feedc0de/wolfssl
1
0
Fork 0
forked from wolfSSL/wolfssl
Code Pull Requests Activity

Merge pull request #593 from JacobBarthelmeh/ARMv8

Browse Source 
ARMv8 : sanity checks
This commit is contained in:
toddouska
2016-10-12 11:23:27 -07:00
committed by GitHub
parent 11102b6726 eb9161d8a7
commit ffb2a8ff12
3 changed files with 395 additions and 356 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
wolfcrypt/benchmark/benchmark.c
View File

@ -120,7 +120,6 @@
#include <sys/syscall.h>
#include <unistd.h>
#define HAVE_GET_CYCLES
static word64 begin_cycles;
static word64 total_cycles;
static int cycles = -1;
@ -2534,7 +2533,7 @@ void bench_ed25519KeySign(void)
#endif /* _WIN32 */
#ifdef HAVE_GET_CYCLES
#if defined(HAVE_GET_CYCLES)
static INLINE word64 get_intel_cycles(void)
{

42
wolfcrypt/src/port/arm/armv8-aes.c
View File

@ -2511,10 +2511,10 @@ int wc_AesGcmEncrypt(Aes* aes, byte* out, const byte* in, word32 sz,
{
/* sanity checks */
if (aes == NULL || (iv == NULL && ivSz > 0) ||
(authTag == NULL && authTagSz > 0) ||
(authIn == NULL && authInSz > 0) ||
(authTag == NULL) ||
(authIn == NULL) ||
(in == NULL && sz > 0) ||
(out == NULL && authTag == NULL)) {
(out == NULL && sz > 0)) {
WOLFSSL_MSG("a NULL parameter passed in when size is larger than 0");
return BAD_FUNC_ARG;
}
@ -2571,10 +2571,10 @@ int wc_AesGcmDecrypt(Aes* aes, byte* out, const byte* in, word32 sz,
/* sanity checks */
if (aes == NULL || (iv == NULL && ivSz > 0) ||
(authTag == NULL && authTagSz > 0) ||
(authIn == NULL && authInSz > 0) ||
(in == NULL && sz > 0) ||
(out == NULL && authTag == NULL)) {
(authTag == NULL) ||
(authIn == NULL) ||
(in == NULL && sz > 0) ||
(out == NULL && sz > 0)) {
WOLFSSL_MSG("a NULL parameter passed in when size is larger than 0");
return BAD_FUNC_ARG;
}
@ -2827,6 +2827,12 @@ int wc_AesGcmDecrypt(Aes* aes, byte* out, const byte* in, word32 sz,
if (partial != 0) {
IncrementGcmCounter(ctr);
wc_AesEncrypt(aes, ctr, scratch);
/* check if pointer is null after main AES-GCM blocks
* helps static analysis */
if (p == NULL || c == NULL) {
return BAD_STATE_E;
}
xorbuf(scratch, c, partial);
XMEMCPY(p, scratch, partial);
}
@ -4157,10 +4163,10 @@ int wc_AesGcmEncrypt(Aes* aes, byte* out, const byte* in, word32 sz,
/* sanity checks */
if (aes == NULL || (iv == NULL && ivSz > 0) ||
(authTag == NULL && authTagSz > 0) ||
(authIn == NULL && authInSz > 0) ||
(authTag == NULL) ||
(authIn == NULL) ||
(in == NULL && sz > 0) ||
(out == NULL && authTag == NULL)) {
(out == NULL && sz > 0)) {
WOLFSSL_MSG("a NULL parameter passed in when size is larger than 0");
return BAD_FUNC_ARG;
}
@ -4234,6 +4240,16 @@ int wc_AesGcmDecrypt(Aes* aes, byte* out, const byte* in, word32 sz,
byte scratch[AES_BLOCK_SIZE];
ctr = counter ;
/* sanity checks */
if (aes == NULL || (iv == NULL && ivSz > 0) ||
(authTag == NULL) ||
(authIn == NULL) ||
(in == NULL && sz > 0) ||
(out == NULL && sz > 0)) {
WOLFSSL_MSG("a NULL parameter passed in when size is larger than 0");
return BAD_FUNC_ARG;
}
XMEMSET(initialCounter, 0, AES_BLOCK_SIZE);
if (ivSz == NONCE_SZ) {
XMEMCPY(initialCounter, iv, ivSz);
@ -4270,6 +4286,12 @@ int wc_AesGcmDecrypt(Aes* aes, byte* out, const byte* in, word32 sz,
if (partial != 0) {
IncrementGcmCounter(ctr);
wc_AesEncrypt(aes, ctr, scratch);
