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. Neither the name of the University nor the names of its contributors
18 * may be used to endorse or promote products derived from this software
19 * without specific prior written permission.
20 *
21 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
31 * SUCH DAMAGE.
32 *
33 * @(#)ar_subs.c 8.2 (Berkeley) 4/18/94
34 * $FreeBSD: src/bin/pax/ar_subs.c,v 1.13.2.1 2001/08/01 05:03:11 obrien Exp $
35 */
36
37 #include <sys/types.h>
38 #include <sys/time.h>
39 #include <sys/stat.h>
40 #include <signal.h>
41 #include <string.h>
42 #include <stdio.h>
43 #include <fcntl.h>
44 #include <errno.h>
45 #include <unistd.h>
46 #include <stdlib.h>
47 #include "pax.h"
48 #include "extern.h"
49
50 static void wr_archive (ARCHD *, int is_app);
51 static int get_arc (void);
52 static int next_head (ARCHD *);
53 extern sigset_t s_mask;
54
55 /*
56 * Routines which control the overall operation modes of pax as specified by
57 * the user: list, append, read ...
58 */
59
60 static char hdbuf[BLKMULT]; /* space for archive header on read */
61 u_long flcnt; /* number of files processed */
62
63 /*
64 * list()
65 * list the contents of an archive which match user supplied pattern(s)
66 * (no pattern matches all).
67 */
68
69 void
list(void)70 list(void)
71 {
72 ARCHD *arcn;
73 int res;
74 ARCHD archd;
75 time_t now;
76
77 arcn = &archd;
78 /*
79 * figure out archive type; pass any format specific options to the
80 * archive option processing routine; call the format init routine. We
81 * also save current time for ls_list() so we do not make a system
82 * call for each file we need to print. If verbose (vflag) start up
83 * the name and group caches.
84 */
85 if ((get_arc() < 0) || ((*frmt->options)() < 0) ||
86 ((*frmt->st_rd)() < 0))
87 return;
88
89 now = time(NULL);
90
91 /*
92 * step through the archive until the format says it is done
93 */
94 while (next_head(arcn) == 0) {
95 /*
96 * check for pattern, and user specified options match.
97 * When all patterns are matched we are done.
98 */
99 if ((res = pat_match(arcn)) < 0)
100 break;
101
102 if ((res == 0) && (sel_chk(arcn) == 0)) {
103 /*
104 * pattern resulted in a selected file
105 */
106 if (pat_sel(arcn) < 0)
107 break;
108
109 /*
110 * modify the name as requested by the user if name
111 * survives modification, do a listing of the file
112 */
113 if ((res = mod_name(arcn)) < 0)
114 break;
115 if (res == 0)
116 ls_list(arcn, now, stdout);
117 }
118
119 /*
120 * skip to next archive format header using values calculated
121 * by the format header read routine
122 */
123 if (rd_skip(arcn->skip + arcn->pad) == 1)
124 break;
125 }
126
127 /*
128 * all done, let format have a chance to cleanup, and make sure that
129 * the patterns supplied by the user were all matched
130 */
131 (*frmt->end_rd)();
132 sigprocmask(SIG_BLOCK, &s_mask, NULL);
133 ar_close();
134 pat_chk();
135 }
136
137 /*
138 * extract()
139 * extract the member(s) of an archive as specified by user supplied
140 * pattern(s) (no patterns extracts all members)
141 */
142
143 void
extract(void)144 extract(void)
145 {
146 ARCHD *arcn;
147 int res;
148 off_t cnt;
149 ARCHD archd;
150 struct stat sb;
151 int fd;
152 time_t now;
153
154 arcn = &archd;
155 /*
156 * figure out archive type; pass any format specific options to the
157 * archive option processing routine; call the format init routine;
158 * start up the directory modification time and access mode database
159 */
160 if ((get_arc() < 0) || ((*frmt->options)() < 0) ||
161 ((*frmt->st_rd)() < 0) || (dir_start() < 0))
162 return;
163
164 /*
165 * When we are doing interactive rename, we store the mapping of names
166 * so we can fix up hard links files later in the archive.
167 */
168 if (iflag && (name_start() < 0))
169 return;
170
171 now = time(NULL);
172
173 /*
174 * step through each entry on the archive until the format read routine
175 * says it is done
176 */
177 while (next_head(arcn) == 0) {
178
179 /*
180 * check for pattern, and user specified options match. When
181 * all the patterns are matched we are done
182 */
183 if ((res = pat_match(arcn)) < 0)
184 break;
185
186 if ((res > 0) || (sel_chk(arcn) != 0)) {
187 /*
188 * file is not selected. skip past any file data and
189 * padding and go back for the next archive member
190 */
191 rd_skip(arcn->skip + arcn->pad);
192 continue;
193 }
194
195 /*
196 * with -u or -D only extract when the archive member is newer
197 * than the file with the same name in the file system (no
198 * test of being the same type is required).
