Import libarchive-2.5.5.
[dragonfly.git] / contrib / libarchive-2 / libarchive / archive_read_support_format_iso9660.c
blob481220dd47b5f76f7673cf9adc14947862fdc1d7
1 /*-
2 * Copyright (c) 2003-2007 Tim Kientzle
3 * All rights reserved.
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.
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: src/lib/libarchive/archive_read_support_format_iso9660.c,v 1.26 2008/05/26 17:00:22 kientzle Exp $");
29 #ifdef HAVE_ERRNO_H
30 #include <errno.h>
31 #endif
32 /* #include <stdint.h> */ /* See archive_platform.h */
33 #include <stdio.h>
34 #ifdef HAVE_STDLIB_H
35 #include <stdlib.h>
36 #endif
37 #ifdef HAVE_STRING_H
38 #include <string.h>
39 #endif
40 #include <time.h>
42 #include "archive.h"
43 #include "archive_entry.h"
44 #include "archive_private.h"
45 #include "archive_read_private.h"
46 #include "archive_string.h"
49 * An overview of ISO 9660 format:
51 * Each disk is laid out as follows:
52 * * 32k reserved for private use
53 * * Volume descriptor table. Each volume descriptor
54 * is 2k and specifies basic format information.
55 * The "Primary Volume Descriptor" (PVD) is defined by the
56 * standard and should always be present; other volume
57 * descriptors include various vendor-specific extensions.
58 * * Files and directories. Each file/dir is specified by
59 * an "extent" (starting sector and length in bytes).
60 * Dirs are just files with directory records packed one
61 * after another. The PVD contains a single dir entry
62 * specifying the location of the root directory. Everything
63 * else follows from there.
65 * This module works by first reading the volume descriptors, then
66 * building a list of directory entries, sorted by starting
67 * sector. At each step, I look for the earliest dir entry that
68 * hasn't yet been read, seek forward to that location and read
69 * that entry. If it's a dir, I slurp in the new dir entries and
70 * add them to the heap; if it's a regular file, I return the
71 * corresponding archive_entry and wait for the client to request
72 * the file body. This strategy allows us to read most compliant
73 * CDs with a single pass through the data, as required by libarchive.
76 /* Structure of on-disk primary volume descriptor. */
77 #define PVD_type_offset 0
78 #define PVD_type_size 1
79 #define PVD_id_offset (PVD_type_offset + PVD_type_size)
80 #define PVD_id_size 5
81 #define PVD_version_offset (PVD_id_offset + PVD_id_size)
82 #define PVD_version_size 1
83 #define PVD_reserved1_offset (PVD_version_offset + PVD_version_size)
84 #define PVD_reserved1_size 1
85 #define PVD_system_id_offset (PVD_reserved1_offset + PVD_reserved1_size)
86 #define PVD_system_id_size 32
87 #define PVD_volume_id_offset (PVD_system_id_offset + PVD_system_id_size)
88 #define PVD_volume_id_size 32
89 #define PVD_reserved2_offset (PVD_volume_id_offset + PVD_volume_id_size)
90 #define PVD_reserved2_size 8
91 #define PVD_volume_space_size_offset (PVD_reserved2_offset + PVD_reserved2_size)
92 #define PVD_volume_space_size_size 8
93 #define PVD_reserved3_offset (PVD_volume_space_size_offset + PVD_volume_space_size_size)
94 #define PVD_reserved3_size 32
95 #define PVD_volume_set_size_offset (PVD_reserved3_offset + PVD_reserved3_size)
96 #define PVD_volume_set_size_size 4
97 #define PVD_volume_sequence_number_offset (PVD_volume_set_size_offset + PVD_volume_set_size_size)
98 #define PVD_volume_sequence_number_size 4
99 #define PVD_logical_block_size_offset (PVD_volume_sequence_number_offset + PVD_volume_sequence_number_size)
100 #define PVD_logical_block_size_size 4
101 #define PVD_path_table_size_offset (PVD_logical_block_size_offset + PVD_logical_block_size_size)
102 #define PVD_path_table_size_size 8