/* check if pointer is null after main AES-GCM blocks
* helps static analysis */
if (p == NULL || c == NULL) {
return BAD_STATE_E;
}
xorbuf(scratch, c, partial);
XMEMCPY(p, scratch, partial);
}

706
wolfcrypt/src/port/arm/armv8-sha256.c
View File

@ -80,6 +80,10 @@ int wc_InitSha256(Sha256* sha256)
{
int ret = 0;
if (sha256 == NULL) {
return BAD_FUNC_ARG;
}
sha256->digest[0] = 0x6A09E667L;
sha256->digest[1] = 0xBB67AE85L;
sha256->digest[2] = 0x3C6EF372L;
@ -116,199 +120,206 @@ int wc_Sha256Update(Sha256* sha256, const byte* data, word32 len)
return BAD_FUNC_ARG;
}
/* fill leftover buffer with data */
add = min(len, SHA256_BLOCK_SIZE - sha256->buffLen);
XMEMCPY((byte*)(sha256->buffer) + sha256->buffLen, data, add);
sha256->buffLen += add;
data += add;
len -= add;
/* only perform actions if a buffer is passed in */
if (len > 0) {
/* fill leftover buffer with data */
add = min(len, SHA256_BLOCK_SIZE - sha256->buffLen);
XMEMCPY((byte*)(sha256->buffer) + sha256->buffLen, data, add);
sha256->buffLen += add;
data += add;
len -= add;
/* number of blocks in a row to complete */
numBlocks = (len + sha256->buffLen)/SHA256_BLOCK_SIZE;
/* number of blocks in a row to complete */
numBlocks = (len + sha256->buffLen)/SHA256_BLOCK_SIZE;
if (numBlocks > 0) {
/* get leftover amount after blocks */
add = (len + sha256->buffLen) - numBlocks * SHA256_BLOCK_SIZE;
__asm__ volatile (
"#load leftover data\n"
"LD1 {v0.2d-v3.2d}, %[buffer] \n"
if (numBlocks > 0) {
/* get leftover amount after blocks */
add = (len + sha256->buffLen) - numBlocks * SHA256_BLOCK_SIZE;
__asm__ volatile (
"#load leftover data\n"
"LD1 {v0.2d-v3.2d}, %[buffer] \n"
"#load current digest\n"
"LD1 {v12.2d-v13.2d}, %[digest] \n"
"MOV w8, %w[blocks] \n"
"REV32 v0.16b, v0.16b \n"
"REV32 v1.16b, v1.16b \n"
"REV32 v2.16b, v2.16b \n"
"REV32 v3.16b, v3.16b \n"
"#load current digest\n"
"LD1 {v12.2d-v13.2d}, %[digest] \n"
"MOV w8, %w[blocks] \n"
"REV32 v0.16b, v0.16b \n"
"REV32 v1.16b, v1.16b \n"
"REV32 v2.16b, v2.16b \n"
"REV32 v3.16b, v3.16b \n"
"#load K values in \n"
"LD1 {v16.4s-v19.4s}, [%[k]], #64 \n"
"LD1 {v20.4s-v23.4s}, [%[k]], #64 \n"
"MOV v14.16b, v12.16b \n" /* store digest for add at the end */
"MOV v15.16b, v13.16b \n"
"LD1 {v24.4s-v27.4s}, [%[k]], #64 \n"
"LD1 {v28.4s-v31.4s}, [%[k]], #64 \n"
"#load K values in \n"
"LD1 {v16.4s-v19.4s}, [%[k]], #64 \n"
"LD1 {v20.4s-v23.4s}, [%[k]], #64 \n"
"MOV v14.16b, v12.16b \n" /* store digest for add at the end */
"MOV v15.16b, v13.16b \n"
"LD1 {v24.4s-v27.4s}, [%[k]], #64 \n"
"LD1 {v28.4s-v31.4s}, [%[k]], #64 \n"
/* begining of SHA256 block operation */
"sha256Start:\n"
/* Round 1 */
"MOV v4.16b, v0.16b \n"
"ADD v0.4s, v0.4s, v16.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* begining of SHA256 block operation */
"sha256Start:\n"
/* Round 1 */
"MOV v4.16b, v0.16b \n"
"ADD v0.4s, v0.4s, v16.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 2 */
"SHA256SU0 v4.4s, v1.4s \n"
"ADD v0.4s, v1.4s, v17.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v4.4s, v2.4s, v3.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 2 */
"SHA256SU0 v4.4s, v1.4s \n"
"ADD v0.4s, v1.4s, v17.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v4.4s, v2.4s, v3.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 3 */
