lib/md4.c

   1 /* Functions to compute MD4 message digest of files or memory blocks.
   2    according to the definition of MD4 in RFC 1320 from April 1992.
   3    Copyright (C) 1995-1997, 1999-2003, 2005-2006, 2008-2018 Free Software
   4    Foundation, Inc.
   5
   6    This program is free software; you can redistribute it and/or modify it
   7    under the terms of the GNU General Public License as published by the
   8    Free Software Foundation; either version 2, or (at your option) any
   9    later version.
  10
  11    This program is distributed in the hope that it will be useful,
  12    but WITHOUT ANY WARRANTY; without even the implied warranty of
  13    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14    GNU General Public License for more details.
  15
  16    You should have received a copy of the GNU General Public License
  17    along with this program; if not, see <https://www.gnu.org/licenses/>.  */
  18
  19 /* Adapted by Simon Josefsson from gnulib md5.? and Libgcrypt
  20    cipher/md4.c . */
  21
  22 #include <config.h>
  23
  24 #include "md4.h"
  25
  26 #include <stdalign.h>
  27 #include <stdint.h>
  28 #include <stdlib.h>
  29 #include <string.h>
  30 #include <sys/types.h>
  31
  32 #if USE_UNLOCKED_IO
  33 # include "unlocked-io.h"
  34 #endif
  35
  36 #include <byteswap.h>
  37 #ifdef WORDS_BIGENDIAN
  38 # define SWAP(n) bswap_32 (n)
  39 #else
  40 # define SWAP(n) (n)
  41 #endif
  42
  43 #define BLOCKSIZE 32768
  44 #if BLOCKSIZE % 64 != 0
  45 # error "invalid BLOCKSIZE"
  46 #endif
  47
  48 /* This array contains the bytes used to pad the buffer to the next
  49    64-byte boundary.  (RFC 1320, 3.1: Step 1)  */
  50 static const unsigned char fillbuf[64] = { 0x80, 0 /* , 0, 0, ...  */  };
  51
  52
  53 /* Initialize structure containing state of computation.
  54    (RFC 1320, 3.3: Step 3)  */
  55 void
  56 md4_init_ctx (struct md4_ctx *ctx)
  57 {
  58   ctx->A = 0x67452301;
  59   ctx->B = 0xefcdab89;
  60   ctx->C = 0x98badcfe;
  61   ctx->D = 0x10325476;
  62
  63   ctx->total[0] = ctx->total[1] = 0;
  64   ctx->buflen = 0;
  65 }
  66
  67 /* Copy the 4 byte value from v into the memory location pointed to by *cp,
  68    If your architecture allows unaligned access this is equivalent to
  69    * (uint32_t *) cp = v  */
  70 static void
  71 set_uint32 (char *cp, uint32_t v)
  72 {
  73   memcpy (cp, &v, sizeof v);
  74 }
  75
  76 /* Put result from CTX in first 16 bytes following RESBUF.  The result
  77    must be in little endian byte order.  */
  78 void *
  79 md4_read_ctx (const struct md4_ctx *ctx, void *resbuf)
  80 {
  81   char *r = resbuf;
  82   set_uint32 (r + 0 * sizeof ctx->A, SWAP (ctx->A));
  83   set_uint32 (r + 1 * sizeof ctx->B, SWAP (ctx->B));
  84   set_uint32 (r + 2 * sizeof ctx->C, SWAP (ctx->C));
  85   set_uint32 (r + 3 * sizeof ctx->D, SWAP (ctx->D));
  86
  87   return resbuf;
  88 }
  89
  90 /* Process the remaining bytes in the internal buffer and the usual
  91    prolog according to the standard and write the result to RESBUF.  */
  92 void *
  93 md4_finish_ctx (struct md4_ctx *ctx, void *resbuf)
  94 {
  95   /* Take yet unprocessed bytes into account.  */
  96   uint32_t bytes = ctx->buflen;
  97   size_t pad;
  98
