src/md5.c

   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, 2002, 2003, 2004,
   4                  2005, 2006, 2007, 2010  Free Software Foundation, Inc.
   5
   6    This file is part of the GNU C Library.
   7
   8    The GNU C Library is free software; you can redistribute it and/or
   9    modify it under the terms of the GNU General Public License as
  10    published by the Free Software Foundation; either version 2 of the
  11    License, or (at your option) any later version.
  12
  13    The GNU C Library 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    General Public License for more details.
  17
  18    You should have received a copy of the GNU General Public License
  19    along with the GNU C Library; see the file COPYING.  If not, write to the
  20    Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
  21    Boston, MA 02110-1301, USA.  */
  22
  23 /* Written by Ulrich Drepper <drepper@gnu.ai.mit.edu>, 1995.  */
  24
  25 #ifdef HAVE_CONFIG_H
  26 # include <config.h>
  27 #endif
  28
  29 #include <sys/types.h>
  30
  31 #if STDC_HEADERS || defined _LIBC
  32 # include <stdlib.h>
  33 # include <string.h>
  34 #else
  35 # ifndef HAVE_MEMCPY
  36 #  define memcpy(d, s, n) bcopy ((s), (d), (n))
  37 # endif
  38 #endif
  39
  40 #ifdef _LIBC
  41 # include <endian.h>
  42 # if __BYTE_ORDER == __BIG_ENDIAN
  43 #  define WORDS_BIGENDIAN 1
  44 # endif
  45 /* We need to keep the namespace clean so define the MD5 function
  46    protected using leading __ .  */
  47 # define md5_init_ctx __md5_init_ctx
  48 # define md5_process_block __md5_process_block
  49 # define md5_process_bytes __md5_process_bytes
  50 # define md5_finish_ctx __md5_finish_ctx
  51 # define md5_read_ctx __md5_read_ctx
  52 # define md5_stream __md5_stream
  53 # define md5_buffer __md5_buffer
  54 #endif
  55
  56 #include "md5.h"
  57
  58 #ifdef WORDS_BIGENDIAN
  59 # define SWAP(n)                                                        \
  60     (((n) << 24) | (((n) & 0xff00) << 8) | (((n) >> 8) & 0xff00) | ((n) >> 24))
  61 #else
  62 # define SWAP(n) (n)
  63 #endif
  64
  65
  66 /* This array contains the bytes used to pad the buffer to the next
  67    64-byte boundary.  (RFC 1321, 3.1: Step 1)  */
  68 static const unsigned char fillbuf[64] = { 0x80, 0 /* , 0, 0, ...  */ };
  69
  70
  71 /* Initialize structure containing state of computation.
  72    (RFC 1321, 3.3: Step 3)  */
  73 void
  74 md5_init_ctx (struct md5_ctx *ctx)
  75 {
  76   ctx->A = 0x67452301;
  77   ctx->B = 0xefcdab89;
  78   ctx->C = 0x98badcfe;
  79   ctx->D = 0x10325476;
  80
  81   ctx->total[0] = ctx->total[1] = 0;
  82   ctx->buflen = 0;
  83 }
  84
  85 /* Put result from CTX in first 16 bytes following RESBUF.  The result
  86    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 bits value.  */
  90 void *
  91 md5_read_ctx (const struct md5_ctx *ctx, void *resbuf)
  92 {
  93   ((md5_uint32 *) resbuf)[0] = SWAP (ctx->A);
  94   ((md5_uint32 *) resbuf)[1] = SWAP (ctx->B);
  95   ((md5_uint32 *) resbuf)[2] = SWAP (ctx->C);
  96   ((md5_uint32 *) resbuf)[3] = SWAP (ctx->D);
  97
  98   return resbuf;
  99 }
 100
 101 /* Process the remaining bytes in the internal buffer and the usual
 102    prolog according to the standard and write the result to RESBUF.
