2 * Copyright (c) 2003-2007 Tim Kientzle
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
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.
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.
26 #include "archive_platform.h"
27 __FBSDID("$FreeBSD$");
32 /* #include <stdint.h> */ /* See archive_platform.h */
41 #include "archive_entry.h"
42 #include "archive_private.h"
43 #include "archive_read_private.h"
49 struct cpio_bin_header {
50 unsigned char c_magic[2];
51 unsigned char c_dev[2];
52 unsigned char c_ino[2];
53 unsigned char c_mode[2];
54 unsigned char c_uid[2];
55 unsigned char c_gid[2];
56 unsigned char c_nlink[2];
57 unsigned char c_rdev[2];
58 unsigned char c_mtime[4];
59 unsigned char c_namesize[2];
60 unsigned char c_filesize[4];
63 struct cpio_odc_header {
77 struct cpio_newc_header {
100 struct links_entry *next;
101 struct links_entry *previous;
108 #define CPIO_MAGIC 0x13141516
111 int (*read_header)(struct archive_read *, struct cpio *,
112 struct archive_entry *, size_t *, size_t *);
113 struct links_entry *links_head;
114 struct archive_string entry_name;
115 struct archive_string entry_linkname;
116 off_t entry_bytes_remaining;
121 static int64_t atol16(const char *, unsigned);
122 static int64_t atol8(const char *, unsigned);
123 static int archive_read_format_cpio_bid(struct archive_read *);
124 static int archive_read_format_cpio_cleanup(struct archive_read *);
125 static int archive_read_format_cpio_read_data(struct archive_read *,
126 const void **, size_t *, off_t *);
127 static int archive_read_format_cpio_read_header(struct archive_read *,
128 struct archive_entry *);
129 static int be4(const unsigned char *);
130 static int find_odc_header(struct archive_read *);
131 static int find_newc_header(struct archive_read *);
132 static int header_bin_be(struct archive_read *, struct cpio *,
133 struct archive_entry *, size_t *, size_t *);
134 static int header_bin_le(struct archive_read *, struct cpio *,
135 struct archive_entry *, size_t *, size_t *);
136 static int header_newc(struct archive_read *, struct cpio *,
137 struct archive_entry *, size_t *, size_t *);
138 static int header_odc(struct archive_read *, struct cpio *,
139 struct archive_entry *, size_t *, size_t *);
140 static int is_octal(const char *, size_t);
141 static int is_hex(const char *, size_t);
142 static int le4(const unsigned char *);
143 static void record_hardlink(struct cpio *cpio, struct archive_entry *entry);
146 archive_read_support_format_cpio(struct archive *_a)
148 struct archive_read *a = (struct archive_read *)_a;
152 cpio = (struct cpio *)malloc(sizeof(*cpio));
154 archive_set_error(&a->archive, ENOMEM, "Can't allocate cpio data");
155 return (ARCHIVE_FATAL);
157 memset(cpio, 0, sizeof(*cpio));
158 cpio->magic = CPIO_MAGIC;
160 r = __archive_read_register_format(a,
163 archive_read_format_cpio_bid,
165 archive_read_format_cpio_read_header,
166 archive_read_format_cpio_read_data,
168 archive_read_format_cpio_cleanup);
177 archive_read_format_cpio_bid(struct archive_read *a)
180 const unsigned char *p;
184 cpio = (struct cpio *)(a->format->data);
186 if ((h = __archive_read_ahead(a, 6, NULL)) == NULL)
189 p = (const unsigned char *)h;
191 if (memcmp(p, "070707", 6) == 0) {
192 /* ASCII cpio archive (odc, POSIX.1) */
193 cpio->read_header = header_odc;
196 * XXX TODO: More verification; Could check that only octal
197 * digits appear in appropriate header locations. XXX
199 } else if (memcmp(p, "070701", 6) == 0) {
200 /* ASCII cpio archive (SVR4 without CRC) */
201 cpio->read_header = header_newc;
204 * XXX TODO: More verification; Could check that only hex
205 * digits appear in appropriate header locations. XXX
207 } else if (memcmp(p, "070702", 6) == 0) {
208 /* ASCII cpio archive (SVR4 with CRC) */