"SHA256SU0 v1.4s, v2.4s \n"
"ADD v0.4s, v2.4s, v18.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v1.4s, v3.4s, v4.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 3 */
"SHA256SU0 v1.4s, v2.4s \n"
"ADD v0.4s, v2.4s, v18.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v1.4s, v3.4s, v4.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 4 */
"SHA256SU0 v2.4s, v3.4s \n"
"ADD v0.4s, v3.4s, v19.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v2.4s, v4.4s, v1.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 4 */
"SHA256SU0 v2.4s, v3.4s \n"
"ADD v0.4s, v3.4s, v19.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v2.4s, v4.4s, v1.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 5 */
"SHA256SU0 v3.4s, v4.4s \n"
"ADD v0.4s, v4.4s, v20.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v3.4s, v1.4s, v2.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 5 */
"SHA256SU0 v3.4s, v4.4s \n"
"ADD v0.4s, v4.4s, v20.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v3.4s, v1.4s, v2.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 6 */
"SHA256SU0 v4.4s, v1.4s \n"
"ADD v0.4s, v1.4s, v21.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v4.4s, v2.4s, v3.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 6 */
"SHA256SU0 v4.4s, v1.4s \n"
"ADD v0.4s, v1.4s, v21.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v4.4s, v2.4s, v3.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 7 */
"SHA256SU0 v1.4s, v2.4s \n"
"ADD v0.4s, v2.4s, v22.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v1.4s, v3.4s, v4.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 7 */
"SHA256SU0 v1.4s, v2.4s \n"
"ADD v0.4s, v2.4s, v22.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v1.4s, v3.4s, v4.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 8 */
"SHA256SU0 v2.4s, v3.4s \n"
"ADD v0.4s, v3.4s, v23.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v2.4s, v4.4s, v1.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 8 */
"SHA256SU0 v2.4s, v3.4s \n"
"ADD v0.4s, v3.4s, v23.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v2.4s, v4.4s, v1.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 9 */
"SHA256SU0 v3.4s, v4.4s \n"
"ADD v0.4s, v4.4s, v24.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v3.4s, v1.4s, v2.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 9 */
"SHA256SU0 v3.4s, v4.4s \n"
"ADD v0.4s, v4.4s, v24.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v3.4s, v1.4s, v2.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 10 */
"SHA256SU0 v4.4s, v1.4s \n"
"ADD v0.4s, v1.4s, v25.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v4.4s, v2.4s, v3.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 10 */
"SHA256SU0 v4.4s, v1.4s \n"
"ADD v0.4s, v1.4s, v25.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v4.4s, v2.4s, v3.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 11 */
"SHA256SU0 v1.4s, v2.4s \n"
"ADD v0.4s, v2.4s, v26.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v1.4s, v3.4s, v4.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 11 */
"SHA256SU0 v1.4s, v2.4s \n"
"ADD v0.4s, v2.4s, v26.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v1.4s, v3.4s, v4.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 12 */