199 * NOTE: this test is done BEFORE name modifications as
200 * specified by pax. this operation can be confusing to the
201 * user who might expect the test to be done on an existing
202 * file AFTER the name mod. In honesty the pax spec is probably
203 * flawed in this respect.
204 */
205 if ((uflag || Dflag) && ((lstat(arcn->name, &sb) == 0))) {
206 if (uflag && Dflag) {
207 if ((arcn->sb.st_mtime <= sb.st_mtime) &&
208 (arcn->sb.st_ctime <= sb.st_ctime)) {
209 rd_skip(arcn->skip + arcn->pad);
210 continue;
211 }
212 } else if (Dflag) {
213 if (arcn->sb.st_ctime <= sb.st_ctime) {
214 rd_skip(arcn->skip + arcn->pad);
215 continue;
216 }
217 } else if (arcn->sb.st_mtime <= sb.st_mtime) {
218 rd_skip(arcn->skip + arcn->pad);
219 continue;
220 }
221 }
222
223 /*
224 * this archive member is now been selected. modify the name.
225 */
226 if ((pat_sel(arcn) < 0) || ((res = mod_name(arcn)) < 0))
227 break;
228 if (res > 0) {
229 /*
230 * a bad name mod, skip and purge name from link table
231 */
232 purg_lnk(arcn);
233 rd_skip(arcn->skip + arcn->pad);
234 continue;
235 }
236
237 /*
238 * Non standard -Y and -Z flag. When the existing file is
239 * same age or newer skip
240 */
241 if ((Yflag || Zflag) && ((lstat(arcn->name, &sb) == 0))) {
242 if (Yflag && Zflag) {
243 if ((arcn->sb.st_mtime <= sb.st_mtime) &&
244 (arcn->sb.st_ctime <= sb.st_ctime)) {
245 rd_skip(arcn->skip + arcn->pad);
246 continue;
247 }
248 } else if (Yflag) {
249 if (arcn->sb.st_ctime <= sb.st_ctime) {
250 rd_skip(arcn->skip + arcn->pad);
251 continue;
252 }
253 } else if (arcn->sb.st_mtime <= sb.st_mtime) {
254 rd_skip(arcn->skip + arcn->pad);
255 continue;
256 }
257 }
258
259 if (vflag) {
260 if (vflag > 1)
261 ls_list(arcn, now, listf);
262 else {
263 fputs(arcn->name, listf);
264 vfpart = 1;
265 }
266 }
267
268 /*
269 * if required, chdir around.
270 */
271 if ((arcn->pat != NULL) && (arcn->pat->chdname != NULL))
272 if (chdir(arcn->pat->chdname) != 0)
273 syswarn(1, errno, "Cannot chdir to %s",
274 arcn->pat->chdname);
275
276 /*
277 * all ok, extract this member based on type
278 */
279 if ((arcn->type != PAX_REG) && (arcn->type != PAX_CTG)) {
280 /*
281 * process archive members that are not regular files.
282 * throw out padding and any data that might follow the
283 * header (as determined by the format).
284 */
285 if ((arcn->type == PAX_HLK) || (arcn->type == PAX_HRG))
286 res = lnk_creat(arcn);
287 else
288 res = node_creat(arcn);
289
290 rd_skip(arcn->skip + arcn->pad);
291 if (res < 0)
292 purg_lnk(arcn);
293
294 if (vflag && vfpart) {
295 putc('\n', listf);
296 vfpart = 0;
297 }
298 continue;
299 }
300 /*
301 * we have a file with data here. If we can not create it, skip
302 * over the data and purge the name from hard link table
303 */
304 if ((fd = file_creat(arcn)) < 0) {
305 rd_skip(arcn->skip + arcn->pad);
306 purg_lnk(arcn);
307 continue;
308 }
309 /*
310 * extract the file from the archive and skip over padding and
311 * any unprocessed data
312 */
313 res = (*frmt->rd_data)(arcn, fd, &cnt);
314 file_close(arcn, fd);
315 if (vflag && vfpart) {
316 putc('\n', listf);
317 vfpart = 0;
318 }
319 if (!res)
320 rd_skip(cnt + arcn->pad);
321
322 /*
323 * if required, chdir around.
