pax.c revision 1.30
1 /* $NetBSD: pax.c,v 1.30 2003/10/13 07:41:22 agc 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 #include <sys/cdefs.h> 37 #if defined(__COPYRIGHT) && !defined(lint) 38 __COPYRIGHT("@(#) Copyright (c) 1992, 1993\n\ 39 The Regents of the University of California. All rights reserved.\n"); 40 #endif /* not lint */ 41 42 #if defined(__RCSID) && !defined(lint) 43 #if 0 44 static char sccsid[] = "@(#)pax.c 8.2 (Berkeley) 4/18/94"; 45 #else 46 __RCSID("$NetBSD: pax.c,v 1.30 2003/10/13 07:41:22 agc Exp $"); 47 #endif 48 #endif /* not lint */ 49 50 #include <sys/types.h> 51 #include <sys/param.h> 52 #include <sys/stat.h> 53 #include <sys/time.h> 54 #include <sys/resource.h> 55 #include <stdio.h> 56 #include <fcntl.h> 57 #include <signal.h> 58 #include <unistd.h> 59 #include <stdlib.h> 60 #include <string.h> 61 #include <errno.h> 62 #include <paths.h> 63 #include "pax.h" 64 #include "extern.h" 65 static int gen_init(void); 66 67 /* 68 * PAX main routines, general globals and some simple start up routines 69 */ 70 71 /* 72 * Variables that can be accessed by any routine within pax 73 */ 74 int act = ERROR; /* read/write/append/copy */ 75 FSUB *frmt = NULL; /* archive format type */ 76 int cflag; /* match all EXCEPT pattern/file */ 77 int cwdfd; /* starting cwd */ 78 int dflag; /* directory member match only */ 79 int iflag; /* interactive file/archive rename */ 80 int jflag; /* pass through bzip2 */ 81 int kflag; /* do not overwrite existing files */ 82 int lflag; /* use hard links when possible */ 83 int nflag; /* select first archive member match */ 84 int tflag; /* restore access time after read */ 85 int uflag; /* ignore older modification time files */ 86 int vflag; /* produce verbose output */ 87 int zflag; /* use gzip */ 88 int Aflag; /* honor absolute path */ 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 Mflag; /* treat stdin as an mtree(8) specfile */ 93 int Xflag; /* archive files with same device id only */ 94 int Yflag; /* same as Dflg except after name mode */ 95 int Zflag; /* same as uflg except after name mode */ 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 pfflags = 1; /* preserve file flags */ 101 int pmode; /* preserve file mode bits */ 102 int pids; /* preserve file uid/gid */ 103 int rmleadslash = 0; /* remove leading '/' from pathnames */ 104 int exit_val; /* exit value */ 105 int docrc; /* check/create file crc */ 106 int to_stdout; /* extract to stdout */ 107 char *dirptr; /* destination dir in a copy */ 108 char *ltmfrmt; /* -v locale time format (if any) */ 109 char *argv0; /* root of argv[0] */ 110 sigset_t s_mask; /* signal mask for cleanup critical sect */ 111 FILE *listf; /* file pointer to print file list to */ 112 char *tempfile; /* tempfile to use for mkstemp(3) */ 113 char *tempbase; /* basename of tempfile to use for mkstemp(3) */ 114 int forcelocal; /* force local operation even if the name 115 * contains a : 116 */ 117 int secure = 1; /* don't extract names that contain .. */ 118 119 /* 120 * PAX - Portable Archive Interchange 121 * 122 * A utility to read, write, and write lists of the members of archive 123 * files and copy directory hierarchies. A variety of archive formats 124 * are supported (some are described in POSIX 1003.1 10.1): 125 * 126 * ustar - 10.1.1 extended tar interchange format 127 * cpio - 10.1.2 extended cpio interchange format 128 * tar - old BSD 4.3 tar format 129 * binary cpio - old cpio with binary header format 130 * sysVR4 cpio - with and without CRC 131 * 132 * This version is a superset of IEEE Std 1003.2b-d3 133 * 134 * Summary of Extensions to the IEEE Standard: 135 * 136 * 1 READ ENHANCEMENTS 137 * 1.1 Operations which read archives will continue to operate even when 138 * processing archives which may be damaged, truncated, or fail to meet 139 * format specs in several different ways. Damaged sections of archives 140 * are detected and avoided if possible. Attempts will be made to resync 141 * archive read operations even with badly damaged media. 