static void sha256_compress(unsigned int* iv, const uint8_t* data) {
unsigned int a, b, c, d, e, f, g, h;
unsigned int s0, s1;
unsigned int t1, t2;
unsigned int work_space[16];
unsigned int n;
unsigned int i;
a = iv[0];
b = iv[1];
c = iv[2];
d = iv[3];
e = iv[4];
f = iv[5];
g = iv[6];
h = iv[7];
for (i = 0; i < 16; ++i) {
n = BigEndian(&data);
t1 = work_space[i] = n;
t1 += h + Sigma1(e) + Ch(e, f, g) + k256[i];
t2 = Sigma0(a) + Maj(a, b, c);
h = g;
g = f;
f = e;
e = d + t1;
d = c;
c = b;
b = a;
a = t1 + t2;
}
for (; i < 64; ++i) {
s0 = work_space[(i + 1) & 0x0f];
s0 = sigma0(s0);
s1 = work_space[(i + 14) & 0x0f];
s1 = sigma1(s1);
t1 = work_space[i & 0xf] += s0 + s1 + work_space[(i + 9) & 0xf];
t1 += h + Sigma1(e) + Ch(e, f, g) + k256[i];
t2 = Sigma0(a) + Maj(a, b, c);
h = g;
g = f;
f = e;
e = d + t1;
d = c;
c = b;
b = a;
a = t1 + t2;
}
iv[0] += a;
iv[1] += b;
iv[2] += c;
iv[3] += d;
iv[4] += e;
iv[5] += f;
iv[6] += g;
iv[7] += h;
}
static void sha256_block (SHA256_CTX *ctx, const void *in, size_t num, int host)
{
unsigned MD32_REG_T a,b,c,d,e,f,g,h,s0,s1,T1,T2;
SHA_LONG X[16];
int i;
const unsigned char *data=in;
while (num--) {
a = ctx->h[0]; b = ctx->h[1]; c = ctx->h[2]; d = ctx->h[3];
e = ctx->h[4]; f = ctx->h[5]; g = ctx->h[6]; h = ctx->h[7];
if (host)
{
const SHA_LONG *W=(const SHA_LONG *)data;
for (i=0;i<16;i++)
{
T1 = X[i] = W[i];
T1 += h + Sigma1(e) + Ch(e,f,g) + K256[i];
T2 = Sigma0(a) + Maj(a,b,c);
h = g; g = f; f = e; e = d + T1;
d = c; c = b; b = a; a = T1 + T2;
}
data += SHA256_CBLOCK;
}
else
{
SHA_LONG l;
for (i=0;i<16;i++)
{
HOST_c2l(data,l); T1 = X[i] = l;
T1 += h + Sigma1(e) + Ch(e,f,g) + K256[i];
T2 = Sigma0(a) + Maj(a,b,c);
h = g; g = f; f = e; e = d + T1;
d = c; c = b; b = a; a = T1 + T2;
}
}
for (;i<64;i++)
{
s0 = X[(i+1)&0x0f]; s0 = sigma0(s0);
s1 = X[(i+14)&0x0f]; s1 = sigma1(s1);
T1 = X[i&0xf] += s0 + s1 + X[(i+9)&0xf];
T1 += h + Sigma1(e) + Ch(e,f,g) + K256[i];
T2 = Sigma0(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;
}
}
/**
* sha512 compression function - 32-bit machines
* @param res The resulting hash value
* @param hash The chaining input value
* @param in The message input
*/
void sha512_comp (hashblock res, const hashblock hash, const messageblock in)
{
const uint64_t *W=in;
uint64_t A,E,T;
uint64_t X[9+80],*F;
uint64_t H[8];
int i;
for (i = 0; i < SHA512_DIGEST_LENGTH/8; i++) {
H[i]=PULL64(hash[i*8]);
}
F = X+80;
A = H[0]; F[1] = H[1];
F[2] = H[2]; F[3] = H[3];
E = H[4]; F[5] = H[5];
F[6] = H[6]; F[7] = H[7];
for (i=0;i<16;i++,F--)
{
#ifdef B_ENDIAN
T = W[i];
#else
T = PULL64(W[i]);
#endif
F[0] = A;
F[4] = E;
F[8] = T;
T += F[7] + Sigma1(E) + Ch(E,F[5],F[6]) + K512[i];
E = F[3] + T;
A = T + Sigma0(A) + Maj(A,F[1],F[2]);
}
for (;i<80;i++,F--)
{
T = sigma0(F[8+16-1]);
T += sigma1(F[8+16-14]);
T += F[8+16] + F[8+16-9];
F[0] = A;
F[4] = E;
F[8] = T;
T += F[7] + Sigma1(E) + Ch(E,F[5],F[6]) + K512[i];
E = F[3] + T;
A = T + Sigma0(A) + Maj(A,F[1],F[2]);
}
H[0] += A; H[1] += F[1];
H[2] += F[2]; H[3] += F[3];
H[4] += E; H[5] += F[5];
H[6] += F[6]; H[7] += F[7];
for (i = 0; i < SHA512_DIGEST_LENGTH/8; i++) {
PUSH64(H[i],res[i*8]);
}
}
/*
* This code should give better results on 32-bit CPU with less than
* ~24 registers, both size and performance wise...
