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. 36 * 37 * @(#)ar_subs.c 8.2 (Berkeley) 4/18/94 38 * $FreeBSD: src/bin/pax/ar_subs.c,v 1.13.2.1 2001/08/01 05:03:11 obrien Exp $ 39 * $DragonFly: src/bin/pax/ar_subs.c,v 1.7 2006/09/27 21:58:08 pavalos Exp $ 40 */ 41 42 #include <sys/types.h> 43 #include <sys/time.h> 44 #include <sys/stat.h> 45 #include <signal.h> 46 #include <string.h> 47 #include <stdio.h> 48 #include <fcntl.h> 49 #include <errno.h> 50 #include <unistd.h> 51 #include <stdlib.h> 52 #include "pax.h" 53 #include "extern.h" 54 55 static void wr_archive (ARCHD *, int is_app); 56 static int get_arc (void); 57 static int next_head (ARCHD *); 58 extern sigset_t s_mask; 59 60 /* 61 * Routines which control the overall operation modes of pax as specified by 62 * the user: list, append, read ... 63 */ 64 65 static char hdbuf[BLKMULT]; /* space for archive header on read */ 66 u_long flcnt; /* number of files processed */ 67 68 /* 69 * list() 70 * list the contents of an archive which match user supplied pattern(s) 71 * (no pattern matches all). 72 */ 73 74 void 75 list(void) 76 { 77 ARCHD *arcn; 78 int res; 79 ARCHD archd; 80 time_t now; 81 82 arcn = &archd; 83 /* 84 * figure out archive type; pass any format specific options to the 85 * archive option processing routine; call the format init routine. We 86 * also save current time for ls_list() so we do not make a system 87 * call for each file we need to print. If verbose (vflag) start up 88 * the name and group caches. 89 */ 90 if ((get_arc() < 0) || ((*frmt->options)() < 0) || 91 ((*frmt->st_rd)() < 0)) 92 return; 93 94 if (vflag && ((uidtb_start() < 0) || (gidtb_start() < 0))) 95 return; 96 97 now = time(NULL); 98 99 /* 100 * step through the archive until the format says it is done 101 */ 102 while (next_head(arcn) == 0) { 103 /* 104 * check for pattern, and user specified options match. 105 * When all patterns are matched we are done. 106 */ 107 if ((res = pat_match(arcn)) < 0) 108 break; 109 110 if ((res == 0) && (sel_chk(arcn) == 0)) { 111 /* 112 * pattern resulted in a selected file 113 */ 114 if (pat_sel(arcn) < 0) 115 break; 116 117 /* 118 * modify the name as requested by the user if name 119 * survives modification, do a listing of the file 120 */ 121 if ((res = mod_name(arcn)) < 0) 122 break; 123 if (res == 0) 124 ls_list(arcn, now, stdout); 125 } 126 127 /* 128 * skip to next archive format header using values calculated 129 * by the format header read routine 130 */ 131 if (rd_skip(arcn->skip + arcn->pad) == 1) 132 break; 133 } 134 135 /* 136 * all done, let format have a chance to cleanup, and make sure that 137 * the patterns supplied by the user were all matched 138 */ 139 (*frmt->end_rd)(); 140 sigprocmask(SIG_BLOCK, &s_mask, NULL); 141 ar_close(); 142 pat_chk(); 143 } 144 145 /* 146 * extract() 147 * extract the member(s) of an archive as specified by user supplied 148 * pattern(s) (no patterns extracts all members) 149 */ 150 151 void 152 extract(void) 153 { 154 ARCHD *arcn; 155 int res; 156 off_t cnt; 157 ARCHD archd; 158 struct stat sb; 159 int fd; 160 time_t now; 161 162 arcn = &archd; 163 /* 164 * figure out archive type; pass any format specific options to the 165 * archive option processing routine; call the format init routine; 166 * start up the directory modification time and access mode database 167 */ 168 if ((get_arc() < 0) || ((*frmt->options)() < 0) || 169 ((*frmt->st_rd)() < 0) || (dir_start() < 0)) 170 return; 171 172 /* 173 * When we are doing interactive rename, we store the mapping of names 174 * so we can fix up hard links files later in the archive. 