src/md5sum.c

   1 /* Compute MD5, SHA1, SHA224, SHA256, SHA384 or SHA512 checksum of files or strings
   2    Copyright (C) 1995-2006 Free Software Foundation, Inc.
   3
   4    This program is free software; you can redistribute it and/or modify
   5    it under the terms of the GNU General Public License as published by
   6    the Free Software Foundation; either version 2, or (at your option)
   7    any later version.
   8
   9    This program is distributed in the hope that it will be useful,
  10    but WITHOUT ANY WARRANTY; without even the implied warranty of
  11    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12    GNU General Public License for more details.
  13
  14    You should have received a copy of the GNU General Public License
  15    along with this program; if not, write to the Free Software Foundation,
  16    Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.  */
  17
  18 /* Written by Ulrich Drepper <drepper@gnu.ai.mit.edu>.  */
  19
  20 #include <config.h>
  21
  22 #include <getopt.h>
  23 #include <sys/types.h>
  24
  25 #include "system.h"
  26
  27 #if HASH_ALGO_MD5
  28 # include "md5.h"
  29 #endif
  30 #if HASH_ALGO_SHA1
  31 # include "sha1.h"
  32 #endif
  33 #if HASH_ALGO_SHA256 || HASH_ALGO_SHA224
  34 # include "sha256.h"
  35 #endif
  36 #if HASH_ALGO_SHA512 || HASH_ALGO_SHA384
  37 # include "sha512.h"
  38 #endif
  39 #include "getline.h"
  40 #include "error.h"
  41 #include "quote.h"
  42 #include "stdio--.h"
  43
  44 /* The official name of this program (e.g., no `g' prefix).  */
  45 #if HASH_ALGO_MD5
  46 # define PROGRAM_NAME "md5sum"
  47 # define DIGEST_TYPE_STRING "MD5"
  48 # define DIGEST_STREAM md5_stream
  49 # define DIGEST_BITS 128
  50 # define DIGEST_REFERENCE "RFC 1321"
  51 # define DIGEST_ALIGN 4
  52 #elif HASH_ALGO_SHA1
  53 # define PROGRAM_NAME "sha1sum"
  54 # define DIGEST_TYPE_STRING "SHA1"
  55 # define DIGEST_STREAM sha1_stream
  56 # define DIGEST_BITS 160
  57 # define DIGEST_REFERENCE "FIPS-180-1"
  58 # define DIGEST_ALIGN 4
  59 #elif HASH_ALGO_SHA256
  60 # define PROGRAM_NAME "sha256sum"
  61 # define DIGEST_TYPE_STRING "SHA256"
  62 # define DIGEST_STREAM sha256_stream
  63 # define DIGEST_BITS 256
  64 # define DIGEST_REFERENCE "FIPS-180-2"
  65 # define DIGEST_ALIGN 4
  66 #elif HASH_ALGO_SHA224
  67 # define PROGRAM_NAME "sha224sum"
  68 # define DIGEST_TYPE_STRING "SHA224"
  69 # define DIGEST_STREAM sha224_stream
  70 # define DIGEST_BITS 224
  71 # define DIGEST_REFERENCE "RFC 3874"
  72 # define DIGEST_ALIGN 4
  73 #elif HASH_ALGO_SHA512
  74 # define PROGRAM_NAME "sha512sum"
  75 # define DIGEST_TYPE_STRING "SHA512"
  76 # define DIGEST_STREAM sha512_stream
  77 # define DIGEST_BITS 512
  78 # define DIGEST_REFERENCE "FIPS-180-2"
  79 # define DIGEST_ALIGN 8
  80 #elif HASH_ALGO_SHA384
  81 # define PROGRAM_NAME "sha384sum"
  82 # define DIGEST_TYPE_STRING "SHA384"
  83 # define DIGEST_STREAM sha384_stream
  84 # define DIGEST_BITS 384
  85 # define DIGEST_REFERENCE "FIPS-180-2"
  86 # define DIGEST_ALIGN 8
  87 #else
  88 # error "Can't decide which hash algorithm to compile."
