archive_read_support_format_tar.c revision 229592
1/*- 2 * Copyright (c) 2003-2007 Tim Kientzle 3 * All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 14 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR 15 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES 16 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. 17 * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, 18 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT 19 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 20 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 21 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 22 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF 23 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 24 */ 25 26#include "archive_platform.h" 27__FBSDID("$FreeBSD: stable/9/contrib/libarchive/libarchive/archive_read_support_format_tar.c 229592 2012-01-05 12:06:54Z mm $"); 28 29#ifdef HAVE_ERRNO_H 30#include <errno.h> 31#endif 32#include <stddef.h> 33/* #include <stdint.h> */ /* See archive_platform.h */ 34#ifdef HAVE_STDLIB_H 35#include <stdlib.h> 36#endif 37#ifdef HAVE_STRING_H 38#include <string.h> 39#endif 40 41/* Obtain suitable wide-character manipulation functions. */ 42#ifdef HAVE_WCHAR_H 43#include <wchar.h> 44#else 45/* Good enough for equality testing, which is all we need. */ 46static int wcscmp(const wchar_t *s1, const wchar_t *s2) 47{ 48 int diff = *s1 - *s2; 49 while (*s1 && diff == 0) 50 diff = (int)*++s1 - (int)*++s2; 51 return diff; 52} 53/* Good enough for equality testing, which is all we need. */ 54static int wcsncmp(const wchar_t *s1, const wchar_t *s2, size_t n) 55{ 56 int diff = *s1 - *s2; 57 while (*s1 && diff == 0 && n-- > 0) 58 diff = (int)*++s1 - (int)*++s2; 59 return diff; 60} 61static size_t wcslen(const wchar_t *s) 62{ 63 const wchar_t *p = s; 64 while (*p) 65 p++; 66 return p - s; 67} 68#endif 69 70#include "archive.h" 71#include "archive_entry.h" 72#include "archive_private.h" 73#include "archive_read_private.h" 74 75#define tar_min(a,b) ((a) < (b) ? (a) : (b)) 76 77/* 78 * Layout of POSIX 'ustar' tar header. 79 */ 80struct archive_entry_header_ustar { 81 char name[100]; 82 char mode[8]; 83 char uid[8]; 84 char gid[8]; 85 char size[12]; 86 char mtime[12]; 87 char checksum[8]; 88 char typeflag[1]; 89 char linkname[100]; /* "old format" header ends here */ 90 char magic[6]; /* For POSIX: "ustar\0" */ 91 char version[2]; /* For POSIX: "00" */ 92 char uname[32]; 93 char gname[32]; 94 char rdevmajor[8]; 95 char rdevminor[8]; 96 char prefix[155]; 97}; 98 99/* 100 * Structure of GNU tar header 101 */ 102struct gnu_sparse { 103 char offset[12]; 104 char numbytes[12]; 105}; 106 107struct archive_entry_header_gnutar { 108 char name[100]; 109 char mode[8]; 110 char uid[8]; 111 char gid[8]; 112 char size[12]; 113 char mtime[12]; 114 char checksum[8]; 115 char typeflag[1]; 116 char linkname[100]; 117 char magic[8]; /* "ustar \0" (note blank/blank/null at end) */ 118 char uname[32]; 119 char gname[32]; 120 char rdevmajor[8]; 121 char rdevminor[8]; 122 char atime[12]; 123 char ctime[12]; 124 char offset[12]; 125 char longnames[4]; 126 char unused[1]; 127 struct gnu_sparse sparse[4]; 128 char isextended[1]; 129 char realsize[12]; 130 /* 131 * Old GNU format doesn't use POSIX 'prefix' field; they use 132 * the 'L' (longname) entry instead. 133 */ 134}; 135 136/* 137 * Data specific to this format. 138 */ 139struct sparse_block { 140 struct sparse_block *next; 141 off_t offset; 142 off_t remaining; 143}; 144 145struct tar { 146 struct archive_string acl_text; 147 struct archive_string entry_pathname; 148 /* For "GNU.sparse.name" and other similar path extensions. */ 149 struct archive_string entry_pathname_override; 150 struct archive_string entry_linkpath; 151 struct archive_string entry_uname; 152 struct archive_string entry_gname; 153 struct archive_string longlink; 154 struct archive_string longname; 155 struct archive_string pax_header; 156 struct archive_string pax_global; 157 struct archive_string line; 158 int pax_hdrcharset_binary; 159 wchar_t *pax_entry; 160 size_t pax_entry_length; 161 int header_recursion_depth; 162 int64_t entry_bytes_remaining; 163 int64_t entry_offset; 164 int64_t entry_padding; 165 int64_t realsize; 166 struct sparse_block *sparse_list; 167 struct sparse_block *sparse_last; 168 int64_t sparse_offset; 169 int64_t sparse_numbytes; 170 int sparse_gnu_major; 171 int sparse_gnu_minor; 172 char sparse_gnu_pending; 173}; 174 175static ssize_t UTF8_mbrtowc(wchar_t *pwc, const char *s, size_t n); 176static int archive_block_is_null(const unsigned char *p); 177static char *base64_decode(const char *, size_t, size_t *); 178static void gnu_add_sparse_entry(struct tar *, 179 off_t offset, off_t remaining); 180static void gnu_clear_sparse_list(struct tar *); 181static int gnu_sparse_old_read(struct archive_read *, struct tar *, 182 const struct archive_entry_header_gnutar *header); 183static void gnu_sparse_old_parse(struct tar *, 184 const struct gnu_sparse *sparse, int length); 185static int gnu_sparse_01_parse(struct tar *, const char *); 186static ssize_t gnu_sparse_10_read(struct archive_read *, struct tar *); 187static int header_Solaris_ACL(struct archive_read *, struct tar *, 188 struct archive_entry *, const void *); 189static int header_common(struct archive_read *, struct tar *, 190 struct archive_entry *, const void *); 191static int header_old_tar(struct archive_read *, struct tar *, 192 struct archive_entry *, const void *); 193static int header_pax_extensions(struct archive_read *, struct tar *, 194 struct archive_entry *, const void *); 195static int header_pax_global(struct archive_read *, struct tar *, 196 struct archive_entry *, const void *h); 197static int header_longlink(struct archive_read *, struct tar *, 198 struct archive_entry *, const void *h); 199static int header_longname(struct archive_read *, struct tar *, 200 struct archive_entry *, const void *h); 201static int header_volume(struct archive_read *, struct tar *, 202 struct archive_entry *, const void *h); 203static int header_ustar(struct archive_read *, struct tar *, 204 struct archive_entry *, const void *h); 205static int header_gnutar(struct archive_read *, struct tar *, 206 struct archive_entry *, const void *h); 207static int archive_read_format_tar_bid(struct archive_read *); 208static int archive_read_format_tar_cleanup(struct archive_read *); 209static int archive_read_format_tar_read_data(struct archive_read *a, 210 const void **buff, size_t *size, off_t *offset); 211static int archive_read_format_tar_skip(struct archive_read *a); 212static int archive_read_format_tar_read_header(struct archive_read *, 213 struct archive_entry *); 214static int checksum(struct archive_read *, const void *); 215static int pax_attribute(struct tar *, struct archive_entry *, 216 char *key, char *value); 217static int pax_header(struct archive_read *, struct tar *, 218 struct archive_entry *, char *attr); 219static void pax_time(const char *, int64_t *sec, long *nanos); 220static ssize_t readline(struct archive_read *, struct tar *, const char **, 221 ssize_t limit); 222static int read_body_to_string(struct archive_read *, struct tar *, 223 struct archive_string *, const void *h); 224static int64_t tar_atol(const char *, unsigned); 225static int64_t tar_atol10(const char *, unsigned); 226static int64_t tar_atol256(const char *, unsigned); 227static int64_t tar_atol8(const char *, unsigned); 228static int tar_read_header(struct archive_read *, struct tar *, 229 struct archive_entry *); 230static int tohex(int c); 231static char *url_decode(const char *); 232static wchar_t *utf8_decode(struct tar *, const char *, size_t length); 233 234int 235archive_read_support_format_gnutar(struct archive *a) 236{ 237 return (archive_read_support_format_tar(a)); 238} 239 240 241int 242archive_read_support_format_tar(struct archive *_a) 243{ 244 struct archive_read *a = (struct archive_read *)_a; 245 struct tar *tar; 246 int r; 247 248 tar = (struct tar *)malloc(sizeof(*tar)); 249 if (tar == NULL) { 250 archive_set_error(&a->archive, ENOMEM, 251 "Can't allocate tar data"); 252 return (ARCHIVE_FATAL); 253 } 254 memset(tar, 0, sizeof(*tar)); 255 256 r = __archive_read_register_format(a, tar, "tar", 257 archive_read_format_tar_bid, 258 NULL, 259 archive_read_format_tar_read_header, 260 archive_read_format_tar_read_data, 261 archive_read_format_tar_skip, 262 archive_read_format_tar_cleanup); 263 264 if (r != ARCHIVE_OK) 265 free(tar); 266 return (ARCHIVE_OK); 267} 268 269static int 270archive_read_format_tar_cleanup(struct archive_read *a) 271{ 272 struct tar *tar; 273 274 tar = (struct tar *)(a->format->data); 275 gnu_clear_sparse_list(tar); 276 archive_string_free(&tar->acl_text); 277 archive_string_free(&tar->entry_pathname); 278 archive_string_free(&tar->entry_pathname_override); 279 archive_string_free(&tar->entry_linkpath); 280 archive_string_free(&tar->entry_uname); 281 archive_string_free(&tar->entry_gname); 282 archive_string_free(&tar->line); 283 archive_string_free(&tar->pax_global); 284 archive_string_free(&tar->pax_header); 285 archive_string_free(&tar->longname); 286 archive_string_free(&tar->longlink); 287 free(tar->pax_entry); 288 free(tar); 289 (a->format->data) = NULL; 290 return (ARCHIVE_OK); 291} 292 293 294static int 295archive_read_format_tar_bid(struct archive_read *a) 296{ 297 int bid; 298 const void *h; 299 const struct archive_entry_header_ustar *header; 300 301 bid = 0; 302 303 /* Now let's look at the actual header and see if it matches. */ 304 h = __archive_read_ahead(a, 512, NULL); 305 if (h == NULL) 306 return (-1); 307 308 /* If it's an end-of-archive mark, we can handle it. */ 309 if ((*(const char *)h) == 0 310 && archive_block_is_null((const unsigned char *)h)) { 311 /* 312 * Usually, I bid the number of bits verified, but 313 * in this case, 4096 seems excessive so I picked 10 as 314 * an arbitrary but reasonable-seeming value. 