142 * 1.2 Blocksize requirements are not strictly enforced on archive read. 143 * Tapes which have variable sized records can be read without errors. 144 * 1.3 The user can specify via the non-standard option flag -E if error 145 * resync operation should stop on a media error, try a specified number 146 * of times to correct, or try to correct forever. 147 * 1.4 Sparse files (lseek holes) stored on the archive (but stored with blocks 148 * of all zeros will be restored with holes appropriate for the target 149 * filesystem 150 * 1.5 The user is notified whenever something is found during archive 151 * read operations which violates spec (but the read will continue). 152 * 1.6 Multiple archive volumes can be read and may span over different 153 * archive devices 154 * 1.7 Rigidly restores all file attributes exactly as they are stored on the 155 * archive. 156 * 1.8 Modification change time ranges can be specified via multiple -T 157 * options. These allow a user to select files whose modification time 158 * lies within a specific time range. 159 * 1.9 Files can be selected based on owner (user name or uid) via one or more 160 * -U options. 161 * 1.10 Files can be selected based on group (group name or gid) via one o 162 * more -G options. 163 * 1.11 File modification time can be checked against existing file after 164 * name modification (-Z) 165 * 166 * 2 WRITE ENHANCEMENTS 167 * 2.1 Write operation will stop instead of allowing a user to create a flawed 168 * flawed archive (due to any problem). 169 * 2.2 Archives written by pax are forced to strictly conform to both the 170 * archive and pax the spceific format specifications. 171 * 2.3 Blocking size and format is rigidly enforced on writes. 172 * 2.4 Formats which may exhibit header overflow problems (they have fields 173 * too small for large file systems, such as inode number storage), use 174 * routines designed to repair this problem. These techniques still 175 * conform to both pax and format specifications, but no longer truncate 176 * these fields. This removes any restrictions on using these archive 177 * formats on large file systems. 178 * 2.5 Multiple archive volumes can be written and may span over different 179 * archive devices 180 * 2.6 A archive volume record limit allows the user to specify the number 181 * of bytes stored on an archive volume. When reached the user is 182 * prompted for the next archive volume. This is specified with the 183 * non-standard -B flag. The limit is rounded up to the next blocksize. 184 * 2.7 All archive padding during write use zero filled sections. This makes 185 * it much easier to pull data out of flawed archive during read 186 * operations. 187 * 2.8 Access time reset with the -t applies to all file nodes (including 188 * directories). 189 * 2.9 Symbolic links can be followed with -L (optional in the spec). 190 * 2.10 Modification or inode change time ranges can be specified via 191 * multiple -T options. These allow a user to select files whose 192 * modification or inode change time lies within a specific time range. 193 * 2.11 Files can be selected based on owner (user name or uid) via one or more 194 * -U options. 195 * 2.12 Files can be selected based on group (group name or gid) via one o 196 * more -G options. 197 * 2.13 Symlinks which appear on the command line can be followed (without 198 * following other symlinks; -H flag) 199 * 200 * 3 COPY ENHANCEMENTS 201 * 3.1 Sparse files (lseek holes) can be copied without expanding the holes 202 * into zero filled blocks. The file copy is created with holes which are 203 * appropriate for the target filesystem 204 * 3.2 Access time as well as modification time on copied file trees can be 205 * preserved with the appropriate -p options. 