pax.c revision 1.43
1 /* $NetBSD: pax.c,v 1.43 2007/07/03 10:20:09 lukem Exp $ */ 2 3 /*- 4 * Copyright (c) 1992 Keith Muller. 5 * Copyright (c) 1992, 1993 6 * The Regents of the University of California. All rights reserved. 7 * 8 * This code is derived from software contributed to Berkeley by 9 * Keith Muller of the University of California, San Diego. 10 * 11 * Redistribution and use in source and binary forms, with or without 12 * modification, are permitted provided that the following conditions 13 * are met: 14 * 1. Redistributions of source code must retain the above copyright 15 * notice, this list of conditions and the following disclaimer. 16 * 2. Redistributions in binary form must reproduce the above copyright 17 * notice, this list of conditions and the following disclaimer in the 18 * documentation and/or other materials provided with the distribution. 19 * 3. Neither the name of the University nor the names of its contributors 20 * may be used to endorse or promote products derived from this software 21 * without specific prior written permission. 22 * 23 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 24 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 25 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 26 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 27 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 28 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 29 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 30 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 31 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 32 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 33 * SUCH DAMAGE. 34 */ 35 36 #if HAVE_NBTOOL_CONFIG_H 37 #include "nbtool_config.h" 38 #endif 39 40 #include <sys/cdefs.h> 41 #if !defined(lint) 42 __COPYRIGHT("@(#) Copyright (c) 1992, 1993\n\ 43 The Regents of the University of California. All rights reserved.\n"); 44 #if 0 45 static char sccsid[] = "@(#)pax.c 8.2 (Berkeley) 4/18/94"; 46 #else 47 __RCSID("$NetBSD: pax.c,v 1.43 2007/07/03 10:20:09 lukem Exp $"); 48 #endif 49 #endif /* not lint */ 50 51 #include <sys/types.h> 52 #include <sys/param.h> 53 #include <sys/stat.h> 54 #include <sys/time.h> 55 #include <sys/resource.h> 56 #include <stdio.h> 57 #include <fcntl.h> 58 #include <signal.h> 59 #include <unistd.h> 60 #include <stdlib.h> 61 #include <string.h> 62 #include <errno.h> 63 #include <paths.h> 64 #include "pax.h" 65 #include "extern.h" 66 static int gen_init(void); 67 68 /* 69 * PAX main routines, general globals and some simple start up routines 70 */ 71 72 /* 73 * Variables that can be accessed by any routine within pax 74 */ 75 int act = ERROR; /* read/write/append/copy */ 76 FSUB *frmt = NULL; /* archive format type */ 77 int cflag; /* match all EXCEPT pattern/file */ 78 int cwdfd = -1; /* starting cwd */ 79 int dflag; /* directory member match only */ 80 int iflag; /* interactive file/archive rename */ 81 int jflag; /* pass through bzip2 */ 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 zflag; /* use gzip */ 89 int Aflag; /* honor absolute path */ 90 int Dflag; /* same as uflag except inode change time */ 91 int Hflag; /* follow command line symlinks (write only) */ 92 int Lflag; /* follow symlinks when writing */ 93 int Mflag; /* treat stdin as an mtree(8) specfile */ 94 int Vflag; /* produce somewhat verbose output (no listing) */ 95 int Xflag; /* archive files with same device id only */ 96 int Yflag; /* same as Dflg except after name mode */ 97 int Zflag; /* same as uflg except after name