175 */ 176 if (iflag && (name_start() < 0)) 177 return; 178 179 now = time(NULL); 180 181 /* 182 * step through each entry on the archive until the format read routine 183 * says it is done 184 */ 185 while (next_head(arcn) == 0) { 186 187 /* 188 * check for pattern, and user specified options match. When 189 * all the patterns are matched we are done 190 */ 191 if ((res = pat_match(arcn)) < 0) 192 break; 193 194 if ((res > 0) || (sel_chk(arcn) != 0)) { 195 /* 196 * file is not selected. skip past any file data and 197 * padding and go back for the next archive member 198 */ 199 rd_skip(arcn->skip + arcn->pad); 200 continue; 201 } 202 203 /* 204 * with -u or -D only extract when the archive member is newer 205 * than the file with the same name in the file system (no 206 * test of being the same type is required). 207 * NOTE: this test is done BEFORE name modifications as 208 * specified by pax. this operation can be confusing to the 209 * user who might expect the test to be done on an existing 210 * file AFTER the name mod. In honesty the pax spec is probably 211 * flawed in this respect. 212 */ 213 if ((uflag || Dflag) && ((lstat(arcn->name, &sb) == 0))) { 214 if (uflag && Dflag) { 215 if ((arcn->sb.st_mtime <= sb.st_mtime) && 216 (arcn->sb.st_ctime <= sb.st_ctime)) { 217 rd_skip(arcn->skip + arcn->pad); 218 continue; 219 } 220 } else if (Dflag) { 221 if (arcn->sb.st_ctime <= sb.st_ctime) { 222 rd_skip(arcn->skip + arcn->pad); 223 continue; 224 } 225 } else if (arcn->sb.st_mtime <= sb.st_mtime) { 226 rd_skip(arcn->skip + arcn->pad); 227 continue; 228 } 229 } 230 231 /* 232 * this archive member is now been selected. modify the name. 233 */ 234 if ((pat_sel(arcn) < 0) || ((res = mod_name(arcn)) < 0)) 235 break; 236 if (res > 0) { 237 /* 238 * a bad name mod, skip and purge name from link table 239 */ 240 purg_lnk(arcn); 241 rd_skip(arcn->skip + arcn->pad); 242 continue; 243 } 244 245 /* 246 * Non standard -Y and -Z flag. When the existing file is 247 * same age or newer skip 248 */ 249 if ((Yflag || Zflag) && ((lstat(arcn->name, &sb) == 0))) { 250 if (Yflag && Zflag) { 251 if ((arcn->sb.st_mtime <= sb.st_mtime) && 252 (arcn->sb.st_ctime <= sb.st_ctime)) { 253 rd_skip(arcn->skip + arcn->pad); 254 continue; 255 } 256 } else if (Yflag) { 257 if (arcn->sb.st_ctime <= sb.st_ctime) { 258 rd_skip(arcn->skip + arcn->pad); 259 continue; 260 } 261 } else if (arcn->sb.st_mtime <= sb.st_mtime) { 262 rd_skip(arcn->skip + arcn->pad); 263 continue; 264 } 265 } 266 267 if (vflag) { 268 if (vflag > 1) 269 ls_list(arcn, now, listf); 270 else { 271 fputs(arcn->name, listf); 272 vfpart = 1; 273 } 274 } 275 276 /* 277 * if required, chdir around. 278 */ 279 if ((arcn->pat != NULL) && (arcn->pat->chdname != NULL)) 280 if (chdir(arcn->pat->chdname) != 0) 281 syswarn(1, errno, "Cannot chdir to %s", 282 arcn->pat->chdname); 283 284 /* 285 * all ok, extract this member based on type 286 */ 287 if ((arcn->type != PAX_REG) && (arcn->type != PAX_CTG)) { 288 /* 289 * process archive members that are not regular files. 290 * throw out padding and any data that might follow the 291 * header (as determined by the format). 292 */ 293 if ((arcn->type == PAX_HLK) || (arcn->type == PAX_HRG)) 294 res = lnk_creat(arcn); 295 else 296 res = node_creat(arcn); 297 298 rd_skip(arcn->skip + arcn->pad); 299 if (res < 0) 300 purg_lnk(arcn); 301 302 if (vflag && vfpart) { 303 putc('\n', listf); 304 vfpart = 0; 305 } 306 continue; 307 } 308 /* 309 * we have a file with data here. If we can not create it, skip 310 * over the data and purge the name from hard link table 311 */ 312 if ((fd = file_creat(arcn)) < 0) { 313 rd_skip(arcn->skip + arcn->pad); 314 purg_lnk(arcn); 315 continue; 316 } 317 /* 318 * extract the file from the archive and skip over padding and 319 * any unprocessed data 320 */ 321 res = (*frmt->rd_data)(arcn, fd, &cnt); 322 file_close(arcn, fd); 323 if (vflag && vfpart) { 324 putc('\n', listf); 325 vfpart = 0; 326 } 327 if (!res) 328 rd_skip(cnt + arcn->pad); 329 330 /* 331 * if required, chdir around. 332 */ 333 if ((arcn->pat != NULL) && (arcn->pat->chdname != NULL)) 334 if (fchdir(cwdfd) != 0) 335 syswarn(1, errno, 336 "Can't fchdir to starting directory"); 337 } 338 339 /* 340 * all done, restore directory modes and times as required; make sure 341 * all patterns supplied by the user were matched; block off signals 342 * to avoid chance for multiple entry into the cleanup code. 343 */ 344 (*frmt->end_rd)(); 345 sigprocmask(SIG_BLOCK, &s_mask, NULL); 346 ar_close(); 347 proc_dir(); 348 pat_chk(); 349 } 350 351 /* 352 * wr_archive() 353 * Write an archive. used in both creating a new archive and appends on 354 * previously written archive. 