2 * Copyright (c) 2007 Kai Wang
3 * Copyright (c) 2007 Tim Kientzle
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following conditions
9 * 1. Redistributions of source code must retain the above copyright
10 * notice, this list of conditions and the following disclaimer
11 * in this position and unchanged.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution.
16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
17 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
18 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
19 * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
20 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
21 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
22 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
23 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
24 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
25 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
28 #include "archive_platform.h"
29 __FBSDID("$FreeBSD: src/lib/libarchive/archive_read_support_format_ar.c,v 1.5 2007/04/15 00:53:38 kientzle Exp $");
31 #ifdef HAVE_SYS_STAT_H
45 #include "archive_entry.h"
46 #include "archive_private.h"
47 #include "archive_read_private.h"
51 off_t entry_bytes_remaining
;
59 * Define structure of the "ar" header.
61 #define AR_name_offset 0
62 #define AR_name_size 16
63 #define AR_date_offset 16
64 #define AR_date_size 12
65 #define AR_uid_offset 28
67 #define AR_gid_offset 34
69 #define AR_mode_offset 40
70 #define AR_mode_size 8
71 #define AR_size_offset 48
72 #define AR_size_size 10
73 #define AR_fmag_offset 58
74 #define AR_fmag_size 2
76 #define isdigit(x) (x) >= '0' && (x) <= '9'
78 static int archive_read_format_ar_bid(struct archive_read
*a
);
79 static int archive_read_format_ar_cleanup(struct archive_read
*a
);
80 static int archive_read_format_ar_read_data(struct archive_read
*a
,
81 const void **buff
, size_t *size
, off_t
*offset
);
82 static int archive_read_format_ar_skip(struct archive_read
*a
);
83 static int archive_read_format_ar_read_header(struct archive_read
*a
,
84 struct archive_entry
*e
);
85 static uint64_t ar_atol8(const char *p
, unsigned char_cnt
);
86 static uint64_t ar_atol10(const char *p
, unsigned char_cnt
);
87 static int ar_parse_gnu_filename_table(struct archive_read
*, struct ar
*,
88 const void *, size_t);
89 static int ar_parse_common_header(struct ar
*ar
, struct archive_entry
*,
93 archive_read_support_format_ar(struct archive
*_a
)
95 struct archive_read
*a
= (struct archive_read
*)_a
;
99 ar
= (struct ar
*)malloc(sizeof(*ar
));
101 archive_set_error(&a
->archive
, ENOMEM
,
102 "Can't allocate ar data");
103 return (ARCHIVE_FATAL
);
105 memset(ar
, 0, sizeof(*ar
));
109 r
= __archive_read_register_format(a
,
111 archive_read_format_ar_bid
,
112 archive_read_format_ar_read_header
,
113 archive_read_format_ar_read_data
,
114 archive_read_format_ar_skip
,
115 archive_read_format_ar_cleanup
);
117 if (r
!= ARCHIVE_OK
) {
125 archive_read_format_ar_cleanup(struct archive_read
*a
)
129 ar
= (struct ar
*)(a
->format
->data
);
133 (a
->format
->data
) = NULL
;
138 archive_read_format_ar_bid(struct archive_read
*a
)
144 if (a
->archive
.archive_format
!= 0 &&
145 (a
->archive
.archive_format
& ARCHIVE_FORMAT_BASE_MASK
) !=
149 ar
= (struct ar
*)(a
->format
->data
);
155 * Verify the 8-byte file signature.
156 * TODO: Do we need to check more than this?
158 bytes_read
= (a
->decompressor
->read_ahead
)(a
, &h
, 8);
161 if (strncmp((const char*)h
, "!<arch>\n", 8) == 0) {
169 archive_read_format_ar_read_header(struct archive_read
*a
,
170 struct archive_entry
*entry
)
172 char filename
[AR_name_size
+ 1];
174 uint64_t number
; /* Used to hold parsed numbers before validation. */
176 size_t bsd_name_length
, entry_size
;
182 ar
= (struct ar
*)(a
->format
->data
);
184 if (a
->archive
.file_position
== 0) {
186 * We are now at the beginning of the archive,
187 * so we need first consume the ar global header.
189 (a
->decompressor
->consume
)(a
, 8);
190 /* Set a default format code for now. */
191 a
->archive
.archive_format
= ARCHIVE_FORMAT_AR
;
194 /* Read the header for the next file entry. */
195 bytes_read
= (a
->decompressor
->read_ahead
)(a
, &b
, 60);
196 if (bytes_read
< 60) {
198 return (ARCHIVE_EOF
);
200 (a
->decompressor
->consume
)(a
, 60);
203 /* Verify the magic signature on the file header. */
204 if (strncmp(h
+ AR_fmag_offset
, "`\n", 2) != 0) {
205 archive_set_error(&a
->archive
, EINVAL
,
206 "Consistency check failed");
207 return (ARCHIVE_WARN
);
210 /* Copy filename into work buffer. */
211 strncpy(filename
, h
+ AR_name_offset
, AR_name_size
);
212 filename
[AR_name_size
] = '\0';
215 * Guess the format variant based on the filename.