315 */ 316 return (10); 317 } 318 319 /* If it's not an end-of-archive mark, it must have a valid checksum.*/ 320 if (!checksum(a, h)) 321 return (0); 322 bid += 48; /* Checksum is usually 6 octal digits. */ 323 324 header = (const struct archive_entry_header_ustar *)h; 325 326 /* Recognize POSIX formats. */ 327 if ((memcmp(header->magic, "ustar\0", 6) == 0) 328 &&(memcmp(header->version, "00", 2)==0)) 329 bid += 56; 330 331 /* Recognize GNU tar format. */ 332 if ((memcmp(header->magic, "ustar ", 6) == 0) 333 &&(memcmp(header->version, " \0", 2)==0)) 334 bid += 56; 335 336 /* Type flag must be null, digit or A-Z, a-z. */ 337 if (header->typeflag[0] != 0 && 338 !( header->typeflag[0] >= '0' && header->typeflag[0] <= '9') && 339 !( header->typeflag[0] >= 'A' && header->typeflag[0] <= 'Z') && 340 !( header->typeflag[0] >= 'a' && header->typeflag[0] <= 'z') ) 341 return (0); 342 bid += 2; /* 6 bits of variation in an 8-bit field leaves 2 bits. */ 343 344 /* Sanity check: Look at first byte of mode field. */ 345 switch (255 & (unsigned)header->mode[0]) { 346 case 0: case 255: 347 /* Base-256 value: No further verification possible! */ 348 break; 349 case ' ': /* Not recommended, but not illegal, either. */ 350 break; 351 case '0': case '1': case '2': case '3': 352 case '4': case '5': case '6': case '7': 353 /* Octal Value. */ 354 /* TODO: Check format of remainder of this field. */ 355 break; 356 default: 357 /* Not a valid mode; bail out here. */ 358 return (0); 359 } 360 /* TODO: Sanity test uid/gid/size/mtime/rdevmajor/rdevminor fields. */ 361 362 return (bid); 363} 364 365/* 366 * The function invoked by archive_read_header(). This 367 * just sets up a few things and then calls the internal 368 * tar_read_header() function below. 369 */ 370static int 371archive_read_format_tar_read_header(struct archive_read *a, 372 struct archive_entry *entry) 373{ 374 /* 375 * When converting tar archives to cpio archives, it is 376 * essential that each distinct file have a distinct inode 377 * number. To simplify this, we keep a static count here to 378 * assign fake dev/inode numbers to each tar entry. Note that 379 * pax format archives may overwrite this with something more 380 * useful. 381 * 382 * Ideally, we would track every file read from the archive so 383 * that we could assign the same dev/ino pair to hardlinks, 384 * but the memory required to store a complete lookup table is 385 * probably not worthwhile just to support the relatively 386 * obscure tar->cpio conversion case. 387 */ 388 static int default_inode; 389 static int default_dev; 390 struct tar *tar; 391 struct sparse_block *sp; 392 const char *p; 393 int r; 394 size_t l; 395 396 /* Assign default device/inode values. */ 397 archive_entry_set_dev(entry, 1 + default_dev); /* Don't use zero. */ 398 archive_entry_set_ino(entry, ++default_inode); /* Don't use zero. */ 399 /* Limit generated st_ino number to 16 bits. */ 400 if (default_inode >= 0xffff) { 401 ++default_dev; 402 default_inode = 0; 403 } 404 405 tar = (struct tar *)(a->format->data); 406 tar->entry_offset = 0; 407 while (tar->sparse_list != NULL) { 408 sp = tar->sparse_list; 409 tar->sparse_list = sp->next; 410 free(sp); 411 } 412 tar->sparse_last = NULL; 413 tar->realsize = -1; /* Mark this as "unset" */ 414 415 r = tar_read_header(a, tar, entry); 416 417 /* 418 * "non-sparse" files are really just sparse files with 419 * a single block. 420 */ 421 if (tar->sparse_list == NULL) 422 gnu_add_sparse_entry(tar, 0, tar->entry_bytes_remaining); 423 424 if (r == ARCHIVE_OK) { 425 /* 426 * "Regular" entry with trailing '/' is really 427 * directory: This is needed for certain old tar 428 * variants and even for some broken newer ones. 429 */ 430 p = archive_entry_pathname(entry); 431 l = strlen(p); 432 if (archive_entry_filetype(entry) == AE_IFREG 433 && p[l-1] == '/') 434 archive_entry_set_filetype(entry, AE_IFDIR); 435 } 436 return (r); 437} 438 439static int 440archive_read_format_tar_read_data(struct archive_read *a, 441 const void **buff, size_t *size, off_t *offset) 442{ 443 ssize_t bytes_read; 444 struct tar *tar; 445 struct sparse_block *p; 446 447 tar = (struct tar *)(a->format->data); 448 449 if (tar->sparse_gnu_pending) { 450 if (tar->sparse_gnu_major == 1 && tar->sparse_gnu_minor == 0) { 451 tar->sparse_gnu_pending = 0; 452 /* Read initial sparse map. */ 453 bytes_read = gnu_sparse_10_read(a, tar); 454 tar->entry_bytes_remaining -= bytes_read; 455 if (bytes_read < 0) 456 return (bytes_read); 457 } else { 458 *size = 0; 459 *offset = 0; 460 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 461 "Unrecognized GNU sparse file format"); 462 return (ARCHIVE_WARN); 463 } 464 tar->sparse_gnu_pending = 0; 465 } 466 467 /* Remove exhausted entries from sparse list. */ 468 while (tar->sparse_list != NULL && 469 tar->sparse_list->remaining == 0) { 470 p = tar->sparse_list; 471 tar->sparse_list = p->next; 472 free(p); 473 } 474 475 /* If we're at end of file, return EOF. */ 476 if (tar->sparse_list == NULL || tar->entry_bytes_remaining == 0) { 477 if (__archive_read_skip(a, tar->entry_padding) < 0) 478 return (ARCHIVE_FATAL); 479 tar->entry_padding = 0; 480 *buff = NULL; 481 *size = 0; 482 *offset = tar->realsize; 483 return (ARCHIVE_EOF); 484 } 485 486 *buff = __archive_read_ahead(a, 1, &bytes_read); 487 if (bytes_read < 0) 488 return (ARCHIVE_FATAL); 489 if (*buff == NULL) { 490 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 491 "Truncated tar archive"); 492 return (ARCHIVE_FATAL); 493 } 494 if (bytes_read > tar->entry_bytes_remaining) 495 bytes_read = tar->entry_bytes_remaining; 496 /* Don't read more than is available in the 497 * current sparse block. */ 498 if (tar->sparse_list->remaining < bytes_read) 499 bytes_read = tar->sparse_list->remaining; 500 *size = bytes_read; 501 *offset = tar->sparse_list->offset; 502 tar->sparse_list->remaining -= bytes_read; 503 tar->sparse_list->offset += bytes_read; 504 tar->entry_bytes_remaining -= bytes_read; 505 __archive_read_consume(a, bytes_read); 506 return (ARCHIVE_OK); 507} 508 509static int 510archive_read_format_tar_skip(struct archive_read *a) 511{ 512 int64_t bytes_skipped; 513 struct tar* tar; 514 515 tar = (struct tar *)(a->format->data); 516 517 /* 518 * Compression layer skip functions are required to either skip the 519 * length requested or fail, so we can rely upon the entire entry 520 * plus padding being skipped. 521 */ 522 bytes_skipped = __archive_read_skip(a, 523 tar->entry_bytes_remaining + tar->entry_padding); 524 if (bytes_skipped < 0) 525 return (ARCHIVE_FATAL); 526 527 tar->entry_bytes_remaining = 0; 528 tar->entry_padding = 0; 529 530 /* Free the sparse list. */ 531 gnu_clear_sparse_list(tar); 532 533 return (ARCHIVE_OK); 534} 535 536/* 537 * This function recursively interprets all of the headers associated 538 * with a single entry. 539 */ 540static int 541tar_read_header(struct archive_read *a, struct tar *tar, 542 struct archive_entry *entry) 543{ 544 ssize_t bytes; 545 int err; 546 const void *h; 547 const struct archive_entry_header_ustar *header; 548 549 /* Read 512-byte header record */ 550 h = __archive_read_ahead(a, 512, &bytes); 551 if (bytes < 0) 552 return (bytes); 553 if (bytes < 512) { /* Short read or EOF. */ 554 /* Try requesting just one byte and see what happens. */ 555 (void)__archive_read_ahead(a, 1, &bytes); 556 if (bytes == 0) { 557 /* 558 * The archive ends at a 512-byte boundary but 559 * without a proper end-of-archive marker. 560 * Yes, there are tar writers that do this; 561 * hold our nose and accept it. 562 */ 563 return (ARCHIVE_EOF); 564 } 565 /* Archive ends with a partial block; this is bad. */ 566 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 567 "Truncated tar archive"); 568 return (ARCHIVE_FATAL); 569 } 570 __archive_read_consume(a, 512); 571 572 573 /* Check for end-of-archive mark. */ 574 if (((*(const char *)h)==0) && archive_block_is_null((const unsigned char *)h)) { 575 /* Try to consume a second all-null record, as well. */ 576 h = __archive_read_ahead(a, 512, NULL); 577 if (h != NULL) 578 __archive_read_consume(a, 512); 579 archive_clear_error(&a->archive); 580 if (a->archive.archive_format_name == NULL) { 581 a->archive.archive_format = ARCHIVE_FORMAT_TAR; 582 a->archive.archive_format_name = "tar"; 583 } 584 return (ARCHIVE_EOF); 585 } 586 587 /* 588 * Note: If the checksum fails and we return ARCHIVE_RETRY, 589 * then the client is likely to just retry. This is a very 590 * crude way to search for the next valid header! 