355 */ 356 357 static void 358 wr_archive(ARCHD *arcn, int is_app) 359 { 360 int res; 361 int hlk; 362 int wr_one; 363 off_t cnt; 364 int (*wrf)(); 365 int fd = -1; 366 time_t now; 367 368 /* 369 * if this format supports hard link storage, start up the database 370 * that detects them. 371 */ 372 if (((hlk = frmt->hlk) == 1) && (lnk_start() < 0)) 373 return; 374 375 /* 376 * start up the file traversal code and format specific write 377 */ 378 if ((ftree_start() < 0) || ((*frmt->st_wr)() < 0)) 379 return; 380 wrf = frmt->wr; 381 382 /* 383 * When we are doing interactive rename, we store the mapping of names 384 * so we can fix up hard links files later in the archive. 385 */ 386 if (iflag && (name_start() < 0)) 387 return; 388 389 /* 390 * if this not append, and there are no files, we do no write a trailer 391 */ 392 wr_one = is_app; 393 394 now = time(NULL); 395 396 /* 397 * while there are files to archive, process them one at at time 398 */ 399 while (next_file(arcn) == 0) { 400 /* 401 * check if this file meets user specified options match. 402 */ 403 if (sel_chk(arcn) != 0) 404 continue; 405 fd = -1; 406 if (uflag) { 407 /* 408 * only archive if this file is newer than a file with 409 * the same name that is already stored on the archive 410 */ 411 if ((res = chk_ftime(arcn)) < 0) 412 break; 413 if (res > 0) 414 continue; 415 } 416 417 /* 418 * this file is considered selected now. see if this is a hard 419 * link to a file already stored 420 */ 421 ftree_sel(arcn); 422 if (hlk && (chk_lnk(arcn) < 0)) 423 break; 424 425 if ((arcn->type == PAX_REG) || (arcn->type == PAX_HRG) || 426 (arcn->type == PAX_CTG)) { 427 /* 428 * we will have to read this file. by opening it now we 429 * can avoid writing a header to the archive for a file 430 * we were later unable to read (we also purge it from 431 * the link table). 432 */ 433 if ((fd = open(arcn->org_name, O_RDONLY, 0)) < 0) { 434 syswarn(1,errno, "Unable to open %s to read", 435 arcn->org_name); 436 purg_lnk(arcn); 437 continue; 438 } 439 } 440 441 /* 442 * Now modify the name as requested by the user 443 */ 444 if ((res = mod_name(arcn)) < 0) { 445 /* 446 * name modification says to skip this file, close the 447 * file and purge link table entry 448 */ 449 rdfile_close(arcn, &fd); 450 purg_lnk(arcn); 451 break; 452 } 453 454 if ((res > 0) || (docrc && (set_crc(arcn, fd) < 0))) { 455 /* 456 * unable to obtain the crc we need, close the file, 457 * purge link table entry 458 */ 459 rdfile_close(arcn, &fd); 460 purg_lnk(arcn); 461 continue; 462 } 463 464 if (vflag) { 465 if (vflag > 1) 466 ls_list(arcn, now, listf); 467 else { 468 fputs(arcn->name, listf); 469 vfpart = 1; 470 } 471 } 472 ++flcnt; 473 474 /* 475 * looks safe to store the file, have the format specific 476 * routine write routine store the file header on the archive 477 */ 478 if ((res = (*wrf)(arcn)) < 0) { 479 rdfile_close(arcn, &fd); 480 break; 481 } 482 wr_one = 1; 483 if (res > 0) { 484 /* 485 * format write says no file data needs to be stored 486 * so we are done messing with this file 487 */ 488 if (vflag && vfpart) { 489 putc('\n', listf); 490 vfpart = 0; 491 } 492 rdfile_close(arcn, &fd); 493 continue; 494 } 495 496 /* 497 * Add file data to the archive, quit on write error. if we 498 * cannot write the entire file contents to the archive we 499 * must pad the archive to replace the missing file data 500 * (otherwise during an extract the file header for the file 501 * which FOLLOWS this one will not be where we expect it to 502 * be). 