Blame view

libmailutils/base/md5.c 13.6 KB
1
/* Functions to compute MD5 message digest of files or memory blocks.
2
   according to the definition of MD5 in RFC 1321 from April 1992.
3
   Copyright (C) 1995,1996,1997,1999,2000,2001,2005,2006,2010
4 5
	Free Software Foundation, Inc.
   This file is part of the GNU C Library.
6 7 8 9 10 11 12 13 14 15 16

   This program is free software; you can redistribute it and/or modify it
   under the terms of the GNU General Public License as published by the
   Free Software Foundation; either version 2, or (at your option) any
   later version.

   This program 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 General Public License for more details.

17
   You should have received a copy of the GNU General Public License
18
   along with this program.  If not, see <http://www.gnu.org/licenses/>. */
19 20 21 22

/* Written by Ulrich Drepper <drepper@gnu.ai.mit.edu>, 1995.  */

#ifdef HAVE_CONFIG_H
23 24 25
# include <config.h>
#endif

26
#include "mailutils/md5.h"
27 28

#include <stddef.h>
29
#include <stdlib.h>
30
#include <string.h>
31
#include <sys/types.h>
32 33 34 35 36 37 38 39 40 41

#if USE_UNLOCKED_IO
# include "unlocked-io.h"
#endif

#ifdef _LIBC
# include <endian.h>
# if __BYTE_ORDER == __BIG_ENDIAN
#  define WORDS_BIGENDIAN 1
# endif
42 43
#endif

44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62
/* We need to keep the namespace clean so define the MD5 function
   protected using leading __ .  */
# define md5_init_ctx __md5_init_ctx
# define md5_process_block __md5_process_block
# define md5_process_bytes __md5_process_bytes
# define md5_finish_ctx __md5_finish_ctx
# define md5_read_ctx __md5_read_ctx
# define md5_stream __md5_stream
# define md5_buffer __md5_buffer

#ifdef WORDS_BIGENDIAN
# define SWAP(n)							\
    (((n) << 24) | (((n) & 0xff00) << 8) | (((n) >> 8) & 0xff00) | ((n) >> 24))
#else
# define SWAP(n) (n)
#endif

#define BLOCKSIZE 4096
#if BLOCKSIZE % 64 != 0
63
# error "invalid BLOCKSIZE"
64
#endif
65

66 67 68
/* This array contains the bytes used to pad the buffer to the next
   64-byte boundary.  (RFC 1321, 3.1: Step 1)  */
static const unsigned char fillbuf[64] = { 0x80, 0 /* , 0, 0, ...  */ };
69

70 71 72

/* Initialize structure containing state of computation.
   (RFC 1321, 3.3: Step 3)  */
73
void
74
md5_init_ctx (struct md5_ctx *ctx)
75
{
76 77 78 79
  ctx->A = 0x67452301;
  ctx->B = 0xefcdab89;
  ctx->C = 0x98badcfe;
  ctx->D = 0x10325476;
80

81 82
  ctx->total[0] = ctx->total[1] = 0;
  ctx->buflen = 0;
83 84
}

85 86
/* Put result from CTX in first 16 bytes following RESBUF.  The result
   must be in little endian byte order.
87

88
   IMPORTANT: On some systems it is required that RESBUF is correctly
89
   aligned for a 32-bit value.  */
90 91
void *
md5_read_ctx (const struct md5_ctx *ctx, void *resbuf)
92
{
93 94 95 96
  ((uint32_t *) resbuf)[0] = SWAP (ctx->A);
  ((uint32_t *) resbuf)[1] = SWAP (ctx->B);
  ((uint32_t *) resbuf)[2] = SWAP (ctx->C);
  ((uint32_t *) resbuf)[3] = SWAP (ctx->D);
97

98
  return resbuf;
99 100
}

101 102 103 104
/* Process the remaining bytes in the internal buffer and the usual
   prolog according to the standard and write the result to RESBUF.

