serpent-encrypt.c 14.2 KB
Newer Older
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 26 27 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
/* serpent-encrypt.c
 *
 * The serpent block cipher.
 *
 * For more details on this algorithm, see the Serpent website at
 * http://www.cl.cam.ac.uk/~rja14/serpent.html
 */

/* nettle, low-level cryptographics library
 *
 * Copyright (C) 2011  Niels Möller
 * Copyright (C) 2010, 2011  Simon Josefsson
 * Copyright (C) 2003, 2004, 2005 Free Software Foundation, Inc.
 *  
 * The nettle library is free software; you can redistribute it and/or modify
 * it under the terms of the GNU Lesser General Public License as published by
 * the Free Software Foundation; either version 2.1 of the License, or (at your
 * option) any later version.
 * 
 * The nettle library is distributed in the hope that it will be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
 * or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU Lesser General Public
 * License for more details.
 * 
 * You should have received a copy of the GNU Lesser General Public License
 * along with the nettle library; see the file COPYING.LIB.  If not, write to
 * the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
 * MA 02111-1307, USA.
 */

/* This file is derived from cipher/serpent.c in Libgcrypt v1.4.6.
   The adaption to Nettle was made by Simon Josefsson on 2010-12-07
   with final touches on 2011-05-30.  Changes include replacing
   libgcrypt with nettle in the license template, renaming
   serpent_context to serpent_ctx, renaming u32 to uint32_t, removing
   libgcrypt stubs and selftests, modifying entry function prototypes,
   using FOR_BLOCKS to iterate through data in encrypt/decrypt, using
   LE_READ_UINT32 and LE_WRITE_UINT32 to access data in
   encrypt/decrypt, and running indent on the code. */

#if HAVE_CONFIG_H
#include "config.h"
#endif

#include <assert.h>
#include <limits.h>

#include "serpent.h"

#include "macros.h"
#include "serpent-internal.h"

/* These are the S-Boxes of Serpent.  They are copied from Serpents
   reference implementation (the optimized one, contained in
   `floppy2') and are therefore:

     Copyright (C) 1998 Ross Anderson, Eli Biham, Lars Knudsen.

  To quote the Serpent homepage
  (http://www.cl.cam.ac.uk/~rja14/serpent.html):

  "Serpent is now completely in the public domain, and we impose no
   restrictions on its use.  This was announced on the 21st August at
   the First AES Candidate Conference. The optimised implementations
   in the submission package are now under the GNU PUBLIC LICENSE
   (GPL), although some comments in the code still say otherwise. You
   are welcome to use Serpent for any application."  */

/* FIXME: Except when used within the key schedule, the inputs are not
   used after the substitution, and hence we could allow them to be
   destroyed. Can this freedom be used to optimize the sboxes? */

/* S0:  3  8 15  1 10  6  5 11 14 13  4  2  7  0  9 12 */
74
/* Could easily let y0, y1 overlap with x0, x1, and possibly also x2 and y2 */
75
#define SBOX0(x0, x1, x2, x3, y0, y1, y2, y3)	\
76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94
  do {							\
    y3  = x1 ^ x2;					\
    y0  = x0 | x3;					\
    y1  = x0 ^ x1;					\
    y3 ^= y0;						\
    y2  = x2 | y3;					\
    x0 ^= x3;						\
    y2 &= x3;						\
    x3 ^= x2;						\
    x2 |= x1;						\
    y0  = y1 & x2;					\
    y2 ^= y0;						\
    y0 &= y2;						\
    y0 ^= x2;						\
    x1 &= x0;						\
    y0 ^= x0;						\
    y0  = ~ y0;						\
    y1  = y0 ^ x1;					\
    y1 ^= x3;						\
95 96
  } while (0)

97
/* FIXME: Arrange for some overlap between inputs and outputs? */
98
/* S1: 15 12  2  7  9  0  5 10  1 11 14  8  6 13  3  4 */
99
#define SBOX1(x0, x1, x2, x3, y0, y1, y2, y3)	\
100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118
  do {							\
    y1  = x0 | x3;					\
    y2  = x2 ^ x3;					\
    y0  = ~ x1;						\
    y3  = x0   ^ x2;					\
    y0 |= x0;						\
    y3 &= x3;						\
    x0 = y1 & y2;					\
    y3 |= x1;						\
    y2 ^= y0;						\
    y3 ^= x0;						\
    x0  = y1 ^ y3;					\
    x0 ^= y2;						\
    y1  = x1   & x3;					\
    y1 ^= x0;						\
    x3  = y3 | y1;					\
    y3  = ~ y3;						\
    y0 &= x3;						\
    y0 ^= x2;						\
119 120
  } while (0)