  99   /* Now count remaining bytes.  */
 100   ctx->total[0] += bytes;
 101   if (ctx->total[0] < bytes)
 102     ++ctx->total[1];
 103
 104   pad = bytes >= 56 ? 64 + 56 - bytes : 56 - bytes;
 105   memcpy (&((char*)ctx->buffer)[bytes], fillbuf, pad);
 106
 107   /* Put the 64-bit file length in *bits* at the end of the buffer.  */
 108   ctx->buffer[(bytes + pad) / 4] = SWAP (ctx->total[0] << 3);
 109   ctx->buffer[(bytes + pad) / 4 + 1] = SWAP ((ctx->total[1] << 3) |
 110                                              (ctx->total[0] >> 29));
 111
 112   /* Process last bytes.  */
 113   md4_process_block (ctx->buffer, bytes + pad + 8, ctx);
 114
 115   return md4_read_ctx (ctx, resbuf);
 116 }
 117
 118 /* Compute MD4 message digest for bytes read from STREAM.  The
 119    resulting message digest number will be written into the 16 bytes
 120    beginning at RESBLOCK.  */
 121 int
 122 md4_stream (FILE * stream, void *resblock)
 123 {
 124   struct md4_ctx ctx;
 125   size_t sum;
 126
 127   char *buffer = malloc (BLOCKSIZE + 72);
 128   if (!buffer)
 129     return 1;
 130
 131   /* Initialize the computation context.  */
 132   md4_init_ctx (&ctx);
 133
 134   /* Iterate over full file contents.  */
 135   while (1)
 136     {
 137       /* We read the file in blocks of BLOCKSIZE bytes.  One call of the
 138          computation function processes the whole buffer so that with the
 139          next round of the loop another block can be read.  */
 140       size_t n;
 141       sum = 0;
 142
 143       /* Read block.  Take care for partial reads.  */
 144       while (1)
 145         {
 146           n = fread (buffer + sum, 1, BLOCKSIZE - sum, stream);
 147
 148           sum += n;
 149
 150           if (sum == BLOCKSIZE)
 151             break;
 152
 153           if (n == 0)
 154             {
 155               /* Check for the error flag IFF N == 0, so that we don't
 156                  exit the loop after a partial read due to e.g., EAGAIN
 157                  or EWOULDBLOCK.  */
 158               if (ferror (stream))
 159                 {
 160                   free (buffer);
 161                   return 1;
 162                 }
 163               goto process_partial_block;
 164             }
 165
 166           /* We've read at least one byte, so ignore errors.  But always
 167              check for EOF, since feof may be true even though N > 0.
 168              Otherwise, we could end up calling fread after EOF.  */
 169           if (feof (stream))
 170             goto process_partial_block;
 171         }
 172
 173       /* Process buffer with BLOCKSIZE bytes.  Note that
 174          BLOCKSIZE % 64 == 0
 175        */
 176       md4_process_block (buffer, BLOCKSIZE, &ctx);
 177     }
 178
 179 process_partial_block:;
 180
 181   /* Process any remaining bytes.  */
 182   if (sum > 0)
 183     md4_process_bytes (buffer, sum, &ctx);
 184
 185   /* Construct result in desired memory.  */
 186   md4_finish_ctx (&ctx, resblock);
 187   free (buffer);
 188   return 0;
 189 }
 190
 191 /* Compute MD4 message digest for LEN bytes beginning at BUFFER.  The
 192    result is always in little endian byte order, so that a byte-wise
 193    output yields to the wanted ASCII representation of the message
