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 #include <sys/cdefs.h>
  38 #ifndef lint
  39 __used static const char copyright[] =
  40 "@(#) Copyright (c) 1992, 1993\n\
  41         The Regents of the University of California.  All rights reserved.\n";
  42 #endif /* not lint */
  43
  44 #ifndef lint
  45 #if 0
  46 static const char sccsid[] = "@(#)pax.c 8.2 (Berkeley) 4/18/94";
  47 #else
  48 __used static const char rcsid[] = "$OpenBSD: pax.c,v 1.28 2005/08/04 10:02:44 mpf Exp $";
  49 #endif
  50 #endif /* not lint */
  51
  52 #include <stdio.h>
  53 #include <sys/types.h>
  54 #include <sys/param.h>
  55 #include <sys/stat.h>
  56 #include <sys/time.h>
  57 #include <sys/resource.h>
  58 #include <signal.h>
  59 #include <unistd.h>
  60 #include <stdlib.h>
  61 #include <string.h>
  62 #include <errno.h>
  63 #include <err.h>
  64 #include <fcntl.h>
  65 #include <paths.h>
  66 #include "pax.h"
  67 #include "extern.h"
  68 static int gen_init(void);
  69
  70 /*
  71  * PAX main routines, general globals and some simple start up routines
  72  */
  73
  74 /*
  75  * Variables that can be accessed by any routine within pax
  76  */
  77 int     act = DEFOP;            /* read/write/append/copy */
  78 const FSUB      *frmt = NULL;           /* archive format type */
  79 int     cflag;                  /* match all EXCEPT pattern/file */
  80 int     cwdfd;                  /* starting cwd */
  81 int     dflag;                  /* directory member match only  */
  82 int     iflag;                  /* interactive file/archive rename */
  83 int     kflag;                  /* do not overwrite existing files */
  84 int     lflag;                  /* use hard links when possible */
  85 int     nflag;                  /* select first archive member match */
  86 int     tflag;                  /* restore access time after read */
  87 int     uflag;                  /* ignore older modification time files */
  88 int     vflag;                  /* produce verbose output */
  89 int     Dflag;                  /* same as uflag except inode change time */
  90 int     Hflag;                  /* follow command line symlinks (write only) */
  91 int     Lflag;                  /* follow symlinks when writing */
  92 int     Xflag;                  /* archive files with same device id only */
  93 int     Yflag;                  /* same as Dflag except after name mode */
  94 int     Zflag;                  /* same as uflag except after name mode */
  95 int     zeroflag;               /* use \0 as pathname terminator */
  96 int     vfpart;                 /* is partial verbose output in progress */
  97 int     patime = 1;             /* preserve file access time */
  98 int     pmtime = 1;             /* preserve file modification times */
  99 int     nodirs;                 /* do not create directories as needed */
 100 int     pmode;                  /* preserve file mode bits */
 101 int     pids;                   /* preserve file uid/gid */
 102 int     rmleadslash = 0;        /* remove leading '/' from pathnames */
 103 int     secure = 1;             /* don't extract names that contain .. */
 104 int     exit_val;               /* exit value */
 105 int     docrc;                  /* check/create file crc */
 106 char    *dirptr;                /* destination dir in a copy */
 107 char    *ltmfrmt;               /* -v locale time format (if any) */
 108 char    *argv0;                 /* root of argv[0] */
 109 sigset_t s_mask;                /* signal mask for cleanup critical sect */
 110 FILE    *listf;                 /* file pointer to print file list to */
 111 char    *tempfile;              /* tempfile to use for mkstemp(3) */
 112 char    *tempbase;              /* basename of tempfile to use for mkstemp(3) */
 113
 114 /*
 115  *      PAX - Portable Archive Interchange
 116  *
 117  *      A utility to read, write, and write lists of the members of archive
 118  *      files and copy directory hierarchies. A variety of archive formats
 119  *      are supported (some are described in POSIX 1003.1 10.1):
 120  *
 121  *              ustar - 10.1.1 extended tar interchange format
 122  *              cpio  - 10.1.2 extended cpio interchange format
 123  *              tar - old BSD 4.3 tar format
 124  *              binary cpio - old cpio with binary header format
 125  *              sysVR4 cpio -  with and without CRC
 126  *
 127  * This version is a superset of IEEE Std 1003.2b-d3
 128  *
 129  * Summary of Extensions to the IEEE Standard:
 130  *
 131  * 1    READ ENHANCEMENTS
 132  * 1.1  Operations which read archives will continue to operate even when
 133  *      processing archives which may be damaged, truncated, or fail to meet
 134  *      format specs in several different ways. Damaged sections of archives
 135  *      are detected and avoided if possible. Attempts will be made to resync
 136  *      archive read operations even with badly damaged media.