503 */ 504 res = (*frmt->wr_data)(arcn, fd, &cnt); 505 rdfile_close(arcn, &fd); 506 if (vflag && vfpart) { 507 putc('\n', listf); 508 vfpart = 0; 509 } 510 if (res < 0) 511 break; 512 513 /* 514 * pad as required, cnt is number of bytes not written 515 */ 516 if (((cnt > 0) && (wr_skip(cnt) < 0)) || 517 ((arcn->pad > 0) && (wr_skip(arcn->pad) < 0))) 518 break; 519 } 520 521 /* 522 * tell format to write trailer; pad to block boundary; reset directory 523 * mode/access times, and check if all patterns supplied by the user 524 * were matched. block off signals to avoid chance for multiple entry 525 * into the cleanup code 526 */ 527 if (wr_one) { 528 (*frmt->end_wr)(); 529 wr_fin(); 530 } 531 sigprocmask(SIG_BLOCK, &s_mask, NULL); 532 ar_close(); 533 if (tflag) 534 proc_dir(); 535 ftree_chk(); 536 } 537 538 /* 539 * append() 540 * Add file to previously written archive. Archive format specified by the 541 * user must agree with archive. The archive is read first to collect 542 * modification times (if -u) and locate the archive trailer. The archive 543 * is positioned in front of the record with the trailer and wr_archive() 544 * is called to add the new members. 545 * PAX IMPLEMENTATION DETAIL NOTE: 546 * -u is implemented by adding the new members to the end of the archive. 547 * Care is taken so that these do not end up as links to the older 548 * version of the same file already stored in the archive. It is expected 549 * when extraction occurs these newer versions will over-write the older 550 * ones stored "earlier" in the archive (this may be a bad assumption as 551 * it depends on the implementation of the program doing the extraction). 552 * It is really difficult to splice in members without either re-writing 553 * the entire archive (from the point were the old version was), or having 554 * assistance of the format specification in terms of a special update 555 * header that invalidates a previous archive record. The POSIX spec left 556 * the method used to implement -u unspecified. This pax is able to 557 * over write existing files that it creates. 558 */ 559 560 void 561 append(void) 562 { 563 ARCHD *arcn; 564 int res; 565 ARCHD archd; 566 FSUB *orgfrmt; 567 int udev; 568 off_t tlen; 569 570 arcn = &archd; 571 orgfrmt = frmt; 572 573 /* 574 * Do not allow an append operation if the actual archive is of a 575 * different format than the user specified format. 576 */ 577 if (get_arc() < 0) 578 return; 579 if ((orgfrmt != NULL) && (orgfrmt != frmt)) { 580 paxwarn(1, "Cannot mix current archive format %s with %s", 581 frmt->name, orgfrmt->name); 582 return; 583 } 584 585 /* 586 * pass the format any options and start up format 587 */ 588 if (((*frmt->options)() < 0) || ((*frmt->st_rd)() < 0)) 589 return; 590 591 /* 592 * if we only are adding members that are newer, we need to save the 593 * mod times for all files we see. 594 */ 595 if (uflag && (ftime_start() < 0)) 596 return; 597 598 /* 599 * some archive formats encode hard links by recording the device and 600 * file serial number (inode) but copy the file anyway (multiple times) 601 * to the archive. When we append, we run the risk that newly added 602 * files may have the same device and inode numbers as those recorded 603 * on the archive but during a previous run. If this happens, when the 604 * archive is extracted we get INCORRECT hard links. We avoid this by 605 * remapping the device numbers so that newly added files will never 606 * use the same device number as one found on the archive. remapping 607 * allows new members to safely have links among themselves. remapping 608 * also avoids problems with file inode (serial number) truncations 609 * when the inode number is larger than storage space in the archive 610 * header. See the remap routines for more details. 611 */ 612 if ((udev = frmt->udev) && (dev_start() < 0)) 613 return; 614 615 /* 616 * reading the archive may take a long time. If verbose tell the user 617 */ 618 if (vflag) { 619 fprintf(listf, 620 "%s: Reading archive to position at the end...", argv0); 621 vfpart = 1; 622 } 623 624 /* 625 * step through the archive until the format says it is done 626 */ 627 while (next_head(arcn) == 0) { 628 /* 629 * check if this file meets user specified options. 