movdqa  %xmm15,%xmm5
-jne    DLOOP_4
-addq    $64, %rsi
+        addq        $64, %rdi
+        addq        $64, %rsi
+        decq        %rcx
+        aesdeclast  %xmm11, %xmm1
+        aesdeclast  %xmm11, %xmm2
+        aesdeclast  %xmm11, %xmm3
+        aesdeclast  %xmm11, %xmm4
+        pxor        %xmm5, %xmm1
+        pxor        %xmm6, %xmm2
+        pxor        %xmm7, %xmm3
+        pxor        %xmm8, %xmm4
+        movdqu      %xmm1, (%rsi)
+        movdqu      %xmm2, 16(%rsi)
+        movdqu      %xmm3, 32(%rsi)
+        movdqu      %xmm4, 48(%rsi)
+        movdqa      %xmm15,%xmm5
+        jne         DLOOP_4
+        addq        $64, %rsi
 DREMAINDER_4:
-cmpq    $0, %r10
-je  DEND_4
+        cmpq        $0, %r10
+        je          DEND_4
 DLOOP_4_2:
-movdqu  (%rdi), %xmm1
-movdqa    %xmm1 ,%xmm15
-addq  $16, %rdi
-pxor  (%r8), %xmm1
-movdqu 160(%r8), %xmm2
-cmpl    $12, %r9d
-aesdec    16(%r8), %xmm1
-aesdec   32(%r8), %xmm1
-aesdec   48(%r8), %xmm1
-aesdec   64(%r8), %xmm1
-aesdec   80(%r8), %xmm1
-aesdec   96(%r8), %xmm1
-aesdec   112(%r8), %xmm1
-aesdec  128(%r8), %xmm1
-aesdec  144(%r8), %xmm1
-jb  DLAST_4_2
-movdqu    192(%r8), %xmm2
-cmpl    $14, %r9d
-aesdec    160(%r8), %xmm1
-aesdec  176(%r8), %xmm1
-jb  DLAST_4_2
-movdqu    224(%r8), %xmm2
-aesdec  192(%r8), %xmm1
-aesdec  208(%r8), %xmm1
+        movdqu      (%rdi), %xmm1
+        movdqa      %xmm1, %xmm15
+        addq        $16, %rdi
+        pxor        (%r8), %xmm1
+        movdqu      160(%r8), %xmm2
+        cmpl        $12, %r9d
+        aesdec      16(%r8), %xmm1
+        aesdec      32(%r8), %xmm1
+        aesdec      48(%r8), %xmm1
+        aesdec      64(%r8), %xmm1
+        aesdec      80(%r8), %xmm1
+        aesdec      96(%r8), %xmm1
+        aesdec      112(%r8), %xmm1
+        aesdec      128(%r8), %xmm1
+        aesdec      144(%r8), %xmm1
+        jb          DLAST_4_2
+        movdqu      192(%r8), %xmm2
+        cmpl        $14, %r9d
+        aesdec      160(%r8), %xmm1
+        aesdec      176(%r8), %xmm1
+        jb          DLAST_4_2
+        movdqu      224(%r8), %xmm2
+        aesdec      192(%r8), %xmm1
+        aesdec      208(%r8), %xmm1
 DLAST_4_2:
-aesdeclast %xmm2, %xmm1
-pxor    %xmm5, %xmm1
-movdqa %xmm15, %xmm5
-movdqu    %xmm1, (%rsi)
-
-addq    $16, %rsi
-decq    %r10
-jne DLOOP_4_2
+        aesdeclast  %xmm2, %xmm1
+        pxor        %xmm5, %xmm1
+        movdqa      %xmm15, %xmm5
+        movdqu      %xmm1, (%rsi)
+        addq        $16, %rsi
+        decq        %r10
+        jne         DLOOP_4_2
 DEND_4:
-ret
+        ret
+
+#elif defined(WOLFSSL_AESNI_BY6)
+
+/*
+AES_CBC_decrypt_by6 (const unsigned char *in,
+  unsigned char *out,
+  unsigned char ivec[16],
+  unsigned long length,
+  const unsigned char *KS,
+  int nr)
+*/
+.globl AES_CBC_decrypt_by6
