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"
45 struct cpio_bin_header {
46 unsigned char c_magic[2];
47 unsigned char c_dev[2];
48 unsigned char c_ino[2];
49 unsigned char c_mode[2];
50 unsigned char c_uid[2];
51 unsigned char c_gid[2];
52 unsigned char c_nlink[2];
53 unsigned char c_rdev[2];
54 unsigned char c_mtime[4];
55 unsigned char c_namesize[2];
56 unsigned char c_filesize[4];
59 struct cpio_odc_header {
73 struct cpio_newc_header {
91 struct links_entry *next;
92 struct links_entry *previous;
99 #define CPIO_MAGIC 0x13141516
102 int (*read_header)(struct archive_read *, struct cpio *,
103 struct archive_entry *, size_t *, size_t *);
104 struct links_entry *links_head;
105 struct archive_string entry_name;
106 struct archive_string entry_linkname;
107 off_t entry_bytes_remaining;
112 static int64_t atol16(const char *, unsigned);
113 static int64_t atol8(const char *, unsigned);
114 static int archive_read_format_cpio_bid(struct archive_read *);
115 static int archive_read_format_cpio_cleanup(struct archive_read *);
116 static int archive_read_format_cpio_read_data(struct archive_read *,
117 const void **, size_t *, off_t *);
118 static int archive_read_format_cpio_read_header(struct archive_read *,
119 struct archive_entry *);
120 static int be4(const unsigned char *);
121 static int find_odc_header(struct archive_read *);
122 static int find_newc_header(struct archive_read *);
123 static int header_bin_be(struct archive_read *, struct cpio *,
124 struct archive_entry *, size_t *, size_t *);
125 static int header_bin_le(struct archive_read *, struct cpio *,
126 struct archive_entry *, size_t *, size_t *);
127 static int header_newc(struct archive_read *, struct cpio *,
128 struct archive_entry *, size_t *, size_t *);
129 static int header_odc(struct archive_read *, struct cpio *,
130 struct archive_entry *, size_t *, size_t *);
131 static int is_octal(const char *, size_t);
132 static int is_hex(const char *, size_t);
133 static int le4(const unsigned char *);
134 static void record_hardlink(struct cpio *cpio, struct archive_entry *entry);
137 archive_read_support_format_cpio(struct archive *_a)
139 struct archive_read *a = (struct archive_read *)_a;
143 cpio = (struct cpio *)malloc(sizeof(*cpio));
145 archive_set_error(&a->archive, ENOMEM, "Can't allocate cpio data");
146 return (ARCHIVE_FATAL);
148 memset(cpio, 0, sizeof(*cpio));
149 cpio->magic = CPIO_MAGIC;
151 r = __archive_read_register_format(a,
154 archive_read_format_cpio_bid,
156 archive_read_format_cpio_read_header,
157 archive_read_format_cpio_read_data,
159 archive_read_format_cpio_cleanup);
168 archive_read_format_cpio_bid(struct archive_read *a)
171 const unsigned char *p;
175 cpio = (struct cpio *)(a->format->data);
177 if ((h = __archive_read_ahead(a, 6, NULL)) == NULL)
180 p = (const unsigned char *)h;
182 if (memcmp(p, "070707", 6) == 0) {
183 /* ASCII cpio archive (odc, POSIX.1) */
184 cpio->read_header = header_odc;
187 * XXX TODO: More verification; Could check that only octal
188 * digits appear in appropriate header locations. XXX
190 } else if (memcmp(p, "070701", 6) == 0) {
191 /* ASCII cpio archive (SVR4 without CRC) */
192 cpio->read_header = header_newc;
195 * XXX TODO: More verification; Could check that only hex
196 * digits appear in appropriate header locations. XXX
198 } else if (memcmp(p, "070702", 6) == 0) {
199 /* ASCII cpio archive (SVR4 with CRC) */
200 /* XXX TODO: Flag that we should check the CRC. XXX */
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 (p[0] * 256 + p[1] == 070707) {