  89 #endif
  90
  91 #define DIGEST_HEX_BYTES (DIGEST_BITS / 4)
  92 #define DIGEST_BIN_BYTES (DIGEST_BITS / 8)
  93
  94 #define AUTHORS "Ulrich Drepper", "Scott Miller", "David Madore"
  95
  96 /* The minimum length of a valid digest line.  This length does
  97    not include any newline character at the end of a line.  */
  98 #define MIN_DIGEST_LINE_LENGTH \
  99   (DIGEST_HEX_BYTES /* length of hexadecimal message digest */ \
 100    + 2 /* blank and binary indicator */ \
 101    + 1 /* minimum filename length */ )
 102
 103 /* True if any of the files read were the standard input. */
 104 static bool have_read_stdin;
 105
 106 /* The minimum length of a valid checksum line for the selected algorithm.  */
 107 static size_t min_digest_line_length;
 108
 109 /* Set to the length of a digest hex string for the selected algorithm.  */
 110 static size_t digest_hex_bytes;
 111
 112 /* With --check, don't generate any output.
 113    The exit code indicates success or failure.  */
 114 static bool status_only = false;
 115
 116 /* With --check, print a message to standard error warning about each
 117    improperly formatted checksum line.  */
 118 static bool warn = false;
 119
 120 /* The name this program was run with.  */
 121 char *program_name;
 122
 123 /* For long options that have no equivalent short option, use a
 124    non-character as a pseudo short option, starting with CHAR_MAX + 1.  */
 125 enum
 126 {
 127   STATUS_OPTION = CHAR_MAX + 1
 128 };
 129
 130 static const struct option long_options[] =
 131 {
 132   { "binary", no_argument, NULL, 'b' },
 133   { "check", no_argument, NULL, 'c' },
 134   { "status", no_argument, NULL, STATUS_OPTION },
 135   { "text", no_argument, NULL, 't' },
 136   { "warn", no_argument, NULL, 'w' },
 137   { GETOPT_HELP_OPTION_DECL },
 138   { GETOPT_VERSION_OPTION_DECL },
 139   { NULL, 0, NULL, 0 }
 140 };
 141
 142 void
 143 usage (int status)
 144 {
 145   if (status != EXIT_SUCCESS)
 146     fprintf (stderr, _("Try `%s --help' for more information.\n"),
 147              program_name);
 148   else
 149     {
 150       printf (_("\
 151 Usage: %s [OPTION] [FILE]...\n\
 152 Print or check %s (%d-bit) checksums.\n\
 153 With no FILE, or when FILE is -, read standard input.\n\
 154 \n\
 155 "),
 156               program_name,
 157               DIGEST_TYPE_STRING,
 158               DIGEST_BITS);
 159       if (O_BINARY)
 160         fputs (_("\
 161   -b, --binary            read in binary mode (default unless reading tty stdin)\n\
 162 "), stdout);
 163       else
 164         fputs (_("\
 165   -b, --binary            read in binary mode\n\
 166 "), stdout);
 167       printf (_("\
 168   -c, --check             read %s sums from the FILEs and check them\n"),
 169               DIGEST_TYPE_STRING);
 170       if (O_BINARY)
 171         fputs (_("\
 172   -t, --text              read in text mode (default if reading tty stdin)\n\
 173 "), stdout);
 174       else
 175         fputs (_("\
 176   -t, --text              read in text mode (default)\n\
 177 "), stdout);
 178       fputs (_("\
 179 \n\
 180 The following two options are useful only when verifying checksums:\n\
 181       --status            don't output anything, status code shows success\n\
 182   -w, --warn              warn about improperly formatted checksum lines\n\
 183 \n\
 184 "), stdout);
 185       fputs (HELP_OPTION_DESCRIPTION, stdout);
 186       fputs (VERSION_OPTION_DESCRIPTION, stdout);
 187       printf (_("\
 188 \n\
 189 The sums are computed as described in %s.  When checking, the input\n\
 190 should be a former output of this program.  The default mode is to print\n\
 191 a line with checksum, a character indicating type (`*' for binary, ` ' for\n\
 192 text), and name for each FILE.\n"),
 193               DIGEST_REFERENCE);
 194       printf (_("\nReport bugs to <%s>.\n"), PACKAGE_BUGREPORT);
 195     }
 196
 197   exit (status);
 198 }
 199
 200 #define ISWHITE(c) ((c) == ' ' || (c) == '\t')
 201
 202 /* Split the checksum string S (of length S_LEN) from a BSD 'md5' or
 203    'sha1' command into two parts: a hexadecimal digest, and the file
 204    name.  S is modified.  Return true if successful.  */
 205
 206 static bool
 207 bsd_split_3 (char *s, size_t s_len, unsigned char **hex_digest, char **file_name)
