Line | Branch | Exec | Source |
---|---|---|---|
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 | |||
23 | #include <string.h> | ||
24 | |||
25 | #include "config.h" | ||
26 | #include "aes.h" | ||
27 | #include "aes_internal.h" | ||
28 | #include "error.h" | ||
29 | #include "intreadwrite.h" | ||
30 | #include "macros.h" | ||
31 | #include "mem.h" | ||
32 | |||
33 | const int av_aes_size= sizeof(AVAES); | ||
34 | |||
35 | 16 | struct AVAES *av_aes_alloc(void) | |
36 | { | ||
37 | 16 | return av_mallocz(sizeof(struct AVAES)); | |
38 | } | ||
39 | |||
40 | static const uint8_t rcon[10] = { | ||
41 | 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x1b, 0x36 | ||
42 | }; | ||
43 | |||
44 | static uint8_t sbox[256]; | ||
45 | static uint8_t inv_sbox[256]; | ||
46 | #if CONFIG_SMALL | ||
47 | static uint32_t enc_multbl[1][256]; | ||
48 | static uint32_t dec_multbl[1][256]; | ||
49 | #else | ||
50 | static uint32_t enc_multbl[4][256]; | ||
51 | static uint32_t dec_multbl[4][256]; | ||
52 | #endif | ||
53 | |||
54 | #if HAVE_BIGENDIAN | ||
55 | # define ROT(x, s) (((x) >> (s)) | ((x) << (32-(s)))) | ||
56 | #else | ||
57 | # define ROT(x, s) (((x) << (s)) | ((x) >> (32-(s)))) | ||
58 | #endif | ||
59 | |||
60 | 47934 | static inline void addkey(av_aes_block *dst, const av_aes_block *src, | |
61 | const av_aes_block *round_key) | ||
62 | { | ||
63 | 47934 | dst->u64[0] = src->u64[0] ^ round_key->u64[0]; | |
64 | 47934 | dst->u64[1] = src->u64[1] ^ round_key->u64[1]; | |
65 | 47934 | } | |
66 | |||
67 | 5326 | static inline void addkey_s(av_aes_block *dst, const uint8_t *src, | |
68 | const av_aes_block *round_key) | ||
69 | { | ||
70 | 5326 | dst->u64[0] = AV_RN64(src) ^ round_key->u64[0]; | |
71 | 5326 | dst->u64[1] = AV_RN64(src + 8) ^ round_key->u64[1]; | |
72 | 5326 | } | |
73 | |||
74 | 5326 | static inline void addkey_d(uint8_t *dst, const av_aes_block *src, | |
75 | const av_aes_block *round_key) | ||
76 | { | ||
77 | 5326 | AV_WN64(dst, src->u64[0] ^ round_key->u64[0]); | |
78 | 5326 | AV_WN64(dst + 8, src->u64[1] ^ round_key->u64[1]); | |
79 | 5326 | } | |
80 | |||
81 | 5344 | static void subshift(av_aes_block s0[2], int s, const uint8_t *box) | |
82 | { | ||
83 | 5344 | unsigned char *s1_dst = (unsigned char*)s0[0].u8 + 3 - s; | |
84 | 5344 | const unsigned char *s1_src = s1_dst + sizeof(*s0); | |
85 | 5344 | unsigned char *s3_dst = (unsigned char*)s0[0].u8 + s + 1; | |
86 | 5344 | const unsigned char *s3_src = s3_dst + sizeof(*s0); | |
87 | |||
88 | 5344 | s0[0].u8[ 0] = box[s0[1].u8[ 0]]; | |
89 | 5344 | s0[0].u8[ 4] = box[s0[1].u8[ 4]]; | |
90 | 5344 | s0[0].u8[ 8] = box[s0[1].u8[ 8]]; | |
91 | 5344 | s0[0].u8[12] = box[s0[1].u8[12]]; | |
92 | 5344 | s1_dst[ 0] = box[s1_src[ 4]]; | |
93 | 5344 | s1_dst[ 4] = box[s1_src[ 8]]; | |
94 | 5344 | s1_dst[ 8] = box[s1_src[12]]; | |
95 | 5344 | s1_dst[12] = box[s1_src[ 0]]; | |
