Dogcows Code - chaz/tar/blob - src/compare.c

   1 /* Diff files from a tar archive.
   2
   3    Copyright (C) 1988, 1992, 1993, 1994, 1996, 1997, 1999, 2000, 2001,
   4    2003, 2004, 2005, 2006, 2007, 2009, 2010 Free Software Foundation, Inc.
   5
   6    Written by John Gilmore, on 1987-04-30.
   7
   8    This program is free software; you can redistribute it and/or modify it
   9    under the terms of the GNU General Public License as published by the
  10    Free Software Foundation; either version 3, or (at your option) any later
  11    version.
  12
  13    This program is distributed in the hope that it will be useful, but
  14    WITHOUT ANY WARRANTY; without even the implied warranty of
  15    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General
  16    Public License for more details.
  17
  18    You should have received a copy of the GNU General Public License along
  19    with this program; if not, write to the Free Software Foundation, Inc.,
  20    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.  */
  21
  22 #include <system.h>
  23 #include <system-ioctl.h>
  24
  25 #if HAVE_LINUX_FD_H
  26 # include <linux/fd.h>
  27 #endif
  28
  29 #include "common.h"
  30 #include <quotearg.h>
  31 #include <rmt.h>
  32 #include <stdarg.h>
  33
  34 /* Nonzero if we are verifying at the moment.  */
  35 bool now_verifying;
  36
  37 /* File descriptor for the file we are diffing.  */
  38 static int diff_handle;
  39
  40 /* Area for reading file contents into.  */
  41 static char *diff_buffer;
  42
  43 /* Initialize for a diff operation.  */
  44 void
  45 diff_init (void)
  46 {
  47   void *ptr;
  48   diff_buffer = page_aligned_alloc (&ptr, record_size);
  49   if (listed_incremental_option)
  50     read_directory_file ();
  51 }
  52
  53 /* Sigh about something that differs by writing a MESSAGE to stdlis,
  54    given MESSAGE is nonzero.  Also set the exit status if not already.  */
  55 void
  56 report_difference (struct tar_stat_info *st, const char *fmt, ...)
  57 {
  58   if (fmt)
  59     {
  60       va_list ap;
  61
  62       fprintf (stdlis, "%s: ", quotearg_colon (st->file_name));
  63       va_start (ap, fmt);
  64       vfprintf (stdlis, fmt, ap);
  65       va_end (ap);
  66       fprintf (stdlis, "\n");
  67     }
  68
  69   set_exit_status (TAREXIT_DIFFERS);
  70 }
  71
  72 /* Take a buffer returned by read_and_process and do nothing with it.  */
  73 static int
  74 process_noop (size_t size __attribute__ ((unused)),
  75               char *data __attribute__ ((unused)))
  76 {
  77   return 1;
  78 }
  79
  80 static int
  81 process_rawdata (size_t bytes, char *buffer)
  82 {
  83   size_t status = blocking_read (diff_handle, diff_buffer, bytes);
  84
  85   if (status != bytes)
  86     {
  87       if (status == SAFE_READ_ERROR)
  88         {
  89           read_error (current_stat_info.file_name);
  90           report_difference (&current_stat_info, NULL);
  91         }
  92       else
  93         {
  94           report_difference (&current_stat_info,
  95                              ngettext ("Could only read %lu of %lu byte",
  96                                        "Could only read %lu of %lu bytes",
  97                                        bytes),
  98                              (unsigned long) status, (unsigned long) bytes);
  99         }
 100       return 0;
 101     }
 102
 103   if (memcmp (buffer, diff_buffer, bytes))
 104     {
 105       report_difference (&current_stat_info, _("Contents differ"));
 106       return 0;
 107     }
 108
 109   return 1;
 110 }
 111
 112 /* Some other routine wants SIZE bytes in the archive.  For each chunk
 113    of the archive, call PROCESSOR with the size of the chunk, and the
 114    address of the chunk it can work with.  The PROCESSOR should return
 115    nonzero for success.  Once it returns error, continue skipping
 116    without calling PROCESSOR anymore.  */
 117
 118 static void
 119 read_and_process (struct tar_stat_info *st, int (*processor) (size_t, char *))
 120 {
 121   union block *data_block;
 122   size_t data_size;
 123   off_t size = st->stat.st_size;
 124
 125   mv_begin_read (st);
 126   while (size)
 127     {