217 if (a
->archive
.archive_format
== ARCHIVE_FORMAT_AR
) {
218 /* We don't already know the variant, so let's guess. */
220 * Biggest clue is presence of '/': GNU starts special
221 * filenames with '/', appends '/' as terminator to
222 * non-special names, so anything with '/' should be
223 * GNU except for BSD long filenames.
225 if (strncmp(filename
, "#1/", 3) == 0)
226 a
->archive
.archive_format
= ARCHIVE_FORMAT_AR_BSD
;
227 else if (strchr(filename
, '/') != NULL
)
228 a
->archive
.archive_format
= ARCHIVE_FORMAT_AR_GNU
;
229 else if (strncmp(filename
, "__.SYMDEF", 9) == 0)
230 a
->archive
.archive_format
= ARCHIVE_FORMAT_AR_BSD
;
232 * XXX Do GNU/SVR4 'ar' programs ever omit trailing '/'
233 * if name exactly fills 16-byte field? If so, we
234 * can't assume entries without '/' are BSD. XXX
238 /* Update format name from the code. */
239 if (a
->archive
.archive_format
== ARCHIVE_FORMAT_AR_GNU
)
240 a
->archive
.archive_format_name
= "ar (GNU/SVR4)";
241 else if (a
->archive
.archive_format
== ARCHIVE_FORMAT_AR_BSD
)
242 a
->archive
.archive_format_name
= "ar (BSD)";
244 a
->archive
.archive_format_name
= "ar";
247 * Remove trailing spaces from the filename. GNU and BSD
248 * variants both pad filename area out with spaces.
249 * This will only be wrong if GNU/SVR4 'ar' implementations
250 * omit trailing '/' for 16-char filenames and we have
251 * a 16-char filename that ends in ' '.
253 p
= filename
+ AR_name_size
- 1;
254 while (p
>= filename
&& *p
== ' ') {
260 * Remove trailing slash unless first character is '/'.
261 * (BSD entries never end in '/', so this will only trim
262 * GNU-format entries. GNU special entries start with '/'
263 * and are not terminated in '/', so we don't trim anything
264 * that starts with '/'.)
266 if (filename
[0] != '/' && *p
== '/')
270 * '//' is the GNU filename table.
271 * Later entries can refer to names in this table.
273 if (strcmp(filename
, "//") == 0) {
274 /* This must come before any call to _read_ahead. */
275 ar_parse_common_header(ar
, entry
, h
);
276 archive_entry_copy_pathname(entry
, filename
);
277 archive_entry_set_mode(entry
,
278 S_IFREG
| (archive_entry_mode(entry
) & 0777));
279 /* Get the size of the filename table. */
280 number
= ar_atol10(h
+ AR_size_offset
, AR_size_size
);
281 if (number
> SIZE_MAX
) {
282 archive_set_error(&a
->archive
, ARCHIVE_ERRNO_MISC
,
283 "Filename table too large");
284 return (ARCHIVE_FATAL
);
286 entry_size
= (size_t)number
;
287 /* Read the filename table into memory. */
288 bytes_read
= (a
->decompressor
->read_ahead
)(a
, &b
, entry_size
);
290 return (ARCHIVE_FATAL
);
291 if ((size_t)bytes_read
< entry_size
) {
292 archive_set_error(&a
->archive
, ARCHIVE_ERRNO_MISC
,
293 "Truncated input file");
294 return (ARCHIVE_FATAL
);
297 * Don't consume the contents, so the client will
298 * also get a shot at reading it.
301 /* Parse the filename table. */
302 return (ar_parse_gnu_filename_table(a
, ar
, b
, entry_size
));
306 * GNU variant handles long filenames by storing /<number>
307 * to indicate a name stored in the filename table.
309 if (filename
[0] == '/' && isdigit(filename
[1])) {
310 number
= ar_atol10(h
+ AR_name_offset
+ 1, AR_name_size
- 1);
312 * If we can't look up the real name, warn and return
313 * the entry with the wrong name.
