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.
27 * This file contains the "essential" portions of the read API, that
28 * is, stuff that will probably always be used by any client that
29 * actually needs to read an archive. Optional pieces have been, as
30 * far as possible, separated out into separate files to avoid
31 * needlessly bloating statically-linked clients.
34 #include "archive_platform.h"
35 __FBSDID("$FreeBSD: src/lib/libarchive/archive_read.c,v 1.30 2007/03/03 07:37:36 kientzle Exp $");
52 #include "archive_entry.h"
53 #include "archive_private.h"
54 #include "archive_read_private.h"
56 static int choose_decompressor(struct archive_read
*, const void*, size_t);
57 static int choose_format(struct archive_read
*);
60 * Allocate, initialize and return a struct archive object.
63 archive_read_new(void)
65 struct archive_read
*a
;
68 a
= (struct archive_read
*)malloc(sizeof(*a
));
71 memset(a
, 0, sizeof(*a
));
72 a
->archive
.magic
= ARCHIVE_READ_MAGIC
;
73 a
->bytes_per_block
= ARCHIVE_DEFAULT_BYTES_PER_BLOCK
;
75 a
->null_length
= 1024;
76 nulls
= (unsigned char *)malloc(a
->null_length
);
78 archive_set_error(&a
->archive
, ENOMEM
,
79 "Can't allocate archive object 'nulls' element");
83 memset(nulls
, 0, a
->null_length
);
86 a
->archive
.state
= ARCHIVE_STATE_NEW
;
87 a
->entry
= archive_entry_new();
89 /* We always support uncompressed archives. */
90 archive_read_support_compression_none(&a
->archive
);
96 * Record the do-not-extract-to file. This belongs in archive_read_extract.c.
99 archive_read_extract_set_skip_file(struct archive
*_a
, dev_t d
, ino_t i
)
101 struct archive_read
*a
= (struct archive_read
*)_a
;
102 __archive_check_magic(_a
, ARCHIVE_READ_MAGIC
, ARCHIVE_STATE_ANY
,
103 "archive_read_extract_set_skip_file");
104 a
->skip_file_dev
= d
;
105 a
->skip_file_ino
= i
;
113 archive_read_open(struct archive
*a
, void *client_data
,
114 archive_open_callback
*client_opener
, archive_read_callback
*client_reader
,
115 archive_close_callback
*client_closer
)
117 /* Old archive_read_open() is just a thin shell around
118 * archive_read_open2. */
119 return archive_read_open2(a
, client_data
, client_opener
,
120 client_reader
, NULL
, client_closer
);
124 archive_read_open2(struct archive
*_a
, void *client_data
,
125 archive_open_callback
*client_opener
,
126 archive_read_callback
*client_reader
,
127 archive_skip_callback
*client_skipper
,
128 archive_close_callback
*client_closer
)
130 struct archive_read
*a
= (struct archive_read
*)_a
;
136 __archive_check_magic(_a
, ARCHIVE_READ_MAGIC
, ARCHIVE_STATE_NEW
, "archive_read_open");
138 if (client_reader
== NULL
)
140 "No reader function provided to archive_read_open");
143 * Set these NULL initially. If the open or initial read fails,
144 * we'll leave them NULL to indicate that the file is invalid.
145 * (In particular, this helps ensure that the closer doesn't
146 * get called more than once.)
148 a
->client_opener
= NULL
;
149 a
->client_reader
= NULL
;
150 a
->client_skipper
= NULL
;
151 a
->client_closer
= NULL
;
152 a
->client_data
= NULL
;
154 /* Open data source. */
155 if (client_opener
!= NULL
) {
156 e
=(client_opener
)(&a
->archive
, client_data
);
158 /* If the open failed, call the closer to clean up. */
160 (client_closer
)(&a
->archive
, client_data
);
165 /* Read first block now for format detection. */
166 bytes_read
= (client_reader
)(&a
->archive
, client_data
, &buffer
);
168 if (bytes_read
< 0) {
169 /* If the first read fails, close before returning error. */
171 (client_closer
)(&a
->archive
, client_data
);
172 /* client_reader should have already set error information. */
173 return (ARCHIVE_FATAL
);
176 /* Now that the client callbacks have worked, remember them. */
177 a
->client_opener
= client_opener
; /* Do we need to remember this? */
178 a
->client_reader
= client_reader
;
179 a
->client_skipper
= client_skipper
;
180 a
->client_closer
= client_closer
;
181 a
->client_data
= client_data
;
183 /* Select a decompression routine. */
184 high_bidder
= choose_decompressor(a
, buffer
, bytes_read
);
186 return (ARCHIVE_FATAL
);
188 /* Initialize decompression routine with the first block of data. */
189 e
= (a
->decompressors
[high_bidder
].init
)(a
, buffer
, bytes_read
);
192 a
->archive
.state
= ARCHIVE_STATE_HEADER
;
198 * Allow each registered decompression routine to bid on whether it
199 * wants to handle this stream. Return index of winning bidder.