 208 {
 209   size_t i;
 210
 211   *file_name = s;
 212
 213   /* Find end of filename. The BSD 'md5' and 'sha1' commands do not escape
 214      filenames, so search backwards for the last ')'. */
 215   i = s_len - 1;
 216   while (i && s[i] != ')')
 217     i--;
 218
 219   if (s[i] != ')')
 220     return false;
 221
 222   s[i++] = '\0';
 223
 224   while (ISWHITE (s[i]))
 225     i++;
 226
 227   if (s[i] != '=')
 228     return false;
 229
 230   i++;
 231
 232   while (ISWHITE (s[i]))
 233     i++;
 234
 235   *hex_digest = (unsigned char *) &s[i];
 236   return true;
 237 }
 238
 239 /* Split the string S (of length S_LEN) into three parts:
 240    a hexadecimal digest, binary flag, and the file name.
 241    S is modified.  Return true if successful.  */
 242
 243 static bool
 244 split_3 (char *s, size_t s_len,
 245          unsigned char **hex_digest, int *binary, char **file_name)
 246 {
 247   size_t i;
 248   bool escaped_filename = false;
 249   size_t algo_name_len;
 250
 251   i = 0;
 252   while (ISWHITE (s[i]))
 253     ++i;
 254
 255   /* Check for BSD-style checksum line. */
 256   algo_name_len = strlen (DIGEST_TYPE_STRING);
 257   if (strncmp (s + i, DIGEST_TYPE_STRING, algo_name_len) == 0)
 258     {
 259       if (strncmp (s + i + algo_name_len, " (", 2) == 0)
 260         {
 261           *binary = 0;
 262           return bsd_split_3 (s +      i + algo_name_len + 2,
 263                               s_len - (i + algo_name_len + 2),
 264                               hex_digest, file_name);
 265         }
 266     }
 267
 268   /* Ignore this line if it is too short.
 269      Each line must have at least `min_digest_line_length - 1' (or one more, if
 270      the first is a backslash) more characters to contain correct message digest
 271      information.  */
 272   if (s_len - i < min_digest_line_length + (s[i] == '\\'))
 273     return false;
 274
 275   if (s[i] == '\\')
 276     {
 277       ++i;
 278       escaped_filename = true;
 279     }
 280   *hex_digest = (unsigned char *) &s[i];
 281
 282   /* The first field has to be the n-character hexadecimal
 283      representation of the message digest.  If it is not followed
 284      immediately by a white space it's an error.  */
 285   i += digest_hex_bytes;
 286   if (!ISWHITE (s[i]))
 287     return false;
 288
 289   s[i++] = '\0';
 290
 291   if (s[i] != ' ' && s[i] != '*')
 292     return false;
 293   *binary = (s[i++] == '*');
 294
 295   /* All characters between the type indicator and end of line are
 296      significant -- that includes leading and trailing white space.  */
 297   *file_name = &s[i];
 298
 299   if (escaped_filename)
 300     {
 301       /* Translate each `\n' string in the file name to a NEWLINE,
 302          and each `\\' string to a backslash.  */
 303
 304       char *dst = &s[i];
 305
 306       while (i < s_len)
 307         {
 308           switch (s[i])
 309             {
 310             case '\\':
 311               if (i == s_len - 1)
 312                 {
 313                   /* A valid line does not end with a backslash.  */
 314                   return false;
 315                 }
 316               ++i;
 317               switch (s[i++])
 318                 {
 319                 case 'n':
 320                   *dst++ = '\n';
 321                   break;
 322                 case '\\':
 323                   *dst++ = '\\';
 324                   break;
 325                 default:
 326                   /* Only `\' or `n' may follow a backslash.  */
 327                   return false;
 328                 }
 329               break;
 330
 331             case '\0':
 332               /* The file name may not contain a NUL.  */
 333               return false;
 334               break;
 335
 336             default:
 337               *dst++ = s[i++];
 338               break;
 339             }
 340         }
 341       *dst = '\0';
 342     }
 343   return true;
 344 }
 345
 346 static bool
 347 hex_digits (unsigned char const *s)
 348 {
 349   while (*s)
 350     {
 351       if (!isxdigit (*s))
 352         return false;
 353       ++s;
 354     }
 355   return true;
 356 }
 357
 358 /* An interface to the function, DIGEST_STREAM.