630 */ 631 if (sel_chk(arcn) != 0) { 632 if (rd_skip(arcn->skip + arcn->pad) == 1) 633 break; 634 continue; 635 } 636 637 if (uflag) { 638 /* 639 * see if this is the newest version of this file has 640 * already been seen, if so skip. 641 */ 642 if ((res = chk_ftime(arcn)) < 0) 643 break; 644 if (res > 0) { 645 if (rd_skip(arcn->skip + arcn->pad) == 1) 646 break; 647 continue; 648 } 649 } 650 651 /* 652 * Store this device number. Device numbers seen during the 653 * read phase of append will cause newly appended files with a 654 * device number seen in the old part of the archive to be 655 * remapped to an unused device number. 656 */ 657 if ((udev && (add_dev(arcn) < 0)) || 658 (rd_skip(arcn->skip + arcn->pad) == 1)) 659 break; 660 } 661 662 /* 663 * done, finish up read and get the number of bytes to back up so we 664 * can add new members. The format might have used the hard link table, 665 * purge it. 666 */ 667 tlen = (*frmt->end_rd)(); 668 lnk_end(); 669 670 /* 671 * try to position for write, if this fails quit. if any error occurs, 672 * we will refuse to write 673 */ 674 if (appnd_start(tlen) < 0) 675 return; 676 677 /* 678 * tell the user we are done reading. 679 */ 680 if (vflag && vfpart) { 681 fputs("done.\n", listf); 682 vfpart = 0; 683 } 684 685 /* 686 * go to the writing phase to add the new members 687 */ 688 wr_archive(arcn, 1); 689 } 690 691 /* 692 * archive() 693 * write a new archive 694 */ 695 696 void 697 archive(void) 698 { 699 ARCHD archd; 700 701 /* 702 * if we only are adding members that are newer, we need to save the 703 * mod times for all files; set up for writing; pass the format any 704 * options write the archive 705 */ 706 if ((uflag && (ftime_start() < 0)) || (wr_start() < 0)) 707 return; 708 if ((*frmt->options)() < 0) 709 return; 710 711 wr_archive(&archd, 0); 712 } 713 714 /* 715 * copy() 716 * copy files from one part of the file system to another. this does not 717 * use any archive storage. The EFFECT OF THE COPY IS THE SAME as if an 718 * archive was written and then extracted in the destination directory 719 * (except the files are forced to be under the destination directory). 720 */ 721 722 void 723 copy(void) 724 { 725 ARCHD *arcn; 726 int res; 727 int fddest; 728 char *dest_pt; 729 int dlen; 730 int drem; 731 int fdsrc = -1; 732 struct stat sb; 733 ARCHD archd; 734 char dirbuf[PAXPATHLEN+1]; 735 736 arcn = &archd; 737 /* 738 * set up the destination dir path and make sure it is a directory. We 739 * make sure we have a trailing / on the destination 740 */ 741 dlen = l_strncpy(dirbuf, dirptr, sizeof(dirbuf) - 1); 742 dest_pt = dirbuf + dlen; 743 if (*(dest_pt-1) != '/') { 744 *dest_pt++ = '/'; 745 ++dlen; 746 } 747 *dest_pt = '\0'; 748 drem = PAXPATHLEN - dlen; 749 750 if (stat(dirptr, &sb) < 0) { 751 syswarn(1, errno, "Cannot access destination directory %s", 752 dirptr); 753 return; 754 } 755 if (!S_ISDIR(sb.st_mode)) { 756 paxwarn(1, "Destination is not a directory %s", dirptr); 757 return; 758 } 759 760 /* 761 * start up the hard link table; file traversal routines and the 762 * modification time and access mode database 763 */ 764 if ((lnk_start() < 0) || (ftree_start() < 0) || (dir_start() < 0)) 765 return; 766 767 /* 768 * When we are doing interactive rename, we store the mapping of names 769 * so we can fix up hard links files later in the archive. 770 */ 771 if (iflag && (name_start() < 0)) 772 return; 773 774 /* 775 * set up to cp file trees 776 */ 777 cp_start(); 778 779 /* 780 * while there are files to archive, process them 781 */ 782 while (next_file(arcn) == 0) { 783 fdsrc = -1; 784 785 /* 786 * check if this file meets user specified options 787 */ 788 if (sel_chk(arcn) != 0) 789 continue; 790 791 /* 792 * if there is already a file in the destination directory with 793 * the same name and it is newer, skip the one stored on the 794 * archive. 