103 #define PVD_type_1_path_table_offset (PVD_path_table_size_offset + PVD_path_table_size_size)
104 #define PVD_type_1_path_table_size 4
105 #define PVD_opt_type_1_path_table_offset (PVD_type_1_path_table_offset + PVD_type_1_path_table_size)
106 #define PVD_opt_type_1_path_table_size 4
107 #define PVD_type_m_path_table_offset (PVD_opt_type_1_path_table_offset + PVD_opt_type_1_path_table_size)
108 #define PVD_type_m_path_table_size 4
109 #define PVD_opt_type_m_path_table_offset (PVD_type_m_path_table_offset + PVD_type_m_path_table_size)
110 #define PVD_opt_type_m_path_table_size 4
111 #define PVD_root_directory_record_offset (PVD_opt_type_m_path_table_offset + PVD_opt_type_m_path_table_size)
112 #define PVD_root_directory_record_size 34
113 #define PVD_volume_set_id_offset (PVD_root_directory_record_offset + PVD_root_directory_record_size)
114 #define PVD_volume_set_id_size 128
115 #define PVD_publisher_id_offset (PVD_volume_set_id_offset + PVD_volume_set_id_size)
116 #define PVD_publisher_id_size 128
117 #define PVD_preparer_id_offset (PVD_publisher_id_offset + PVD_publisher_id_size)
118 #define PVD_preparer_id_size 128
119 #define PVD_application_id_offset (PVD_preparer_id_offset + PVD_preparer_id_size)
120 #define PVD_application_id_size 128
121 #define PVD_copyright_file_id_offset (PVD_application_id_offset + PVD_application_id_size)
122 #define PVD_copyright_file_id_size 37
123 #define PVD_abstract_file_id_offset (PVD_copyright_file_id_offset + PVD_copyright_file_id_size)
124 #define PVD_abstract_file_id_size 37
125 #define PVD_bibliographic_file_id_offset (PVD_abstract_file_id_offset + PVD_abstract_file_id_size)
126 #define PVD_bibliographic_file_id_size 37
127 #define PVD_creation_date_offset (PVD_bibliographic_file_id_offset + PVD_bibliographic_file_id_size)
128 #define PVD_creation_date_size 17
129 #define PVD_modification_date_offset (PVD_creation_date_offset + PVD_creation_date_size)
130 #define PVD_modification_date_size 17
131 #define PVD_expiration_date_offset (PVD_modification_date_offset + PVD_modification_date_size)
132 #define PVD_expiration_date_size 17
133 #define PVD_effective_date_offset (PVD_expiration_date_offset + PVD_expiration_date_size)
134 #define PVD_effective_date_size 17
135 #define PVD_file_structure_version_offset (PVD_effective_date_offset + PVD_effective_date_size)
136 #define PVD_file_structure_version_size 1
137 #define PVD_reserved4_offset (PVD_file_structure_version_offset + PVD_file_structure_version_size)
138 #define PVD_reserved4_size 1
139 #define PVD_application_data_offset (PVD_reserved4_offset + PVD_reserved4_size)
140 #define PVD_application_data_size 512
142 /* Structure of an on-disk directory record. */
143 /* Note: ISO9660 stores each multi-byte integer twice, once in
144 * each byte order. The sizes here are the size of just one
145 * of the two integers. (This is why the offset of a field isn't
146 * the same as the offset+size of the previous field.) */
147 #define DR_length_offset 0
148 #define DR_length_size 1
149 #define DR_ext_attr_length_offset 1
150 #define DR_ext_attr_length_size 1
151 #define DR_extent_offset 2
152 #define DR_extent_size 4
153 #define DR_size_offset 10
154 #define DR_size_size 4
155 #define DR_date_offset 18
156 #define DR_date_size 7
157 #define DR_flags_offset 25
158 #define DR_flags_size 1
159 #define DR_file_unit_size_offset 26
160 #define DR_file_unit_size_size 1
161 #define DR_interleave_offset 27
162 #define DR_interleave_size 1
163 #define DR_volume_sequence_number_offset 28
164 #define DR_volume_sequence_number_size 2
165 #define DR_name_len_offset 32
166 #define DR_name_len_size 1
167 #define DR_name_offset 33
170 * Our private data.