*/
void sha512_block_data_order(uint64_t *state, const uint64_t *W, size_t num) {
uint64_t A, E, T;
uint64_t X[9 + 80], *F;
int i;
while (num--) {
F = X + 80;
A = state[0];
F[1] = state[1];
F[2] = state[2];
F[3] = state[3];
E = state[4];
F[5] = state[5];
F[6] = state[6];
F[7] = state[7];
for (i = 0; i < 16; i++, F--) {
T = from_be_u64(W[i]);
F[0] = A;
F[4] = E;
F[8] = T;
T += F[7] + Sigma1(E) + Ch(E, F[5], F[6]) + K512[i];
E = F[3] + T;
A = T + Sigma0(A) + Maj(A, F[1], F[2]);
}
for (; i < 80; i++, F--) {
T = sigma0(F[8 + 16 - 1]);
T += sigma1(F[8 + 16 - 14]);
T += F[8 + 16] + F[8 + 16 - 9];
F[0] = A;
F[4] = E;
F[8] = T;
T += F[7] + Sigma1(E) + Ch(E, F[5], F[6]) + K512[i];
E = F[3] + T;
A = T + Sigma0(A) + Maj(A, F[1], F[2]);
}
state[0] += A;
state[1] += F[1];
state[2] += F[2];
state[3] += F[3];
state[4] += E;
state[5] += F[5];
state[6] += F[6];
state[7] += F[7];
W += 16;
}
}
/**
* sha512 compression function - 64-bit machines
* @param res The resulting hash value
* @param hash The chaining input value
* @param in The message input
*/
void sha512_comp (hashblock res, const hashblock hash, const messageblock in)
{
// CHANGE type casting added due to c++
const uint64_t *W=reinterpret_cast<const uint64_t*>(in);
uint64_t a,b,c,d,e,f,g,h,s0,s1,T1,T2;
uint64_t X[16];
uint64_t H[8];
int i;
for (i = 0; i < SHA512_DIGEST_LENGTH/8; i++) {
H[i]=PULL64(hash[i*8]);
}
a = H[0]; b = H[1]; c = H[2]; d = H[3];
e = H[4]; f = H[5]; g = H[6]; h = H[7];
for (i=0;i<16;i++)
{
#ifdef B_ENDIAN
T1 = X[i] = W[i];
#else
T1 = X[i] = PULL64(W[i]);
#endif
T1 += h + Sigma1(e) + Ch(e,f,g) + K512[i];
T2 = Sigma0(a) + Maj(a,b,c);
h = g; g = f; f = e; e = d + T1;
d = c; c = b; b = a; a = T1 + T2;
}
for (;i<80;i++)
{
s0 = X[(i+1)&0x0f]; s0 = sigma0(s0);
s1 = X[(i+14)&0x0f]; s1 = sigma1(s1);
T1 = X[i&0xf] += s0 + s1 + X[(i+9)&0xf];
T1 += h + Sigma1(e) + Ch(e,f,g) + K512[i];
T2 = Sigma0(a) + Maj(a,b,c);
h = g; g = f; f = e; e = d + T1;
d = c; c = b; b = a; a = T1 + T2;
}
H[0] += a; H[1] += b; H[2] += c; H[3] += d;
H[4] += e; H[5] += f; H[6] += g; H[7] += h;
for (i = 0; i < SHA512_DIGEST_LENGTH/8; i++) {
PUSH64(H[i],res[i*8]);
}
}
void setupCasadiVars(const std::vector<Matrix<C_DIM> >& X, const std::vector<Matrix<U_DIM> >& U, double* XU_arr, double* Sigma0_arr, double* l_arr, double* params_arr)
{
int index = 0;
for(int t = 0; t < T-1; ++t) {
for(int i=0; i < C_DIM; ++i) {
XU_arr[index++] = X[t][i];
}
for(int i=0; i < U_DIM; ++i) {
XU_arr[index++] = U[t][i];
}
}
for(int i=0; i < C_DIM; ++i) {
XU_arr[index++] = X[T-1][i];
}
Matrix<X_DIM,X_DIM> Sigma0 = SqrtSigma0*SqrtSigma0;
index = 0;
for(int i=0; i < X_DIM; ++i) {
for(int j=0; j < X_DIM; ++j) {
Sigma0_arr[index++] = Sigma0(i,j);
}
}
index = 0;
for(int i=C_DIM; i < X_DIM; ++i) {
l_arr[index++] = x0[i];
}
params_arr[0] = alpha_belief;
params_arr[1] = alpha_control;
params_arr[2] = alpha_final_belief;
}
/** One round of SHA-256. */
void inline __attribute__((always_inline)) Round(__m256i a, __m256i b, __m256i c, __m256i& d, __m256i e, __m256i f, __m256i g, __m256i& h, __m256i k)
{
__m256i t1 = Add(h, Sigma1(e), Ch(e, f, g), k);
__m256i t2 = Add(Sigma0(a), Maj(a, b, c));
d = Add(d, t1);
h = Add(t1, t2);
}
/*****************************************
* sha256 compression function *
* *
* H points to chaining input *
* in points to the message input *
* *
*****************************************/
void sha256_comp (hashblock res, const hashblock hash, const void *in)
{
uint32_t a,b,c,d,e,f,g,h,s0,s1,T1,T2;