"SHA256SU0 v2.4s, v3.4s \n"
"ADD v0.4s, v3.4s, v27.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v2.4s, v4.4s, v1.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 12 */
"SHA256SU0 v2.4s, v3.4s \n"
"ADD v0.4s, v3.4s, v27.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v2.4s, v4.4s, v1.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 13 */
"SHA256SU0 v3.4s, v4.4s \n"
"ADD v0.4s, v4.4s, v28.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v3.4s, v1.4s, v2.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 13 */
"SHA256SU0 v3.4s, v4.4s \n"
"ADD v0.4s, v4.4s, v28.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256SU1 v3.4s, v1.4s, v2.4s \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 14 */
"ADD v0.4s, v1.4s, v29.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 14 */
"ADD v0.4s, v1.4s, v29.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 15 */
"ADD v0.4s, v2.4s, v30.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 15 */
"ADD v0.4s, v2.4s, v30.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 16 */
"ADD v0.4s, v3.4s, v31.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
/* Round 16 */
"ADD v0.4s, v3.4s, v31.4s \n"
"MOV v11.16b, v12.16b \n"
"SHA256H q12, q13, v0.4s \n"
"SHA256H2 q13, q11, v0.4s \n"
"#Add working vars back into digest state \n"
"SUB w8, w8, #1 \n"
"ADD v12.4s, v12.4s, v14.4s \n"
"ADD v13.4s, v13.4s, v15.4s \n"
"#Add working vars back into digest state \n"
"SUB w8, w8, #1 \n"
"ADD v12.4s, v12.4s, v14.4s \n"
"ADD v13.4s, v13.4s, v15.4s \n"
"#check if more blocks should be done\n"
"CBZ w8, sha256End \n"
"#check if more blocks should be done\n"
"CBZ w8, sha256End \n"
"#load in message and schedual updates \n"
"LD1 {v0.2d-v3.2d}, [%[dataIn]], #64 \n"
"MOV v14.16b, v12.16b \n"
"MOV v15.16b, v13.16b \n"
"REV32 v0.16b, v0.16b \n"
"REV32 v1.16b, v1.16b \n"
"REV32 v2.16b, v2.16b \n"
"REV32 v3.16b, v3.16b \n"
"B sha256Start \n" /* do another block */
"#load in message and schedual updates \n"
"LD1 {v0.2d-v3.2d}, [%[dataIn]], #64 \n"
"MOV v14.16b, v12.16b \n"
"MOV v15.16b, v13.16b \n"
"REV32 v0.16b, v0.16b \n"
"REV32 v1.16b, v1.16b \n"
"REV32 v2.16b, v2.16b \n"
"REV32 v3.16b, v3.16b \n"
"B sha256Start \n" /* do another block */
"sha256End:\n"
"STP q12, q13, %[out] \n"
"sha256End:\n"
"STP q12, q13, %[out] \n"
: [out] "=m" (sha256->digest), "=m" (sha256->buffer), "=r" (numBlocks),
"=r" (data)
: [k] "r" (K), [digest] "m" (sha256->digest), [buffer] "m" (sha256->buffer),
[blocks] "2" (numBlocks), [dataIn] "3" (data)
: "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7",
"v8", "v9", "v10", "v11", "v12", "v13", "v14",
"v15", "w8"
);
: [out] "=m" (sha256->digest), "=m" (sha256->buffer), "=r" (numBlocks),
"=r" (data)
: [k] "r" (K), [digest] "m" (sha256->digest), [buffer] "m" (sha256->buffer),
[blocks] "2" (numBlocks), [dataIn] "3" (data)
: "cc", "memory", "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7",
"v8", "v9", "v10", "v11", "v12", "v13", "v14",
"v15", "w8"
);
AddLength(sha256, SHA256_BLOCK_SIZE * numBlocks);
AddLength(sha256, SHA256_BLOCK_SIZE * numBlocks);
/* copy over any remaining data leftover */
XMEMCPY(sha256->buffer, data, add);
sha256->buffLen = add;
/* copy over any remaining data leftover */
XMEMCPY(sha256->buffer, data, add);
