static void DoubleBlockSHA256(const void* pin, void* pad, const void *pre, unsigned int thash[9][NPAR], const void *init)
{
unsigned int* In = (unsigned int*)pin;
unsigned int* Pad = (unsigned int*)pad;
unsigned int* hPre = (unsigned int*)pre;
unsigned int* hInit = (unsigned int*)init;
unsigned int /* i, j, */ k;
/* vectors used in calculation */
vector unsigned int w0, w1, w2, w3, w4, w5, w6, w7, w8, w9, w10, w11, w12, w13, w14, w15;
vector unsigned int T1;
vector unsigned int a, b, c, d, e, f, g, h;
vector unsigned int nonce, preNonce;
/* nonce offset for vector */
vector unsigned int offset = (vector unsigned int)(0, 1, 2, 3);
preNonce = vec_add((vector unsigned int)(In[3],In[3],In[3],In[3]), offset);
for(k = 0; k<NPAR; k+=4)
{
w0 = (vector unsigned int)(In[0],In[0],In[0],In[0]);
w1 = (vector unsigned int)(In[1],In[1],In[1],In[1]);
w2 = (vector unsigned int)(In[2],In[2],In[2],In[2]);
//w3 = (vector unsigned int)(In[3],In[3],In[3],In[3]); nonce will be later hacked into the hash
w4 = (vector unsigned int)(In[4],In[4],In[4],In[4]);
w5 = (vector unsigned int)(In[5],In[5],In[5],In[5]);
w6 = (vector unsigned int)(In[6],In[6],In[6],In[6]);
w7 = (vector unsigned int)(In[7],In[7],In[7],In[7]);
w8 = (vector unsigned int)(In[8],In[8],In[8],In[8]);
w9 = (vector unsigned int)(In[9],In[9],In[9],In[9]);
w10 = (vector unsigned int)(In[10],In[10],In[10],In[10]);
w11 = (vector unsigned int)(In[11],In[11],In[11],In[11]);
w12 = (vector unsigned int)(In[12],In[12],In[12],In[12]);
w13 = (vector unsigned int)(In[13],In[13],In[13],In[13]);
w14 = (vector unsigned int)(In[14],In[14],In[14],In[14]);
w15 = (vector unsigned int)(In[15],In[15],In[15],In[15]);
/* hack nonce into lowest byte of w3 */
nonce = vec_add(preNonce, (vector unsigned int)(k,k,k,k));
w3 = nonce;
//printf ("W3: %08vlx\n", w3);
a = (vector unsigned int)(hPre[0],hPre[0],hPre[0],hPre[0]);
b = (vector unsigned int)(hPre[1],hPre[1],hPre[1],hPre[1]);
c = (vector unsigned int)(hPre[2],hPre[2],hPre[2],hPre[2]);
d = (vector unsigned int)(hPre[3],hPre[3],hPre[3],hPre[3]);
e = (vector unsigned int)(hPre[4],hPre[4],hPre[4],hPre[4]);
f = (vector unsigned int)(hPre[5],hPre[5],hPre[5],hPre[5]);
g = (vector unsigned int)(hPre[6],hPre[6],hPre[6],hPre[6]);
h = (vector unsigned int)(hPre[7],hPre[7],hPre[7],hPre[7]);
SHA256ROUND(a, b, c, d, e, f, g, h, 0, w0);
SHA256ROUND(h, a, b, c, d, e, f, g, 1, w1);
SHA256ROUND(g, h, a, b, c, d, e, f, 2, w2);
SHA256ROUND(f, g, h, a, b, c, d, e, 3, w3);
SHA256ROUND(e, f, g, h, a, b, c, d, 4, w4);
SHA256ROUND(d, e, f, g, h, a, b, c, 5, w5);
SHA256ROUND(c, d, e, f, g, h, a, b, 6, w6);
SHA256ROUND(b, c, d, e, f, g, h, a, 7, w7);
SHA256ROUND(a, b, c, d, e, f, g, h, 8, w8);
SHA256ROUND(h, a, b, c, d, e, f, g, 9, w9);
SHA256ROUND(g, h, a, b, c, d, e, f, 10, w10);
SHA256ROUND(f, g, h, a, b, c, d, e, 11, w11);
SHA256ROUND(e, f, g, h, a, b, c, d, 12, w12);
SHA256ROUND(d, e, f, g, h, a, b, c, 13, w13);