121
/* FIXME: Arrange for some overlap between inputs and outputs? */
122
/* S2:  8  6  7  9  3 12 10 15 13  1 14  4  0 11  5  2 */
123
#define SBOX2(x0, x1, x2, x3, y0, y1, y2, y3)	\
124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140
  do {							\
    y2  = x0 | x2;					\
    y1  = x0 ^ x1;					\
    y3  = x3 ^ y2;					\
    y0  = y1 ^ y3;					\
    x3 |= x0;						\
    x2 ^= y0;						\
    x0  = x1 ^ x2;					\
    x2 |= x1;						\
    x0 &= y2;						\
    y3 ^= x2;						\
    y1 |= y3;						\
    y1 ^= x0;						\
    y2  = y3 ^ y1;					\
    y2 ^= x1;						\
    y3  = ~ y3;						\
    y2 ^= x3;						\
141 142 143
  } while (0)

/* S3:  0 15 11  8 12  9  6  3 13  1  2  4 10  7  5 14 */
144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164
/* Original single-assignment form:
   
     t01 = x0  ^ x2;
     t02 = x0  | x3;
     t03 = x0  & x3;
     t04 = t01 & t02;
     t05 = x1  | t03;
     t06 = x0  & x1;
     t07 = x3  ^ t04;
     t08 = x2  | t06;
     t09 = x1  ^ t07;
     t10 = x3  & t05;
     t11 = t02 ^ t10;
     y3  = t08 ^ t09;
     t13 = x3  | y3;
     t14 = x0  | t07;
     t15 = x1  & t13;
     y2  = t08 ^ t11;
     y0  = t14 ^ t15;
     y1  = t05 ^ t04;
*/
165
#define SBOX3(x0, x1, x2, x3, y0, y1, y2, y3)	\
166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184
  do {							\
    y2  = x0 ^ x2;					\
    y0  = x0 | x3;					\
    y3  = x0 & x3;					\
    y1  = y2 & y0;					\
    y3 |= x1;						\
    y2  = x0   & x1;					\
    y2 |= x2;						\
    x2  = x3   ^ y1;					\
    y1 ^= y3;						\
    x0 |= x2;						\
    x2 ^= x1;						\
    y3 &= x3;						\
    y0 ^= y3;						\
    y3  = y2 ^ x2;					\
    y2 ^= y0;						\
    x3 |= y3;						\
    x1 &= x3;						\
    y0  = x0 ^ x1;					\
185 186
  } while (0)

187

188
/* S4:  1 15  8  3 12  0 11  6  2  5  4 10  9 14  7 13 */
189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209
/* Original single-assignment form:
    t01 = x0  | x1;
    t02 = x1  | x2;
    t03 = x0  ^ t02;
    t04 = x1  ^ x3;
    t05 = x3  | t03;
    t06 = x3  & t01;
    y3  = t03 ^ t06;
    t08 = y3  & t04;
    t09 = t04 & t05;
    t10 = x2  ^ t06;
    t11 = x1  & x2;
    t12 = t04 ^ t08;
    t13 = t11 | t03;
    t14 = t10 ^ t09;
    t15 = x0  & t05;
    t16 = t11 | t12;
    y2  = t13 ^ t08;
    y1  = t15 ^ t16;
    y0  =     ~ t14;
*/
210
#define SBOX4(x0, x1, x2, x3, y0, y1, y2, y3)	\
211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230
  do {							\
    y3  = x0 | x1;					\
    y2  = x1 | x2;					\
    y2  = x0 ^ y2;					\
    y3 &= x3;						\
    y0  = x1 ^ x3;					\
    x3 |= y2;						\
    x0 &= x3;						\
    x1 &= x2;						\
    x2 ^= y3;						\
    y3 ^= y2;						\
    y2 |= x1;						\
    y1  = y3 & y0;					\
    y2 ^= y1;						\
    y1 ^= y0;						\
    y1 |= x1;						\
    y1 ^= x0;						\
    y0 &= x3;						\
    y0 ^= x2;						\
    y0  = ~y0;						\
231 232 233
  } while (0)

