DXR is a code search and navigation tool aimed at making sense of large projects. It supports full-text and regex searches as well as structural queries.

Mercurial (c68fe15a81fc)

VCS Links

Line Code
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218
/* This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * file, You can obtain one at http://mozilla.org/MPL/2.0/. */

#ifdef FREEBL_NO_DEPEND
#include "stubs.h"
#endif
#endif
#include "rijndael.h"
#include "secerr.h"

#include <wmmintrin.h> /* aes-ni */


#define EXPAND_KEY128(k, rcon, res)                   \
    tmp_key = _mm_aeskeygenassist_si128(k, rcon);     \
    tmp_key = _mm_shuffle_epi32(tmp_key, 0xFF);       \
    tmp = _mm_xor_si128(k, _mm_slli_si128(k, 4));     \
    tmp = _mm_xor_si128(tmp, _mm_slli_si128(tmp, 4)); \
    tmp = _mm_xor_si128(tmp, _mm_slli_si128(tmp, 4)); \
    res = _mm_xor_si128(tmp, tmp_key)
    res = _mm_xor_si128(tmp, tmp_key)

static void
native_key_expansion128(AESContext *cx, const unsigned char *key)
{
    __m128i *keySchedule = cx->k.keySchedule;
    __m128i *keySchedule = cx->k.keySchedule;
    pre_align __m128i tmp_key post_align;
    pre_align __m128i tmp post_align;
    keySchedule[0] = _mm_loadu_si128((__m128i *)key);
    EXPAND_KEY128(keySchedule[0], 0x01, keySchedule[1]);
    EXPAND_KEY128(keySchedule[1], 0x02, keySchedule[2]);
    EXPAND_KEY128(keySchedule[1], 0x02, keySchedule[2]);
    EXPAND_KEY128(keySchedule[2], 0x04, keySchedule[3]);
    EXPAND_KEY128(keySchedule[3], 0x08, keySchedule[4]);
    EXPAND_KEY128(keySchedule[4], 0x10, keySchedule[5]);
    EXPAND_KEY128(keySchedule[5], 0x20, keySchedule[6]);
    EXPAND_KEY128(keySchedule[6], 0x40, keySchedule[7]);
    EXPAND_KEY128(keySchedule[7], 0x80, keySchedule[8]);
    EXPAND_KEY128(keySchedule[8], 0x1B, keySchedule[9]);
    EXPAND_KEY128(keySchedule[9], 0x36, keySchedule[10]);
    EXPAND_KEY128(keySchedule[9], 0x36, keySchedule[10]);
}

#define EXPAND_KEY192_PART1(res, k0, kt, rcon)                                \
    tmp2 = _mm_slli_si128(k0, 4);                                             \
    tmp1 = _mm_xor_si128(k0, tmp2);                                           \
    tmp1 = _mm_xor_si128(k0, tmp2);                                           \
    tmp2 = _mm_slli_si128(tmp2, 4);                                           \
    tmp1 = _mm_xor_si128(_mm_xor_si128(tmp1, tmp2), _mm_slli_si128(tmp2, 4)); \
    tmp2 = _mm_aeskeygenassist_si128(kt, rcon);                               \
    res = _mm_xor_si128(tmp1, _mm_shuffle_epi32(tmp2, 0x55))


#define EXPAND_KEY192_PART2(res, k1, k2)             \
    tmp2 = _mm_xor_si128(k1, _mm_slli_si128(k1, 4)); \
    res = _mm_xor_si128(tmp2, _mm_shuffle_epi32(k2, 0xFF))

#define EXPAND_KEY192(k0, res1, res2, res3, carry, rcon1, rcon2)         \
#define EXPAND_KEY192(k0, res1, res2, res3, carry, rcon1, rcon2)         \
    EXPAND_KEY192_PART1(tmp3, k0, res1, rcon1);                          \
    EXPAND_KEY192_PART2(carry, res1, tmp3);                              \
    res1 = _mm_castpd_si128(_mm_shuffle_pd(_mm_castsi128_pd(res1),       \
                                           _mm_castsi128_pd(tmp3), 0));  \
    res2 = _mm_castpd_si128(_mm_shuffle_pd(_mm_castsi128_pd(tmp3),       \
    res2 = _mm_castpd_si128(_mm_shuffle_pd(_mm_castsi128_pd(tmp3),       \
                                           _mm_castsi128_pd(carry), 1)); \
    EXPAND_KEY192_PART1(res3, tmp3, carry, rcon2)

static void
native_key_expansion192(AESContext *cx, const unsigned char *key)
{
    __m128i *keySchedule = cx->k.keySchedule;
    __m128i *keySchedule = cx->k.keySchedule;
    pre_align __m128i tmp1 post_align;
    pre_align __m128i tmp2 post_align;
    pre_align __m128i tmp3 post_align;
    pre_align __m128i carry post_align;
    keySchedule[0] = _mm_loadu_si128((__m128i *)key);
    keySchedule[1] = _mm_loadu_si128((__m128i *)(key + 16));
    EXPAND_KEY192(keySchedule[0], keySchedule[1], keySchedule[2],
                  keySchedule[3], carry, 0x1, 0x2);
    EXPAND_KEY192_PART2(keySchedule[4], carry, keySchedule[3]);
    EXPAND_KEY192_PART2(keySchedule[4], carry, keySchedule[3]);
    EXPAND_KEY192(keySchedule[3], keySchedule[4], keySchedule[5],
                  keySchedule[6], carry, 0x4, 0x8);
    EXPAND_KEY192_PART2(keySchedule[7], carry, keySchedule[6]);
    EXPAND_KEY192(keySchedule[6], keySchedule[7], keySchedule[8],
                  keySchedule[9], carry, 0x10, 0x20);
                  keySchedule[9], carry, 0x10, 0x20);
    EXPAND_KEY192_PART2(keySchedule[10], carry, keySchedule[9]);
    EXPAND_KEY192(keySchedule[9], keySchedule[10], keySchedule[11],
                  keySchedule[12], carry, 0x40, 0x80);
}