uint32_t H[8];
uint32_t X[16],l;
int i;
// CHANGE type casting added due to c++
const unsigned char *data=static_cast<const unsigned char*>(in);
for (i = 0; i < SHA256_DIGEST_LENGTH/4; i++) {
HOST_c2l(hash, H[i]);
}
a = H[0]; b = H[1]; c = H[2]; d = H[3];
e = H[4]; f = H[5]; g = H[6]; h = H[7];
for (i=0;i<16;i++)
{
HOST_c2l(data,l); T1 = X[i] = l;
T1 += h + Sigma1(e) + Ch(e,f,g) + K256[i];
T2 = Sigma0(a) + Maj(a,b,c);
h = g;
g = f;
f = e;
e = d + T1;
d = c; c = b; b = a; a = T1 + T2;
}
for (;i<64;i++)
{
s0 = X[(i+1)&0x0f]; s0 = sigma0(s0);
s1 = X[(i+14)&0x0f]; s1 = sigma1(s1);
T1 = X[i&0xf] += s0 + s1 + X[(i+9)&0xf];
T1 += h + Sigma1(e) + Ch(e,f,g) + K256[i];
T2 = Sigma0(a) + Maj(a,b,c);
h = g; g = f; f = e; e = d + T1;
d = c; c = b; b = a; a = T1 + T2;
}
H[0] += a; H[1] += b; H[2] += c; H[3] += d;
H[4] += e; H[5] += f; H[6] += g; H[7] += h;
for (i = 0; i < SHA256_DIGEST_LENGTH/4; i++) {
HOST_l2c(H[i], res);
}
}
void sha512_compress(psDigestContext_t * md, unsigned char *buf)
#endif
{
uint64 S[8], W[80], t0, t1;
int i;
/* copy state into S */
for (i = 0; i < 8; i++) {
S[i] = md->sha512.state[i];
}
/* copy the state into 1024-bits into W[0..15] */
for (i = 0; i < 16; i++) {
LOAD64H(W[i], buf + (8*i));
}
/* fill W[16..79] */
for (i = 16; i < 80; i++) {
W[i] = Gamma1(W[i - 2]) + W[i - 7] + Gamma0(W[i - 15]) + W[i - 16];
}
/* Compress */
#ifndef PS_SHA512_IMPROVE_PERF_INCREASE_CODESIZE
for (i = 0; i < 80; i++) {
t0 = S[7] + Sigma1(S[4]) + Ch(S[4], S[5], S[6]) + K[i] + W[i];
t1 = Sigma0(S[0]) + Maj(S[0], S[1], S[2]);
S[7] = S[6];
S[6] = S[5];
S[5] = S[4];
S[4] = S[3] + t0;
S[3] = S[2];
S[2] = S[1];
S[1] = S[0];
S[0] = t0 + t1;
}
#else
#define RND(a,b,c,d,e,f,g,h,i) \
t0 = h + Sigma1(e) + Ch(e, f, g) + K[i] + W[i]; \
t1 = Sigma0(a) + Maj(a, b, c); \
d += t0; \
h = t0 + t1;
for (i = 0; i < 80; i += 8) {
RND(S[0],S[1],S[2],S[3],S[4],S[5],S[6],S[7],i+0);
RND(S[7],S[0],S[1],S[2],S[3],S[4],S[5],S[6],i+1);
RND(S[6],S[7],S[0],S[1],S[2],S[3],S[4],S[5],i+2);
RND(S[5],S[6],S[7],S[0],S[1],S[2],S[3],S[4],i+3);
RND(S[4],S[5],S[6],S[7],S[0],S[1],S[2],S[3],i+4);
RND(S[3],S[4],S[5],S[6],S[7],S[0],S[1],S[2],i+5);
RND(S[2],S[3],S[4],S[5],S[6],S[7],S[0],S[1],i+6);
RND(S[1],S[2],S[3],S[4],S[5],S[6],S[7],S[0],i+7);
}
#endif /* PS_SHA512_IMPROVE_PERF_INCREASE_CODESIZE */
/* feedback */
for (i = 0; i < 8; i++) {
md->sha512.state[i] = md->sha512.state[i] + S[i];
}
}
void sha256_block_data_order (SHA256_CTX *ctx, const void *in)
{
unsigned MD32_REG_T a,b,c,d,e,f,g,h,s0,s1,T1,T2,t;
SHA_LONG X[16],l,Ki;
int i;
const unsigned char *data=in;
a = ctx->h[0]; b = ctx->h[1]; c = ctx->h[2]; d = ctx->h[3];
e = ctx->h[4]; f = ctx->h[5]; g = ctx->h[6]; h = ctx->h[7];
for (i=0;i<16;i++)
{
HOST_c2l(data,l); X[i] = l;
Ki=K256[i];
T1 = l + h + Sigma1(e) + Ch(e,f,g) + Ki;
T2 = Sigma0(a) + Maj(a,b,c);
h = g; g = f; f = e; e = d + T1;
d = c; c = b; b = a; a = T1 + T2;
}
for (;i<64;i++)
{
s0 = X[(i+1)&0x0f]; s0 = sigma0(s0);
s1 = X[(i+14)&0x0f]; s1 = sigma1(s1);
T1 = X[i&0xf];
t = X[(i+9)&0xf];
T1 += s0 + s1 + t;
X[i&0xf] = T1;
Ki=K256[i];
T1 += h + Sigma1(e) + Ch(e,f,g) + Ki;
T2 = Sigma0(a) + Maj(a,b,c);
h = g; g = f; f = e; e = d + T1;
d = c; c = b; b = a; a = T1 + T2;
}
t=ctx->h[0]; ctx->h[0]=t+a;
t=ctx->h[1]; ctx->h[1]=t+b;
t=ctx->h[2]; ctx->h[2]=t+c;
t=ctx->h[3]; ctx->h[3]=t+d;
t=ctx->h[4]; ctx->h[4]=t+e;
t=ctx->h[5]; ctx->h[5]=t+f;
t=ctx->h[6]; ctx->h[6]=t+g;
t=ctx->h[7]; ctx->h[7]=t+h;
return;
}
static void sha256_compress(hash_state * md)