 137  * 1.2  Blocksize requirements are not strictly enforced on archive read.
 138  *      Tapes which have variable sized records can be read without errors.
 139  * 1.3  The user can specify via the non-standard option flag -E if error
 140  *      resync operation should stop on a media error, try a specified number
 141  *      of times to correct, or try to correct forever.
 142  * 1.4  Sparse files (lseek holes) stored on the archive (but stored with blocks
 143  *      of all zeros will be restored with holes appropriate for the target
 144  *      filesystem
 145  * 1.5  The user is notified whenever something is found during archive
 146  *      read operations which violates spec (but the read will continue).
 147  * 1.6  Multiple archive volumes can be read and may span over different
 148  *      archive devices
 149  * 1.7  Rigidly restores all file attributes exactly as they are stored on the
 150  *      archive.
 151  * 1.8  Modification change time ranges can be specified via multiple -T
 152  *      options. These allow a user to select files whose modification time
 153  *      lies within a specific time range.
 154  * 1.9  Files can be selected based on owner (user name or uid) via one or more
 155  *      -U options.
 156  * 1.10 Files can be selected based on group (group name or gid) via one o
 157  *      more -G options.
 158  * 1.11 File modification time can be checked against existing file after
 159  *      name modification (-Z)
 160  *
 161  * 2    WRITE ENHANCEMENTS
 162  * 2.1  Write operation will stop instead of allowing a user to create a flawed
 163  *      flawed archive (due to any problem).
 164  * 2.2  Archives written by pax are forced to strictly conform to both the
 165  *      archive and pax the specific format specifications.
 166  * 2.3  Blocking size and format is rigidly enforced on writes.
 167  * 2.4  Formats which may exhibit header overflow problems (they have fields
 168  *      too small for large file systems, such as inode number storage), use
 169  *      routines designed to repair this problem. These techniques still
 170  *      conform to both pax and format specifications, but no longer truncate
 171  *      these fields. This removes any restrictions on using these archive
 172  *      formats on large file systems.
 173  * 2.5  Multiple archive volumes can be written and may span over different
 174  *      archive devices
 175  * 2.6  A archive volume record limit allows the user to specify the number
 176  *      of bytes stored on an archive volume. When reached the user is
 177  *      prompted for the next archive volume. This is specified with the
 178  *      non-standard -B flag. The limit is rounded up to the next blocksize.
 179  * 2.7  All archive padding during write use zero filled sections. This makes
 180  *      it much easier to pull data out of flawed archive during read
 181  *      operations.
 182  * 2.8  Access time reset with the -t applies to all file nodes (including
 183  *      directories).
 184  * 2.9  Symbolic links can be followed with -L (optional in the spec).
 185  * 2.10 Modification or inode change time ranges can be specified via
 186  *      multiple -T options. These allow a user to select files whose
 187  *      modification or inode change time lies within a specific time range.
 188  * 2.11 Files can be selected based on owner (user name or uid) via one or more
 189  *      -U options.
 190  * 2.12 Files can be selected based on group (group name or gid) via one o
 191  *      more -G options.
 192  * 2.13 Symlinks which appear on the command line can be followed (without
 193  *      following other symlinks; -H flag)
 194  *
 195  * 3    COPY ENHANCEMENTS
 196  * 3.1  Sparse files (lseek holes) can be copied without expanding the holes
 197  *      into zero filled blocks. The file copy is created with holes which are
 198  *      appropriate for the target filesystem
 199  * 3.2  Access time as well as modification time on copied file trees can be
 200  *      preserved with the appropriate -p options.
 201  * 3.3  Access time reset with the -t applies to all file nodes (including
 202  *      directories).
 203  * 3.4  Symbolic links can be followed with -L (optional in the spec).
 204  * 3.5  Modification or inode change time ranges can be specified via
 205  *      multiple -T options. These allow a user to select files whose
 206  *      modification or inode change time lies within a specific time range.
 207  * 3.6  Files can be selected based on owner (user name or uid) via one or more
 208  *      -U options.
 209  * 3.7  Files can be selected based on group (group name or gid) via one o
 210  *      more -G options.