202 choose_decompressor(struct archive_read
*a
,
203 const void *buffer
, size_t bytes_read
)
205 int decompression_slots
, i
, bid
, best_bid
, best_bid_slot
;
207 decompression_slots
= sizeof(a
->decompressors
) /
208 sizeof(a
->decompressors
[0]);
213 for (i
= 0; i
< decompression_slots
; i
++) {
214 if (a
->decompressors
[i
].bid
) {
215 bid
= (a
->decompressors
[i
].bid
)(buffer
, bytes_read
);
216 if ((bid
> best_bid
) || (best_bid_slot
< 0)) {
224 * There were no bidders; this is a serious programmer error
225 * and demands a quick and definitive abort.
227 if (best_bid_slot
< 0)
228 __archive_errx(1, "No decompressors were registered; you "
229 "must call at least one "
230 "archive_read_support_compression_XXX function in order "
231 "to successfully read an archive.");
234 * There were bidders, but no non-zero bids; this means we can't
235 * support this stream.
238 archive_set_error(&a
->archive
, ARCHIVE_ERRNO_FILE_FORMAT
,
239 "Unrecognized archive format");
240 return (ARCHIVE_FATAL
);
243 return (best_bid_slot
);
247 * Read header of next entry.
250 archive_read_next_header(struct archive
*_a
, struct archive_entry
**entryp
)
252 struct archive_read
*a
= (struct archive_read
*)_a
;
253 struct archive_entry
*entry
;
256 __archive_check_magic(_a
, ARCHIVE_READ_MAGIC
,
257 ARCHIVE_STATE_HEADER
| ARCHIVE_STATE_DATA
,
258 "archive_read_next_header");
262 archive_entry_clear(entry
);
263 archive_clear_error(&a
->archive
);
266 * If client didn't consume entire data, skip any remainder
267 * (This is especially important for GNU incremental directories.)
269 if (a
->archive
.state
== ARCHIVE_STATE_DATA
) {
270 ret
= archive_read_data_skip(&a
->archive
);
271 if (ret
== ARCHIVE_EOF
) {
272 archive_set_error(&a
->archive
, EIO
, "Premature end-of-file.");
273 a
->archive
.state
= ARCHIVE_STATE_FATAL
;
274 return (ARCHIVE_FATAL
);
276 if (ret
!= ARCHIVE_OK
)
280 /* Record start-of-header. */
281 a
->header_position
= a
->archive
.file_position
;
283 slot
= choose_format(a
);
285 a
->archive
.state
= ARCHIVE_STATE_FATAL
;
286 return (ARCHIVE_FATAL
);
288 a
->format
= &(a
->formats
[slot
]);
289 a
->pformat_data
= &(a
->format
->format_data
);
290 ret
= (a
->format
->read_header
)(a
, entry
);
293 * EOF and FATAL are persistent at this layer. By
294 * modifying the state, we guarantee that future calls to
295 * read a header or read data will fail.
299 a
->archive
.state
= ARCHIVE_STATE_EOF
;
302 a
->archive
.state
= ARCHIVE_STATE_DATA
;
305 a
->archive
.state
= ARCHIVE_STATE_DATA
;
310 a
->archive
.state
= ARCHIVE_STATE_FATAL
;
315 a
->read_data_output_offset
= 0;
316 a
->read_data_remaining
= 0;
321 * Allow each registered format to bid on whether it wants to handle
322 * the next entry. Return index of winning bidder.
325 choose_format(struct archive_read
*a
)
332 slots
= sizeof(a
->formats
) / sizeof(a
->formats
[0]);
336 /* Set up a->format and a->pformat_data for convenience of bidders. */
337 a
->format
= &(a
->formats
[0]);
338 for (i
= 0; i
< slots
; i
++, a
->format
++) {
339 if (a
->format
->bid
) {
340 a
->pformat_data
= &(a
->format
->format_data
);
341 bid
= (a
->format
->bid
)(a
);
342 if (bid
== ARCHIVE_FATAL
)
343 return (ARCHIVE_FATAL
);
344 if ((bid
> best_bid
) || (best_bid_slot
< 0)) {
352 * There were no bidders; this is a serious programmer error
353 * and demands a quick and definitive abort.