+AES_CBC_decrypt_by6:
+# parameter 1: %rdi - in
+# parameter 2: %rsi - out
+# parameter 3: %rdx - ivec
+# parameter 4: %rcx - length
+# parameter 5: %r8  - KS
+# parameter 6: %r9d - nr
+
+        movq        %rcx, %r10
+        shrq        $4, %rcx
+        shlq        $60, %r10
+        je          DNO_PARTS_6
+        addq        $1, %rcx
+DNO_PARTS_6:
+        movq        %rax, %r12
+        movq        %rdx, %r13
+        movq        %rbx, %r14
+        movq        $0, %rdx
+        movq        %rcx, %rax
+        movq        $6, %rbx
+        div         %rbx
+        movq        %rax, %rcx
+        movq        %rdx, %r10
+        movq        %r12, %rax
+        movq        %r13, %rdx
+        movq        %r14, %rbx
+        cmpq        $0, %rcx
+        movdqu      (%rdx), %xmm7
+        je          DREMAINDER_6
+        subq        $96, %rsi
+DLOOP_6:
+        movdqu      (%rdi), %xmm1
+        movdqu      16(%rdi), %xmm2
+        movdqu      32(%rdi), %xmm3
+        movdqu      48(%rdi), %xmm4
+        movdqu      64(%rdi), %xmm5
+        movdqu      80(%rdi), %xmm6
+        movdqa      (%r8), %xmm8
+        movdqa      16(%r8), %xmm9
+        movdqa      32(%r8), %xmm10
+        movdqa      48(%r8), %xmm11
+        pxor        %xmm8, %xmm1
+        pxor        %xmm8, %xmm2
+        pxor        %xmm8, %xmm3
+        pxor        %xmm8, %xmm4
+        pxor        %xmm8, %xmm5
+        pxor        %xmm8, %xmm6
+        aesdec      %xmm9, %xmm1
+        aesdec      %xmm9, %xmm2
+        aesdec      %xmm9, %xmm3
+        aesdec      %xmm9, %xmm4
+        aesdec      %xmm9, %xmm5
+        aesdec      %xmm9, %xmm6
+        aesdec      %xmm10, %xmm1
+        aesdec      %xmm10, %xmm2
+        aesdec      %xmm10, %xmm3
+        aesdec      %xmm10, %xmm4
+        aesdec      %xmm10, %xmm5
+        aesdec      %xmm10, %xmm6
+        aesdec      %xmm11, %xmm1
+        aesdec      %xmm11, %xmm2
+        aesdec      %xmm11, %xmm3
+        aesdec      %xmm11, %xmm4
+        aesdec      %xmm11, %xmm5
+        aesdec      %xmm11, %xmm6
+        movdqa      64(%r8), %xmm8
+        movdqa      80(%r8), %xmm9
+        movdqa      96(%r8), %xmm10
+        movdqa      112(%r8), %xmm11
+        aesdec      %xmm8, %xmm1
+        aesdec      %xmm8, %xmm2
+        aesdec      %xmm8, %xmm3
+        aesdec      %xmm8, %xmm4
+        aesdec      %xmm8, %xmm5
+        aesdec      %xmm8, %xmm6
+        aesdec      %xmm9, %xmm1
+        aesdec      %xmm9, %xmm2
+        aesdec      %xmm9, %xmm3
+        aesdec      %xmm9, %xmm4
+        aesdec      %xmm9, %xmm5
+        aesdec      %xmm9, %xmm6
+        aesdec      %xmm10, %xmm1
+        aesdec      %xmm10, %xmm2
+        aesdec      %xmm10, %xmm3
+        aesdec      %xmm10, %xmm4
+        aesdec      %xmm10, %xmm5
+        aesdec      %xmm10, %xmm6