96 | 5344 | s0[0].u8[ 2] = box[s0[1].u8[10]]; | |
97 | 5344 | s0[0].u8[10] = box[s0[1].u8[ 2]]; | |
98 | 5344 | s0[0].u8[ 6] = box[s0[1].u8[14]]; | |
99 | 5344 | s0[0].u8[14] = box[s0[1].u8[ 6]]; | |
100 | 5344 | s3_dst[ 0] = box[s3_src[12]]; | |
101 | 5344 | s3_dst[12] = box[s3_src[ 8]]; | |
102 | 5344 | s3_dst[ 8] = box[s3_src[ 4]]; | |
103 | 5344 | s3_dst[ 4] = box[s3_src[ 0]]; | |
104 | 5344 | } | |
105 | |||
106 | 191808 | static inline int mix_core(uint32_t multbl[][256], int a, int b, int c, int d) | |
107 | { | ||
108 | #if CONFIG_SMALL | ||
109 | return multbl[0][a] ^ ROT(multbl[0][b], 8) ^ ROT(multbl[0][c], 16) ^ ROT(multbl[0][d], 24); | ||
110 | #else | ||
111 | 191808 | return multbl[0][a] ^ multbl[1][b] ^ multbl[2][c] ^ multbl[3][d]; | |
112 | #endif | ||
113 | } | ||
114 | |||
115 | 47952 | static inline void mix(av_aes_block state[2], uint32_t multbl[][256], int s1, int s3) | |
116 | { | ||
117 | 47952 | uint8_t (*src)[4] = state[1].u8x4; | |
118 | 47952 | state[0].u32[0] = mix_core(multbl, src[0][0], src[s1 ][1], src[2][2], src[s3 ][3]); | |
119 | 47952 | state[0].u32[1] = mix_core(multbl, src[1][0], src[s3 - 1][1], src[3][2], src[s1 - 1][3]); | |
120 | 47952 | state[0].u32[2] = mix_core(multbl, src[2][0], src[s3 ][1], src[0][2], src[s1 ][3]); | |
121 | 47952 | state[0].u32[3] = mix_core(multbl, src[3][0], src[s1 - 1][1], src[1][2], src[s3 - 1][3]); | |
122 | 47952 | } | |
123 | |||
124 | 5326 | static inline void aes_crypt(AVAES *a, int s, const uint8_t *sbox, | |
125 | uint32_t multbl[][256]) | ||
126 | { | ||
127 | int r; | ||
128 | |||
129 |
2/2✓ Branch 0 taken 47934 times.
✓ Branch 1 taken 5326 times.
|
53260 | for (r = a->rounds - 1; r > 0; r--) { |
130 | 47934 | mix(a->state, multbl, 3 - s, 1 + s); | |
131 | 47934 | addkey(&a->state[1], &a->state[0], &a->round_key[r]); | |
132 | } | ||
133 | |||
134 | 5326 | subshift(&a->state[0], s, sbox); | |
135 | 5326 | } | |
136 | |||
137 | 5324 | static void aes_encrypt(AVAES *a, uint8_t *dst, const uint8_t *src, | |
138 | int count, uint8_t *iv, int rounds) | ||
139 | { | ||
140 |
2/2✓ Branch 0 taken 5324 times.
✓ Branch 1 taken 5324 times.
|
10648 | while (count--) { |
141 | 5324 | addkey_s(&a->state[1], src, &a->round_key[rounds]); | |
142 |
1/2✗ Branch 0 not taken.
✓ Branch 1 taken 5324 times.
|
5324 | if (iv) |
143 | ✗ | addkey_s(&a->state[1], iv, &a->state[1]); | |
144 | 5324 | aes_crypt(a, 2, sbox, enc_multbl); | |
145 | 5324 | addkey_d(dst, &a->state[0], &a->round_key[0]); | |
146 |
1/2✗ Branch 0 not taken.
✓ Branch 1 taken 5324 times.
|
5324 | if (iv) |
147 | ✗ | memcpy(iv, dst, 16); | |
148 | 5324 | src += 16; | |
149 | 5324 | dst += 16; | |
150 | } | ||
151 | 5324 | } | |
152 | |||
153 | 2 | static void aes_decrypt(AVAES *a, uint8_t *dst, const uint8_t *src, | |
154 | int count, uint8_t *iv, int rounds) | ||
155 | { | ||
156 |
2/2✓ Branch 0 taken 2 times.
