aes.c 8.0 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243
  1. /*
  2. * copyright (c) 2007 Michael Niedermayer <michaelni@gmx.at>
  3. *
  4. * some optimization ideas from aes128.c by Reimar Doeffinger
  5. *
  6. * This file is part of FFmpeg.
  7. *
  8. * FFmpeg is free software; you can redistribute it and/or
  9. * modify it under the terms of the GNU Lesser General Public
  10. * License as published by the Free Software Foundation; either
  11. * version 2.1 of the License, or (at your option) any later version.
  12. *
  13. * FFmpeg is distributed in the hope that it will be useful,
  14. * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15. * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
  16. * Lesser General Public License for more details.
  17. *
  18. * You should have received a copy of the GNU Lesser General Public
  19. * License along with FFmpeg; if not, write to the Free Software
  20. * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  21. */
  22. #include "common.h"
  23. #include "aes.h"
  24. typedef struct AVAES{
  25. // Note: round_key[16] is accessed in the init code, but this only
  26. // overwrites state, which does not matter (see also r7471).
  27. uint8_t round_key[15][4][4];
  28. uint8_t state[2][4][4];
  29. int rounds;
  30. }AVAES;
  31. const int av_aes_size= sizeof(AVAES);
  32. static const uint8_t rcon[10] = {
  33. 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x1b, 0x36
  34. };
  35. static uint8_t sbox[256];
  36. static uint8_t inv_sbox[256];
  37. #if CONFIG_SMALL
  38. static uint32_t enc_multbl[1][256];
  39. static uint32_t dec_multbl[1][256];
  40. #else
  41. static uint32_t enc_multbl[4][256];
  42. static uint32_t dec_multbl[4][256];
  43. #endif
  44. static inline void addkey(uint64_t dst[2], const uint64_t src[2], const uint64_t round_key[2]){
  45. dst[0] = src[0] ^ round_key[0];
  46. dst[1] = src[1] ^ round_key[1];
  47. }
  48. static void subshift(uint8_t s0[2][16], int s, const uint8_t *box){
  49. uint8_t (*s1)[16]= s0[0] - s;
  50. uint8_t (*s3)[16]= s0[0] + s;
  51. s0[0][0]=box[s0[1][ 0]]; s0[0][ 4]=box[s0[1][ 4]]; s0[0][ 8]=box[s0[1][ 8]]; s0[0][12]=box[s0[1][12]];
  52. s1[0][3]=box[s1[1][ 7]]; s1[0][ 7]=box[s1[1][11]]; s1[0][11]=box[s1[1][15]]; s1[0][15]=box[s1[1][ 3]];
  53. s0[0][2]=box[s0[1][10]]; s0[0][10]=box[s0[1][ 2]]; s0[0][ 6]=box[s0[1][14]]; s0[0][14]=box[s0[1][ 6]];
  54. s3[0][1]=box[s3[1][13]]; s3[0][13]=box[s3[1][ 9]]; s3[0][ 9]=box[s3[1][ 5]]; s3[0][ 5]=box[s3[1][ 1]];
  55. }
  56. static inline int mix_core(uint32_t multbl[4][256], int a, int b, int c, int d){
  57. #if CONFIG_SMALL
  58. #define ROT(x,s) ((x<<s)|(x>>(32-s)))
  59. return multbl[0][a] ^ ROT(multbl[0][b], 8) ^ ROT(multbl[0][c], 16) ^ ROT(multbl[0][d], 24);
  60. #else
