pqc/crypto_kem/mceliece6960119f/avx/vec256.c

/*
  This file is for functions related to 256-bit vectors
  including functions for bitsliced field operations
*/

#include "vec256.h"

extern void PQCLEAN_MCELIECE6960119F_AVX_vec256_mul_asm(vec256 *, vec256 *, const vec256 *);

vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_set1_16b(uint16_t a) {
    return _mm256_set1_epi16(a);
}

vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_setzero(void) {
    return  _mm256_setzero_si256();
}

vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_set4x(uint64_t a0, uint64_t a1, uint64_t a2, uint64_t a3) {
    return  _mm256_set_epi64x(a3, a2, a1, a0);
}

int PQCLEAN_MCELIECE6960119F_AVX_vec256_testz(vec256 a) {
    return _mm256_testz_si256(a, a);
}

vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_and(vec256 a, vec256 b) {
    return _mm256_and_si256(a, b);
}

vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(vec256 a, vec256 b) {
    return _mm256_xor_si256(a, b);
}

vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_or(vec256 a, vec256 b) {
    return _mm256_or_si256(a, b);
}

vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_sll_4x(vec256 a, int s) {
    return _mm256_slli_epi64(a, s);
}

vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_srl_4x(vec256 a, int s) {
    return _mm256_srli_epi64(a, s);
}

vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_unpack_low(vec256 a, vec256 b) {
    return _mm256_permute2x128_si256 (a, b, 0x20);
}

vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_unpack_high(vec256 a, vec256 b) {
    return _mm256_permute2x128_si256 (a, b, 0x31);
}

vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_unpack_low_2x(vec256 a, vec256 b) {
    return _mm256_unpacklo_epi64 (a, b);
}

vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_unpack_high_2x(vec256 a, vec256 b) {
    return _mm256_unpackhi_epi64 (a, b);
}

vec256 PQCLEAN_MCELIECE6960119F_AVX_vec256_or_reduce(const vec256 *a) {
    int i;
    vec256 ret;

    ret = a[0];
    for (i = 1; i < GFBITS; i++) {
        ret = PQCLEAN_MCELIECE6960119F_AVX_vec256_or(ret, a[i]);
    }

    return ret;
}

void PQCLEAN_MCELIECE6960119F_AVX_vec256_copy(vec256 *dest, const vec256 *src) {
    int i;

    for (i = 0; i < GFBITS; i++) {
        dest[i] = src[i];
    }
}


void PQCLEAN_MCELIECE6960119F_AVX_vec256_mul(vec256 *h, vec256 *f, const vec256 *g) {
    PQCLEAN_MCELIECE6960119F_AVX_vec256_mul_asm(h, f, g);
}

/* bitsliced field squarings */
void PQCLEAN_MCELIECE6960119F_AVX_vec256_sq(vec256 *out, const vec256 *in) {
    int i;
    vec256 result[GFBITS], t;

    t = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(in[11], in[12]);

    result[0] = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(in[0], in[11]);
    result[1] = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(in[7], t);
    result[2] = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(in[1], in[7]);
    result[3] = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(in[8], t);
    result[4] = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(in[2], in[7]);
    result[4] = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(result[4], in[8]);
    result[4] = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(result[4], t);
    result[5] = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(in[7], in[9]);
    result[6] = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(in[3], in[8]);
    result[6] = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(result[6], in[9]);
    result[6] = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(result[6], in[12]);
    result[7] = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(in[8], in[10]);
    result[8] = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(in[4], in[9]);
    result[8] = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(result[8], in[10]);
    result[9] = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(in[9], in[11]);
    result[10] = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(in[5], in[10]);
    result[10] = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(result[10], in[11]);
    result[11] = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(in[10], in[12]);
    result[12] = PQCLEAN_MCELIECE6960119F_AVX_vec256_xor(in[6], t);

    for (i = 0; i < GFBITS; i++) {
        out[i] = result[i];
    }
}

/* bitsliced field inverses */
void PQCLEAN_MCELIECE6960119F_AVX_vec256_inv(vec256 *out, const vec256 *in) {
    vec256 tmp_11[ GFBITS ];
    vec256 tmp_1111[ GFBITS ];

    PQCLEAN_MCELIECE6960119F_AVX_vec256_copy(out, in);

    PQCLEAN_MCELIECE6960119F_AVX_vec256_sq(out, out);
    PQCLEAN_MCELIECE6960119F_AVX_vec256_mul(tmp_11, out, in); // ^11

    PQCLEAN_MCELIECE6960119F_AVX_vec256_sq(out, tmp_11);
    PQCLEAN_MCELIECE6960119F_AVX_vec256_sq(out, out);
    PQCLEAN_MCELIECE6960119F_AVX_vec256_mul(tmp_1111, out, tmp_11); // ^1111

    PQCLEAN_MCELIECE6960119F_AVX_vec256_sq(out, tmp_1111);
    PQCLEAN_MCELIECE6960119F_AVX_vec256_sq(out, out);
    PQCLEAN_MCELIECE6960119F_AVX_vec256_sq(out, out);
    PQCLEAN_MCELIECE6960119F_AVX_vec256_sq(out, out);
    PQCLEAN_MCELIECE6960119F_AVX_vec256_mul(out, out, tmp_1111); // ^11111111

    PQCLEAN_MCELIECE6960119F_AVX_vec256_sq(out, out);
    PQCLEAN_MCELIECE6960119F_AVX_vec256_sq(out, out);
    PQCLEAN_MCELIECE6960119F_AVX_vec256_sq(out, out);
    PQCLEAN_MCELIECE6960119F_AVX_vec256_sq(out, out);
    PQCLEAN_MCELIECE6960119F_AVX_vec256_mul(out, out, tmp_1111); // ^111111111111

    PQCLEAN_MCELIECE6960119F_AVX_vec256_sq(out, out); // ^1111111111110
}