209 /* XXX TODO: Flag that we should check the CRC. XXX */
210 cpio->read_header = header_newc;
213 * XXX TODO: More verification; Could check that only hex
214 * digits appear in appropriate header locations. XXX
216 } else if (p[0] * 256 + p[1] == 070707) {
217 /* big-endian binary cpio archives */
218 cpio->read_header = header_bin_be;
220 /* Is more verification possible here? */
221 } else if (p[0] + p[1] * 256 == 070707) {
222 /* little-endian binary cpio archives */
223 cpio->read_header = header_bin_le;
225 /* Is more verification possible here? */
227 return (ARCHIVE_WARN);
233 archive_read_format_cpio_read_header(struct archive_read *a,
234 struct archive_entry *entry)
242 cpio = (struct cpio *)(a->format->data);
243 r = (cpio->read_header(a, cpio, entry, &namelength, &name_pad));
245 if (r < ARCHIVE_WARN)
248 /* Read name from buffer. */
249 h = __archive_read_ahead(a, namelength + name_pad, NULL);
251 return (ARCHIVE_FATAL);
252 __archive_read_consume(a, namelength + name_pad);
253 archive_strncpy(&cpio->entry_name, (const char *)h, namelength);
254 archive_entry_set_pathname(entry, cpio->entry_name.s);
255 cpio->entry_offset = 0;
257 /* If this is a symlink, read the link contents. */
258 if (archive_entry_filetype(entry) == AE_IFLNK) {
259 h = __archive_read_ahead(a, cpio->entry_bytes_remaining, NULL);
261 return (ARCHIVE_FATAL);
262 __archive_read_consume(a, cpio->entry_bytes_remaining);
263 archive_strncpy(&cpio->entry_linkname, (const char *)h,
264 cpio->entry_bytes_remaining);
265 archive_entry_set_symlink(entry, cpio->entry_linkname.s);
266 cpio->entry_bytes_remaining = 0;
269 /* XXX TODO: If the full mode is 0160200, then this is a Solaris
270 * ACL description for the following entry. Read this body
271 * and parse it as a Solaris-style ACL, then read the next
274 /* Compare name to "TRAILER!!!" to test for end-of-archive. */
275 if (namelength == 11 && strcmp((const char *)h, "TRAILER!!!") == 0) {
276 /* TODO: Store file location of start of block. */
277 archive_clear_error(&a->archive);
278 return (ARCHIVE_EOF);
281 /* Detect and record hardlinks to previously-extracted entries. */
282 record_hardlink(cpio, entry);
288 archive_read_format_cpio_read_data(struct archive_read *a,
289 const void **buff, size_t *size, off_t *offset)
294 cpio = (struct cpio *)(a->format->data);
295 if (cpio->entry_bytes_remaining > 0) {
296 *buff = __archive_read_ahead(a, 1, &bytes_read);
298 return (ARCHIVE_FATAL);
299 if (bytes_read > cpio->entry_bytes_remaining)
300 bytes_read = cpio->entry_bytes_remaining;
302 *offset = cpio->entry_offset;
303 cpio->entry_offset += bytes_read;
304 cpio->entry_bytes_remaining -= bytes_read;
305 __archive_read_consume(a, bytes_read);
308 while (cpio->entry_padding > 0) {
309 *buff = __archive_read_ahead(a, 1, &bytes_read);
311 return (ARCHIVE_FATAL);
312 if (bytes_read > cpio->entry_padding)
313 bytes_read = cpio->entry_padding;
314 __archive_read_consume(a, bytes_read);
315 cpio->entry_padding -= bytes_read;
319 *offset = cpio->entry_offset;
320 return (ARCHIVE_EOF);
325 * Skip forward to the next cpio newc header by searching for the
326 * 07070[12] string. This should be generalized and merged with
327 * find_odc_header below.
330 is_hex(const char *p, size_t len)
333 if ((*p >= '0' && *p <= '9')
334 || (*p >= 'a' && *p <= 'f')
335 || (*p >= 'A' && *p <= 'F'))
344 find_newc_header(struct archive_read *a)
348 size_t skip, skipped = 0;
352 h = __archive_read_ahead(a, sizeof(struct cpio_newc_header), &bytes);
354 return (ARCHIVE_FATAL);
358 /* Try the typical case first, then go into the slow search.*/
359 if (memcmp("07070", p, 5) == 0
360 && (p[5] == '1' || p[5] == '2')
361 && is_hex(p, sizeof(struct cpio_newc_header)))
365 * Scan ahead until we find something that looks
366 * like an odc header.