SHA256ROUND(c, d, e, f, g, h, a, b, 14, w14);
SHA256ROUND(b, c, d, e, f, g, h, a, 15, w15);
w0 = add4(SIGMA1_256(w14), w9, SIGMA0_256(w1), w0);
SHA256ROUND(a, b, c, d, e, f, g, h, 16, w0);
w1 = add4(SIGMA1_256(w15), w10, SIGMA0_256(w2), w1);
SHA256ROUND(h, a, b, c, d, e, f, g, 17, w1);
w2 = add4(SIGMA1_256(w0), w11, SIGMA0_256(w3), w2);
SHA256ROUND(g, h, a, b, c, d, e, f, 18, w2);
w3 = add4(SIGMA1_256(w1), w12, SIGMA0_256(w4), w3);
SHA256ROUND(f, g, h, a, b, c, d, e, 19, w3);
w4 = add4(SIGMA1_256(w2), w13, SIGMA0_256(w5), w4);
SHA256ROUND(e, f, g, h, a, b, c, d, 20, w4);
w5 = add4(SIGMA1_256(w3), w14, SIGMA0_256(w6), w5);
SHA256ROUND(d, e, f, g, h, a, b, c, 21, w5);
w6 = add4(SIGMA1_256(w4), w15, SIGMA0_256(w7), w6);
SHA256ROUND(c, d, e, f, g, h, a, b, 22, w6);
w7 = add4(SIGMA1_256(w5), w0, SIGMA0_256(w8), w7);
SHA256ROUND(b, c, d, e, f, g, h, a, 23, w7);
w8 = add4(SIGMA1_256(w6), w1, SIGMA0_256(w9), w8);
SHA256ROUND(a, b, c, d, e, f, g, h, 24, w8);
w9 = add4(SIGMA1_256(w7), w2, SIGMA0_256(w10), w9);
SHA256ROUND(h, a, b, c, d, e, f, g, 25, w9);
w10 = add4(SIGMA1_256(w8), w3, SIGMA0_256(w11), w10);
SHA256ROUND(g, h, a, b, c, d, e, f, 26, w10);
w11 = add4(SIGMA1_256(w9), w4, SIGMA0_256(w12), w11);
SHA256ROUND(f, g, h, a, b, c, d, e, 27, w11);
w12 = add4(SIGMA1_256(w10), w5, SIGMA0_256(w13), w12);
SHA256ROUND(e, f, g, h, a, b, c, d, 28, w12);
w13 = add4(SIGMA1_256(w11), w6, SIGMA0_256(w14), w13);
SHA256ROUND(d, e, f, g, h, a, b, c, 29, w13);
w14 = add4(SIGMA1_256(w12), w7, SIGMA0_256(w15), w14);
SHA256ROUND(c, d, e, f, g, h, a, b, 30, w14);
w15 = add4(SIGMA1_256(w13), w8, SIGMA0_256(w0), w15);
SHA256ROUND(b, c, d, e, f, g, h, a, 31, w15);
w0 = add4(SIGMA1_256(w14), w9, SIGMA0_256(w1), w0);
SHA256ROUND(a, b, c, d, e, f, g, h, 32, w0);
w1 = add4(SIGMA1_256(w15), w10, SIGMA0_256(w2), w1);
SHA256ROUND(h, a, b, c, d, e, f, g, 33, w1);
w2 = add4(SIGMA1_256(w0), w11, SIGMA0_256(w3), w2);
SHA256ROUND(g, h, a, b, c, d, e, f, 34, w2);
w3 = add4(SIGMA1_256(w1), w12, SIGMA0_256(w4), w3);
SHA256ROUND(f, g, h, a, b, c, d, e, 35, w3);
w4 = add4(SIGMA1_256(w2), w13, SIGMA0_256(w5), w4);
SHA256ROUND(e, f, g, h, a, b, c, d, 36, w4);
w5 = add4(SIGMA1_256(w3), w14, SIGMA0_256(w6), w5);
SHA256ROUND(d, e, f, g, h, a, b, c, 37, w5);
w6 = add4(SIGMA1_256(w4), w15, SIGMA0_256(w7), w6);
SHA256ROUND(c, d, e, f, g, h, a, b, 38, w6);
w7 = add4(SIGMA1_256(w5), w0, SIGMA0_256(w8), w7);
SHA256ROUND(b, c, d, e, f, g, h, a, 39, w7);
w8 = add4(SIGMA1_256(w6), w1, SIGMA0_256(w9), w8);
SHA256ROUND(a, b, c, d, e, f, g, h, 40, w8);
w9 = add4(SIGMA1_256(w7), w2, SIGMA0_256(w10), w9);
SHA256ROUND(h, a, b, c, d, e, f, g, 41, w9);
w10 = add4(SIGMA1_256(w8), w3, SIGMA0_256(w11), w10);
SHA256ROUND(g, h, a, b, c, d, e, f, 42, w10);
w11 = add4(SIGMA1_256(w9), w4, SIGMA0_256(w12), w11);
SHA256ROUND(f, g, h, a, b, c, d, e, 43, w11);
w12 = add4(SIGMA1_256(w10), w5, SIGMA0_256(w13), w12);
SHA256ROUND(e, f, g, h, a, b, c, d, 44, w12);
