adler32: make test program more useful and add fate test
[libav.git] / libavutil / aes.c
CommitLineData
3306dfd5 1/*
034f5fba 2 * copyright (c) 2007 Michael Niedermayer <michaelni@gmx.at>
3306dfd5 3 *
2eb697a4
DB
4 * some optimization ideas from aes128.c by Reimar Doeffinger
5 *
2912e87a 6 * This file is part of Libav.
3306dfd5 7 *
2912e87a 8 * Libav is free software; you can redistribute it and/or
3306dfd5
MN
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 *
2912e87a 13 * Libav is distributed in the hope that it will be useful,
3306dfd5
MN
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
2912e87a 19 * License along with Libav; if not, write to the Free Software
3306dfd5
MN
20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21 */
22
23#include "common.h"
3306dfd5
MN
24#include "aes.h"
25
dd025f25
RD
26typedef union {
27 uint64_t u64[2];
28 uint32_t u32[4];
29 uint8_t u8x4[4][4];
30 uint8_t u8[16];
31} av_aes_block;
32
3306dfd5 33typedef struct AVAES{
ef0d7a0d
RD
34 // Note: round_key[16] is accessed in the init code, but this only
35 // overwrites state, which does not matter (see also r7471).
dd025f25
RD
36 av_aes_block round_key[15];
37 av_aes_block state[2];
3306dfd5
MN
38 int rounds;
39}AVAES;
40
6572e1a4
MN
41const int av_aes_size= sizeof(AVAES);
42
a6e923c9
MN
43static const uint8_t rcon[10] = {
44 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x1b, 0x36
3306dfd5
MN
45};
46
3306dfd5
MN
47static uint8_t sbox[256];
48static uint8_t inv_sbox[256];
b250f9c6 49#if CONFIG_SMALL
cd0968c4
MN
50static uint32_t enc_multbl[1][256];
51static uint32_t dec_multbl[1][256];
52#else
53static uint32_t enc_multbl[4][256];
54static uint32_t dec_multbl[4][256];
55#endif
3306dfd5 56
dd025f25
RD
57static inline void addkey(av_aes_block *dst, const av_aes_block *src, const av_aes_block *round_key){
58 dst->u64[0] = src->u64[0] ^ round_key->u64[0];
59 dst->u64[1] = src->u64[1] ^ round_key->u64[1];
3306dfd5
MN
60}
61
dd025f25
RD
62static void subshift(av_aes_block s0[2], int s, const uint8_t *box){
63 av_aes_block *s1= (av_aes_block *)(s0[0].u8 - s);
64 av_aes_block *s3= (av_aes_block *)(s0[0].u8 + s);
65 s0[0].u8[0]=box[s0[1].u8[ 0]]; s0[0].u8[ 4]=box[s0[1].u8[ 4]]; s0[0].u8[ 8]=box[s0[1].u8[ 8]]; s0[0].u8[12]=box[s0[1].u8[12]];
66 s1[0].u8[3]=box[s1[1].u8[ 7]]; s1[0].u8[ 7]=box[s1[1].u8[11]]; s1[0].u8[11]=box[s1[1].u8[15]]; s1[0].u8[15]=box[s1[1].u8[ 3]];
67 s0[0].u8[2]=box[s0[1].u8[10]]; s0[0].u8[10]=box[s0[1].u8[ 2]]; s0[0].u8[ 6]=box[s0[1].u8[14]]; s0[0].u8[14]=box[s0[1].u8[ 6]];
68 s3[0].u8[1]=box[s3[1].u8[13]]; s3[0].u8[13]=box[s3[1].u8[ 9]]; s3[0].u8[ 9]=box[s3[1].u8[ 5]]; s3[0].u8[ 5]=box[s3[1].u8[ 1]];
1aa72a70 69}
0458e790 70
17c78b08 71static inline int mix_core(uint32_t multbl[][256], int a, int b, int c, int d){
b250f9c6 72#if CONFIG_SMALL
97f6d154
MN
73#define ROT(x,s) ((x<<s)|(x>>(32-s)))
74 return multbl[0][a] ^ ROT(multbl[0][b], 8) ^ ROT(multbl[0][c], 16) ^ ROT(multbl[0][d], 24);
cd0968c4 75#else
97f6d154 76 return multbl[0][a] ^ multbl[1][b] ^ multbl[2][c] ^ multbl[3][d];
cd0968c4 77#endif
3306dfd5 78}
298dd6b7 79
dd025f25 80static inline void mix(av_aes_block state[2], uint32_t multbl[][256], int s1, int s3){
19c039ee