 128       data_block = find_next_block ();
 129       if (! data_block)
 130         {
 131           ERROR ((0, 0, _("Unexpected EOF in archive")));
 132           return;
 133         }
 134
 135       data_size = available_space_after (data_block);
 136       if (data_size > size)
 137         data_size = size;
 138       if (!(*processor) (data_size, data_block->buffer))
 139         processor = process_noop;
 140       set_next_block_after ((union block *)
 141                             (data_block->buffer + data_size - 1));
 142       size -= data_size;
 143       mv_size_left (size);
 144     }
 145   mv_end ();
 146 }
 147
 148 /* Call either stat or lstat over STAT_DATA, depending on
 149    --dereference (-h), for a file which should exist.  Diagnose any
 150    problem.  Return nonzero for success, zero otherwise.  */
 151 static int
 152 get_stat_data (char const *file_name, struct stat *stat_data)
 153 {
 154   int status = deref_stat (file_name, stat_data);
 155
 156   if (status != 0)
 157     {
 158       if (errno == ENOENT)
 159         stat_warn (file_name);
 160       else
 161         stat_error (file_name);
 162       report_difference (&current_stat_info, NULL);
 163       return 0;
 164     }
 165
 166   return 1;
 167 }
 168
 169 \f
 170 static void
 171 diff_dir (void)
 172 {
 173   struct stat stat_data;
 174
 175   if (!get_stat_data (current_stat_info.file_name, &stat_data))
 176     return;
 177
 178   if (!S_ISDIR (stat_data.st_mode))
 179     report_difference (&current_stat_info, _("File type differs"));
 180   else if ((current_stat_info.stat.st_mode & MODE_ALL) !=
 181            (stat_data.st_mode & MODE_ALL))
 182     report_difference (&current_stat_info, _("Mode differs"));
 183 }
 184
 185 static void
 186 diff_file (void)
 187 {
 188   char const *file_name = current_stat_info.file_name;
 189   struct stat stat_data;
 190
 191   if (!get_stat_data (file_name, &stat_data))
 192     skip_member ();
 193   else if (!S_ISREG (stat_data.st_mode))
 194     {
 195       report_difference (&current_stat_info, _("File type differs"));
 196       skip_member ();
 197     }
 198   else
 199     {
 200       if ((current_stat_info.stat.st_mode & MODE_ALL) !=
 201           (stat_data.st_mode & MODE_ALL))
 202         report_difference (&current_stat_info, _("Mode differs"));
 203
 204       if (!sys_compare_uid (&stat_data, &current_stat_info.stat))
 205         report_difference (&current_stat_info, _("Uid differs"));
 206       if (!sys_compare_gid (&stat_data, &current_stat_info.stat))
 207         report_difference (&current_stat_info, _("Gid differs"));
 208
 209       if (tar_timespec_cmp (get_stat_mtime (&stat_data),
 210                             current_stat_info.mtime))
 211         report_difference (&current_stat_info, _("Mod time differs"));
 212       if (current_header->header.typeflag != GNUTYPE_SPARSE
 213           && stat_data.st_size != current_stat_info.stat.st_size)
 214         {
 215           report_difference (&current_stat_info, _("Size differs"));
 216           skip_member ();
 217         }
 218       else
 219         {
 220           diff_handle = openat (chdir_fd, file_name, open_read_flags);
 221
 222           if (diff_handle < 0)
 223             {
 224               open_error (file_name);
 225               skip_member ();
 226               report_difference (&current_stat_info, NULL);
 227             }
 228           else
 229             {
 230               int status;
 231
 232               if (current_stat_info.is_sparse)
 233                 sparse_diff_file (diff_handle, &current_stat_info);
 234               else
 235                 read_and_process (&current_stat_info, process_rawdata);
 236
 237               if (atime_preserve_option == replace_atime_preserve
 238                   && stat_data.st_size != 0)
 239                 {
 240                   struct timespec atime = get_stat_atime (&stat_data);
 241                   if (set_file_atime (diff_handle, chdir_fd, file_name, atime)
 242                       != 0)
 243                     utime_error (file_name);
 244                 }
 245
 246               status = close (diff_handle);
 247               if (status != 0)