 359    Operate on FILENAME (it may be "-").
 360
 361    *BINARY indicates whether the file is binary.  BINARY < 0 means it
 362    depends on whether binary mode makes any difference and the file is
 363    a terminal; in that case, clear *BINARY if the file was treated as
 364    text because it was a terminal.
 365
 366    Put the checksum in *BIN_RESULT, which must be properly aligned.
 367    Return true if successful.  */
 368
 369 static bool
 370 digest_file (const char *filename, int *binary, unsigned char *bin_result)
 371 {
 372   FILE *fp;
 373   int err;
 374   bool is_stdin = STREQ (filename, "-");
 375
 376   if (is_stdin)
 377     {
 378       have_read_stdin = true;
 379       fp = stdin;
 380       if (O_BINARY && *binary)
 381         {
 382           if (*binary < 0)
 383             *binary = ! isatty (STDIN_FILENO);
 384           if (*binary)
 385             freopen (NULL, "rb", stdin);
 386         }
 387     }
 388   else
 389     {
 390       fp = fopen (filename, (O_BINARY && *binary ? "rb" : "r"));
 391       if (fp == NULL)
 392         {
 393           error (0, errno, "%s", filename);
 394           return false;
 395         }
 396     }
 397
 398   err = DIGEST_STREAM (fp, bin_result);
 399   if (err)
 400     {
 401       error (0, errno, "%s", filename);
 402       if (fp != stdin)
 403         fclose (fp);
 404       return false;
 405     }
 406
 407   if (!is_stdin && fclose (fp) != 0)
 408     {
 409       error (0, errno, "%s", filename);
 410       return false;
 411     }
 412
 413   return true;
 414 }
 415
 416 static bool
 417 digest_check (const char *checkfile_name)
 418 {
 419   FILE *checkfile_stream;
 420   uintmax_t n_properly_formatted_lines = 0;
 421   uintmax_t n_mismatched_checksums = 0;
 422   uintmax_t n_open_or_read_failures = 0;
 423   unsigned char bin_buffer_unaligned[DIGEST_BIN_BYTES + DIGEST_ALIGN];
 424   /* Make sure bin_buffer is properly aligned. */
 425   unsigned char *bin_buffer = ptr_align (bin_buffer_unaligned, DIGEST_ALIGN);
 426   uintmax_t line_number;
 427   char *line;
 428   size_t line_chars_allocated;
 429   bool is_stdin = STREQ (checkfile_name, "-");
 430
 431   if (is_stdin)
 432     {
 433       have_read_stdin = true;
 434       checkfile_name = _("standard input");
 435       checkfile_stream = stdin;
 436     }
 437   else
 438     {
 439       checkfile_stream = fopen (checkfile_name, "r");
 440       if (checkfile_stream == NULL)
 441         {
 442           error (0, errno, "%s", checkfile_name);
 443           return false;
 444         }
 445     }
 446
 447   line_number = 0;
 448   line = NULL;
 449   line_chars_allocated = 0;
 450   do
 451     {
 452       char *filename;
 453       int binary;
 454       unsigned char *hex_digest IF_LINT (= NULL);
 455       ssize_t line_length;
 456
 457       ++line_number;
 458       if (line_number == 0)
 459         error (EXIT_FAILURE, 0, _("%s: too many checksum lines"),
 460                checkfile_name);
 461
 462       line_length = getline (&line, &line_chars_allocated, checkfile_stream);
 463       if (line_length <= 0)
 464         break;
 465
 466       /* Ignore comment lines, which begin with a '#' character.  */
 467       if (line[0] == '#')
 468         continue;
 469
 470       /* Remove any trailing newline.  */
 471       if (line[line_length - 1] == '\n')
 472         line[--line_length] = '\0';
 473
 474       if (! (split_3 (line, line_length, &hex_digest, &binary, &filename)
 475              && ! (is_stdin && STREQ (filename, "-"))
 476              && hex_digits (hex_digest)))
 477         {
 478           if (warn)
 479             {
 480               error (0, 0,
 481                      _("%s: %" PRIuMAX
 482                        ": improperly formatted %s checksum line"),
 483                      checkfile_name, line_number,
 484                      DIGEST_TYPE_STRING);
 485             }
 486         }
 487       else
 488         {
