pax/pax.c

   1 /*      $OpenBSD: pax.c,v 1.28 2005/08/04 10:02:44 mpf Exp $    */
   2 /*      $NetBSD: pax.c,v 1.5 1996/03/26 23:54:20 mrg Exp $      */
   3
   4 /*-
   5  * Copyright (c) 1992 Keith Muller.
   6  * Copyright (c) 1992, 1993
   7  *      The Regents of the University of California.  All rights reserved.
   8  *
   9  * This code is derived from software contributed to Berkeley by
  10  * Keith Muller of the University of California, San Diego.
  11  *
  12  * Redistribution and use in source and binary forms, with or without
  13  * modification, are permitted provided that the following conditions
  14  * are met:
  15  * 1. Redistributions of source code must retain the above copyright
  16  *    notice, this list of conditions and the following disclaimer.
  17  * 2. Redistributions in binary form must reproduce the above copyright
  18  *    notice, this list of conditions and the following disclaimer in the
  19  *    documentation and/or other materials provided with the distribution.
  20  * 3. Neither the name of the University nor the names of its contributors
  21  *    may be used to endorse or promote products derived from this software
  22  *    without specific prior written permission.
  23  *
  24  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  25  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  26  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  27  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  28  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  29  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  30  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  31  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  32  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  33  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  34  * SUCH DAMAGE.
  35  */
  36
  37 #ifndef lint
  38 static const char copyright[] =
  39 "@(#) Copyright (c) 1992, 1993\n\
  40         The Regents of the University of California.  All rights reserved.\n";
  41 #endif /* not lint */
  42
  43 #ifndef lint
  44 #if 0
  45 static const char sccsid[] = "@(#)pax.c 8.2 (Berkeley) 4/18/94";
  46 #else
  47 static const char rcsid[] = "$OpenBSD: pax.c,v 1.28 2005/08/04 10:02:44 mpf Exp $";
  48 #endif
  49 #endif /* not lint */
  50
  51 #include <stdio.h>
  52 #include <sys/types.h>
  53 #include <sys/param.h>
  54 #include <sys/stat.h>
  55 #include <sys/time.h>
  56 #include <sys/resource.h>
  57 #include <signal.h>
  58 #include <unistd.h>
  59 #include <stdlib.h>
  60 #include <string.h>
  61 #include <errno.h>
  62 #include <err.h>
  63 #include <fcntl.h>
  64 #include <paths.h>
  65 #include "pax.h"
  66 #include "extern.h"
  67 static int gen_init(void);
  68
  69 /*
  70  * PAX main routines, general globals and some simple start up routines
  71  */
  72
  73 /*
  74  * Variables that can be accessed by any routine within pax
  75  */
  76 int     act = DEFOP;            /* read/write/append/copy */
  77 FSUB    *frmt = NULL;           /* archive format type */
  78 int     cflag;                  /* match all EXCEPT pattern/file */
  79 int     cwdfd;                  /* starting cwd */
  80 int     dflag;                  /* directory member match only  */
  81 int     iflag;                  /* interactive file/archive rename */
  82 int     kflag;                  /* do not overwrite existing files */
  83 int     lflag;                  /* use hard links when possible */
  84 int     nflag;                  /* select first archive member match */
  85 int     tflag;                  /* restore access time after read */
  86 int     uflag;                  /* ignore older modification time files */
  87 int     vflag;                  /* produce verbose output */
  88 int     Dflag;                  /* same as uflag except inode change time */
  89 int     Hflag;                  /* follow command line symlinks (write only) */
  90 int     Lflag;                  /* follow symlinks when writing */
  91 int     Xflag;                  /* archive files with same device id only */
  92 int     Yflag;                  /* same as Dflag except after name mode */
  93 int     Zflag;                  /* same as uflag except after name mode */
  94 int     zeroflag;               /* use \0 as pathname terminator */
  95 int     vfpart;                 /* is partial verbose output in progress */
  96 int     patime = 1;             /* preserve file access time */
  97 int     pmtime = 1;             /* preserve file modification times */
  98 int     nodirs;                 /* do not create directories as needed */
  99 int     pmode;                  /* preserve file mode bits */
 100 int     pids;                   /* preserve file uid/gid */
 101 int     rmleadslash = 0;        /* remove leading '/' from pathnames */
 102 int     secure = 1;             /* don't extract names that contain .. */
 103 int     exit_val;               /* exit value */
 104 int     docrc;                  /* check/create file crc */
 105 char    *dirptr;                /* destination dir in a copy */
 106 char    *ltmfrmt;               /* -v locale time format (if any) */
 107 char    *argv0;                 /* root of argv[0] */
 108 sigset_t s_mask;                /* signal mask for cleanup critical sect */
 109 FILE    *listf;                 /* file pointer to print file list to */
 110 char    *tempfile;              /* tempfile to use for mkstemp(3) */
 111 char    *tempbase;              /* basename of tempfile to use for mkstemp(3) */
 112
 113 /*
 114  *      PAX - Portable Archive Interchange
 115  *
 116  *      A utility to read, write, and write lists of the members of archive
 117  *      files and copy directory hierarchies. A variety of archive formats
 118  *      are supported (some are described in POSIX 1003.1 10.1):
 119  *
 120  *              ustar - 10.1.1 extended tar interchange format
 121  *              cpio  - 10.1.2 extended cpio interchange format
 122  *              tar - old BSD 4.3 tar format
 123  *              binary cpio - old cpio with binary header format
 124  *              sysVR4 cpio -  with and without CRC
 125  *
 126  * This version is a superset of IEEE Std 1003.2b-d3
 127  *
 128  * Summary of Extensions to the IEEE Standard:
 129  *
 130  * 1    READ ENHANCEMENTS
 131  * 1.1  Operations which read archives will continue to operate even when
 132  *      processing archives which may be damaged, truncated, or fail to meet
 133  *      format specs in several different ways. Damaged sections of archives
 134  *      are detected and avoided if possible. Attempts will be made to resync
 135  *      archive read operations even with badly damaged media.