RD
81 uint8_t (*src)[4] = state[1].u8x4;
82 state[0].u32[0] = mix_core(multbl, src[0][0], src[s1 ][1], src[2][2], src[s3 ][3]);
83 state[0].u32[1] = mix_core(multbl, src[1][0], src[s3-1][1], src[3][2], src[s1-1][3]);
84 state[0].u32[2] = mix_core(multbl, src[2][0], src[s3 ][1], src[0][2], src[s1 ][3]);
85 state[0].u32[3] = mix_core(multbl, src[3][0], src[s1-1][1], src[1][2], src[s3-1][3]);
1aa72a70 86}
3306dfd5 87
17c78b08 88static inline void crypt(AVAES *a, int s, const uint8_t *sbox, uint32_t multbl[][256]){
14bc3849 89 int r;
3306dfd5 90
7465619a 91 for(r=a->rounds-1; r>0; r--){
298dd6b7 92 mix(a->state, multbl, 3-s, 1+s);
dd025f25 93 addkey(&a->state[1], &a->state[0], &a->round_key[r]);
3306dfd5 94 }
dd025f25 95 subshift(&a->state[0], s, sbox);
3306dfd5
MN
96}
97
dd025f25
RD
98void av_aes_crypt(AVAES *a, uint8_t *dst_, const uint8_t *src_, int count, uint8_t *iv_, int decrypt){
99 av_aes_block *dst = (av_aes_block *)dst_;
100 const av_aes_block *src = (const av_aes_block *)src_;
101 av_aes_block *iv = (av_aes_block *)iv_;
97e3458c 102 while(count--){
dd025f25 103 addkey(&a->state[1], src, &a->round_key[a->rounds]);
97e3458c
MN
104 if(decrypt) {
105 crypt(a, 0, inv_sbox, dec_multbl);
106 if(iv){
dd025f25 107 addkey(&a->state[0], &a->state[0], iv);
97e3458c
MN
108 memcpy(iv, src, 16);
109 }
dd025f25 110 addkey(dst, &a->state[0], &a->round_key[0]);
97e3458c 111 }else{
dd025f25 112 if(iv) addkey(&a->state[1], &a->state[1], iv);
97e3458c 113 crypt(a, 2, sbox, enc_multbl);
dd025f25 114 addkey(dst, &a->state[0], &a->round_key[0]);
97e3458c
MN
115 if(iv) memcpy(iv, dst, 16);
116 }
dd025f25
RD
117 src++;
118 dst++;
97e3458c 119 }
ff8f6594
MN
120}
121
7b484a7a 122static void init_multbl2(uint8_t tbl[1024], const int c[4], const uint8_t *log8, const uint8_t *alog8, const uint8_t *sbox){
597e3232 123 int i, j;
0458e790 124 for(i=0; i<1024; i++){
96b10308 125 int x= sbox[i>>2];
0458e790
MN
126 if(x) tbl[i]= alog8[ log8[x] + log8[c[i&3]] ];
127 }
b250f9c6 128#if !CONFIG_SMALL
597e3232
MN
129 for(j=256; j<1024; j++)
130 for(i=0; i<4; i++)
131 tbl[4*j+i]= tbl[4*j + ((i-1)&3) - 1024];
7200f5d0 132#endif
0458e790
MN
133}
134
3306dfd5 135// this is based on the reference AES code by Paulo Barreto and Vincent Rijmen
57d7b036 136int av_aes_init(AVAES *a, const uint8_t *key, int key_bits, int decrypt) {
3306dfd5
MN
137 int i, j, t, rconpointer = 0;
138 uint8_t tk[8][4];
96b10308 139 int KC= key_bits>>5;
d43a2975 140 int rounds= KC + 6;
cd0968c4
MN
141 uint8_t log8[256];
142 uint8_t alog8[512];
3306dfd5 143
c184bb44 144 if(!enc_multbl[FF_ARRAY_ELEMS(enc_multbl)-1][FF_ARRAY_ELEMS(enc_multbl[0])-1]){
3306dfd5
MN
145 j=1;
146 for(i=0; i<255; i++){
147 alog8[i]=
148 alog8[i+255]= j;
149 log8[j]= i;
150 j^= j+j;
151 if(j>255) j^= 0x11B;
152 }
3306dfd5
MN
153 for(i=0; i<256; i++){
154 j= i ? alog8[255-log8[i]] : 0;
2c3427eb
MN
155 j ^= (j<<1) ^ (j<<2) ^ (j<<3) ^ (j<<4);
156 j = (j ^ (j>>8) ^ 99) & 255;
3306dfd5
MN
157 inv_sbox[j]= i;
158 sbox [i]= j;
3306dfd5 159 }
7b484a7a
RD
160 init_multbl2(dec_multbl[0], (const int[4]){0xe, 0x9, 0xd, 0xb}, log8, alog8, inv_sbox);
161 init_multbl2(enc_multbl[0], (const int[4]){0x2, 0x1, 0x1, 0x3}, log8, alog8, sbox);
3306dfd5