591 * 592 * TODO: Improve this by implementing a real header scan. 593 */ 594 if (!checksum(a, h)) { 595 archive_set_error(&a->archive, EINVAL, "Damaged tar archive"); 596 return (ARCHIVE_RETRY); /* Retryable: Invalid header */ 597 } 598 599 if (++tar->header_recursion_depth > 32) { 600 archive_set_error(&a->archive, EINVAL, "Too many special headers"); 601 return (ARCHIVE_WARN); 602 } 603 604 /* Determine the format variant. */ 605 header = (const struct archive_entry_header_ustar *)h; 606 switch(header->typeflag[0]) { 607 case 'A': /* Solaris tar ACL */ 608 a->archive.archive_format = ARCHIVE_FORMAT_TAR_PAX_INTERCHANGE; 609 a->archive.archive_format_name = "Solaris tar"; 610 err = header_Solaris_ACL(a, tar, entry, h); 611 break; 612 case 'g': /* POSIX-standard 'g' header. */ 613 a->archive.archive_format = ARCHIVE_FORMAT_TAR_PAX_INTERCHANGE; 614 a->archive.archive_format_name = "POSIX pax interchange format"; 615 err = header_pax_global(a, tar, entry, h); 616 break; 617 case 'K': /* Long link name (GNU tar, others) */ 618 err = header_longlink(a, tar, entry, h); 619 break; 620 case 'L': /* Long filename (GNU tar, others) */ 621 err = header_longname(a, tar, entry, h); 622 break; 623 case 'V': /* GNU volume header */ 624 err = header_volume(a, tar, entry, h); 625 break; 626 case 'X': /* Used by SUN tar; same as 'x'. */ 627 a->archive.archive_format = ARCHIVE_FORMAT_TAR_PAX_INTERCHANGE; 628 a->archive.archive_format_name = 629 "POSIX pax interchange format (Sun variant)"; 630 err = header_pax_extensions(a, tar, entry, h); 631 break; 632 case 'x': /* POSIX-standard 'x' header. */ 633 a->archive.archive_format = ARCHIVE_FORMAT_TAR_PAX_INTERCHANGE; 634 a->archive.archive_format_name = "POSIX pax interchange format"; 635 err = header_pax_extensions(a, tar, entry, h); 636 break; 637 default: 638 if (memcmp(header->magic, "ustar \0", 8) == 0) { 639 a->archive.archive_format = ARCHIVE_FORMAT_TAR_GNUTAR; 640 a->archive.archive_format_name = "GNU tar format"; 641 err = header_gnutar(a, tar, entry, h); 642 } else if (memcmp(header->magic, "ustar", 5) == 0) { 643 if (a->archive.archive_format != ARCHIVE_FORMAT_TAR_PAX_INTERCHANGE) { 644 a->archive.archive_format = ARCHIVE_FORMAT_TAR_USTAR; 645 a->archive.archive_format_name = "POSIX ustar format"; 646 } 647 err = header_ustar(a, tar, entry, h); 648 } else { 649 a->archive.archive_format = ARCHIVE_FORMAT_TAR; 650 a->archive.archive_format_name = "tar (non-POSIX)"; 651 err = header_old_tar(a, tar, entry, h); 652 } 653 } 654 --tar->header_recursion_depth; 655 /* We return warnings or success as-is. Anything else is fatal. */ 656 if (err == ARCHIVE_WARN || err == ARCHIVE_OK) 657 return (err); 658 if (err == ARCHIVE_EOF) 659 /* EOF when recursively reading a header is bad. */ 660 archive_set_error(&a->archive, EINVAL, "Damaged tar archive"); 661 return (ARCHIVE_FATAL); 662} 663 664/* 665 * Return true if block checksum is correct. 666 */ 667static int 668checksum(struct archive_read *a, const void *h) 669{ 670 const unsigned char *bytes; 671 const struct archive_entry_header_ustar *header; 672 int check, i, sum; 673 674 (void)a; /* UNUSED */ 675 bytes = (const unsigned char *)h; 676 header = (const struct archive_entry_header_ustar *)h; 677 678 /* 679 * Test the checksum. Note that POSIX specifies _unsigned_ 680 * bytes for this calculation. 681 */ 682 sum = tar_atol(header->checksum, sizeof(header->checksum)); 683 check = 0; 684 for (i = 0; i < 148; i++) 685 check += (unsigned char)bytes[i]; 686 for (; i < 156; i++) 687 check += 32; 688 for (; i < 512; i++) 689 check += (unsigned char)bytes[i]; 690 if (sum == check) 691 return (1); 692 693 /* 694 * Repeat test with _signed_ bytes, just in case this archive 695 * was created by an old BSD, Solaris, or HP-UX tar with a 696 * broken checksum calculation. 697 */ 698 check = 0; 699 for (i = 0; i < 148; i++) 700 check += (signed char)bytes[i]; 701 for (; i < 156; i++) 702 check += 32; 703 for (; i < 512; i++) 704 check += (signed char)bytes[i]; 705 if (sum == check) 706 return (1); 707 708 return (0); 709} 710 711/* 712 * Return true if this block contains only nulls. 713 */ 714static int 715archive_block_is_null(const unsigned char *p) 716{ 717 unsigned i; 718 719 for (i = 0; i < 512; i++) 720 if (*p++) 721 return (0); 722 return (1); 723} 724 725/* 726 * Interpret 'A' Solaris ACL header 727 */ 728static int 729header_Solaris_ACL(struct archive_read *a, struct tar *tar, 730 struct archive_entry *entry, const void *h) 731{ 732 const struct archive_entry_header_ustar *header; 733 size_t size; 734 int err; 735 int64_t type; 736 char *acl, *p; 737 wchar_t *wp; 738 739 /* 740 * read_body_to_string adds a NUL terminator, but we need a little 741 * more to make sure that we don't overrun acl_text later. 742 */ 743 header = (const struct archive_entry_header_ustar *)h; 744 size = tar_atol(header->size, sizeof(header->size)); 745 err = read_body_to_string(a, tar, &(tar->acl_text), h); 746 if (err != ARCHIVE_OK) 747 return (err); 748 /* Recursively read next header */ 749 err = tar_read_header(a, tar, entry); 750 if ((err != ARCHIVE_OK) && (err != ARCHIVE_WARN)) 751 return (err); 752 753 /* TODO: Examine the first characters to see if this 754 * is an AIX ACL descriptor. We'll likely never support 755 * them, but it would be polite to recognize and warn when 756 * we do see them. */ 757 758 /* Leading octal number indicates ACL type and number of entries. */ 759 p = acl = tar->acl_text.s; 760 type = 0; 761 while (*p != '\0' && p < acl + size) { 762 if (*p < '0' || *p > '7') { 763 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 764 "Malformed Solaris ACL attribute (invalid digit)"); 765 return(ARCHIVE_WARN); 766 } 767 type <<= 3; 768 type += *p - '0'; 769 if (type > 077777777) { 770 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 771 "Malformed Solaris ACL attribute (count too large)"); 772 return (ARCHIVE_WARN); 773 } 774 p++; 775 } 776 switch ((int)type & ~0777777) { 777 case 01000000: 778 /* POSIX.1e ACL */ 779 break; 780 case 03000000: 781 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 782 "Solaris NFSv4 ACLs not supported"); 783 return (ARCHIVE_WARN); 784 default: 785 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 786 "Malformed Solaris ACL attribute (unsupported type %o)", 787 (int)type); 788 return (ARCHIVE_WARN); 789 } 790 p++; 791 792 if (p >= acl + size) { 793 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 794 "Malformed Solaris ACL attribute (body overflow)"); 795 return(ARCHIVE_WARN); 796 } 797 798 /* ACL text is null-terminated; find the end. */ 799 size -= (p - acl); 800 acl = p; 801 802 while (*p != '\0' && p < acl + size) 803 p++; 804 805 wp = utf8_decode(tar, acl, p - acl); 806 err = __archive_entry_acl_parse_w(entry, wp, 807 ARCHIVE_ENTRY_ACL_TYPE_ACCESS); 808 if (err != ARCHIVE_OK) 809 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 810 "Malformed Solaris ACL attribute (unparsable)"); 811 return (err); 812} 813 814/* 815 * Interpret 'K' long linkname header. 816 */ 817static int 818header_longlink(struct archive_read *a, struct tar *tar, 819 struct archive_entry *entry, const void *h) 820{ 821 int err; 822 823 err = read_body_to_string(a, tar, &(tar->longlink), h); 824 if (err != ARCHIVE_OK) 825 return (err); 826 err = tar_read_header(a, tar, entry); 827 if ((err != ARCHIVE_OK) && (err != ARCHIVE_WARN)) 828 return (err); 829 /* Set symlink if symlink already set, else hardlink. */ 830 archive_entry_copy_link(entry, tar->longlink.s); 831 return (ARCHIVE_OK); 832} 833 834/* 835 * Interpret 'L' long filename header. 836 */ 837static int 838header_longname(struct archive_read *a, struct tar *tar, 839 struct archive_entry *entry, const void *h) 840{ 841 int err; 842 843 err = read_body_to_string(a, tar, &(tar->longname), h); 844 if (err != ARCHIVE_OK) 845 return (err); 846 /* Read and parse "real" header, then override name. */ 847 err = tar_read_header(a, tar, entry); 848 if ((err != ARCHIVE_OK) && (err != ARCHIVE_WARN)) 849 return (err); 850 archive_entry_copy_pathname(entry, tar->longname.s); 851 return (ARCHIVE_OK); 852} 853 854 855/* 856 * Interpret 'V' GNU tar volume header. 857 */ 858static int 859header_volume(struct archive_read *a, struct tar *tar, 860 struct archive_entry *entry, const void *h) 861{ 862 (void)h; 863 864 /* Just skip this and read the next header. */ 865 return (tar_read_header(a, tar, entry)); 866} 867 868/* 869 * Read body of an archive entry into an archive_string object. 870 */ 871static int 872read_body_to_string(struct archive_read *a, struct tar *tar, 873 struct archive_string *as, const void *h) 874{ 875 off_t size, padded_size; 876 const struct archive_entry_header_ustar *header; 877 const void *src; 878 879 (void)tar; /* UNUSED */ 880 header = (const struct archive_entry_header_ustar *)h; 881 size = tar_atol(header->size, sizeof(header->size)); 882 if ((size > 1048576) || (size < 0)) { 883 archive_set_error(&a->archive, EINVAL, 884 "Special header too large"); 885 return (ARCHIVE_FATAL); 886 } 887 888 /* Fail if we can't make our buffer big enough. */ 889 if (archive_string_ensure(as, size+1) == NULL) { 890 archive_set_error(&a->archive, ENOMEM, 891 "No memory"); 892 return (ARCHIVE_FATAL); 893 } 894 895 /* Read the body into the string. */ 896 padded_size = (size + 511) & ~ 511; 897 src = __archive_read_ahead(a, padded_size, NULL); 898 if (src == NULL) 899 return (ARCHIVE_FATAL); 900 memcpy(as->s, src, size); 901 __archive_read_consume(a, padded_size); 902 as->s[size] = '\0'; 903 return (ARCHIVE_OK); 904} 905 906/* 907 * Parse out common header elements. 