#endif
{
unsigned long S[8], W[64], t0, t1;
int i;
_ARGCHK(md != NULL);
/* copy state into S */
for (i = 0; i < 8; i++)
S[i] = md->sha256.state[i];
/* copy the state into 512-bits into W[0..15] */
for (i = 0; i < 16; i++) {
LOAD32H(W[i], md->sha256.buf + (4*i));
}
/* fill W[16..63] */
for (i = 16; i < 64; i++) {
W[i] = Gamma1(W[i - 2]) + W[i - 7] + Gamma0(W[i - 15]) + W[i - 16];
}
/* Compress */
for (i = 0; i < 64; i++) {
t0 = S[7] + Sigma1(S[4]) + Ch(S[4], S[5], S[6]) + K[i] + W[i];
t1 = Sigma0(S[0]) + Maj(S[0], S[1], S[2]);
S[7] = S[6];
S[6] = S[5];
S[5] = S[4];
S[4] = S[3] + t0;
S[3] = S[2];
S[2] = S[1];
S[1] = S[0];
S[0] = t0 + t1;
}
/* feedback */
for (i = 0; i < 8; i++) {
md->sha256.state[i] = md->sha256.state[i] + S[i];
}
}
static int sha512_compress(hash_state * md, unsigned char *buf)
#endif
{
ulong64 S[8], W[80], t0, t1;
int i;
/* copy state into S */
for (i = 0; i < 8; i++) {
S[i] = md->sha512.state[i];
}
/* copy the state into 1024-bits into W[0..15] */
for (i = 0; i < 16; i++) {
LOAD64H(W[i], buf + (8*i));
}
/* fill W[16..79] */
for (i = 16; i < 80; i++) {
W[i] = Gamma1(W[i - 2]) + W[i - 7] + Gamma0(W[i - 15]) + W[i - 16];
}
/* Compress */
#ifdef LTC_SMALL_CODE
for (i = 0; i < 80; i++) {
t0 = S[7] + Sigma1(S[4]) + Ch(S[4], S[5], S[6]) + K[i] + W[i];
t1 = Sigma0(S[0]) + Maj(S[0], S[1], S[2]);
S[7] = S[6];
S[6] = S[5];
S[5] = S[4];
S[4] = S[3] + t0;
S[3] = S[2];
S[2] = S[1];
S[1] = S[0];
S[0] = t0 + t1;
}
#else
#define RND(a,b,c,d,e,f,g,h,i) \
t0 = h + Sigma1(e) + Ch(e, f, g) + K[i] + W[i]; \
t1 = Sigma0(a) + Maj(a, b, c); \
d += t0; \
h = t0 + t1;
for (i = 0; i < 80; i += 8) {
RND(S[0],S[1],S[2],S[3],S[4],S[5],S[6],S[7],i+0);
RND(S[7],S[0],S[1],S[2],S[3],S[4],S[5],S[6],i+1);
RND(S[6],S[7],S[0],S[1],S[2],S[3],S[4],S[5],i+2);
RND(S[5],S[6],S[7],S[0],S[1],S[2],S[3],S[4],i+3);
RND(S[4],S[5],S[6],S[7],S[0],S[1],S[2],S[3],i+4);
RND(S[3],S[4],S[5],S[6],S[7],S[0],S[1],S[2],i+5);
RND(S[2],S[3],S[4],S[5],S[6],S[7],S[0],S[1],i+6);
RND(S[1],S[2],S[3],S[4],S[5],S[6],S[7],S[0],i+7);
}
#endif
/* feedback */
for (i = 0; i < 8; i++) {
md->sha512.state[i] = md->sha512.state[i] + S[i];
}
return CRYPT_OK;
}
static inline void sha256_transform(__m128i *state, __m128i *block, __m128i *dst)
{
__m128i W[64], t1, t2;
W[0] = block[ 0];
W[1] = block[ 1];
W[2] = block[ 2];
W[3] = block[ 3];
W[4] = block[ 4];
W[5] = block[ 5];
W[6] = block[ 6];
W[7] = block[ 7];
W[8] = block[ 8];
W[9] = block[ 9];
W[10] = block[10];
W[11] = block[11];
W[12] = block[12];
W[13] = block[13];
W[14] = block[14];
W[15] = block[15];
W[16] = add4(sigma1(W[16 - 2]), W[16 - 7], sigma0(W[16 - 15]), W[16 - 16]);
W[17] = add4(sigma1(W[17 - 2]), W[17 - 7], sigma0(W[17 - 15]), W[17 - 16]);
W[18] = add4(sigma1(W[18 - 2]), W[18 - 7], sigma0(W[18 - 15]), W[18 - 16]);
W[19] = add4(sigma1(W[19 - 2]), W[19 - 7], sigma0(W[19 - 15]), W[19 - 16]);
W[20] = add4(sigma1(W[20 - 2]), W[20 - 7], sigma0(W[20 - 15]), W[20 - 16]);
W[21] = add4(sigma1(W[21 - 2]), W[21 - 7], sigma0(W[21 - 15]), W[21 - 16]);
W[22] = add4(sigma1(W[22 - 2]), W[22 - 7], sigma0(W[22 - 15]), W[22 - 16]);
W[23] = add4(sigma1(W[23 - 2]), W[23 - 7], sigma0(W[23 - 15]), W[23 - 16]);
W[24] = add4(sigma1(W[24 - 2]), W[24 - 7], sigma0(W[24 - 15]), W[24 - 16]);
W[25] = add4(sigma1(W[25 - 2]), W[25 - 7], sigma0(W[25 - 15]), W[25 - 16]);
W[26] = add4(sigma1(W[26 - 2]), W[26 - 7], sigma0(W[26 - 15]), W[26 - 16]);
W[27] = add4(sigma1(W[27 - 2]), W[27 - 7], sigma0(W[27 - 15]), W[27 - 16]);