  61. return multbl[0][a] ^ multbl[1][b] ^ multbl[2][c] ^ multbl[3][d];
  62. #endif
  63. }
  64. static inline void mix(uint8_t state[2][4][4], uint32_t multbl[4][256], int s1, int s3){
  65. ((uint32_t *)(state))[0] = mix_core(multbl, state[1][0][0], state[1][s1 ][1], state[1][2][2], state[1][s3 ][3]);
  66. ((uint32_t *)(state))[1] = mix_core(multbl, state[1][1][0], state[1][s3-1][1], state[1][3][2], state[1][s1-1][3]);
  67. ((uint32_t *)(state))[2] = mix_core(multbl, state[1][2][0], state[1][s3 ][1], state[1][0][2], state[1][s1 ][3]);
  68. ((uint32_t *)(state))[3] = mix_core(multbl, state[1][3][0], state[1][s1-1][1], state[1][1][2], state[1][s3-1][3]);
  69. }
  70. static inline void crypt(AVAES *a, int s, const uint8_t *sbox, const uint32_t *multbl){
  71. int r;
  72. for(r=a->rounds-1; r>0; r--){
  73. mix(a->state, multbl, 3-s, 1+s);
  74. addkey(a->state[1], a->state[0], a->round_key[r]);
  75. }
  76. subshift(a->state[0][0], s, sbox);
  77. }
  78. void av_aes_crypt(AVAES *a, uint8_t *dst, const uint8_t *src, int count, uint8_t *iv, int decrypt){
  79. while(count--){
  80. addkey(a->state[1], src, a->round_key[a->rounds]);
  81. if(decrypt) {
  82. crypt(a, 0, inv_sbox, dec_multbl);
  83. if(iv){
  84. addkey(a->state[0], a->state[0], iv);
  85. memcpy(iv, src, 16);
  86. }
  87. addkey(dst, a->state[0], a->round_key[0]);
  88. }else{
  89. if(iv) addkey(a->state[1], a->state[1], iv);
  90. crypt(a, 2, sbox, enc_multbl);
  91. addkey(dst, a->state[0], a->round_key[0]);
  92. if(iv) memcpy(iv, dst, 16);
  93. }
  94. src+=16;
  95. dst+=16;
  96. }
  97. }
  98. static void init_multbl2(uint8_t tbl[1024], const int c[4], const uint8_t *log8, const uint8_t *alog8, const uint8_t *sbox){
  99. int i, j;
  100. for(i=0; i<1024; i++){
  101. int x= sbox[i>>2];
  102. if(x) tbl[i]= alog8[ log8[x] + log8[c[i&3]] ];
  103. }
  104. #if !CONFIG_SMALL
  105. for(j=256; j<1024; j++)
  106. for(i=0; i<4; i++)
  107. tbl[4*j+i]= tbl[4*j + ((i-1)&3) - 1024];
  108. #endif
  109. }
  110. // this is based on the reference AES code by Paulo Barreto and Vincent Rijmen
  111. int av_aes_init(AVAES *a, const uint8_t *key, int key_bits, int decrypt) {
  112. int i, j, t, rconpointer = 0;
  113. uint8_t tk[8][4];
  114. int KC= key_bits>>5;
  115. int rounds= KC + 6;
  116. uint8_t log8[256];
  117. uint8_t alog8[512];
  118. if(!enc_multbl[0][sizeof(enc_multbl)/sizeof(enc_multbl[0][0])-1]){
  119. j=1;
  120. for(i=0; i<255; i++){
  121. alog8[i]=
  122. alog8[i+255]= j;
  123. log8[j]= i;
  124. j^= j+j;
  125. if(j>255) j^= 0x11B;
  126. }
  127. for(i=0; i<256; i++){
  128. j= i ? alog8[255-log8[i]] : 0;
  129. j ^= (j<<1) ^ (j<<2) ^ (j<<3) ^ (j<<4);
  130. j = (j ^ (j>>8) ^ 99) & 255;
  131. inv_sbox[j]= i;
  132. sbox [i]= j;
  133. }