315 if (ar
->strtab
== NULL
|| number
> ar
->strtab_size
) {
316 archive_set_error(&a
->archive
, EINVAL
,
317 "Can't find long filename for entry");
318 archive_entry_copy_pathname(entry
, filename
);
319 /* Parse the time, owner, mode, size fields. */
320 ar_parse_common_header(ar
, entry
, h
);
321 return (ARCHIVE_WARN
);
324 archive_entry_copy_pathname(entry
, &ar
->strtab
[(size_t)number
]);
325 /* Parse the time, owner, mode, size fields. */
326 return (ar_parse_common_header(ar
, entry
, h
));
330 * BSD handles long filenames by storing "#1/" followed by the
331 * length of filename as a decimal number, then prepends the
332 * the filename to the file contents.
334 if (strncmp(filename
, "#1/", 3) == 0) {
335 /* Parse the time, owner, mode, size fields. */
336 /* This must occur before _read_ahead is called again. */
337 ar_parse_common_header(ar
, entry
, h
);
339 /* Parse the size of the name, adjust the file size. */
340 number
= ar_atol10(h
+ AR_name_offset
+ 3, AR_name_size
- 3);
341 if ((off_t
)number
> ar
->entry_bytes_remaining
) {
342 archive_set_error(&a
->archive
, ARCHIVE_ERRNO_MISC
,
343 "Bad input file size");
344 return (ARCHIVE_FATAL
);
346 bsd_name_length
= (size_t)number
;
347 ar
->entry_bytes_remaining
-= bsd_name_length
;
348 /* Adjust file size reported to client. */
349 archive_entry_set_size(entry
, ar
->entry_bytes_remaining
);
351 /* Read the long name into memory. */
352 bytes_read
= (a
->decompressor
->read_ahead
)(a
, &b
, bsd_name_length
);
354 return (ARCHIVE_FATAL
);
355 if ((size_t)bytes_read
< bsd_name_length
) {
356 archive_set_error(&a
->archive
, ARCHIVE_ERRNO_MISC
,
357 "Truncated input file");
358 return (ARCHIVE_FATAL
);
360 (a
->decompressor
->consume
)(a
, bsd_name_length
);
362 /* Store it in the entry. */
363 p
= (char *)malloc(bsd_name_length
+ 1);
365 archive_set_error(&a
->archive
, ENOMEM
,
366 "Can't allocate fname buffer");
367 return (ARCHIVE_FATAL
);
369 strncpy(p
, b
, bsd_name_length
);
370 p
[bsd_name_length
] = '\0';
371 archive_entry_copy_pathname(entry
, p
);
377 * "/" is the SVR4/GNU archive symbol table.
379 if (strcmp(filename
, "/") == 0) {
380 archive_entry_copy_pathname(entry
, "/");
381 /* Parse the time, owner, mode, size fields. */
382 r
= ar_parse_common_header(ar
, entry
, h
);
383 /* Force the file type to a regular file. */
384 archive_entry_set_mode(entry
,
385 S_IFREG
| (archive_entry_mode(entry
) & 0777));
390 * "__.SYMDEF" is a BSD archive symbol table.
392 if (strcmp(filename
, "__.SYMDEF") == 0) {
393 archive_entry_copy_pathname(entry
, filename
);
394 /* Parse the time, owner, mode, size fields. */
395 return (ar_parse_common_header(ar
, entry
, h
));
399 * Otherwise, this is a standard entry. The filename
400 * has already been trimmed as much as possible, based
401 * on our current knowledge of the format.
403 archive_entry_copy_pathname(entry
, filename
);
404 return (ar_parse_common_header(ar
, entry
, h
));
408 ar_parse_common_header(struct ar
*ar
, struct archive_entry
*entry
,
413 /* Copy remaining header */
414 archive_entry_set_mtime(entry
,
415 (time_t)ar_atol10(h
+ AR_date_offset
, AR_date_size
), 0L);
416 archive_entry_set_uid(entry
,
417 (uid_t
)ar_atol10(h
+ AR_uid_offset
, AR_uid_size
));
418 archive_entry_set_gid(entry
,
419 (gid_t
)ar_atol10(h
+ AR_gid_offset
, AR_gid_size
));
420 archive_entry_set_mode(entry
,
421 (mode_t
)ar_atol8(h
+ AR_mode_offset
, AR_mode_size
));
422 n
= ar_atol10(h
+ AR_size_offset
, AR_size_size
);
424 ar
->entry_offset
= 0;
425 ar
->entry_padding
= n
% 2;
426 archive_entry_set_size(entry
, n
);
427 ar
->entry_bytes_remaining
= n
;