   IMPORTANT: On some systems it is required that RESBUF is correctly
105
   aligned for a 32-bit value.  */
106 107
void *
md5_finish_ctx (struct md5_ctx *ctx, void *resbuf)
108
{
109
  /* Take yet unprocessed bytes into account.  */
110 111
  uint32_t bytes = ctx->buflen;
  size_t size = (bytes < 56) ? 64 / 4 : 64 * 2 / 4;
112

113 114 115 116
  /* Now count remaining bytes.  */
  ctx->total[0] += bytes;
  if (ctx->total[0] < bytes)
    ++ctx->total[1];
117

118
  /* Put the 64-bit file length in *bits* at the end of the buffer.  */
119 120 121 122
  ctx->buffer[size - 2] = SWAP (ctx->total[0] << 3);
  ctx->buffer[size - 1] = SWAP ((ctx->total[1] << 3) | (ctx->total[0] >> 29));

  memcpy (&((char *) ctx->buffer)[bytes], fillbuf, (size - 2) * 4 - bytes);
123

124
  /* Process last bytes.  */
125
  md5_process_block (ctx->buffer, size * 4, ctx);
126

127
  return md5_read_ctx (ctx, resbuf);
128 129
}

130 131 132 133 134
/* Compute MD5 message digest for bytes read from STREAM.  The
   resulting message digest number will be written into the 16 bytes
   beginning at RESBLOCK.  */
int
md5_stream (FILE *stream, void *resblock)
135
{
136 137 138 139 140 141 142 143 144 145 146
  struct md5_ctx ctx;
  char buffer[BLOCKSIZE + 72];
  size_t sum;

  /* Initialize the computation context.  */
  md5_init_ctx (&ctx);

  /* Iterate over full file contents.  */
  while (1)
    {
      /* We read the file in blocks of BLOCKSIZE bytes.  One call of the
147 148
         computation function processes the whole buffer so that with the
         next round of the loop another block can be read.  */
149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164
      size_t n;
      sum = 0;

      /* Read block.  Take care for partial reads.  */
      while (1)
	{
	  n = fread (buffer + sum, 1, BLOCKSIZE - sum, stream);

	  sum += n;

	  if (sum == BLOCKSIZE)
	    break;

	  if (n == 0)
	    {
	      /* Check for the error flag IFF N == 0, so that we don't
165 166
	         exit the loop after a partial read due to e.g., EAGAIN
	         or EWOULDBLOCK.  */
167 168 169 170 171 172 173 174 175 176 177 178 179
	      if (ferror (stream))
		return 1;
	      goto process_partial_block;
	    }

	  /* We've read at least one byte, so ignore errors.  But always
	     check for EOF, since feof may be true even though N > 0.
	     Otherwise, we could end up calling fread after EOF.  */
	  if (feof (stream))
	    goto process_partial_block;
	}

      /* Process buffer with BLOCKSIZE bytes.  Note that
180
         BLOCKSIZE % 64 == 0
181 182
       */
      md5_process_block (buffer, BLOCKSIZE, &ctx);
183 184
    }

185
process_partial_block:
186 187 188 189 190 191 192 193

  /* Process any remaining bytes.  */
  if (sum > 0)
    md5_process_bytes (buffer, sum, &ctx);

  /* Construct result in desired memory.  */
  md5_finish_ctx (&ctx, resblock);
  return 0;
194 195
}

196 197 198 199 200 201 202 203
/* Compute MD5 message digest for LEN bytes beginning at BUFFER.  The
   result is always in little endian byte order, so that a byte-wise
   output yields to the wanted ASCII representation of the message
   digest.  */
void *
md5_buffer (const char *buffer, size_t len, void *resblock)
{
  struct md5_ctx ctx;
204

205 206
  /* Initialize the computation context.  */
  md5_init_ctx (&ctx);
207

208 209 210 211 212 213
  /* Process whole buffer but last len % 64 bytes.  */
  md5_process_bytes (buffer, len, &ctx);