#define EXPAND_KEY256_PART(res, rconx, k1x, k2x, X)                           \
    tmp_key = _mm_shuffle_epi32(_mm_aeskeygenassist_si128(k2x, rconx), X);    \
    tmp2 = _mm_slli_si128(k1x, 4);                                            \
    tmp1 = _mm_xor_si128(k1x, tmp2);                                          \
    tmp2 = _mm_slli_si128(tmp2, 4);                                           \
    tmp2 = _mm_slli_si128(tmp2, 4);                                           \
    tmp1 = _mm_xor_si128(_mm_xor_si128(tmp1, tmp2), _mm_slli_si128(tmp2, 4)); \
    res = _mm_xor_si128(tmp1, tmp_key);

#define EXPAND_KEY256(res1, res2, k1, k2, rcon)   \
    EXPAND_KEY256_PART(res1, rcon, k1, k2, 0xFF); \
    EXPAND_KEY256_PART(res1, rcon, k1, k2, 0xFF); \
    EXPAND_KEY256_PART(res2, 0x00, k2, res1, 0xAA)

static void
native_key_expansion256(AESContext *cx, const unsigned char *key)
{
{
    __m128i *keySchedule = cx->k.keySchedule;
    pre_align __m128i tmp_key post_align;
    pre_align __m128i tmp1 post_align;
    pre_align __m128i tmp2 post_align;
    keySchedule[0] = _mm_loadu_si128((__m128i *)key);
    keySchedule[1] = _mm_loadu_si128((__m128i *)(key + 16));
    EXPAND_KEY256(keySchedule[2], keySchedule[3], keySchedule[0],
    EXPAND_KEY256(keySchedule[2], keySchedule[3], keySchedule[0],
                  keySchedule[1], 0x01);
    EXPAND_KEY256(keySchedule[4], keySchedule[5], keySchedule[2],
                  keySchedule[3], 0x02);
    EXPAND_KEY256(keySchedule[6], keySchedule[7], keySchedule[4],
    EXPAND_KEY256(keySchedule[6], keySchedule[7], keySchedule[4],
                  keySchedule[5], 0x04);
    EXPAND_KEY256(keySchedule[8], keySchedule[9], keySchedule[6],
                  keySchedule[7], 0x08);
    EXPAND_KEY256(keySchedule[10], keySchedule[11], keySchedule[8],
                  keySchedule[9], 0x10);
                  keySchedule[9], 0x10);
    EXPAND_KEY256(keySchedule[12], keySchedule[13], keySchedule[10],
                  keySchedule[11], 0x20);
    EXPAND_KEY256_PART(keySchedule[14], 0x40, keySchedule[12],
                       keySchedule[13], 0xFF);
                       keySchedule[13], 0xFF);
}

/*
 * AES key expansion using aes-ni instructions.
 */
 */
void
rijndael_native_key_expansion(AESContext *cx, const unsigned char *key,
                              unsigned int Nk)
{
    switch (Nk) {
    switch (Nk) {
        case 4:
            native_key_expansion128(cx, key);
            return;
        case 6:
            native_key_expansion192(cx, key);
            native_key_expansion192(cx, key);
            return;
        case 8:
            native_key_expansion256(cx, key);
            return;
        default:
        default:
            /* This shouldn't happen (checked by the caller). */
            return;
    }
}


void
rijndael_native_encryptBlock(AESContext *cx,
                             unsigned char *output,
                             const unsigned char *input)
{
    unsigned int i;
    pre_align __m128i m post_align = _mm_loadu_si128((__m128i *)input);
    pre_align __m128i m post_align = _mm_loadu_si128((__m128i *)input);
    m = _mm_xor_si128(m, cx->k.keySchedule[0]);
    for (i = 1; i < cx->Nr; ++i) {
        m = _mm_aesenc_si128(m, cx->k.keySchedule[i]);
    }
    m = _mm_aesenclast_si128(m, cx->k.keySchedule[cx->Nr]);
    m = _mm_aesenclast_si128(m, cx->k.keySchedule[cx->Nr]);
    _mm_storeu_si128((__m128i *)output, m);
}

void
rijndael_native_decryptBlock(AESContext *cx,
rijndael_native_decryptBlock(AESContext *cx,
                             unsigned char *output,
                             const unsigned char *input)
{
    int i;
    pre_align __m128i m post_align = _mm_loadu_si128((__m128i *)input);
    pre_align __m128i m post_align = _mm_loadu_si128((__m128i *)input);
    m = _mm_xor_si128(m, cx->k.keySchedule[cx->Nr]);
    for (i = cx->Nr - 1; i > 0; --i) {
        m = _mm_aesdec_si128(m, cx->k.keySchedule[i]);
    }
    m = _mm_aesdeclast_si128(m, cx->k.keySchedule[0]);
    m = _mm_aesdeclast_si128(m, cx->k.keySchedule[0]);
    _mm_storeu_si128((__m128i *)output, m);
}

// out = a ^ b
void
void
native_xorBlock(unsigned char *out,
                const unsigned char *a,
                const unsigned char *b)
{
    pre_align __m128i post_align in1 = _mm_loadu_si128((__m128i *)(a));
    pre_align __m128i post_align in1 = _mm_loadu_si128((__m128i *)(a));
    pre_align __m128i post_align in2 = _mm_loadu_si128((__m128i *)(b));
    in1 = _mm_xor_si128(in1, in2);
    _mm_storeu_si128((__m128i *)(out), in1);
}