internal-photon256.c 17.4 KB
Newer Older
Rhys Weatherley committed
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
/*
 * Copyright (C) 2020 Southern Storm Software, Pty Ltd.
 *
 * Permission is hereby granted, free of charge, to any person obtaining a
 * copy of this software and associated documentation files (the "Software"),
 * to deal in the Software without restriction, including without limitation
 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
 * and/or sell copies of the Software, and to permit persons to whom the
 * Software is furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included
 * in all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
 * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
 * DEALINGS IN THE SOFTWARE.
 */

#include "internal-photon256.h"
#include "internal-util.h"

26 27
#if !defined(__AVR__)

Rhys Weatherley committed
28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481
/**
 * \brief Number of rounds in the PHOTON-256 permutation in bit-sliced form.
 */
#define PHOTON256_ROUNDS 12

/* Round constants for PHOTON-256 */
static uint32_t const photon256_rc[PHOTON256_ROUNDS] = {
    0x96d2f0e1, 0xb4f0d2c3, 0xf0b49687, 0x692d0f1e,
    0x5a1e3c2d, 0x3c785a4b, 0xe1a58796, 0x4b0f2d3c,
    0x1e5a7869, 0xa5e1c3d2, 0xd296b4a5, 0x2d694b5a
};

/**
 * \brief Evaluates the PHOTON-256 S-box in bit-sliced form.
 *
 * \param x0 Slice with bit 0 of all nibbles.
 * \param x1 Slice with bit 1 of all nibbles.
 * \param x2 Slice with bit 2 of all nibbles.
 * \param x3 Slice with bit 3 of all nibbles.
 *
 * This bit-sliced S-box implementation is based on the AVR version
 * "add_avr8_bitslice_asm" from the PHOTON-Beetle reference code.
 */
#define photon256_sbox(x0, x1, x2, x3) \
    do { \
        x1 ^= x2; \
        x3 ^= (x2 & x1); \
        t1 = x3; \
        x3 = (x3 & x1) ^ x2; \
        t2 = x3; \
        x3 ^= x0; \
        x3 = ~(x3); \
        x2 = x3; \
        t2 |= x0; \
        x0 ^= t1; \
        x1 ^= x0; \
        x2 |= x1; \
        x2 ^= t1; \
        x1 ^= t2; \
        x3 ^= x1; \
    } while (0)

/**
 * \brief Performs a field multiplication on the 8 nibbles in a row.
 *
 * \param a Field constant to multiply by.
 * \param x Bit-sliced form of the row, with bits 0..3 of each nibble
 * in bytes 0..3 of the word.
 *
 * \return a * x packed into the bytes of a word.
 */
static uint32_t photon256_field_multiply(uint8_t a, uint32_t x)
{
    /* For each 4-bit nibble we need to do this:
     *
     *      result = 0;
     *      for (bit = 0; bit < 4; ++ bit) {
     *          if ((a & (1 << bit)) != 0)
     *              result ^= x;
     *          if ((x & 0x08) != 0) {
     *              x = (x << 1) ^ 3;
     *          } else {
     *              x = (x << 1);
     *          }
     *      }
     *
     * We don't need to worry about constant time for "a" because it is a
     * known constant that isn't data-dependent.  But we do need to worry
     * about constant time for "x" as it is data.
     */
    uint32_t result = 0;
    uint32_t t;
    #define PARALLEL_CONDITIONAL_ADD(bit) \
        do { \
            if ((a) & (1 << (bit))) \
                result ^= x; \
        } while (0)
    #define PARALELL_ROTATE() \
        do { \
            t = x >> 24; \
            x = (x << 8) ^ t ^ (t << 8); \
        } while (0)
    PARALLEL_CONDITIONAL_ADD(0);
    PARALELL_ROTATE();
    PARALLEL_CONDITIONAL_ADD(1);
    PARALELL_ROTATE();
    PARALLEL_CONDITIONAL_ADD(2);
    PARALELL_ROTATE();
    PARALLEL_CONDITIONAL_ADD(3);
    return result;
}