W[28] = add4(sigma1(W[28 - 2]), W[28 - 7], sigma0(W[28 - 15]), W[28 - 16]);
W[29] = add4(sigma1(W[29 - 2]), W[29 - 7], sigma0(W[29 - 15]), W[29 - 16]);
W[30] = add4(sigma1(W[30 - 2]), W[30 - 7], sigma0(W[30 - 15]), W[30 - 16]);
W[31] = add4(sigma1(W[31 - 2]), W[31 - 7], sigma0(W[31 - 15]), W[31 - 16]);
W[32] = add4(sigma1(W[32 - 2]), W[32 - 7], sigma0(W[32 - 15]), W[32 - 16]);
W[33] = add4(sigma1(W[33 - 2]), W[33 - 7], sigma0(W[33 - 15]), W[33 - 16]);
W[34] = add4(sigma1(W[34 - 2]), W[34 - 7], sigma0(W[34 - 15]), W[34 - 16]);
W[35] = add4(sigma1(W[35 - 2]), W[35 - 7], sigma0(W[35 - 15]), W[35 - 16]);
W[36] = add4(sigma1(W[36 - 2]), W[36 - 7], sigma0(W[36 - 15]), W[36 - 16]);
W[37] = add4(sigma1(W[37 - 2]), W[37 - 7], sigma0(W[37 - 15]), W[37 - 16]);
W[38] = add4(sigma1(W[38 - 2]), W[38 - 7], sigma0(W[38 - 15]), W[38 - 16]);
W[39] = add4(sigma1(W[39 - 2]), W[39 - 7], sigma0(W[39 - 15]), W[39 - 16]);
W[40] = add4(sigma1(W[40 - 2]), W[40 - 7], sigma0(W[40 - 15]), W[40 - 16]);
W[41] = add4(sigma1(W[41 - 2]), W[41 - 7], sigma0(W[41 - 15]), W[41 - 16]);
W[42] = add4(sigma1(W[42 - 2]), W[42 - 7], sigma0(W[42 - 15]), W[42 - 16]);
W[43] = add4(sigma1(W[43 - 2]), W[43 - 7], sigma0(W[43 - 15]), W[43 - 16]);
W[44] = add4(sigma1(W[44 - 2]), W[44 - 7], sigma0(W[44 - 15]), W[44 - 16]);
W[45] = add4(sigma1(W[45 - 2]), W[45 - 7], sigma0(W[45 - 15]), W[45 - 16]);
W[46] = add4(sigma1(W[46 - 2]), W[46 - 7], sigma0(W[46 - 15]), W[46 - 16]);
W[47] = add4(sigma1(W[47 - 2]), W[47 - 7], sigma0(W[47 - 15]), W[47 - 16]);
W[48] = add4(sigma1(W[48 - 2]), W[48 - 7], sigma0(W[48 - 15]), W[48 - 16]);
W[49] = add4(sigma1(W[49 - 2]), W[49 - 7], sigma0(W[49 - 15]), W[49 - 16]);
W[50] = add4(sigma1(W[50 - 2]), W[50 - 7], sigma0(W[50 - 15]), W[50 - 16]);
W[51] = add4(sigma1(W[51 - 2]), W[51 - 7], sigma0(W[51 - 15]), W[51 - 16]);
W[52] = add4(sigma1(W[52 - 2]), W[52 - 7], sigma0(W[52 - 15]), W[52 - 16]);
W[53] = add4(sigma1(W[53 - 2]), W[53 - 7], sigma0(W[53 - 15]), W[53 - 16]);
W[54] = add4(sigma1(W[54 - 2]), W[54 - 7], sigma0(W[54 - 15]), W[54 - 16]);
W[55] = add4(sigma1(W[55 - 2]), W[55 - 7], sigma0(W[55 - 15]), W[55 - 16]);
W[56] = add4(sigma1(W[56 - 2]), W[56 - 7], sigma0(W[56 - 15]), W[56 - 16]);
W[57] = add4(sigma1(W[57 - 2]), W[57 - 7], sigma0(W[57 - 15]), W[57 - 16]);
W[58] = add4(sigma1(W[58 - 2]), W[58 - 7], sigma0(W[58 - 15]), W[58 - 16]);
W[59] = add4(sigma1(W[59 - 2]), W[59 - 7], sigma0(W[59 - 15]), W[59 - 16]);
W[60] = add4(sigma1(W[60 - 2]), W[60 - 7], sigma0(W[60 - 15]), W[60 - 16]);
W[61] = add4(sigma1(W[61 - 2]), W[61 - 7], sigma0(W[61 - 15]), W[61 - 16]);
W[62] = add4(sigma1(W[62 - 2]), W[62 - 7], sigma0(W[62 - 15]), W[62 - 16]);
W[63] = add4(sigma1(W[63 - 2]), W[63 - 7], sigma0(W[63 - 15]), W[63 - 16]);
// read existing state
__m128i a = state[0];
__m128i b = state[1];
__m128i c = state[2];
__m128i d = state[3];
__m128i e = state[4];
__m128i f = state[5];
__m128i g = state[6];
__m128i h = state[7];
t1 = add5(h, Sigma1(e), Ch(e, f, g), _mm_set1_epi32(0x428a2f98), W[0]);
t2 = add2(Sigma0(a), Maj(a, b, c));
d = add2(d, t1);
h = add2(t1, t2);
t1 = add5(g, Sigma1(d), Ch(d, e, f), _mm_set1_epi32(0x71374491), W[1]);
t2 = add2(Sigma0(h), Maj(h, a, b));
c = add2(c, t1);
g = add2(t1, t2);
t1 = add5(f, Sigma1(c), Ch(c, d, e), _mm_set1_epi32(0xb5c0fbcf), W[2]);
t2 = add2(Sigma0(g), Maj(g, h, a));
b = add2(b, t1);
f = add2(t1, t2);
t1 = add5(e, Sigma1(b), Ch(b, c, d), _mm_set1_epi32(0xe9b5dba5), W[3]);
t2 = add2(Sigma0(f), Maj(f, g, h));