w13 = add4(SIGMA1_256(w11), w6, SIGMA0_256(w14), w13);
SHA256ROUND(d, e, f, g, h, a, b, c, 45, w13);
w14 = add4(SIGMA1_256(w12), w7, SIGMA0_256(w15), w14);
SHA256ROUND(c, d, e, f, g, h, a, b, 46, w14);
w15 = add4(SIGMA1_256(w13), w8, SIGMA0_256(w0), w15);
SHA256ROUND(b, c, d, e, f, g, h, a, 47, w15);
w0 = add4(SIGMA1_256(w14), w9, SIGMA0_256(w1), w0);
SHA256ROUND(a, b, c, d, e, f, g, h, 48, w0);
w1 = add4(SIGMA1_256(w15), w10, SIGMA0_256(w2), w1);
SHA256ROUND(h, a, b, c, d, e, f, g, 49, w1);
w2 = add4(SIGMA1_256(w0), w11, SIGMA0_256(w3), w2);
SHA256ROUND(g, h, a, b, c, d, e, f, 50, w2);
w3 = add4(SIGMA1_256(w1), w12, SIGMA0_256(w4), w3);
SHA256ROUND(f, g, h, a, b, c, d, e, 51, w3);
w4 = add4(SIGMA1_256(w2), w13, SIGMA0_256(w5), w4);
SHA256ROUND(e, f, g, h, a, b, c, d, 52, w4);
w5 = add4(SIGMA1_256(w3), w14, SIGMA0_256(w6), w5);
SHA256ROUND(d, e, f, g, h, a, b, c, 53, w5);
w6 = add4(SIGMA1_256(w4), w15, SIGMA0_256(w7), w6);
SHA256ROUND(c, d, e, f, g, h, a, b, 54, w6);
w7 = add4(SIGMA1_256(w5), w0, SIGMA0_256(w8), w7);
SHA256ROUND(b, c, d, e, f, g, h, a, 55, w7);
w8 = add4(SIGMA1_256(w6), w1, SIGMA0_256(w9), w8);
SHA256ROUND(a, b, c, d, e, f, g, h, 56, w8);
w9 = add4(SIGMA1_256(w7), w2, SIGMA0_256(w10), w9);
SHA256ROUND(h, a, b, c, d, e, f, g, 57, w9);
w10 = add4(SIGMA1_256(w8), w3, SIGMA0_256(w11), w10);
SHA256ROUND(g, h, a, b, c, d, e, f, 58, w10);
w11 = add4(SIGMA1_256(w9), w4, SIGMA0_256(w12), w11);
SHA256ROUND(f, g, h, a, b, c, d, e, 59, w11);
w12 = add4(SIGMA1_256(w10), w5, SIGMA0_256(w13), w12);
SHA256ROUND(e, f, g, h, a, b, c, d, 60, w12);
w13 = add4(SIGMA1_256(w11), w6, SIGMA0_256(w14), w13);
SHA256ROUND(d, e, f, g, h, a, b, c, 61, w13);
w14 = add4(SIGMA1_256(w12), w7, SIGMA0_256(w15), w14);
SHA256ROUND(c, d, e, f, g, h, a, b, 62, w14);
w15 = add4(SIGMA1_256(w13), w8, SIGMA0_256(w0), w15);
SHA256ROUND(b, c, d, e, f, g, h, a, 63, w15);
#define store_load(x, i, dest) \
T1 = (vector unsigned int)((hPre)[i],(hPre)[i],(hPre)[i],(hPre)[i]); \
dest = vec_add(T1, x);
store_load(a, 0, w0);
store_load(b, 1, w1);
store_load(c, 2, w2);
store_load(d, 3, w3);
store_load(e, 4, w4);
store_load(f, 5, w5);
store_load(g, 6, w6);
store_load(h, 7, w7);
/* end of first SHA256 round */
w8 = (vector unsigned int)(Pad[8],Pad[8],Pad[8],Pad[8]);
w9 = (vector unsigned int)(Pad[9],Pad[9],Pad[9],Pad[9]);
w10 = (vector unsigned int)(Pad[10],Pad[10],Pad[10],Pad[10]);
w11 = (vector unsigned int)(Pad[11],Pad[11],Pad[11],Pad[11]);
w12 = (vector unsigned int)(Pad[12],Pad[12],Pad[12],Pad[12]);
w13 = (vector unsigned int)(Pad[13],Pad[13],Pad[13],Pad[13]);
w14 = (vector unsigned int)(Pad[14],Pad[14],Pad[14],Pad[14]);
w15 = (vector unsigned int)(Pad[15],Pad[15],Pad[15],Pad[15]);
a = (vector unsigned int)(hInit[0],hInit[0],hInit[0],hInit[0]);
b = (vector unsigned int)(hInit[1],hInit[1],hInit[1],hInit[1]);
c = (vector unsigned int)(hInit[2],hInit[2],hInit[2],hInit[2]);
