Add the DragonFly cvs id and perform general cleanups on cvs/rcs/sccs ids. Most
[dragonfly.git] / bin / pax / pax.c
CommitLineData
984263bc
MD
1/*-
2 * Copyright (c) 1992 Keith Muller.
3 * Copyright (c) 1992, 1993
4 * The Regents of the University of California. All rights reserved.
5 *
6 * This code is derived from software contributed to Berkeley by
7 * Keith Muller of the University of California, San Diego.
8 *
9 * Redistribution and use in source and binary forms, with or without
10 * modification, are permitted provided that the following conditions
11 * are met:
12 * 1. Redistributions of source code must retain the above copyright
13 * notice, this list of conditions and the following disclaimer.
14 * 2. Redistributions in binary form must reproduce the above copyright
15 * notice, this list of conditions and the following disclaimer in the
16 * documentation and/or other materials provided with the distribution.
17 * 3. All advertising materials mentioning features or use of this software
18 * must display the following acknowledgement:
19 * This product includes software developed by the University of
20 * California, Berkeley and its contributors.
21 * 4. Neither the name of the University nor the names of its contributors
22 * may be used to endorse or promote products derived from this software
23 * without specific prior written permission.
24 *
25 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
26 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
27 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
28 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
29 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
30 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
31 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
32 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
33 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
34 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
35 * SUCH DAMAGE.
1de703da
MD
36 *
37 * @(#) Copyright (c) 1992, 1993 The Regents of the University of California. All rights reserved.
38 * @(#)pax.c 8.2 (Berkeley) 4/18/94
39 * $FreeBSD: src/bin/pax/pax.c,v 1.13.2.4 2002/11/07 15:29:53 imp Exp $
40 * $DragonFly: src/bin/pax/pax.c,v 1.2 2003/06/17 04:22:50 dillon Exp $
984263bc
MD
41 */
42
984263bc
MD
43#include <sys/types.h>
44#include <sys/stat.h>
45#include <sys/time.h>
46#include <sys/resource.h>
47#include <err.h>
48#include <errno.h>
49#include <fcntl.h>
50#include <locale.h>
51#include <paths.h>
52#include <signal.h>
53#include <stdio.h>
54#include <stdlib.h>
55#include <string.h>
56#include <unistd.h>
57#include "pax.h"
58#include "extern.h"
59static int gen_init __P((void));
60
61/*
62 * PAX main routines, general globals and some simple start up routines
63 */
64
65/*
66 * Variables that can be accessed by any routine within pax
67 */
68int act = DEFOP; /* read/write/append/copy */
69FSUB *frmt = NULL; /* archive format type */
70int cflag; /* match all EXCEPT pattern/file */
71int cwdfd; /* starting cwd */
72int dflag; /* directory member match only */
73int iflag; /* interactive file/archive rename */
74int kflag; /* do not overwrite existing files */
75int lflag; /* use hard links when possible */
76int nflag; /* select first archive member match */
77int tflag; /* restore access time after read */
78int uflag; /* ignore older modification time files */
79int vflag; /* produce verbose output */
80int Dflag; /* same as uflag except inode change time */
81int Hflag; /* follow command line symlinks (write only) */
82int Lflag; /* follow symlinks when writing */
83int Xflag; /* archive files with same device id only */
84int Yflag; /* same as Dflg except after name mode */
85int Zflag; /* same as uflg except after name mode */
86int vfpart; /* is partial verbose output in progress */
87int patime = 1; /* preserve file access time */
88int pmtime = 1; /* preserve file modification times */
89int nodirs; /* do not create directories as needed */
90int pmode; /* preserve file mode bits */
91int pids; /* preserve file uid/gid */
92int rmleadslash = 0; /* remove leading '/' from pathnames */
93int exit_val; /* exit value */
94int docrc; /* check/create file crc */
95char *dirptr; /* destination dir in a copy */
96char *argv0; /* root of argv[0] */
97sigset_t s_mask; /* signal mask for cleanup critical sect */
98FILE *listf; /* file pointer to print file list to */
99char *tempfile; /* tempfile to use for mkstemp(3) */
100char *tempbase; /* basename of tempfile to use for mkstemp(3) */
101
102/*
103 * PAX - Portable Archive Interchange
104 *
105 * A utility to read, write, and write lists of the members of archive
106 * files and copy directory hierarchies. A variety of archive formats
107 * are supported (some are described in POSIX 1003.1 10.1):
108 *
109 * ustar - 10.1.1 extended tar interchange format
110 * cpio - 10.1.2 extended cpio interchange format
111 * tar - old BSD 4.3 tar format
112 * binary cpio - old cpio with binary header format
113 * sysVR4 cpio - with and without CRC
114 *
115 * This version is a superset of IEEE Std 1003.2b-d3
116 *
117 * Summary of Extensions to the IEEE Standard:
118 *
119 * 1 READ ENHANCEMENTS
120 * 1.1 Operations which read archives will continue to operate even when
121 * processing archives which may be damaged, truncated, or fail to meet
122 * format specs in several different ways. Damaged sections of archives
123 * are detected and avoided if possible. Attempts will be made to resync
124 * archive read operations even with badly damaged media.