173 /* In-memory storage for a directory record. */
174 struct file_info {
175 struct file_info *parent;
176 int refcount;
177 uint64_t offset; /* Offset on disk. */
178 uint64_t size; /* File size in bytes. */
179 uint64_t ce_offset; /* Offset of CE */
180 uint64_t ce_size; /* Size of CE */
181 time_t mtime; /* File last modified time. */
182 time_t atime; /* File last accessed time. */
183 time_t ctime; /* File creation time. */
184 uint64_t rdev; /* Device number */
185 mode_t mode;
186 uid_t uid;
187 gid_t gid;
188 ino_t inode;
189 int nlinks;
190 char *name; /* Null-terminated filename. */
191 struct archive_string symlink;
195 struct iso9660 {
196 int magic;
197 #define ISO9660_MAGIC 0x96609660
198 struct archive_string pathname;
199 char seenRockridge; /* Set true if RR extensions are used. */
200 unsigned char suspOffset;
202 uint64_t previous_offset;
203 uint64_t previous_size;
204 struct archive_string previous_pathname;
206 /* TODO: Make this a heap for fast inserts and deletions. */
207 struct file_info **pending_files;
208 int pending_files_allocated;
209 int pending_files_used;
211 uint64_t current_position;
212 ssize_t logical_block_size;
214 off_t entry_sparse_offset;
215 int64_t entry_bytes_remaining;
218 static void add_entry(struct iso9660 *iso9660, struct file_info *file);
219 static int archive_read_format_iso9660_bid(struct archive_read *);
220 static int archive_read_format_iso9660_cleanup(struct archive_read *);
221 static int archive_read_format_iso9660_read_data(struct archive_read *,
222 const void **, size_t *, off_t *);
223 static int archive_read_format_iso9660_read_data_skip(struct archive_read *);
224 static int archive_read_format_iso9660_read_header(struct archive_read *,
225 struct archive_entry *);
226 static const char *build_pathname(struct archive_string *, struct file_info *);
227 static void dump_isodirrec(FILE *, const unsigned char *isodirrec);
228 static time_t time_from_tm(struct tm *);
229 static time_t isodate17(const unsigned char *);
230 static time_t isodate7(const unsigned char *);
231 static int isPVD(struct iso9660 *, const unsigned char *);
232 static struct file_info *next_entry(struct iso9660 *);
233 static int next_entry_seek(struct archive_read *a, struct iso9660 *iso9660,
234 struct file_info **pfile);
235 static struct file_info *
236 parse_file_info(struct iso9660 *iso9660,
237 struct file_info *parent, const unsigned char *isodirrec);
238 static void parse_rockridge(struct iso9660 *iso9660,
239 struct file_info *file, const unsigned char *start,
240 const unsigned char *end);
241 static void release_file(struct iso9660 *, struct file_info *);
242 static unsigned toi(const void *p, int n);
245 archive_read_support_format_iso9660(struct archive *_a)
247 struct archive_read *a = (struct archive_read *)_a;
248 struct iso9660 *iso9660;
249 int r;
251 iso9660 = (struct iso9660 *)malloc(sizeof(*iso9660));
252 if (iso9660 == NULL) {
253 archive_set_error(&a->archive, ENOMEM, "Can't allocate iso9660 data");
254 return (ARCHIVE_FATAL);
256 memset(iso9660, 0, sizeof(*iso9660));
257 iso9660->magic = ISO9660_MAGIC;
259 r = __archive_read_register_format(a,
260 iso9660,
261 archive_read_format_iso9660_bid,
262 archive_read_format_iso9660_read_header,
263 archive_read_format_iso9660_read_data,
264 archive_read_format_iso9660_read_data_skip,
265 archive_read_format_iso9660_cleanup);
267 if (r != ARCHIVE_OK) {
268 free(iso9660);
269 return (r);
271 return (ARCHIVE_OK);
275 static int
276 archive_read_format_iso9660_bid(struct archive_read *a)
278 struct iso9660 *iso9660;
279 ssize_t bytes_read;
280 const void *h;
281 const unsigned char *p;
282 int bid;
284 iso9660 = (struct iso9660 *)(a->format->data);
287 * Skip the first 32k (reserved area) and get the first
288 * 8 sectors of the volume descriptor table. Of course,
289 * if the I/O layer gives us more, we'll take it.