sha256->buffLen = add;
}
}
/* account for possiblity of not used if len = 0 */
(void)add;
(void)numBlocks;
return 0;
}
@ -672,217 +683,224 @@ int wc_Sha256Update(Sha256* sha256, const byte* data, word32 len)
return BAD_FUNC_ARG;
}
/* fill leftover buffer with data */
add = min(len, SHA256_BLOCK_SIZE - sha256->buffLen);
XMEMCPY((byte*)(sha256->buffer) + sha256->buffLen, data, add);
sha256->buffLen += add;
data += add;
len -= add;
/* only perform actions if a buffer is passed in */
if (len > 0) {
/* fill leftover buffer with data */
add = min(len, SHA256_BLOCK_SIZE - sha256->buffLen);
XMEMCPY((byte*)(sha256->buffer) + sha256->buffLen, data, add);
sha256->buffLen += add;
data += add;
len -= add;
/* number of blocks in a row to complete */
numBlocks = (len + sha256->buffLen)/SHA256_BLOCK_SIZE;
/* number of blocks in a row to complete */
numBlocks = (len + sha256->buffLen)/SHA256_BLOCK_SIZE;
if (numBlocks > 0) {
word32* bufPt = sha256->buffer;
word32* digPt = sha256->digest;
/* get leftover amount after blocks */
add = (len + sha256->buffLen) - numBlocks * SHA256_BLOCK_SIZE;
__asm__ volatile (
"#load leftover data\n"
"VLDM %[buffer]!, {q0-q3} \n"
if (numBlocks > 0) {
word32* bufPt = sha256->buffer;
word32* digPt = sha256->digest;
/* get leftover amount after blocks */
add = (len + sha256->buffLen) - numBlocks * SHA256_BLOCK_SIZE;
__asm__ volatile (
"#load leftover data\n"
"VLDM %[buffer]!, {q0-q3} \n"
"#load current digest\n"
"VLDM %[digest], {q12-q13} \n"
"MOV r8, %r[blocks] \n"
"VREV32.8 q0, q0 \n"
"VREV32.8 q1, q1 \n"
"VREV32.8 q2, q2 \n"
"VREV32.8 q3, q3 \n"
"VLDM %[k]! ,{q5-q8} \n"
"VLDM %[k]! ,{q9}\n"
"#load current digest\n"
"VLDM %[digest], {q12-q13} \n"
"MOV r8, %r[blocks] \n"
"VREV32.8 q0, q0 \n"
"VREV32.8 q1, q1 \n"
"VREV32.8 q2, q2 \n"
"VREV32.8 q3, q3 \n"
"VLDM %[k]! ,{q5-q8} \n"
"VLDM %[k]! ,{q9}\n"
"VMOV.32 q14, q12 \n" /* store digest for add at the end */
"VMOV.32 q15, q13 \n"
"VMOV.32 q14, q12 \n" /* store digest for add at the end */
"VMOV.32 q15, q13 \n"
/* begining of SHA256 block operation */
"sha256Start:\n"
/* begining of SHA256 block operation */
"sha256Start:\n"
/* Round 1 */
"VMOV.32 q4, q0 \n"
"VADD.i32 q0, q0, q5 \n"
"VMOV.32 q11, q12 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 1 */
"VMOV.32 q4, q0 \n"
"VADD.i32 q0, q0, q5 \n"
"VMOV.32 q11, q12 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 2 */
"SHA256SU0.32 q4, q1 \n"
"VADD.i32 q0, q1, q6 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q4, q2, q3 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 2 */
"SHA256SU0.32 q4, q1 \n"
"VADD.i32 q0, q1, q6 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q4, q2, q3 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 3 */
"SHA256SU0.32 q1, q2 \n"
"VADD.i32 q0, q2, q7 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q1, q3, q4 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 3 */
"SHA256SU0.32 q1, q2 \n"
"VADD.i32 q0, q2, q7 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q1, q3, q4 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 4 */
"SHA256SU0.32 q2, q3 \n"
"VADD.i32 q0, q3, q8 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q2, q4, q1 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 4 */
"SHA256SU0.32 q2, q3 \n"