 489           static const char bin2hex[] = { '0', '1', '2', '3',
 490                                           '4', '5', '6', '7',
 491                                           '8', '9', 'a', 'b',
 492                                           'c', 'd', 'e', 'f' };
 493           bool ok;
 494
 495           ++n_properly_formatted_lines;
 496
 497           ok = digest_file (filename, &binary, bin_buffer);
 498
 499           if (!ok)
 500             {
 501               ++n_open_or_read_failures;
 502               if (!status_only)
 503                 {
 504                   printf (_("%s: FAILED open or read\n"), filename);
 505                   fflush (stdout);
 506                 }
 507             }
 508           else
 509             {
 510               size_t digest_bin_bytes = digest_hex_bytes / 2;
 511               size_t cnt;
 512               /* Compare generated binary number with text representation
 513                  in check file.  Ignore case of hex digits.  */
 514               for (cnt = 0; cnt < digest_bin_bytes; ++cnt)
 515                 {
 516                   if (tolower (hex_digest[2 * cnt])
 517                       != bin2hex[bin_buffer[cnt] >> 4]
 518                       || (tolower (hex_digest[2 * cnt + 1])
 519                           != (bin2hex[bin_buffer[cnt] & 0xf])))
 520                     break;
 521                 }
 522               if (cnt != digest_bin_bytes)
 523                 ++n_mismatched_checksums;
 524
 525               if (!status_only)
 526                 {
 527                   printf ("%s: %s\n", filename,
 528                           (cnt != digest_bin_bytes ? _("FAILED") : _("OK")));
 529                   fflush (stdout);
 530                 }
 531             }
 532         }
 533     }
 534   while (!feof (checkfile_stream) && !ferror (checkfile_stream));
 535
 536   free (line);
 537
 538   if (ferror (checkfile_stream))
 539     {
 540       error (0, 0, _("%s: read error"), checkfile_name);
 541       return false;
 542     }
 543
 544   if (!is_stdin && fclose (checkfile_stream) != 0)
 545     {
 546       error (0, errno, "%s", checkfile_name);
 547       return false;
 548     }
 549
 550   if (n_properly_formatted_lines == 0)
 551     {
 552       /* Warn if no tests are found.  */
 553       error (0, 0, _("%s: no properly formatted %s checksum lines found"),
 554              checkfile_name, DIGEST_TYPE_STRING);
 555     }
 556   else
 557     {
 558       if (!status_only)
 559         {
 560           if (n_open_or_read_failures != 0)
 561             error (0, 0,
 562                    ngettext ("WARNING: %" PRIuMAX " of %" PRIuMAX
 563                              " listed file could not be read",
 564                              "WARNING: %" PRIuMAX " of %" PRIuMAX
 565                              " listed files could not be read",
 566                              select_plural (n_properly_formatted_lines)),
 567                    n_open_or_read_failures, n_properly_formatted_lines);
 568
 569           if (n_mismatched_checksums != 0)
 570             {
 571               uintmax_t n_computed_checksums =
 572                 (n_properly_formatted_lines - n_open_or_read_failures);
 573               error (0, 0,
 574                      ngettext ("WARNING: %" PRIuMAX " of %" PRIuMAX
 575                                " computed checksum did NOT match",
 576                                "WARNING: %" PRIuMAX " of %" PRIuMAX
 577                                " computed checksums did NOT match",
 578                                select_plural (n_computed_checksums)),
 579                      n_mismatched_checksums, n_computed_checksums);
 580             }
 581         }
 582     }
 583
 584   return (n_properly_formatted_lines != 0
 585           && n_mismatched_checksums == 0
 586           && n_open_or_read_failures == 0);
 587 }
 588
 589 int
 590 main (int argc, char **argv)
 591 {
