1/* 2 * Copyright (c) 2016 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/* see inner.h */ 28void 29br_aes_ct64_bitslice_Sbox(uint64_t *q) 30{ 31 /* 32 * This S-box implementation is a straightforward translation of 33 * the circuit described by Boyar and Peralta in "A new 34 * combinational logic minimization technique with applications 35 * to cryptology" (https://eprint.iacr.org/2009/191.pdf). 36 * 37 * Note that variables x* (input) and s* (output) are numbered 38 * in "reverse" order (x0 is the high bit, x7 is the low bit). 39 */ 40 41 uint64_t x0, x1, x2, x3, x4, x5, x6, x7; 42 uint64_t y1, y2, y3, y4, y5, y6, y7, y8, y9; 43 uint64_t y10, y11, y12, y13, y14, y15, y16, y17, y18, y19; 44 uint64_t y20, y21; 45 uint64_t z0, z1, z2, z3, z4, z5, z6, z7, z8, z9; 46 uint64_t z10, z11, z12, z13, z14, z15, z16, z17; 47 uint64_t t0, t1, t2, t3, t4, t5, t6, t7, t8, t9; 48 uint64_t t10, t11, t12, t13, t14, t15, t16, t17, t18, t19; 49 uint64_t t20, t21, t22, t23, t24, t25, t26, t27, t28, t29; 50 uint64_t t30, t31, t32, t33, t34, t35, t36, t37, t38, t39; 51 uint64_t t40, t41, t42, t43, t44, t45, t46, t47, t48, t49; 52 uint64_t t50, t51, t52, t53, t54, t55, t56, t57, t58, t59; 53 uint64_t t60, t61, t62, t63, t64, t65, t66, t67; 54 uint64_t s0, s1, s2, s3, s4, s5, s6, s7; 55 56 x0 = q[7]; 57 x1 = q[6]; 58 x2 = q[5]; 59 x3 = q[4]; 60 x4 = q[3]; 61 x5 = q[2]; 62 x6 = q[1]; 63 x7 = q[0]; 64 65 /* 66 * Top linear transformation. 67 */ 68 y14 = x3 ^ x5; 69 y13 = x0 ^ x6; 70 y9 = x0 ^ x3; 71 y8 = x0 ^ x5; 72 t0 = x1 ^ x2; 73 y1 = t0 ^ x7; 74 y4 = y1 ^ x3; 75 y12 = y13 ^ y14; 76 y2 = y1 ^ x0; 77 y5 = y1 ^ x6; 78 y3 = y5 ^ y8; 79 t1 = x4 ^ y12; 80 y15 = t1 ^ x5; 81 y20 = t1 ^ x1; 82 y6 = y15 ^ x7; 83 y10 = y15 ^ t0; 84 y11 = y20 ^ y9; 85 y7 = x7 ^ y11; 86 y17 = y10 ^ y11; 87 y19 = y10 ^ y8; 88 y16 = t0 ^ y11; 89 y21 = y13 ^ y16; 90 y18 = x0 ^ y16; 91 92 /* 93 * Non-linear section. 94 */ 95 t2 = y12 & y15; 96 t3 = y3 & y6; 97 t4 = t3 ^ t2; 98 t5 = y4 & x7; 99 t6 = t5 ^ t2; 100 t7 = y13 & y16; 101 t8 = y5 & y1; 102 t9 = t8 ^ t7; 103 t10 = y2 & y7; 104 t11 = t10 ^ t7; 105 t12 = y9 & y11; 106 t13 = y14 & y17; 107 t14 = t13 ^ t12; 108 t15 = y8 & y10; 109 t16 = t15 ^ t12; 110 t17 = t4 ^ t14; 111 t18 = t6 ^ t16; 112 t19 = t9 ^ t14; 113 t20 = t11 ^ t16; 114 t21 = t17 ^ y20; 115 t22 = t18 ^ y19; 116 t23 = t19 ^ y21; 117 t24 = t20 ^ y18; 118 119 t25 = t21 ^ t22; 120 t26 = t21 & t23; 121 t27 = t24 ^ t26; 122 t28 = t25 & t27; 123 t29 = t28 ^ t22; 124 t30 = t23 ^ t24; 125 t31 = t22 ^ t26; 126 t32 = t31 & t30; 127 t33 = t32 ^ t24; 128 t34 = t23 ^ t33; 129 t35 = t27 ^ t33; 130 t36 = t24 & t35; 131 t37 = t36 ^ t34; 132 t38 = t27 ^ t36; 133 t39 = t29 & t38; 134 t40 = t25 ^ t39; 135 136 t41 = t40 ^ t37; 137 t42 = t29 ^ t33; 138 t43 = t29 ^ t40; 139 t44 = t33 ^ t37; 140 t45 = t42 ^ t41; 141 z0 = t44 & y15; 142 z1 = t37 & y6; 143 z2 = t33 & x7; 144 z3 = t43 & y16; 145 z4 = t40 & y1; 146 z5 = t29 & y7; 147 z6 = t42 & y11; 148 z7 = t45 & y17; 149 z8 = t41 & y10; 150 z9 = t44 & y12; 151 z10 = t37 & y3; 152 z11 = t33 & y4; 153 z12 = t43 & y13; 154 z13 = t40 & y5; 155 z14 = t29 & y2; 156 z15 = t42 & y9; 157 z16 = t45 & y14; 158 z17 = t41 & y8; 159 160 /* 161 * Bottom linear transformation. 