// Run 16 parallel hashes
void sha2Hash(vecu16 pointers, int totalBlocks, vecu16 outHashes)
{
// Initial H values
vecu16 A = __builtin_vp_makevectori(0x6A09E667);
vecu16 B = __builtin_vp_makevectori(0xBB67AE85);
vecu16 C = __builtin_vp_makevectori(0x3C6EF372);
vecu16 D = __builtin_vp_makevectori(0xA54FF53A);
vecu16 E = __builtin_vp_makevectori(0x510E527F);
vecu16 F = __builtin_vp_makevectori(0x9B05688C);
vecu16 G = __builtin_vp_makevectori(0x1F83D9AB);
vecu16 H = __builtin_vp_makevectori(0x5BE0CD19);
for (int i = 0; i < totalBlocks; i++)
{
vecu16 W[64];
for (int index = 0; index < 16; index++)
{
W[index] = __builtin_vp_gather_loadi(pointers);
pointers += __builtin_vp_makevectori(4);
}
for (int index = 16; index < 64; index++)
W[index] = SIG1(W[index - 2]) + W[index - 7] + SIG0(W[index - 15]) + W[index - 16];
for (int round = 0; round < 64; round++)
{
vecu16 temp1 = H + SIG1(E) + CH(E, F, G) + __builtin_vp_makevectori(K[round]) + W[round];
vecu16 temp2 = SIG0(A) + MA(A, B, C);
H = G;
G = F;
F = E;
E = D + temp1;
D = C;
C = B;
B = A;
A = temp1 + temp2;
}
}
// doesn't add padding or length fields to end...
__builtin_vp_scatter_storei(outHashes, A);
__builtin_vp_scatter_storei(outHashes + __builtin_vp_makevectori(4), B);
__builtin_vp_scatter_storei(outHashes + __builtin_vp_makevectori(8), C);
__builtin_vp_scatter_storei(outHashes + __builtin_vp_makevectori(12), D);
__builtin_vp_scatter_storei(outHashes + __builtin_vp_makevectori(16), E);
__builtin_vp_scatter_storei(outHashes + __builtin_vp_makevectori(20), F);
__builtin_vp_scatter_storei(outHashes + __builtin_vp_makevectori(24), G);
__builtin_vp_scatter_storei(outHashes + __builtin_vp_makevectori(28), H);
}
static void Sha256_Transform(Sha256ContextType *context, uint8_t * data)
{
uint32_t a,b,c,d,e,f,g,h,i,j,t1,t2,m[64];
for (i = 0, j = 0; i < 16; ++i, j += 4)
{
m[i] = (data[j] << 24) | (data[j+1] << 16) | (data[j+2] << 8) | (data[j+3]);
}
for (; i < 64; i++)
{
m[i] = SIG1(m[i-2]) + m[i-7] + SIG0(m[i-15]) + m[i-16];
}
a = context->State[0];
b = context->State[1];
c = context->State[2];
d = context->State[3];
e = context->State[4];
f = context->State[5];
g = context->State[6];
h = context->State[7];
for (i = 0; i < 64; i++)
{
t1 = h + EP1(e) + CH(e,f,g) + k[i] + m[i];
t2 = EP0(a) + MAJ(a,b,c);
h = g;
g = f;
f = e;
e = d + t1;
d = c;
c = b;
b = a;
a = t1 + t2;
}
context->State[0] += a;
context->State[1] += b;
context->State[2] += c;
context->State[3] += d;
context->State[4] += e;
context->State[5] += f;
context->State[6] += g;
context->State[7] += h;
}
static void SHA256_transform( SHA256_ctx* ctx )
{
int t;
unsigned int A = ctx->H[ 0 ];
unsigned int B = ctx->H[ 1 ];
unsigned int C = ctx->H[ 2 ];
unsigned int D = ctx->H[ 3 ];
unsigned int E = ctx->H[ 4 ];
unsigned int F = ctx->H[ 5 ];
unsigned int G = ctx->H[ 6 ];
unsigned int H = ctx->H[ 7 ];
unsigned int T1, T2;
unsigned int W[ 64 ];
memcpy( W, ctx->M, 64 );
for ( t = 16; t < 64; t++ )
{
W[ t ] = sig1(W[t-2]) + W[t-7] + sig0(W[t-15]) + W[t-16];
}
for ( t = 0; t < 64; t++ )
{
T1 = H + SIG1(E) + Ch(E,F,G) + K[t] + W[t];
T2 = SIG0(A) + Maj(A,B,C);
H = G;
G = F;
F = E;
E = D + T1;
D = C;
C = B;
B = A;
A = T1 + T2;
}
ctx->H[ 0 ] += A;
ctx->H[ 1 ] += B;
ctx->H[ 2 ] += C;
ctx->H[ 3 ] += D;
ctx->H[ 4 ] += E;
ctx->H[ 5 ] += F;
ctx->H[ 6 ] += G;
ctx->H[ 7 ] += H;
}
// Run 16 parallel hashes
void sha2Hash(vecu16_t pointers, int totalBlocks, vecu16_t outHashes)
{
// Initial H values
vecu16_t h0 = __builtin_nyuzi_makevectori(0x6A09E667);
vecu16_t h1 = __builtin_nyuzi_makevectori(0xBB67AE85);
vecu16_t h2 = __builtin_nyuzi_makevectori(0x3C6EF372);
vecu16_t h3 = __builtin_nyuzi_makevectori(0xA54FF53A);
vecu16_t h4 = __builtin_nyuzi_makevectori(0x510E527F);
vecu16_t h5 = __builtin_nyuzi_makevectori(0x9B05688C);
vecu16_t h6 = __builtin_nyuzi_makevectori(0x1F83D9AB);
vecu16_t h7 = __builtin_nyuzi_makevectori(0x5BE0CD19);
for (int i = 0; i < totalBlocks; i++)
{
vecu16_t w[64];
for (int index = 0; index < 16; index++)
{
w[index] = __builtin_nyuzi_gather_loadi(pointers);
pointers += __builtin_nyuzi_makevectori(4);
}
for (int index = 16; index < 64; index++)
w[index] = SIG1(w[index - 2]) + w[index - 7] + SIG0(w[index - 15]) + w[index - 16];
vecu16_t a = h0;
vecu16_t b = h1;
vecu16_t c = h2;
vecu16_t d = h3;
vecu16_t e = h4;
vecu16_t f = h5;
vecu16_t g = h6;
vecu16_t h = h7;
for (int round = 0; round < 64; round++)
{
vecu16_t temp1 = h + SIG1(e) + CH(e, f, g) + __builtin_nyuzi_makevectori(K[round]) + w[round];
vecu16_t temp2 = SIG0(a) + MA(a, b, c);
h = g;
g = f;
f = e;
e = d + temp1;
d = c;
c = b;
b = a;
a = temp1 + temp2;
}
h0 += a;
h1 += b;
h2 += c;
h3 += d;
h4 += e;
h5 += f;
h6 += g;
h7 += h;
}
// doesn't add padding or length fields to end...