+        aesdec      %xmm11, %xmm1
+        aesdec      %xmm11, %xmm2
+        aesdec      %xmm11, %xmm3
+        aesdec      %xmm11, %xmm4
+        aesdec      %xmm11, %xmm5
+        aesdec      %xmm11, %xmm6
+        movdqa      128(%r8), %xmm8
+        movdqa      144(%r8), %xmm9
+        movdqa      160(%r8), %xmm10
+        cmpl        $12, %r9d
+        aesdec      %xmm8, %xmm1
+        aesdec      %xmm8, %xmm2
+        aesdec      %xmm8, %xmm3
+        aesdec      %xmm8, %xmm4
+        aesdec      %xmm8, %xmm5
+        aesdec      %xmm8, %xmm6
+        aesdec      %xmm9, %xmm1
+        aesdec      %xmm9, %xmm2
+        aesdec      %xmm9, %xmm3
+        aesdec      %xmm9, %xmm4
+        aesdec      %xmm9, %xmm5
+        aesdec      %xmm9, %xmm6
+        jb          DLAST_6
+        movdqa      160(%r8), %xmm8
+        movdqa      176(%r8), %xmm9
+        movdqa      192(%r8), %xmm10
+        cmpl        $14, %r9d
+        aesdec      %xmm8, %xmm1
+        aesdec      %xmm8, %xmm2
+        aesdec      %xmm8, %xmm3
+        aesdec      %xmm8, %xmm4
+        aesdec      %xmm8, %xmm5
+        aesdec      %xmm8, %xmm6
+        aesdec      %xmm9, %xmm1
+        aesdec      %xmm9, %xmm2
+        aesdec      %xmm9, %xmm3
+        aesdec      %xmm9, %xmm4
+        aesdec      %xmm9, %xmm5
+        aesdec      %xmm9, %xmm6
+        jb          DLAST_6
+        movdqa      192(%r8), %xmm8
+        movdqa      208(%r8), %xmm9
+        movdqa      224(%r8), %xmm10
+        aesdec      %xmm8, %xmm1
+        aesdec      %xmm8, %xmm2
+        aesdec      %xmm8, %xmm3
+        aesdec      %xmm8, %xmm4
+        aesdec      %xmm8, %xmm5
+        aesdec      %xmm8, %xmm6
+        aesdec      %xmm9, %xmm1
+        aesdec      %xmm9, %xmm2
+        aesdec      %xmm9, %xmm3
+        aesdec      %xmm9, %xmm4
+        aesdec      %xmm9, %xmm5
+        aesdec      %xmm9, %xmm6
+DLAST_6:
+        addq        $96, %rsi
+        aesdeclast  %xmm10, %xmm1
+        aesdeclast  %xmm10, %xmm2
+        aesdeclast  %xmm10, %xmm3
+        aesdeclast  %xmm10, %xmm4
+        aesdeclast  %xmm10, %xmm5
+        aesdeclast  %xmm10, %xmm6
+        movdqu      (%rdi), %xmm8
+        movdqu      16(%rdi), %xmm9
+        movdqu      32(%rdi), %xmm10
+        movdqu      48(%rdi), %xmm11
+        movdqu      64(%rdi), %xmm12
+        movdqu      80(%rdi), %xmm13
+        pxor        %xmm7, %xmm1
+        pxor        %xmm8, %xmm2
+        pxor        %xmm9, %xmm3
+        pxor        %xmm10, %xmm4
+        pxor        %xmm11, %xmm5
+        pxor        %xmm12, %xmm6
+        movdqu      %xmm13, %xmm7
+        movdqu      %xmm1, (%rsi)
+        movdqu      %xmm2, 16(%rsi)
+        movdqu      %xmm3, 32(%rsi)
+        movdqu      %xmm4, 48(%rsi)
+        movdqu      %xmm5, 64(%rsi)
+        movdqu      %xmm6, 80(%rsi)