  /* Put result in desired memory area.  */
  return md5_finish_ctx (&ctx, resblock);
}
214 215


216 217
void
md5_process_bytes (const void *buffer, size_t len, struct md5_ctx *ctx)
218
{
219 220 221 222 223 224 225
  /* When we already have some bits in our internal buffer concatenate
     both inputs first.  */
  if (ctx->buflen != 0)
    {
      size_t left_over = ctx->buflen;
      size_t add = 128 - left_over > len ? len : 128 - left_over;

226
      memcpy (&((char *) ctx->buffer)[left_over], buffer, add);
227 228 229 230 231 232 233 234
      ctx->buflen += add;

      if (ctx->buflen > 64)
	{
	  md5_process_block (ctx->buffer, ctx->buflen & ~63, ctx);

	  ctx->buflen &= 63;
	  /* The regions in the following copy operation cannot overlap.  */
235 236
	  memcpy (ctx->buffer,
		  &((char *) ctx->buffer)[(left_over + add) & ~63],
237 238 239 240 241 242
		  ctx->buflen);
	}

      buffer = (const char *) buffer + add;
      len -= add;
    }
243

244 245 246 247
  /* Process available complete blocks.  */
  if (len >= 64)
    {
#if !_STRING_ARCH_unaligned
248 249
# define alignof(type) offsetof (struct { char c; type x; }, x)
# define UNALIGNED_P(p) (((size_t) p) % alignof (uint32_t) != 0)
250 251 252 253 254 255 256 257
      if (UNALIGNED_P (buffer))
	while (len > 64)
	  {
	    md5_process_block (memcpy (ctx->buffer, buffer, 64), 64, ctx);
	    buffer = (const char *) buffer + 64;
	    len -= 64;
	  }
      else
258
#endif
259 260 261 262 263 264 265 266 267 268 269 270
	{
	  md5_process_block (buffer, len & ~63, ctx);
	  buffer = (const char *) buffer + (len & ~63);
	  len &= 63;
	}
    }

  /* Move remaining bytes in internal buffer.  */
  if (len > 0)
    {
      size_t left_over = ctx->buflen;

271
      memcpy (&((char *) ctx->buffer)[left_over], buffer, len);
272 273 274 275 276
      left_over += len;
      if (left_over >= 64)
	{
	  md5_process_block (ctx->buffer, 64, ctx);
	  left_over -= 64;
277
	  memcpy (ctx->buffer, &ctx->buffer[16], left_over);
278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294
	}
      ctx->buflen = left_over;
    }
}


/* These are the four functions used in the four steps of the MD5 algorithm
   and defined in the RFC 1321.  The first function is a little bit optimized
   (as found in Colin Plumbs public domain implementation).  */
/* #define FF(b, c, d) ((b & c) | (~b & d)) */
#define FF(b, c, d) (d ^ (b & (c ^ d)))
#define FG(b, c, d) FF (d, b, c)
#define FH(b, c, d) (b ^ c ^ d)
#define FI(b, c, d) (c ^ (b | ~d))

/* Process LEN bytes of BUFFER, accumulating context into CTX.
   It is assumed that LEN % 64 == 0.  */
295 296

void
297
md5_process_block (const void *buffer, size_t len, struct md5_ctx *ctx)
298
{
299 300 301 302 303 304 305 306
  uint32_t correct_words[16];
  const uint32_t *words = buffer;
  size_t nwords = len / sizeof (uint32_t);
  const uint32_t *endp = words + nwords;
  uint32_t A = ctx->A;
  uint32_t B = ctx->B;
  uint32_t C = ctx->C;
  uint32_t D = ctx->D;
307 308 309 310 311 312 313 314 315 316 317 318

  /* First increment the byte count.  RFC 1321 specifies the possible
     length of the file up to 2^64 bits.  Here we only compute the
     number of bytes.  Do a double word increment.  */
  ctx->total[0] += len;
  if (ctx->total[0] < len)
    ++ctx->total[1];