mode */ 98 int vfpart; /* is partial verbose output in progress */ 99 int patime = 1; /* preserve file access time */ 100 int pmtime = 1; /* preserve file modification times */ 101 int nodirs; /* do not create directories as needed */ 102 int pfflags = 1; /* preserve file flags */ 103 int pmode; /* preserve file mode bits */ 104 int pids; /* preserve file uid/gid */ 105 int rmleadslash = 0; /* remove leading '/' from pathnames */ 106 int exit_val; /* exit value */ 107 int docrc; /* check/create file crc */ 108 int to_stdout; /* extract to stdout */ 109 char *dirptr; /* destination dir in a copy */ 110 char *ltmfrmt; /* -v locale time format (if any) */ 111 const char *argv0; /* root of argv[0] */ 112 sigset_t s_mask; /* signal mask for cleanup critical sect */ 113 FILE *listf; /* file pointer to print file list to */ 114 char *tempfile; /* tempfile to use for mkstemp(3) */ 115 char *tempbase; /* basename of tempfile to use for mkstemp(3) */ 116 int forcelocal; /* force local operation even if the name 117 * contains a : 118 */ 119 int secure = 1; /* don't extract names that contain .. */ 120 121 /* 122 * PAX - Portable Archive Interchange 123 * 124 * A utility to read, write, and write lists of the members of archive 125 * files and copy directory hierarchies. A variety of archive formats 126 * are supported (some are described in POSIX 1003.1 10.1): 127 * 128 * ustar - 10.1.1 extended tar interchange format 129 * cpio - 10.1.2 extended cpio interchange format 130 * tar - old BSD 4.3 tar format 131 * binary cpio - old cpio with binary header format 132 * sysVR4 cpio - with and without CRC 133 * 134 * This version is a superset of IEEE Std 1003.2b-d3 135 * 136 * Summary of Extensions to the IEEE Standard: 137 * 138 * 1 READ ENHANCEMENTS 139 * 1.1 Operations which read archives will continue to operate even when 140 * processing archives which may be damaged, truncated, or fail to meet 141 * format specs in several different ways. Damaged sections of archives 142 * are detected and avoided if possible. Attempts will be made to resync 143 * archive read operations even with badly damaged media. 144 * 1.2 Blocksize requirements are not strictly enforced on archive read. 145 * Tapes which have variable sized records can be read without errors. 146 * 1.3 The user can specify via the non-standard option flag -E if error 147 * resync operation should stop on a media error, try a specified number 148 * of times to correct, or try to correct forever. 149 * 1.4 Sparse files (lseek holes) stored on the archive (but stored with blocks 150 * of all zeros will be restored with holes appropriate for the target 151 * filesystem 152 * 1.5 The user is notified whenever something is found during archive 153 * read operations which violates spec (but the read will continue). 154 * 1.6 Multiple archive volumes can be read and may span over different 155 * archive devices 156 * 1.7 Rigidly restores all file attributes exactly as they are stored on the 157 * archive. 158 * 1.8 Modification change time ranges can be specified via multiple -T 159 * options. These allow a user to select files whose modification time 160 * lies within a specific time range. 161 * 1.9 Files can be selected based on owner (user name or uid) via one or more 162 * -U options. 163 * 1.10 Files can be selected based on group (group name or gid) via one o 164 * more -G options. 165 * 1.11 File modification time can be checked against existing file after 166 * name modification (-Z) 167 * 168 * 2 WRITE ENHANCEMENTS 169 * 2.1 Write operation will stop instead of allowing a user to create a flawed 170 * flawed archive (due to any problem). 