795 * NOTE: this test is done BEFORE name modifications as 796 * specified by pax. this can be confusing to the user who 797 * might expect the test to be done on an existing file AFTER 798 * the name mod. In honesty the pax spec is probably flawed in 799 * this respect 800 */ 801 if (uflag || Dflag) { 802 /* 803 * create the destination name 804 */ 805 if (*(arcn->name) == '/') 806 res = 1; 807 else 808 res = 0; 809 if ((arcn->nlen - res) > drem) { 810 paxwarn(1, "Destination pathname too long %s", 811 arcn->name); 812 continue; 813 } 814 strncpy(dest_pt, arcn->name + res, drem); 815 dirbuf[PAXPATHLEN] = '\0'; 816 817 /* 818 * if existing file is same age or newer skip 819 */ 820 res = lstat(dirbuf, &sb); 821 *dest_pt = '\0'; 822 823 if (res == 0) { 824 if (uflag && Dflag) { 825 if ((arcn->sb.st_mtime<=sb.st_mtime) && 826 (arcn->sb.st_ctime<=sb.st_ctime)) 827 continue; 828 } else if (Dflag) { 829 if (arcn->sb.st_ctime <= sb.st_ctime) 830 continue; 831 } else if (arcn->sb.st_mtime <= sb.st_mtime) 832 continue; 833 } 834 } 835 836 /* 837 * this file is considered selected. See if this is a hard link 838 * to a previous file; modify the name as requested by the 839 * user; set the final destination. 840 */ 841 ftree_sel(arcn); 842 if ((chk_lnk(arcn) < 0) || ((res = mod_name(arcn)) < 0)) 843 break; 844 if ((res > 0) || (set_dest(arcn, dirbuf, dlen) < 0)) { 845 /* 846 * skip file, purge from link table 847 */ 848 purg_lnk(arcn); 849 continue; 850 } 851 852 /* 853 * Non standard -Y and -Z flag. When the existing file is 854 * same age or newer skip 855 */ 856 if ((Yflag || Zflag) && ((lstat(arcn->name, &sb) == 0))) { 857 if (Yflag && Zflag) { 858 if ((arcn->sb.st_mtime <= sb.st_mtime) && 859 (arcn->sb.st_ctime <= sb.st_ctime)) 860 continue; 861 } else if (Yflag) { 862 if (arcn->sb.st_ctime <= sb.st_ctime) 863 continue; 864 } else if (arcn->sb.st_mtime <= sb.st_mtime) 865 continue; 866 } 867 868 if (vflag) { 869 fputs(arcn->name, listf); 870 vfpart = 1; 871 } 872 ++flcnt; 873 874 /* 875 * try to create a hard link to the src file if requested 876 * but make sure we are not trying to overwrite ourselves. 877 */ 878 if (lflag) 879 res = cross_lnk(arcn); 880 else 881 res = chk_same(arcn); 882 if (res <= 0) { 883 if (vflag && vfpart) { 884 putc('\n', listf); 885 vfpart = 0; 886 } 887 continue; 888 } 889 890 /* 891 * have to create a new file 892 */ 893 if ((arcn->type != PAX_REG) && (arcn->type != PAX_CTG)) { 894 /* 895 * create a link or special file 896 */ 897 if ((arcn->type == PAX_HLK) || (arcn->type == PAX_HRG)) 898 res = lnk_creat(arcn); 899 else 900 res = node_creat(arcn); 901 if (res < 0) 902 purg_lnk(arcn); 903 if (vflag && vfpart) { 904 putc('\n', listf); 905 vfpart = 0; 906 } 907 continue; 908 } 909 910 /* 911 * have to copy a regular file to the destination directory. 912 * first open source file and then create the destination file 913 */ 914 if ((fdsrc = open(arcn->org_name, O_RDONLY, 0)) < 0) { 915 syswarn(1, errno, "Unable to open %s to read", 916 arcn->org_name); 917 purg_lnk(arcn); 918 continue; 919 } 920 if ((fddest = file_creat(arcn)) < 0) { 921 rdfile_close(arcn, &fdsrc); 922 purg_lnk(arcn); 923 continue; 924 } 925 926 /* 927 * copy source file data to the destination file 928 */ 929 cp_file(arcn, fdsrc, fddest); 930 file_close(arcn, fddest); 931 rdfile_close(arcn, &fdsrc); 932 933 if (vflag && vfpart) { 934 putc('\n', listf); 935 vfpart = 0; 936 } 937 } 938 939 /* 940 * restore directory modes and times as required; make sure all 941 * patterns were selected block off signals to avoid chance for 942 * multiple entry into the cleanup code. 943 */ 944 sigprocmask(SIG_BLOCK, &s_mask, NULL); 945 ar_close(); 946 proc_dir(); 947 ftree_chk(); 948 } 949 950 /* 951 * next_head() 952 * try to find a valid header in the archive. Uses format specific 953 * routines to extract the header and id the trailer. Trailers may be 954 * located within a valid header or in an invalid header (the location 955 * is format specific. The inhead field from the option table tells us 956 * where to look for the trailer). 