"VADD.i32 q0, q3, q8 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q2, q4, q1 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 5 */
"SHA256SU0.32 q3, q4 \n"
"VADD.i32 q0, q4, q9 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q3, q1, q2 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 5 */
"SHA256SU0.32 q3, q4 \n"
"VADD.i32 q0, q4, q9 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q3, q1, q2 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 6 */
"VLD1.32 {q10}, [%[k]]! \n"
"SHA256SU0.32 q4, q1 \n"
"VADD.i32 q0, q1, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q4, q2, q3 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 6 */
"VLD1.32 {q10}, [%[k]]! \n"
"SHA256SU0.32 q4, q1 \n"
"VADD.i32 q0, q1, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q4, q2, q3 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 7 */
"VLD1.32 {q10}, [%[k]]! \n"
"SHA256SU0.32 q1, q2 \n"
"VADD.i32 q0, q2, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q1, q3, q4 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 7 */
"VLD1.32 {q10}, [%[k]]! \n"
"SHA256SU0.32 q1, q2 \n"
"VADD.i32 q0, q2, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q1, q3, q4 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 8 */
"VLD1.32 {q10}, [%[k]]! \n"
"SHA256SU0.32 q2, q3 \n"
"VADD.i32 q0, q3, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q2, q4, q1 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 8 */
"VLD1.32 {q10}, [%[k]]! \n"
"SHA256SU0.32 q2, q3 \n"
"VADD.i32 q0, q3, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q2, q4, q1 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 9 */
"VLD1.32 {q10}, [%[k]]! \n"
"SHA256SU0.32 q3, q4 \n"
"VADD.i32 q0, q4, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q3, q1, q2 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 9 */
"VLD1.32 {q10}, [%[k]]! \n"
"SHA256SU0.32 q3, q4 \n"
"VADD.i32 q0, q4, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q3, q1, q2 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 10 */
"VLD1.32 {q10}, [%[k]]! \n"
"SHA256SU0.32 q4, q1 \n"
"VADD.i32 q0, q1, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q4, q2, q3 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 10 */
"VLD1.32 {q10}, [%[k]]! \n"
"SHA256SU0.32 q4, q1 \n"
"VADD.i32 q0, q1, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q4, q2, q3 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 11 */
"VLD1.32 {q10}, [%[k]]! \n"
"SHA256SU0.32 q1, q2 \n"
"VADD.i32 q0, q2, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q1, q3, q4 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 11 */
"VLD1.32 {q10}, [%[k]]! \n"
"SHA256SU0.32 q1, q2 \n"
"VADD.i32 q0, q2, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q1, q3, q4 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 12 */
"VLD1.32 {q10}, [%[k]]! \n"
"SHA256SU0.32 q2, q3 \n"
"VADD.i32 q0, q3, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q2, q4, q1 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 12 */
"VLD1.32 {q10}, [%[k]]! \n"
"SHA256SU0.32 q2, q3 \n"
"VADD.i32 q0, q3, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q2, q4, q1 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 13 */
"VLD1.32 {q10}, [%[k]]! \n"
"SHA256SU0.32 q3, q4 \n"