a = add2(a, t1);
e = add2(t1, t2);
t1 = add5(d, Sigma1(a), Ch(a, b, c), _mm_set1_epi32(0x3956c25b), W[4]);
t2 = add2(Sigma0(e), Maj(e, f, g));
h = add2(h, t1);
d = add2(t1, t2);
t1 = add5(c, Sigma1(h), Ch(h, a, b), _mm_set1_epi32(0x59f111f1), W[5]);
t2 = add2(Sigma0(d), Maj(d, e, f));
g = add2(g, t1);
c = add2(t1, t2);
t1 = add5(b, Sigma1(g), Ch(g, h, a), _mm_set1_epi32(0x923f82a4), W[6]);
t2 = add2(Sigma0(c), Maj(c, d, e));
f = add2(f, t1);
b = add2(t1, t2);
t1 = add5(a, Sigma1(f), Ch(f, g, h), _mm_set1_epi32(0xab1c5ed5), W[7]);
t2 = add2(Sigma0(b), Maj(b, c, d));
e = add2(e, t1);
a = add2(t1, t2);
t1 = add5(h, Sigma1(e), Ch(e, f, g), _mm_set1_epi32(0xd807aa98), W[8]);
t2 = add2(Sigma0(a), Maj(a, b, c));
d = add2(d, t1);
h = add2(t1, t2);
t1 = add5(g, Sigma1(d), Ch(d, e, f), _mm_set1_epi32(0x12835b01), W[9]);
t2 = add2(Sigma0(h), Maj(h, a, b));
c = add2(c, t1);
g = add2(t1, t2);
t1 = add5(f, Sigma1(c), Ch(c, d, e), _mm_set1_epi32(0x243185be), W[10]);
t2 = add2(Sigma0(g), Maj(g, h, a));
b = add2(b, t1);
f = add2(t1, t2);
t1 = add5(e, Sigma1(b), Ch(b, c, d), _mm_set1_epi32(0x550c7dc3), W[11]);
t2 = add2(Sigma0(f), Maj(f, g, h));
a = add2(a, t1);
e = add2(t1, t2);
t1 = add5(d, Sigma1(a), Ch(a, b, c), _mm_set1_epi32(0x72be5d74), W[12]);
t2 = add2(Sigma0(e), Maj(e, f, g));
h = add2(h, t1);
d = add2(t1, t2);
t1 = add5(c, Sigma1(h), Ch(h, a, b), _mm_set1_epi32(0x80deb1fe), W[13]);
t2 = add2(Sigma0(d), Maj(d, e, f));
g = add2(g, t1);
c = add2(t1, t2);
t1 = add5(b, Sigma1(g), Ch(g, h, a), _mm_set1_epi32(0x9bdc06a7), W[14]);
t2 = add2(Sigma0(c), Maj(c, d, e));
f = add2(f, t1);
b = add2(t1, t2);
t1 = add5(a, Sigma1(f), Ch(f, g, h), _mm_set1_epi32(0xc19bf174), W[15]);
t2 = add2(Sigma0(b), Maj(b, c, d));
e = add2(e, t1);
a = add2(t1, t2);
t1 = add5(h, Sigma1(e), Ch(e, f, g), _mm_set1_epi32(0xe49b69c1), W[16]);
t2 = add2(Sigma0(a), Maj(a, b, c));
d = add2(d, t1);
h = add2(t1, t2);
t1 = add5(g, Sigma1(d), Ch(d, e, f), _mm_set1_epi32(0xefbe4786), W[17]);
t2 = add2(Sigma0(h), Maj(h, a, b));
c = add2(c, t1);
g = add2(t1, t2);
t1 = add5(f, Sigma1(c), Ch(c, d, e), _mm_set1_epi32(0x0fc19dc6), W[18]);
t2 = add2(Sigma0(g), Maj(g, h, a));
b = add2(b, t1);
f = add2(t1, t2);
t1 = add5(e, Sigma1(b), Ch(b, c, d), _mm_set1_epi32(0x240ca1cc), W[19]);
t2 = add2(Sigma0(f), Maj(f, g, h));
a = add2(a, t1);
e = add2(t1, t2);
t1 = add5(d, Sigma1(a), Ch(a, b, c), _mm_set1_epi32(0x2de92c6f), W[20]);
t2 = add2(Sigma0(e), Maj(e, f, g));
h = add2(h, t1);
d = add2(t1, t2);
t1 = add5(c, Sigma1(h), Ch(h, a, b), _mm_set1_epi32(0x4a7484aa), W[21]);
t2 = add2(Sigma0(d), Maj(d, e, f));
g = add2(g, t1);
c = add2(t1, t2);
t1 = add5(b, Sigma1(g), Ch(g, h, a), _mm_set1_epi32(0x5cb0a9dc), W[22]);
t2 = add2(Sigma0(c), Maj(c, d, e));
f = add2(f, t1);
b = add2(t1, t2);
t1 = add5(a, Sigma1(f), Ch(f, g, h), _mm_set1_epi32(0x76f988da), W[23]);
t2 = add2(Sigma0(b), Maj(b, c, d));
e = add2(e, t1);
a = add2(t1, t2);
t1 = add5(h, Sigma1(e), Ch(e, f, g), _mm_set1_epi32(0x983e5152), W[24]);
t2 = add2(Sigma0(a), Maj(a, b, c));
d = add2(d, t1);
h = add2(t1, t2);
t1 = add5(g, Sigma1(d), Ch(d, e, f), _mm_set1_epi32(0xa831c66d), W[25]);
t2 = add2(Sigma0(h), Maj(h, a, b));
c = add2(c, t1);
g = add2(t1, t2);
t1 = add5(f, Sigma1(c), Ch(c, d, e), _mm_set1_epi32(0xb00327c8), W[26]);
t2 = add2(Sigma0(g), Maj(g, h, a));
b = add2(b, t1);
f = add2(t1, t2);
t1 = add5(e, Sigma1(b), Ch(b, c, d), _mm_set1_epi32(0xbf597fc7), W[27]);
t2 = add2(Sigma0(f), Maj(f, g, h));
a = add2(a, t1);