125 * 1.2 Blocksize requirements are not strictly enforced on archive read.
126 * Tapes which have variable sized records can be read without errors.
127 * 1.3 The user can specify via the non-standard option flag -E if error
128 * resync operation should stop on a media error, try a specified number
129 * of times to correct, or try to correct forever.
130 * 1.4 Sparse files (lseek holes) stored on the archive (but stored with blocks
131 * of all zeros will be restored with holes appropriate for the target
132 * filesystem
133 * 1.5 The user is notified whenever something is found during archive
134 * read operations which violates spec (but the read will continue).
135 * 1.6 Multiple archive volumes can be read and may span over different
136 * archive devices
137 * 1.7 Rigidly restores all file attributes exactly as they are stored on the
138 * archive.
139 * 1.8 Modification change time ranges can be specified via multiple -T
140 * options. These allow a user to select files whose modification time
141 * lies within a specific time range.
142 * 1.9 Files can be selected based on owner (user name or uid) via one or more
143 * -U options.
144 * 1.10 Files can be selected based on group (group name or gid) via one o
145 * more -G options.
146 * 1.11 File modification time can be checked against existing file after
147 * name modification (-Z)
148 *
149 * 2 WRITE ENHANCEMENTS
150 * 2.1 Write operation will stop instead of allowing a user to create a flawed
151 * flawed archive (due to any problem).
152 * 2.2 Archives written by pax are forced to strictly conform to both the
153 * archive and pax the specific format specifications.
154 * 2.3 Blocking size and format is rigidly enforced on writes.
155 * 2.4 Formats which may exhibit header overflow problems (they have fields
156 * too small for large file systems, such as inode number storage), use
157 * routines designed to repair this problem. These techniques still
158 * conform to both pax and format specifications, but no longer truncate
159 * these fields. This removes any restrictions on using these archive
160 * formats on large file systems.
161 * 2.5 Multiple archive volumes can be written and may span over different
162 * archive devices
163 * 2.6 A archive volume record limit allows the user to specify the number
164 * of bytes stored on an archive volume. When reached the user is
165 * prompted for the next archive volume. This is specified with the
166 * non-standard -B flag. The limit is rounded up to the next blocksize.
167 * 2.7 All archive padding during write use zero filled sections. This makes
168 * it much easier to pull data out of flawed archive during read
169 * operations.
170 * 2.8 Access time reset with the -t applies to all file nodes (including
171 * directories).
172 * 2.9 Symbolic links can be followed with -L (optional in the spec).
173 * 2.10 Modification or inode change time ranges can be specified via
174 * multiple -T options. These allow a user to select files whose
175 * modification or inode change time lies within a specific time range.
176 * 2.11 Files can be selected based on owner (user name or uid) via one or more
177 * -U options.
178 * 2.12 Files can be selected based on group (group name or gid) via one o
179 * more -G options.
180 * 2.13 Symlinks which appear on the command line can be followed (without
181 * following other symlinks; -H flag)
182 *
183 * 3 COPY ENHANCEMENTS
184 * 3.1 Sparse files (lseek holes) can be copied without expanding the holes
185 * into zero filled blocks. The file copy is created with holes which are
186 * appropriate for the target filesystem
187 * 3.2 Access time as well as modification time on copied file trees can be
188 * preserved with the appropriate -p options.
189 * 3.3 Access time reset with the -t applies to all file nodes (including
190 * directories).