 136  * 1.2  Blocksize requirements are not strictly enforced on archive read.
 137  *      Tapes which have variable sized records can be read without errors.
 138  * 1.3  The user can specify via the non-standard option flag -E if error
 139  *      resync operation should stop on a media error, try a specified number
 140  *      of times to correct, or try to correct forever.
 141  * 1.4  Sparse files (lseek holes) stored on the archive (but stored with blocks
 142  *      of all zeros will be restored with holes appropriate for the target
 143  *      filesystem
 144  * 1.5  The user is notified whenever something is found during archive
 145  *      read operations which violates spec (but the read will continue).
 146  * 1.6  Multiple archive volumes can be read and may span over different
 147  *      archive devices
 148  * 1.7  Rigidly restores all file attributes exactly as they are stored on the
 149  *      archive.
 150  * 1.8  Modification change time ranges can be specified via multiple -T
 151  *      options. These allow a user to select files whose modification time
 152  *      lies within a specific time range.
 153  * 1.9  Files can be selected based on owner (user name or uid) via one or more
 154  *      -U options.
 155  * 1.10 Files can be selected based on group (group name or gid) via one o
 156  *      more -G options.
 157  * 1.11 File modification time can be checked against existing file after
 158  *      name modification (-Z)
 159  *
 160  * 2    WRITE ENHANCEMENTS
 161  * 2.1  Write operation will stop instead of allowing a user to create a flawed
 162  *      flawed archive (due to any problem).
 163  * 2.2  Archives written by pax are forced to strictly conform to both the
 164  *      archive and pax the specific format specifications.
 165  * 2.3  Blocking size and format is rigidly enforced on writes.
 166  * 2.4  Formats which may exhibit header overflow problems (they have fields
 167  *      too small for large file systems, such as inode number storage), use
 168  *      routines designed to repair this problem. These techniques still
 169  *      conform to both pax and format specifications, but no longer truncate
 170  *      these fields. This removes any restrictions on using these archive
 171  *      formats on large file systems.
 172  * 2.5  Multiple archive volumes can be written and may span over different
 173  *      archive devices
 174  * 2.6  A archive volume record limit allows the user to specify the number
 175  *      of bytes stored on an archive volume. When reached the user is
 176  *      prompted for the next archive volume. This is specified with the
 177  *      non-standard -B flag. The limit is rounded up to the next blocksize.
 178  * 2.7  All archive padding during write use zero filled sections. This makes
 179  *      it much easier to pull data out of flawed archive during read
 180  *      operations.
 181  * 2.8  Access time reset with the -t applies to all file nodes (including
 182  *      directories).
 183  * 2.9  Symbolic links can be followed with -L (optional in the spec).
 184  * 2.10 Modification or inode change time ranges can be specified via
 185  *      multiple -T options. These allow a user to select files whose
 186  *      modification or inode change time lies within a specific time range.
 187  * 2.11 Files can be selected based on owner (user name or uid) via one or more
 188  *      -U options.
 189  * 2.12 Files can be selected based on group (group name or gid) via one o
 190  *      more -G options.
 191  * 2.13 Symlinks which appear on the command line can be followed (without
 192  *      following other symlinks; -H flag)
 193  *
 194  * 3    COPY ENHANCEMENTS
 195  * 3.1  Sparse files (lseek holes) can be copied without expanding the holes
 196  *      into zero filled blocks. The file copy is created with holes which are
 197  *      appropriate for the target filesystem
 198  * 3.2  Access time as well as modification time on copied file trees can be
 199  *      preserved with the appropriate -p options.