d = (vector unsigned int)(hInit[3],hInit[3],hInit[3],hInit[3]);
e = (vector unsigned int)(hInit[4],hInit[4],hInit[4],hInit[4]);
f = (vector unsigned int)(hInit[5],hInit[5],hInit[5],hInit[5]);
g = (vector unsigned int)(hInit[6],hInit[6],hInit[6],hInit[6]);
h = (vector unsigned int)(hInit[7],hInit[7],hInit[7],hInit[7]);
SHA256ROUND(a, b, c, d, e, f, g, h, 0, w0);
SHA256ROUND(h, a, b, c, d, e, f, g, 1, w1);
SHA256ROUND(g, h, a, b, c, d, e, f, 2, w2);
SHA256ROUND(f, g, h, a, b, c, d, e, 3, w3);
SHA256ROUND(e, f, g, h, a, b, c, d, 4, w4);
SHA256ROUND(d, e, f, g, h, a, b, c, 5, w5);
SHA256ROUND(c, d, e, f, g, h, a, b, 6, w6);
SHA256ROUND(b, c, d, e, f, g, h, a, 7, w7);
SHA256ROUND(a, b, c, d, e, f, g, h, 8, w8);
SHA256ROUND(h, a, b, c, d, e, f, g, 9, w9);
SHA256ROUND(g, h, a, b, c, d, e, f, 10, w10);
SHA256ROUND(f, g, h, a, b, c, d, e, 11, w11);
SHA256ROUND(e, f, g, h, a, b, c, d, 12, w12);
SHA256ROUND(d, e, f, g, h, a, b, c, 13, w13);
SHA256ROUND(c, d, e, f, g, h, a, b, 14, w14);
SHA256ROUND(b, c, d, e, f, g, h, a, 15, w15);
w0 = add4(SIGMA1_256(w14), w9, SIGMA0_256(w1), w0);
SHA256ROUND(a, b, c, d, e, f, g, h, 16, w0);
w1 = add4(SIGMA1_256(w15), w10, SIGMA0_256(w2), w1);
SHA256ROUND(h, a, b, c, d, e, f, g, 17, w1);
w2 = add4(SIGMA1_256(w0), w11, SIGMA0_256(w3), w2);
SHA256ROUND(g, h, a, b, c, d, e, f, 18, w2);
w3 = add4(SIGMA1_256(w1), w12, SIGMA0_256(w4), w3);
SHA256ROUND(f, g, h, a, b, c, d, e, 19, w3);
w4 = add4(SIGMA1_256(w2), w13, SIGMA0_256(w5), w4);
SHA256ROUND(e, f, g, h, a, b, c, d, 20, w4);
w5 = add4(SIGMA1_256(w3), w14, SIGMA0_256(w6), w5);
SHA256ROUND(d, e, f, g, h, a, b, c, 21, w5);
w6 = add4(SIGMA1_256(w4), w15, SIGMA0_256(w7), w6);
SHA256ROUND(c, d, e, f, g, h, a, b, 22, w6);
w7 = add4(SIGMA1_256(w5), w0, SIGMA0_256(w8), w7);
SHA256ROUND(b, c, d, e, f, g, h, a, 23, w7);
w8 = add4(SIGMA1_256(w6), w1, SIGMA0_256(w9), w8);
SHA256ROUND(a, b, c, d, e, f, g, h, 24, w8);
w9 = add4(SIGMA1_256(w7), w2, SIGMA0_256(w10), w9);
SHA256ROUND(h, a, b, c, d, e, f, g, 25, w9);
w10 = add4(SIGMA1_256(w8), w3, SIGMA0_256(w11), w10);
SHA256ROUND(g, h, a, b, c, d, e, f, 26, w10);
w11 = add4(SIGMA1_256(w9), w4, SIGMA0_256(w12), w11);
SHA256ROUND(f, g, h, a, b, c, d, e, 27, w11);
w12 = add4(SIGMA1_256(w10), w5, SIGMA0_256(w13), w12);
SHA256ROUND(e, f, g, h, a, b, c, d, 28, w12);
w13 = add4(SIGMA1_256(w11), w6, SIGMA0_256(w14), w13);
SHA256ROUND(d, e, f, g, h, a, b, c, 29, w13);
w14 = add4(SIGMA1_256(w12), w7, SIGMA0_256(w15), w14);
SHA256ROUND(c, d, e, f, g, h, a, b, 30, w14);
w15 = add4(SIGMA1_256(w13), w8, SIGMA0_256(w0), w15);
SHA256ROUND(b, c, d, e, f, g, h, a, 31, w15);
w0 = add4(SIGMA1_256(w14), w9, SIGMA0_256(w1), w0);
SHA256ROUND(a, b, c, d, e, f, g, h, 32, w0);
w1 = add4(SIGMA1_256(w15), w10, SIGMA0_256(w2), w1);
SHA256ROUND(h, a, b, c, d, e, f, g, 33, w1);
w2 = add4(SIGMA1_256(w0), w11, SIGMA0_256(w3), w2);
SHA256ROUND(g, h, a, b, c, d, e, f, 34, w2);