324 */
325 if ((arcn->pat != NULL) && (arcn->pat->chdname != NULL))
326 if (fchdir(cwdfd) != 0)
327 syswarn(1, errno,
328 "Can't fchdir to starting directory");
329 }
330
331 /*
332 * all done, restore directory modes and times as required; make sure
333 * all patterns supplied by the user were matched; block off signals
334 * to avoid chance for multiple entry into the cleanup code.
335 */
336 (*frmt->end_rd)();
337 sigprocmask(SIG_BLOCK, &s_mask, NULL);
338 ar_close();
339 proc_dir();
340 pat_chk();
341 }
342
343 /*
344 * wr_archive()
345 * Write an archive. used in both creating a new archive and appends on
346 * previously written archive.
347 */
348
349 static void
wr_archive(ARCHD * arcn,int is_app)350 wr_archive(ARCHD *arcn, int is_app)
351 {
352 int res;
353 int hlk;
354 int wr_one;
355 off_t cnt;
356 int (*wrf)();
357 int fd = -1;
358 time_t now;
359
360 /*
361 * if this format supports hard link storage, start up the database
362 * that detects them.
363 */
364 if (((hlk = frmt->hlk) == 1) && (lnk_start() < 0))
365 return;
366
367 /*
368 * start up the file traversal code and format specific write
369 */
370 if ((ftree_start() < 0) || ((*frmt->st_wr)() < 0))
371 return;
372 wrf = frmt->wr;
373
374 /*
375 * When we are doing interactive rename, we store the mapping of names
376 * so we can fix up hard links files later in the archive.
377 */
378 if (iflag && (name_start() < 0))
379 return;
380
381 /*
382 * if this not append, and there are no files, we do no write a trailer
383 */
384 wr_one = is_app;
385
386 now = time(NULL);
387
388 /*
389 * while there are files to archive, process them one at at time
390 */
391 while (next_file(arcn) == 0) {
392 /*
393 * check if this file meets user specified options match.
394 */
395 if (sel_chk(arcn) != 0)
396 continue;
397 fd = -1;
398 if (uflag) {
399 /*
400 * only archive if this file is newer than a file with
401 * the same name that is already stored on the archive
402 */
403 if ((res = chk_ftime(arcn)) < 0)
404 break;
405 if (res > 0)
406 continue;
407 }
408
409 /*
410 * this file is considered selected now. see if this is a hard
411 * link to a file already stored
412 */
413 ftree_sel(arcn);
414 if (hlk && (chk_lnk(arcn) < 0))
415 break;
416
417 if ((arcn->type == PAX_REG) || (arcn->type == PAX_HRG) ||
418 (arcn->type == PAX_CTG)) {
419 /*
420 * we will have to read this file. by opening it now we
421 * can avoid writing a header to the archive for a file
422 * we were later unable to read (we also purge it from
423 * the link table).
424 */
425 if ((fd = open(arcn->org_name, O_RDONLY, 0)) < 0) {
426 syswarn(1,errno, "Unable to open %s to read",
427 arcn->org_name);
428 purg_lnk(arcn);
429 continue;
430 }
431 }
432
433 /*
434 * Now modify the name as requested by the user
435 */
436 if ((res = mod_name(arcn)) < 0) {
437 /*
438 * name modification says to skip this file, close the
439 * file and purge link table entry
440 */
441 rdfile_close(arcn, &fd);
442 purg_lnk(arcn);
443 break;
444 }
445
446 if ((res > 0) || (docrc && (set_crc(arcn, fd) < 0))) {
447 /*
448 * unable to obtain the crc we need, close the file,
449 * purge link table entry
450 */
451 rdfile_close(arcn, &fd);
452 purg_lnk(arcn);
453 continue;
454 }
455
456 if (vflag) {
457 if (vflag > 1)
458 ls_list(arcn, now, listf);
459 else {
460 fputs(arcn->name, listf);
461 vfpart = 1;
462 }
463 }
464 ++flcnt;
465
466 /*
467 * looks safe to store the file, have the format specific
468 * routine write routine store the file header on the archive
469 */
470 if ((res = (*wrf)(arcn)) < 0) {
471 rdfile_close(arcn, &fd);
472 break;
473 }
474 wr_one = 1;
475 if (res > 0) {
476 /*
477 * format write says no file data needs to be stored
478 * so we are done messing with this file
479 */
480 if (vflag && vfpart) {
481 putc('\n', listf);
482 vfpart = 0;
483 }
484 rdfile_close(arcn, &fd);
485 continue;
486 }
487
488 /*
489 * Add file data to the archive, quit on write error. if we
490 * cannot write the entire file contents to the archive we
491 * must pad the archive to replace the missing file data
492 * (otherwise during an extract the file header for the file
493 * which FOLLOWS this one will not be where we expect it to
494 * be).