+        addq        $96, %rdi
+        decq        %rcx
+        jne         DLOOP_6
+        addq        $96, %rsi
+DREMAINDER_6:
+        cmpq        $0, %r10
+        je          DEND_6
+DLOOP_6_2:
+        movdqu      (%rdi), %xmm1
+        movdqa      %xmm1, %xmm10
+        addq        $16, %rdi
+        pxor        (%r8), %xmm1
+        movdqu      160(%r8), %xmm2
+        cmpl        $12, %r9d
+        aesdec      16(%r8), %xmm1
+        aesdec      32(%r8), %xmm1
+        aesdec      48(%r8), %xmm1
+        aesdec      64(%r8), %xmm1
+        aesdec      80(%r8), %xmm1
+        aesdec      96(%r8), %xmm1
+        aesdec      112(%r8), %xmm1
+        aesdec      128(%r8), %xmm1
+        aesdec      144(%r8), %xmm1
+        jb          DLAST_6_2
+        movdqu      192(%r8), %xmm2
+        cmpl        $14, %r9d
+        aesdec      160(%r8), %xmm1
+        aesdec      176(%r8), %xmm1
+        jb          DLAST_6_2
+        movdqu      224(%r8), %xmm2
+        aesdec      192(%r8), %xmm1
+        aesdec      208(%r8), %xmm1
+DLAST_6_2:
+        aesdeclast  %xmm2, %xmm1
+        pxor        %xmm7, %xmm1
+        movdqa      %xmm10, %xmm7
+        movdqu      %xmm1, (%rsi)
+        addq        $16, %rsi
+        decq        %r10
+        jne         DLOOP_6_2
+DEND_6:
+        ret
+
+#else /* WOLFSSL_AESNI_BYx */
+
+/*
+AES_CBC_decrypt_by8 (const unsigned char *in,
+  unsigned char *out,
+  unsigned char ivec[16],
+  unsigned long length,
+  const unsigned char *KS,
+  int nr)
+*/
+.globl AES_CBC_decrypt_by8
+AES_CBC_decrypt_by8:
+# parameter 1: %rdi - in
+# parameter 2: %rsi - out
+# parameter 3: %rdx - ivec
+# parameter 4: %rcx - length
+# parameter 5: %r8  - KS
+# parameter 6: %r9d - nr
+
+        movq        %rcx, %r10
+        shrq        $4, %rcx
+        shlq        $60, %r10
+        je          DNO_PARTS_8
+        addq        $1, %rcx
+DNO_PARTS_8:
+        movq        %rcx, %r10
+        shlq        $61, %r10
+        shrq        $61, %r10
+        shrq        $3, %rcx
+        movdqu      (%rdx), %xmm9
+        je          DREMAINDER_8
+        subq        $128, %rsi
+DLOOP_8:
+        movdqu      (%rdi), %xmm1
+        movdqu      16(%rdi), %xmm2
+        movdqu      32(%rdi), %xmm3
+        movdqu      48(%rdi), %xmm4
+        movdqu      64(%rdi), %xmm5
+        movdqu      80(%rdi), %xmm6
+        movdqu      96(%rdi), %xmm7
+        movdqu      112(%rdi), %xmm8
+        movdqa      (%r8), %xmm10
+        movdqa      16(%r8), %xmm11
+        movdqa      32(%r8), %xmm12
+        movdqa      48(%r8), %xmm13
+        pxor        %xmm10, %xmm1
+        pxor        %xmm10, %xmm2
+        pxor        %xmm10, %xmm3
+        pxor        %xmm10, %xmm4
+        pxor        %xmm10, %xmm5
+        pxor        %xmm10, %xmm6
+        pxor        %xmm10, %xmm7