✓ Branch 1 taken 2 times.
|
4 | while (count--) { |
157 | 2 | addkey_s(&a->state[1], src, &a->round_key[rounds]); | |
158 | 2 | aes_crypt(a, 0, inv_sbox, dec_multbl); | |
159 |
1/2✗ Branch 0 not taken.
✓ Branch 1 taken 2 times.
|
2 | if (iv) { |
160 | ✗ | addkey_s(&a->state[0], iv, &a->state[0]); | |
161 | ✗ | memcpy(iv, src, 16); | |
162 | } | ||
163 | 2 | addkey_d(dst, &a->state[0], &a->round_key[0]); | |
164 | 2 | src += 16; | |
165 | 2 | dst += 16; | |
166 | } | ||
167 | 2 | } | |
168 | |||
169 | 5326 | void av_aes_crypt(AVAES *a, uint8_t *dst, const uint8_t *src, | |
170 | int count, uint8_t *iv, int decrypt) | ||
171 | { | ||
172 | 5326 | a->crypt(a, dst, src, count, iv, a->rounds); | |
173 | 5326 | } | |
174 | |||
175 | 12 | static void init_multbl2(uint32_t tbl[][256], const int c[4], | |
176 | const uint8_t *log8, const uint8_t *alog8, | ||
177 | const uint8_t *sbox) | ||
178 | { | ||
179 | int i; | ||
180 | |||
181 |
2/2✓ Branch 0 taken 3072 times.
✓ Branch 1 taken 12 times.
|
3084 | for (i = 0; i < 256; i++) { |
182 | 3072 | int x = sbox[i]; | |
183 |
2/2✓ Branch 0 taken 3060 times.
✓ Branch 1 taken 12 times.
|
3072 | if (x) { |
184 | int k, l, m, n; | ||
185 | 3060 | x = log8[x]; | |
186 | 3060 | k = alog8[x + log8[c[0]]]; | |
187 | 3060 | l = alog8[x + log8[c[1]]]; | |
188 | 3060 | m = alog8[x + log8[c[2]]]; | |
189 | 3060 | n = alog8[x + log8[c[3]]]; | |
190 | 3060 | tbl[0][i] = AV_NE(MKBETAG(k, l, m, n), MKTAG(k, l, m, n)); | |
191 | #if !CONFIG_SMALL | ||
192 | 3060 | tbl[1][i] = ROT(tbl[0][i], 8); | |
193 | 3060 | tbl[2][i] = ROT(tbl[0][i], 16); | |
194 | 3060 | tbl[3][i] = ROT(tbl[0][i], 24); | |
195 | #endif | ||
196 | } | ||
197 | } | ||
198 | 12 | } | |
199 | |||
200 | // this is based on the reference AES code by Paulo Barreto and Vincent Rijmen | ||
201 | 40 | int av_aes_init(AVAES *a, const uint8_t *key, int key_bits, int decrypt) | |
202 | { | ||
203 | 40 | int i, j, t, rconpointer = 0; | |
204 | uint8_t tk[8][4]; | ||
205 | 40 | int KC = key_bits >> 5; | |
206 | 40 | int rounds = KC + 6; | |
207 | uint8_t log8[256]; | ||
208 | uint8_t alog8[512]; | ||
209 | |||
210 |
2/2✓ Branch 0 taken 2 times.
✓ Branch 1 taken 38 times.
|
40 | a->crypt = decrypt ? aes_decrypt : aes_encrypt; |
211 | |||
212 |
2/2✓ Branch 0 taken 6 times.
✓ Branch 1 taken 34 times.
|
40 | if (!enc_multbl[FF_ARRAY_ELEMS(enc_multbl) - 1][FF_ARRAY_ELEMS(enc_multbl[0]) - 1]) { |
213 | 6 | j = 1; | |
214 |
2/2✓ Branch 0 taken 1530 times.
✓ Branch 1 taken 6 times.
|
1536 | for (i = 0; i < 255; i++) { |
215 | 1530 | alog8[i] = alog8[i + 255] = j; | |
216 | 1530 | log8[j] = i; | |
217 | 1530 | j ^= j + j; | |
218 |
2/2✓ Branch 0 taken 768 times.
✓ Branch 1 taken 762 times.
|
1530 | if (j > 255) |
219 | 768 | j ^= 0x11B; | |
220 | } | ||
221 |
2/2✓ Branch 0 taken 1536 times.