495 */
496 res = (*frmt->wr_data)(arcn, fd, &cnt);
497 rdfile_close(arcn, &fd);
498 if (vflag && vfpart) {
499 putc('\n', listf);
500 vfpart = 0;
501 }
502 if (res < 0)
503 break;
504
505 /*
506 * pad as required, cnt is number of bytes not written
507 */
508 if (((cnt > 0) && (wr_skip(cnt) < 0)) ||
509 ((arcn->pad > 0) && (wr_skip(arcn->pad) < 0)))
510 break;
511 }
512
513 /*
514 * tell format to write trailer; pad to block boundary; reset directory
515 * mode/access times, and check if all patterns supplied by the user
516 * were matched. block off signals to avoid chance for multiple entry
517 * into the cleanup code
518 */
519 if (wr_one) {
520 (*frmt->end_wr)();
521 wr_fin();
522 }
523 sigprocmask(SIG_BLOCK, &s_mask, NULL);
524 ar_close();
525 if (tflag)
526 proc_dir();
527 ftree_chk();
528 }
529
530 /*
531 * append()
532 * Add file to previously written archive. Archive format specified by the
533 * user must agree with archive. The archive is read first to collect
534 * modification times (if -u) and locate the archive trailer. The archive
535 * is positioned in front of the record with the trailer and wr_archive()
536 * is called to add the new members.
537 * PAX IMPLEMENTATION DETAIL NOTE:
538 * -u is implemented by adding the new members to the end of the archive.
539 * Care is taken so that these do not end up as links to the older
540 * version of the same file already stored in the archive. It is expected
541 * when extraction occurs these newer versions will over-write the older
542 * ones stored "earlier" in the archive (this may be a bad assumption as
543 * it depends on the implementation of the program doing the extraction).
544 * It is really difficult to splice in members without either re-writing
545 * the entire archive (from the point were the old version was), or having
546 * assistance of the format specification in terms of a special update
547 * header that invalidates a previous archive record. The POSIX spec left
548 * the method used to implement -u unspecified. This pax is able to
549 * over write existing files that it creates.
550 */
551
552 void
append(void)553 append(void)
554 {
555 ARCHD *arcn;
556 int res;
557 ARCHD archd;
558 FSUB *orgfrmt;
559 int udev;
560 off_t tlen;
561
562 arcn = &archd;
563 orgfrmt = frmt;
564
565 /*
566 * Do not allow an append operation if the actual archive is of a
567 * different format than the user specified format.
568 */
569 if (get_arc() < 0)
570 return;
571 if ((orgfrmt != NULL) && (orgfrmt != frmt)) {
572 paxwarn(1, "Cannot mix current archive format %s with %s",
573 frmt->name, orgfrmt->name);
574 return;
575 }
576
577 /*
578 * pass the format any options and start up format
579 */
580 if (((*frmt->options)() < 0) || ((*frmt->st_rd)() < 0))
581 return;
582
583 /*
584 * if we only are adding members that are newer, we need to save the
585 * mod times for all files we see.
586 */
587 if (uflag && (ftime_start() < 0))
588 return;
589
590 /*
591 * some archive formats encode hard links by recording the device and
592 * file serial number (inode) but copy the file anyway (multiple times)
593 * to the archive. When we append, we run the risk that newly added
594 * files may have the same device and inode numbers as those recorded
595 * on the archive but during a previous run. If this happens, when the
596 * archive is extracted we get INCORRECT hard links. We avoid this by
597 * remapping the device numbers so that newly added files will never
598 * use the same device number as one found on the archive. remapping
599 * allows new members to safely have links among themselves. remapping
600 * also avoids problems with file inode (serial number) truncations
601 * when the inode number is larger than storage space in the archive
602 * header. See the remap routines for more details.
603 */
604 if ((udev = frmt->udev) && (dev_start() < 0))
605 return;
606
607 /*
608 * reading the archive may take a long time. If verbose tell the user
609 */
610 if (vflag) {
611 fprintf(listf,
612 "%s: Reading archive to position at the end...", argv0);
613 vfpart = 1;
614 }
615
616 /*
617 * step through the archive until the format says it is done
618 */
619 while (next_head(arcn) == 0) {
620 /*
621 * check if this file meets user specified options.
622 */
623 if (sel_chk(arcn) != 0) {
624 if (rd_skip(arcn->skip + arcn->pad) == 1)
625 break;
626 continue;
627 }
628
629 if (uflag) {
630 /*
631 * see if this is the newest version of this file has
632 * already been seen, if so skip.