 194    digest.  */
 195 void *
 196 md4_buffer (const char *buffer, size_t len, void *resblock)
 197 {
 198   struct md4_ctx ctx;
 199
 200   /* Initialize the computation context.  */
 201   md4_init_ctx (&ctx);
 202
 203   /* Process whole buffer but last len % 64 bytes.  */
 204   md4_process_bytes (buffer, len, &ctx);
 205
 206   /* Put result in desired memory area.  */
 207   return md4_finish_ctx (&ctx, resblock);
 208 }
 209
 210 void
 211 md4_process_bytes (const void *buffer, size_t len, struct md4_ctx *ctx)
 212 {
 213   /* When we already have some bits in our internal buffer concatenate
 214      both inputs first.  */
 215   if (ctx->buflen != 0)
 216     {
 217       size_t left_over = ctx->buflen;
 218       size_t add = 128 - left_over > len ? len : 128 - left_over;
 219
 220       memcpy (&((char*)ctx->buffer)[left_over], buffer, add);
 221       ctx->buflen += add;
 222
 223       if (ctx->buflen > 64)
 224         {
 225           md4_process_block (ctx->buffer, ctx->buflen & ~63, ctx);
 226
 227           ctx->buflen &= 63;
 228           /* The regions in the following copy operation cannot overlap.  */
 229           memcpy (ctx->buffer, &((char*)ctx->buffer)[(left_over + add) & ~63],
 230                   ctx->buflen);
 231         }
 232
 233       buffer = (const char *) buffer + add;
 234       len -= add;
 235     }
 236
 237   /* Process available complete blocks.  */
 238   if (len >= 64)
 239     {
 240 #if !(_STRING_ARCH_unaligned || _STRING_INLINE_unaligned)
 241 # define UNALIGNED_P(p) ((uintptr_t) (p) % alignof (uint32_t) != 0)
 242       if (UNALIGNED_P (buffer))
 243         while (len > 64)
 244           {
 245             md4_process_block (memcpy (ctx->buffer, buffer, 64), 64, ctx);
 246             buffer = (const char *) buffer + 64;
 247             len -= 64;
 248           }
 249       else
 250 #endif
 251         {
 252           md4_process_block (buffer, len & ~63, ctx);
 253           buffer = (const char *) buffer + (len & ~63);
 254           len &= 63;
 255         }
 256     }
 257
 258   /* Move remaining bytes in internal buffer.  */
 259   if (len > 0)
 260     {
 261       size_t left_over = ctx->buflen;
 262
 263       memcpy (&((char*)ctx->buffer)[left_over], buffer, len);
 264       left_over += len;
 265       if (left_over >= 64)
 266         {
 267           md4_process_block (ctx->buffer, 64, ctx);
 268           left_over -= 64;
 269           memcpy (ctx->buffer, &ctx->buffer[16], left_over);
 270         }
 271       ctx->buflen = left_over;
 272     }
 273 }
 274
 275 /* --- Code below is the primary difference between md5.c and md4.c --- */
 276
 277 /* MD4 round constants */
 278 #define K1 0x5a827999
 279 #define K2 0x6ed9eba1
 280
 281 /* Round functions.  */
 282 #define F(x, y, z) ((z) ^ ((x) & ((y) ^ (z))))
 283 #define G(x, y, z) (((x) & (y)) | ((x) & (z)) | ((y) & (z)))
 284 #define H(x, y, z) ((x) ^ (y) ^ (z))
 285 #define rol(x, n) (((x) << (n)) | ((uint32_t) (x) >> (32 - (n))))
 286 #define R1(a,b,c,d,k,s) a=rol(a+F(b,c,d)+x[k],s);
 287 #define R2(a,b,c,d,k,s) a=rol(a+G(b,c,d)+x[k]+K1,s);
 288 #define R3(a,b,c,d,k,s) a=rol(a+H(b,c,d)+x[k]+K2,s);
 289
 290 /* Process LEN bytes of BUFFER, accumulating context into CTX.