"VADD.i32 q0, q4, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q3, q1, q2 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 13 */
"VLD1.32 {q10}, [%[k]]! \n"
"SHA256SU0.32 q3, q4 \n"
"VADD.i32 q0, q4, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256SU1.32 q3, q1, q2 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 14 */
"VLD1.32 {q10}, [%[k]]! \n"
"VADD.i32 q0, q1, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 14 */
"VLD1.32 {q10}, [%[k]]! \n"
"VADD.i32 q0, q1, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 15 */
"VLD1.32 {q10}, [%[k]]! \n"
"VADD.i32 q0, q2, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 15 */
"VLD1.32 {q10}, [%[k]]! \n"
"VADD.i32 q0, q2, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 16 */
"VLD1.32 {q10}, [%[k]] \n"
"SUB r8, r8, #1 \n"
"VADD.i32 q0, q3, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
/* Round 16 */
"VLD1.32 {q10}, [%[k]] \n"
"SUB r8, r8, #1 \n"
"VADD.i32 q0, q3, q10 \n"
"VMOV.32 q11, q12 \n"
"SHA256H.32 q12, q13, q0 \n"
"SHA256H2.32 q13, q11, q0 \n"
"#Add working vars back into digest state \n"
"VADD.i32 q12, q12, q14 \n"
"VADD.i32 q13, q13, q15 \n"
"#Add working vars back into digest state \n"
"VADD.i32 q12, q12, q14 \n"
"VADD.i32 q13, q13, q15 \n"
"#check if more blocks should be done\n"
"CMP r8, #0 \n"
"BEQ sha256End \n"
"#check if more blocks should be done\n"
"CMP r8, #0 \n"
"BEQ sha256End \n"
"#load in message and schedual updates \n"
"VLD1.32 {q0}, [%[dataIn]]! \n"
"VLD1.32 {q1}, [%[dataIn]]! \n"
"VLD1.32 {q2}, [%[dataIn]]! \n"
"VLD1.32 {q3}, [%[dataIn]]! \n"
"#load in message and schedual updates \n"
"VLD1.32 {q0}, [%[dataIn]]! \n"
"VLD1.32 {q1}, [%[dataIn]]! \n"
"VLD1.32 {q2}, [%[dataIn]]! \n"
"VLD1.32 {q3}, [%[dataIn]]! \n"
/* reset K pointer */
"SUB %[k], %[k], #160 \n"
"VREV32.8 q0, q0 \n"
"VREV32.8 q1, q1 \n"
"VREV32.8 q2, q2 \n"
"VREV32.8 q3, q3 \n"
"VMOV.32 q14, q12 \n"
"VMOV.32 q15, q13 \n"
"B sha256Start \n" /* do another block */
/* reset K pointer */
"SUB %[k], %[k], #160 \n"
"VREV32.8 q0, q0 \n"
"VREV32.8 q1, q1 \n"
"VREV32.8 q2, q2 \n"
"VREV32.8 q3, q3 \n"
"VMOV.32 q14, q12 \n"
"VMOV.32 q15, q13 \n"
"B sha256Start \n" /* do another block */
"sha256End:\n"
"VST1.32 {q12, q13}, [%[out]] \n"
"sha256End:\n"
"VST1.32 {q12, q13}, [%[out]] \n"
: [out] "=r" (digPt), "=r" (bufPt), "=r" (numBlocks),
"=r" (data)
: [k] "r" (K), [digest] "0" (digPt), [buffer] "1" (bufPt),
[blocks] "2" (numBlocks), [dataIn] "3" (data)
: "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7",
"q8", "q9", "q10", "q11", "q12", "q13", "q14",
"q15", "r8"
);
: [out] "=r" (digPt), "=r" (bufPt), "=r" (numBlocks),
"=r" (data)
: [k] "r" (K), [digest] "0" (digPt), [buffer] "1" (bufPt),
[blocks] "2" (numBlocks), [dataIn] "3" (data)
: "cc", "memory", "q0", "q1", "q2", "q3", "q4", "q5", "q6", "q7",
"q8", "q9", "q10", "q11", "q12", "q13", "q14",
"q15", "r8"
);
AddLength(sha256, SHA256_BLOCK_SIZE * numBlocks);
AddLength(sha256, SHA256_BLOCK_SIZE * numBlocks);
/* copy over any remaining data leftover */
XMEMCPY(sha256->buffer, data, add);
sha256->buffLen = add;
/* copy over any remaining data leftover */
XMEMCPY(sha256->buffer, data, add);
sha256->buffLen = add;
}
}
/* account for possiblity of not used if len = 0 */
(void)add;
(void)numBlocks;
return 0;
}