633 */
634 if ((res = chk_ftime(arcn)) < 0)
635 break;
636 if (res > 0) {
637 if (rd_skip(arcn->skip + arcn->pad) == 1)
638 break;
639 continue;
640 }
641 }
642
643 /*
644 * Store this device number. Device numbers seen during the
645 * read phase of append will cause newly appended files with a
646 * device number seen in the old part of the archive to be
647 * remapped to an unused device number.
648 */
649 if ((udev && (add_dev(arcn) < 0)) ||
650 (rd_skip(arcn->skip + arcn->pad) == 1))
651 break;
652 }
653
654 /*
655 * done, finish up read and get the number of bytes to back up so we
656 * can add new members. The format might have used the hard link table,
657 * purge it.
658 */
659 tlen = (*frmt->end_rd)();
660 lnk_end();
661
662 /*
663 * try to position for write, if this fails quit. if any error occurs,
664 * we will refuse to write
665 */
666 if (appnd_start(tlen) < 0)
667 return;
668
669 /*
670 * tell the user we are done reading.
671 */
672 if (vflag && vfpart) {
673 fputs("done.\n", listf);
674 vfpart = 0;
675 }
676
677 /*
678 * go to the writing phase to add the new members
679 */
680 wr_archive(arcn, 1);
681 }
682
683 /*
684 * archive()
685 * write a new archive
686 */
687
688 void
archive(void)689 archive(void)
690 {
691 ARCHD archd;
692
693 /*
694 * if we only are adding members that are newer, we need to save the
695 * mod times for all files; set up for writing; pass the format any
696 * options write the archive
697 */
698 if ((uflag && (ftime_start() < 0)) || (wr_start() < 0))
699 return;
700 if ((*frmt->options)() < 0)
701 return;
702
703 wr_archive(&archd, 0);
704 }
705
706 /*
707 * copy()
708 * copy files from one part of the file system to another. this does not
709 * use any archive storage. The EFFECT OF THE COPY IS THE SAME as if an
710 * archive was written and then extracted in the destination directory
711 * (except the files are forced to be under the destination directory).
712 */
713
714 void
copy(void)715 copy(void)
716 {
717 ARCHD *arcn;
718 int res;
719 int fddest;
720 char *dest_pt;
721 int dlen;
722 int drem;
723 int fdsrc = -1;
724 struct stat sb;
725 ARCHD archd;
726 char dirbuf[PAXPATHLEN+1];
727
728 arcn = &archd;
729 /*
730 * set up the destination dir path and make sure it is a directory. We
731 * make sure we have a trailing / on the destination
732 */
733 dlen = l_strncpy(dirbuf, dirptr, sizeof(dirbuf) - 1);
734 dest_pt = dirbuf + dlen;
735 if (*(dest_pt-1) != '/') {
736 *dest_pt++ = '/';
737 ++dlen;
738 }
739 *dest_pt = '\0';
740 drem = PAXPATHLEN - dlen;
741
742 if (stat(dirptr, &sb) < 0) {
743 syswarn(1, errno, "Cannot access destination directory %s",
744 dirptr);
745 return;
746 }
747 if (!S_ISDIR(sb.st_mode)) {
748 paxwarn(1, "Destination is not a directory %s", dirptr);
749 return;
750 }
751
752 /*
753 * start up the hard link table; file traversal routines and the
754 * modification time and access mode database
755 */
756 if ((lnk_start() < 0) || (ftree_start() < 0) || (dir_start() < 0))
757 return;
758
759 /*
760 * When we are doing interactive rename, we store the mapping of names
761 * so we can fix up hard links files later in the archive.
762 */
763 if (iflag && (name_start() < 0))
764 return;
765
766 /*
767 * set up to cp file trees
768 */
769 cp_start();
770
771 /*
772 * while there are files to archive, process them
773 */
774 while (next_file(arcn) == 0) {
775 fdsrc = -1;
776
777 /*
778 * check if this file meets user specified options
779 */
780 if (sel_chk(arcn) != 0)
781 continue;
782
783 /*
784 * if there is already a file in the destination directory with
785 * the same name and it is newer, skip the one stored on the
786 * archive.