206 * 3.3 Access time reset with the -t applies to all file nodes (including 207 * directories). 208 * 3.4 Symbolic links can be followed with -L (optional in the spec). 209 * 3.5 Modification or inode change time ranges can be specified via 210 * multiple -T options. These allow a user to select files whose 211 * modification or inode change time lies within a specific time range. 212 * 3.6 Files can be selected based on owner (user name or uid) via one or more 213 * -U options. 214 * 3.7 Files can be selected based on group (group name or gid) via one o 215 * more -G options. 216 * 3.8 Symlinks which appear on the command line can be followed (without 217 * following other symlinks; -H flag) 218 * 3.9 File inode change time can be checked against existing file before 219 * name modification (-D) 220 * 3.10 File inode change time can be checked against existing file after 221 * name modification (-Y) 222 * 3.11 File modification time can be checked against existing file after 223 * name modification (-Z) 224 * 225 * 4 GENERAL ENHANCEMENTS 226 * 4.1 Internal structure is designed to isolate format dependent and 227 * independent functions. Formats are selected via a format driver table. 228 * This encourages the addition of new archive formats by only having to 229 * write those routines which id, read and write the archive header. 230 */ 231 232 /* 233 * main() 234 * parse options, set up and operate as specified by the user. 235 * any operational flaw will set exit_val to non-zero 236 * Return: 0 if ok, 1 otherwise 237 */ 238 239 int 240 main(int argc, char **argv) 241 { 242 char *tmpdir; 243 size_t tdlen; 244 245 setprogname(argv[0]); 246 247 listf = stderr; 248 249 /* 250 * Keep a reference to cwd, so we can always come back home. 251 */ 252 cwdfd = open(".", O_RDONLY); 253 if (cwdfd < 0) { 254 syswarn(0, errno, "Can't open current working directory."); 255 return(exit_val); 256 } 257 258 /* 259 * Where should we put temporary files? 260 */ 261 if ((tmpdir = getenv("TMPDIR")) == NULL || *tmpdir == '\0') 262 tmpdir = _PATH_TMP; 263 tdlen = strlen(tmpdir); 264 while(tdlen > 0 && tmpdir[tdlen - 1] == '/') 265 tdlen--; 266 tempfile = malloc(tdlen + 1 + sizeof(_TFILE_BASE)); 267 if (tempfile == NULL) { 268 tty_warn(1, "Cannot allocate memory for temp file name."); 269 return(exit_val); 270 } 271 if (tdlen) 272 memcpy(tempfile, tmpdir, tdlen); 273 tempbase = tempfile + tdlen; 274 *tempbase++ = '/'; 275 276 /* 277 * parse options, determine operational mode, general init 278 */ 279 options(argc, argv); 280 if ((gen_init() < 0) || (tty_init() < 0)) 281 return(exit_val); 282 283 (void)time(&starttime); 284 #ifdef SIGINFO 285 (void)signal(SIGINFO, ar_summary); 286 #endif 287 /* 288 * select a primary operation mode 289 */ 290 switch(act) { 291 case EXTRACT: 292 extract(); 293 break; 294 case ARCHIVE: 295 archive(); 296 break; 297 case APPND: 298 if (gzip_program != NULL) 299 err(1, "cannot gzip while appending"); 300 append(); 301 /* 302 * Check if we tried to append on an empty file and 303 * turned into ARCHIVE mode. 304 */ 305 if (act == -ARCHIVE) { 306 act = ARCHIVE; 307 archive(); 308 } 309 break; 310 case COPY: 311 copy(); 312 break; 313 default: 314 case LIST: 315 list(); 316 break; 317 } 318 return(exit_val); 319 } 320 321 /* 322 * sig_cleanup() 323 * when interrupted we try to do whatever delayed processing we can. 324 * This is not critical, but we really ought to limit our damage when we 325 * are aborted by the user. 326 * Return: 327 * never.... 