__builtin_nyuzi_scatter_storei(outHashes, h0);
__builtin_nyuzi_scatter_storei(outHashes + __builtin_nyuzi_makevectori(4), h1);
__builtin_nyuzi_scatter_storei(outHashes + __builtin_nyuzi_makevectori(8), h2);
__builtin_nyuzi_scatter_storei(outHashes + __builtin_nyuzi_makevectori(12), h3);
__builtin_nyuzi_scatter_storei(outHashes + __builtin_nyuzi_makevectori(16), h4);
__builtin_nyuzi_scatter_storei(outHashes + __builtin_nyuzi_makevectori(20), h5);
__builtin_nyuzi_scatter_storei(outHashes + __builtin_nyuzi_makevectori(24), h6);
__builtin_nyuzi_scatter_storei(outHashes + __builtin_nyuzi_makevectori(28), h7);
}
void *subchain_hash_thread(void *pthread_arg) {
/*
* subchain_hash_thread()
* Given a hash value and its position (index) in the chain, determine
* the corresponding final hash. This is then used to find a candidate
* chain in the main table
*/
PthreadData *mydata;
mydata = (PthreadData*)pthread_arg;
// TableHeader *header = mydata->header;
TableEntry *entry = mydata->entry;
uint thread_idx = mydata->entry_idx;
uint8_t M[64]; // Initial string - zero padded and length in bits appended
uint32_t W[64]; // Expanded Key Schedule
uint32_t H[8]; // Hash
int i = 0; // working index
uint64_t l = 0; // length of message
uint8_t B[64]; // store initial and working passwords here to protect original data
uint8_t *in,*out;
int reduction_idx,count;
if(thread_idx<LINKS) {
// set up a pointer to input_hash & final_hash
TableEntry *data = entry + thread_idx;
// move target hash to H
for(i=0;i<8;i++) H[i] = data->input_hash[i];
reduction_idx = thread_idx;
count = LINKS - thread_idx - 1;
while(count > 0) {
// Reduce hash to zero terminated password in B
reduce_hash(H,B,reduction_idx);
// copy zero terminated string from B to M and note length
in = B;
out = M;
i=0; l=0;
while(in[i] != 0x00) {
out[i] = in[i];
i++;
l++;
}
out[i++] = 0x80;
// zero fill
while(i < 56) out[i++]=0x00;
/*
* The hash algorithm uses 32 bit (4 byte words).
* On little endian machines (Intel) the constants
* are stored lsb->msb internally. To match this the WORDS
* of the input message are subject to endian swap.
*/
uint8_t *x = M;
int y;
for(y=0; y<14; y++) {
// long swap
*(x+3) ^= *x;
*x ^= *(x+3);
*(x+3) ^= *x;
// short swap
*(x+2) ^= *(x+1);
*(x+1) ^= *(x+2);
*(x+2) ^= *(x+1);
// move pointer up
x += 4;
}
// need a 32 bit pointer to store length as 2 words
l*=8; //length in bits
uint32_t *p = (uint32_t*)&l;
uint32_t *q = (uint32_t*)&out[i];
*q = *(p+1);
*(q+1) = *p;
// The 64 bytes in the message block can now be used
// to initialise the 64 4-byte words in the message schedule W[64]
// REUSE i
uint8_t *r = (uint8_t*)M;
uint8_t *s = (uint8_t*)W;
for(i=0;i<64;i++) s[i] = r[i];
for(i=16;i<64;i++) W[i] = SIG1(W[i-2]) + W[i-7] + SIG0(W[i-15]) + W[i-16];
// set initial hash values
initHash(H);
// Now calc the hash
sha256_transform(W,H);
// update the counters
reduction_idx += 1;
count -= 1;
} // while(count>0)
// copy comp_hash to final hash
for(i=0;i<8;i++) data->final_hash[i] = H[i];
data->sublinks = thread_idx;
} // if(thread_idx<LINKS)
void *void_ptr=NULL;
return(void_ptr);
} // hash_calculate