  /* Process all bytes in the buffer with 64 bytes in each round of
     the loop.  */
  while (words < endp)
    {
319 320 321 322 323
      uint32_t *cwp = correct_words;
      uint32_t A_save = A;
      uint32_t B_save = B;
      uint32_t C_save = C;
      uint32_t D_save = D;
324 325

      /* First round: using the given function, the context and a constant
326 327 328 329 330
         the next context is computed.  Because the algorithms processing
         unit is a 32-bit word and it is determined to work on words in
         little endian byte order we perhaps have to change the byte order
         before the computation.  To reduce the work for the next steps
         we store the swapped words in the array CORRECT_WORDS.  */
331 332 333 334 335 336

#define OP(a, b, c, d, s, T)						\
      do								\
        {								\
	  a += FF (b, c, d) + (*cwp++ = SWAP (*words)) + T;		\
	  ++words;							\
337
	  CYCLIC (a, s);						\
338 339 340 341
	  a += b;							\
        }								\
      while (0)

342 343 344 345
      /* It is unfortunate that C does not provide an operator for
         cyclic rotation.  Hope the C compiler is smart enough.  */
#define CYCLIC(w, s) (w = (w << s) | (w >> (32 - s)))

346
      /* Before we start, one word to the strange constants.
347
         They are defined in RFC 1321 as
348

349 350 351 352 353
         T[i] = (int) (4294967296.0 * fabs (sin (i))), i=1..64

         Here is an equivalent invocation using Perl:

         perl -e 'foreach(1..64){printf "0x%08x\n", int (4294967296 * abs (sin $_))}'
354 355 356
       */

      /* Round 1.  */
357
      OP (A, B, C, D, 7, 0xd76aa478);
358 359 360
      OP (D, A, B, C, 12, 0xe8c7b756);
      OP (C, D, A, B, 17, 0x242070db);
      OP (B, C, D, A, 22, 0xc1bdceee);
361
      OP (A, B, C, D, 7, 0xf57c0faf);
362 363 364
      OP (D, A, B, C, 12, 0x4787c62a);
      OP (C, D, A, B, 17, 0xa8304613);
      OP (B, C, D, A, 22, 0xfd469501);
365
      OP (A, B, C, D, 7, 0x698098d8);
366 367 368
      OP (D, A, B, C, 12, 0x8b44f7af);
      OP (C, D, A, B, 17, 0xffff5bb1);
      OP (B, C, D, A, 22, 0x895cd7be);
369
      OP (A, B, C, D, 7, 0x6b901122);
370 371 372 373 374
      OP (D, A, B, C, 12, 0xfd987193);
      OP (C, D, A, B, 17, 0xa679438e);
      OP (B, C, D, A, 22, 0x49b40821);

      /* For the second to fourth round we have the possibly swapped words
375 376
         in CORRECT_WORDS.  Redefine the macro to take an additional first
         argument specifying the function to use.  */
377 378
#undef OP
#define OP(f, a, b, c, d, k, s, T)					\
379
      do								\
380 381
	{								\
	  a += f (b, c, d) + correct_words[k] + T;			\
382
	  CYCLIC (a, s);						\
383 384 385 386 387
	  a += b;							\
	}								\
      while (0)