 200  * 3.3  Access time reset with the -t applies to all file nodes (including
 201  *      directories).
 202  * 3.4  Symbolic links can be followed with -L (optional in the spec).
 203  * 3.5  Modification or inode change time ranges can be specified via
 204  *      multiple -T options. These allow a user to select files whose
 205  *      modification or inode change time lies within a specific time range.
 206  * 3.6  Files can be selected based on owner (user name or uid) via one or more
 207  *      -U options.
 208  * 3.7  Files can be selected based on group (group name or gid) via one o
 209  *      more -G options.
 210  * 3.8  Symlinks which appear on the command line can be followed (without
 211  *      following other symlinks; -H flag)
 212  * 3.9  File inode change time can be checked against existing file before
 213  *      name modification (-D)
 214  * 3.10 File inode change time can be checked against existing file after
 215  *      name modification (-Y)
 216  * 3.11 File modification time can be checked against existing file after
 217  *      name modification (-Z)
 218  *
 219  * 4    GENERAL ENHANCEMENTS
 220  * 4.1  Internal structure is designed to isolate format dependent and
 221  *      independent functions. Formats are selected via a format driver table.
 222  *      This encourages the addition of new archive formats by only having to
 223  *      write those routines which id, read and write the archive header.
 224  */
 225
 226 /*
 227  * main()
 228  *      parse options, set up and operate as specified by the user.
 229  *      any operational flaw will set exit_val to non-zero
 230  * Return: 0 if ok, 1 otherwise
 231  */
 232
 233 int
 234 main(int argc, char **argv)
 235 {
 236         char *tmpdir;
 237         size_t tdlen;
 238
 239         listf = stderr;
 240         /*
 241          * Keep a reference to cwd, so we can always come back home.
 242          */
 243         cwdfd = open(".", O_RDONLY);
 244         if (cwdfd < 0) {
 245                 syswarn(1, errno, "Can't open current working directory.");
 246                 return(exit_val);
 247         }
 248
 249         if (updatepath() == -1)
 250                 return exit_val;
 251         /*
 252          * Where should we put temporary files?
 253          */
 254         if ((tmpdir = getenv("TMPDIR")) == NULL || *tmpdir == '\0')
 255                 tmpdir = _PATH_TMP;
 256         tdlen = strlen(tmpdir);
 257         while (tdlen > 0 && tmpdir[tdlen - 1] == '/')
 258                 tdlen--;
 259         tempfile = malloc(tdlen + 1 + sizeof(_TFILE_BASE));
 260         if (tempfile == NULL) {
 261                 paxwarn(1, "Cannot allocate memory for temp file name.");
 262                 return(exit_val);
 263         }
 264         if (tdlen)
 265                 memcpy(tempfile, tmpdir, tdlen);
 266         tempbase = tempfile + tdlen;
 267         *tempbase++ = '/';
 268
 269         /*
 270          * parse options, determine operational mode, general init
 271          */
 272         options(argc, argv);
 273         if ((gen_init() < 0) || (tty_init() < 0))
 274                 return(exit_val);
 275
 276         /*
 277          * select a primary operation mode
 278          */
 279         switch (act) {
 280         case EXTRACT:
 281                 extract();
 282                 break;
 283         case ARCHIVE:
 284                 archive();
 285                 break;
 286         case APPND:
 287                 if (gzip_program != NULL)
 288                         errx(1, "can not gzip while appending");
 289                 append();
 290                 break;
 291         case COPY:
 292                 copy();
 293                 break;
 294         default:
 295         case LIST:
 296                 list();
 297                 break;
 298         }
 299         return(exit_val);
 300 }
 301
 302 /*
 303  * sig_cleanup()
 304  *      when interrupted we try to do whatever delayed processing we can.
 305  *      This is not critical, but we really ought to limit our damage when we
 306  *      are aborted by the user.
 307  * Return:
 308  *      never....
 309  */
 310
 311 void
 312 sig_cleanup(int which_sig)
 313 {
 314         /* XXX signal races */
 315
 316         /*
 317          * restore modes and times for any dirs we may have created
 318          * or any dirs we may have read. Set vflag and vfpart so the user
 319          * will clearly see the message on a line by itself.
 320          */
 321         vflag = vfpart = 1;
 322         if (which_sig == SIGXCPU)
 323                 paxwarn(0, "Cpu time limit reached, cleaning up.");
 324         else
 325                 paxwarn(0, "Signal caught, cleaning up.");
 326
 327         ar_close();
 328         proc_dir();
 329         if (tflag)
 330                 atdir_end();
 331         exit(1);
 332 }
 333
 334 /*
 335  * gen_init()
 336  *      general setup routines. Not all are required, but they really help
 337  *      when dealing with a medium to large sized archives.