787 * NOTE: this test is done BEFORE name modifications as
788 * specified by pax. this can be confusing to the user who
789 * might expect the test to be done on an existing file AFTER
790 * the name mod. In honesty the pax spec is probably flawed in
791 * this respect
792 */
793 if (uflag || Dflag) {
794 /*
795 * create the destination name
796 */
797 if (*(arcn->name) == '/')
798 res = 1;
799 else
800 res = 0;
801 if ((arcn->nlen - res) > drem) {
802 paxwarn(1, "Destination pathname too long %s",
803 arcn->name);
804 continue;
805 }
806 strncpy(dest_pt, arcn->name + res, drem);
807 dirbuf[PAXPATHLEN] = '\0';
808
809 /*
810 * if existing file is same age or newer skip
811 */
812 res = lstat(dirbuf, &sb);
813 *dest_pt = '\0';
814
815 if (res == 0) {
816 if (uflag && Dflag) {
817 if ((arcn->sb.st_mtime<=sb.st_mtime) &&
818 (arcn->sb.st_ctime<=sb.st_ctime))
819 continue;
820 } else if (Dflag) {
821 if (arcn->sb.st_ctime <= sb.st_ctime)
822 continue;
823 } else if (arcn->sb.st_mtime <= sb.st_mtime)
824 continue;
825 }
826 }
827
828 /*
829 * this file is considered selected. See if this is a hard link
830 * to a previous file; modify the name as requested by the
831 * user; set the final destination.
832 */
833 ftree_sel(arcn);
834 if ((chk_lnk(arcn) < 0) || ((res = mod_name(arcn)) < 0))
835 break;
836 if ((res > 0) || (set_dest(arcn, dirbuf, dlen) < 0)) {
837 /*
838 * skip file, purge from link table
839 */
840 purg_lnk(arcn);
841 continue;
842 }
843
844 /*
845 * Non standard -Y and -Z flag. When the existing file is
846 * same age or newer skip
847 */
848 if ((Yflag || Zflag) && ((lstat(arcn->name, &sb) == 0))) {
849 if (Yflag && Zflag) {
850 if ((arcn->sb.st_mtime <= sb.st_mtime) &&
851 (arcn->sb.st_ctime <= sb.st_ctime))
852 continue;
853 } else if (Yflag) {
854 if (arcn->sb.st_ctime <= sb.st_ctime)
855 continue;
856 } else if (arcn->sb.st_mtime <= sb.st_mtime)
857 continue;
858 }
859
860 if (vflag) {
861 fputs(arcn->name, listf);
862 vfpart = 1;
863 }
864 ++flcnt;
865
866 /*
867 * try to create a hard link to the src file if requested
868 * but make sure we are not trying to overwrite ourselves.
869 */
870 if (lflag)
871 res = cross_lnk(arcn);
872 else
873 res = chk_same(arcn);
874 if (res <= 0) {
875 if (vflag && vfpart) {
876 putc('\n', listf);
877 vfpart = 0;
878 }
879 continue;
880 }
881
882 /*
883 * have to create a new file
884 */
885 if ((arcn->type != PAX_REG) && (arcn->type != PAX_CTG)) {
886 /*
887 * create a link or special file
888 */
889 if ((arcn->type == PAX_HLK) || (arcn->type == PAX_HRG))
890 res = lnk_creat(arcn);
891 else
892 res = node_creat(arcn);
893 if (res < 0)
894 purg_lnk(arcn);
895 if (vflag && vfpart) {
896 putc('\n', listf);
897 vfpart = 0;
898 }
899 continue;
900 }
901
902 /*
903 * have to copy a regular file to the destination directory.
904 * first open source file and then create the destination file
905 */
906 if ((fdsrc = open(arcn->org_name, O_RDONLY, 0)) < 0) {
907 syswarn(1, errno, "Unable to open %s to read",
908 arcn->org_name);
909 purg_lnk(arcn);
910 continue;
911 }
912 if ((fddest = file_creat(arcn)) < 0) {
913 rdfile_close(arcn, &fdsrc);
914 purg_lnk(arcn);
915 continue;
916 }
917
918 /*
919 * copy source file data to the destination file
920 */
921 cp_file(arcn, fdsrc, fddest);
922 file_close(arcn, fddest);
923 rdfile_close(arcn, &fdsrc);
924
925 if (vflag && vfpart) {
926 putc('\n', listf);
927 vfpart = 0;
928 }
929 }
930
931 /*
932 * restore directory modes and times as required; make sure all
933 * patterns were selected block off signals to avoid chance for
934 * multiple entry into the cleanup code.
935 */
936 sigprocmask(SIG_BLOCK, &s_mask, NULL);
937 ar_close();
938 proc_dir();
939 ftree_chk();
940 }
941
942 /*
943 * next_head()
944 * try to find a valid header in the archive. Uses format specific
945 * routines to extract the header and id the trailer. Trailers may be
946 * located within a valid header or in an invalid header (the location
947 * is format specific. The inhead field from the option table tells us
948 * where to look for the trailer).