/* http://programming.sirrida.de/perm_fn.html#bit_permute_step */
#define bit_permute_step(_y, mask, shift) \
    do { \
        uint32_t y = (_y); \
        uint32_t t = ((y >> (shift)) ^ y) & (mask); \
        (_y) = (y ^ t) ^ (t << (shift)); \
    } while (0)

/**
 * \brief Converts a PHOTON-256 state into bit-sliced form.
 *
 * \param out Points to the converted output.
 * \param in Points to the PHOTON-256 state to convert.
 */
static void photon256_to_sliced
    (uint32_t out[PHOTON256_STATE_SIZE / 4],
     const unsigned char in[PHOTON256_STATE_SIZE])
{
    /* We first scatter bits 0..3 of the nibbles to bytes 0..3 of the words.
     * Then we rearrange the bytes to group all bits N into word N.
     *
     * Permutation generated with "http://programming.sirrida.de/calcperm.php".
     *
     * P = [0 8 16 24 1 9 17 25 2 10 18 26 3 11 19 27
     *      4 12 20 28 5 13 21 29 6 14 22 30 7 15 23 31]
     */
    uint32_t t0, t1, t2, t3;
    #define TO_BITSLICED_PERM(x) \
        do { \
            bit_permute_step(x, 0x0a0a0a0a, 3); \
            bit_permute_step(x, 0x00cc00cc, 6); \
            bit_permute_step(x, 0x0000f0f0, 12); \
            bit_permute_step(x, 0x0000ff00, 8); \
        } while (0)
    #define FROM_BITSLICED_PERM(x) \
        do { \
            bit_permute_step(x, 0x00aa00aa, 7); \
            bit_permute_step(x, 0x0000cccc, 14); \
            bit_permute_step(x, 0x00f000f0, 4); \
            bit_permute_step(x, 0x0000ff00, 8); \
        } while (0)
    t0 = le_load_word32(in);
    t1 = le_load_word32(in + 4);
    t2 = le_load_word32(in + 8);
    t3 = le_load_word32(in + 12);
    TO_BITSLICED_PERM(t0);
    TO_BITSLICED_PERM(t1);
    TO_BITSLICED_PERM(t2);
    TO_BITSLICED_PERM(t3);
    out[0] = (t0 & 0x000000FFU) | ((t1 << 8) & 0x0000FF00U) |
             ((t2 << 16) & 0x00FF0000U) | ((t3 << 24) & 0xFF000000U);
    out[1] = ((t0 >> 8) & 0x000000FFU) | (t1 & 0x0000FF00U) |
             ((t2 << 8) & 0x00FF0000U) | ((t3 << 16) & 0xFF000000U);
    out[2] = ((t0 >> 16) & 0x000000FFU) | ((t1 >> 8) & 0x0000FF00U) |
             (t2 & 0x00FF0000U) | ((t3 << 8) & 0xFF000000U);
    out[3] = ((t0 >> 24) & 0x000000FFU) | ((t1 >> 16) & 0x0000FF00U) |
             ((t2 >> 8) & 0x00FF0000U) | (t3 & 0xFF000000U);
    t0 = le_load_word32(in + 16);
    t1 = le_load_word32(in + 20);
    t2 = le_load_word32(in + 24);
    t3 = le_load_word32(in + 28);
    TO_BITSLICED_PERM(t0);
    TO_BITSLICED_PERM(t1);
    TO_BITSLICED_PERM(t2);
    TO_BITSLICED_PERM(t3);
    out[4] = (t0 & 0x000000FFU) | ((t1 << 8) & 0x0000FF00U) |
             ((t2 << 16) & 0x00FF0000U) | ((t3 << 24) & 0xFF000000U);
    out[5] = ((t0 >> 8) & 0x000000FFU) | (t1 & 0x0000FF00U) |
             ((t2 << 8) & 0x00FF0000U) | ((t3 << 16) & 0xFF000000U);
    out[6] = ((t0 >> 16) & 0x000000FFU) | ((t1 >> 8) & 0x0000FF00U) |
             (t2 & 0x00FF0000U) | ((t3 << 8) & 0xFF000000U);
    out[7] = ((t0 >> 24) & 0x000000FFU) | ((t1 >> 16) & 0x0000FF00U) |
             ((t2 >> 8) & 0x00FF0000U) | (t3 & 0xFF000000U);
}