 248                 close_error (file_name);
 249             }
 250         }
 251     }
 252 }
 253
 254 static void
 255 diff_link (void)
 256 {
 257   struct stat file_data;
 258   struct stat link_data;
 259
 260   if (get_stat_data (current_stat_info.file_name, &file_data)
 261       && get_stat_data (current_stat_info.link_name, &link_data)
 262       && !sys_compare_links (&file_data, &link_data))
 263     report_difference (&current_stat_info,
 264                        _("Not linked to %s"),
 265                        quote (current_stat_info.link_name));
 266 }
 267
 268 #ifdef HAVE_READLINK
 269 static void
 270 diff_symlink (void)
 271 {
 272   size_t len = strlen (current_stat_info.link_name);
 273   char *linkbuf = alloca (len + 1);
 274
 275   int status = readlinkat (chdir_fd, current_stat_info.file_name,
 276                            linkbuf, len + 1);
 277
 278   if (status < 0)
 279     {
 280       if (errno == ENOENT)
 281         readlink_warn (current_stat_info.file_name);
 282       else
 283         readlink_error (current_stat_info.file_name);
 284       report_difference (&current_stat_info, NULL);
 285     }
 286   else if (status != len
 287            || strncmp (current_stat_info.link_name, linkbuf, len) != 0)
 288     report_difference (&current_stat_info, _("Symlink differs"));
 289 }
 290 #endif
 291
 292 static void
 293 diff_special (void)
 294 {
 295   struct stat stat_data;
 296
 297   /* FIXME: deal with umask.  */
 298
 299   if (!get_stat_data (current_stat_info.file_name, &stat_data))
 300     return;
 301
 302   if (current_header->header.typeflag == CHRTYPE
 303       ? !S_ISCHR (stat_data.st_mode)
 304       : current_header->header.typeflag == BLKTYPE
 305       ? !S_ISBLK (stat_data.st_mode)
 306       : /* current_header->header.typeflag == FIFOTYPE */
 307       !S_ISFIFO (stat_data.st_mode))
 308     {
 309       report_difference (&current_stat_info, _("File type differs"));
 310       return;
 311     }
 312
 313   if ((current_header->header.typeflag == CHRTYPE
 314        || current_header->header.typeflag == BLKTYPE)
 315       && current_stat_info.stat.st_rdev != stat_data.st_rdev)
 316     {
 317       report_difference (&current_stat_info, _("Device number differs"));
 318       return;
 319     }
 320
 321   if ((current_stat_info.stat.st_mode & MODE_ALL) !=
 322       (stat_data.st_mode & MODE_ALL))
 323     report_difference (&current_stat_info, _("Mode differs"));
 324 }
 325
 326 static int
 327 dumpdir_cmp (const char *a, const char *b)
 328 {
 329   size_t len;
 330
 331   while (*a)
 332     switch (*a)
 333       {
 334       case 'Y':
 335       case 'N':
 336         if (!strchr ("YN", *b))
 337           return 1;
 338         if (strcmp(a + 1, b + 1))
 339           return 1;
 340         len = strlen (a) + 1;
 341         a += len;
 342         b += len;
 343         break;
 344
 345       case 'D':
 346         if (strcmp(a, b))
 347           return 1;
 348         len = strlen (a) + 1;
 349         a += len;
 350         b += len;
 351         break;
 352
 353       case 'R':
 354       case 'T':
 355       case 'X':
 356         return *b;
 357       }
 358   return *b;
 359 }
 360
 361 static void
 362 diff_dumpdir (struct tar_stat_info *dir)
 363 {
 364   const char *dumpdir_buffer;
 365   struct stat stat_data;
 366
 367   if (deref_stat (dir->file_name, &stat_data) != 0)
 368     {
 369       if (errno == ENOENT)
 370         stat_warn (dir->file_name);
 371       else
 372         stat_error (dir->file_name);
 373     }
 374
 375   if (dir->fd == 0)
 376     {
 377       void (*diag) (char const *) = NULL;
 378       int fd = subfile_open (dir->parent, dir->orig_file_name, open_read_flags);
 379       if (fd < 0)
 380         diag = open_diag;
 381       else if (fstat (fd, &dir->stat))
 382         diag = stat_diag;
 383       else
 384         dir->fd = fd;
 385       if (diag)
 386         {
 387           file_removed_diag (dir->orig_file_name, false, diag);
 388           return;
 389         }
 390     }
 391   dumpdir_buffer = directory_contents (scan_directory (dir));
 392
 393   if (dumpdir_buffer)
 394     {
 395       if (dumpdir_cmp (dir->dumpdir, dumpdir_buffer))