 211  * 3.8  Symlinks which appear on the command line can be followed (without
 212  *      following other symlinks; -H flag)
 213  * 3.9  File inode change time can be checked against existing file before
 214  *      name modification (-D)
 215  * 3.10 File inode change time can be checked against existing file after
 216  *      name modification (-Y)
 217  * 3.11 File modification time can be checked against existing file after
 218  *      name modification (-Z)
 219  *
 220  * 4    GENERAL ENHANCEMENTS
 221  * 4.1  Internal structure is designed to isolate format dependent and
 222  *      independent functions. Formats are selected via a format driver table.
 223  *      This encourages the addition of new archive formats by only having to
 224  *      write those routines which id, read and write the archive header.
 225  */
 226
 227 /*
 228  * main()
 229  *      parse options, set up and operate as specified by the user.
 230  *      any operational flaw will set exit_val to non-zero
 231  * Return: 0 if ok, 1 otherwise
 232  */
 233
 234 int
 235 main(int argc, char **argv)
 236 {
 237         char *tmpdir;
 238         size_t tdlen;
 239
 240         listf = stderr;
 241         /*
 242          * Keep a reference to cwd, so we can always come back home.
 243          */
 244         cwdfd = open(".", O_RDONLY);
 245         if (cwdfd < 0) {
 246                 syswarn(1, errno, "Can't open current working directory.");
 247                 return(exit_val);
 248         }
 249
 250         if (updatepath() == -1)
 251                 return exit_val;
 252         /*
 253          * Where should we put temporary files?
 254          */
 255         if ((tmpdir = getenv("TMPDIR")) == NULL || *tmpdir == '\0')
 256                 tmpdir = _PATH_TMP;
 257         tdlen = strlen(tmpdir);
 258         while (tdlen > 0 && tmpdir[tdlen - 1] == '/')
 259                 tdlen--;
 260         tempfile = malloc(tdlen + 1 + sizeof(_TFILE_BASE));
 261         if (tempfile == NULL) {
 262                 paxwarn(1, "Cannot allocate memory for temp file name.");
 263                 return(exit_val);
 264         }
 265         if (tdlen)
 266                 memcpy(tempfile, tmpdir, tdlen);
 267         tempbase = tempfile + tdlen;
 268         *tempbase++ = '/';
 269
 270         /*
 271          * parse options, determine operational mode, general init
 272          */
 273         options(argc, argv);
 274         if ((gen_init() < 0) || (tty_init() < 0))
 275                 return(exit_val);
 276
 277         /*
 278          * select a primary operation mode
 279          */
 280         switch (act) {
 281         case EXTRACT:
 282                 extract();
 283                 break;
 284         case ARCHIVE:
 285                 archive();
 286                 break;
 287         case APPND:
 288                 if (gzip_program != NULL)
 289                         errx(1, "can not gzip while appending");
 290                 append();
 291                 break;
 292         case COPY:
 293                 copy();
 294                 break;
 295         default:
 296         case LIST:
 297                 list();
 298                 break;
 299         }
 300         return(exit_val);
 301 }
 302
 303 /*
 304  * sig_cleanup()
 305  *      when interrupted we try to do whatever delayed processing we can.
 306  *      This is not critical, but we really ought to limit our damage when we
 307  *      are aborted by the user.
 308  * Return:
 309  *      never....
 310  */
 311
 312 void
 313 sig_cleanup(int which_sig)
 314 {
 315         /* XXX signal races */
 316
 317         /*
 318          * restore modes and times for any dirs we may have created
 319          * or any dirs we may have read. Set vflag and vfpart so the user
 320          * will clearly see the message on a line by itself.
 321          */
 322         vflag = vfpart = 1;
 323         if (which_sig == SIGXCPU)
 324                 paxwarn(0, "Cpu time limit reached, cleaning up.");
 325         else
 326                 paxwarn(0, "Signal caught, cleaning up.");
 327
 328         ar_close();
 329         proc_dir();
 330         if (tflag)
 331                 atdir_end();
 332         exit(1);
 333 }
 334
 335 /*
 336  * gen_init()
 337  *      general setup routines. Not all are required, but they really help
 338  *      when dealing with a medium to large sized archives.
 339  */
 340
 341 static int
 342 gen_init(void)
 343 {
 344         struct rlimit reslimit;
 345         struct sigaction n_hand;
 346         struct sigaction o_hand;
 347
 348         /*
 349          * Really needed to handle large archives. We can run out of memory for
 350          * internal tables really fast when we have a whole lot of files...