957 * We keep reading (and resyncing) until we get enough contiguous data 958 * to check for a header. If we cannot find one, we shift by a byte 959 * add a new byte from the archive to the end of the buffer and try again. 960 * If we get a read error, we throw out what we have (as we must have 961 * contiguous data) and start over again. 962 * ASSUMED: headers fit within a BLKMULT header. 963 * Return: 964 * 0 if we got a header, -1 if we are unable to ever find another one 965 * (we reached the end of input, or we reached the limit on retries. see 966 * the specs for rd_wrbuf() for more details) 967 */ 968 969 static int 970 next_head(ARCHD *arcn) 971 { 972 int ret; 973 char *hdend; 974 int res; 975 int shftsz; 976 int hsz; 977 int in_resync = 0; /* set when we are in resync mode */ 978 int cnt = 0; /* counter for trailer function */ 979 int first = 1; /* on 1st read, EOF isn't premature. */ 980 981 /* 982 * set up initial conditions, we want a whole frmt->hsz block as we 983 * have no data yet. 984 */ 985 res = hsz = frmt->hsz; 986 hdend = hdbuf; 987 shftsz = hsz - 1; 988 for(;;) { 989 /* 990 * keep looping until we get a contiguous FULL buffer 991 * (frmt->hsz is the proper size) 992 */ 993 for (;;) { 994 if ((ret = rd_wrbuf(hdend, res)) == res) 995 break; 996 997 /* 998 * If we read 0 bytes (EOF) from an archive when we 999 * expect to find a header, we have stepped upon 1000 * an archive without the customary block of zeroes 1001 * end marker. It's just stupid to error out on 1002 * them, so exit gracefully. 1003 */ 1004 if (first && ret == 0) 1005 return(-1); 1006 first = 0; 1007 1008 /* 1009 * some kind of archive read problem, try to resync the 1010 * storage device, better give the user the bad news. 1011 */ 1012 if ((ret == 0) || (rd_sync() < 0)) { 1013 paxwarn(1,"Premature end of file on archive read"); 1014 return(-1); 1015 } 1016 if (!in_resync) { 1017 if (act == APPND) { 1018 paxwarn(1, 1019 "Archive I/O error, cannot continue"); 1020 return(-1); 1021 } 1022 paxwarn(1,"Archive I/O error. Trying to recover."); 1023 ++in_resync; 1024 } 1025 1026 /* 1027 * oh well, throw it all out and start over 1028 */ 1029 res = hsz; 1030 hdend = hdbuf; 1031 } 1032 1033 /* 1034 * ok we have a contiguous buffer of the right size. Call the 1035 * format read routine. If this was not a valid header and this 1036 * format stores trailers outside of the header, call the 1037 * format specific trailer routine to check for a trailer. We 1038 * have to watch out that we do not mis-identify file data or 1039 * block padding as a header or trailer. Format specific 1040 * trailer functions must NOT check for the trailer while we 1041 * are running in resync mode. Some trailer functions may tell 1042 * us that this block cannot contain a valid header either, so 1043 * we then throw out the entire block and start over. 1044 */ 1045 if ((*frmt->rd)(arcn, hdbuf) == 0) 1046 break; 1047 1048 if (!frmt->inhead) { 1049 /* 1050 * this format has trailers outside of valid headers 1051 */ 1052 if ((ret = (*frmt->trail)(hdbuf,in_resync,&cnt)) == 0){ 1053 /* 1054 * valid trailer found, drain input as required 1055 */ 1056 ar_drain(); 1057 return(-1); 1058 } 1059 1060 if (ret == 1) { 1061 /* 1062 * we are in resync and we were told to throw 1063 * the whole block out because none of the 1064 * bytes in this block can be used to form a 1065 * valid header 1066 */ 1067 res = hsz; 1068 hdend = hdbuf; 1069 continue; 1070 } 1071 } 1072 1073 /* 1074 * Brute force section. 1075 * not a valid header. We may be able to find a header yet. So 1076 * we shift over by one byte, and set up to read one byte at a 1077 * time from the archive and place it at the end of the buffer. 