908 * 909 * This would be the same as header_old_tar, except that the 910 * filename is handled slightly differently for old and POSIX 911 * entries (POSIX entries support a 'prefix'). This factoring 912 * allows header_old_tar and header_ustar 913 * to handle filenames differently, while still putting most of the 914 * common parsing into one place. 915 */ 916static int 917header_common(struct archive_read *a, struct tar *tar, 918 struct archive_entry *entry, const void *h) 919{ 920 const struct archive_entry_header_ustar *header; 921 char tartype; 922 923 (void)a; /* UNUSED */ 924 925 header = (const struct archive_entry_header_ustar *)h; 926 if (header->linkname[0]) 927 archive_strncpy(&(tar->entry_linkpath), header->linkname, 928 sizeof(header->linkname)); 929 else 930 archive_string_empty(&(tar->entry_linkpath)); 931 932 /* Parse out the numeric fields (all are octal) */ 933 archive_entry_set_mode(entry, tar_atol(header->mode, sizeof(header->mode))); 934 archive_entry_set_uid(entry, tar_atol(header->uid, sizeof(header->uid))); 935 archive_entry_set_gid(entry, tar_atol(header->gid, sizeof(header->gid))); 936 tar->entry_bytes_remaining = tar_atol(header->size, sizeof(header->size)); 937 tar->realsize = tar->entry_bytes_remaining; 938 archive_entry_set_size(entry, tar->entry_bytes_remaining); 939 archive_entry_set_mtime(entry, tar_atol(header->mtime, sizeof(header->mtime)), 0); 940 941 /* Handle the tar type flag appropriately. */ 942 tartype = header->typeflag[0]; 943 944 switch (tartype) { 945 case '1': /* Hard link */ 946 archive_entry_copy_hardlink(entry, tar->entry_linkpath.s); 947 /* 948 * The following may seem odd, but: Technically, tar 949 * does not store the file type for a "hard link" 950 * entry, only the fact that it is a hard link. So, I 951 * leave the type zero normally. But, pax interchange 952 * format allows hard links to have data, which 953 * implies that the underlying entry is a regular 954 * file. 955 */ 956 if (archive_entry_size(entry) > 0) 957 archive_entry_set_filetype(entry, AE_IFREG); 958 959 /* 960 * A tricky point: Traditionally, tar readers have 961 * ignored the size field when reading hardlink 962 * entries, and some writers put non-zero sizes even 963 * though the body is empty. POSIX blessed this 964 * convention in the 1988 standard, but broke with 965 * this tradition in 2001 by permitting hardlink 966 * entries to store valid bodies in pax interchange 967 * format, but not in ustar format. Since there is no 968 * hard and fast way to distinguish pax interchange 969 * from earlier archives (the 'x' and 'g' entries are 970 * optional, after all), we need a heuristic. 971 */ 972 if (archive_entry_size(entry) == 0) { 973 /* If the size is already zero, we're done. */ 974 } else if (a->archive.archive_format 975 == ARCHIVE_FORMAT_TAR_PAX_INTERCHANGE) { 976 /* Definitely pax extended; must obey hardlink size. */ 977 } else if (a->archive.archive_format == ARCHIVE_FORMAT_TAR 978 || a->archive.archive_format == ARCHIVE_FORMAT_TAR_GNUTAR) 979 { 980 /* Old-style or GNU tar: we must ignore the size. */ 981 archive_entry_set_size(entry, 0); 982 tar->entry_bytes_remaining = 0; 983 } else if (archive_read_format_tar_bid(a) > 50) { 984 /* 985 * We don't know if it's pax: If the bid 986 * function sees a valid ustar header 987 * immediately following, then let's ignore 988 * the hardlink size. 989 */ 990 archive_entry_set_size(entry, 0); 991 tar->entry_bytes_remaining = 0; 992 } 993 /* 994 * TODO: There are still two cases I'd like to handle: 995 * = a ustar non-pax archive with a hardlink entry at 996 * end-of-archive. (Look for block of nulls following?) 997 * = a pax archive that has not seen any pax headers 998 * and has an entry which is a hardlink entry storing 999 * a body containing an uncompressed tar archive. 1000 * The first is worth addressing; I don't see any reliable 1001 * way to deal with the second possibility. 1002 */ 1003 break; 1004 case '2': /* Symlink */ 1005 archive_entry_set_filetype(entry, AE_IFLNK); 1006 archive_entry_set_size(entry, 0); 1007 tar->entry_bytes_remaining = 0; 1008 archive_entry_copy_symlink(entry, tar->entry_linkpath.s); 1009 break; 1010 case '3': /* Character device */ 1011 archive_entry_set_filetype(entry, AE_IFCHR); 1012 archive_entry_set_size(entry, 0); 1013 tar->entry_bytes_remaining = 0; 1014 break; 1015 case '4': /* Block device */ 1016 archive_entry_set_filetype(entry, AE_IFBLK); 1017 archive_entry_set_size(entry, 0); 1018 tar->entry_bytes_remaining = 0; 1019 break; 1020 case '5': /* Dir */ 1021 archive_entry_set_filetype(entry, AE_IFDIR); 1022 archive_entry_set_size(entry, 0); 1023 tar->entry_bytes_remaining = 0; 1024 break; 1025 case '6': /* FIFO device */ 1026 archive_entry_set_filetype(entry, AE_IFIFO); 1027 archive_entry_set_size(entry, 0); 1028 tar->entry_bytes_remaining = 0; 1029 break; 1030 case 'D': /* GNU incremental directory type */ 1031 /* 1032 * No special handling is actually required here. 1033 * It might be nice someday to preprocess the file list and 1034 * provide it to the client, though. 1035 */ 1036 archive_entry_set_filetype(entry, AE_IFDIR); 1037 break; 1038 case 'M': /* GNU "Multi-volume" (remainder of file from last archive)*/ 1039 /* 1040 * As far as I can tell, this is just like a regular file 1041 * entry, except that the contents should be _appended_ to 1042 * the indicated file at the indicated offset. This may 1043 * require some API work to fully support. 1044 */ 1045 break; 1046 case 'N': /* Old GNU "long filename" entry. */ 1047 /* The body of this entry is a script for renaming 1048 * previously-extracted entries. Ugh. It will never 1049 * be supported by libarchive. */ 1050 archive_entry_set_filetype(entry, AE_IFREG); 1051 break; 1052 case 'S': /* GNU sparse files */ 1053 /* 1054 * Sparse files are really just regular files with 1055 * sparse information in the extended area. 1056 */ 1057 /* FALLTHROUGH */ 1058 default: /* Regular file and non-standard types */ 1059 /* 1060 * Per POSIX: non-recognized types should always be 1061 * treated as regular files. 1062 */ 1063 archive_entry_set_filetype(entry, AE_IFREG); 1064 break; 1065 } 1066 return (0); 1067} 1068 1069/* 1070 * Parse out header elements for "old-style" tar archives. 1071 */ 1072static int 1073header_old_tar(struct archive_read *a, struct tar *tar, 1074 struct archive_entry *entry, const void *h) 1075{ 1076 const struct archive_entry_header_ustar *header; 1077 1078 /* Copy filename over (to ensure null termination). */ 1079 header = (const struct archive_entry_header_ustar *)h; 1080 archive_strncpy(&(tar->entry_pathname), header->name, sizeof(header->name)); 1081 archive_entry_copy_pathname(entry, tar->entry_pathname.s); 1082 1083 /* Grab rest of common fields */ 1084 header_common(a, tar, entry, h); 1085 1086 tar->entry_padding = 0x1ff & (-tar->entry_bytes_remaining); 1087 return (0); 1088} 1089 1090/* 1091 * Parse a file header for a pax extended archive entry. 1092 */ 1093static int 1094header_pax_global(struct archive_read *a, struct tar *tar, 1095 struct archive_entry *entry, const void *h) 1096{ 1097 int err; 1098 1099 err = read_body_to_string(a, tar, &(tar->pax_global), h); 1100 if (err != ARCHIVE_OK) 1101 return (err); 1102 err = tar_read_header(a, tar, entry); 1103 return (err); 1104} 1105 1106static int 1107header_pax_extensions(struct archive_read *a, struct tar *tar, 1108 struct archive_entry *entry, const void *h) 1109{ 1110 int err, err2; 1111 1112 err = read_body_to_string(a, tar, &(tar->pax_header), h); 1113 if (err != ARCHIVE_OK) 1114 return (err); 1115 1116 /* Parse the next header. */ 1117 err = tar_read_header(a, tar, entry); 1118 if ((err != ARCHIVE_OK) && (err != ARCHIVE_WARN)) 1119 return (err); 1120 1121 /* 1122 * TODO: Parse global/default options into 'entry' struct here 1123 * before handling file-specific options. 1124 * 1125 * This design (parse standard header, then overwrite with pax 1126 * extended attribute data) usually works well, but isn't ideal; 1127 * it would be better to parse the pax extended attributes first 1128 * and then skip any fields in the standard header that were 1129 * defined in the pax header. 1130 */ 1131 err2 = pax_header(a, tar, entry, tar->pax_header.s); 1132 err = err_combine(err, err2); 1133 tar->entry_padding = 0x1ff & (-tar->entry_bytes_remaining); 1134 return (err); 1135} 1136 1137 1138/* 1139 * Parse a file header for a Posix "ustar" archive entry. This also 1140 * handles "pax" or "extended ustar" entries. 