 103
 104    IMPORTANT: On some systems it is required that RESBUF is correctly
 105    aligned for a 32 bits value.  */
 106 void *
 107 md5_finish_ctx (struct md5_ctx *ctx, void *resbuf)
 108 {
 109   /* Take yet unprocessed bytes into account.  */
 110   md5_uint32 bytes = ctx->buflen;
 111   size_t pad;
 112
 113   /* Now count remaining bytes.  */
 114   ctx->total[0] += bytes;
 115   if (ctx->total[0] < bytes)
 116     ++ctx->total[1];
 117
 118   pad = bytes >= 56 ? 64 + 56 - bytes : 56 - bytes;
 119   memcpy (&ctx->buffer[bytes], fillbuf, pad);
 120
 121   /* Put the 64-bit file length in *bits* at the end of the buffer.  */
 122   *(md5_uint32 *) &ctx->buffer[bytes + pad] = SWAP (ctx->total[0] << 3);
 123   *(md5_uint32 *) &ctx->buffer[bytes + pad + 4] = SWAP ((ctx->total[1] << 3) |
 124                                                         (ctx->total[0] >> 29));
 125
 126   /* Process last bytes.  */
 127   md5_process_block (ctx->buffer, bytes + pad + 8, ctx);
 128
 129   return md5_read_ctx (ctx, resbuf);
 130 }
 131
 132 /* Compute MD5 message digest for bytes read from STREAM.  The
 133    resulting message digest number will be written into the 16 bytes
 134    beginning at RESBLOCK.  */
 135 int
 136 md5_stream (FILE *stream, void *resblock)
 137 {
 138   /* Important: BLOCKSIZE must be a multiple of 64.  */
 139 #define BLOCKSIZE 4096
 140   struct md5_ctx ctx;
 141   char buffer[BLOCKSIZE + 72];
 142   size_t sum;
 143
 144   /* Initialize the computation context.  */
 145   md5_init_ctx (&ctx);
 146
 147   /* Iterate over full file contents.  */
 148   while (1)
 149     {
 150       /* We read the file in blocks of BLOCKSIZE bytes.  One call of the
 151          computation function processes the whole buffer so that with the
 152          next round of the loop another block can be read.  */
 153       size_t n;
 154       sum = 0;
 155
 156       /* Read block.  Take care for partial reads.  */
 157       do
 158         {
 159           n = fread (buffer + sum, 1, BLOCKSIZE - sum, stream);
 160
 161           sum += n;
 162         }
 163       while (sum < BLOCKSIZE && n != 0);
 164       if (n == 0 && ferror (stream))
 165         return 1;
 166
 167       /* If end of file is reached, end the loop.  */
 168       if (n == 0)
 169         break;
 170
 171       /* Process buffer with BLOCKSIZE bytes.  Note that
 172                         BLOCKSIZE % 64 == 0
 173        */
 174       md5_process_block (buffer, BLOCKSIZE, &ctx);
 175     }
 176
 177   /* Add the last bytes if necessary.  */
 178   if (sum > 0)
 179     md5_process_bytes (buffer, sum, &ctx);
 180
 181   /* Construct result in desired memory.  */
 182   md5_finish_ctx (&ctx, resblock);
 183   return 0;
 184 }
 185
 186 /* Compute MD5 message digest for LEN bytes beginning at BUFFER.  The
 187    result is always in little endian byte order, so that a byte-wise
 188    output yields to the wanted ASCII representation of the message
 189    digest.  */
 190 void *
 191 md5_buffer (const char *buffer, size_t len, void *resblock)
 192 {
 193   struct md5_ctx ctx;
 194
 195   /* Initialize the computation context.  */
 196   md5_init_ctx (&ctx);
 197
 198   /* Process whole buffer but last len % 64 bytes.  */
 199   md5_process_bytes (buffer, len, &ctx);
 200
 201   /* Put result in desired memory area.  */
 202   return md5_finish_ctx (&ctx, resblock);
 203 }
 204
 205
 206 void
 207 md5_process_bytes (const void *buffer, size_t len, struct md5_ctx *ctx)
 208 {
 209   /* const void aligned_buffer = buffer; */
 210
 211   /* When we already have some bits in our internal buffer concatenate
 212      both inputs first.  */
 213   if (ctx->buflen != 0)
 214     {
 215       size_t left_over = ctx->buflen;
 216       size_t add = 128 - left_over > len ? len : 128 - left_over;
 217
 218       /* Only put full words in the buffer.  */
 219       add -= add % __alignof__ (md5_uint32);
 220
 221       memcpy (&ctx->buffer[left_over], buffer, add);
 222       ctx->buflen += add;
 223
 224       if (ctx->buflen > 64)
 225         {
 226           md5_process_block (ctx->buffer, ctx->buflen & ~63, ctx);
 227
 228           ctx->buflen &= 63;
 229           /* The regions in the following copy operation cannot overlap.  */
 230           memcpy (ctx->buffer, &ctx->buffer[(left_over + add) & ~63],
 231                   ctx->buflen);
 232         }
 233
 234       buffer = (const char *) buffer + add;
 235       len -= add;
 236     }
 237
 238   /* Process available complete blocks.  */
 239   if (len > 64)
 240     {
 241       md5_process_block (buffer, len & ~63, ctx);
 242       buffer = (const char *) buffer + (len & ~63);
 243       len &= 63;
 244     }
 245
 246   /* Move remaining bytes in internal buffer.  */
 247   if (len > 0)
 248     {
 249       size_t left_over = ctx->buflen;
 250
 251       memcpy (&ctx->buffer[left_over], buffer, len);
 252       left_over += len;
 253       if (left_over >= 64)
 254         {
 255           md5_process_block (ctx->buffer, 64, ctx);
 256           left_over -= 64;
 257           memcpy (ctx->buffer, &ctx->buffer[64], left_over);
 258         }
 259       ctx->buflen = left_over;
 260     }
 261 }
 262
 263
 264 /* These are the four functions used in the four steps of the MD5 algorithm
 265    and defined in the RFC 1321.  The first function is a little bit optimized
 266    (as found in Colin Plumbs public domain implementation).  */
 267 /* #define FF(b, c, d) ((b & c) | (~b & d)) */
 268 #define FF(b, c, d) (d ^ (b & (c ^ d)))
 269 #define FG(b, c, d) FF (d, b, c)
 270 #define FH(b, c, d) (b ^ c ^ d)
 271 #define FI(b, c, d) (c ^ (b | ~d))
 272
 273 /* Process LEN bytes of BUFFER, accumulating context into CTX.