291 bytes_read = (a->decompressor->read_ahead)(a, &h, 32768 + 8*2048);
292 if (bytes_read < 32768 + 8*2048)
293 return (-1);
294 p = (const unsigned char *)h;
296 /* Skip the reserved area. */
297 bytes_read -= 32768;
298 p += 32768;
300 /* Check each volume descriptor to locate the PVD. */
301 for (; bytes_read > 2048; bytes_read -= 2048, p += 2048) {
302 bid = isPVD(iso9660, p);
303 if (bid > 0)
304 return (bid);
305 if (*p == '\177') /* End-of-volume-descriptor marker. */
306 break;
309 /* We didn't find a valid PVD; return a bid of zero. */
310 return (0);
313 static int
314 isPVD(struct iso9660 *iso9660, const unsigned char *h)
316 struct file_info *file;
318 if (h[0] != 1)
319 return (0);
320 if (memcmp(h+1, "CD001", 5) != 0)
321 return (0);
323 iso9660->logical_block_size = toi(h + PVD_logical_block_size_offset, 2);
325 /* Store the root directory in the pending list. */
326 file = parse_file_info(iso9660, NULL, h + PVD_root_directory_record_offset);
327 add_entry(iso9660, file);
328 return (48);
331 static int
332 archive_read_format_iso9660_read_header(struct archive_read *a,
333 struct archive_entry *entry)
335 struct iso9660 *iso9660;
336 struct file_info *file;
337 ssize_t bytes_read;
338 int r;
340 iso9660 = (struct iso9660 *)(a->format->data);
342 if (!a->archive.archive_format) {
343 a->archive.archive_format = ARCHIVE_FORMAT_ISO9660;
344 a->archive.archive_format_name = "ISO9660";
347 /* Get the next entry that appears after the current offset. */
348 r = next_entry_seek(a, iso9660, &file);
349 if (r != ARCHIVE_OK)
350 return (r);
352 iso9660->entry_bytes_remaining = file->size;
353 iso9660->entry_sparse_offset = 0; /* Offset for sparse-file-aware clients. */
355 /* Set up the entry structure with information about this entry. */
356 archive_entry_set_mode(entry, file->mode);
357 archive_entry_set_uid(entry, file->uid);
358 archive_entry_set_gid(entry, file->gid);
359 archive_entry_set_nlink(entry, file->nlinks);
360 archive_entry_set_ino(entry, file->inode);
361 archive_entry_set_mtime(entry, file->mtime, 0);
362 archive_entry_set_ctime(entry, file->ctime, 0);
363 archive_entry_set_atime(entry, file->atime, 0);
364 /* N.B.: Rock Ridge supports 64-bit device numbers. */
365 archive_entry_set_rdev(entry, (dev_t)file->rdev);
366 archive_entry_set_size(entry, iso9660->entry_bytes_remaining);
367 archive_string_empty(&iso9660->pathname);
368 archive_entry_set_pathname(entry,
369 build_pathname(&iso9660->pathname, file));
370 if (file->symlink.s != NULL)
371 archive_entry_copy_symlink(entry, file->symlink.s);
373 /* If this entry points to the same data as the previous
374 * entry, convert this into a hardlink to that entry.
375 * But don't bother for zero-length files. */
376 if (file->offset == iso9660->previous_offset
377 && file->size == iso9660->previous_size
378 && file->size > 0) {
379 archive_entry_set_hardlink(entry,
380 iso9660->previous_pathname.s);
381 iso9660->entry_bytes_remaining = 0;
382 iso9660->entry_sparse_offset = 0;
383 release_file(iso9660, file);
384 return (ARCHIVE_OK);
387 /* If the offset is before our current position, we can't
388 * seek backwards to extract it, so issue a warning. */
389 if (file->offset < iso9660->current_position) {
390 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC,
391 "Ignoring out-of-order file");
392 iso9660->entry_bytes_remaining = 0;
393 iso9660->entry_sparse_offset = 0;
394 release_file(iso9660, file);
395 return (ARCHIVE_WARN);
398 iso9660->previous_size = file->size;
399 iso9660->previous_offset = file->offset;
400 archive_strcpy(&iso9660->previous_pathname, iso9660->pathname.s);
402 /* If this is a directory, read in all of the entries right now. */
403 if (archive_entry_filetype(entry) == AE_IFDIR) {
404 while (iso9660->entry_bytes_remaining > 0) {
405 const void *block;
406 const unsigned char *p;
407 ssize_t step = iso9660->logical_block_size;