171 * 2.2 Archives written by pax are forced to strictly conform to both the 172 * archive and pax the specific format specifications. 173 * 2.3 Blocking size and format is rigidly enforced on writes. 174 * 2.4 Formats which may exhibit header overflow problems (they have fields 175 * too small for large file systems, such as inode number storage), use 176 * routines designed to repair this problem. These techniques still 177 * conform to both pax and format specifications, but no longer truncate 178 * these fields. This removes any restrictions on using these archive 179 * formats on large file systems. 180 * 2.5 Multiple archive volumes can be written and may span over different 181 * archive devices 182 * 2.6 A archive volume record limit allows the user to specify the number 183 * of bytes stored on an archive volume. When reached the user is 184 * prompted for the next archive volume. This is specified with the 185 * non-standard -B flag. The limit is rounded up to the next blocksize. 186 * 2.7 All archive padding during write use zero filled sections. This makes 187 * it much easier to pull data out of flawed archive during read 188 * operations. 189 * 2.8 Access time reset with the -t applies to all file nodes (including 190 * directories). 191 * 2.9 Symbolic links can be followed with -L (optional in the spec). 192 * 2.10 Modification or inode change time ranges can be specified via 193 * multiple -T options. These allow a user to select files whose 194 * modification or inode change time lies within a specific time range. 195 * 2.11 Files can be selected based on owner (user name or uid) via one or more 196 * -U options. 197 * 2.12 Files can be selected based on group (group name or gid) via one o 198 * more -G options. 199 * 2.13 Symlinks which appear on the command line can be followed (without 200 * following other symlinks; -H flag) 201 * 202 * 3 COPY ENHANCEMENTS 203 * 3.1 Sparse files (lseek holes) can be copied without expanding the holes 204 * into zero filled blocks. The file copy is created with holes which are 205 * appropriate for the target filesystem 206 * 3.2 Access time as well as modification time on copied file trees can be 207 * preserved with the appropriate -p options. 208 * 3.3 Access time reset with the -t applies to all file nodes (including 209 * directories). 210 * 3.4 Symbolic links can be followed with -L (optional in the spec). 211 * 3.5 Modification or inode change time ranges can be specified via 212 * multiple -T options. These allow a user to select files whose 213 * modification or inode change time lies within a specific time range. 214 * 3.6 Files can be selected based on owner (user name or uid) via one or more 215 * -U options. 216 * 3.7 Files can be selected based on group (group name or gid) via one o 217 * more -G options. 218 * 3.8 Symlinks which appear on the command line can be followed (without 219 * following other symlinks; -H flag) 220 * 3.9 File inode change time can be checked against existing file before 221 * name modification (-D) 222 * 3.10 File inode change time can be checked against existing file after 223 * name modification (-Y) 224 * 3.11 File modification time can be checked against existing file after 225 * name modification (-Z) 226 * 227 * 4 GENERAL ENHANCEMENTS 228 * 4.1 Internal structure is designed to isolate format dependent and 229 * independent functions. Formats are selected via a format driver table. 