/* S5: 15  5  2 11  4 10  9 12  0  3 14  8 13  6  7  1 */
234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252
/* Original single-assignment form:
    t01 = x1  ^ x3;
    t02 = x1  | x3;
    t03 = x0  & t01;
    t04 = x2  ^ t02;
    t05 = t03 ^ t04;
    y0  =     ~ t05;
    t07 = x0  ^ t01;
    t08 = x3  | y0;
    t09 = x1  | t05;
    t10 = x3  ^ t08;
    t11 = x1  | t07;
    t12 = t03 | y0;
    t13 = t07 | t10;
    t14 = t01 ^ t11;
    y2  = t09 ^ t13;
    y1  = t07 ^ t08;
    y3  = t12 ^ t14;
*/
253
#define SBOX5(x0, x1, x2, x3, y0, y1, y2, y3)	\
254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271
  do {							\
    y0  = x1 | x3;					\
    y0 ^= x2;						\
    x2  = x1 ^ x3;					\
    y2  = x0 ^ x2;					\
    x0 &= x2;						\
    y0 ^= x0;						\
    y3  = x1 | y2;					\
    x1 |= y0;						\
    y0  = ~y0;						\
    x0 |= y0;						\
    y3 ^= x2;						\
    y3 ^= x0;						\
    y1  = x3 | y0;					\
    x3 ^= y1;						\
    y1 ^= y2;						\
    y2 |= x3;						\
    y2 ^= x1;						\
272 273 274
  } while (0)

/* S6:  7  2 12  5  8  4  6 11 14  9  1 15 13  3 10  0 */
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
/* Original single-assignment form:
    t01 = x0  & x3;
    t02 = x1  ^ x2;
    t03 = x0  ^ x3;
    t04 = t01 ^ t02;
    t05 = x1  | x2;
    y1  =     ~ t04;
    t07 = t03 & t05;
    t08 = x1  & y1;
    t09 = x0  | x2;
    t10 = t07 ^ t08;
    t11 = x1  | x3;
    t12 = x2  ^ t11;
    t13 = t09 ^ t10;
    y2  =     ~ t13;
    t15 = y1  & t03;
    y3  = t12 ^ t07;
    t17 = x0  ^ x1;
    t18 = y2  ^ t15;
    y0  = t17 ^ t18;
*/
#define SBOX6(x0, x1, x2, x3, y0, y1, y2, y3)	\
  do {							\
    y0  = x0 ^ x3;					\
    y1  = x0 & x3;					\
    y2  = x0 | x2;					\
    x3 |= x1;						\
    x3 ^= x2;						\
    x0 ^= x1;						\
    y3  = x1 | x2;					\
    x2 ^= x1;						\
    y3 &= y0;						\
    y1 ^= x2;						\
    y1  = ~y1;						\
    y0 &= y1;						\
    x1 &= y1;						\
    x1 ^= y3;						\
    y3 ^= x3;						\
    y2 ^= x1;						\
    y2  = ~y2;						\
    y0 ^= y2;						\
    y0 ^= x0;						\
317 318 319
  } while (0)

/* S7:  1 13 15  0 14  8  2 11  7  4 12 10  9  3  5  6 */
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
/* Original single-assignment form:
    t01 = x0  & x2;
    t02 =     ~ x3;
    t03 = x0  & t02;
    t04 = x1  | t01;
    t05 = x0  & x1;
    t06 = x2  ^ t04;
    y3  = t03 ^ t06;
    t08 = x2  | y3;
    t09 = x3  | t05;
    t10 = x0  ^ t08;
    t11 = t04 & y3;
    y1  = t09 ^ t10;
    t13 = x1  ^ y1;
    t14 = t01 ^ y1;
    t15 = x2  ^ t05;
    t16 = t11 | t13;
    t17 = t02 | t14;
    y0  = t15 ^ t17;
    y2  = x0  ^ t16;
*/
/* It appears impossible to do this with only 8 registers. We
   recompute t02, and t04 (if we have spare registers, hopefully the
   compiler can recognize it them as common subexpressions). */
#define SBOX7(x0, x1, x2, x3, y0, y1, y2, y3)	\
  do {							\
    y0  = x0 & x2;					\
    y3  = x1 | y0; /* t04 */				\
    y3 ^= x2;						\
    y1  = ~x3;     /* t02 */				\
    y1 &= x0;						\
    y3 ^= y1;						\
    y1  = x2 | y3;					\
    y1 ^= x0;						\
    y2  = x0 & x1;					\
    x2 ^= y2;						\
    y2 |= x3;						\
    y1 ^= y2;						\
    y2  = x1 | y0; /* t04 */				\
    y2 &= y3;						\
    x1 ^= y1;						\
    y2 |= x1;						\
    y2 ^= x0;						\
    y0 ^= y1;						\
    x3  = ~x3;     /* t02 */				\
    y0 |= x3;						\
    y0 ^= x2;						\
367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388
  } while (0)