208 /* big-endian binary cpio archives */
209 cpio->read_header = header_bin_be;
211 /* Is more verification possible here? */
212 } else if (p[0] + p[1] * 256 == 070707) {
213 /* little-endian binary cpio archives */
214 cpio->read_header = header_bin_le;
216 /* Is more verification possible here? */
218 return (ARCHIVE_WARN);
224 archive_read_format_cpio_read_header(struct archive_read *a,
225 struct archive_entry *entry)
233 cpio = (struct cpio *)(a->format->data);
234 r = (cpio->read_header(a, cpio, entry, &namelength, &name_pad));
236 if (r < ARCHIVE_WARN)
239 /* Read name from buffer. */
240 h = __archive_read_ahead(a, namelength + name_pad, NULL);
242 return (ARCHIVE_FATAL);
243 __archive_read_consume(a, namelength + name_pad);
244 archive_strncpy(&cpio->entry_name, (const char *)h, namelength);
245 archive_entry_set_pathname(entry, cpio->entry_name.s);
246 cpio->entry_offset = 0;
248 /* If this is a symlink, read the link contents. */
249 if (archive_entry_filetype(entry) == AE_IFLNK) {
250 h = __archive_read_ahead(a, cpio->entry_bytes_remaining, NULL);
252 return (ARCHIVE_FATAL);
253 __archive_read_consume(a, cpio->entry_bytes_remaining);
254 archive_strncpy(&cpio->entry_linkname, (const char *)h,
255 cpio->entry_bytes_remaining);
256 archive_entry_set_symlink(entry, cpio->entry_linkname.s);
257 cpio->entry_bytes_remaining = 0;
260 /* XXX TODO: If the full mode is 0160200, then this is a Solaris
261 * ACL description for the following entry. Read this body
262 * and parse it as a Solaris-style ACL, then read the next
265 /* Compare name to "TRAILER!!!" to test for end-of-archive. */
266 if (namelength == 11 && strcmp((const char *)h, "TRAILER!!!") == 0) {
267 /* TODO: Store file location of start of block. */
268 archive_set_error(&a->archive, 0, NULL);
269 return (ARCHIVE_EOF);
272 /* Detect and record hardlinks to previously-extracted entries. */
273 record_hardlink(cpio, entry);
279 archive_read_format_cpio_read_data(struct archive_read *a,
280 const void **buff, size_t *size, off_t *offset)
285 cpio = (struct cpio *)(a->format->data);
286 if (cpio->entry_bytes_remaining > 0) {
287 *buff = __archive_read_ahead(a, 1, &bytes_read);
289 return (ARCHIVE_FATAL);
290 if (bytes_read > cpio->entry_bytes_remaining)
291 bytes_read = cpio->entry_bytes_remaining;
293 *offset = cpio->entry_offset;
294 cpio->entry_offset += bytes_read;
295 cpio->entry_bytes_remaining -= bytes_read;
296 __archive_read_consume(a, bytes_read);
299 while (cpio->entry_padding > 0) {
300 *buff = __archive_read_ahead(a, 1, &bytes_read);
302 return (ARCHIVE_FATAL);
303 if (bytes_read > cpio->entry_padding)
304 bytes_read = cpio->entry_padding;
305 __archive_read_consume(a, bytes_read);
306 cpio->entry_padding -= bytes_read;
310 *offset = cpio->entry_offset;
311 return (ARCHIVE_EOF);
316 * Skip forward to the next cpio newc header by searching for the
317 * 07070[12] string. This should be generalized and merged with
318 * find_odc_header below.
321 is_hex(const char *p, size_t len)
324 if ((*p >= '0' && *p <= '9')
325 || (*p >= 'a' && *p <= 'f')
326 || (*p >= 'A' && *p <= 'F'))
335 find_newc_header(struct archive_read *a)
339 size_t skip, skipped = 0;
343 h = __archive_read_ahead(a, sizeof(struct cpio_newc_header), &bytes);
345 return (ARCHIVE_FATAL);
349 /* Try the typical case first, then go into the slow search.*/
350 if (memcmp("07070", p, 5) == 0
351 && (p[5] == '1' || p[5] == '2')
352 && is_hex(p, sizeof(struct cpio_newc_header)))
356 * Scan ahead until we find something that looks
357 * like an odc header.