e = add2(t1, t2);
t1 = add5(d, Sigma1(a), Ch(a, b, c), _mm_set1_epi32(0xc6e00bf3), W[28]);
t2 = add2(Sigma0(e), Maj(e, f, g));
h = add2(h, t1);
d = add2(t1, t2);
t1 = add5(c, Sigma1(h), Ch(h, a, b), _mm_set1_epi32(0xd5a79147), W[29]);
t2 = add2(Sigma0(d), Maj(d, e, f));
g = add2(g, t1);
c = add2(t1, t2);
t1 = add5(b, Sigma1(g), Ch(g, h, a), _mm_set1_epi32(0x06ca6351), W[30]);
t2 = add2(Sigma0(c), Maj(c, d, e));
f = add2(f, t1);
b = add2(t1, t2);
t1 = add5(a, Sigma1(f), Ch(f, g, h), _mm_set1_epi32(0x14292967), W[31]);
t2 = add2(Sigma0(b), Maj(b, c, d));
e = add2(e, t1);
a = add2(t1, t2);
t1 = add5(h, Sigma1(e), Ch(e, f, g), _mm_set1_epi32(0x27b70a85), W[32]);
t2 = add2(Sigma0(a), Maj(a, b, c));
d = add2(d, t1);
h = add2(t1, t2);
t1 = add5(g, Sigma1(d), Ch(d, e, f), _mm_set1_epi32(0x2e1b2138), W[33]);
t2 = add2(Sigma0(h), Maj(h, a, b));
c = add2(c, t1);
g = add2(t1, t2);
t1 = add5(f, Sigma1(c), Ch(c, d, e), _mm_set1_epi32(0x4d2c6dfc), W[34]);
t2 = add2(Sigma0(g), Maj(g, h, a));
b = add2(b, t1);
f = add2(t1, t2);
t1 = add5(e, Sigma1(b), Ch(b, c, d), _mm_set1_epi32(0x53380d13), W[35]);
t2 = add2(Sigma0(f), Maj(f, g, h));
a = add2(a, t1);
e = add2(t1, t2);
t1 = add5(d, Sigma1(a), Ch(a, b, c), _mm_set1_epi32(0x650a7354), W[36]);
t2 = add2(Sigma0(e), Maj(e, f, g));
h = add2(h, t1);
d = add2(t1, t2);
t1 = add5(c, Sigma1(h), Ch(h, a, b), _mm_set1_epi32(0x766a0abb), W[37]);
t2 = add2(Sigma0(d), Maj(d, e, f));
g = add2(g, t1);
c = add2(t1, t2);
t1 = add5(b, Sigma1(g), Ch(g, h, a), _mm_set1_epi32(0x81c2c92e), W[38]);
t2 = add2(Sigma0(c), Maj(c, d, e));
f = add2(f, t1);
b = add2(t1, t2);
t1 = add5(a, Sigma1(f), Ch(f, g, h), _mm_set1_epi32(0x92722c85), W[39]);
t2 = add2(Sigma0(b), Maj(b, c, d));
e = add2(e, t1);
a = add2(t1, t2);
t1 = add5(h, Sigma1(e), Ch(e, f, g), _mm_set1_epi32(0xa2bfe8a1), W[40]);
t2 = add2(Sigma0(a), Maj(a, b, c));
d = add2(d, t1);
h = add2(t1, t2);
t1 = add5(g, Sigma1(d), Ch(d, e, f), _mm_set1_epi32(0xa81a664b), W[41]);
t2 = add2(Sigma0(h), Maj(h, a, b));
c = add2(c, t1);
g = add2(t1, t2);
t1 = add5(f, Sigma1(c), Ch(c, d, e), _mm_set1_epi32(0xc24b8b70), W[42]);
t2 = add2(Sigma0(g), Maj(g, h, a));
b = add2(b, t1);
f = add2(t1, t2);
t1 = add5(e, Sigma1(b), Ch(b, c, d), _mm_set1_epi32(0xc76c51a3), W[43]);
t2 = add2(Sigma0(f), Maj(f, g, h));
a = add2(a, t1);
e = add2(t1, t2);
t1 = add5(d, Sigma1(a), Ch(a, b, c), _mm_set1_epi32(0xd192e819), W[44]);
t2 = add2(Sigma0(e), Maj(e, f, g));
h = add2(h, t1);
d = add2(t1, t2);
t1 = add5(c, Sigma1(h), Ch(h, a, b), _mm_set1_epi32(0xd6990624), W[45]);
t2 = add2(Sigma0(d), Maj(d, e, f));
g = add2(g, t1);
c = add2(t1, t2);
t1 = add5(b, Sigma1(g), Ch(g, h, a), _mm_set1_epi32(0xf40e3585), W[46]);
t2 = add2(Sigma0(c), Maj(c, d, e));
f = add2(f, t1);
b = add2(t1, t2);
t1 = add5(a, Sigma1(f), Ch(f, g, h), _mm_set1_epi32(0x106aa070), W[47]);
t2 = add2(Sigma0(b), Maj(b, c, d));
e = add2(e, t1);
a = add2(t1, t2);
t1 = add5(h, Sigma1(e), Ch(e, f, g), _mm_set1_epi32(0x19a4c116), W[48]);
t2 = add2(Sigma0(a), Maj(a, b, c));
d = add2(d, t1);
h = add2(t1, t2);
t1 = add5(g, Sigma1(d), Ch(d, e, f), _mm_set1_epi32(0x1e376c08), W[49]);
t2 = add2(Sigma0(h), Maj(h, a, b));
c = add2(c, t1);
g = add2(t1, t2);
t1 = add5(f, Sigma1(c), Ch(c, d, e), _mm_set1_epi32(0x2748774c), W[50]);
t2 = add2(Sigma0(g), Maj(g, h, a));
b = add2(b, t1);
f = add2(t1, t2);
t1 = add5(e, Sigma1(b), Ch(b, c, d), _mm_set1_epi32(0x34b0bcb5), W[51]);
t2 = add2(Sigma0(f), Maj(f, g, h));
a = add2(a, t1);