1141 */ 1142static int 1143header_ustar(struct archive_read *a, struct tar *tar, 1144 struct archive_entry *entry, const void *h) 1145{ 1146 const struct archive_entry_header_ustar *header; 1147 struct archive_string *as; 1148 1149 header = (const struct archive_entry_header_ustar *)h; 1150 1151 /* Copy name into an internal buffer to ensure null-termination. */ 1152 as = &(tar->entry_pathname); 1153 if (header->prefix[0]) { 1154 archive_strncpy(as, header->prefix, sizeof(header->prefix)); 1155 if (as->s[archive_strlen(as) - 1] != '/') 1156 archive_strappend_char(as, '/'); 1157 archive_strncat(as, header->name, sizeof(header->name)); 1158 } else 1159 archive_strncpy(as, header->name, sizeof(header->name)); 1160 1161 archive_entry_copy_pathname(entry, as->s); 1162 1163 /* Handle rest of common fields. */ 1164 header_common(a, tar, entry, h); 1165 1166 /* Handle POSIX ustar fields. */ 1167 archive_strncpy(&(tar->entry_uname), header->uname, 1168 sizeof(header->uname)); 1169 archive_entry_copy_uname(entry, tar->entry_uname.s); 1170 1171 archive_strncpy(&(tar->entry_gname), header->gname, 1172 sizeof(header->gname)); 1173 archive_entry_copy_gname(entry, tar->entry_gname.s); 1174 1175 /* Parse out device numbers only for char and block specials. */ 1176 if (header->typeflag[0] == '3' || header->typeflag[0] == '4') { 1177 archive_entry_set_rdevmajor(entry, 1178 tar_atol(header->rdevmajor, sizeof(header->rdevmajor))); 1179 archive_entry_set_rdevminor(entry, 1180 tar_atol(header->rdevminor, sizeof(header->rdevminor))); 1181 } 1182 1183 tar->entry_padding = 0x1ff & (-tar->entry_bytes_remaining); 1184 1185 return (0); 1186} 1187 1188 1189/* 1190 * Parse the pax extended attributes record. 1191 * 1192 * Returns non-zero if there's an error in the data. 1193 */ 1194static int 1195pax_header(struct archive_read *a, struct tar *tar, 1196 struct archive_entry *entry, char *attr) 1197{ 1198 size_t attr_length, l, line_length; 1199 char *p; 1200 char *key, *value; 1201 int err, err2; 1202 1203 attr_length = strlen(attr); 1204 tar->pax_hdrcharset_binary = 0; 1205 archive_string_empty(&(tar->entry_gname)); 1206 archive_string_empty(&(tar->entry_linkpath)); 1207 archive_string_empty(&(tar->entry_pathname)); 1208 archive_string_empty(&(tar->entry_pathname_override)); 1209 archive_string_empty(&(tar->entry_uname)); 1210 err = ARCHIVE_OK; 1211 while (attr_length > 0) { 1212 /* Parse decimal length field at start of line. */ 1213 line_length = 0; 1214 l = attr_length; 1215 p = attr; /* Record start of line. */ 1216 while (l>0) { 1217 if (*p == ' ') { 1218 p++; 1219 l--; 1220 break; 1221 } 1222 if (*p < '0' || *p > '9') { 1223 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 1224 "Ignoring malformed pax extended attributes"); 1225 return (ARCHIVE_WARN); 1226 } 1227 line_length *= 10; 1228 line_length += *p - '0'; 1229 if (line_length > 999999) { 1230 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 1231 "Rejecting pax extended attribute > 1MB"); 1232 return (ARCHIVE_WARN); 1233 } 1234 p++; 1235 l--; 1236 } 1237 1238 /* 1239 * Parsed length must be no bigger than available data, 1240 * at least 1, and the last character of the line must 1241 * be '\n'. 1242 */ 1243 if (line_length > attr_length 1244 || line_length < 1 1245 || attr[line_length - 1] != '\n') 1246 { 1247 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 1248 "Ignoring malformed pax extended attribute"); 1249 return (ARCHIVE_WARN); 1250 } 1251 1252 /* Null-terminate the line. */ 1253 attr[line_length - 1] = '\0'; 1254 1255 /* Find end of key and null terminate it. */ 1256 key = p; 1257 if (key[0] == '=') 1258 return (-1); 1259 while (*p && *p != '=') 1260 ++p; 1261 if (*p == '\0') { 1262 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 1263 "Invalid pax extended attributes"); 1264 return (ARCHIVE_WARN); 1265 } 1266 *p = '\0'; 1267 1268 /* Identify null-terminated 'value' portion. */ 1269 value = p + 1; 1270 1271 /* Identify this attribute and set it in the entry. */ 1272 err2 = pax_attribute(tar, entry, key, value); 1273 err = err_combine(err, err2); 1274 1275 /* Skip to next line */ 1276 attr += line_length; 1277 attr_length -= line_length; 1278 } 1279 if (archive_strlen(&(tar->entry_gname)) > 0) { 1280 value = tar->entry_gname.s; 1281 if (tar->pax_hdrcharset_binary) 1282 archive_entry_copy_gname(entry, value); 1283 else { 1284 if (!archive_entry_update_gname_utf8(entry, value)) { 1285 err = ARCHIVE_WARN; 1286 archive_set_error(&a->archive, 1287 ARCHIVE_ERRNO_FILE_FORMAT, 1288 "Gname in pax header can't " 1289 "be converted to current locale."); 1290 } 1291 } 1292 } 1293 if (archive_strlen(&(tar->entry_linkpath)) > 0) { 1294 value = tar->entry_linkpath.s; 1295 if (tar->pax_hdrcharset_binary) 1296 archive_entry_copy_link(entry, value); 1297 else { 1298 if (!archive_entry_update_link_utf8(entry, value)) { 1299 err = ARCHIVE_WARN; 1300 archive_set_error(&a->archive, 1301 ARCHIVE_ERRNO_FILE_FORMAT, 1302 "Linkname in pax header can't " 1303 "be converted to current locale."); 1304 } 1305 } 1306 } 1307 /* 1308 * Some extensions (such as the GNU sparse file extensions) 1309 * deliberately store a synthetic name under the regular 'path' 1310 * attribute and the real file name under a different attribute. 1311 * Since we're supposed to not care about the order, we 1312 * have no choice but to store all of the various filenames 1313 * we find and figure it all out afterwards. This is the 1314 * figuring out part. 1315 */ 1316 value = NULL; 1317 if (archive_strlen(&(tar->entry_pathname_override)) > 0) 1318 value = tar->entry_pathname_override.s; 1319 else if (archive_strlen(&(tar->entry_pathname)) > 0) 1320 value = tar->entry_pathname.s; 1321 if (value != NULL) { 1322 if (tar->pax_hdrcharset_binary) 1323 archive_entry_copy_pathname(entry, value); 1324 else { 1325 if (!archive_entry_update_pathname_utf8(entry, value)) { 1326 err = ARCHIVE_WARN; 1327 archive_set_error(&a->archive, 1328 ARCHIVE_ERRNO_FILE_FORMAT, 1329 "Pathname in pax header can't be " 1330 "converted to current locale."); 1331 } 1332 } 1333 } 1334 if (archive_strlen(&(tar->entry_uname)) > 0) { 1335 value = tar->entry_uname.s; 1336 if (tar->pax_hdrcharset_binary) 1337 archive_entry_copy_uname(entry, value); 1338 else { 1339 if (!archive_entry_update_uname_utf8(entry, value)) { 1340 err = ARCHIVE_WARN; 1341 archive_set_error(&a->archive, 1342 ARCHIVE_ERRNO_FILE_FORMAT, 1343 "Uname in pax header can't " 1344 "be converted to current locale."); 1345 } 1346 } 1347 } 1348 return (err); 1349} 1350 1351static int 1352pax_attribute_xattr(struct archive_entry *entry, 1353 char *name, char *value) 1354{ 1355 char *name_decoded; 1356 void *value_decoded; 1357 size_t value_len; 1358 1359 if (strlen(name) < 18 || (strncmp(name, "LIBARCHIVE.xattr.", 17)) != 0) 1360 return 3; 1361 1362 name += 17; 1363 1364 /* URL-decode name */ 1365 name_decoded = url_decode(name); 1366 if (name_decoded == NULL) 1367 return 2; 1368 1369 /* Base-64 decode value */ 1370 value_decoded = base64_decode(value, strlen(value), &value_len); 1371 if (value_decoded == NULL) { 1372 free(name_decoded); 1373 return 1; 1374 } 1375 1376 archive_entry_xattr_add_entry(entry, name_decoded, 1377 value_decoded, value_len); 1378 1379 free(name_decoded); 1380 free(value_decoded); 1381 return 0; 1382} 1383 1384/* 1385 * Parse a single key=value attribute. key/value pointers are 1386 * assumed to point into reasonably long-lived storage. 1387 * 1388 * Note that POSIX reserves all-lowercase keywords. Vendor-specific 1389 * extensions should always have keywords of the form "VENDOR.attribute" 1390 * In particular, it's quite feasible to support many different 1391 * vendor extensions here. I'm using "LIBARCHIVE" for extensions 1392 * unique to this library. 1393 * 1394 * Investigate other vendor-specific extensions and see if 1395 * any of them look useful. 1396 */ 1397static int 1398pax_attribute(struct tar *tar, struct archive_entry *entry, 1399 char *key, char *value) 1400{ 1401 int64_t s; 1402 long n; 1403 wchar_t *wp; 1404 1405 switch (key[0]) { 1406 case 'G': 1407 /* GNU "0.0" sparse pax format. */ 1408 if (strcmp(key, "GNU.sparse.numblocks") == 0) { 1409 tar->sparse_offset = -1; 1410 tar->sparse_numbytes = -1; 1411 tar->sparse_gnu_major = 0; 1412 tar->sparse_gnu_minor = 0; 1413 } 1414 if (strcmp(key, "GNU.sparse.offset") == 0) { 1415 tar->sparse_offset = tar_atol10(value, strlen(value)); 1416 if (tar->sparse_numbytes != -1) { 1417 gnu_add_sparse_entry(tar, 1418 tar->sparse_offset, tar->sparse_numbytes); 1419 tar->sparse_offset = -1; 1420 tar->sparse_numbytes = -1; 1421 } 1422 } 1423 if (strcmp(key, "GNU.sparse.numbytes") == 0) { 1424 tar->sparse_numbytes = tar_atol10(value, strlen(value)); 1425 if (tar->sparse_numbytes != -1) { 1426 gnu_add_sparse_entry(tar, 1427 tar->sparse_offset, tar->sparse_numbytes); 1428 tar->sparse_offset = -1; 1429 tar->sparse_numbytes = -1; 1430 } 1431 } 1432 if (strcmp(key, "GNU.sparse.size") == 0) { 1433 tar->realsize = tar_atol10(value, strlen(value)); 1434 archive_entry_set_size(entry, tar->realsize); 1435 } 1436 1437 /* GNU "0.1" sparse pax format. */ 1438 if (strcmp(key, "GNU.sparse.map") == 0) { 1439 tar->sparse_gnu_major = 0; 1440 tar->sparse_gnu_minor = 1; 1441 if (gnu_sparse_01_parse(tar, value) != ARCHIVE_OK) 1442 return (ARCHIVE_WARN); 1443 } 1444 1445 /* GNU "1.0" sparse pax format */ 1446 if (strcmp(key, "GNU.sparse.major") == 0) { 1447 tar->sparse_gnu_major = tar_atol10(value, strlen(value)); 1448 tar->sparse_gnu_pending = 1; 1449 } 1450 if (strcmp(key, "GNU.sparse.minor") == 0) { 1451 tar->sparse_gnu_minor = tar_atol10(value, strlen(value)); 1452 tar->sparse_gnu_pending = 1; 1453 } 1454 if (strcmp(key, "GNU.sparse.name") == 0) { 1455 /* 1456 * The real filename; when storing sparse 1457 * files, GNU tar puts a synthesized name into 1458 * the regular 'path' attribute in an attempt 1459 * to limit confusion. ;-) 1460 */ 1461 archive_strcpy(&(tar->entry_pathname_override), value); 1462 } 1463 if (strcmp(key, "GNU.sparse.realsize") == 0) { 1464 tar->realsize = tar_atol10(value, strlen(value)); 1465 archive_entry_set_size(entry, tar->realsize); 1466 } 1467 break; 1468 case 'L': 1469 /* Our extensions */ 1470/* TODO: Handle arbitrary extended attributes... */ 1471/* 1472 if (strcmp(key, "LIBARCHIVE.xxxxxxx")==0) 1473 archive_entry_set_xxxxxx(entry, value); 1474*/ 1475 if (strcmp(key, "LIBARCHIVE.creationtime")==0) { 1476 pax_time(value, &s, &n); 1477 archive_entry_set_birthtime(entry, s, n); 1478 } 1479 if (strncmp(key, "LIBARCHIVE.xattr.", 17)==0) 1480 pax_attribute_xattr(entry, key, value); 1481 break; 1482 case 'S': 1483 /* We support some keys used by the "star" archiver */ 1484 if (strcmp(key, "SCHILY.acl.access")==0) { 1485 wp = utf8_decode(tar, value, strlen(value)); 1486 /* TODO: if (wp == NULL) */ 1487 __archive_entry_acl_parse_w(entry, wp, 1488 ARCHIVE_ENTRY_ACL_TYPE_ACCESS); 1489 } else if (strcmp(key, "SCHILY.acl.default")==0) { 1490 wp = utf8_decode(tar, value, strlen(value)); 1491 /* TODO: if (wp == NULL) */ 1492 __archive_entry_acl_parse_w(entry, wp, 1493 ARCHIVE_ENTRY_ACL_TYPE_DEFAULT); 1494 } else if (strcmp(key, "SCHILY.devmajor")==0) { 1495 archive_entry_set_rdevmajor(entry, 1496 tar_atol10(value, strlen(value))); 1497 } else if (strcmp(key, "SCHILY.devminor")==0) { 1498 archive_entry_set_rdevminor(entry, 1499 tar_atol10(value, strlen(value))); 1500 } else if (strcmp(key, "SCHILY.fflags")==0) { 1501 archive_entry_copy_fflags_text(entry, value); 1502 } else if (strcmp(key, "SCHILY.dev")==0) { 1503 archive_entry_set_dev(entry, 1504 tar_atol10(value, strlen(value))); 1505 } else if (strcmp(key, "SCHILY.ino")==0) { 1506 archive_entry_set_ino(entry, 1507 tar_atol10(value, strlen(value))); 1508 } else if (strcmp(key, "SCHILY.nlink")==0) { 1509 archive_entry_set_nlink(entry, 1510 tar_atol10(value, strlen(value))); 1511 } else if (strcmp(key, "SCHILY.realsize")==0) { 1512 tar->realsize = tar_atol10(value, strlen(value)); 1513 archive_entry_set_size(entry, tar->realsize); 1514 } 1515 break; 1516 case 'a': 1517 if (strcmp(key, "atime")==0) { 1518 pax_time(value, &s, &n); 1519 archive_entry_set_atime(entry, s, n); 1520 } 1521 break; 1522 case 'c': 1523 if (strcmp(key, "ctime")==0) { 1524 pax_time(value, &s, &n); 1525 archive_entry_set_ctime(entry, s, n); 1526 } else if (strcmp(key, "charset")==0) { 1527 /* TODO: Publish charset information in entry. */ 1528 } else if (strcmp(key, "comment")==0) { 1529 /* TODO: Publish comment in entry. */ 1530 } 1531 break; 1532 case 'g': 1533 if (strcmp(key, "gid")==0) { 1534 archive_entry_set_gid(entry, 1535 tar_atol10(value, strlen(value))); 1536 } else if (strcmp(key, "gname")==0) { 1537 archive_strcpy(&(tar->entry_gname), value); 1538 } 1539 break; 1540 case 'h': 1541 if (strcmp(key, "hdrcharset") == 0) { 1542 if (strcmp(value, "BINARY") == 0) 1543 tar->pax_hdrcharset_binary = 1; 1544 else if (strcmp(value, "ISO-IR 10646 2000 UTF-8") == 0) 1545 tar->pax_hdrcharset_binary = 0; 1546 else { 1547 /* TODO: Warn about unsupported hdrcharset */ 1548 } 1549 } 1550 break; 1551 case 'l': 1552 /* pax interchange doesn't distinguish hardlink vs. symlink. */ 1553 if (strcmp(key, "linkpath")==0) { 1554 archive_strcpy(&(tar->entry_linkpath), value); 1555 } 1556 break; 1557 case 'm': 1558 if (strcmp(key, "mtime")==0) { 1559 pax_time(value, &s, &n); 1560 archive_entry_set_mtime(entry, s, n); 1561 } 1562 break; 1563 case 'p': 1564 if (strcmp(key, "path")==0) { 1565 archive_strcpy(&(tar->entry_pathname), value); 1566 } 1567 break; 1568 case 'r': 1569 /* POSIX has reserved 'realtime.*' */ 1570 break; 1571 case 's': 1572 /* POSIX has reserved 'security.*' */ 1573 /* Someday: if (strcmp(key, "security.acl")==0) { ... } */ 1574 if (strcmp(key, "size")==0) { 1575 /* "size" is the size of the data in the entry. */ 1576 tar->entry_bytes_remaining 1577 = tar_atol10(value, strlen(value)); 1578 /* 1579 * But, "size" is not necessarily the size of 1580 * the file on disk; if this is a sparse file, 1581 * the disk size may have already been set from 1582 * GNU.sparse.realsize or GNU.sparse.size or 1583 * an old GNU header field or SCHILY.realsize 1584 * or .... 1585 */ 1586 if (tar->realsize < 0) { 1587 archive_entry_set_size(entry, 1588 tar->entry_bytes_remaining); 1589 tar->realsize 1590 = tar->entry_bytes_remaining; 1591 } 1592 } 1593 break; 1594 case 'u': 1595 if (strcmp(key, "uid")==0) { 1596 archive_entry_set_uid(entry, 1597 tar_atol10(value, strlen(value))); 1598 } else if (strcmp(key, "uname")==0) { 1599 archive_strcpy(&(tar->entry_uname), value); 1600 } 1601 break; 1602 } 1603 return (0); 1604} 1605 1606 1607 1608/* 1609 * parse a decimal time value, which may include a fractional portion 1610 */ 1611static void 1612pax_time(const char *p, int64_t *ps, long *pn) 1613{ 1614 char digit; 1615 int64_t s; 1616 unsigned long l; 1617 int sign; 1618 int64_t limit, last_digit_limit; 1619 1620 limit = INT64_MAX / 10; 1621 last_digit_limit = INT64_MAX % 10; 1622 1623 s = 0; 1624 sign = 1; 1625 if (*p == '-') { 1626 sign = -1; 1627 p++; 1628 } 1629 while (*p >= '0' && *p <= '9') { 1630 digit = *p - '0'; 1631 if (s > limit || 1632 (s == limit && digit > last_digit_limit)) { 1633 s = INT64_MAX; 1634 break; 1635 } 1636 s = (s * 10) + digit; 1637 ++p; 1638 } 1639 1640 *ps = s * sign; 1641 1642 /* Calculate nanoseconds. */ 1643 *pn = 0; 1644 1645 if (*p != '.') 1646 return; 1647 1648 l = 100000000UL; 1649 do { 1650 ++p; 1651 if (*p >= '0' && *p <= '9') 1652 *pn += (*p - '0') * l; 1653 else 1654 break; 1655 } while (l /= 10); 1656} 1657 1658/* 1659 * Parse GNU tar header 1660 */ 1661static int 1662header_gnutar(struct archive_read *a, struct tar *tar, 1663 struct archive_entry *entry, const void *h) 1664{ 1665 const struct archive_entry_header_gnutar *header; 1666 1667 (void)a; 1668 1669 /* 1670 * GNU header is like POSIX ustar, except 'prefix' is 1671 * replaced with some other fields. This also means the 1672 * filename is stored as in old-style archives. 1673 */ 1674 1675 /* Grab fields common to all tar variants. */ 1676 header_common(a, tar, entry, h); 1677 1678 /* Copy filename over (to ensure null termination). */ 1679 header = (const struct archive_entry_header_gnutar *)h; 1680 archive_strncpy(&(tar->entry_pathname), header->name, 1681 sizeof(header->name)); 1682 archive_entry_copy_pathname(entry, tar->entry_pathname.s); 1683 1684 /* Fields common to ustar and GNU */ 1685 /* XXX Can the following be factored out since it's common 1686 * to ustar and gnu tar? Is it okay to move it down into 1687 * header_common, perhaps? */ 1688 archive_strncpy(&(tar->entry_uname), 1689 header->uname, sizeof(header->uname)); 1690 archive_entry_copy_uname(entry, tar->entry_uname.s); 1691 1692 archive_strncpy(&(tar->entry_gname), 1693 header->gname, sizeof(header->gname)); 1694 archive_entry_copy_gname(entry, tar->entry_gname.s); 1695 1696 /* Parse out device numbers only for char and block specials */ 1697 if (header->typeflag[0] == '3' || header->typeflag[0] == '4') { 1698 archive_entry_set_rdevmajor(entry, 1699 tar_atol(header->rdevmajor, sizeof(header->rdevmajor))); 1700 archive_entry_set_rdevminor(entry, 1701 tar_atol(header->rdevminor, sizeof(header->rdevminor))); 1702 } else 1703 archive_entry_set_rdev(entry, 0); 1704 1705 tar->entry_padding = 0x1ff & (-tar->entry_bytes_remaining); 1706 1707 /* Grab GNU-specific fields. */ 1708 archive_entry_set_atime(entry, 1709 tar_atol(header->atime, sizeof(header->atime)), 0); 1710 archive_entry_set_ctime(entry, 1711 tar_atol(header->ctime, sizeof(header->ctime)), 0); 1712 if (header->realsize[0] != 0) { 1713 tar->realsize 1714 = tar_atol(header->realsize, sizeof(header->realsize)); 1715 archive_entry_set_size(entry, tar->realsize); 1716 } 1717 1718 if (header->sparse[0].offset[0] != 0) { 1719 gnu_sparse_old_read(a, tar, header); 1720 } else { 1721 if (header->isextended[0] != 0) { 1722 /* XXX WTF? XXX */ 1723 } 1724 } 1725 1726 return (0); 1727} 1728 1729static void 1730gnu_add_sparse_entry(struct tar *tar, off_t offset, off_t remaining) 1731{ 1732 struct sparse_block *p; 1733 1734 p = (struct sparse_block *)malloc(sizeof(*p)); 1735 if (p == NULL) 1736 __archive_errx(1, "Out of memory"); 1737 memset(p, 0, sizeof(*p)); 1738 if (tar->sparse_last != NULL) 1739 tar->sparse_last->next = p; 1740 else 1741 tar->sparse_list = p; 1742 tar->sparse_last = p; 1743 p->offset = offset; 1744 p->remaining = remaining; 1745} 1746 1747static void 1748gnu_clear_sparse_list(struct tar *tar) 1749{ 1750 struct sparse_block *p; 1751 1752 while (tar->sparse_list != NULL) { 1753 p = tar->sparse_list; 1754 tar->sparse_list = p->next; 1755 free(p); 1756 } 1757 tar->sparse_last = NULL; 1758} 1759 1760/* 1761 * GNU tar old-format sparse data. 1762 * 1763 * GNU old-format sparse data is stored in a fixed-field 1764 * format. Offset/size values are 11-byte octal fields (same 1765 * format as 'size' field in ustart header). These are 1766 * stored in the header, allocating subsequent header blocks 1767 * as needed. Extending the header in this way is a pretty 1768 * severe POSIX violation; this design has earned GNU tar a 1769 * lot of criticism. 