162 */ 163 t46 = z15 ^ z16; 164 t47 = z10 ^ z11; 165 t48 = z5 ^ z13; 166 t49 = z9 ^ z10; 167 t50 = z2 ^ z12; 168 t51 = z2 ^ z5; 169 t52 = z7 ^ z8; 170 t53 = z0 ^ z3; 171 t54 = z6 ^ z7; 172 t55 = z16 ^ z17; 173 t56 = z12 ^ t48; 174 t57 = t50 ^ t53; 175 t58 = z4 ^ t46; 176 t59 = z3 ^ t54; 177 t60 = t46 ^ t57; 178 t61 = z14 ^ t57; 179 t62 = t52 ^ t58; 180 t63 = t49 ^ t58; 181 t64 = z4 ^ t59; 182 t65 = t61 ^ t62; 183 t66 = z1 ^ t63; 184 s0 = t59 ^ t63; 185 s6 = t56 ^ ~t62; 186 s7 = t48 ^ ~t60; 187 t67 = t64 ^ t65; 188 s3 = t53 ^ t66; 189 s4 = t51 ^ t66; 190 s5 = t47 ^ t65; 191 s1 = t64 ^ ~s3; 192 s2 = t55 ^ ~t67; 193 194 q[7] = s0; 195 q[6] = s1; 196 q[5] = s2; 197 q[4] = s3; 198 q[3] = s4; 199 q[2] = s5; 200 q[1] = s6; 201 q[0] = s7; 202} 203 204/* see inner.h */ 205void 206br_aes_ct64_ortho(uint64_t *q) 207{ 208#define SWAPN(cl, ch, s, x, y) do { \ 209 uint64_t a, b; \ 210 a = (x); \ 211 b = (y); \ 212 (x) = (a & (uint64_t)cl) | ((b & (uint64_t)cl) << (s)); \ 213 (y) = ((a & (uint64_t)ch) >> (s)) | (b & (uint64_t)ch); \ 214 } while (0) 215 216#define SWAP2(x, y) SWAPN(0x5555555555555555, 0xAAAAAAAAAAAAAAAA, 1, x, y) 217#define SWAP4(x, y) SWAPN(0x3333333333333333, 0xCCCCCCCCCCCCCCCC, 2, x, y) 218#define SWAP8(x, y) SWAPN(0x0F0F0F0F0F0F0F0F, 0xF0F0F0F0F0F0F0F0, 4, x, y) 219 220 SWAP2(q[0], q[1]); 221 SWAP2(q[2], q[3]); 222 SWAP2(q[4], q[5]); 223 SWAP2(q[6], q[7]); 224 225 SWAP4(q[0], q[2]); 226 SWAP4(q[1], q[3]); 227 SWAP4(q[4], q[6]); 228 SWAP4(q[5], q[7]); 229 230 SWAP8(q[0], q[4]); 231 SWAP8(q[1], q[5]); 232 SWAP8(q[2], q[6]); 233 SWAP8(q[3], q[7]); 234} 235 236/* see inner.h */ 237void 238br_aes_ct64_interleave_in(uint64_t *q0, uint64_t *q1, const uint32_t *w) 239{ 240 uint64_t x0, x1, x2, x3; 241 242 x0 = w[0]; 243 x1 = w[1]; 244 x2 = w[2]; 245 x3 = w[3]; 246 x0 |= (x0 << 16); 247 x1 |= (x1 << 16); 248 x2 |= (x2 << 16); 249 x3 |= (x3 << 16); 250 x0 &= (uint64_t)0x0000FFFF0000FFFF; 251 x1 &= (uint64_t)0x0000FFFF0000FFFF; 252 x2 &= (uint64_t)0x0000FFFF0000FFFF; 253 x3 &= (uint64_t)0x0000FFFF0000FFFF; 254 x0 |= (x0 << 8); 255 x1 |= (x1 << 8); 256 x2 |= (x2 << 8); 257 x3 |= (x3 << 8); 258 x0 &= (uint64_t)0x00FF00FF00FF00FF; 259 x1 &= (uint64_t)0x00FF00FF00FF00FF; 260 x2 &= (uint64_t)0x00FF00FF00FF00FF; 261 x3 &= (uint64_t)0x00FF00FF00FF00FF; 262 *q0 = x0 | (x2 << 8); 263 *q1 = x1 | (x3 << 8); 264} 265 266/* see inner.h */ 267void 268br_aes_ct64_interleave_out(uint32_t *w, uint64_t q0, uint64_t q1) 269{ 270 uint64_t x0, x1, x2, x3; 271 272 x0 = q0 & (uint64_t)0x00FF00FF00FF00FF; 273 x1 = q1 & (uint64_t)0x00FF00FF00FF00FF; 274 x2 = (q0 >> 8) & (uint64_t)0x00FF00FF00FF00FF; 275 x3 = (q1 >> 8) & (uint64_t)0x00FF00FF00FF00FF; 276 x0 |= (x0 >> 8); 277 x1 |= (x1 >> 8); 278 x2 |= (x2 >> 