      /* Round 2.  */
388 389
      OP (FG, A, B, C, D, 1, 5, 0xf61e2562);
      OP (FG, D, A, B, C, 6, 9, 0xc040b340);
390
      OP (FG, C, D, A, B, 11, 14, 0x265e5a51);
391 392 393
      OP (FG, B, C, D, A, 0, 20, 0xe9b6c7aa);
      OP (FG, A, B, C, D, 5, 5, 0xd62f105d);
      OP (FG, D, A, B, C, 10, 9, 0x02441453);
394
      OP (FG, C, D, A, B, 15, 14, 0xd8a1e681);
395 396 397 398 399 400 401 402
      OP (FG, B, C, D, A, 4, 20, 0xe7d3fbc8);
      OP (FG, A, B, C, D, 9, 5, 0x21e1cde6);
      OP (FG, D, A, B, C, 14, 9, 0xc33707d6);
      OP (FG, C, D, A, B, 3, 14, 0xf4d50d87);
      OP (FG, B, C, D, A, 8, 20, 0x455a14ed);
      OP (FG, A, B, C, D, 13, 5, 0xa9e3e905);
      OP (FG, D, A, B, C, 2, 9, 0xfcefa3f8);
      OP (FG, C, D, A, B, 7, 14, 0x676f02d9);
403 404 405
      OP (FG, B, C, D, A, 12, 20, 0x8d2a4c8a);

      /* Round 3.  */
406 407
      OP (FH, A, B, C, D, 5, 4, 0xfffa3942);
      OP (FH, D, A, B, C, 8, 11, 0x8771f681);
408 409
      OP (FH, C, D, A, B, 11, 16, 0x6d9d6122);
      OP (FH, B, C, D, A, 14, 23, 0xfde5380c);
410 411 412
      OP (FH, A, B, C, D, 1, 4, 0xa4beea44);
      OP (FH, D, A, B, C, 4, 11, 0x4bdecfa9);
      OP (FH, C, D, A, B, 7, 16, 0xf6bb4b60);
413
      OP (FH, B, C, D, A, 10, 23, 0xbebfbc70);
414 415 416 417 418
      OP (FH, A, B, C, D, 13, 4, 0x289b7ec6);
      OP (FH, D, A, B, C, 0, 11, 0xeaa127fa);
      OP (FH, C, D, A, B, 3, 16, 0xd4ef3085);
      OP (FH, B, C, D, A, 6, 23, 0x04881d05);
      OP (FH, A, B, C, D, 9, 4, 0xd9d4d039);
419 420
      OP (FH, D, A, B, C, 12, 11, 0xe6db99e5);
      OP (FH, C, D, A, B, 15, 16, 0x1fa27cf8);
421
      OP (FH, B, C, D, A, 2, 23, 0xc4ac5665);
422 423

      /* Round 4.  */
424 425
      OP (FI, A, B, C, D, 0, 6, 0xf4292244);
      OP (FI, D, A, B, C, 7, 10, 0x432aff97);
426
      OP (FI, C, D, A, B, 14, 15, 0xab9423a7);
427 428 429
      OP (FI, B, C, D, A, 5, 21, 0xfc93a039);
      OP (FI, A, B, C, D, 12, 6, 0x655b59c3);
      OP (FI, D, A, B, C, 3, 10, 0x8f0ccc92);
430
      OP (FI, C, D, A, B, 10, 15, 0xffeff47d);
431 432
      OP (FI, B, C, D, A, 1, 21, 0x85845dd1);
      OP (FI, A, B, C, D, 8, 6, 0x6fa87e4f);
433
      OP (FI, D, A, B, C, 15, 10, 0xfe2ce6e0);
434
      OP (FI, C, D, A, B, 6, 15, 0xa3014314);
435
      OP (FI, B, C, D, A, 13, 21, 0x4e0811a1);
436
      OP (FI, A, B, C, D, 4, 6, 0xf7537e82);
437
      OP (FI, D, A, B, C, 11, 10, 0xbd3af235);
438 439
      OP (FI, C, D, A, B, 2, 15, 0x2ad7d2bb);
      OP (FI, B, C, D, A, 9, 21, 0xeb86d391);
440 441 442 443 444 445 446

      /* Add the starting values of the context.  */
      A += A_save;
      B += B_save;
      C += C_save;
      D += D_save;
    }
447

448 449 450 451 452 453
  /* Put checksum in context given as argument.  */
  ctx->A = A;
  ctx->B = B;
  ctx->C = C;
  ctx->D = D;
}