/* computes x/R == x (mod N) via Montgomery Reduction */
int32 pstm_montgomery_reduce(psPool_t *pool, pstm_int *a, pstm_int *m,
pstm_digit mp, pstm_digit *paD, uint32 paDlen)
{
pstm_digit *c, *_c, *tmpm, mu;
int32 oldused, x, y;
int16 pa;
pa = m->used;
if (paD && paDlen >= (uint32)2*pa+1) {
c = paD;
memset(c, 0x0, paDlen);
} else {
c = psMalloc(pool, 2*pa+1);
if (c == NULL) {
return PS_MEM_FAIL;
} else {
memset(c, 0x0, 2*pa+1);
}
}
/* copy the input */
oldused = a->used;
for (x = 0; x < oldused; x++) {
c[x] = a->dp[x];
}
MONT_START;
for (x = 0; x < pa; x++) {
pstm_digit cy = 0;
/* get Mu for this round */
LOOP_START;
_c = c + x;
tmpm = m->dp;
y = 0;
for (; y < pa; y++) {
INNERMUL;
++_c;
}
LOOP_END;
while (cy) {
PROPCARRY;
++_c;
}
}
/* now copy out */
_c = c + pa;
tmpm = a->dp;
for (x = 0; x < pa+1; x++) {
*tmpm++ = *_c++;
}
for (; x < oldused; x++) {
*tmpm++ = 0;
}
MONT_FINI;
a->used = pa+1;
pstm_clamp(a);
/* reuse x as return code */
x = PSTM_OKAY;
/* if A >= m then A = A - m */
if (pstm_cmp_mag (a, m) != PSTM_LT) {
if (s_pstm_sub (a, m, a) != PSTM_OKAY) {
x = PS_MEM_FAIL;
}
}
if (paDlen < (uint32)2*pa+1) {
psFree(c);
}
return x;
}
/**
computes x/R == x (mod N) via Montgomery Reduction.
*/
int32_t pstm_montgomery_reduce(psPool_t *pool, pstm_int *a, const pstm_int *m,
pstm_digit mp, pstm_digit *paD, uint16_t paDlen)
{
pstm_digit *c, *_c, *tmpm, mu;
int32 oldused, x, y;
int16 pa;
uint32 cSize;
pa = m->used;
if (pa > a->alloc) {
/* Sanity test for bad numbers. This will confirm no buffer overruns */
return PS_LIMIT_FAIL;
}
cSize = (2 * pa + 1) * sizeof(pstm_digit);
if (paD && paDlen >= cSize) {
c = paD;
memset(c, 0x0, paDlen);
} else {
c = psMalloc(pool, cSize);
if (c == NULL) {
return PS_MEM_FAIL;
} else {
memset(c, 0x0, cSize);
}
}
/* copy the input */
oldused = a->used;
for (x = 0; x < oldused; x++) {
c[x] = a->dp[x];
}
MONT_START;
for (x = 0; x < pa; x++) {
pstm_digit cy = 0;
/* get Mu for this round */
LOOP_START;
_c = c + x;
tmpm = m->dp;
y = 0;
#ifdef PSTM_X86_64
for (; y < (pa & ~7); y += 8) {
INNERMUL8;
_c += 8;
tmpm += 8;
}
#endif /* PSTM_X86_64 */
for (; y < pa; y++) {
INNERMUL;
++_c;
}
LOOP_END;
while (cy) {
PROPCARRY;
++_c;
}
}
/* now copy out */
_c = c + pa;
tmpm = a->dp;
for (x = 0; x < pa+1; x++) {
*tmpm++ = *_c++;
}
for (; x < oldused; x++) {
*tmpm++ = 0;
}
MONT_FINI;
a->used = pa+1;
pstm_clamp(a);
/* reuse x as return code */
x = PSTM_OKAY;
/* if A >= m then A = A - m */
if (pstm_cmp_mag (a, m) != PSTM_LT) {
if (s_pstm_sub (a, m, a) != PSTM_OKAY) {
x = PS_MEM_FAIL;
}
}
if (paDlen < cSize) {
psFree(c, pool);
}
return x;
}