✓ Branch 1 taken 6 times.
|
1542 | for (i = 0; i < 256; i++) { |
222 |
2/2✓ Branch 0 taken 1530 times.
✓ Branch 1 taken 6 times.
|
1536 | j = i ? alog8[255 - log8[i]] : 0; |
223 | 1536 | j ^= (j << 1) ^ (j << 2) ^ (j << 3) ^ (j << 4); | |
224 | 1536 | j = (j ^ (j >> 8) ^ 99) & 255; | |
225 | 1536 | inv_sbox[j] = i; | |
226 | 1536 | sbox[i] = j; | |
227 | } | ||
228 | 6 | init_multbl2(dec_multbl, (const int[4]) { 0xe, 0x9, 0xd, 0xb }, | |
229 | log8, alog8, inv_sbox); | ||
230 | 6 | init_multbl2(enc_multbl, (const int[4]) { 0x2, 0x1, 0x1, 0x3 }, | |
231 | log8, alog8, sbox); | ||
232 | } | ||
233 | |||
234 |
1/6✗ Branch 0 not taken.
✓ Branch 1 taken 40 times.
✗ Branch 2 not taken.
✗ Branch 3 not taken.
✗ Branch 4 not taken.
✗ Branch 5 not taken.
|
40 | if (key_bits != 128 && key_bits != 192 && key_bits != 256) |
235 | ✗ | return AVERROR(EINVAL); | |
236 | |||
237 | 40 | a->rounds = rounds; | |
238 | |||
239 | 40 | memcpy(tk, key, KC * 4); | |
240 | 40 | memcpy(a->round_key[0].u8, key, KC * 4); | |
241 | |||
242 |
2/2✓ Branch 0 taken 400 times.
✓ Branch 1 taken 40 times.
|
440 | for (t = KC * 4; t < (rounds + 1) * 16; t += KC * 4) { |
243 |
2/2✓ Branch 0 taken 1600 times.
✓ Branch 1 taken 400 times.
|
2000 | for (i = 0; i < 4; i++) |
244 | 1600 | tk[0][i] ^= sbox[tk[KC - 1][(i + 1) & 3]]; | |
245 | 400 | tk[0][0] ^= rcon[rconpointer++]; | |
246 | |||
247 |
2/2✓ Branch 0 taken 1200 times.
✓ Branch 1 taken 400 times.
|
1600 | for (j = 1; j < KC; j++) { |
248 |
1/4✗ Branch 0 not taken.
✓ Branch 1 taken 1200 times.
✗ Branch 2 not taken.
✗ Branch 3 not taken.
|
1200 | if (KC != 8 || j != KC >> 1) |
249 |
2/2✓ Branch 0 taken 4800 times.
✓ Branch 1 taken 1200 times.
|
6000 | for (i = 0; i < 4; i++) |
250 | 4800 | tk[j][i] ^= tk[j - 1][i]; | |
251 | else | ||
252 | ✗ | for (i = 0; i < 4; i++) | |
253 | ✗ | tk[j][i] ^= sbox[tk[j - 1][i]]; | |
254 | } | ||
255 | |||
256 | 400 | memcpy((unsigned char*)a->round_key + t, tk, KC * 4); | |
257 | } | ||
258 | |||
259 |
2/2✓ Branch 0 taken 2 times.
✓ Branch 1 taken 38 times.
|
40 | if (decrypt) { |
260 |
2/2✓ Branch 0 taken 18 times.
✓ Branch 1 taken 2 times.
|
20 | for (i = 1; i < rounds; i++) { |
261 | av_aes_block tmp[3]; | ||
262 | 18 | tmp[2] = a->round_key[i]; | |
263 | 18 | subshift(&tmp[1], 0, sbox); | |
264 | 18 | mix(tmp, dec_multbl, 1, 3); | |
265 | 18 | a->round_key[i] = tmp[0]; | |
266 | } | ||
267 | } else { | ||
268 |
2/2✓ Branch 0 taken 190 times.
✓ Branch 1 taken 38 times.
|
228 | for (i = 0; i < (rounds + 1) >> 1; i++) |
269 | 190 | FFSWAP(av_aes_block, a->round_key[i], a->round_key[rounds - i]); | |
270 | } | ||
271 | |||
272 | 40 | return 0; | |
273 | } | ||
274 | |||
275 |