230 * This encourages the addition of new archive formats by only having to 231 * write those routines which id, read and write the archive header. 232 */ 233 234 /* 235 * main() 236 * parse options, set up and operate as specified by the user. 237 * any operational flaw will set exit_val to non-zero 238 * Return: 0 if ok, 1 otherwise 239 */ 240 241 int 242 main(int argc, char **argv) 243 { 244 const char *tmpdir; 245 size_t tdlen; 246 int rval; 247 248 setprogname(argv[0]); 249 250 listf = stderr; 251 252 /* 253 * parse options, determine operational mode 254 */ 255 options(argc, argv); 256 257 /* 258 * general init 259 */ 260 if ((gen_init() < 0) || (tty_init() < 0)) 261 return exit_val; 262 263 /* 264 * Keep a reference to cwd, so we can always come back home. 265 */ 266 cwdfd = open(".", O_RDONLY); 267 if (cwdfd < 0) { 268 syswarn(1, errno, "Can't open current working directory."); 269 return exit_val; 270 } 271 if (updatepath() == -1) 272 return exit_val; 273 274 /* 275 * Where should we put temporary files? 276 */ 277 if ((tmpdir = getenv("TMPDIR")) == NULL || *tmpdir == '\0') 278 tmpdir = _PATH_TMP; 279 tdlen = strlen(tmpdir); 280 while(tdlen > 0 && tmpdir[tdlen - 1] == '/') 281 tdlen--; 282 tempfile = malloc(tdlen + 1 + sizeof(_TFILE_BASE)); 283 if (tempfile == NULL) { 284 tty_warn(1, "Cannot allocate memory for temp file name."); 285 return exit_val; 286 } 287 if (tdlen) 288 memcpy(tempfile, tmpdir, tdlen); 289 tempbase = tempfile + tdlen; 290 *tempbase++ = '/'; 291 292 (void)time(&starttime); 293 #ifdef SIGINFO 294 (void)signal(SIGINFO, ar_summary); 295 #endif 296 /* 297 * select a primary operation mode 298 */ 299 switch (act) { 300 case EXTRACT: 301 rval = extract(); 302 break; 303 case ARCHIVE: 304 rval = archive(); 305 break; 306 case APPND: 307 if (gzip_program != NULL) 308 err(1, "cannot gzip while appending"); 309 rval = append(); 310 /* 311 * Check if we tried to append on an empty file and 312 * turned into ARCHIVE mode. 313 */ 314 if (act == -ARCHIVE) { 315 act = ARCHIVE; 316 rval = archive(); 317 } 318 break; 319 case COPY: 320 rval = copy(); 321 break; 322 default: 323 case LIST: 324 rval = list(); 325 break; 326 } 327 if (rval != 0) 328 exit_val = 1; 329 return exit_val; 330 } 331 332 /* 333 * sig_cleanup() 334 * when interrupted we try to do whatever delayed processing we can. 335 * This is not critical, but we really ought to limit our damage when we 336 * are aborted by the user. 337 * Return: 338 * never.... 339 */ 340 341 void 342 sig_cleanup(int which_sig) 343 { 344 struct sigaction n_hand; 345 sigset_t n_mask; 346 347 /* 348 * restore modes and times for any dirs we may have created 349 * or any dirs we may have read. Set vflag and vfpart so the user 350 * will clearly see the message on a line by itself. 351 */ 352 vflag = vfpart = 1; 353 #ifdef SIGXCPU 354 if (which_sig == SIGXCPU) 355 tty_warn(1, "CPU time limit reached, cleaning up."); 356 else 357 #endif 358 tty_warn(1, "Signal caught, cleaning up."); 359 360 /* delete any open temporary file */ 361 if (xtmp_name) 362 (void)unlink(xtmp_name); 363 ar_close(); 364 proc_dir(); 365 if (tflag) 366 atdir_end(); 367 368 memset(&n_hand, 0, sizeof n_hand); 369 sigemptyset(&n_hand.sa_mask); 370 n_hand.sa_handler = SIG_DFL; 371 if ((sigaction(which_sig, &n_hand, NULL) == 0) && 372 (sigemptyset(&n_mask) == 0) && 373 (sigaddset(&n_mask, which_sig) == 0) && 374 (sigprocmask(SIG_UNBLOCK, &n_mask, 0) == 0)) { 375 raise(which_sig); 376 } 377 exit(1); 378 } 379 380 /* 381 * gen_init() 382 * general setup routines. Not all are required, but they really help 383 * when dealing with a medium to large sized archives. 