w3 = add4(SIGMA1_256(w1), w12, SIGMA0_256(w4), w3);
SHA256ROUND(f, g, h, a, b, c, d, e, 35, w3);
w4 = add4(SIGMA1_256(w2), w13, SIGMA0_256(w5), w4);
SHA256ROUND(e, f, g, h, a, b, c, d, 36, w4);
w5 = add4(SIGMA1_256(w3), w14, SIGMA0_256(w6), w5);
SHA256ROUND(d, e, f, g, h, a, b, c, 37, w5);
w6 = add4(SIGMA1_256(w4), w15, SIGMA0_256(w7), w6);
SHA256ROUND(c, d, e, f, g, h, a, b, 38, w6);
w7 = add4(SIGMA1_256(w5), w0, SIGMA0_256(w8), w7);
SHA256ROUND(b, c, d, e, f, g, h, a, 39, w7);
w8 = add4(SIGMA1_256(w6), w1, SIGMA0_256(w9), w8);
SHA256ROUND(a, b, c, d, e, f, g, h, 40, w8);
w9 = add4(SIGMA1_256(w7), w2, SIGMA0_256(w10), w9);
SHA256ROUND(h, a, b, c, d, e, f, g, 41, w9);
w10 = add4(SIGMA1_256(w8), w3, SIGMA0_256(w11), w10);
SHA256ROUND(g, h, a, b, c, d, e, f, 42, w10);
w11 = add4(SIGMA1_256(w9), w4, SIGMA0_256(w12), w11);
SHA256ROUND(f, g, h, a, b, c, d, e, 43, w11);
w12 = add4(SIGMA1_256(w10), w5, SIGMA0_256(w13), w12);
SHA256ROUND(e, f, g, h, a, b, c, d, 44, w12);
w13 = add4(SIGMA1_256(w11), w6, SIGMA0_256(w14), w13);
SHA256ROUND(d, e, f, g, h, a, b, c, 45, w13);
w14 = add4(SIGMA1_256(w12), w7, SIGMA0_256(w15), w14);
SHA256ROUND(c, d, e, f, g, h, a, b, 46, w14);
w15 = add4(SIGMA1_256(w13), w8, SIGMA0_256(w0), w15);
SHA256ROUND(b, c, d, e, f, g, h, a, 47, w15);
w0 = add4(SIGMA1_256(w14), w9, SIGMA0_256(w1), w0);
SHA256ROUND(a, b, c, d, e, f, g, h, 48, w0);
w1 = add4(SIGMA1_256(w15), w10, SIGMA0_256(w2), w1);
SHA256ROUND(h, a, b, c, d, e, f, g, 49, w1);
w2 = add4(SIGMA1_256(w0), w11, SIGMA0_256(w3), w2);
SHA256ROUND(g, h, a, b, c, d, e, f, 50, w2);
w3 = add4(SIGMA1_256(w1), w12, SIGMA0_256(w4), w3);
SHA256ROUND(f, g, h, a, b, c, d, e, 51, w3);
w4 = add4(SIGMA1_256(w2), w13, SIGMA0_256(w5), w4);
SHA256ROUND(e, f, g, h, a, b, c, d, 52, w4);
w5 = add4(SIGMA1_256(w3), w14, SIGMA0_256(w6), w5);
SHA256ROUND(d, e, f, g, h, a, b, c, 53, w5);
w6 = add4(SIGMA1_256(w4), w15, SIGMA0_256(w7), w6);
SHA256ROUND(c, d, e, f, g, h, a, b, 54, w6);
w7 = add4(SIGMA1_256(w5), w0, SIGMA0_256(w8), w7);
SHA256ROUND(b, c, d, e, f, g, h, a, 55, w7);
w8 = add4(SIGMA1_256(w6), w1, SIGMA0_256(w9), w8);
SHA256ROUND(a, b, c, d, e, f, g, h, 56, w8);
w9 = add4(SIGMA1_256(w7), w2, SIGMA0_256(w10), w9);
SHA256ROUND(h, a, b, c, d, e, f, g, 57, w9);
w10 = add4(SIGMA1_256(w8), w3, SIGMA0_256(w11), w10);
SHA256ROUND(g, h, a, b, c, d, e, f, 58, w10);
w11 = add4(SIGMA1_256(w9), w4, SIGMA0_256(w12), w11);
SHA256ROUND(f, g, h, a, b, c, d, e, 59, w11);
w12 = add4(SIGMA1_256(w10), w5, SIGMA0_256(w13), w12);
SHA256ROUND(e, f, g, h, a, b, c, d, 60, w12);
/* Skip last 3-rounds; not necessary for H==0 */
/*#if 0
w13 = add4(SIGMA1_256(w11), w6, SIGMA0_256(w14), w13);
SHA256ROUND(d, e, f, g, h, a, b, c, 61, w13);
w14 = add4(SIGMA1_256(w12), w7, SIGMA0_256(w15), w14);
SHA256ROUND(c, d, e, f, g, h, a, b, 62, w14);
w15 = add4(SIGMA1_256(w13), w8, SIGMA0_256(w0), w15);
SHA256ROUND(b, c, d, e, f, g, h, a, 63, w15);
#endif*/