+        pxor        %xmm10, %xmm8
+        aesdec      %xmm11, %xmm1
+        aesdec      %xmm11, %xmm2
+        aesdec      %xmm11, %xmm3
+        aesdec      %xmm11, %xmm4
+        aesdec      %xmm11, %xmm5
+        aesdec      %xmm11, %xmm6
+        aesdec      %xmm11, %xmm7
+        aesdec      %xmm11, %xmm8
+        aesdec      %xmm12, %xmm1
+        aesdec      %xmm12, %xmm2
+        aesdec      %xmm12, %xmm3
+        aesdec      %xmm12, %xmm4
+        aesdec      %xmm12, %xmm5
+        aesdec      %xmm12, %xmm6
+        aesdec      %xmm12, %xmm7
+        aesdec      %xmm12, %xmm8
+        aesdec      %xmm13, %xmm1
+        aesdec      %xmm13, %xmm2
+        aesdec      %xmm13, %xmm3
+        aesdec      %xmm13, %xmm4
+        aesdec      %xmm13, %xmm5
+        aesdec      %xmm13, %xmm6
+        aesdec      %xmm13, %xmm7
+        aesdec      %xmm13, %xmm8
+        movdqa      64(%r8), %xmm10
+        movdqa      80(%r8), %xmm11
+        movdqa      96(%r8), %xmm12
+        movdqa      112(%r8), %xmm13
+        aesdec      %xmm10, %xmm1
+        aesdec      %xmm10, %xmm2
+        aesdec      %xmm10, %xmm3
+        aesdec      %xmm10, %xmm4
+        aesdec      %xmm10, %xmm5
+        aesdec      %xmm10, %xmm6
+        aesdec      %xmm10, %xmm7
+        aesdec      %xmm10, %xmm8
+        aesdec      %xmm11, %xmm1
+        aesdec      %xmm11, %xmm2
+        aesdec      %xmm11, %xmm3
+        aesdec      %xmm11, %xmm4
+        aesdec      %xmm11, %xmm5
+        aesdec      %xmm11, %xmm6
+        aesdec      %xmm11, %xmm7
+        aesdec      %xmm11, %xmm8
+        aesdec      %xmm12, %xmm1
+        aesdec      %xmm12, %xmm2
+        aesdec      %xmm12, %xmm3
+        aesdec      %xmm12, %xmm4
+        aesdec      %xmm12, %xmm5
+        aesdec      %xmm12, %xmm6
+        aesdec      %xmm12, %xmm7
+        aesdec      %xmm12, %xmm8
+        aesdec      %xmm13, %xmm1
+        aesdec      %xmm13, %xmm2
+        aesdec      %xmm13, %xmm3
+        aesdec      %xmm13, %xmm4
+        aesdec      %xmm13, %xmm5
+        aesdec      %xmm13, %xmm6
+        aesdec      %xmm13, %xmm7
+        aesdec      %xmm13, %xmm8
+        movdqa      128(%r8), %xmm10
+        movdqa      144(%r8), %xmm11
+        movdqa      160(%r8), %xmm12
+        cmpl        $12, %r9d
+        aesdec      %xmm10, %xmm1
+        aesdec      %xmm10, %xmm2
+        aesdec      %xmm10, %xmm3
+        aesdec      %xmm10, %xmm4
+        aesdec      %xmm10, %xmm5
+        aesdec      %xmm10, %xmm6
+        aesdec      %xmm10, %xmm7
+        aesdec      %xmm10, %xmm8
+        aesdec      %xmm11, %xmm1
+        aesdec      %xmm11, %xmm2
+        aesdec      %xmm11, %xmm3
+        aesdec      %xmm11, %xmm4
+        aesdec      %xmm11, %xmm5
+        aesdec      %xmm11, %xmm6
+        aesdec      %xmm11, %xmm7
