static void scrypt_1024_1_1_256_3way(const uint32_t *input,
uint32_t *output, uint32_t *midstate, unsigned char *scratchpad, int N)
{
uint32_t tstate[3 * 8], ostate[3 * 8];
uint32_t X[3 * 32] __attribute__((aligned(64)));
uint32_t *V;
V = (uint32_t *)(((uintptr_t)(scratchpad) + 63) & ~ (uintptr_t)(63));
memcpy(tstate + 0, midstate, 32);
memcpy(tstate + 8, midstate, 32);
memcpy(tstate + 16, midstate, 32);
HMAC_SHA256_80_init(input + 0, tstate + 0, ostate + 0);
HMAC_SHA256_80_init(input + 20, tstate + 8, ostate + 8);
HMAC_SHA256_80_init(input + 40, tstate + 16, ostate + 16);
PBKDF2_SHA256_80_128(tstate + 0, ostate + 0, input + 0, X + 0);
PBKDF2_SHA256_80_128(tstate + 8, ostate + 8, input + 20, X + 32);
PBKDF2_SHA256_80_128(tstate + 16, ostate + 16, input + 40, X + 64);
scrypt_core_3way(X, V, N);
PBKDF2_SHA256_128_32(tstate + 0, ostate + 0, X + 0, output + 0);
PBKDF2_SHA256_128_32(tstate + 8, ostate + 8, X + 32, output + 8);
PBKDF2_SHA256_128_32(tstate + 16, ostate + 16, X + 64, output + 16);
}
static void scrypt_1024_1_1_256(const uint32_t *input, uint32_t *output,
uint32_t *midstate, unsigned char *scratchpad, int N)
{
uint32_t tstate[8], ostate[8];
uint32_t X[32] __attribute__((aligned(128)));
uint32_t *V;
V = (uint32_t *)(((uintptr_t)(scratchpad) + 63) & ~ (uintptr_t)(63));
memcpy(tstate, midstate, 32);
HMAC_SHA256_80_init(input, tstate, ostate);
PBKDF2_SHA256_80_128(tstate, ostate, input, X);
scrypt_core(X, V, N);
PBKDF2_SHA256_128_32(tstate, ostate, X, output);
}
/* cpu and memory intensive function to transform a 80 byte buffer into a 32 byte output
scratchpad size needs to be at least 63 + (128 * r * p) + (256 * r + 64) + (128 * r * N) bytes
*/
static uint32_t scrypt_1024_1_1_256_sp(const uint32_t* input, char* scratchpad)
{
uint32_t * V;
uint32_t X[32];
uint32_t i;
uint32_t j;
uint32_t k;
uint64_t *p1, *p2;
p1 = (uint64_t *)X;
V = (uint32_t *)(((uintptr_t)(scratchpad) + 63) & ~ (uintptr_t)(63));
PBKDF2_SHA256_80_128(input, X);
for (i = 0; i < 1024; i += 2) {
memcpy(&V[i * 32], X, 128);
salsa20_8(&X[0], &X[16]);
salsa20_8(&X[16], &X[0]);
memcpy(&V[(i + 1) * 32], X, 128);
salsa20_8(&X[0], &X[16]);
salsa20_8(&X[16], &X[0]);
}
for (i = 0; i < 1024; i += 2) {
j = X[16] & 1023;
p2 = (uint64_t *)(&V[j * 32]);
for(k = 0; k < 16; k++)
p1[k] ^= p2[k];
salsa20_8(&X[0], &X[16]);
salsa20_8(&X[16], &X[0]);
j = X[16] & 1023;
p2 = (uint64_t *)(&V[j * 32]);
for(k = 0; k < 16; k++)
p1[k] ^= p2[k];
salsa20_8(&X[0], &X[16]);
salsa20_8(&X[16], &X[0]);
}
return PBKDF2_SHA256_80_128_32(input, X);
}
static void SHA256_before(uint32_t *tstate, uint32_t *ostate, uint32_t *input, uint32_t* X){
HMAC_SHA256_80_init(input, tstate, ostate);
PBKDF2_SHA256_80_128(tstate, ostate, input, X);
}