/* In-place linear transformation.  */
#define LINEAR_TRANSFORMATION(x0,x1,x2,x3)		 \
  do {                                                   \
    x0 = ROL32 (x0, 13);                    \
    x2 = ROL32 (x2, 3);                     \
    x1 = x1 ^ x0 ^ x2;        \
    x3 = x3 ^ x2 ^ (x0 << 3); \
    x1 = ROL32 (x1, 1);                     \
    x3 = ROL32 (x3, 7);                     \
    x0 = x0 ^ x1 ^ x3;        \
    x2 = x2 ^ x3 ^ (x1 << 7); \
    x0 = ROL32 (x0, 5);                     \
    x2 = ROL32 (x2, 22);                    \
  } while (0)

/* Round inputs are x0,x1,x2,x3 (destroyed), and round outputs are
   y0,y1,y2,y3. */
#define ROUND(which, subkey, x0,x1,x2,x3, y0,y1,y2,y3) \
  do {						       \
    KEYXOR(x0,x1,x2,x3, subkey);		       \
389
    SBOX##which(x0,x1,x2,x3, y0,y1,y2,y3);	       \
390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411
    LINEAR_TRANSFORMATION(y0,y1,y2,y3);		       \
  } while (0)

#if HAVE_NATIVE_64_BIT

#define LINEAR_TRANSFORMATION64(x0,x1,x2,x3)		 \
  do {                                                   \
    x0 = ROL64 (x0, 13);                    \
    x2 = ROL64 (x2, 3);                     \
    x1 = x1 ^ x0 ^ x2;        \
    x3 = x3 ^ x2 ^ RSHIFT64(x0, 3);	    \
    x1 = ROL64 (x1, 1);                     \
    x3 = ROL64 (x3, 7);                     \
    x0 = x0 ^ x1 ^ x3;        \
    x2 = x2 ^ x3 ^ RSHIFT64(x1, 7);	    \
    x0 = ROL64 (x0, 5);                     \
    x2 = ROL64 (x2, 22);                    \
  } while (0)

#define ROUND64(which, subkey, x0,x1,x2,x3, y0,y1,y2,y3) \
  do {						       \
    KEYXOR64(x0,x1,x2,x3, subkey);		       \
412
    SBOX##which(x0,x1,x2,x3, y0,y1,y2,y3);	       \
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
    LINEAR_TRANSFORMATION64(y0,y1,y2,y3);		       \
  } while (0)

#endif /* HAVE_NATIVE_64_BIT */

void
serpent_encrypt (const struct serpent_ctx *ctx,
		 unsigned length, uint8_t * dst, const uint8_t * src)
{
  assert( !(length % SERPENT_BLOCK_SIZE));
  
#if HAVE_NATIVE_64_BIT
  if (length & SERPENT_BLOCK_SIZE)
#else
  while (length >= SERPENT_BLOCK_SIZE)
#endif
    {
      uint32_t x0,x1,x2,x3, y0,y1,y2,y3;
      unsigned k;

      x0 = LE_READ_UINT32 (src);
      x1 = LE_READ_UINT32 (src + 4);
      x2 = LE_READ_UINT32 (src + 8);
      x3 = LE_READ_UINT32 (src + 12);