+        aesdec      %xmm11, %xmm8
+        jb          DLAST_8
+        movdqa      160(%r8), %xmm10
+        movdqa      176(%r8), %xmm11
+        movdqa      192(%r8), %xmm12
+        cmpl        $14, %r9d
+        aesdec      %xmm10, %xmm1
+        aesdec      %xmm10, %xmm2
+        aesdec      %xmm10, %xmm3
+        aesdec      %xmm10, %xmm4
+        aesdec      %xmm10, %xmm5
+        aesdec      %xmm10, %xmm6
+        aesdec      %xmm10, %xmm7
+        aesdec      %xmm10, %xmm8
+        aesdec      %xmm11, %xmm1
+        aesdec      %xmm11, %xmm2
+        aesdec      %xmm11, %xmm3
+        aesdec      %xmm11, %xmm4
+        aesdec      %xmm11, %xmm5
+        aesdec      %xmm11, %xmm6
+        aesdec      %xmm11, %xmm7
+        aesdec      %xmm11, %xmm8
+        jb          DLAST_8
+        movdqa      192(%r8), %xmm10
+        movdqa      208(%r8), %xmm11
+        movdqa      224(%r8), %xmm12
+        aesdec      %xmm10, %xmm1
+        aesdec      %xmm10, %xmm2
+        aesdec      %xmm10, %xmm3
+        aesdec      %xmm10, %xmm4
+        aesdec      %xmm10, %xmm5
+        aesdec      %xmm10, %xmm6
+        aesdec      %xmm10, %xmm7
+        aesdec      %xmm10, %xmm8
+        aesdec      %xmm11, %xmm1
+        aesdec      %xmm11, %xmm2
+        aesdec      %xmm11, %xmm3
+        aesdec      %xmm11, %xmm4
+        aesdec      %xmm11, %xmm5
+        aesdec      %xmm11, %xmm6
+        aesdec      %xmm11, %xmm7
+        aesdec      %xmm11, %xmm8
+DLAST_8:
+        addq        $128, %rsi
+        aesdeclast  %xmm12, %xmm1
+        aesdeclast  %xmm12, %xmm2
+        aesdeclast  %xmm12, %xmm3
+        aesdeclast  %xmm12, %xmm4
+        aesdeclast  %xmm12, %xmm5
+        aesdeclast  %xmm12, %xmm6
+        aesdeclast  %xmm12, %xmm7
+        aesdeclast  %xmm12, %xmm8
+        movdqu      (%rdi), %xmm10
+        movdqu      16(%rdi), %xmm11
+        movdqu      32(%rdi), %xmm12
+        movdqu      48(%rdi), %xmm13
+        pxor        %xmm9, %xmm1
+        pxor        %xmm10, %xmm2
+        pxor        %xmm11, %xmm3
+        pxor        %xmm12, %xmm4
+        pxor        %xmm13, %xmm5
+        movdqu      64(%rdi), %xmm10
+        movdqu      80(%rdi), %xmm11
+        movdqu      96(%rdi), %xmm12
+        movdqu      112(%rdi), %xmm9
+        pxor        %xmm10, %xmm6
+        pxor        %xmm11, %xmm7
+        pxor        %xmm12, %xmm8
+        movdqu      %xmm1, (%rsi)
+        movdqu      %xmm2, 16(%rsi)
+        movdqu      %xmm3, 32(%rsi)
+        movdqu      %xmm4, 48(%rsi)
+        movdqu      %xmm5, 64(%rsi)
+        movdqu      %xmm6, 80(%rsi)
+        movdqu      %xmm7, 96(%rsi)
+        movdqu      %xmm8, 112(%rsi)
+        addq        $128, %rdi
+        decq        %rcx
+        jne         DLOOP_8
+        addq        $128, %rsi
+DREMAINDER_8:
+        cmpq        $0, %r10