 338  */
 339
 340 static int
 341 gen_init(void)
 342 {
 343         struct rlimit reslimit;
 344         struct sigaction n_hand;
 345         struct sigaction o_hand;
 346
 347         /*
 348          * Really needed to handle large archives. We can run out of memory for
 349          * internal tables really fast when we have a whole lot of files...
 350          */
 351         if (getrlimit(RLIMIT_DATA , &reslimit) == 0){
 352                 reslimit.rlim_cur = reslimit.rlim_max;
 353                 (void)setrlimit(RLIMIT_DATA , &reslimit);
 354         }
 355
 356         /*
 357          * should file size limits be waived? if the os limits us, this is
 358          * needed if we want to write a large archive
 359          */
 360         if (getrlimit(RLIMIT_FSIZE , &reslimit) == 0){
 361                 reslimit.rlim_cur = reslimit.rlim_max;
 362                 (void)setrlimit(RLIMIT_FSIZE , &reslimit);
 363         }
 364
 365         /*
 366          * increase the size the stack can grow to
 367          */
 368         if (getrlimit(RLIMIT_STACK , &reslimit) == 0){
 369                 reslimit.rlim_cur = reslimit.rlim_max;
 370                 (void)setrlimit(RLIMIT_STACK , &reslimit);
 371         }
 372
 373         /*
 374          * not really needed, but doesn't hurt
 375          */
 376         if (getrlimit(RLIMIT_RSS , &reslimit) == 0){
 377                 reslimit.rlim_cur = reslimit.rlim_max;
 378                 (void)setrlimit(RLIMIT_RSS , &reslimit);
 379         }
 380
 381         /*
 382          * Handle posix locale
 383          *
 384          * set user defines time printing format for -v option
 385          */
 386         ltmfrmt = getenv("LC_TIME");
 387
 388         /*
 389          * signal handling to reset stored directory times and modes. Since
 390          * we deal with broken pipes via failed writes we ignore it. We also
 391          * deal with any file size limit through failed writes. Cpu time
 392          * limits are caught and a cleanup is forced.
 393          */
 394         if ((sigemptyset(&s_mask) < 0) || (sigaddset(&s_mask, SIGTERM) < 0) ||
 395             (sigaddset(&s_mask,SIGINT) < 0)||(sigaddset(&s_mask,SIGHUP) < 0) ||
 396             (sigaddset(&s_mask,SIGPIPE) < 0)||(sigaddset(&s_mask,SIGQUIT)<0) ||
 397             (sigaddset(&s_mask,SIGXCPU) < 0)||(sigaddset(&s_mask,SIGXFSZ)<0)) {
 398                 paxwarn(1, "Unable to set up signal mask");
 399                 return(-1);
 400         }
 401         memset(&n_hand, 0, sizeof n_hand);
 402         n_hand.sa_mask = s_mask;
 403         n_hand.sa_flags = 0;
 404         n_hand.sa_handler = sig_cleanup;
 405
 406         if ((sigaction(SIGHUP, &n_hand, &o_hand) < 0) &&
 407             (o_hand.sa_handler == SIG_IGN) &&
 408             (sigaction(SIGHUP, &o_hand, &o_hand) < 0))
 409                 goto out;
 410
 411         if ((sigaction(SIGTERM, &n_hand, &o_hand) < 0) &&
 412             (o_hand.sa_handler == SIG_IGN) &&
 413             (sigaction(SIGTERM, &o_hand, &o_hand) < 0))
 414                 goto out;
 415
 416         if ((sigaction(SIGINT, &n_hand, &o_hand) < 0) &&
 417             (o_hand.sa_handler == SIG_IGN) &&
 418             (sigaction(SIGINT, &o_hand, &o_hand) < 0))
 419                 goto out;
 420
 421         if ((sigaction(SIGQUIT, &n_hand, &o_hand) < 0) &&
 422             (o_hand.sa_handler == SIG_IGN) &&
 423             (sigaction(SIGQUIT, &o_hand, &o_hand) < 0))
 424                 goto out;
 425
 426         if ((sigaction(SIGXCPU, &n_hand, &o_hand) < 0) &&
 427             (o_hand.sa_handler == SIG_IGN) &&
 428             (sigaction(SIGXCPU, &o_hand, &o_hand) < 0))
 429                 goto out;
 430
 431         n_hand.sa_handler = SIG_IGN;
 432         if ((sigaction(SIGPIPE, &n_hand, &o_hand) < 0) ||
 433             (sigaction(SIGXFSZ, &n_hand, &o_hand) < 0))
 434                 goto out;
 435         return(0);
 436
 437     out:
 438         syswarn(1, errno, "Unable to set up signal handler");
 439         return(-1);
 440 }