359 while (p + sizeof(struct cpio_newc_header) <= q) {
363 if (memcmp("07070", p, 5) == 0
364 && is_hex(p, sizeof(struct cpio_newc_header))) {
365 skip = p - (const char *)h;
366 __archive_read_consume(a, skip);
369 archive_set_error(&a->archive,
371 "Skipped %d bytes before "
372 "finding valid header",
374 return (ARCHIVE_WARN);
388 skip = p - (const char *)h;
389 __archive_read_consume(a, skip);
395 header_newc(struct archive_read *a, struct cpio *cpio,
396 struct archive_entry *entry, size_t *namelength, size_t *name_pad)
399 const struct cpio_newc_header *header;
402 r = find_newc_header(a);
403 if (r < ARCHIVE_WARN)
406 /* Read fixed-size portion of header. */
407 h = __archive_read_ahead(a, sizeof(struct cpio_newc_header), NULL);
409 return (ARCHIVE_FATAL);
410 __archive_read_consume(a, sizeof(struct cpio_newc_header));
412 /* Parse out hex fields. */
413 header = (const struct cpio_newc_header *)h;
415 if (memcmp(header->c_magic, "070701", 6) == 0) {
416 a->archive.archive_format = ARCHIVE_FORMAT_CPIO_SVR4_NOCRC;
417 a->archive.archive_format_name = "ASCII cpio (SVR4 with no CRC)";
418 } else if (memcmp(header->c_magic, "070702", 6) == 0) {
419 a->archive.archive_format = ARCHIVE_FORMAT_CPIO_SVR4_CRC;
420 a->archive.archive_format_name = "ASCII cpio (SVR4 with CRC)";
422 /* TODO: Abort here? */
425 archive_entry_set_devmajor(entry, atol16(header->c_devmajor, sizeof(header->c_devmajor)));
426 archive_entry_set_devminor(entry, atol16(header->c_devminor, sizeof(header->c_devminor)));
427 archive_entry_set_ino(entry, atol16(header->c_ino, sizeof(header->c_ino)));
428 archive_entry_set_mode(entry, atol16(header->c_mode, sizeof(header->c_mode)));
429 archive_entry_set_uid(entry, atol16(header->c_uid, sizeof(header->c_uid)));
430 archive_entry_set_gid(entry, atol16(header->c_gid, sizeof(header->c_gid)));
431 archive_entry_set_nlink(entry, atol16(header->c_nlink, sizeof(header->c_nlink)));
432 archive_entry_set_rdevmajor(entry, atol16(header->c_rdevmajor, sizeof(header->c_rdevmajor)));
433 archive_entry_set_rdevminor(entry, atol16(header->c_rdevminor, sizeof(header->c_rdevminor)));
434 archive_entry_set_mtime(entry, atol16(header->c_mtime, sizeof(header->c_mtime)), 0);
435 *namelength = atol16(header->c_namesize, sizeof(header->c_namesize));
436 /* Pad name to 2 more than a multiple of 4. */
437 *name_pad = (2 - *namelength) & 3;
440 * Note: entry_bytes_remaining is at least 64 bits and
441 * therefore guaranteed to be big enough for a 33-bit file
444 cpio->entry_bytes_remaining =
445 atol16(header->c_filesize, sizeof(header->c_filesize));
446 archive_entry_set_size(entry, cpio->entry_bytes_remaining);
447 /* Pad file contents to a multiple of 4. */
448 cpio->entry_padding = 3 & -cpio->entry_bytes_remaining;
453 * Skip forward to the next cpio odc header by searching for the
454 * 070707 string. This is a hand-optimized search that could
455 * probably be easily generalized to handle all character-based
459 is_octal(const char *p, size_t len)
462 if (*p < '0' || *p > '7')
470 find_odc_header(struct archive_read *a)
474 size_t skip, skipped = 0;
478 h = __archive_read_ahead(a, sizeof(struct cpio_odc_header), &bytes);
480 return (ARCHIVE_FATAL);
484 /* Try the typical case first, then go into the slow search.*/
485 if (memcmp("070707", p, 6) == 0
486 && is_octal(p, sizeof(struct cpio_odc_header)))
490 * Scan ahead until we find something that looks
491 * like an odc header.