+        je          DEND_8
+DLOOP_8_2:
+        movdqu      (%rdi), %xmm1
+        movdqa      %xmm1, %xmm10
+        addq        $16, %rdi
+        pxor        (%r8), %xmm1
+        movdqu      160(%r8), %xmm2
+        cmpl        $12, %r9d
+        aesdec      16(%r8), %xmm1
+        aesdec      32(%r8), %xmm1
+        aesdec      48(%r8), %xmm1
+        aesdec      64(%r8), %xmm1
+        aesdec      80(%r8), %xmm1
+        aesdec      96(%r8), %xmm1
+        aesdec      112(%r8), %xmm1
+        aesdec      128(%r8), %xmm1
+        aesdec      144(%r8), %xmm1
+        jb          DLAST_8_2
+        movdqu      192(%r8), %xmm2
+        cmpl        $14, %r9d
+        aesdec      160(%r8), %xmm1
+        aesdec      176(%r8), %xmm1
+        jb          DLAST_8_2
+        movdqu      224(%r8), %xmm2
+        aesdec      192(%r8), %xmm1
+        aesdec      208(%r8), %xmm1
+DLAST_8_2:
+        aesdeclast  %xmm2, %xmm1
+        pxor        %xmm9, %xmm1
+        movdqa      %xmm10, %xmm9
+        movdqu      %xmm1, (%rsi)
+        addq        $16, %rsi
+        decq        %r10
+        jne         DLOOP_8_2
+DEND_8:
+        ret
+
+#endif /* WOLFSSL_AESNI_BYx */
 
 
 /*
diff --git a/wolfcrypt/test/test.c b/wolfcrypt/test/test.c
index 6930ed96c..9e831323c 100644
--- a/wolfcrypt/test/test.c
+++ b/wolfcrypt/test/test.c
@@ -2669,6 +2669,94 @@ int aes_test(void)
     if (memcmp(cipher, verify, AES_BLOCK_SIZE))
         return -61;
 
+#if defined(WOLFSSL_AESNI) && defined(HAVE_AES_DECRYPT)
+    {
+        const byte bigMsg[] = {
+            /* "All work and no play makes Jack a dull boy. " */
+            0x41,0x6c,0x6c,0x20,0x77,0x6f,0x72,0x6b,
+            0x20,0x61,0x6e,0x64,0x20,0x6e,0x6f,0x20,
+            0x70,0x6c,0x61,0x79,0x20,0x6d,0x61,0x6b,
+            0x65,0x73,0x20,0x4a,0x61,0x63,0x6b,0x20,
+            0x61,0x20,0x64,0x75,0x6c,0x6c,0x20,0x62,
+            0x6f,0x79,0x2e,0x20,0x41,0x6c,0x6c,0x20,
+            0x77,0x6f,0x72,0x6b,0x20,0x61,0x6e,0x64,
+            0x20,0x6e,0x6f,0x20,0x70,0x6c,0x61,0x79,
+            0x20,0x6d,0x61,0x6b,0x65,0x73,0x20,0x4a,
+            0x61,0x63,0x6b,0x20,0x61,0x20,0x64,0x75,
+            0x6c,0x6c,0x20,0x62,0x6f,0x79,0x2e,0x20,
+            0x41,0x6c,0x6c,0x20,0x77,0x6f,0x72,0x6b,
+            0x20,0x61,0x6e,0x64,0x20,0x6e,0x6f,0x20,
+            0x70,0x6c,0x61,0x79,0x20,0x6d,0x61,0x6b,
+            0x65,0x73,0x20,0x4a,0x61,0x63,0x6b,0x20,
+            0x61,0x20,0x64,0x75,0x6c,0x6c,0x20,0x62,
+            0x6f,0x79,0x2e,0x20,0x41,0x6c,0x6c,0x20,
+            0x77,0x6f,0x72,0x6b,0x20,0x61,0x6e,0x64,
+            0x20,0x6e,0x6f,0x20,0x70,0x6c,0x61,0x79,
+            0x20,0x6d,0x61,0x6b,0x65,0x73,0x20,0x4a,
+            0x61,0x63,0x6b,0x20,0x61,0x20,0x64,0x75,
+            0x6c,0x6c,0x20,0x62,0x6f,0x79,0x2e,0x20,
+            0x41,0x6c,0x6c,0x20,0x77,0x6f,0x72,0x6b,