/**
 * \brief Converts a PHOTON-256 state from bit-sliced form.
 *
 * \param out Points to the converted output.
 * \param in Points to the PHOTON-256 state to convert.
 */
static void photon256_from_sliced
    (unsigned char out[PHOTON256_STATE_SIZE],
     const unsigned char in[PHOTON256_STATE_SIZE])
{
    /* Do the reverse of photon256_to_sliced() */
    uint32_t x0, x1, x2, x3;
    x0 =  ((uint32_t)(in[0])) |
         (((uint32_t)(in[4]))  << 8) |
         (((uint32_t)(in[8]))  << 16) |
         (((uint32_t)(in[12])) << 24);
    x1 =  ((uint32_t)(in[1])) |
         (((uint32_t)(in[5]))  << 8) |
         (((uint32_t)(in[9]))  << 16) |
         (((uint32_t)(in[13])) << 24);
    x2 =  ((uint32_t)(in[2])) |
         (((uint32_t)(in[6]))  << 8) |
         (((uint32_t)(in[10])) << 16) |
         (((uint32_t)(in[14])) << 24);
    x3 =  ((uint32_t)(in[3])) |
         (((uint32_t)(in[7]))  << 8) |
         (((uint32_t)(in[11])) << 16) |
         (((uint32_t)(in[15])) << 24);
    FROM_BITSLICED_PERM(x0);
    FROM_BITSLICED_PERM(x1);
    FROM_BITSLICED_PERM(x2);
    FROM_BITSLICED_PERM(x3);
    le_store_word32(out,      x0);
    le_store_word32(out + 4,  x1);
    le_store_word32(out + 8,  x2);
    le_store_word32(out + 12, x3);
    x0 =  ((uint32_t)(in[16])) |
         (((uint32_t)(in[20])) << 8) |
         (((uint32_t)(in[24])) << 16) |
         (((uint32_t)(in[28])) << 24);
    x1 =  ((uint32_t)(in[17])) |
         (((uint32_t)(in[21])) << 8) |
         (((uint32_t)(in[25])) << 16) |
         (((uint32_t)(in[29])) << 24);
    x2 =  ((uint32_t)(in[18])) |
         (((uint32_t)(in[22])) << 8) |
         (((uint32_t)(in[26])) << 16) |
         (((uint32_t)(in[30])) << 24);
    x3 =  ((uint32_t)(in[19])) |
         (((uint32_t)(in[23])) << 8) |
         (((uint32_t)(in[27])) << 16) |
         (((uint32_t)(in[31])) << 24);
    FROM_BITSLICED_PERM(x0);
    FROM_BITSLICED_PERM(x1);
    FROM_BITSLICED_PERM(x2);
    FROM_BITSLICED_PERM(x3);
    le_store_word32(out + 16, x0);
    le_store_word32(out + 20, x1);
    le_store_word32(out + 24, x2);
    le_store_word32(out + 28, x3);
}