e = add2(t1, t2);
t1 = add5(d, Sigma1(a), Ch(a, b, c), _mm_set1_epi32(0x391c0cb3), W[52]);
t2 = add2(Sigma0(e), Maj(e, f, g));
h = add2(h, t1);
d = add2(t1, t2);
t1 = add5(c, Sigma1(h), Ch(h, a, b), _mm_set1_epi32(0x4ed8aa4a), W[53]);
t2 = add2(Sigma0(d), Maj(d, e, f));
g = add2(g, t1);
c = add2(t1, t2);
t1 = add5(b, Sigma1(g), Ch(g, h, a), _mm_set1_epi32(0x5b9cca4f), W[54]);
t2 = add2(Sigma0(c), Maj(c, d, e));
f = add2(f, t1);
b = add2(t1, t2);
t1 = add5(a, Sigma1(f), Ch(f, g, h), _mm_set1_epi32(0x682e6ff3), W[55]);
t2 = add2(Sigma0(b), Maj(b, c, d));
e = add2(e, t1);
a = add2(t1, t2);
t1 = add5(h, Sigma1(e), Ch(e, f, g), _mm_set1_epi32(0x748f82ee), W[56]);
t2 = add2(Sigma0(a), Maj(a, b, c));
d = add2(d, t1);
h = add2(t1, t2);
t1 = add5(g, Sigma1(d), Ch(d, e, f), _mm_set1_epi32(0x78a5636f), W[57]);
t2 = add2(Sigma0(h), Maj(h, a, b));
c = add2(c, t1);
g = add2(t1, t2);
t1 = add5(f, Sigma1(c), Ch(c, d, e), _mm_set1_epi32(0x84c87814), W[58]);
t2 = add2(Sigma0(g), Maj(g, h, a));
b = add2(b, t1);
f = add2(t1, t2);
t1 = add5(e, Sigma1(b), Ch(b, c, d), _mm_set1_epi32(0x8cc70208), W[59]);
t2 = add2(Sigma0(f), Maj(f, g, h));
a = add2(a, t1);
e = add2(t1, t2);
t1 = add5(d, Sigma1(a), Ch(a, b, c), _mm_set1_epi32(0x90befffa), W[60]);
t2 = add2(Sigma0(e), Maj(e, f, g));
h = add2(h, t1);
d = add2(t1, t2);
t1 = add5(c, Sigma1(h), Ch(h, a, b), _mm_set1_epi32(0xa4506ceb), W[61]);
t2 = add2(Sigma0(d), Maj(d, e, f));
g = add2(g, t1);
c = add2(t1, t2);
t1 = add5(b, Sigma1(g), Ch(g, h, a), _mm_set1_epi32(0xbef9a3f7), W[62]);
t2 = add2(Sigma0(c), Maj(c, d, e));
f = add2(f, t1);
b = add2(t1, t2);
t1 = add5(a, Sigma1(f), Ch(f, g, h), _mm_set1_epi32(0xc67178f2), W[63]);
t2 = add2(Sigma0(b), Maj(b, c, d));
e = add2(e, t1);
a = add2(t1, t2);
dst[0] = add2(state[0], a);
dst[1] = add2(state[1], b);
dst[2] = add2(state[2], c);
dst[3] = add2(state[3], d);
dst[4] = add2(state[4], e);
dst[5] = add2(state[5], f);
dst[6] = add2(state[6], g);
dst[7] = add2(state[7], h);
}
/* compress 1024-bits */
void ccsha512_ltc_compress(ccdigest_state_t state, unsigned long nblocks, const void *in)
{
uint64_t S[8], W[80], t0, t1;
int i;
uint64_t *s = ccdigest_u64(state);
const unsigned char *buf = in;
while(nblocks--) {
/* copy state into S */
for (i = 0; i < 8; i++) {
S[i] = s[i];
}
/* copy the state into 1024-bits into W[0..15] */
for (i = 0; i < 16; i++) {
CC_LOAD64_BE(W[i], buf + (8*i));
}
/* fill W[16..79] */
for (i = 16; i < 80; i++) {
W[i] = Gamma1(W[i - 2]) + W[i - 7] + Gamma0(W[i - 15]) + W[i - 16];
}
/* Compress */
#ifdef CC_SMALL_CODE
for (i = 0; i < 80; i++) {
t0 = S[7] + Sigma1(S[4]) + Ch(S[4], S[5], S[6]) + K[i] + W[i];
t1 = Sigma0(S[0]) + Maj(S[0], S[1], S[2]);
S[7] = S[6];
S[6] = S[5];
S[5] = S[4];
S[4] = S[3] + t0;
S[3] = S[2];
S[2] = S[1];
S[1] = S[0];
S[0] = t0 + t1;
}
#else
#define RND(a,b,c,d,e,f,g,h,i) \
t0 = h + Sigma1(e) + Ch(e, f, g) + K[i] + W[i]; \
t1 = Sigma0(a) + Maj(a, b, c); \
d += t0; \
h = t0 + t1;
for (i = 0; i < 80; i += 8) {
RND(S[0],S[1],S[2],S[3],S[4],S[5],S[6],S[7],i+0);
RND(S[7],S[0],S[1],S[2],S[3],S[4],S[5],S[6],i+1);
RND(S[6],S[7],S[0],S[1],S[2],S[3],S[4],S[5],i+2);
RND(S[5],S[6],S[7],S[0],S[1],S[2],S[3],S[4],i+3);
RND(S[4],S[5],S[6],S[7],S[0],S[1],S[2],S[3],i+4);
RND(S[3],S[4],S[5],S[6],S[7],S[0],S[1],S[2],i+5);
RND(S[2],S[3],S[4],S[5],S[6],S[7],S[0],S[1],i+6);
RND(S[1],S[2],S[3],S[4],S[5],S[6],S[7],S[0],i+7);
}
#endif
/* feedback */
for (i = 0; i < 8; i++) {
s[i] = s[i] + S[i];
}
buf+=CCSHA512_BLOCK_SIZE;
}
}