432 archive_read_format_ar_read_data(struct archive_read
*a
,
433 const void **buff
, size_t *size
, off_t
*offset
)
438 ar
= (struct ar
*)(a
->format
->data
);
440 if (ar
->entry_bytes_remaining
> 0) {
441 bytes_read
= (a
->decompressor
->read_ahead
)(a
, buff
, 1);
442 if (bytes_read
== 0) {
443 archive_set_error(&a
->archive
, ARCHIVE_ERRNO_MISC
,
444 "Truncated ar archive");
445 return (ARCHIVE_FATAL
);
448 return (ARCHIVE_FATAL
);
449 if (bytes_read
> ar
->entry_bytes_remaining
)
450 bytes_read
= (ssize_t
)ar
->entry_bytes_remaining
;
452 *offset
= ar
->entry_offset
;
453 ar
->entry_offset
+= bytes_read
;
454 ar
->entry_bytes_remaining
-= bytes_read
;
455 (a
->decompressor
->consume
)(a
, (size_t)bytes_read
);
458 while (ar
->entry_padding
> 0) {
459 bytes_read
= (a
->decompressor
->read_ahead
)(a
, buff
, 1);
461 return (ARCHIVE_FATAL
);
462 if (bytes_read
> ar
->entry_padding
)
463 bytes_read
= (ssize_t
)ar
->entry_padding
;
464 (a
->decompressor
->consume
)(a
, (size_t)bytes_read
);
465 ar
->entry_padding
-= bytes_read
;
469 *offset
= ar
->entry_offset
;
470 return (ARCHIVE_EOF
);
475 archive_read_format_ar_skip(struct archive_read
*a
)
480 const void *b
; /* Dummy variables */
484 ar
= (struct ar
*)(a
->format
->data
);
485 if (a
->decompressor
->skip
== NULL
) {
486 while (r
== ARCHIVE_OK
)
487 r
= archive_read_format_ar_read_data(a
, &b
, &s
, &o
);
491 bytes_skipped
= (a
->decompressor
->skip
)(a
, ar
->entry_bytes_remaining
+
493 if (bytes_skipped
< 0)
494 return (ARCHIVE_FATAL
);
496 ar
->entry_bytes_remaining
= 0;
497 ar
->entry_padding
= 0;
503 ar_parse_gnu_filename_table(struct archive_read
*a
, struct ar
*ar
,
504 const void *h
, size_t size
)
508 if (ar
->strtab
!= NULL
) {
509 archive_set_error(&a
->archive
, EINVAL
,
510 "More than one string tables exist");
511 return (ARCHIVE_WARN
);
515 archive_set_error(&a
->archive
, EINVAL
, "Invalid string table");
516 return (ARCHIVE_WARN
);
519 ar
->strtab_size
= size
;
520 ar
->strtab
= malloc(size
);
521 if (ar
->strtab
== NULL
) {
522 archive_set_error(&a
->archive
, ENOMEM
,
523 "Can't allocate string table buffer");
524 return (ARCHIVE_FATAL
);
527 (void)memcpy(ar
->strtab
, h
, size
);
528 for (p
= ar
->strtab
; p
< ar
->strtab
+ size
- 1; ++p
) {
532 goto bad_string_table
;
537 * Sanity check, last two chars must be `/\n' or '\n\n',
538 * depending on whether the string table is padded by a '\n'
539 * (string table produced by GNU ar always has a even size).
541 if (p
!= ar
->strtab
+ size
&& *p
!= '\n')
542 goto bad_string_table
;
544 /* Enforce zero termination. */
545 ar
->strtab
[size
- 1] = '\0';
550 archive_set_error(&a
->archive
, EINVAL
,
551 "Invalid string table");
554 return (ARCHIVE_WARN
);
558 ar_atol8(const char *p
, unsigned char_cnt
)
560 uint64_t l
, limit
, last_digit_limit
;
561 unsigned int digit
, base
;
564 limit
= UINT64_MAX
/ base
;
565 last_digit_limit
= UINT64_MAX
% base
;
567 while ((*p
== ' ' || *p
== '\t') && char_cnt
-- > 0)
572 while (*p
>= '0' && digit
< base
&& char_cnt
-- > 0) {
573 if (l
>limit
|| (l
== limit
&& digit
> last_digit_limit
)) {
574 l
= UINT64_MAX
; /* Truncate on overflow. */
577 l
= (l
* base
) + digit
;
584 ar_atol10(const char *p
, unsigned char_cnt
)
586 uint64_t l
, limit
, last_digit_limit
;
587 unsigned int base
, digit
;
590 limit
= UINT64_MAX
/ base
;
591 last_digit_limit
= UINT64_MAX
% base
;
593 while ((*p
== ' ' || *p
== '\t') && char_cnt
-- > 0)
597 while (*p
>= '0' && digit
< base
&& char_cnt
-- > 0) {
598 if (l
> limit
|| (l
== limit
&& digit
> last_digit_limit
)) {
599 l
= UINT64_MAX
; /* Truncate on overflow. */
602 l
= (l
* base
) + digit
;