#if defined(LW_UTIL_LITTLE_ENDIAN)
/* Index the bit-sliced state bytes in little-endian byte order */
#define READ_ROW0() \
     (((uint32_t)(S.bytes[0])) | \
     (((uint32_t)(S.bytes[4]))  << 8)  | \
     (((uint32_t)(S.bytes[8]))  << 16) | \
     (((uint32_t)(S.bytes[12])) << 24))
#define READ_ROW1() \
     (((uint32_t)(S.bytes[1])) | \
     (((uint32_t)(S.bytes[5]))  << 8)  | \
     (((uint32_t)(S.bytes[9]))  << 16) | \
     (((uint32_t)(S.bytes[13])) << 24))
#define READ_ROW2() \
     (((uint32_t)(S.bytes[2])) | \
     (((uint32_t)(S.bytes[6]))  << 8)  | \
     (((uint32_t)(S.bytes[10])) << 16) | \
     (((uint32_t)(S.bytes[14])) << 24))
#define READ_ROW3() \
     (((uint32_t)(S.bytes[3])) | \
     (((uint32_t)(S.bytes[7]))  << 8)  | \
     (((uint32_t)(S.bytes[11])) << 16) | \
     (((uint32_t)(S.bytes[15])) << 24))
#define READ_ROW4() \
     (((uint32_t)(S.bytes[16])) | \
     (((uint32_t)(S.bytes[20])) << 8)  | \
     (((uint32_t)(S.bytes[24])) << 16) | \
     (((uint32_t)(S.bytes[28])) << 24))
#define READ_ROW5() \
     (((uint32_t)(S.bytes[17])) | \
     (((uint32_t)(S.bytes[21])) << 8)  | \
     (((uint32_t)(S.bytes[25])) << 16) | \
     (((uint32_t)(S.bytes[29])) << 24))
#define READ_ROW6() \
     (((uint32_t)(S.bytes[18])) | \
     (((uint32_t)(S.bytes[22])) << 8)  | \
     (((uint32_t)(S.bytes[26])) << 16) | \
     (((uint32_t)(S.bytes[30])) << 24))
#define READ_ROW7() \
     (((uint32_t)(S.bytes[19])) | \
     (((uint32_t)(S.bytes[23])) << 8)  | \
     (((uint32_t)(S.bytes[27])) << 16) | \
     (((uint32_t)(S.bytes[31])) << 24))
#define WRITE_ROW(row, value) \
    do { \
        if ((row) < 4) { \
            S.bytes[(row)]      = (uint8_t)(value); \
            S.bytes[(row) + 4]  = (uint8_t)((value) >> 8); \
            S.bytes[(row) + 8]  = (uint8_t)((value) >> 16); \
            S.bytes[(row) + 12] = (uint8_t)((value) >> 24); \
        } else { \
            S.bytes[(row) + 12] = (uint8_t)(value); \
            S.bytes[(row) + 16] = (uint8_t)((value) >> 8); \
            S.bytes[(row) + 20] = (uint8_t)((value) >> 16); \
            S.bytes[(row) + 24] = (uint8_t)((value) >> 24); \
        } \
    } while (0)
#else
/* Index the bit-sliced state bytes in big-endian byte order */
#define READ_ROW0() \
     (((uint32_t)(S.bytes[3])) | \
     (((uint32_t)(S.bytes[7]))  << 8)  | \
     (((uint32_t)(S.bytes[11])) << 16) | \
     (((uint32_t)(S.bytes[15])) << 24))
#define READ_ROW1() \
     (((uint32_t)(S.bytes[2])) | \
     (((uint32_t)(S.bytes[6]))  << 8)  | \
     (((uint32_t)(S.bytes[10])) << 16) | \
     (((uint32_t)(S.bytes[14])) << 24))
#define READ_ROW2() \
     (((uint32_t)(S.bytes[1])) | \
     (((uint32_t)(S.bytes[5]))  << 8)  | \
     (((uint32_t)(S.bytes[9]))  << 16) | \
     (((uint32_t)(S.bytes[13])) << 24))
#define READ_ROW3() \
     (((uint32_t)(S.bytes[0])) | \
     (((uint32_t)(S.bytes[4]))  << 8)  | \
     (((uint32_t)(S.bytes[8]))  << 16) | \
     (((uint32_t)(S.bytes[12])) << 24))
#define READ_ROW4() \
     (((uint32_t)(S.bytes[19])) | \
     (((uint32_t)(S.bytes[23])) << 8)  | \
     (((uint32_t)(S.bytes[27])) << 16) | \
     (((uint32_t)(S.bytes[31])) << 24))
#define READ_ROW5() \
     (((uint32_t)(S.bytes[18])) | \
     (((uint32_t)(S.bytes[22])) << 8)  | \
     (((uint32_t)(S.bytes[26])) << 16) | \
     (((uint32_t)(S.bytes[30])) << 24))
#define READ_ROW6() \
     (((uint32_t)(S.bytes[17])) | \
     (((uint32_t)(S.bytes[21])) << 8)  | \
     (((uint32_t)(S.bytes[25])) << 16) | \
     (((uint32_t)(S.bytes[29])) << 24))
#define READ_ROW7() \
     (((uint32_t)(S.bytes[16])) | \
     (((uint32_t)(S.bytes[20])) << 8)  | \
     (((uint32_t)(S.bytes[24])) << 16) | \
     (((uint32_t)(S.bytes[28])) << 24))
#define WRITE_ROW(row, value) \
    do { \
        if ((row) < 4) { \
            S.bytes[3  - (row)] = (uint8_t)(value); \
            S.bytes[7  - (row)] = (uint8_t)((value) >> 8); \
            S.bytes[11 - (row)] = (uint8_t)((value) >> 16); \
            S.bytes[15 - (row)] = (uint8_t)((value) >> 24); \
        } else { \
            S.bytes[20 - (row)] = (uint8_t)(value); \
            S.bytes[24 - (row)] = (uint8_t)((value) >> 8); \
            S.bytes[28 - (row)] = (uint8_t)((value) >> 16); \
            S.bytes[32 - (row)] = (uint8_t)((value) >> 24); \
        } \
    } while (0)
#endif