 592   unsigned char bin_buffer_unaligned[DIGEST_BIN_BYTES + DIGEST_ALIGN];
 593   /* Make sure bin_buffer is properly aligned. */
 594   unsigned char *bin_buffer = ptr_align (bin_buffer_unaligned, DIGEST_ALIGN);
 595   bool do_check = false;
 596   int opt;
 597   bool ok = true;
 598   int binary = -1;
 599
 600   /* Setting values of global variables.  */
 601   initialize_main (&argc, &argv);
 602   program_name = argv[0];
 603   setlocale (LC_ALL, "");
 604   bindtextdomain (PACKAGE, LOCALEDIR);
 605   textdomain (PACKAGE);
 606
 607   atexit (close_stdout);
 608
 609   while ((opt = getopt_long (argc, argv, "bctw", long_options, NULL)) != -1)
 610     switch (opt)
 611       {
 612       case 'b':
 613         binary = 1;
 614         break;
 615       case 'c':
 616         do_check = true;
 617         break;
 618       case STATUS_OPTION:
 619         status_only = true;
 620         warn = false;
 621         break;
 622       case 't':
 623         binary = 0;
 624         break;
 625       case 'w':
 626         status_only = false;
 627         warn = true;
 628         break;
 629       case_GETOPT_HELP_CHAR;
 630       case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
 631       default:
 632         usage (EXIT_FAILURE);
 633       }
 634
 635   min_digest_line_length = MIN_DIGEST_LINE_LENGTH;
 636   digest_hex_bytes = DIGEST_HEX_BYTES;
 637
 638   if (0 <= binary && do_check)
 639     {
 640       error (0, 0, _("the --binary and --text options are meaningless when "
 641                      "verifying checksums"));
 642       usage (EXIT_FAILURE);
 643     }
 644
 645   if (status_only & !do_check)
 646     {
 647       error (0, 0,
 648        _("the --status option is meaningful only when verifying checksums"));
 649       usage (EXIT_FAILURE);
 650     }
 651
 652   if (warn & !do_check)
 653     {
 654       error (0, 0,
 655        _("the --warn option is meaningful only when verifying checksums"));
 656       usage (EXIT_FAILURE);
 657     }
 658
 659   if (!O_BINARY && binary < 0)
 660     binary = 0;
 661
 662   if (optind == argc)
 663     argv[argc++] = "-";
 664
 665   for (; optind < argc; ++optind)
 666     {
 667       char *file = argv[optind];
 668
 669       if (do_check)
 670         ok &= digest_check (file);
 671       else
 672         {
 673           int file_is_binary = binary;
 674
 675           if (! digest_file (file, &file_is_binary, bin_buffer))
 676             ok = false;
 677           else
 678             {
 679               size_t i;
 680
 681               /* Output a leading backslash if the file name contains
 682                  a newline or backslash.  */
 683               if (strchr (file, '\n') || strchr (file, '\\'))
 684                 putchar ('\\');
 685
 686               for (i = 0; i < (digest_hex_bytes / 2); ++i)
 687                 printf ("%02x", bin_buffer[i]);
 688
 689               putchar (' ');
 690               if (file_is_binary)
 691                 putchar ('*');
 692               else
 693                 putchar (' ');
 694
 695               /* Translate each NEWLINE byte to the string, "\\n",
 696                  and each backslash to "\\\\".  */
 697               for (i = 0; i < strlen (file); ++i)
 698                 {
 699                   switch (file[i])
 700                     {
 701                     case '\n':
 702                       fputs ("\\n", stdout);
 703                       break;
 704
 705                     case '\\':
 706                       fputs ("\\\\", stdout);
 707                       break;
 708
 709                     default:
 710                       putchar (file[i]);
 711                       break;
 712                     }
 713                 }
 714               putchar ('\n');
 715             }
 716         }
 717     }
 718
 719   if (have_read_stdin && fclose (stdin) == EOF)
 720     error (EXIT_FAILURE, errno, _("standard input"));
 721
 722   exit (ok ? EXIT_SUCCESS : EXIT_FAILURE);
 723 }