328 */ 329 330 void 331 sig_cleanup(int which_sig) 332 { 333 /* 334 * restore modes and times for any dirs we may have created 335 * or any dirs we may have read. Set vflag and vfpart so the user 336 * will clearly see the message on a line by itself. 337 */ 338 vflag = vfpart = 1; 339 if (which_sig == SIGXCPU) 340 tty_warn(0, "Cpu time limit reached, cleaning up."); 341 else 342 tty_warn(0, "Signal caught, cleaning up."); 343 344 ar_close(); 345 proc_dir(); 346 if (tflag) 347 atdir_end(); 348 exit(1); 349 } 350 351 /* 352 * gen_init() 353 * general setup routines. Not all are required, but they really help 354 * when dealing with a medium to large sized archives. 355 */ 356 357 static int 358 gen_init(void) 359 { 360 struct rlimit reslimit; 361 struct sigaction n_hand; 362 struct sigaction o_hand; 363 364 /* 365 * Really needed to handle large archives. We can run out of memory for 366 * internal tables really fast when we have a whole lot of files... 367 */ 368 if (getrlimit(RLIMIT_DATA , &reslimit) == 0){ 369 reslimit.rlim_cur = reslimit.rlim_max; 370 (void)setrlimit(RLIMIT_DATA , &reslimit); 371 } 372 373 /* 374 * should file size limits be waived? if the os limits us, this is 375 * needed if we want to write a large archive 376 */ 377 if (getrlimit(RLIMIT_FSIZE , &reslimit) == 0){ 378 reslimit.rlim_cur = reslimit.rlim_max; 379 (void)setrlimit(RLIMIT_FSIZE , &reslimit); 380 } 381 382 /* 383 * increase the size the stack can grow to 384 */ 385 if (getrlimit(RLIMIT_STACK , &reslimit) == 0){ 386 reslimit.rlim_cur = reslimit.rlim_max; 387 (void)setrlimit(RLIMIT_STACK , &reslimit); 388 } 389 390 #ifdef RLIMIT_RSS 391 /* 392 * not really needed, but doesn't hurt 393 */ 394 if (getrlimit(RLIMIT_RSS , &reslimit) == 0){ 395 reslimit.rlim_cur = reslimit.rlim_max; 396 (void)setrlimit(RLIMIT_RSS , &reslimit); 397 } 398 #endif 399 400 /* 401 * Handle posix locale 402 * 403 * set user defines time printing format for -v option 404 */ 405 ltmfrmt = getenv("LC_TIME"); 406 407 /* 408 * signal handling to reset stored directory times and modes. Since 409 * we deal with broken pipes via failed writes we ignore it. We also 410 * deal with any file size limit thorugh failed writes. Cpu time 411 * limits are caught and a cleanup is forced. 412 */ 413 if ((sigemptyset(&s_mask) < 0) || (sigaddset(&s_mask, SIGTERM) < 0) || 414 (sigaddset(&s_mask,SIGINT) < 0)||(sigaddset(&s_mask,SIGHUP) < 0) || 415 (sigaddset(&s_mask,SIGPIPE) < 0)||(sigaddset(&s_mask,SIGQUIT)<0) || 416 (sigaddset(&s_mask,SIGXCPU) < 0)||(sigaddset(&s_mask,SIGXFSZ)<0)) { 417 tty_warn(1, "Unable to set up signal mask"); 418 return(-1); 419 } 420 memset(&n_hand, 0, sizeof n_hand); 421 n_hand.sa_mask = s_mask; 422 n_hand.sa_flags = 0; 423 n_hand.sa_handler = sig_cleanup; 424 425 if ((sigaction(SIGHUP, &n_hand, &o_hand) < 0) && 426 (o_hand.sa_handler == SIG_IGN) && 427 (sigaction(SIGHUP, &o_hand, &o_hand) < 0)) 428 goto out; 429 430 if ((sigaction(SIGTERM, &n_hand, &o_hand) < 0) && 431 (o_hand.sa_handler == SIG_IGN) && 432 (sigaction(SIGTERM, &o_hand, &o_hand) < 0)) 433 goto out; 434 435 if ((sigaction(SIGINT, &n_hand, &o_hand) < 0) && 436 (o_hand.sa_handler == SIG_IGN) && 437 (sigaction(SIGINT, &o_hand, &o_hand) < 0)) 438 goto out; 439 440 if ((sigaction(SIGQUIT, &n_hand, &o_hand) < 0) && 441 (o_hand.sa_handler == SIG_IGN) && 442 (sigaction(SIGQUIT, &o_hand, &o_hand) < 0)) 443 goto out; 444 445 if ((sigaction(SIGXCPU, &n_hand, &o_hand) < 0) && 446 (o_hand.sa_handler == SIG_IGN) && 447 (sigaction(SIGXCPU, &o_hand, &o_hand) < 0)) 448 goto out; 449 450 n_hand.sa_handler = SIG_IGN; 451 if ((sigaction(SIGPIPE, &n_hand, &o_hand) < 0) || 452 (sigaction(SIGXFSZ, &n_hand, &o_hand) < 0)) 453 goto out; 454 return(0); 455 456 out: 457 syswarn(1, errno, "Unable to set up signal handler"); 458 return(-1); 459 } 460
Indexes created Mon Sep 29 21:09:56 GMT 2025