368 while (p + sizeof(struct cpio_newc_header) <= q) {
372 if (memcmp("07070", p, 5) == 0
373 && is_hex(p, sizeof(struct cpio_newc_header))) {
374 skip = p - (const char *)h;
375 __archive_read_consume(a, skip);
378 archive_set_error(&a->archive,
380 "Skipped %d bytes before "
381 "finding valid header",
383 return (ARCHIVE_WARN);
397 skip = p - (const char *)h;
398 __archive_read_consume(a, skip);
404 header_newc(struct archive_read *a, struct cpio *cpio,
405 struct archive_entry *entry, size_t *namelength, size_t *name_pad)
408 const struct cpio_newc_header *header;
411 r = find_newc_header(a);
412 if (r < ARCHIVE_WARN)
415 /* Read fixed-size portion of header. */
416 h = __archive_read_ahead(a, sizeof(struct cpio_newc_header), NULL);
418 return (ARCHIVE_FATAL);
419 __archive_read_consume(a, sizeof(struct cpio_newc_header));
421 /* Parse out hex fields. */
422 header = (const struct cpio_newc_header *)h;
424 if (memcmp(header->c_magic, "070701", 6) == 0) {
425 a->archive.archive_format = ARCHIVE_FORMAT_CPIO_SVR4_NOCRC;
426 a->archive.archive_format_name = "ASCII cpio (SVR4 with no CRC)";
427 } else if (memcmp(header->c_magic, "070702", 6) == 0) {
428 a->archive.archive_format = ARCHIVE_FORMAT_CPIO_SVR4_CRC;
429 a->archive.archive_format_name = "ASCII cpio (SVR4 with CRC)";
431 /* TODO: Abort here? */
434 archive_entry_set_devmajor(entry, atol16(header->c_devmajor, sizeof(header->c_devmajor)));
435 archive_entry_set_devminor(entry, atol16(header->c_devminor, sizeof(header->c_devminor)));
436 archive_entry_set_ino(entry, atol16(header->c_ino, sizeof(header->c_ino)));
437 archive_entry_set_mode(entry, atol16(header->c_mode, sizeof(header->c_mode)));
438 archive_entry_set_uid(entry, atol16(header->c_uid, sizeof(header->c_uid)));
439 archive_entry_set_gid(entry, atol16(header->c_gid, sizeof(header->c_gid)));
440 archive_entry_set_nlink(entry, atol16(header->c_nlink, sizeof(header->c_nlink)));
441 archive_entry_set_rdevmajor(entry, atol16(header->c_rdevmajor, sizeof(header->c_rdevmajor)));
442 archive_entry_set_rdevminor(entry, atol16(header->c_rdevminor, sizeof(header->c_rdevminor)));
443 archive_entry_set_mtime(entry, atol16(header->c_mtime, sizeof(header->c_mtime)), 0);
444 *namelength = atol16(header->c_namesize, sizeof(header->c_namesize));
445 /* Pad name to 2 more than a multiple of 4. */
446 *name_pad = (2 - *namelength) & 3;
449 * Note: entry_bytes_remaining is at least 64 bits and
450 * therefore guaranteed to be big enough for a 33-bit file
453 cpio->entry_bytes_remaining =
454 atol16(header->c_filesize, sizeof(header->c_filesize));
455 archive_entry_set_size(entry, cpio->entry_bytes_remaining);
456 /* Pad file contents to a multiple of 4. */
457 cpio->entry_padding = 3 & -cpio->entry_bytes_remaining;
462 * Skip forward to the next cpio odc header by searching for the
463 * 070707 string. This is a hand-optimized search that could
464 * probably be easily generalized to handle all character-based
468 is_octal(const char *p, size_t len)
471 if (*p < '0' || *p > '7')
479 find_odc_header(struct archive_read *a)
483 size_t skip, skipped = 0;
487 h = __archive_read_ahead(a, sizeof(struct cpio_odc_header), &bytes);
489 return (ARCHIVE_FATAL);
493 /* Try the typical case first, then go into the slow search.*/
494 if (memcmp("070707", p, 6) == 0
495 && is_octal(p, sizeof(struct cpio_odc_header)))
499 * Scan ahead until we find something that looks
500 * like an odc header.