8); 279 x3 |= (x3 >> 8); 280 x0 &= (uint64_t)0x0000FFFF0000FFFF; 281 x1 &= (uint64_t)0x0000FFFF0000FFFF; 282 x2 &= (uint64_t)0x0000FFFF0000FFFF; 283 x3 &= (uint64_t)0x0000FFFF0000FFFF; 284 w[0] = (uint32_t)x0 | (uint32_t)(x0 >> 16); 285 w[1] = (uint32_t)x1 | (uint32_t)(x1 >> 16); 286 w[2] = (uint32_t)x2 | (uint32_t)(x2 >> 16); 287 w[3] = (uint32_t)x3 | (uint32_t)(x3 >> 16); 288} 289 290static const unsigned char Rcon[] = { 291 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x1B, 0x36 292}; 293 294static uint32_t 295sub_word(uint32_t x) 296{ 297 uint64_t q[8]; 298 299 memset(q, 0, sizeof q); 300 q[0] = x; 301 br_aes_ct64_ortho(q); 302 br_aes_ct64_bitslice_Sbox(q); 303 br_aes_ct64_ortho(q); 304 return (uint32_t)q[0]; 305} 306 307/* see inner.h */ 308unsigned 309br_aes_ct64_keysched(uint64_t *comp_skey, const void *key, size_t key_len) 310{ 311 unsigned num_rounds; 312 int i, j, k, nk, nkf; 313 uint32_t tmp; 314 uint32_t skey[60]; 315 316 switch (key_len) { 317 case 16: 318 num_rounds = 10; 319 break; 320 case 24: 321 num_rounds = 12; 322 break; 323 case 32: 324 num_rounds = 14; 325 break; 326 default: 327 /* abort(); */ 328 return 0; 329 } 330 nk = (int)(key_len >> 2); 331 nkf = (int)((num_rounds + 1) << 2); 332 br_range_dec32le(skey, (key_len >> 2), key); 333 tmp = skey[(key_len >> 2) - 1]; 334 for (i = nk, j = 0, k = 0; i < nkf; i ++) { 335 if (j == 0) { 336 tmp = (tmp << 24) | (tmp >> 8); 337 tmp = sub_word(tmp) ^ Rcon[k]; 338 } else if (nk > 6 && j == 4) { 339 tmp = sub_word(tmp); 340 } 341 tmp ^= skey[i - nk]; 342 skey[i] = tmp; 343 if (++ j == nk) { 344 j = 0; 345 k ++; 346 } 347 } 348 349 for (i = 0, j = 0; i < nkf; i += 4, j += 2) { 350 uint64_t q[8]; 351 352 br_aes_ct64_interleave_in(&q[0], &q[4], skey + i); 353 q[1] = q[0]; 354 q[2] = q[0]; 355 q[3] = q[0]; 356 q[5] = q[4]; 357 q[6] = q[4]; 358 q[7] = q[4]; 359 br_aes_ct64_ortho(q); 360 comp_skey[j + 0] = 361 (q[0] & (uint64_t)0x1111111111111111) 362 | (q[1] & (uint64_t)0x2222222222222222) 363 | (q[2] & (uint64_t)0x4444444444444444) 364 | (q[3] & (uint64_t)0x8888888888888888); 365 comp_skey[j + 1] = 366 (q[4] & (uint64_t)0x1111111111111111) 367 | (q[5] & (uint64_t)0x2222222222222222) 368 | (q[6] & (uint64_t)0x4444444444444444) 369 | (q[7] & (uint64_t)0x8888888888888888); 370 } 371 return num_rounds; 372} 373 374/* see inner.h */ 375void 376br_aes_ct64_skey_expand(uint64_t *skey, 377 unsigned num_rounds, const uint64_t *comp_skey) 378{ 379 unsigned u, v, n; 380 381 n = (num_rounds + 1) << 1; 382 for (u = 0, v = 0; u < n; u ++, v += 4) { 383 uint64_t x0, x1, x2, x3; 384 385 x0 = x1 = x2 = x3 = comp_skey[u]; 386 x0 &= (uint64_t)0x1111111111111111; 387 x1 &= (uint64_t)0x2222222222222222; 388 x2 &= (uint64_t)0x4444444444444444; 389 x3 &= (uint64_t)0x8888888888888888; 390 x1 >>= 1; 391 x2 >>= 2; 392 x3 >>= 3; 393 skey[v + 0] = (x0 << 4) - x0; 394 skey[v + 1] = (x1 << 4) - x1; 395 skey[v + 2] = (x2 << 4) - x2; 396 skey[v + 3] = (x3 << 4) - x3; 397 } 398} 399