949 * We keep reading (and resyncing) until we get enough contiguous data
950 * to check for a header. If we cannot find one, we shift by a byte
951 * add a new byte from the archive to the end of the buffer and try again.
952 * If we get a read error, we throw out what we have (as we must have
953 * contiguous data) and start over again.
954 * ASSUMED: headers fit within a BLKMULT header.
955 * Return:
956 * 0 if we got a header, -1 if we are unable to ever find another one
957 * (we reached the end of input, or we reached the limit on retries. see
958 * the specs for rd_wrbuf() for more details)
959 */
960
961 static int
next_head(ARCHD * arcn)962 next_head(ARCHD *arcn)
963 {
964 int ret;
965 char *hdend;
966 int res;
967 int shftsz;
968 int hsz;
969 int in_resync = 0; /* set when we are in resync mode */
970 int cnt = 0; /* counter for trailer function */
971 int first = 1; /* on 1st read, EOF isn't premature. */
972
973 /*
974 * set up initial conditions, we want a whole frmt->hsz block as we
975 * have no data yet.
976 */
977 res = hsz = frmt->hsz;
978 hdend = hdbuf;
979 shftsz = hsz - 1;
980 for(;;) {
981 /*
982 * keep looping until we get a contiguous FULL buffer
983 * (frmt->hsz is the proper size)
984 */
985 for (;;) {
986 if ((ret = rd_wrbuf(hdend, res)) == res)
987 break;
988
989 /*
990 * If we read 0 bytes (EOF) from an archive when we
991 * expect to find a header, we have stepped upon
992 * an archive without the customary block of zeroes
993 * end marker. It's just stupid to error out on
994 * them, so exit gracefully.
995 */
996 if (first && ret == 0)
997 return(-1);
998 first = 0;
999
1000 /*
1001 * some kind of archive read problem, try to resync the
1002 * storage device, better give the user the bad news.
1003 */
1004 if ((ret == 0) || (rd_sync() < 0)) {
1005 paxwarn(1,"Premature end of file on archive read");
1006 return(-1);
1007 }
1008 if (!in_resync) {
1009 if (act == APPND) {
1010 paxwarn(1,
1011 "Archive I/O error, cannot continue");
1012 return(-1);
1013 }
1014 paxwarn(1,"Archive I/O error. Trying to recover.");
1015 ++in_resync;
1016 }
1017
1018 /*
1019 * oh well, throw it all out and start over
1020 */
1021 res = hsz;
1022 hdend = hdbuf;
1023 }
1024
1025 /*
1026 * ok we have a contiguous buffer of the right size. Call the
1027 * format read routine. If this was not a valid header and this
1028 * format stores trailers outside of the header, call the
1029 * format specific trailer routine to check for a trailer. We
1030 * have to watch out that we do not mis-identify file data or
1031 * block padding as a header or trailer. Format specific
1032 * trailer functions must NOT check for the trailer while we
1033 * are running in resync mode. Some trailer functions may tell
1034 * us that this block cannot contain a valid header either, so
1035 * we then throw out the entire block and start over.
1036 */
1037 if ((*frmt->rd)(arcn, hdbuf) == 0)
1038 break;
1039
1040 if (!frmt->inhead) {
1041 /*
1042 * this format has trailers outside of valid headers
1043 */
1044 if ((ret = (*frmt->trail)(hdbuf,in_resync,&cnt)) == 0){
1045 /*
1046 * valid trailer found, drain input as required
1047 */
1048 ar_drain();
1049 return(-1);
1050 }
1051
1052 if (ret == 1) {
1053 /*
1054 * we are in resync and we were told to throw
1055 * the whole block out because none of the
1056 * bytes in this block can be used to form a
1057 * valid header
1058 */
1059 res = hsz;
1060 hdend = hdbuf;
1061 continue;
1062 }
1063 }
1064
1065 /*
1066 * Brute force section.
1067 * not a valid header. We may be able to find a header yet. So
1068 * we shift over by one byte, and set up to read one byte at a
1069 * time from the archive and place it at the end of the buffer.
1070 * We will keep moving byte at a time until we find a header or
1071 * get a read error and have to start over.
1072 */
1073 if (!in_resync) {
1074 if (act == APPND) {
1075 paxwarn(1,"Unable to append, archive header flaw");
1076 return(-1);
1077 }
1078 paxwarn(1,"Invalid header, starting valid header search.");
1079 ++in_resync;
1080 }
1081 memmove(hdbuf, hdbuf+1, shftsz);
1082 res = 1;
1083 hdend = hdbuf + shftsz;
1084 }
1085
1086 /*
1087 * ok got a valid header, check for trailer if format encodes it in the
1088 * the header. NOTE: the parameters are different than trailer routines
1089 * which encode trailers outside of the header!