502 while (p + sizeof(struct cpio_odc_header) <= q) {
505 if (memcmp("070707", p, 6) == 0
506 && is_octal(p, sizeof(struct cpio_odc_header))) {
507 skip = p - (const char *)h;
508 __archive_read_consume(a, skip);
511 archive_set_error(&a->archive,
513 "Skipped %d bytes before "
514 "finding valid header",
516 return (ARCHIVE_WARN);
530 skip = p - (const char *)h;
531 __archive_read_consume(a, skip);
537 header_odc(struct archive_read *a, struct cpio *cpio,
538 struct archive_entry *entry, size_t *namelength, size_t *name_pad)
542 const struct cpio_odc_header *header;
544 a->archive.archive_format = ARCHIVE_FORMAT_CPIO_POSIX;
545 a->archive.archive_format_name = "POSIX octet-oriented cpio";
547 /* Find the start of the next header. */
548 r = find_odc_header(a);
549 if (r < ARCHIVE_WARN)
552 /* Read fixed-size portion of header. */
553 h = __archive_read_ahead(a, sizeof(struct cpio_odc_header), NULL);
555 return (ARCHIVE_FATAL);
556 __archive_read_consume(a, sizeof(struct cpio_odc_header));
558 /* Parse out octal fields. */
559 header = (const struct cpio_odc_header *)h;
561 archive_entry_set_dev(entry, atol8(header->c_dev, sizeof(header->c_dev)));
562 archive_entry_set_ino(entry, atol8(header->c_ino, sizeof(header->c_ino)));
563 archive_entry_set_mode(entry, atol8(header->c_mode, sizeof(header->c_mode)));
564 archive_entry_set_uid(entry, atol8(header->c_uid, sizeof(header->c_uid)));
565 archive_entry_set_gid(entry, atol8(header->c_gid, sizeof(header->c_gid)));
566 archive_entry_set_nlink(entry, atol8(header->c_nlink, sizeof(header->c_nlink)));
567 archive_entry_set_rdev(entry, atol8(header->c_rdev, sizeof(header->c_rdev)));
568 archive_entry_set_mtime(entry, atol8(header->c_mtime, sizeof(header->c_mtime)), 0);
569 *namelength = atol8(header->c_namesize, sizeof(header->c_namesize));
570 *name_pad = 0; /* No padding of filename. */
573 * Note: entry_bytes_remaining is at least 64 bits and
574 * therefore guaranteed to be big enough for a 33-bit file
577 cpio->entry_bytes_remaining =
578 atol8(header->c_filesize, sizeof(header->c_filesize));
579 archive_entry_set_size(entry, cpio->entry_bytes_remaining);
580 cpio->entry_padding = 0;
585 header_bin_le(struct archive_read *a, struct cpio *cpio,
586 struct archive_entry *entry, size_t *namelength, size_t *name_pad)
589 const struct cpio_bin_header *header;
591 a->archive.archive_format = ARCHIVE_FORMAT_CPIO_BIN_LE;
592 a->archive.archive_format_name = "cpio (little-endian binary)";
594 /* Read fixed-size portion of header. */
595 h = __archive_read_ahead(a, sizeof(struct cpio_bin_header), NULL);
597 return (ARCHIVE_FATAL);
598 __archive_read_consume(a, sizeof(struct cpio_bin_header));
600 /* Parse out binary fields. */
601 header = (const struct cpio_bin_header *)h;
603 archive_entry_set_dev(entry, header->c_dev[0] + header->c_dev[1] * 256);
604 archive_entry_set_ino(entry, header->c_ino[0] + header->c_ino[1] * 256);
605 archive_entry_set_mode(entry, header->c_mode[0] + header->c_mode[1] * 256);
606 archive_entry_set_uid(entry, header->c_uid[0] + header->c_uid[1] * 256);
607 archive_entry_set_gid(entry, header->c_gid[0] + header->c_gid[1] * 256);
608 archive_entry_set_nlink(entry, header->c_nlink[0] + header->c_nlink[1] * 256);
609 archive_entry_set_rdev(entry, header->c_rdev[0] + header->c_rdev[1] * 256);
610 archive_entry_set_mtime(entry, le4(header->c_mtime), 0);
611 *namelength = header->c_namesize[0] + header->c_namesize[1] * 256;
612 *name_pad = *namelength & 1; /* Pad to even. */
614 cpio->entry_bytes_remaining = le4(header->c_filesize);
615 archive_entry_set_size(entry, cpio->entry_bytes_remaining);
616 cpio->entry_padding = cpio->entry_bytes_remaining & 1; /* Pad to even. */
621 header_bin_be(struct archive_read *a, struct cpio *cpio,