+            0x20,0x61,0x6e,0x64,0x20,0x6e,0x6f,0x20,
+            0x70,0x6c,0x61,0x79,0x20,0x6d,0x61,0x6b,
+            0x65,0x73,0x20,0x4a,0x61,0x63,0x6b,0x20,
+            0x61,0x20,0x64,0x75,0x6c,0x6c,0x20,0x62,
+            0x6f,0x79,0x2e,0x20,0x41,0x6c,0x6c,0x20,
+            0x77,0x6f,0x72,0x6b,0x20,0x61,0x6e,0x64,
+            0x20,0x6e,0x6f,0x20,0x70,0x6c,0x61,0x79,
+            0x20,0x6d,0x61,0x6b,0x65,0x73,0x20,0x4a,
+            0x61,0x63,0x6b,0x20,0x61,0x20,0x64,0x75,
+            0x6c,0x6c,0x20,0x62,0x6f,0x79,0x2e,0x20,
+            0x41,0x6c,0x6c,0x20,0x77,0x6f,0x72,0x6b,
+            0x20,0x61,0x6e,0x64,0x20,0x6e,0x6f,0x20,
+            0x70,0x6c,0x61,0x79,0x20,0x6d,0x61,0x6b,
+            0x65,0x73,0x20,0x4a,0x61,0x63,0x6b,0x20,
+            0x61,0x20,0x64,0x75,0x6c,0x6c,0x20,0x62,
+            0x6f,0x79,0x2e,0x20,0x41,0x6c,0x6c,0x20,
+            0x77,0x6f,0x72,0x6b,0x20,0x61,0x6e,0x64,
+            0x20,0x6e,0x6f,0x20,0x70,0x6c,0x61,0x79,
+            0x20,0x6d,0x61,0x6b,0x65,0x73,0x20,0x4a,
+            0x61,0x63,0x6b,0x20,0x61,0x20,0x64,0x75,
+            0x6c,0x6c,0x20,0x62,0x6f,0x79,0x2e,0x20,
+            0x41,0x6c,0x6c,0x20,0x77,0x6f,0x72,0x6b,
+            0x20,0x61,0x6e,0x64,0x20,0x6e,0x6f,0x20,
+            0x70,0x6c,0x61,0x79,0x20,0x6d,0x61,0x6b,
+            0x65,0x73,0x20,0x4a,0x61,0x63,0x6b,0x20
+        };
+        const byte bigKey[] = "0123456789abcdeffedcba9876543210";
+        byte bigCipher[sizeof(bigMsg)];
+        byte bigPlain[sizeof(bigMsg)];
+        word32 keySz, msgSz;
+
+        /* Iterate from one AES_BLOCK_SIZE of bigMsg through the whole
+         * message by AES_BLOCK_SIZE for each size of AES key. */
+        for (keySz = 16; keySz <= 32; keySz += 8) {
+            for (msgSz = AES_BLOCK_SIZE;
+                 msgSz <= sizeof(bigMsg);
+                 msgSz += AES_BLOCK_SIZE) {
+
+                memset(bigCipher, 0, sizeof(bigCipher));
+                memset(bigPlain, 0, sizeof(bigPlain));
+                ret = wc_AesSetKey(&enc, bigKey, keySz, iv, AES_ENCRYPTION);
+                if (ret != 0)
+                    return -1030;
+                ret = wc_AesSetKey(&dec, bigKey, keySz, iv, AES_DECRYPTION);
+                if (ret != 0)
+                    return -1031;
+
+                ret = wc_AesCbcEncrypt(&enc, bigCipher, bigMsg, msgSz);
+                if (ret != 0)
+                    return -1032;
+                ret = wc_AesCbcDecrypt(&dec, bigPlain, bigCipher, msgSz);
+                if (ret != 0)
+                    return -1033;
+
+                if (memcmp(bigPlain, bigMsg, msgSz))
+                    return -1034;
+            }
+        }
+    }
+#endif /* WOLFSSL_AESNI HAVE_AES_DECRYPT */
+
 #ifdef HAVE_CAVIUM
         wc_AesFreeCavium(&enc);
         wc_AesFreeCavium(&dec);