1090 */
1091 if (frmt->inhead && ((*frmt->trail)(arcn) == 0)) {
1092 /*
1093 * valid trailer found, drain input as required
1094 */
1095 ar_drain();
1096 return(-1);
1097 }
1098
1099 ++flcnt;
1100 return(0);
1101 }
1102
1103 /*
1104 * get_arc()
1105 * Figure out what format an archive is. Handles archive with flaws by
1106 * brute force searches for a legal header in any supported format. The
1107 * format id routines have to be careful to NOT mis-identify a format.
1108 * ASSUMED: headers fit within a BLKMULT header.
1109 * Return:
1110 * 0 if archive found -1 otherwise
1111 */
1112
1113 static int
get_arc(void)1114 get_arc(void)
1115 {
1116 int i;
1117 int hdsz = 0;
1118 int res;
1119 int minhd = BLKMULT;
1120 char *hdend;
1121 int notice = 0;
1122
1123 /*
1124 * find the smallest header size in all archive formats and then set up
1125 * to read the archive.
1126 */
1127 for (i = 0; ford[i] >= 0; ++i) {
1128 if (fsub[ford[i]].hsz < minhd)
1129 minhd = fsub[ford[i]].hsz;
1130 }
1131 if (rd_start() < 0)
1132 return(-1);
1133 res = BLKMULT;
1134 hdsz = 0;
1135 hdend = hdbuf;
1136 for(;;) {
1137 for (;;) {
1138 /*
1139 * fill the buffer with at least the smallest header
1140 */
1141 i = rd_wrbuf(hdend, res);
1142 if (i > 0)
1143 hdsz += i;
1144 if (hdsz >= minhd)
1145 break;
1146
1147 /*
1148 * if we cannot recover from a read error quit
1149 */
1150 if ((i == 0) || (rd_sync() < 0))
1151 goto out;
1152
1153 /*
1154 * when we get an error none of the data we already
1155 * have can be used to create a legal header (we just
1156 * got an error in the middle), so we throw it all out
1157 * and refill the buffer with fresh data.
1158 */
1159 res = BLKMULT;
1160 hdsz = 0;
1161 hdend = hdbuf;
1162 if (!notice) {
1163 if (act == APPND)
1164 return(-1);
1165 paxwarn(1,"Cannot identify format. Searching...");
1166 ++notice;
1167 }
1168 }
1169
1170 /*
1171 * we have at least the size of the smallest header in any
1172 * archive format. Look to see if we have a match. The array
1173 * ford[] is used to specify the header id order to reduce the
1174 * chance of incorrectly id'ing a valid header (some formats
1175 * may be subsets of each other and the order would then be
1176 * important).
1177 */
1178 for (i = 0; ford[i] >= 0; ++i) {
1179 if ((*fsub[ford[i]].id)(hdbuf, hdsz) < 0)
1180 continue;
1181 frmt = &(fsub[ford[i]]);
1182 /*
1183 * yuck, to avoid slow special case code in the extract
1184 * routines, just push this header back as if it was
1185 * not seen. We have left extra space at start of the
1186 * buffer for this purpose. This is a bit ugly, but
1187 * adding all the special case code is far worse.
1188 */
1189 pback(hdbuf, hdsz);
1190 return(0);
1191 }
1192
1193 /*
1194 * We have a flawed archive, no match. we start searching, but
1195 * we never allow additions to flawed archives
1196 */
1197 if (!notice) {
1198 if (act == APPND)
1199 return(-1);
1200 paxwarn(1, "Cannot identify format. Searching...");
1201 ++notice;
1202 }
1203
1204 /*
1205 * brute force search for a header that we can id.
1206 * we shift through byte at a time. this is slow, but we cannot
1207 * determine the nature of the flaw in the archive in a
1208 * portable manner
1209 */
1210 if (--hdsz > 0) {
1211 memmove(hdbuf, hdbuf+1, hdsz);
1212 res = BLKMULT - hdsz;
1213 hdend = hdbuf + hdsz;
1214 } else {
1215 res = BLKMULT;
1216 hdend = hdbuf;
1217 hdsz = 0;
1218 }
1219 }
1220
1221 out:
1222 /*
1223 * we cannot find a header, bow, apologize and quit
1224 */
1225 paxwarn(1, "Sorry, unable to determine archive format.");
1226 return(-1);
1227 }
1228