 291    It is assumed that LEN % 64 == 0.  */
 292
 293 void
 294 md4_process_block (const void *buffer, size_t len, struct md4_ctx *ctx)
 295 {
 296   const uint32_t *words = buffer;
 297   size_t nwords = len / sizeof (uint32_t);
 298   const uint32_t *endp = words + nwords;
 299   uint32_t x[16];
 300   uint32_t A = ctx->A;
 301   uint32_t B = ctx->B;
 302   uint32_t C = ctx->C;
 303   uint32_t D = ctx->D;
 304   uint32_t lolen = len;
 305
 306   /* First increment the byte count.  RFC 1320 specifies the possible
 307      length of the file up to 2^64 bits.  Here we only compute the
 308      number of bytes.  Do a double word increment.  */
 309   ctx->total[0] += lolen;
 310   ctx->total[1] += (len >> 31 >> 1) + (ctx->total[0] < lolen);
 311
 312   /* Process all bytes in the buffer with 64 bytes in each round of
 313      the loop.  */
 314   while (words < endp)
 315     {
 316       int t;
 317       for (t = 0; t < 16; t++)
 318         {
 319           x[t] = SWAP (*words);
 320           words++;
 321         }
 322
 323       /* Round 1.  */
 324       R1 (A, B, C, D, 0, 3);
 325       R1 (D, A, B, C, 1, 7);
 326       R1 (C, D, A, B, 2, 11);
 327       R1 (B, C, D, A, 3, 19);
 328       R1 (A, B, C, D, 4, 3);
 329       R1 (D, A, B, C, 5, 7);
 330       R1 (C, D, A, B, 6, 11);
 331       R1 (B, C, D, A, 7, 19);
 332       R1 (A, B, C, D, 8, 3);
 333       R1 (D, A, B, C, 9, 7);
 334       R1 (C, D, A, B, 10, 11);
 335       R1 (B, C, D, A, 11, 19);
 336       R1 (A, B, C, D, 12, 3);
 337       R1 (D, A, B, C, 13, 7);
 338       R1 (C, D, A, B, 14, 11);
 339       R1 (B, C, D, A, 15, 19);
 340
 341       /* Round 2.  */
 342       R2 (A, B, C, D, 0, 3);
 343       R2 (D, A, B, C, 4, 5);
 344       R2 (C, D, A, B, 8, 9);
 345       R2 (B, C, D, A, 12, 13);
 346       R2 (A, B, C, D, 1, 3);
 347       R2 (D, A, B, C, 5, 5);
 348       R2 (C, D, A, B, 9, 9);
 349       R2 (B, C, D, A, 13, 13);
 350       R2 (A, B, C, D, 2, 3);
 351       R2 (D, A, B, C, 6, 5);
 352       R2 (C, D, A, B, 10, 9);
 353       R2 (B, C, D, A, 14, 13);
 354       R2 (A, B, C, D, 3, 3);
 355       R2 (D, A, B, C, 7, 5);
 356       R2 (C, D, A, B, 11, 9);
 357       R2 (B, C, D, A, 15, 13);
 358
 359       /* Round 3.  */
 360       R3 (A, B, C, D, 0, 3);
 361       R3 (D, A, B, C, 8, 9);
 362       R3 (C, D, A, B, 4, 11);
 363       R3 (B, C, D, A, 12, 15);
 364       R3 (A, B, C, D, 2, 3);
 365       R3 (D, A, B, C, 10, 9);
 366       R3 (C, D, A, B, 6, 11);
 367       R3 (B, C, D, A, 14, 15);
 368       R3 (A, B, C, D, 1, 3);
 369       R3 (D, A, B, C, 9, 9);
 370       R3 (C, D, A, B, 5, 11);
 371       R3 (B, C, D, A, 13, 15);
 372       R3 (A, B, C, D, 3, 3);
 373       R3 (D, A, B, C, 11, 9);
 374       R3 (C, D, A, B, 7, 11);
 375       R3 (B, C, D, A, 15, 15);
 376
 377       A = ctx->A += A;
 378       B = ctx->B += B;
 379       C = ctx->C += C;
 380       D = ctx->D += D;
 381     }
 382 }