/* store resulsts directly in thash */
#define store_2(x,i) \
w0 = (vector unsigned int)(hInit[i],hInit[i],hInit[i],hInit[i]); \
vec_st(vec_add(w0, x), 0 ,&thash[i][k]);
store_2(a, 0);
store_2(b, 1);
store_2(c, 2);
store_2(d, 3);
store_2(e, 4);
store_2(f, 5);
store_2(g, 6);
store_2(h, 7);
vec_st(nonce, 0 ,&thash[8][k]);
/* writing the results into the array is time intensive */
/* -> try if it´s faster to compare the results with the target inside this function */
}
}
static void
SHA256Transform(SHA2_CTX *ctx, const uint8_t *blk)
{
uint32_t a = ctx->state.s32[0];
uint32_t b = ctx->state.s32[1];
uint32_t c = ctx->state.s32[2];
uint32_t d = ctx->state.s32[3];
uint32_t e = ctx->state.s32[4];
uint32_t f = ctx->state.s32[5];
uint32_t g = ctx->state.s32[6];
uint32_t h = ctx->state.s32[7];
uint32_t w0, w1, w2, w3, w4, w5, w6, w7;
uint32_t w8, w9, w10, w11, w12, w13, w14, w15;
uint32_t T1, T2;
#if defined(__sparc)
static const uint32_t sha256_consts[] = {
SHA256_CONST_0, SHA256_CONST_1, SHA256_CONST_2,
SHA256_CONST_3, SHA256_CONST_4, SHA256_CONST_5,
SHA256_CONST_6, SHA256_CONST_7, SHA256_CONST_8,
SHA256_CONST_9, SHA256_CONST_10, SHA256_CONST_11,
SHA256_CONST_12, SHA256_CONST_13, SHA256_CONST_14,
SHA256_CONST_15, SHA256_CONST_16, SHA256_CONST_17,
SHA256_CONST_18, SHA256_CONST_19, SHA256_CONST_20,
SHA256_CONST_21, SHA256_CONST_22, SHA256_CONST_23,
SHA256_CONST_24, SHA256_CONST_25, SHA256_CONST_26,
SHA256_CONST_27, SHA256_CONST_28, SHA256_CONST_29,
SHA256_CONST_30, SHA256_CONST_31, SHA256_CONST_32,
SHA256_CONST_33, SHA256_CONST_34, SHA256_CONST_35,
SHA256_CONST_36, SHA256_CONST_37, SHA256_CONST_38,
SHA256_CONST_39, SHA256_CONST_40, SHA256_CONST_41,
SHA256_CONST_42, SHA256_CONST_43, SHA256_CONST_44,
SHA256_CONST_45, SHA256_CONST_46, SHA256_CONST_47,
SHA256_CONST_48, SHA256_CONST_49, SHA256_CONST_50,
SHA256_CONST_51, SHA256_CONST_52, SHA256_CONST_53,
SHA256_CONST_54, SHA256_CONST_55, SHA256_CONST_56,
SHA256_CONST_57, SHA256_CONST_58, SHA256_CONST_59,
SHA256_CONST_60, SHA256_CONST_61, SHA256_CONST_62,
SHA256_CONST_63
};
#endif /* __sparc */
if ((uintptr_t)blk & 0x3) { /* not 4-byte aligned? */
bcopy(blk, ctx->buf_un.buf32, sizeof (ctx->buf_un.buf32));
blk = (uint8_t *)ctx->buf_un.buf32;
}
/* LINTED E_BAD_PTR_CAST_ALIGN */
w0 = LOAD_BIG_32(blk + 4 * 0);
SHA256ROUND(a, b, c, d, e, f, g, h, 0, w0);
/* LINTED E_BAD_PTR_CAST_ALIGN */
w1 = LOAD_BIG_32(blk + 4 * 1);
SHA256ROUND(h, a, b, c, d, e, f, g, 1, w1);
/* LINTED E_BAD_PTR_CAST_ALIGN */
w2 = LOAD_BIG_32(blk + 4 * 2);
SHA256ROUND(g, h, a, b, c, d, e, f, 2, w2);
/* LINTED E_BAD_PTR_CAST_ALIGN */
w3 = LOAD_BIG_32(blk + 4 * 3);
SHA256ROUND(f, g, h, a, b, c, d, e, 3, w3);
/* LINTED E_BAD_PTR_CAST_ALIGN */
w4 = LOAD_BIG_32(blk + 4 * 4);
SHA256ROUND(e, f, g, h, a, b, c, d, 4, w4);
/* LINTED E_BAD_PTR_CAST_ALIGN */
w5 = LOAD_BIG_32(blk + 4 * 5);
SHA256ROUND(d, e, f, g, h, a, b, c, 5, w5);
/* LINTED E_BAD_PTR_CAST_ALIGN */
w6 = LOAD_BIG_32(blk + 4 * 6);