MN
162 }
163
d43a2975 164 if(key_bits!=128 && key_bits!=192 && key_bits!=256)
6572e1a4 165 return -1;
3306dfd5 166
d43a2975 167 a->rounds= rounds;
eade86e2 168
3306dfd5
MN
169 memcpy(tk, key, KC*4);
170
5d1b5393 171 for(t= 0; t < (rounds+1)*16;) {
dd025f25 172 memcpy(a->round_key[0].u8+t, tk, KC*4);
5d1b5393 173 t+= KC*4;
3306dfd5
MN
174
175 for(i = 0; i < 4; i++)
84c72fb1 176 tk[0][i] ^= sbox[tk[KC-1][(i+1)&3]];
3306dfd5
MN
177 tk[0][0] ^= rcon[rconpointer++];
178
179 for(j = 1; j < KC; j++){
96b10308 180 if(KC != 8 || j != KC>>1)
b1e5c294 181 for(i = 0; i < 4; i++) tk[j][i] ^= tk[j-1][i];
3306dfd5 182 else
b1e5c294 183 for(i = 0; i < 4; i++) tk[j][i] ^= sbox[tk[j-1][i]];
3306dfd5
MN
184 }
185 }
f12cf3d2 186
954bd264 187 if(decrypt){
28b51235 188 for(i=1; i<rounds; i++){
dd025f25
RD
189 av_aes_block tmp[3];
190 memcpy(&tmp[2], &a->round_key[i], 16);
2ad4a3bc 191 subshift(&tmp[1], 0, sbox);
298dd6b7 192 mix(tmp, dec_multbl, 1, 3);
dd025f25 193 memcpy(&a->round_key[i], &tmp[0], 16);
28b51235 194 }
f25b650a 195 }else{
96b10308 196 for(i=0; i<(rounds+1)>>1; i++){
f25b650a 197 for(j=0; j<16; j++)
dd025f25 198 FFSWAP(int, a->round_key[i].u8[j], a->round_key[rounds-i].u8[j]);
f25b650a 199 }
cc858970 200 }
f12cf3d2 201
6572e1a4 202 return 0;
3306dfd5
MN
203}
204
205#ifdef TEST
294eaa26 206#include "lfg.h"
0d80bd2f 207#include "log.h"
3306dfd5 208
f8a80fd6 209int main(void){
1b105243 210 int i,j;
6572e1a4 211 AVAES ae, ad, b;
d65dc0cd
MN
212 uint8_t rkey[2][16]= {
213 {0},
214 {0x10, 0xa5, 0x88, 0x69, 0xd7, 0x4b, 0xe5, 0xa3, 0x74, 0xcf, 0x86, 0x7c, 0xfb, 0x47, 0x38, 0x59}};
215 uint8_t pt[16], rpt[2][16]= {
216 {0x6a, 0x84, 0x86, 0x7c, 0xd7, 0x7e, 0x12, 0xad, 0x07, 0xea, 0x1b, 0xe8, 0x95, 0xc5, 0x3f, 0xa3},
217 {0}};
218 uint8_t rct[2][16]= {
219 {0x73, 0x22, 0x81, 0xc0, 0xa0, 0xaa, 0xb8, 0xf7, 0xa5, 0x4a, 0x0c, 0x67, 0xa0, 0xc4, 0x5e, 0xcf},
220 {0x6d, 0x25, 0x1e, 0x69, 0x44, 0xb0, 0x51, 0xe0, 0x4e, 0xaa, 0x6f, 0xb4, 0xdb, 0xf7, 0x84, 0x65}};
7465619a 221 uint8_t temp[16];
64bde197 222 AVLFG prng;
3306dfd5 223
6572e1a4
MN
224 av_aes_init(&ae, "PI=3.141592654..", 128, 0);
225 av_aes_init(&ad, "PI=3.141592654..", 128, 1);
b56e34c5 226 av_log_set_level(AV_LOG_DEBUG);
64bde197 227 av_lfg_init(&prng, 1);
cf5773e9 228
d65dc0cd 229 for(i=0; i<2; i++){
6572e1a4 230 av_aes_init(&b, rkey[i], 128, 1);
ca2560df 231 av_aes_crypt(&b, temp, rct[i], 1, NULL, 1);
d65dc0cd 232 for(j=0; j<16; j++)
7465619a
MN
233 if(rpt[i][j] != temp[j])
234 av_log(NULL, AV_LOG_ERROR, "%d %02X %02X\n", j, rpt[i][j], temp[j]);
d65dc0cd 235 }
0458e790 236
3306dfd5 237 for(i=0; i<10000; i++){
ff8f6594 238 for(j=0; j<16; j++){
64bde197 239 pt[j] = av_lfg_get(&prng);
ff8f6594 240 }
cf5773e9 241{START_TIMER
ca2560df 242 av_aes_crypt(&ae, temp, pt, 1, NULL, 0);
ff8f6594 243 if(!(i&(i-1)))
7465619a 244 av_log(NULL, AV_LOG_ERROR, "%02X %02X %02X %02X\n", temp[0], temp[5], temp[10], temp[15]);
ca2560df 245 av_aes_crypt(&ad, temp, temp, 1, NULL, 1);
cf5773e9 246STOP_TIMER("aes")}
ff8f6594 247 for(j=0; j<16; j++){
7465619a
MN
248 if(pt[j] != temp[j]){
249 av_log(NULL, AV_LOG_ERROR, "%d %d %02X %02X\n", i,j, pt[j], temp[j]);
ff8f6594
MN
250 }
251 }
3306dfd5
MN
252 }
253 return 0;
254}
255#endif