      for (k = 0; ; k += 8)
	{
	  ROUND (0, ctx->keys[k+0], x0,x1,x2,x3, y0,y1,y2,y3);
	  ROUND (1, ctx->keys[k+1], y0,y1,y2,y3, x0,x1,x2,x3);
	  ROUND (2, ctx->keys[k+2], x0,x1,x2,x3, y0,y1,y2,y3);
	  ROUND (3, ctx->keys[k+3], y0,y1,y2,y3, x0,x1,x2,x3);
	  ROUND (4, ctx->keys[k+4], x0,x1,x2,x3, y0,y1,y2,y3);
	  ROUND (5, ctx->keys[k+5], y0,y1,y2,y3, x0,x1,x2,x3);
	  ROUND (6, ctx->keys[k+6], x0,x1,x2,x3, y0,y1,y2,y3);
	  if (k == 24)
	    break;
	  ROUND (7, ctx->keys[k+7], y0,y1,y2,y3, x0,x1,x2,x3);
	}

      /* Special final round, using two subkeys. */
      KEYXOR (y0,y1,y2,y3, ctx->keys[31]);
454
      SBOX7 (y0,y1,y2,y3, x0,x1,x2,x3);
455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497
      KEYXOR (x0,x1,x2,x3, ctx->keys[32]);
    
      LE_WRITE_UINT32 (dst, x0);
      LE_WRITE_UINT32 (dst + 4, x1);
      LE_WRITE_UINT32 (dst + 8, x2);
      LE_WRITE_UINT32 (dst + 12, x3);

      src += SERPENT_BLOCK_SIZE;
      dst += SERPENT_BLOCK_SIZE;
      length -= SERPENT_BLOCK_SIZE;
    }
#if HAVE_NATIVE_64_BIT
  FOR_BLOCKS(length, dst, src, 2*SERPENT_BLOCK_SIZE)
    {
      uint64_t x0,x1,x2,x3, y0,y1,y2,y3;
      unsigned k;

      x0 = LE_READ_UINT32 (src);
      x1 = LE_READ_UINT32 (src + 4);
      x2 = LE_READ_UINT32 (src + 8);
      x3 = LE_READ_UINT32 (src + 12);

      x0 <<= 32; x0 |= LE_READ_UINT32 (src + 16);
      x1 <<= 32; x1 |= LE_READ_UINT32 (src + 20);
      x2 <<= 32; x2 |= LE_READ_UINT32 (src + 24);
      x3 <<= 32; x3 |= LE_READ_UINT32 (src + 28);

      for (k = 0; ; k += 8)
	{
	  ROUND64 (0, ctx->keys[k+0], x0,x1,x2,x3, y0,y1,y2,y3);
	  ROUND64 (1, ctx->keys[k+1], y0,y1,y2,y3, x0,x1,x2,x3);
	  ROUND64 (2, ctx->keys[k+2], x0,x1,x2,x3, y0,y1,y2,y3);
	  ROUND64 (3, ctx->keys[k+3], y0,y1,y2,y3, x0,x1,x2,x3);
	  ROUND64 (4, ctx->keys[k+4], x0,x1,x2,x3, y0,y1,y2,y3);
	  ROUND64 (5, ctx->keys[k+5], y0,y1,y2,y3, x0,x1,x2,x3);
	  ROUND64 (6, ctx->keys[k+6], x0,x1,x2,x3, y0,y1,y2,y3);
	  if (k == 24)
	    break;
	  ROUND64 (7, ctx->keys[k+7], y0,y1,y2,y3, x0,x1,x2,x3);
	}

      /* Special final round, using two subkeys. */
      KEYXOR64 (y0,y1,y2,y3, ctx->keys[31]);
498
      SBOX7 (y0,y1,y2,y3, x0,x1,x2,x3);
499 500 501 502 503 504 505 506 507 508 509 510 511
      KEYXOR64 (x0,x1,x2,x3, ctx->keys[32]);
    
      LE_WRITE_UINT32 (dst + 16, x0);
      LE_WRITE_UINT32 (dst + 20, x1);
      LE_WRITE_UINT32 (dst + 24, x2);
      LE_WRITE_UINT32 (dst + 28, x3);
      x0 >>= 32; LE_WRITE_UINT32 (dst, x0);
      x1 >>= 32; LE_WRITE_UINT32 (dst + 4, x1);
      x2 >>= 32; LE_WRITE_UINT32 (dst + 8, x2);
      x3 >>= 32; LE_WRITE_UINT32 (dst + 12, x3);
    }
#endif /* HAVE_NATIVE_64_BIT */
}