_ Git - BearSSL/blob - src/symcipher/aes_x86ni_cbcenc.c

   1 /*
   2  * Copyright (c) 2017 Thomas Pornin <pornin@bolet.org>
   3  *
   4  * Permission is hereby granted, free of charge, to any person obtaining
   5  * a copy of this software and associated documentation files (the
   6  * "Software"), to deal in the Software without restriction, including
   7  * without limitation the rights to use, copy, modify, merge, publish,
   8  * distribute, sublicense, and/or sell copies of the Software, and to
   9  * permit persons to whom the Software is furnished to do so, subject to
  10  * the following conditions:
  11  *
  12  * The above copyright notice and this permission notice shall be
  13  * included in all copies or substantial portions of the Software.
  14  *
  15  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
  16  * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
  17  * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
  18  * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
  19  * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
  20  * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  21  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
  22  * SOFTWARE.
  23  */
  24
  25 #include "inner.h"
  26
  27 #if BR_AES_X86NI
  28
  29 #if BR_AES_X86NI_GCC
  30 #if BR_AES_X86NI_GCC_OLD
  31 #pragma GCC target("sse2,sse4.1,aes,pclmul")
  32 #endif
  33 #include <wmmintrin.h>
  34 #endif
  35
  36 #if BR_AES_X86NI_MSC
  37 #include <intrin.h>
  38 #endif
  39
  40 /* see bearssl_block.h */
  41 void
  42 br_aes_x86ni_cbcenc_init(br_aes_x86ni_cbcenc_keys *ctx,
  43         const void *key, size_t len)
  44 {
  45         ctx->vtable = &br_aes_x86ni_cbcenc_vtable;
  46         ctx->num_rounds = br_aes_x86ni_keysched_enc(ctx->skey.skni, key, len);
  47 }
  48
  49 /* see bearssl_block.h */
  50 BR_TARGET("sse2,aes")
  51 void
  52 br_aes_x86ni_cbcenc_run(const br_aes_x86ni_cbcenc_keys *ctx,
  53         void *iv, void *data, size_t len)
  54 {
  55         unsigned char *buf;
  56         unsigned num_rounds;
  57         __m128i sk[15], ivx;
  58         unsigned u;
  59
  60         buf = data;
  61         ivx = _mm_loadu_si128(iv);
  62         num_rounds = ctx->num_rounds;
  63         for (u = 0; u <= num_rounds; u ++) {
  64                 sk[u] = _mm_loadu_si128((void *)(ctx->skey.skni + (u << 4)));
  65         }
  66         while (len > 0) {
  67                 __m128i x;
  68
  69                 x = _mm_xor_si128(_mm_loadu_si128((void *)buf), ivx);
  70                 x = _mm_xor_si128(x, sk[0]);
  71                 x = _mm_aesenc_si128(x, sk[1]);
  72                 x = _mm_aesenc_si128(x, sk[2]);
  73                 x = _mm_aesenc_si128(x, sk[3]);
  74                 x = _mm_aesenc_si128(x, sk[4]);
  75                 x = _mm_aesenc_si128(x, sk[5]);
  76                 x = _mm_aesenc_si128(x, sk[6]);
  77                 x = _mm_aesenc_si128(x, sk[7]);
  78                 x = _mm_aesenc_si128(x, sk[8]);
  79                 x = _mm_aesenc_si128(x, sk[9]);
  80                 if (num_rounds == 10) {
  81                         x = _mm_aesenclast_si128(x, sk[10]);
  82                 } else if (num_rounds == 12) {
  83                         x = _mm_aesenc_si128(x, sk[10]);
  84                         x = _mm_aesenc_si128(x, sk[11]);
  85                         x = _mm_aesenclast_si128(x, sk[12]);
  86                 } else {
  87                         x = _mm_aesenc_si128(x, sk[10]);
  88                         x = _mm_aesenc_si128(x, sk[11]);
  89                         x = _mm_aesenc_si128(x, sk[12]);
  90                         x = _mm_aesenc_si128(x, sk[13]);
  91                         x = _mm_aesenclast_si128(x, sk[14]);
  92                 }
  93                 ivx = x;
  94                 _mm_storeu_si128((void *)buf, x);
  95                 buf += 16;
  96                 len -= 16;
  97         }
  98         _mm_storeu_si128(iv, ivx);
  99 }
 100
 101 /* see bearssl_block.h */
 102 const br_block_cbcenc_class br_aes_x86ni_cbcenc_vtable = {
 103         sizeof(br_aes_x86ni_cbcenc_keys),
 104         16,
 105         4,
 106         (void (*)(const br_block_cbcenc_class **, const void *, size_t))
 107                 &br_aes_x86ni_cbcenc_init,
 108         (void (*)(const br_block_cbcenc_class *const *, void *, void *, size_t))
 109                 &br_aes_x86ni_cbcenc_run
 110 };
 111
 112 /* see bearssl_block.h */
 113 const br_block_cbcenc_class *
 114 br_aes_x86ni_cbcenc_get_vtable(void)
 115 {
 116         return br_aes_x86ni_supported() ? &br_aes_x86ni_cbcenc_vtable : NULL;
 117 }
 118
 119 #else
 120
 121 /* see bearssl_block.h */
 122 const br_block_cbcenc_class *
 123 br_aes_x86ni_cbcenc_get_vtable(void)
 124 {
 125         return NULL;
 126 }
 127
 128 #endif