 274    It is assumed that LEN % 64 == 0.  */
 275
 276 void
 277 md5_process_block (const void *buffer, size_t len, struct md5_ctx *ctx)
 278 {
 279   md5_uint32 correct_words[16];
 280   const md5_uint32 *words = buffer;
 281   size_t nwords = len / sizeof (md5_uint32);
 282   const md5_uint32 *endp = words + nwords;
 283   md5_uint32 A = ctx->A;
 284   md5_uint32 B = ctx->B;
 285   md5_uint32 C = ctx->C;
 286   md5_uint32 D = ctx->D;
 287
 288   /* First increment the byte count.  RFC 1321 specifies the possible
 289      length of the file up to 2^64 bits.  Here we only compute the
 290      number of bytes.  Do a double word increment.  */
 291   ctx->total[0] += len;
 292   if (ctx->total[0] < len)
 293     ++ctx->total[1];
 294
 295   /* Process all bytes in the buffer with 64 bytes in each round of
 296      the loop.  */
 297   while (words < endp)
 298     {
 299       md5_uint32 *cwp = correct_words;
 300       md5_uint32 A_save = A;
 301       md5_uint32 B_save = B;
 302       md5_uint32 C_save = C;
 303       md5_uint32 D_save = D;
 304
 305       /* First round: using the given function, the context and a constant
 306          the next context is computed.  Because the algorithms processing
 307          unit is a 32-bit word and it is determined to work on words in
 308          little endian byte order we perhaps have to change the byte order
 309          before the computation.  To reduce the work for the next steps
 310          we store the swapped words in the array CORRECT_WORDS.  */
 311
 312 #define OP(a, b, c, d, s, T)                                            \
 313       do                                                                \
 314         {                                                               \
 315           a += FF (b, c, d) + (*cwp++ = SWAP (*words)) + T;             \
 316           ++words;                                                      \
 317           CYCLIC (a, s);                                                \
 318           a += b;                                                       \
 319         }                                                               \
 320       while (0)
 321
 322       /* It is unfortunate that C does not provide an operator for
 323          cyclic rotation.  Hope the C compiler is smart enough.  */
 324 #define CYCLIC(w, s) (w = (w << s) | (w >> (32 - s)))
 325
 326       /* Before we start, one word to the strange constants.
 327          They are defined in RFC 1321 as
 328
 329          T[i] = (int) (4294967296.0 * fabs (sin (i))), i=1..64
 330        */
 331
 332       /* Round 1.  */
 333       OP (A, B, C, D,  7, 0xd76aa478);
 334       OP (D, A, B, C, 12, 0xe8c7b756);
 335       OP (C, D, A, B, 17, 0x242070db);
 336       OP (B, C, D, A, 22, 0xc1bdceee);
 337       OP (A, B, C, D,  7, 0xf57c0faf);
 338       OP (D, A, B, C, 12, 0x4787c62a);
 339       OP (C, D, A, B, 17, 0xa8304613);
 340       OP (B, C, D, A, 22, 0xfd469501);
 341       OP (A, B, C, D,  7, 0x698098d8);
 342       OP (D, A, B, C, 12, 0x8b44f7af);
 343       OP (C, D, A, B, 17, 0xffff5bb1);
 344       OP (B, C, D, A, 22, 0x895cd7be);
 345       OP (A, B, C, D,  7, 0x6b901122);
 346       OP (D, A, B, C, 12, 0xfd987193);
 347       OP (C, D, A, B, 17, 0xa679438e);
 348       OP (B, C, D, A, 22, 0x49b40821);
 349
 350       /* For the second to fourth round we have the possibly swapped words
 351          in CORRECT_WORDS.  Redefine the macro to take an additional first