 351          */
 352         if (getrlimit(RLIMIT_DATA , &reslimit) == 0){
 353                 reslimit.rlim_cur = reslimit.rlim_max;
 354                 (void)setrlimit(RLIMIT_DATA , &reslimit);
 355         }
 356
 357         /*
 358          * should file size limits be waived? if the os limits us, this is
 359          * needed if we want to write a large archive
 360          */
 361         if (getrlimit(RLIMIT_FSIZE , &reslimit) == 0){
 362                 reslimit.rlim_cur = reslimit.rlim_max;
 363                 (void)setrlimit(RLIMIT_FSIZE , &reslimit);
 364         }
 365
 366         /*
 367          * increase the size the stack can grow to
 368          */
 369         if (getrlimit(RLIMIT_STACK , &reslimit) == 0){
 370                 reslimit.rlim_cur = reslimit.rlim_max;
 371                 (void)setrlimit(RLIMIT_STACK , &reslimit);
 372         }
 373
 374         /*
 375          * not really needed, but doesn't hurt
 376          */
 377         if (getrlimit(RLIMIT_RSS , &reslimit) == 0){
 378                 reslimit.rlim_cur = reslimit.rlim_max;
 379                 (void)setrlimit(RLIMIT_RSS , &reslimit);
 380         }
 381
 382         /*
 383          * Handle posix locale
 384          *
 385          * set user defines time printing format for -v option
 386          */
 387         ltmfrmt = getenv("LC_TIME");
 388
 389         /*
 390          * signal handling to reset stored directory times and modes. Since
 391          * we deal with broken pipes via failed writes we ignore it. We also
 392          * deal with any file size limit through failed writes. Cpu time
 393          * limits are caught and a cleanup is forced.
 394          */
 395         if ((sigemptyset(&s_mask) < 0) || (sigaddset(&s_mask, SIGTERM) < 0) ||
 396             (sigaddset(&s_mask,SIGINT) < 0)||(sigaddset(&s_mask,SIGHUP) < 0) ||
 397             (sigaddset(&s_mask,SIGPIPE) < 0)||(sigaddset(&s_mask,SIGQUIT)<0) ||
 398             (sigaddset(&s_mask,SIGXCPU) < 0)||(sigaddset(&s_mask,SIGXFSZ)<0)) {
 399                 paxwarn(1, "Unable to set up signal mask");
 400                 return(-1);
 401         }
 402         memset(&n_hand, 0, sizeof n_hand);
 403         n_hand.sa_mask = s_mask;
 404         n_hand.sa_flags = 0;
 405         n_hand.sa_handler = sig_cleanup;
 406
 407         if ((sigaction(SIGHUP, &n_hand, &o_hand) < 0) &&
 408             (o_hand.sa_handler == SIG_IGN) &&
 409             (sigaction(SIGHUP, &o_hand, &o_hand) < 0))
 410                 goto out;
 411
 412         if ((sigaction(SIGTERM, &n_hand, &o_hand) < 0) &&
 413             (o_hand.sa_handler == SIG_IGN) &&
 414             (sigaction(SIGTERM, &o_hand, &o_hand) < 0))
 415                 goto out;
 416
 417         if ((sigaction(SIGINT, &n_hand, &o_hand) < 0) &&
 418             (o_hand.sa_handler == SIG_IGN) &&
 419             (sigaction(SIGINT, &o_hand, &o_hand) < 0))
 420                 goto out;
 421
 422         if ((sigaction(SIGQUIT, &n_hand, &o_hand) < 0) &&
 423             (o_hand.sa_handler == SIG_IGN) &&
 424             (sigaction(SIGQUIT, &o_hand, &o_hand) < 0))
 425                 goto out;
 426
 427         if ((sigaction(SIGXCPU, &n_hand, &o_hand) < 0) &&
 428             (o_hand.sa_handler == SIG_IGN) &&
 429             (sigaction(SIGXCPU, &o_hand, &o_hand) < 0))
 430                 goto out;
 431
 432         n_hand.sa_handler = SIG_IGN;
 433         if ((sigaction(SIGPIPE, &n_hand, &o_hand) < 0) ||
 434             (sigaction(SIGXFSZ, &n_hand, &o_hand) < 0))
 435                 goto out;
 436         return(0);
 437
 438     out:
 439         syswarn(1, errno, "Unable to set up signal handler");
 440         return(-1);
 441 }