void F512(uint32_t *h, uint32_t *m) {
int i;
uint32_t roundConstantAddingValue;
uint32_t roundConstant;
uint32_t x[2*COLS512];
for (i = 0; i < 2*COLS512; i++) {
x[i] = h[i]^m[i];
}
roundConstantAddingValue = 0x01010101;
roundConstant = 0;
for(i=0;i<10;i++)
{
RND512Q(m, roundConstant);
roundConstant += roundConstantAddingValue;
}
roundConstant = 0;
for(i=0;i<10;i++)
{
RND512P(x, roundConstant);
roundConstant += roundConstantAddingValue;
}
for (i = 0; i < 2*COLS512; i++) {
h[i] ^= x[i]^m[i];
}
}
/* compute compression function (short variants) */
static void F512(uint32_t *h, const uint32_t *m) {
int i;
uint32_t Ptmp[2*COLS512];
uint32_t Qtmp[2*COLS512];
uint32_t y[2*COLS512];
uint32_t z[2*COLS512];
for (i = 0; i < 2*COLS512; i++) {
z[i] = m[i];
Ptmp[i] = h[i]^m[i];
}
/* compute Q(m) */
RND512Q((uint8_t*)z, y, 0x00000000);
RND512Q((uint8_t*)y, z, 0x01000000);
RND512Q((uint8_t*)z, y, 0x02000000);
RND512Q((uint8_t*)y, z, 0x03000000);
RND512Q((uint8_t*)z, y, 0x04000000);
RND512Q((uint8_t*)y, z, 0x05000000);
RND512Q((uint8_t*)z, y, 0x06000000);
RND512Q((uint8_t*)y, z, 0x07000000);
RND512Q((uint8_t*)z, y, 0x08000000);
RND512Q((uint8_t*)y, Qtmp, 0x09000000);
/* compute P(h+m) */
RND512P((uint8_t*)Ptmp, y, 0x00000000);
RND512P((uint8_t*)y, z, 0x00000001);
RND512P((uint8_t*)z, y, 0x00000002);
RND512P((uint8_t*)y, z, 0x00000003);
RND512P((uint8_t*)z, y, 0x00000004);
RND512P((uint8_t*)y, z, 0x00000005);
RND512P((uint8_t*)z, y, 0x00000006);
RND512P((uint8_t*)y, z, 0x00000007);
RND512P((uint8_t*)z, y, 0x00000008);
RND512P((uint8_t*)y, Ptmp, 0x00000009);
/* compute P(h+m) + Q(m) + h */
for (i = 0; i < 2*COLS512; i++) {
h[i] ^= Ptmp[i]^Qtmp[i];
}
}
void Q(uint32_t *h, uint32_t *m)
{
int i;
uint32_t roundConstantAddingValue;
uint32_t roundConstant;
roundConstantAddingValue = 0x01010101;
roundConstant = 0;
for(i=0;i<10;i++)
{
RND512Q(m, roundConstant);
roundConstant += roundConstantAddingValue;
}
for (i = 0; i < 2*COLS512; i++) {
h[i] ^= m[i];
}
}
/* the compression function (short variants) */
inline void F512(u64 *h, const u64 *m, u64 *c) { //modified
u64 y[COLS512] __attribute__((aligned(16)));
u64 z[COLS512] __attribute__((aligned(16)));
u64 outQ[COLS512] __attribute__((aligned(16)));
u64 inP[COLS512] __attribute__((aligned(16)));
int i,j=0;
//ADDED By Gurpreet
u64 ml[COLS512] __attribute__((aligned(16))); //msg_left
u64 mr[COLS512] __attribute__((aligned(16))); //msg_right
j=0;
/*divide msg into two 512 blocks*/
for (i = 0; i < COLS1024; i++) { //modified
if(i<COLS512){
ml[i] = m[i];
}
else {
mr[j] = m[i];
j++;
}
}
//---------------
/* compute c+ml
h+mr */
for (i = 0; i < COLS512; i++) {
inP[i] = c[i] ^ ml[i];
z[i] = h[i]^mr[i];
}
/* compute Q(h+mr) */
RND512Q(z,y,U64BIG(0x0000000000000000ull));
RND512Q(y,z,U64BIG(0x0000000000000001ull));
RND512Q(z,y,U64BIG(0x0000000000000002ull));
RND512Q(y,z,U64BIG(0x0000000000000003ull));
RND512Q(z,y,U64BIG(0x0000000000000004ull));
RND512Q(y,z,U64BIG(0x0000000000000005ull));
RND512Q(z,y,U64BIG(0x0000000000000006ull));
RND512Q(y,z,U64BIG(0x0000000000000007ull));
RND512Q(z,y,U64BIG(0x0000000000000008ull));
RND512Q(y,outQ,U64BIG(0x0000000000000009ull));
/* compute P(c+ml) */
RND512P(inP,y,U64BIG(0x0000000000000000ull));
RND512P(y,z, U64BIG(0x0100000000000000ull));
RND512P(z,y, U64BIG(0x0200000000000000ull));
RND512P(y,z, U64BIG(0x0300000000000000ull));
RND512P(z,y, U64BIG(0x0400000000000000ull));
RND512P(y,z, U64BIG(0x0500000000000000ull));
RND512P(z,y, U64BIG(0x0600000000000000ull));
RND512P(y,z, U64BIG(0x0700000000000000ull));
RND512P(z,y, U64BIG(0x0800000000000000ull));
RND512P(y,inP, U64BIG(0x0900000000000000ull));
//ADDED by GURPREET
/* compute P(c+ml) + Q(h+mr) */
for (i = 0; i < COLS512; i++) {
inP[i] = inP[i]^outQ[i];
}
/* compute P(P(h+ml) + Q(h+mr)) */
RND512P(inP,z,U64BIG(0x0000000000000000ull));
RND512P(z,y, U64BIG(0x0100000000000000ull));
RND512P(y,z, U64BIG(0x0200000000000000ull));
RND512P(z,y, U64BIG(0x0300000000000000ull));
RND512P(y,z, U64BIG(0x0400000000000000ull));
RND512P(z,y, U64BIG(0x0500000000000000ull));
RND512P(y,z, U64BIG(0x0600000000000000ull));
RND512P(z,y, U64BIG(0x0700000000000000ull));
RND512P(y,z, U64BIG(0x0800000000000000ull));
RND512P(z,inP, U64BIG(0x0900000000000000ull));
/* compute P(P(h+ml) + Q(h+mr)) + Q(h+mr) + h */
#pragma vector aligned
for (i = 0; i < COLS512; i++) {
h[i] ^= inP[i] ^ outQ[i];
}
}