SHA256ROUND(c, d, e, f, g, h, a, b, 6, w6);
/* LINTED E_BAD_PTR_CAST_ALIGN */
w7 = LOAD_BIG_32(blk + 4 * 7);
SHA256ROUND(b, c, d, e, f, g, h, a, 7, w7);
/* LINTED E_BAD_PTR_CAST_ALIGN */
w8 = LOAD_BIG_32(blk + 4 * 8);
SHA256ROUND(a, b, c, d, e, f, g, h, 8, w8);
/* LINTED E_BAD_PTR_CAST_ALIGN */
w9 = LOAD_BIG_32(blk + 4 * 9);
SHA256ROUND(h, a, b, c, d, e, f, g, 9, w9);
/* LINTED E_BAD_PTR_CAST_ALIGN */
w10 = LOAD_BIG_32(blk + 4 * 10);
SHA256ROUND(g, h, a, b, c, d, e, f, 10, w10);
/* LINTED E_BAD_PTR_CAST_ALIGN */
w11 = LOAD_BIG_32(blk + 4 * 11);
SHA256ROUND(f, g, h, a, b, c, d, e, 11, w11);
/* LINTED E_BAD_PTR_CAST_ALIGN */
w12 = LOAD_BIG_32(blk + 4 * 12);
SHA256ROUND(e, f, g, h, a, b, c, d, 12, w12);
/* LINTED E_BAD_PTR_CAST_ALIGN */
w13 = LOAD_BIG_32(blk + 4 * 13);
SHA256ROUND(d, e, f, g, h, a, b, c, 13, w13);
/* LINTED E_BAD_PTR_CAST_ALIGN */
w14 = LOAD_BIG_32(blk + 4 * 14);
SHA256ROUND(c, d, e, f, g, h, a, b, 14, w14);
/* LINTED E_BAD_PTR_CAST_ALIGN */
w15 = LOAD_BIG_32(blk + 4 * 15);
SHA256ROUND(b, c, d, e, f, g, h, a, 15, w15);
w0 = SIGMA1_256(w14) + w9 + SIGMA0_256(w1) + w0;
SHA256ROUND(a, b, c, d, e, f, g, h, 16, w0);
w1 = SIGMA1_256(w15) + w10 + SIGMA0_256(w2) + w1;
SHA256ROUND(h, a, b, c, d, e, f, g, 17, w1);
w2 = SIGMA1_256(w0) + w11 + SIGMA0_256(w3) + w2;
SHA256ROUND(g, h, a, b, c, d, e, f, 18, w2);
w3 = SIGMA1_256(w1) + w12 + SIGMA0_256(w4) + w3;
SHA256ROUND(f, g, h, a, b, c, d, e, 19, w3);
w4 = SIGMA1_256(w2) + w13 + SIGMA0_256(w5) + w4;
SHA256ROUND(e, f, g, h, a, b, c, d, 20, w4);
w5 = SIGMA1_256(w3) + w14 + SIGMA0_256(w6) + w5;
SHA256ROUND(d, e, f, g, h, a, b, c, 21, w5);
w6 = SIGMA1_256(w4) + w15 + SIGMA0_256(w7) + w6;
SHA256ROUND(c, d, e, f, g, h, a, b, 22, w6);
w7 = SIGMA1_256(w5) + w0 + SIGMA0_256(w8) + w7;
SHA256ROUND(b, c, d, e, f, g, h, a, 23, w7);
w8 = SIGMA1_256(w6) + w1 + SIGMA0_256(w9) + w8;
SHA256ROUND(a, b, c, d, e, f, g, h, 24, w8);
w9 = SIGMA1_256(w7) + w2 + SIGMA0_256(w10) + w9;
SHA256ROUND(h, a, b, c, d, e, f, g, 25, w9);
w10 = SIGMA1_256(w8) + w3 + SIGMA0_256(w11) + w10;
SHA256ROUND(g, h, a, b, c, d, e, f, 26, w10);
w11 = SIGMA1_256(w9) + w4 + SIGMA0_256(w12) + w11;
SHA256ROUND(f, g, h, a, b, c, d, e, 27, w11);
w12 = SIGMA1_256(w10) + w5 + SIGMA0_256(w13) + w12;
SHA256ROUND(e, f, g, h, a, b, c, d, 28, w12);
w13 = SIGMA1_256(w11) + w6 + SIGMA0_256(w14) + w13;
SHA256ROUND(d, e, f, g, h, a, b, c, 29, w13);
w14 = SIGMA1_256(w12) + w7 + SIGMA0_256(w15) + w14;
SHA256ROUND(c, d, e, f, g, h, a, b, 30, w14);
w15 = SIGMA1_256(w13) + w8 + SIGMA0_256(w0) + w15;
SHA256ROUND(b, c, d, e, f, g, h, a, 31, w15);
w0 = SIGMA1_256(w14) + w9 + SIGMA0_256(w1) + w0;
SHA256ROUND(a, b, c, d, e, f, g, h, 32, w0);
w1 = SIGMA1_256(w15) + w10 + SIGMA0_256(w2) + w1;
SHA256ROUND(h, a, b, c, d, e, f, g, 33, w1);
w2 = SIGMA1_256(w0) + w11 + SIGMA0_256(w3) + w2;
SHA256ROUND(g, h, a, b, c, d, e, f, 34, w2);
w3 = SIGMA1_256(w1) + w12 + SIGMA0_256(w4) + w3;