622 struct archive_entry *entry, size_t *namelength, size_t *name_pad)
625 const struct cpio_bin_header *header;
627 a->archive.archive_format = ARCHIVE_FORMAT_CPIO_BIN_BE;
628 a->archive.archive_format_name = "cpio (big-endian binary)";
630 /* Read fixed-size portion of header. */
631 h = __archive_read_ahead(a, sizeof(struct cpio_bin_header), NULL);
633 return (ARCHIVE_FATAL);
634 __archive_read_consume(a, sizeof(struct cpio_bin_header));
636 /* Parse out binary fields. */
637 header = (const struct cpio_bin_header *)h;
638 archive_entry_set_dev(entry, header->c_dev[0] * 256 + header->c_dev[1]);
639 archive_entry_set_ino(entry, header->c_ino[0] * 256 + header->c_ino[1]);
640 archive_entry_set_mode(entry, header->c_mode[0] * 256 + header->c_mode[1]);
641 archive_entry_set_uid(entry, header->c_uid[0] * 256 + header->c_uid[1]);
642 archive_entry_set_gid(entry, header->c_gid[0] * 256 + header->c_gid[1]);
643 archive_entry_set_nlink(entry, header->c_nlink[0] * 256 + header->c_nlink[1]);
644 archive_entry_set_rdev(entry, header->c_rdev[0] * 256 + header->c_rdev[1]);
645 archive_entry_set_mtime(entry, be4(header->c_mtime), 0);
646 *namelength = header->c_namesize[0] * 256 + header->c_namesize[1];
647 *name_pad = *namelength & 1; /* Pad to even. */
649 cpio->entry_bytes_remaining = be4(header->c_filesize);
650 archive_entry_set_size(entry, cpio->entry_bytes_remaining);
651 cpio->entry_padding = cpio->entry_bytes_remaining & 1; /* Pad to even. */
656 archive_read_format_cpio_cleanup(struct archive_read *a)
660 cpio = (struct cpio *)(a->format->data);
661 /* Free inode->name map */
662 while (cpio->links_head != NULL) {
663 struct links_entry *lp = cpio->links_head->next;
665 if (cpio->links_head->name)
666 free(cpio->links_head->name);
667 free(cpio->links_head);
668 cpio->links_head = lp;
670 archive_string_free(&cpio->entry_name);
672 (a->format->data) = NULL;
677 le4(const unsigned char *p)
679 return ((p[0]<<16) + (p[1]<<24) + (p[2]<<0) + (p[3]<<8));
684 be4(const unsigned char *p)
686 return ((p[0]<<24) + (p[1]<<16) + (p[2]<<8) + (p[3]));
690 * Note that this implementation does not (and should not!) obey
691 * locale settings; you cannot simply substitute strtol here, since
692 * it does obey locale.
695 atol8(const char *p, unsigned char_cnt)
701 while (char_cnt-- > 0) {
702 if (*p >= '0' && *p <= '7')
714 atol16(const char *p, unsigned char_cnt)
720 while (char_cnt-- > 0) {
721 if (*p >= 'a' && *p <= 'f')
722 digit = *p - 'a' + 10;
723 else if (*p >= 'A' && *p <= 'F')
724 digit = *p - 'A' + 10;
725 else if (*p >= '0' && *p <= '9')
737 record_hardlink(struct cpio *cpio, struct archive_entry *entry)
739 struct links_entry *le;
743 if (archive_entry_nlink(entry) <= 1)
746 dev = archive_entry_dev(entry);
747 ino = archive_entry_ino64(entry);
750 * First look in the list of multiply-linked files. If we've
751 * already dumped it, convert this entry to a hard link entry.
753 for (le = cpio->links_head; le; le = le->next) {
754 if (le->dev == dev && le->ino == ino) {
755 archive_entry_copy_hardlink(entry, le->name);
757 if (--le->links <= 0) {
758 if (le->previous != NULL)
759 le->previous->next = le->next;
760 if (le->next != NULL)
761 le->next->previous = le->previous;
762 if (cpio->links_head == le)
763 cpio->links_head = le->next;
772 le = (struct links_entry *)malloc(sizeof(struct links_entry));
774 __archive_errx(1, "Out of memory adding file to list");
775 if (cpio->links_head != NULL)
776 cpio->links_head->previous = le;
777 le->next = cpio->links_head;
779 cpio->links_head = le;
782 le->links = archive_entry_nlink(entry) - 1;
783 le->name = strdup(archive_entry_pathname(entry));
784 if (le->name == NULL)
785 __archive_errx(1, "Out of memory adding file to list");