 352          argument specifying the function to use.  */
 353 #undef OP
 354 #define OP(f, a, b, c, d, k, s, T)                                      \
 355       do                                                                \
 356         {                                                               \
 357           a += f (b, c, d) + correct_words[k] + T;                      \
 358           CYCLIC (a, s);                                                \
 359           a += b;                                                       \
 360         }                                                               \
 361       while (0)
 362
 363       /* Round 2.  */
 364       OP (FG, A, B, C, D,  1,  5, 0xf61e2562);
 365       OP (FG, D, A, B, C,  6,  9, 0xc040b340);
 366       OP (FG, C, D, A, B, 11, 14, 0x265e5a51);
 367       OP (FG, B, C, D, A,  0, 20, 0xe9b6c7aa);
 368       OP (FG, A, B, C, D,  5,  5, 0xd62f105d);
 369       OP (FG, D, A, B, C, 10,  9, 0x02441453);
 370       OP (FG, C, D, A, B, 15, 14, 0xd8a1e681);
 371       OP (FG, B, C, D, A,  4, 20, 0xe7d3fbc8);
 372       OP (FG, A, B, C, D,  9,  5, 0x21e1cde6);
 373       OP (FG, D, A, B, C, 14,  9, 0xc33707d6);
 374       OP (FG, C, D, A, B,  3, 14, 0xf4d50d87);
 375       OP (FG, B, C, D, A,  8, 20, 0x455a14ed);
 376       OP (FG, A, B, C, D, 13,  5, 0xa9e3e905);
 377       OP (FG, D, A, B, C,  2,  9, 0xfcefa3f8);
 378       OP (FG, C, D, A, B,  7, 14, 0x676f02d9);
 379       OP (FG, B, C, D, A, 12, 20, 0x8d2a4c8a);
 380
 381       /* Round 3.  */
 382       OP (FH, A, B, C, D,  5,  4, 0xfffa3942);
 383       OP (FH, D, A, B, C,  8, 11, 0x8771f681);
 384       OP (FH, C, D, A, B, 11, 16, 0x6d9d6122);
 385       OP (FH, B, C, D, A, 14, 23, 0xfde5380c);
 386       OP (FH, A, B, C, D,  1,  4, 0xa4beea44);
 387       OP (FH, D, A, B, C,  4, 11, 0x4bdecfa9);
 388       OP (FH, C, D, A, B,  7, 16, 0xf6bb4b60);
 389       OP (FH, B, C, D, A, 10, 23, 0xbebfbc70);
 390       OP (FH, A, B, C, D, 13,  4, 0x289b7ec6);
 391       OP (FH, D, A, B, C,  0, 11, 0xeaa127fa);
 392       OP (FH, C, D, A, B,  3, 16, 0xd4ef3085);
 393       OP (FH, B, C, D, A,  6, 23, 0x04881d05);
 394       OP (FH, A, B, C, D,  9,  4, 0xd9d4d039);
 395       OP (FH, D, A, B, C, 12, 11, 0xe6db99e5);
 396       OP (FH, C, D, A, B, 15, 16, 0x1fa27cf8);
 397       OP (FH, B, C, D, A,  2, 23, 0xc4ac5665);
 398
 399       /* Round 4.  */
 400       OP (FI, A, B, C, D,  0,  6, 0xf4292244);
 401       OP (FI, D, A, B, C,  7, 10, 0x432aff97);
 402       OP (FI, C, D, A, B, 14, 15, 0xab9423a7);
 403       OP (FI, B, C, D, A,  5, 21, 0xfc93a039);
 404       OP (FI, A, B, C, D, 12,  6, 0x655b59c3);
 405       OP (FI, D, A, B, C,  3, 10, 0x8f0ccc92);
 406       OP (FI, C, D, A, B, 10, 15, 0xffeff47d);
 407       OP (FI, B, C, D, A,  1, 21, 0x85845dd1);
 408       OP (FI, A, B, C, D,  8,  6, 0x6fa87e4f);
 409       OP (FI, D, A, B, C, 15, 10, 0xfe2ce6e0);
 410       OP (FI, C, D, A, B,  6, 15, 0xa3014314);
 411       OP (FI, B, C, D, A, 13, 21, 0x4e0811a1);
 412       OP (FI, A, B, C, D,  4,  6, 0xf7537e82);
 413       OP (FI, D, A, B, C, 11, 10, 0xbd3af235);
 414       OP (FI, C, D, A, B,  2, 15, 0x2ad7d2bb);
 415       OP (FI, B, C, D, A,  9, 21, 0xeb86d391);
 416
 417       /* Add the starting values of the context.  */
 418       A += A_save;
 419       B += B_save;
 420       C += C_save;
 421       D += D_save;
 422     }
 423
 424   /* Put checksum in context given as argument.  */
 425   ctx->A = A;
 426   ctx->B = B;
 427   ctx->C = C;
 428   ctx->D = D;
 429 }
 430