pqc/crypto_kem/mceliece460896/sse/vec128.c
Thom Wiggers b3f9d4f8d6
Classic McEliece (#259)
* Add McEliece reference implementations

* Add Vec implementations of McEliece

* Add sse implementations

* Add AVX2 implementations

* Get rid of stuff not supported by Mac ABI

* restrict to two cores

* Ditch .data files

* Remove .hidden from all .S files

* speed up duplicate consistency tests by batching

* make cpuinfo more robust

* Hope to stabilize macos cpuinfo without ccache

* Revert "Hope to stabilize macos cpuinfo without ccache"

This reverts commit 6129c3cabe1abbc8b956bc87e902a698e32bf322.

* Just hardcode what's available at travis

* Fixed-size types in api.h

* namespace all header files in mceliece

* Ditch operations.h

* Get rid of static inline functions

* fixup! Ditch operations.h
2020-02-05 13:09:56 +01:00

153 lines
4.9 KiB
C

/*
This file is for functions related to 128-bit vectors
including functions for bitsliced field operations
*/
#include "vec128.h"
#include "params.h"
vec128 PQCLEAN_MCELIECE460896_SSE_vec128_set1_16b(uint16_t a) {
return _mm_set1_epi16(a);
}
int PQCLEAN_MCELIECE460896_SSE_vec128_testz(vec128 a) {
return _mm_testz_si128(a, a);
}
vec128 PQCLEAN_MCELIECE460896_SSE_vec128_setzero(void) {
return _mm_setzero_si128();
}
vec128 PQCLEAN_MCELIECE460896_SSE_vec128_and(vec128 a, vec128 b) {
return _mm_and_si128(a, b);
}
vec128 PQCLEAN_MCELIECE460896_SSE_vec128_xor(vec128 a, vec128 b) {
return _mm_xor_si128(a, b);
}
vec128 PQCLEAN_MCELIECE460896_SSE_vec128_or(vec128 a, vec128 b) {
return _mm_or_si128(a, b);
}
vec128 PQCLEAN_MCELIECE460896_SSE_vec128_sll_2x(vec128 a, int s) {
return _mm_slli_epi64(a, s);
}
vec128 PQCLEAN_MCELIECE460896_SSE_vec128_srl_2x(vec128 a, int s) {
return _mm_srli_epi64(a, s);
}
vec128 PQCLEAN_MCELIECE460896_SSE_vec128_set2x(uint64_t a0, uint64_t a1) {
return _mm_set_epi64x(a1, a0);
}
vec128 PQCLEAN_MCELIECE460896_SSE_vec128_unpack_low(vec128 a, vec128 b) {
return _mm_unpacklo_epi64(a, b);
}
vec128 PQCLEAN_MCELIECE460896_SSE_vec128_unpack_high(vec128 a, vec128 b) {
return _mm_unpackhi_epi64(a, b);
}
vec128 PQCLEAN_MCELIECE460896_SSE_vec128_setbits(uint64_t a) {
return _mm_set1_epi64x(-a);
}
void PQCLEAN_MCELIECE460896_SSE_vec128_copy(vec128 *dest, const vec128 *src) {
int i;
for (i = 0; i < GFBITS; i++) {
dest[i] = src[i];
}
}
void PQCLEAN_MCELIECE460896_SSE_vec128_add(vec128 *c, const vec128 *a, const vec128 *b) {
int i;
for (i = 0; i < GFBITS; i++) {
c[i] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(a[i], b[i]);
}
}
vec128 PQCLEAN_MCELIECE460896_SSE_vec128_or_reduce(const vec128 *a) {
int i;
vec128 ret;
ret = a[0];
for (i = 1; i < GFBITS; i++) {
ret = PQCLEAN_MCELIECE460896_SSE_vec128_or(ret, a[i]);
}
return ret;
}
/* bitsliced field multiplications */
void PQCLEAN_MCELIECE460896_SSE_vec128_mul(vec128 *h, vec128 *f, const vec128 *g) {
PQCLEAN_MCELIECE460896_SSE_vec128_mul_asm(h, f, g, 16);
}
/* bitsliced field squarings */
void PQCLEAN_MCELIECE460896_SSE_vec128_sq(vec128 *out, const vec128 *in) {
int i;
vec128 result[GFBITS], t;
t = PQCLEAN_MCELIECE460896_SSE_vec128_xor(in[11], in[12]);
result[0] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(in[0], in[11]);
result[1] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(in[7], t);
result[2] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(in[1], in[7]);
result[3] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(in[8], t);
result[4] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(in[2], in[7]);
result[4] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(result[4], in[8]);
result[4] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(result[4], t);
result[5] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(in[7], in[9]);
result[6] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(in[3], in[8]);
result[6] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(result[6], in[9]);
result[6] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(result[6], in[12]);
result[7] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(in[8], in[10]);
result[8] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(in[4], in[9]);
result[8] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(result[8], in[10]);
result[9] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(in[9], in[11]);
result[10] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(in[5], in[10]);
result[10] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(result[10], in[11]);
result[11] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(in[10], in[12]);
result[12] = PQCLEAN_MCELIECE460896_SSE_vec128_xor(in[6], t);
for (i = 0; i < GFBITS; i++) {
out[i] = result[i];
}
}
/* bitsliced field inverses */
void PQCLEAN_MCELIECE460896_SSE_vec128_inv(vec128 *out, const vec128 *in) {
vec128 tmp_11[ GFBITS ];
vec128 tmp_1111[ GFBITS ];
PQCLEAN_MCELIECE460896_SSE_vec128_copy(out, in);
PQCLEAN_MCELIECE460896_SSE_vec128_sq(out, out);
PQCLEAN_MCELIECE460896_SSE_vec128_mul(tmp_11, out, in); // ^11
PQCLEAN_MCELIECE460896_SSE_vec128_sq(out, tmp_11);
PQCLEAN_MCELIECE460896_SSE_vec128_sq(out, out);
PQCLEAN_MCELIECE460896_SSE_vec128_mul(tmp_1111, out, tmp_11); // ^1111
PQCLEAN_MCELIECE460896_SSE_vec128_sq(out, tmp_1111);
PQCLEAN_MCELIECE460896_SSE_vec128_sq(out, out);
PQCLEAN_MCELIECE460896_SSE_vec128_sq(out, out);
PQCLEAN_MCELIECE460896_SSE_vec128_sq(out, out);
PQCLEAN_MCELIECE460896_SSE_vec128_mul(out, out, tmp_1111); // ^11111111
PQCLEAN_MCELIECE460896_SSE_vec128_sq(out, out);
PQCLEAN_MCELIECE460896_SSE_vec128_sq(out, out);
PQCLEAN_MCELIECE460896_SSE_vec128_sq(out, out);
PQCLEAN_MCELIECE460896_SSE_vec128_sq(out, out);
PQCLEAN_MCELIECE460896_SSE_vec128_mul(out, out, tmp_1111); // ^111111111111
PQCLEAN_MCELIECE460896_SSE_vec128_sq(out, out); // ^1111111111110
}