void photon256_permute(unsigned char state[PHOTON256_STATE_SIZE])
{
    union {
        uint32_t words[PHOTON256_STATE_SIZE / 4];
        uint8_t bytes[PHOTON256_STATE_SIZE];
    } S;
    uint32_t t0, t1, t2, t3, t4, t5, t6, t7, t8;
    uint8_t round;

    /* Convert the state into bit-sliced form */
    photon256_to_sliced(S.words, state);

    /* Perform all 12 permutation rounds */
    for (round = 0; round < PHOTON256_ROUNDS; ++round) {
        /* Add the constants for this round */
        t0 = photon256_rc[round];
        S.words[0] ^= t0 & 0x01010101U;
        t0 >>= 1;
        S.words[1] ^= t0 & 0x01010101U;
        t0 >>= 1;
        S.words[2] ^= t0 & 0x01010101U;
        t0 >>= 1;
        S.words[3] ^= t0 & 0x01010101U;
        t0 >>= 1;
        S.words[4] ^= t0 & 0x01010101U;
        t0 >>= 1;
        S.words[5] ^= t0 & 0x01010101U;
        t0 >>= 1;
        S.words[6] ^= t0 & 0x01010101U;
        t0 >>= 1;
        S.words[7] ^= t0 & 0x01010101U;

        /* Apply the sbox to all nibbles in the state */
        photon256_sbox(S.words[0], S.words[1], S.words[2], S.words[3]);
        photon256_sbox(S.words[4], S.words[5], S.words[6], S.words[7]);

        /* Rotate all rows left by the row number.
         *
         * We do this by applying permutations to the top and bottom words
         * to rearrange the bits into the rotated form.  Permutations
         * generated with "http://programming.sirrida.de/calcperm.php".
         *
         * P_top = [0 1 2 3 4 5 6 7 15 8 9 10 11 12 13 14 22 23
         *          16 17 18 19 20 21 29 30 31 24 25 26 27 28]
         * P_bot = [4 5 6 7 0 1 2 3 11 12 13 14 15 8 9 10 18 19
         *          20 21 22 23 16 17 25 26 27 28 29 30 31 24
         */
        #define TOP_ROTATE_PERM(x) \
            do { \
                t1 = (x); \
                bit_permute_step(t1, 0x07030100, 4); \
                bit_permute_step(t1, 0x22331100, 2); \
                bit_permute_step(t1, 0x55005500, 1); \
                (x) = t1; \
            } while (0)
        #define BOTTOM_ROTATE_PERM(x) \
            do { \
                t1 = (x); \
                bit_permute_step(t1, 0x080c0e0f, 4); \
                bit_permute_step(t1, 0x22331100, 2); \
                bit_permute_step(t1, 0x55005500, 1); \
                (x) = t1; \
            } while (0)
        TOP_ROTATE_PERM(S.words[0]);
        TOP_ROTATE_PERM(S.words[1]);
        TOP_ROTATE_PERM(S.words[2]);
        TOP_ROTATE_PERM(S.words[3]);
        BOTTOM_ROTATE_PERM(S.words[4]);
        BOTTOM_ROTATE_PERM(S.words[5]);
        BOTTOM_ROTATE_PERM(S.words[6]);
        BOTTOM_ROTATE_PERM(S.words[7]);