SHA256ROUND(f, g, h, a, b, c, d, e, 35, w3);
w4 = SIGMA1_256(w2) + w13 + SIGMA0_256(w5) + w4;
SHA256ROUND(e, f, g, h, a, b, c, d, 36, w4);
w5 = SIGMA1_256(w3) + w14 + SIGMA0_256(w6) + w5;
SHA256ROUND(d, e, f, g, h, a, b, c, 37, w5);
w6 = SIGMA1_256(w4) + w15 + SIGMA0_256(w7) + w6;
SHA256ROUND(c, d, e, f, g, h, a, b, 38, w6);
w7 = SIGMA1_256(w5) + w0 + SIGMA0_256(w8) + w7;
SHA256ROUND(b, c, d, e, f, g, h, a, 39, w7);
w8 = SIGMA1_256(w6) + w1 + SIGMA0_256(w9) + w8;
SHA256ROUND(a, b, c, d, e, f, g, h, 40, w8);
w9 = SIGMA1_256(w7) + w2 + SIGMA0_256(w10) + w9;
SHA256ROUND(h, a, b, c, d, e, f, g, 41, w9);
w10 = SIGMA1_256(w8) + w3 + SIGMA0_256(w11) + w10;
SHA256ROUND(g, h, a, b, c, d, e, f, 42, w10);
w11 = SIGMA1_256(w9) + w4 + SIGMA0_256(w12) + w11;
SHA256ROUND(f, g, h, a, b, c, d, e, 43, w11);
w12 = SIGMA1_256(w10) + w5 + SIGMA0_256(w13) + w12;
SHA256ROUND(e, f, g, h, a, b, c, d, 44, w12);
w13 = SIGMA1_256(w11) + w6 + SIGMA0_256(w14) + w13;
SHA256ROUND(d, e, f, g, h, a, b, c, 45, w13);
w14 = SIGMA1_256(w12) + w7 + SIGMA0_256(w15) + w14;
SHA256ROUND(c, d, e, f, g, h, a, b, 46, w14);
w15 = SIGMA1_256(w13) + w8 + SIGMA0_256(w0) + w15;
SHA256ROUND(b, c, d, e, f, g, h, a, 47, w15);
w0 = SIGMA1_256(w14) + w9 + SIGMA0_256(w1) + w0;
SHA256ROUND(a, b, c, d, e, f, g, h, 48, w0);
w1 = SIGMA1_256(w15) + w10 + SIGMA0_256(w2) + w1;
SHA256ROUND(h, a, b, c, d, e, f, g, 49, w1);
w2 = SIGMA1_256(w0) + w11 + SIGMA0_256(w3) + w2;
SHA256ROUND(g, h, a, b, c, d, e, f, 50, w2);
w3 = SIGMA1_256(w1) + w12 + SIGMA0_256(w4) + w3;
SHA256ROUND(f, g, h, a, b, c, d, e, 51, w3);
w4 = SIGMA1_256(w2) + w13 + SIGMA0_256(w5) + w4;
SHA256ROUND(e, f, g, h, a, b, c, d, 52, w4);
w5 = SIGMA1_256(w3) + w14 + SIGMA0_256(w6) + w5;
SHA256ROUND(d, e, f, g, h, a, b, c, 53, w5);
w6 = SIGMA1_256(w4) + w15 + SIGMA0_256(w7) + w6;
SHA256ROUND(c, d, e, f, g, h, a, b, 54, w6);
w7 = SIGMA1_256(w5) + w0 + SIGMA0_256(w8) + w7;
SHA256ROUND(b, c, d, e, f, g, h, a, 55, w7);
w8 = SIGMA1_256(w6) + w1 + SIGMA0_256(w9) + w8;
SHA256ROUND(a, b, c, d, e, f, g, h, 56, w8);
w9 = SIGMA1_256(w7) + w2 + SIGMA0_256(w10) + w9;
SHA256ROUND(h, a, b, c, d, e, f, g, 57, w9);
w10 = SIGMA1_256(w8) + w3 + SIGMA0_256(w11) + w10;
SHA256ROUND(g, h, a, b, c, d, e, f, 58, w10);
w11 = SIGMA1_256(w9) + w4 + SIGMA0_256(w12) + w11;
SHA256ROUND(f, g, h, a, b, c, d, e, 59, w11);
w12 = SIGMA1_256(w10) + w5 + SIGMA0_256(w13) + w12;
SHA256ROUND(e, f, g, h, a, b, c, d, 60, w12);
w13 = SIGMA1_256(w11) + w6 + SIGMA0_256(w14) + w13;
SHA256ROUND(d, e, f, g, h, a, b, c, 61, w13);
w14 = SIGMA1_256(w12) + w7 + SIGMA0_256(w15) + w14;
SHA256ROUND(c, d, e, f, g, h, a, b, 62, w14);
w15 = SIGMA1_256(w13) + w8 + SIGMA0_256(w0) + w15;
SHA256ROUND(b, c, d, e, f, g, h, a, 63, w15);
ctx->state.s32[0] += a;
ctx->state.s32[1] += b;
ctx->state.s32[2] += c;
ctx->state.s32[3] += d;
ctx->state.s32[4] += e;
ctx->state.s32[5] += f;
ctx->state.s32[6] += g;
ctx->state.s32[7] += h;
}