1770 */ 1771 1772static int 1773gnu_sparse_old_read(struct archive_read *a, struct tar *tar, 1774 const struct archive_entry_header_gnutar *header) 1775{ 1776 ssize_t bytes_read; 1777 const void *data; 1778 struct extended { 1779 struct gnu_sparse sparse[21]; 1780 char isextended[1]; 1781 char padding[7]; 1782 }; 1783 const struct extended *ext; 1784 1785 gnu_sparse_old_parse(tar, header->sparse, 4); 1786 if (header->isextended[0] == 0) 1787 return (ARCHIVE_OK); 1788 1789 do { 1790 data = __archive_read_ahead(a, 512, &bytes_read); 1791 if (bytes_read < 0) 1792 return (ARCHIVE_FATAL); 1793 if (bytes_read < 512) { 1794 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1795 "Truncated tar archive " 1796 "detected while reading sparse file data"); 1797 return (ARCHIVE_FATAL); 1798 } 1799 __archive_read_consume(a, 512); 1800 ext = (const struct extended *)data; 1801 gnu_sparse_old_parse(tar, ext->sparse, 21); 1802 } while (ext->isextended[0] != 0); 1803 if (tar->sparse_list != NULL) 1804 tar->entry_offset = tar->sparse_list->offset; 1805 return (ARCHIVE_OK); 1806} 1807 1808static void 1809gnu_sparse_old_parse(struct tar *tar, 1810 const struct gnu_sparse *sparse, int length) 1811{ 1812 while (length > 0 && sparse->offset[0] != 0) { 1813 gnu_add_sparse_entry(tar, 1814 tar_atol(sparse->offset, sizeof(sparse->offset)), 1815 tar_atol(sparse->numbytes, sizeof(sparse->numbytes))); 1816 sparse++; 1817 length--; 1818 } 1819} 1820 1821/* 1822 * GNU tar sparse format 0.0 1823 * 1824 * Beginning with GNU tar 1.15, sparse files are stored using 1825 * information in the pax extended header. The GNU tar maintainers 1826 * have gone through a number of variations in the process of working 1827 * out this scheme; furtunately, they're all numbered. 1828 * 1829 * Sparse format 0.0 uses attribute GNU.sparse.numblocks to store the 1830 * number of blocks, and GNU.sparse.offset/GNU.sparse.numbytes to 1831 * store offset/size for each block. The repeated instances of these 1832 * latter fields violate the pax specification (which frowns on 1833 * duplicate keys), so this format was quickly replaced. 1834 */ 1835 1836/* 1837 * GNU tar sparse format 0.1 1838 * 1839 * This version replaced the offset/numbytes attributes with 1840 * a single "map" attribute that stored a list of integers. This 1841 * format had two problems: First, the "map" attribute could be very 1842 * long, which caused problems for some implementations. More 1843 * importantly, the sparse data was lost when extracted by archivers 1844 * that didn't recognize this extension. 1845 */ 1846 1847static int 1848gnu_sparse_01_parse(struct tar *tar, const char *p) 1849{ 1850 const char *e; 1851 off_t offset = -1, size = -1; 1852 1853 for (;;) { 1854 e = p; 1855 while (*e != '\0' && *e != ',') { 1856 if (*e < '0' || *e > '9') 1857 return (ARCHIVE_WARN); 1858 e++; 1859 } 1860 if (offset < 0) { 1861 offset = tar_atol10(p, e - p); 1862 if (offset < 0) 1863 return (ARCHIVE_WARN); 1864 } else { 1865 size = tar_atol10(p, e - p); 1866 if (size < 0) 1867 return (ARCHIVE_WARN); 1868 gnu_add_sparse_entry(tar, offset, size); 1869 offset = -1; 1870 } 1871 if (*e == '\0') 1872 return (ARCHIVE_OK); 1873 p = e + 1; 1874 } 1875} 1876 1877/* 1878 * GNU tar sparse format 1.0 1879 * 1880 * The idea: The offset/size data is stored as a series of base-10 1881 * ASCII numbers prepended to the file data, so that dearchivers that 1882 * don't support this format will extract the block map along with the 1883 * data and a separate post-process can restore the sparseness. 1884 * 1885 * Unfortunately, GNU tar 1.16 had a bug that added unnecessary 1886 * padding to the body of the file when using this format. GNU tar 1887 * 1.17 corrected this bug without bumping the version number, so 1888 * it's not possible to support both variants. This code supports 1889 * the later variant at the expense of not supporting the former. 1890 * 1891 * This variant also replaced GNU.sparse.size with GNU.sparse.realsize 1892 * and introduced the GNU.sparse.major/GNU.sparse.minor attributes. 1893 */ 1894 1895/* 1896 * Read the next line from the input, and parse it as a decimal 1897 * integer followed by '\n'. Returns positive integer value or 1898 * negative on error. 1899 */ 1900static int64_t 1901gnu_sparse_10_atol(struct archive_read *a, struct tar *tar, 1902 ssize_t *remaining) 1903{ 1904 int64_t l, limit, last_digit_limit; 1905 const char *p; 1906 ssize_t bytes_read; 1907 int base, digit; 1908 1909 base = 10; 1910 limit = INT64_MAX / base; 1911 last_digit_limit = INT64_MAX % base; 1912 1913 /* 1914 * Skip any lines starting with '#'; GNU tar specs 1915 * don't require this, but they should. 1916 */ 1917 do { 1918 bytes_read = readline(a, tar, &p, tar_min(*remaining, 100)); 1919 if (bytes_read <= 0) 1920 return (ARCHIVE_FATAL); 1921 *remaining -= bytes_read; 1922 } while (p[0] == '#'); 1923 1924 l = 0; 1925 while (bytes_read > 0) { 1926 if (*p == '\n') 1927 return (l); 1928 if (*p < '0' || *p >= '0' + base) 1929 return (ARCHIVE_WARN); 1930 digit = *p - '0'; 1931 if (l > limit || (l == limit && digit > last_digit_limit)) 1932 l = INT64_MAX; /* Truncate on overflow. */ 1933 else 1934 l = (l * base) + digit; 1935 p++; 1936 bytes_read--; 1937 } 1938 /* TODO: Error message. */ 1939 return (ARCHIVE_WARN); 1940} 1941 1942/* 1943 * Returns length (in bytes) of the sparse data description 1944 * that was read. 1945 */ 1946static ssize_t 1947gnu_sparse_10_read(struct archive_read *a, struct tar *tar) 1948{ 1949 ssize_t remaining, bytes_read; 1950 int entries; 1951 off_t offset, size, to_skip; 1952 1953 /* Clear out the existing sparse list. */ 1954 gnu_clear_sparse_list(tar); 1955 1956 remaining = tar->entry_bytes_remaining; 1957 1958 /* Parse entries. */ 1959 entries = gnu_sparse_10_atol(a, tar, &remaining); 1960 if (entries < 0) 1961 return (ARCHIVE_FATAL); 1962 /* Parse the individual entries. */ 1963 while (entries-- > 0) { 1964 /* Parse offset/size */ 1965 offset = gnu_sparse_10_atol(a, tar, &remaining); 1966 if (offset < 0) 1967 return (ARCHIVE_FATAL); 1968 size = gnu_sparse_10_atol(a, tar, &remaining); 1969 if (size < 0) 1970 return (ARCHIVE_FATAL); 1971 /* Add a new sparse entry. */ 1972 gnu_add_sparse_entry(tar, offset, size); 1973 } 1974 /* Skip rest of block... */ 1975 bytes_read = tar->entry_bytes_remaining - remaining; 1976 to_skip = 0x1ff & -bytes_read; 1977 if (to_skip != __archive_read_skip(a, to_skip)) 1978 return (ARCHIVE_FATAL); 1979 return (bytes_read + to_skip); 1980} 1981 1982/*- 1983 * Convert text->integer. 1984 * 1985 * Traditional tar formats (including POSIX) specify base-8 for 1986 * all of the standard numeric fields. This is a significant limitation 1987 * in practice: 1988 * = file size is limited to 8GB 1989 * = rdevmajor and rdevminor are limited to 21 bits 1990 * = uid/gid are limited to 21 bits 1991 * 1992 * There are two workarounds for this: 1993 * = pax extended headers, which use variable-length string fields 1994 * = GNU tar and STAR both allow either base-8 or base-256 in 1995 * most fields. The high bit is set to indicate base-256. 1996 * 1997 * On read, this implementation supports both extensions. 1998 */ 1999static int64_t 2000tar_atol(const char *p, unsigned char_cnt) 2001{ 2002 /* 2003 * Technically, GNU tar considers a field to be in base-256 2004 * only if the first byte is 0xff or 0x80. 2005 */ 2006 if (*p & 0x80) 2007 return (tar_atol256(p, char_cnt)); 2008 return (tar_atol8(p, char_cnt)); 2009} 2010 2011/* 2012 * Note that this implementation does not (and should not!) obey 2013 * locale settings; you cannot simply substitute strtol here, since 2014 * it does obey locale. 2015 */ 2016static int64_t 2017tar_atol8(const char *p, unsigned char_cnt) 2018{ 2019 int64_t l, limit, last_digit_limit; 2020 int digit, sign, base; 2021 2022 base = 8; 2023 limit = INT64_MAX / base; 2024 last_digit_limit = INT64_MAX % base; 2025 2026 while (*p == ' ' || *p == '\t') 2027 p++; 2028 if (*p == '-') { 2029 sign = -1; 2030 p++; 2031 } else 2032 sign = 1; 2033 2034 l = 0; 2035 digit = *p - '0'; 2036 while (digit >= 0 && digit < base && char_cnt-- > 0) { 2037 if (l>limit || (l == limit && digit > last_digit_limit)) { 2038 l = INT64_MAX; /* Truncate on overflow. */ 2039 break; 2040 } 2041 l = (l * base) + digit; 2042 digit = *++p - '0'; 2043 } 2044 return (sign < 0) ? -l : l; 2045} 2046 2047/* 2048 * Note that this implementation does not (and should not!) obey 2049 * locale settings; you cannot simply substitute strtol here, since 2050 * it does obey locale. 2051 */ 2052static int64_t 2053tar_atol10(const char *p, unsigned char_cnt) 2054{ 2055 int64_t l, limit, last_digit_limit; 2056 int base, digit, sign; 2057 2058 base = 10; 2059 limit = INT64_MAX / base; 2060 last_digit_limit = INT64_MAX % base; 2061 2062 while (*p == ' ' || *p == '\t') 2063 p++; 2064 if (*p == '-') { 2065 sign = -1; 2066 p++; 2067 } else 2068 sign = 1; 2069 2070 l = 0; 2071 digit = *p - '0'; 2072 while (digit >= 0 && digit < base && char_cnt-- > 0) { 2073 if (l > limit || (l == limit && digit > last_digit_limit)) { 2074 l = INT64_MAX; /* Truncate on overflow. */ 2075 break; 2076 } 2077 l = (l * base) + digit; 2078 digit = *++p - '0'; 2079 } 2080 return (sign < 0) ? -l : l; 2081} 2082 2083/* 2084 * Parse a base-256 integer. This is just a straight signed binary 2085 * value in big-endian order, except that the high-order bit is 2086 * ignored. 