355 if (best_bid_slot
< 0)
356 __archive_errx(1, "No formats were registered; you must "
357 "invoke at least one archive_read_support_format_XXX "
358 "function in order to successfully read an archive.");
361 * There were bidders, but no non-zero bids; this means we
362 * can't support this stream.
365 archive_set_error(&a
->archive
, ARCHIVE_ERRNO_FILE_FORMAT
,
366 "Unrecognized archive format");
367 return (ARCHIVE_FATAL
);
370 return (best_bid_slot
);
374 * Return the file offset (within the uncompressed data stream) where
375 * the last header started.
378 archive_read_header_position(struct archive
*_a
)
380 struct archive_read
*a
= (struct archive_read
*)_a
;
381 __archive_check_magic(_a
, ARCHIVE_READ_MAGIC
,
382 ARCHIVE_STATE_ANY
, "archive_read_header_position");
383 return (a
->header_position
);
387 * Read data from an archive entry, using a read(2)-style interface.
388 * This is a convenience routine that just calls
389 * archive_read_data_block and copies the results into the client
390 * buffer, filling any gaps with zero bytes. Clients using this
391 * API can be completely ignorant of sparse-file issues; sparse files
392 * will simply be padded with nulls.
394 * DO NOT intermingle calls to this function and archive_read_data_block
395 * to read a single entry body.
398 archive_read_data(struct archive
*_a
, void *buff
, size_t s
)
400 struct archive_read
*a
= (struct archive_read
*)_a
;
410 if (a
->read_data_remaining
<= 0) {
411 r
= archive_read_data_block(&a
->archive
,
412 (const void **)&a
->read_data_block
,
413 &a
->read_data_remaining
,
414 &a
->read_data_offset
);
415 if (r
== ARCHIVE_EOF
)
418 * Error codes are all negative, so the status
419 * return here cannot be confused with a valid
420 * byte count. (ARCHIVE_OK is zero.)
426 if (a
->read_data_offset
< a
->read_data_output_offset
) {
427 archive_set_error(&a
->archive
, ARCHIVE_ERRNO_FILE_FORMAT
,
428 "Encountered out-of-order sparse blocks");
429 return (ARCHIVE_RETRY
);
432 /* Compute the amount of zero padding needed. */
433 if (a
->read_data_output_offset
+ (off_t
)s
<
434 a
->read_data_offset
) {
436 } else if (a
->read_data_output_offset
<
437 a
->read_data_offset
) {
438 len
= a
->read_data_offset
-
439 a
->read_data_output_offset
;
444 memset(dest
, 0, len
);
446 a
->read_data_output_offset
+= len
;
450 /* Copy data if there is any space left. */
452 len
= a
->read_data_remaining
;
455 memcpy(dest
, a
->read_data_block
, len
);
457 a
->read_data_block
+= len
;
458 a
->read_data_remaining
-= len
;
459 a
->read_data_output_offset
+= len
;
460 a
->read_data_offset
+= len
;
469 * Skip over all remaining data in this entry.
472 archive_read_data_skip(struct archive
*_a
)
474 struct archive_read
*a
= (struct archive_read
*)_a
;
480 __archive_check_magic(_a
, ARCHIVE_READ_MAGIC
, ARCHIVE_STATE_DATA
,
481 "archive_read_data_skip");
483 if (a
->format
->read_data_skip
!= NULL
)
484 r
= (a
->format
->read_data_skip
)(a
);
486 while ((r
= archive_read_data_block(&a
->archive
,
487 &buff
, &size
, &offset
))
492 if (r
== ARCHIVE_EOF
)
495 a
->archive
.state
= ARCHIVE_STATE_HEADER
;
500 * Read the next block of entry data from the archive.
501 * This is a zero-copy interface; the client receives a pointer,
502 * size, and file offset of the next available block of data.
504 * Returns ARCHIVE_OK if the operation is successful, ARCHIVE_EOF if
505 * the end of entry is encountered.
508 archive_read_data_block(struct archive
*_a
,
509 const void **buff
, size_t *size
, off_t
*offset
)
511 struct archive_read
*a
= (struct archive_read
*)_a
;
512 __archive_check_magic(_a
, ARCHIVE_READ_MAGIC
, ARCHIVE_STATE_DATA
,
513 "archive_read_data_block");
515 if (a
->format
->read_data
== NULL
) {
516 archive_set_error(&a
->archive
, ARCHIVE_ERRNO_PROGRAMMER
,
518 "No format_read_data_block function registered");
519 return (ARCHIVE_FATAL
);
522 return (a
->format
->read_data
)(a
, buff
, size
, offset
);
526 * Close the file and release most resources.
528 * Be careful: client might just call read_new and then read_finish.