  134. init_multbl2(dec_multbl[0], (const int[4]){0xe, 0x9, 0xd, 0xb}, log8, alog8, inv_sbox);
  135. init_multbl2(enc_multbl[0], (const int[4]){0x2, 0x1, 0x1, 0x3}, log8, alog8, sbox);
  136. }
  137. if(key_bits!=128 && key_bits!=192 && key_bits!=256)
  138. return -1;
  139. a->rounds= rounds;
  140. memcpy(tk, key, KC*4);
  141. for(t= 0; t < (rounds+1)*16;) {
  142. memcpy(a->round_key[0][0]+t, tk, KC*4);
  143. t+= KC*4;
  144. for(i = 0; i < 4; i++)
  145. tk[0][i] ^= sbox[tk[KC-1][(i+1)&3]];
  146. tk[0][0] ^= rcon[rconpointer++];
  147. for(j = 1; j < KC; j++){
  148. if(KC != 8 || j != KC>>1)
  149. for(i = 0; i < 4; i++) tk[j][i] ^= tk[j-1][i];
  150. else
  151. for(i = 0; i < 4; i++) tk[j][i] ^= sbox[tk[j-1][i]];
  152. }
  153. }
  154. if(decrypt){
  155. for(i=1; i<rounds; i++){
  156. uint8_t tmp[3][16];
  157. memcpy(tmp[2], a->round_key[i][0], 16);
  158. subshift(tmp[1], 0, sbox);
  159. mix(tmp, dec_multbl, 1, 3);
  160. memcpy(a->round_key[i][0], tmp[0], 16);
  161. }
  162. }else{
  163. for(i=0; i<(rounds+1)>>1; i++){
  164. for(j=0; j<16; j++)
  165. FFSWAP(int, a->round_key[i][0][j], a->round_key[rounds-i][0][j]);
  166. }
  167. }
  168. return 0;
  169. }
  170. #ifdef TEST
  171. #include "log.h"
  172. #undef random
  173. int main(void){
  174. int i,j;
  175. AVAES ae, ad, b;
  176. uint8_t rkey[2][16]= {
  177. {0},
  178. {0x10, 0xa5, 0x88, 0x69, 0xd7, 0x4b, 0xe5, 0xa3, 0x74, 0xcf, 0x86, 0x7c, 0xfb, 0x47, 0x38, 0x59}};
  179. uint8_t pt[16], rpt[2][16]= {
  180. {0x6a, 0x84, 0x86, 0x7c, 0xd7, 0x7e, 0x12, 0xad, 0x07, 0xea, 0x1b, 0xe8, 0x95, 0xc5, 0x3f, 0xa3},
  181. {0}};
  182. uint8_t rct[2][16]= {
  183. {0x73, 0x22, 0x81, 0xc0, 0xa0, 0xaa, 0xb8, 0xf7, 0xa5, 0x4a, 0x0c, 0x67, 0xa0, 0xc4, 0x5e, 0xcf},
  184. {0x6d, 0x25, 0x1e, 0x69, 0x44, 0xb0, 0x51, 0xe0, 0x4e, 0xaa, 0x6f, 0xb4, 0xdb, 0xf7, 0x84, 0x65}};
  185. uint8_t temp[16];
  186. av_aes_init(&ae, "PI=3.141592654..", 128, 0);
  187. av_aes_init(&ad, "PI=3.141592654..", 128, 1);
  188. av_log_level= AV_LOG_DEBUG;
  189. for(i=0; i<2; i++){
  190. av_aes_init(&b, rkey[i], 128, 1);
  191. av_aes_crypt(&b, temp, rct[i], 1, NULL, 1);
  192. for(j=0; j<16; j++)
  193. if(rpt[i][j] != temp[j])
  194. av_log(NULL, AV_LOG_ERROR, "%d %02X %02X\n", j, rpt[i][j], temp[j]);
  195. }
  196. for(i=0; i<10000; i++){
  197. for(j=0; j<16; j++){
  198. pt[j]= random();
  199. }
  200. {START_TIMER
  201. av_aes_crypt(&ae, temp, pt, 1, NULL, 0);
  202. if(!(i&(i-1)))
  203. av_log(NULL, AV_LOG_ERROR, "%02X %02X %02X %02X\n", temp[0], temp[5], temp[10], temp[15]);
  204. av_aes_crypt(&ad, temp, temp, 1, NULL, 1);
  205. STOP_TIMER("aes")}
  206. for(j=0; j<16; j++){
  207. if(pt[j] != temp[j]){
  208. av_log(NULL, AV_LOG_ERROR, "%d %d %02X %02X\n", i,j, pt[j], temp[j]);
  209. }
  210. }
  211. }
  212. return 0;
  213. }
  214. #endif