191 * 3.4 Symbolic links can be followed with -L (optional in the spec).
192 * 3.5 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 * 3.6 Files can be selected based on owner (user name or uid) via one or more
196 * -U options.
197 * 3.7 Files can be selected based on group (group name or gid) via one o
198 * more -G options.
199 * 3.8 Symlinks which appear on the command line can be followed (without
200 * following other symlinks; -H flag)
201 * 3.9 File inode change time can be checked against existing file before
202 * name modification (-D)
203 * 3.10 File inode change time can be checked against existing file after
204 * name modification (-Y)
205 * 3.11 File modification time can be checked against existing file after
206 * name modification (-Z)
207 *
208 * 4 GENERAL ENHANCEMENTS
209 * 4.1 Internal structure is designed to isolate format dependent and
210 * independent functions. Formats are selected via a format driver table.
211 * This encourages the addition of new archive formats by only having to
212 * write those routines which id, read and write the archive header.
213 */
214
215/*
216 * main()
217 * parse options, set up and operate as specified by the user.
218 * any operational flaw will set exit_val to non-zero
219 * Return: 0 if ok, 1 otherwise
220 */
221
222#ifdef __STDC__
223int
224main(int argc, char **argv)
225#else
226int
227main(argc, argv)
228 int argc;
229 char **argv;
230#endif
231{
232 char *tmpdir;
233 size_t tdlen;
234
235 (void) setlocale(LC_ALL, "");
236 listf = stderr;
237 /*
238 * Keep a reference to cwd, so we can always come back home.
239 */
240 cwdfd = open(".", O_RDONLY);
241 if (cwdfd < 0) {
242 syswarn(0, errno, "Can't open current working directory.");
243 return(exit_val);
244 }
245
246 /*
247 * Where should we put temporary files?
248 */
249 if ((tmpdir = getenv("TMPDIR")) == NULL || *tmpdir == '\0')
250 tmpdir = _PATH_TMP;
251 tdlen = strlen(tmpdir);
252 while(tdlen > 0 && tmpdir[tdlen - 1] == '/')
253 tdlen--;
254 tempfile = malloc(tdlen + 1 + sizeof(_TFILE_BASE));
255 if (tempfile == NULL) {
256 paxwarn(1, "Cannot allocate memory for temp file name.");
257 return(exit_val);
258 }
259 if (tdlen)
260 memcpy(tempfile, tmpdir, tdlen);
261 tempbase = tempfile + tdlen;
262 *tempbase++ = '/';
263
264 /*
265 * parse options, determine operational mode, general init
266 */
267 options(argc, argv);
268 if ((gen_init() < 0) || (tty_init() < 0))
269 return(exit_val);
270
271 /*
272 * select a primary operation mode
273 */
274 switch(act) {
275 case EXTRACT:
276 extract();
277 break;
278 case ARCHIVE:
279 archive();
280 break;
281 case APPND:
282 if (gzip_program != NULL)
283 err(1, "can not gzip while appending");
284 append();
285 break;
286 case COPY:
287 copy();
288 break;
289 default:
290 case LIST:
291 list();
292 break;
293 }
294 return(exit_val);
295}
296
297/*
298 * sig_cleanup()
299 * when interrupted we try to do whatever delayed processing we can.
300 * This is not critical, but we really ought to limit our damage when we
301 * are aborted by the user.
302 * Return:
303 * never....
304 */
305
306#ifdef __STDC__
307void
308sig_cleanup(int which_sig)
309#else
310void
311sig_cleanup(which_sig)
312 int which_sig;
313#endif
314{
315 /*
316 * restore modes and times for any dirs we may have created
317 * or any dirs we may have read. Set vflag and vfpart so the user
318 * will clearly see the message on a line by itself.
319 */
320 vflag = vfpart = 1;
321 if (which_sig == SIGXCPU)
322 paxwarn(0, "Cpu time limit reached, cleaning up.");
323 else
324 paxwarn(0, "Signal caught, cleaning up.");
325
326 ar_close();
327 proc_dir();
328 if (tflag)
329 atdir_end();
330 exit(1);
331}
332
333/*
334 * gen_init()
335 * general setup routines. Not all are required, but they really help
336 * when dealing with a medium to large sized archives.