493 while (p + sizeof(struct cpio_odc_header) <= q) {
496 if (memcmp("070707", p, 6) == 0
497 && is_octal(p, sizeof(struct cpio_odc_header))) {
498 skip = p - (const char *)h;
499 __archive_read_consume(a, skip);
502 archive_set_error(&a->archive,
504 "Skipped %d bytes before "
505 "finding valid header",
507 return (ARCHIVE_WARN);
521 skip = p - (const char *)h;
522 __archive_read_consume(a, skip);
528 header_odc(struct archive_read *a, struct cpio *cpio,
529 struct archive_entry *entry, size_t *namelength, size_t *name_pad)
533 const struct cpio_odc_header *header;
535 a->archive.archive_format = ARCHIVE_FORMAT_CPIO_POSIX;
536 a->archive.archive_format_name = "POSIX octet-oriented cpio";
538 /* Find the start of the next header. */
539 r = find_odc_header(a);
540 if (r < ARCHIVE_WARN)
543 /* Read fixed-size portion of header. */
544 h = __archive_read_ahead(a, sizeof(struct cpio_odc_header), NULL);
546 return (ARCHIVE_FATAL);
547 __archive_read_consume(a, sizeof(struct cpio_odc_header));
549 /* Parse out octal fields. */
550 header = (const struct cpio_odc_header *)h;
552 archive_entry_set_dev(entry, atol8(header->c_dev, sizeof(header->c_dev)));
553 archive_entry_set_ino(entry, atol8(header->c_ino, sizeof(header->c_ino)));
554 archive_entry_set_mode(entry, atol8(header->c_mode, sizeof(header->c_mode)));
555 archive_entry_set_uid(entry, atol8(header->c_uid, sizeof(header->c_uid)));
556 archive_entry_set_gid(entry, atol8(header->c_gid, sizeof(header->c_gid)));
557 archive_entry_set_nlink(entry, atol8(header->c_nlink, sizeof(header->c_nlink)));
558 archive_entry_set_rdev(entry, atol8(header->c_rdev, sizeof(header->c_rdev)));
559 archive_entry_set_mtime(entry, atol8(header->c_mtime, sizeof(header->c_mtime)), 0);
560 *namelength = atol8(header->c_namesize, sizeof(header->c_namesize));
561 *name_pad = 0; /* No padding of filename. */
564 * Note: entry_bytes_remaining is at least 64 bits and
565 * therefore guaranteed to be big enough for a 33-bit file
568 cpio->entry_bytes_remaining =
569 atol8(header->c_filesize, sizeof(header->c_filesize));
570 archive_entry_set_size(entry, cpio->entry_bytes_remaining);
571 cpio->entry_padding = 0;
576 header_bin_le(struct archive_read *a, struct cpio *cpio,
577 struct archive_entry *entry, size_t *namelength, size_t *name_pad)
580 const struct cpio_bin_header *header;
582 a->archive.archive_format = ARCHIVE_FORMAT_CPIO_BIN_LE;
583 a->archive.archive_format_name = "cpio (little-endian binary)";
585 /* Read fixed-size portion of header. */
586 h = __archive_read_ahead(a, sizeof(struct cpio_bin_header), NULL);
588 return (ARCHIVE_FATAL);
589 __archive_read_consume(a, sizeof(struct cpio_bin_header));
591 /* Parse out binary fields. */
592 header = (const struct cpio_bin_header *)h;
594 archive_entry_set_dev(entry, header->c_dev[0] + header->c_dev[1] * 256);
595 archive_entry_set_ino(entry, header->c_ino[0] + header->c_ino[1] * 256);
596 archive_entry_set_mode(entry, header->c_mode[0] + header->c_mode[1] * 256);
597 archive_entry_set_uid(entry, header->c_uid[0] + header->c_uid[1] * 256);
598 archive_entry_set_gid(entry, header->c_gid[0] + header->c_gid[1] * 256);
599 archive_entry_set_nlink(entry, header->c_nlink[0] + header->c_nlink[1] * 256);
600 archive_entry_set_rdev(entry, header->c_rdev[0] + header->c_rdev[1] * 256);
601 archive_entry_set_mtime(entry, le4(header->c_mtime), 0);
602 *namelength = header->c_namesize[0] + header->c_namesize[1] * 256;
603 *name_pad = *namelength & 1; /* Pad to even. */
605 cpio->entry_bytes_remaining = le4(header->c_filesize);
606 archive_entry_set_size(entry, cpio->entry_bytes_remaining);
607 cpio->entry_padding = cpio->entry_bytes_remaining & 1; /* Pad to even. */
612 header_bin_be(struct archive_read *a, struct cpio *cpio,