2087 */ 2088static int64_t 2089tar_atol256(const char *_p, unsigned char_cnt) 2090{ 2091 int64_t l, upper_limit, lower_limit; 2092 const unsigned char *p = (const unsigned char *)_p; 2093 2094 upper_limit = INT64_MAX / 256; 2095 lower_limit = INT64_MIN / 256; 2096 2097 /* Pad with 1 or 0 bits, depending on sign. */ 2098 if ((0x40 & *p) == 0x40) 2099 l = (int64_t)-1; 2100 else 2101 l = 0; 2102 l = (l << 6) | (0x3f & *p++); 2103 while (--char_cnt > 0) { 2104 if (l > upper_limit) { 2105 l = INT64_MAX; /* Truncate on overflow */ 2106 break; 2107 } else if (l < lower_limit) { 2108 l = INT64_MIN; 2109 break; 2110 } 2111 l = (l << 8) | (0xff & (int64_t)*p++); 2112 } 2113 return (l); 2114} 2115 2116/* 2117 * Returns length of line (including trailing newline) 2118 * or negative on error. 'start' argument is updated to 2119 * point to first character of line. This avoids copying 2120 * when possible. 2121 */ 2122static ssize_t 2123readline(struct archive_read *a, struct tar *tar, const char **start, 2124 ssize_t limit) 2125{ 2126 ssize_t bytes_read; 2127 ssize_t total_size = 0; 2128 const void *t; 2129 const char *s; 2130 void *p; 2131 2132 t = __archive_read_ahead(a, 1, &bytes_read); 2133 if (bytes_read <= 0) 2134 return (ARCHIVE_FATAL); 2135 s = t; /* Start of line? */ 2136 p = memchr(t, '\n', bytes_read); 2137 /* If we found '\n' in the read buffer, return pointer to that. */ 2138 if (p != NULL) { 2139 bytes_read = 1 + ((const char *)p) - s; 2140 if (bytes_read > limit) { 2141 archive_set_error(&a->archive, 2142 ARCHIVE_ERRNO_FILE_FORMAT, 2143 "Line too long"); 2144 return (ARCHIVE_FATAL); 2145 } 2146 __archive_read_consume(a, bytes_read); 2147 *start = s; 2148 return (bytes_read); 2149 } 2150 /* Otherwise, we need to accumulate in a line buffer. */ 2151 for (;;) { 2152 if (total_size + bytes_read > limit) { 2153 archive_set_error(&a->archive, 2154 ARCHIVE_ERRNO_FILE_FORMAT, 2155 "Line too long"); 2156 return (ARCHIVE_FATAL); 2157 } 2158 if (archive_string_ensure(&tar->line, total_size + bytes_read) == NULL) { 2159 archive_set_error(&a->archive, ENOMEM, 2160 "Can't allocate working buffer"); 2161 return (ARCHIVE_FATAL); 2162 } 2163 memcpy(tar->line.s + total_size, t, bytes_read); 2164 __archive_read_consume(a, bytes_read); 2165 total_size += bytes_read; 2166 /* If we found '\n', clean up and return. */ 2167 if (p != NULL) { 2168 *start = tar->line.s; 2169 return (total_size); 2170 } 2171 /* Read some more. */ 2172 t = __archive_read_ahead(a, 1, &bytes_read); 2173 if (bytes_read <= 0) 2174 return (ARCHIVE_FATAL); 2175 s = t; /* Start of line? */ 2176 p = memchr(t, '\n', bytes_read); 2177 /* If we found '\n', trim the read. */ 2178 if (p != NULL) { 2179 bytes_read = 1 + ((const char *)p) - s; 2180 } 2181 } 2182} 2183 2184static wchar_t * 2185utf8_decode(struct tar *tar, const char *src, size_t length) 2186{ 2187 wchar_t *dest; 2188 ssize_t n; 2189 2190 /* Ensure pax_entry buffer is big enough. */ 2191 if (tar->pax_entry_length <= length) { 2192 wchar_t *old_entry; 2193 2194 if (tar->pax_entry_length <= 0) 2195 tar->pax_entry_length = 1024; 2196 while (tar->pax_entry_length <= length + 1) 2197 tar->pax_entry_length *= 2; 2198 2199 old_entry = tar->pax_entry; 2200 tar->pax_entry = (wchar_t *)realloc(tar->pax_entry, 2201 tar->pax_entry_length * sizeof(wchar_t)); 2202 if (tar->pax_entry == NULL) { 2203 free(old_entry); 2204 /* TODO: Handle this error. */ 2205 return (NULL); 2206 } 2207 } 2208 2209 dest = tar->pax_entry; 2210 while (length > 0) { 2211 n = UTF8_mbrtowc(dest, src, length); 2212 if (n < 0) 2213 return (NULL); 2214 if (n == 0) 2215 break; 2216 dest++; 2217 src += n; 2218 length -= n; 2219 } 2220 *dest = L'\0'; 2221 return (tar->pax_entry); 2222} 2223 2224/* 2225 * Copied and simplified from FreeBSD libc/locale. 2226 */ 2227static ssize_t 2228UTF8_mbrtowc(wchar_t *pwc, const char *s, size_t n) 2229{ 2230 int ch, i, len, mask; 2231 unsigned long wch; 2232 2233 if (s == NULL || n == 0 || pwc == NULL) 2234 return (0); 2235 2236 /* 2237 * Determine the number of octets that make up this character from 2238 * the first octet, and a mask that extracts the interesting bits of 2239 * the first octet. 2240 */ 2241 ch = (unsigned char)*s; 2242 if ((ch & 0x80) == 0) { 2243 mask = 0x7f; 2244 len = 1; 2245 } else if ((ch & 0xe0) == 0xc0) { 2246 mask = 0x1f; 2247 len = 2; 2248 } else if ((ch & 0xf0) == 0xe0) { 2249 mask = 0x0f; 2250 len = 3; 2251 } else if ((ch & 0xf8) == 0xf0) { 2252 mask = 0x07; 2253 len = 4; 2254 } else { 2255 /* Invalid first byte. */ 2256 return (-1); 2257 } 2258 2259 if (n < (size_t)len) { 2260 /* Valid first byte but truncated. */ 2261 return (-2); 2262 } 2263 2264 /* 2265 * Decode the octet sequence representing the character in chunks 2266 * of 6 bits, most significant first. 2267 */ 2268 wch = (unsigned char)*s++ & mask; 2269 i = len; 2270 while (--i != 0) { 2271 if ((*s & 0xc0) != 0x80) { 2272 /* Invalid intermediate byte; consume one byte and 2273 * emit '?' */ 2274 *pwc = '?'; 2275 return (1); 2276 } 2277 wch <<= 6; 2278 wch |= *s++ & 0x3f; 2279 } 2280 2281 /* Assign the value to the output; out-of-range values 2282 * just get truncated. */ 2283 *pwc = (wchar_t)wch; 2284#ifdef WCHAR_MAX 2285 /* 2286 * If platform has WCHAR_MAX, we can do something 2287 * more sensible with out-of-range values. 2288 */ 2289 if (wch >= WCHAR_MAX) 2290 *pwc = '?'; 2291#endif 2292 /* Return number of bytes input consumed: 0 for end-of-string. */ 2293 return (wch == L'\0' ? 0 : len); 2294} 2295 2296 2297/* 2298 * base64_decode - Base64 decode 2299 * 2300 * This accepts most variations of base-64 encoding, including: 2301 * * with or without line breaks 2302 * * with or without the final group padded with '=' or '_' characters 2303 * (The most economical Base-64 variant does not pad the last group and 2304 * omits line breaks; RFC1341 used for MIME requires both.) 2305 */ 2306static char * 2307base64_decode(const char *s, size_t len, size_t *out_len) 2308{ 2309 static const unsigned char digits[64] = { 2310 'A','B','C','D','E','F','G','H','I','J','K','L','M','N', 2311 'O','P','Q','R','S','T','U','V','W','X','Y','Z','a','b', 2312 'c','d','e','f','g','h','i','j','k','l','m','n','o','p', 2313 'q','r','s','t','u','v','w','x','y','z','0','1','2','3', 2314 '4','5','6','7','8','9','+','/' }; 2315 static unsigned char decode_table[128]; 2316 char *out, *d; 2317 const unsigned char *src = (const unsigned char *)s; 2318 2319 /* If the decode table is not yet initialized, prepare it. */ 2320 if (decode_table[digits[1]] != 1) { 2321 unsigned i; 2322 memset(decode_table, 0xff, sizeof(decode_table)); 2323 for (i = 0; i < sizeof(digits); i++) 2324 decode_table[digits[i]] = i; 2325 } 2326 2327 /* Allocate enough space to hold the entire output. */ 2328 /* Note that we may not use all of this... */ 2329 out = (char *)malloc(len - len / 4 + 1); 2330 if (out == NULL) { 2331 *out_len = 0; 2332 return (NULL); 2333 } 2334 d = out; 2335 2336 while (len > 0) { 2337 /* Collect the next group of (up to) four characters. */ 2338 int v = 0; 2339 int group_size = 0; 2340 while (group_size < 4 && len > 0) { 2341 /* '=' or '_' padding indicates final group. */ 2342 if (*src == '=' || *src == '_') { 2343 len = 0; 2344 break; 2345 } 2346 /* Skip illegal characters (including line breaks) */ 2347 if (*src > 127 || *src < 32 2348 || decode_table[*src] == 0xff) { 2349 len--; 2350 src++; 2351 continue; 2352 } 2353 v <<= 6; 2354 v |= decode_table[*src++]; 2355 len --; 2356 group_size++; 2357 } 2358 /* Align a short group properly. */ 2359 v <<= 6 * (4 - group_size); 2360 /* Unpack the group we just collected. */ 2361 switch (group_size) { 2362 case 4: d[2] = v & 0xff; 2363 /* FALLTHROUGH */ 2364 case 3: d[1] = (v >> 8) & 0xff; 2365 /* FALLTHROUGH */ 2366 case 2: d[0] = (v >> 16) & 0xff; 2367 break; 2368 case 1: /* this is invalid! */ 2369 break; 2370 } 2371 d += group_size * 3 / 4; 2372 } 2373 2374 *out_len = d - out; 2375 return (out); 2376} 2377 2378static char * 2379url_decode(const char *in) 2380{ 2381 char *out, *d; 2382 const char *s; 2383 2384 out = (char *)malloc(strlen(in) + 1); 2385 if (out == NULL) 2386 return (NULL); 2387 for (s = in, d = out; *s != '\0'; ) { 2388 if (s[0] == '%' && s[1] != '\0' && s[2] != '\0') { 2389 /* Try to convert % escape */ 2390 int digit1 = tohex(s[1]); 2391 int digit2 = tohex(s[2]); 2392 if (digit1 >= 0 && digit2 >= 0) { 2393 /* Looks good, consume three chars */ 2394 s += 3; 2395 /* Convert output */ 2396 *d++ = ((digit1 << 4) | digit2); 2397 continue; 2398 } 2399 /* Else fall through and treat '%' as normal char */ 2400 } 2401 *d++ = *s++; 2402 } 2403 *d = '\0'; 2404 return (out); 2405} 2406 2407static int 2408tohex(int c) 2409{ 2410 if (c >= '0' && c <= '9') 2411 return (c - '0'); 2412 else if (c >= 'A' && c <= 'F') 2413 return (c - 'A' + 10); 2414 else if (c >= 'a' && c <= 'f') 2415 return (c - 'a' + 10); 2416 else 2417 return (-1); 2418} 2419