1078 * We will keep moving byte at a time until we find a header or 1079 * get a read error and have to start over. 1080 */ 1081 if (!in_resync) { 1082 if (act == APPND) { 1083 paxwarn(1,"Unable to append, archive header flaw"); 1084 return(-1); 1085 } 1086 paxwarn(1,"Invalid header, starting valid header search."); 1087 ++in_resync; 1088 } 1089 memmove(hdbuf, hdbuf+1, shftsz); 1090 res = 1; 1091 hdend = hdbuf + shftsz; 1092 } 1093 1094 /* 1095 * ok got a valid header, check for trailer if format encodes it in the 1096 * the header. NOTE: the parameters are different than trailer routines 1097 * which encode trailers outside of the header! 1098 */ 1099 if (frmt->inhead && ((*frmt->trail)(arcn) == 0)) { 1100 /* 1101 * valid trailer found, drain input as required 1102 */ 1103 ar_drain(); 1104 return(-1); 1105 } 1106 1107 ++flcnt; 1108 return(0); 1109 } 1110 1111 /* 1112 * get_arc() 1113 * Figure out what format an archive is. Handles archive with flaws by 1114 * brute force searches for a legal header in any supported format. The 1115 * format id routines have to be careful to NOT mis-identify a format. 1116 * ASSUMED: headers fit within a BLKMULT header. 1117 * Return: 1118 * 0 if archive found -1 otherwise 1119 */ 1120 1121 static int 1122 get_arc(void) 1123 { 1124 int i; 1125 int hdsz = 0; 1126 int res; 1127 int minhd = BLKMULT; 1128 char *hdend; 1129 int notice = 0; 1130 1131 /* 1132 * find the smallest header size in all archive formats and then set up 1133 * to read the archive. 1134 */ 1135 for (i = 0; ford[i] >= 0; ++i) { 1136 if (fsub[ford[i]].hsz < minhd) 1137 minhd = fsub[ford[i]].hsz; 1138 } 1139 if (rd_start() < 0) 1140 return(-1); 1141 res = BLKMULT; 1142 hdsz = 0; 1143 hdend = hdbuf; 1144 for(;;) { 1145 for (;;) { 1146 /* 1147 * fill the buffer with at least the smallest header 1148 */ 1149 i = rd_wrbuf(hdend, res); 1150 if (i > 0) 1151 hdsz += i; 1152 if (hdsz >= minhd) 1153 break; 1154 1155 /* 1156 * if we cannot recover from a read error quit 1157 */ 1158 if ((i == 0) || (rd_sync() < 0)) 1159 goto out; 1160 1161 /* 1162 * when we get an error none of the data we already 1163 * have can be used to create a legal header (we just 1164 * got an error in the middle), so we throw it all out 1165 * and refill the buffer with fresh data. 1166 */ 1167 res = BLKMULT; 1168 hdsz = 0; 1169 hdend = hdbuf; 1170 if (!notice) { 1171 if (act == APPND) 1172 return(-1); 1173 paxwarn(1,"Cannot identify format. Searching..."); 1174 ++notice; 1175 } 1176 } 1177 1178 /* 1179 * we have at least the size of the smallest header in any 1180 * archive format. Look to see if we have a match. The array 1181 * ford[] is used to specify the header id order to reduce the 1182 * chance of incorrectly id'ing a valid header (some formats 1183 * may be subsets of each other and the order would then be 1184 * important). 1185 */ 1186 for (i = 0; ford[i] >= 0; ++i) { 1187 if ((*fsub[ford[i]].id)(hdbuf, hdsz) < 0) 1188 continue; 1189 frmt = &(fsub[ford[i]]); 1190 /* 1191 * yuck, to avoid slow special case code in the extract 1192 * routines, just push this header back as if it was 1193 * not seen. We have left extra space at start of the 1194 * buffer for this purpose. This is a bit ugly, but 1195 * adding all the special case code is far worse. 1196 */ 1197 pback(hdbuf, hdsz); 1198 return(0); 1199 } 1200 1201 /* 1202 * We have a flawed archive, no match. we start searching, but 1203 * we never allow additions to flawed archives 1204 */ 1205 if (!notice) { 1206 if (act == APPND) 1207 return(-1); 1208 paxwarn(1, "Cannot identify format. Searching..."); 1209 ++notice; 1210 } 1211 1212 /* 1213 * brute force search for a header that we can id. 1214 * we shift through byte at a time. this is slow, but we cannot 1215 * determine the nature of the flaw in the archive in a 1216 * portable manner 1217 */ 1218 if (--hdsz > 0) { 1219 memmove(hdbuf, hdbuf+1, hdsz); 1220 res = BLKMULT - hdsz; 1221 hdend = hdbuf + hdsz; 1222 } else { 1223 res = BLKMULT; 1224 hdend = hdbuf; 1225 hdsz = 0; 1226 } 1227 } 1228 1229 out: 1230 /* 1231 * we cannot find a header, bow, apologize and quit 1232 */ 1233 paxwarn(1, "Sorry, unable to determine archive format."); 1234 return(-1); 1235 } 1236