 396         report_difference (dir, _("Contents differ"));
 397     }
 398   else
 399     read_and_process (dir, process_noop);
 400 }
 401
 402 static void
 403 diff_multivol (void)
 404 {
 405   struct stat stat_data;
 406   int fd, status;
 407   off_t offset;
 408
 409   if (current_stat_info.had_trailing_slash)
 410     {
 411       diff_dir ();
 412       return;
 413     }
 414
 415   if (!get_stat_data (current_stat_info.file_name, &stat_data))
 416     return;
 417
 418   if (!S_ISREG (stat_data.st_mode))
 419     {
 420       report_difference (&current_stat_info, _("File type differs"));
 421       skip_member ();
 422       return;
 423     }
 424
 425   offset = OFF_FROM_HEADER (current_header->oldgnu_header.offset);
 426   if (stat_data.st_size != current_stat_info.stat.st_size + offset)
 427     {
 428       report_difference (&current_stat_info, _("Size differs"));
 429       skip_member ();
 430       return;
 431     }
 432
 433
 434   fd = openat (chdir_fd, current_stat_info.file_name, open_read_flags);
 435
 436   if (fd < 0)
 437     {
 438       open_error (current_stat_info.file_name);
 439       report_difference (&current_stat_info, NULL);
 440       skip_member ();
 441       return;
 442     }
 443
 444   if (lseek (fd, offset, SEEK_SET) < 0)
 445     {
 446       seek_error_details (current_stat_info.file_name, offset);
 447       report_difference (&current_stat_info, NULL);
 448       return;
 449     }
 450
 451   read_and_process (&current_stat_info, process_rawdata);
 452
 453   status = close (fd);
 454   if (status != 0)
 455     close_error (current_stat_info.file_name);
 456 }
 457
 458 /* Diff a file against the archive.  */
 459 void
 460 diff_archive (void)
 461 {
 462
 463   set_next_block_after (current_header);
 464
 465   /* Print the block from current_header and current_stat_info.  */
 466
 467   if (verbose_option)
 468     {
 469       if (now_verifying)
 470         fprintf (stdlis, _("Verify "));
 471       print_header (&current_stat_info, current_header, -1);
 472     }
 473
 474   switch (current_header->header.typeflag)
 475     {
 476     default:
 477       ERROR ((0, 0, _("%s: Unknown file type '%c', diffed as normal file"),
 478               quotearg_colon (current_stat_info.file_name),
 479               current_header->header.typeflag));
 480       /* Fall through.  */
 481
 482     case AREGTYPE:
 483     case REGTYPE:
 484     case GNUTYPE_SPARSE:
 485     case CONTTYPE:
 486
 487       /* Appears to be a file.  See if it's really a directory.  */
 488
 489       if (current_stat_info.had_trailing_slash)
 490         diff_dir ();
 491       else
 492         diff_file ();
 493       break;
 494
 495     case LNKTYPE:
 496       diff_link ();
 497       break;
 498
 499 #ifdef HAVE_READLINK
 500     case SYMTYPE:
 501       diff_symlink ();
 502       break;
 503 #endif
 504
 505     case CHRTYPE:
 506     case BLKTYPE:
 507     case FIFOTYPE:
 508       diff_special ();
 509       break;
 510
 511     case GNUTYPE_DUMPDIR:
 512     case DIRTYPE:
 513       if (is_dumpdir (&current_stat_info))
 514         diff_dumpdir (&current_stat_info);
 515       diff_dir ();
 516       break;
 517
 518     case GNUTYPE_VOLHDR:
 519       break;
 520
 521     case GNUTYPE_MULTIVOL:
 522       diff_multivol ();
 523     }
 524 }
 525
 526 void
 527 verify_volume (void)
 528 {
 529   int may_fail = 0;
 530   if (removed_prefixes_p ())
 531     {
 532       WARN((0, 0,
 533             _("Archive contains file names with leading prefixes removed.")));
 534       may_fail = 1;
 535     }
 536   if (transform_program_p ())
 537     {
 538       WARN((0, 0,
 539             _("Archive contains transformed file names.")));
 540       may_fail = 1;
 541     }
 542   if (may_fail)
 543     WARN((0, 0,
 544           _("Verification may fail to locate original files.")));
 545
 546   clear_directory_table ();
 547
 548   if (!diff_buffer)
 549     diff_init ();
 550
 551   /* Verifying an archive is meant to check if the physical media got it
 552      correctly, so try to defeat clever in-memory buffering pertaining to
 553      this particular media.  On Linux, for example, the floppy drive would
 554      not even be accessed for the whole verification.