        /* Mix the columns */
        #define MUL(a, x) (photon256_field_multiply((a), (x)))
        t0 = READ_ROW0();
        t1 = READ_ROW1();
        t2 = READ_ROW2();
        t3 = READ_ROW3();
        t4 = READ_ROW4();
        t5 = READ_ROW5();
        t6 = READ_ROW6();
        t7 = READ_ROW7();
        t8 = MUL(0x02, t0) ^ MUL(0x04, t1) ^ MUL(0x02, t2) ^ MUL(0x0b, t3) ^
             MUL(0x02, t4) ^ MUL(0x08, t5) ^ MUL(0x05, t6) ^ MUL(0x06, t7);
        WRITE_ROW(0, t8);
        t8 = MUL(0x0c, t0) ^ MUL(0x09, t1) ^ MUL(0x08, t2) ^ MUL(0x0d, t3) ^
             MUL(0x07, t4) ^ MUL(0x07, t5) ^ MUL(0x05, t6) ^ MUL(0x02, t7);
        WRITE_ROW(1, t8);
        t8 = MUL(0x04, t0) ^ MUL(0x04, t1) ^ MUL(0x0d, t2) ^ MUL(0x0d, t3) ^
             MUL(0x09, t4) ^ MUL(0x04, t5) ^ MUL(0x0d, t6) ^ MUL(0x09, t7);
        WRITE_ROW(2, t8);
        t8 = MUL(0x01, t0) ^ MUL(0x06, t1) ^ MUL(0x05, t2) ^ MUL(0x01, t3) ^
             MUL(0x0c, t4) ^ MUL(0x0d, t5) ^ MUL(0x0f, t6) ^ MUL(0x0e, t7);
        WRITE_ROW(3, t8);
        t8 = MUL(0x0f, t0) ^ MUL(0x0c, t1) ^ MUL(0x09, t2) ^ MUL(0x0d, t3) ^
             MUL(0x0e, t4) ^ MUL(0x05, t5) ^ MUL(0x0e, t6) ^ MUL(0x0d, t7);
        WRITE_ROW(4, t8);
        t8 = MUL(0x09, t0) ^ MUL(0x0e, t1) ^ MUL(0x05, t2) ^ MUL(0x0f, t3) ^
             MUL(0x04, t4) ^ MUL(0x0c, t5) ^ MUL(0x09, t6) ^ MUL(0x06, t7);
        WRITE_ROW(5, t8);
        t8 = MUL(0x0c, t0) ^ MUL(0x02, t1) ^ MUL(0x02, t2) ^ MUL(0x0a, t3) ^
             MUL(0x03, t4) ^ MUL(0x01, t5) ^ MUL(0x01, t6) ^ MUL(0x0e, t7);
        WRITE_ROW(6, t8);
        t8 = MUL(0x0f, t0) ^ MUL(0x01, t1) ^ MUL(0x0d, t2) ^ MUL(0x0a, t3) ^
             MUL(0x05, t4) ^ MUL(0x0a, t5) ^ MUL(0x02, t6) ^ MUL(0x03, t7);
        WRITE_ROW(7, t8);
    }

    /* Convert back from bit-sliced form to regular form */
    photon256_from_sliced(state, S.bytes);
}
482 483

#endif /* !__AVR__ */