337 */
338
339#ifdef __STDC__
340static int
341gen_init(void)
342#else
343static int
344gen_init()
345#endif
346{
347 struct rlimit reslimit;
348 struct sigaction n_hand;
349 struct sigaction o_hand;
350
351 /*
352 * Really needed to handle large archives. We can run out of memory for
353 * internal tables really fast when we have a whole lot of files...
354 */
355 if (getrlimit(RLIMIT_DATA , &reslimit) == 0){
356 reslimit.rlim_cur = reslimit.rlim_max;
357 (void)setrlimit(RLIMIT_DATA , &reslimit);
358 }
359
360 /*
361 * should file size limits be waived? if the os limits us, this is
362 * needed if we want to write a large archive
363 */
364 if (getrlimit(RLIMIT_FSIZE , &reslimit) == 0){
365 reslimit.rlim_cur = reslimit.rlim_max;
366 (void)setrlimit(RLIMIT_FSIZE , &reslimit);
367 }
368
369 /*
370 * increase the size the stack can grow to
371 */
372 if (getrlimit(RLIMIT_STACK , &reslimit) == 0){
373 reslimit.rlim_cur = reslimit.rlim_max;
374 (void)setrlimit(RLIMIT_STACK , &reslimit);
375 }
376
377 /*
378 * not really needed, but doesn't hurt
379 */
380 if (getrlimit(RLIMIT_RSS , &reslimit) == 0){
381 reslimit.rlim_cur = reslimit.rlim_max;
382 (void)setrlimit(RLIMIT_RSS , &reslimit);
383 }
384
385 /*
386 * signal handling to reset stored directory times and modes. Since
387 * we deal with broken pipes via failed writes we ignore it. We also
388 * deal with any file size limit thorugh failed writes. Cpu time
389 * limits are caught and a cleanup is forced.
390 */
391 if ((sigemptyset(&s_mask) < 0) || (sigaddset(&s_mask, SIGTERM) < 0) ||
392 (sigaddset(&s_mask,SIGINT) < 0)||(sigaddset(&s_mask,SIGHUP) < 0) ||
393 (sigaddset(&s_mask,SIGPIPE) < 0)||(sigaddset(&s_mask,SIGQUIT)<0) ||
394 (sigaddset(&s_mask,SIGXCPU) < 0)||(sigaddset(&s_mask,SIGXFSZ)<0)) {
395 paxwarn(1, "Unable to set up signal mask");
396 return(-1);
397 }
398 memset(&n_hand, 0, sizeof n_hand);
399 n_hand.sa_mask = s_mask;
400 n_hand.sa_flags = 0;
401 n_hand.sa_handler = sig_cleanup;
402
403 if ((sigaction(SIGHUP, &n_hand, &o_hand) < 0) &&
404 (o_hand.sa_handler == SIG_IGN) &&
405 (sigaction(SIGHUP, &o_hand, &o_hand) < 0))
406 goto out;
407
408 if ((sigaction(SIGTERM, &n_hand, &o_hand) < 0) &&
409 (o_hand.sa_handler == SIG_IGN) &&
410 (sigaction(SIGTERM, &o_hand, &o_hand) < 0))
411 goto out;
412
413 if ((sigaction(SIGINT, &n_hand, &o_hand) < 0) &&
414 (o_hand.sa_handler == SIG_IGN) &&
415 (sigaction(SIGINT, &o_hand, &o_hand) < 0))
416 goto out;
417
418 if ((sigaction(SIGQUIT, &n_hand, &o_hand) < 0) &&
419 (o_hand.sa_handler == SIG_IGN) &&
420 (sigaction(SIGQUIT, &o_hand, &o_hand) < 0))
421 goto out;
422
423 if ((sigaction(SIGXCPU, &n_hand, &o_hand) < 0) &&
424 (o_hand.sa_handler == SIG_IGN) &&
425 (sigaction(SIGXCPU, &o_hand, &o_hand) < 0))
426 goto out;
427
428 n_hand.sa_handler = SIG_IGN;
429 if ((sigaction(SIGPIPE, &n_hand, &o_hand) < 0) ||
430 (sigaction(SIGXFSZ, &n_hand, &o_hand) < 0))
431 goto out;
432 return(0);
433
434 out:
435 syswarn(1, errno, "Unable to set up signal handler");
436 return(-1);
437}