613 struct archive_entry *entry, size_t *namelength, size_t *name_pad)
616 const struct cpio_bin_header *header;
618 a->archive.archive_format = ARCHIVE_FORMAT_CPIO_BIN_BE;
619 a->archive.archive_format_name = "cpio (big-endian binary)";
621 /* Read fixed-size portion of header. */
622 h = __archive_read_ahead(a, sizeof(struct cpio_bin_header), NULL);
624 return (ARCHIVE_FATAL);
625 __archive_read_consume(a, sizeof(struct cpio_bin_header));
627 /* Parse out binary fields. */
628 header = (const struct cpio_bin_header *)h;
629 archive_entry_set_dev(entry, header->c_dev[0] * 256 + header->c_dev[1]);
630 archive_entry_set_ino(entry, header->c_ino[0] * 256 + header->c_ino[1]);
631 archive_entry_set_mode(entry, header->c_mode[0] * 256 + header->c_mode[1]);
632 archive_entry_set_uid(entry, header->c_uid[0] * 256 + header->c_uid[1]);
633 archive_entry_set_gid(entry, header->c_gid[0] * 256 + header->c_gid[1]);
634 archive_entry_set_nlink(entry, header->c_nlink[0] * 256 + header->c_nlink[1]);
635 archive_entry_set_rdev(entry, header->c_rdev[0] * 256 + header->c_rdev[1]);
636 archive_entry_set_mtime(entry, be4(header->c_mtime), 0);
637 *namelength = header->c_namesize[0] * 256 + header->c_namesize[1];
638 *name_pad = *namelength & 1; /* Pad to even. */
640 cpio->entry_bytes_remaining = be4(header->c_filesize);
641 archive_entry_set_size(entry, cpio->entry_bytes_remaining);
642 cpio->entry_padding = cpio->entry_bytes_remaining & 1; /* Pad to even. */
647 archive_read_format_cpio_cleanup(struct archive_read *a)
651 cpio = (struct cpio *)(a->format->data);
652 /* Free inode->name map */
653 while (cpio->links_head != NULL) {
654 struct links_entry *lp = cpio->links_head->next;
656 if (cpio->links_head->name)
657 free(cpio->links_head->name);
658 free(cpio->links_head);
659 cpio->links_head = lp;
661 archive_string_free(&cpio->entry_name);
663 (a->format->data) = NULL;
668 le4(const unsigned char *p)
670 return ((p[0]<<16) + (p[1]<<24) + (p[2]<<0) + (p[3]<<8));
675 be4(const unsigned char *p)
677 return ((p[0]<<24) + (p[1]<<16) + (p[2]<<8) + (p[3]));
681 * Note that this implementation does not (and should not!) obey
682 * locale settings; you cannot simply substitute strtol here, since
683 * it does obey locale.
686 atol8(const char *p, unsigned char_cnt)
692 while (char_cnt-- > 0) {
693 if (*p >= '0' && *p <= '7')
705 atol16(const char *p, unsigned char_cnt)
711 while (char_cnt-- > 0) {
712 if (*p >= 'a' && *p <= 'f')
713 digit = *p - 'a' + 10;
714 else if (*p >= 'A' && *p <= 'F')
715 digit = *p - 'A' + 10;
716 else if (*p >= '0' && *p <= '9')
728 record_hardlink(struct cpio *cpio, struct archive_entry *entry)
730 struct links_entry *le;
734 if (archive_entry_nlink(entry) <= 1)
737 dev = archive_entry_dev(entry);
738 ino = archive_entry_ino64(entry);
741 * First look in the list of multiply-linked files. If we've
742 * already dumped it, convert this entry to a hard link entry.
744 for (le = cpio->links_head; le; le = le->next) {
745 if (le->dev == dev && le->ino == ino) {
746 archive_entry_copy_hardlink(entry, le->name);
748 if (--le->links <= 0) {
749 if (le->previous != NULL)
750 le->previous->next = le->next;
751 if (le->next != NULL)
752 le->next->previous = le->previous;
753 if (cpio->links_head == le)
754 cpio->links_head = le->next;
763 le = (struct links_entry *)malloc(sizeof(struct links_entry));
765 __archive_errx(1, "Out of memory adding file to list");
766 if (cpio->links_head != NULL)
767 cpio->links_head->previous = le;
768 le->next = cpio->links_head;
770 cpio->links_head = le;
773 le->links = archive_entry_nlink(entry) - 1;
774 le->name = strdup(archive_entry_pathname(entry));
775 if (le->name == NULL)
776 __archive_errx(1, "Out of memory adding file to list");