 555
 556      The code was using fsync only when the ioctl is unavailable, but
 557      Marty Leisner says that the ioctl does not work when not preceded by
 558      fsync.  So, until we know better, or maybe to please Marty, let's do it
 559      the unbelievable way :-).  */
 560
 561 #if HAVE_FSYNC
 562   fsync (archive);
 563 #endif
 564 #ifdef FDFLUSH
 565   ioctl (archive, FDFLUSH);
 566 #endif
 567
 568 #ifdef MTIOCTOP
 569   {
 570     struct mtop operation;
 571     int status;
 572
 573     operation.mt_op = MTBSF;
 574     operation.mt_count = 1;
 575     if (status = rmtioctl (archive, MTIOCTOP, (char *) &operation), status < 0)
 576       {
 577         if (errno != EIO
 578             || (status = rmtioctl (archive, MTIOCTOP, (char *) &operation),
 579                 status < 0))
 580           {
 581 #endif
 582             if (rmtlseek (archive, (off_t) 0, SEEK_SET) != 0)
 583               {
 584                 /* Lseek failed.  Try a different method.  */
 585                 seek_warn (archive_name_array[0]);
 586                 return;
 587               }
 588 #ifdef MTIOCTOP
 589           }
 590       }
 591   }
 592 #endif
 593
 594   access_mode = ACCESS_READ;
 595   now_verifying = 1;
 596
 597   flush_read ();
 598   while (1)
 599     {
 600       enum read_header status = read_header (&current_header,
 601                                              &current_stat_info,
 602                                              read_header_auto);
 603
 604       if (status == HEADER_FAILURE)
 605         {
 606           int counter = 0;
 607
 608           do
 609             {
 610               counter++;
 611               set_next_block_after (current_header);
 612               status = read_header (&current_header, &current_stat_info,
 613                                     read_header_auto);
 614             }
 615           while (status == HEADER_FAILURE);
 616
 617           ERROR ((0, 0,
 618                   ngettext ("VERIFY FAILURE: %d invalid header detected",
 619                             "VERIFY FAILURE: %d invalid headers detected",
 620                             counter), counter));
 621         }
 622       if (status == HEADER_END_OF_FILE)
 623         break;
 624       if (status == HEADER_ZERO_BLOCK)
 625         {
 626           set_next_block_after (current_header);
 627           if (!ignore_zeros_option)
 628             {
 629               char buf[UINTMAX_STRSIZE_BOUND];
 630
 631               status = read_header (&current_header, &current_stat_info,
 632                                     read_header_auto);
 633               if (status == HEADER_ZERO_BLOCK)
 634                 break;
 635               WARNOPT (WARN_ALONE_ZERO_BLOCK,
 636                        (0, 0, _("A lone zero block at %s"),
 637                         STRINGIFY_BIGINT (current_block_ordinal (), buf)));
 638             }
 639           continue;
 640         }
 641
 642       decode_header (current_header, &current_stat_info, &current_format, 1);
 643       diff_archive ();
 644       tar_stat_destroy (&current_stat_info);
 645     }
 646
 647   access_mode = ACCESS_WRITE;
 648   now_verifying = 0;
 649 }