529 * Don't assume we actually read anything or performed any non-trivial
533 archive_read_close(struct archive
*_a
)
535 struct archive_read
*a
= (struct archive_read
*)_a
;
536 int r
= ARCHIVE_OK
, r1
= ARCHIVE_OK
;
538 __archive_check_magic(&a
->archive
, ARCHIVE_READ_MAGIC
,
539 ARCHIVE_STATE_ANY
, "archive_read_close");
540 a
->archive
.state
= ARCHIVE_STATE_CLOSED
;
542 /* Call cleanup functions registered by optional components. */
543 if (a
->cleanup_archive_extract
!= NULL
)
544 r
= (a
->cleanup_archive_extract
)(a
);
546 /* TODO: Finish the format processing. */
548 /* Close the input machinery. */
549 if (a
->compression_finish
!= NULL
) {
550 r1
= (a
->compression_finish
)(a
);
559 * Release memory and other resources.
561 #if ARCHIVE_API_VERSION > 1
564 /* Temporarily allow library to compile with either 1.x or 2.0 API. */
567 archive_read_finish(struct archive
*_a
)
569 struct archive_read
*a
= (struct archive_read
*)_a
;
574 __archive_check_magic(_a
, ARCHIVE_READ_MAGIC
, ARCHIVE_STATE_ANY
,
575 "archive_read_finish");
576 if (a
->archive
.state
!= ARCHIVE_STATE_CLOSED
)
577 r
= archive_read_close(&a
->archive
);
579 /* Cleanup format-specific data. */
580 slots
= sizeof(a
->formats
) / sizeof(a
->formats
[0]);
581 for (i
= 0; i
< slots
; i
++) {
582 a
->pformat_data
= &(a
->formats
[i
].format_data
);
583 if (a
->formats
[i
].cleanup
)
584 (a
->formats
[i
].cleanup
)(a
);
587 /* Casting a pointer to int allows us to remove 'const.' */
588 free((void *)(uintptr_t)(const void *)a
->nulls
);
589 archive_string_free(&a
->archive
.error_string
);
591 archive_entry_free(a
->entry
);
592 a
->archive
.magic
= 0;
594 #if ARCHIVE_API_VERSION > 1
600 * Used internally by read format handlers to register their bid and
601 * initialization functions.
604 __archive_read_register_format(struct archive_read
*a
,
606 int (*bid
)(struct archive_read
*),
607 int (*read_header
)(struct archive_read
*, struct archive_entry
*),
608 int (*read_data
)(struct archive_read
*, const void **, size_t *, off_t
*),
609 int (*read_data_skip
)(struct archive_read
*),
610 int (*cleanup
)(struct archive_read
*))
614 __archive_check_magic(&a
->archive
,
615 ARCHIVE_READ_MAGIC
, ARCHIVE_STATE_NEW
,
616 "__archive_read_register_format");
618 number_slots
= sizeof(a
->formats
) / sizeof(a
->formats
[0]);
620 for (i
= 0; i
< number_slots
; i
++) {
621 if (a
->formats
[i
].bid
== bid
)
622 return (ARCHIVE_WARN
); /* We've already installed */
623 if (a
->formats
[i
].bid
== NULL
) {
624 a
->formats
[i
].bid
= bid
;
625 a
->formats
[i
].read_header
= read_header
;
626 a
->formats
[i
].read_data
= read_data
;
627 a
->formats
[i
].read_data_skip
= read_data_skip
;
628 a
->formats
[i
].cleanup
= cleanup
;
629 a
->formats
[i
].format_data
= format_data
;
634 __archive_errx(1, "Not enough slots for format registration");
635 return (ARCHIVE_FATAL
); /* Never actually called. */
639 * Used internally by decompression routines to register their bid and
640 * initialization functions.
643 __archive_read_register_compression(struct archive_read
*a
,
644 int (*bid
)(const void *, size_t),
645 int (*init
)(struct archive_read
*, const void *, size_t))
649 __archive_check_magic(&a
->archive
,
650 ARCHIVE_READ_MAGIC
, ARCHIVE_STATE_NEW
,
651 "__archive_read_register_compression");
653 number_slots
= sizeof(a
->decompressors
) / sizeof(a
->decompressors
[0]);
655 for (i
= 0; i
< number_slots
; i
++) {
656 if (a
->decompressors
[i
].bid
== bid
)
657 return (ARCHIVE_OK
); /* We've already installed */
658 if (a
->decompressors
[i
].bid
== NULL
) {
659 a
->decompressors
[i
].bid
= bid
;
660 a
->decompressors
[i
].init
= init
;
665 __archive_errx(1, "Not enough slots for compression registration");
666 return (ARCHIVE_FATAL
); /* Never actually executed. */