384 */ 385 386 static int 387 gen_init(void) 388 { 389 struct rlimit reslimit; 390 struct sigaction n_hand; 391 struct sigaction o_hand; 392 393 /* 394 * Really needed to handle large archives. We can run out of memory for 395 * internal tables really fast when we have a whole lot of files... 396 */ 397 if (getrlimit(RLIMIT_DATA , &reslimit) == 0){ 398 reslimit.rlim_cur = reslimit.rlim_max; 399 (void)setrlimit(RLIMIT_DATA , &reslimit); 400 } 401 402 /* 403 * should file size limits be waived? if the os limits us, this is 404 * needed if we want to write a large archive 405 */ 406 if (getrlimit(RLIMIT_FSIZE , &reslimit) == 0){ 407 reslimit.rlim_cur = reslimit.rlim_max; 408 (void)setrlimit(RLIMIT_FSIZE , &reslimit); 409 } 410 411 /* 412 * increase the size the stack can grow to 413 */ 414 if (getrlimit(RLIMIT_STACK , &reslimit) == 0){ 415 reslimit.rlim_cur = reslimit.rlim_max; 416 (void)setrlimit(RLIMIT_STACK , &reslimit); 417 } 418 419 #ifdef RLIMIT_RSS 420 /* 421 * not really needed, but doesn't hurt 422 */ 423 if (getrlimit(RLIMIT_RSS , &reslimit) == 0){ 424 reslimit.rlim_cur = reslimit.rlim_max; 425 (void)setrlimit(RLIMIT_RSS , &reslimit); 426 } 427 #endif 428 429 /* 430 * Handle posix locale 431 * 432 * set user defines time printing format for -v option 433 */ 434 ltmfrmt = getenv("LC_TIME"); 435 436 /* 437 * signal handling to reset stored directory times and modes. Since 438 * we deal with broken pipes via failed writes we ignore it. We also 439 * deal with any file size limit through failed writes. CPU time 440 * limits are caught and a cleanup is forced. 441 */ 442 if ((sigemptyset(&s_mask) < 0) || (sigaddset(&s_mask, SIGTERM) < 0) || 443 (sigaddset(&s_mask,SIGINT) < 0)||(sigaddset(&s_mask,SIGHUP) < 0) || 444 (sigaddset(&s_mask,SIGPIPE) < 0)||(sigaddset(&s_mask,SIGQUIT)<0)){ 445 tty_warn(1, "Unable to set up signal mask"); 446 return -1; 447 } 448 #ifdef SIGXCPU 449 if (sigaddset(&s_mask,SIGXCPU) < 0) { 450 tty_warn(1, "Unable to set up signal mask"); 451 return -1; 452 } 453 #endif 454 #ifdef SIGXFSZ 455 if (sigaddset(&s_mask,SIGXFSZ) < 0) { 456 tty_warn(1, "Unable to set up signal mask"); 457 return -1; 458 } 459 #endif 460 461 memset(&n_hand, 0, sizeof n_hand); 462 n_hand.sa_mask = s_mask; 463 n_hand.sa_flags = 0; 464 n_hand.sa_handler = sig_cleanup; 465 466 if ((sigaction(SIGHUP, &n_hand, &o_hand) < 0) && 467 (o_hand.sa_handler == SIG_IGN) && 468 (sigaction(SIGHUP, &o_hand, &o_hand) < 0)) 469 goto out; 470 471 if ((sigaction(SIGTERM, &n_hand, &o_hand) < 0) && 472 (o_hand.sa_handler == SIG_IGN) && 473 (sigaction(SIGTERM, &o_hand, &o_hand) < 0)) 474 goto out; 475 476 if ((sigaction(SIGINT, &n_hand, &o_hand) < 0) && 477 (o_hand.sa_handler == SIG_IGN) && 478 (sigaction(SIGINT, &o_hand, &o_hand) < 0)) 479 goto out; 480 481 if ((sigaction(SIGQUIT, &n_hand, &o_hand) < 0) && 482 (o_hand.sa_handler == SIG_IGN) && 483 (sigaction(SIGQUIT, &o_hand, &o_hand) < 0)) 484 goto out; 485 486 #ifdef SIGXCPU 487 if ((sigaction(SIGXCPU, &n_hand, &o_hand) < 0) && 488 (o_hand.sa_handler == SIG_IGN) && 489 (sigaction(SIGXCPU, &o_hand, &o_hand) < 0)) 490 goto out; 491 #endif 492 n_hand.sa_handler = SIG_IGN; 493 if (sigaction(SIGPIPE, &n_hand, &o_hand) < 0) 494 goto out; 495 #ifdef SIGXFSZ 496 if (sigaction(SIGXFSZ, &n_hand, &o_hand) < 0) 497 goto out; 498 #endif 499 return 0; 500 501 out: 502 syswarn(1, errno, "Unable to set up signal handler"); 503 return -1; 504 } 505
Indexes created Sun Sep 28 12:09:53 GMT 2025