408 if (step > iso9660->entry_bytes_remaining)
409 step = iso9660->entry_bytes_remaining;
410 bytes_read = (a->decompressor->read_ahead)(a, &block, step);
411 if (bytes_read < step) {
412 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC,
413 "Failed to read full block when scanning ISO9660 directory list");
414 release_file(iso9660, file);
415 return (ARCHIVE_FATAL);
417 if (bytes_read > step)
418 bytes_read = step;
419 (a->decompressor->consume)(a, bytes_read);
420 iso9660->current_position += bytes_read;
421 iso9660->entry_bytes_remaining -= bytes_read;
422 for (p = (const unsigned char *)block;
423 *p != 0 && p < (const unsigned char *)block + bytes_read;
424 p += *p) {
425 struct file_info *child;
427 /* Skip '.' entry. */
428 if (*(p + DR_name_len_offset) == 1
429 && *(p + DR_name_offset) == '\0')
430 continue;
431 /* Skip '..' entry. */
432 if (*(p + DR_name_len_offset) == 1
433 && *(p + DR_name_offset) == '\001')
434 continue;
435 child = parse_file_info(iso9660, file, p);
436 add_entry(iso9660, child);
437 if (iso9660->seenRockridge) {
438 a->archive.archive_format =
439 ARCHIVE_FORMAT_ISO9660_ROCKRIDGE;
440 a->archive.archive_format_name =
441 "ISO9660 with Rockridge extensions";
447 release_file(iso9660, file);
448 return (ARCHIVE_OK);
451 static int
452 archive_read_format_iso9660_read_data_skip(struct archive_read *a)
454 /* Because read_next_header always does an explicit skip
455 * to the next entry, we don't need to do anything here. */
456 (void)a; /* UNUSED */
457 return (ARCHIVE_OK);
460 static int
461 archive_read_format_iso9660_read_data(struct archive_read *a,
462 const void **buff, size_t *size, off_t *offset)
464 ssize_t bytes_read;
465 struct iso9660 *iso9660;
467 iso9660 = (struct iso9660 *)(a->format->data);
468 if (iso9660->entry_bytes_remaining <= 0) {
469 *buff = NULL;
470 *size = 0;
471 *offset = iso9660->entry_sparse_offset;
472 return (ARCHIVE_EOF);
475 bytes_read = (a->decompressor->read_ahead)(a, buff, 1);
476 if (bytes_read == 0)
477 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC,
478 "Truncated input file");
479 if (bytes_read <= 0)
480 return (ARCHIVE_FATAL);
481 if (bytes_read > iso9660->entry_bytes_remaining)
482 bytes_read = iso9660->entry_bytes_remaining;
483 *size = bytes_read;
484 *offset = iso9660->entry_sparse_offset;
485 iso9660->entry_sparse_offset += bytes_read;
486 iso9660->entry_bytes_remaining -= bytes_read;
487 iso9660->current_position += bytes_read;
488 (a->decompressor->consume)(a, bytes_read);
489 return (ARCHIVE_OK);
492 static int
493 archive_read_format_iso9660_cleanup(struct archive_read *a)
495 struct iso9660 *iso9660;
496 struct file_info *file;
498 iso9660 = (struct iso9660 *)(a->format->data);
499 while ((file = next_entry(iso9660)) != NULL)
500 release_file(iso9660, file);
501 archive_string_free(&iso9660->pathname);
502 archive_string_free(&iso9660->previous_pathname);
503 if (iso9660->pending_files)
504 free(iso9660->pending_files);
505 free(iso9660);
506 (a->format->data) = NULL;
507 return (ARCHIVE_OK);
511 * This routine parses a single ISO directory record, makes sense
512 * of any extensions, and stores the result in memory.
514 static struct file_info *
515 parse_file_info(struct iso9660 *iso9660, struct file_info *parent,
516 const unsigned char *isodirrec)
518 struct file_info *file;
519 size_t name_len;
520 int flags;
522 /* TODO: Sanity check that name_len doesn't exceed length, etc. */
524 /* Create a new file entry and copy data from the ISO dir record. */
525 file = (struct file_info *)malloc(sizeof(*file));
526 if (file == NULL)
527 return (NULL);
528 memset(file, 0, sizeof(*file));
529 file->parent = parent;
530 if (parent != NULL)
531 parent->refcount++;
532 file->offset = toi(isodirrec + DR_extent_offset, DR_extent_size)
533 * iso9660->logical_block_size;
534 file->size = toi(isodirrec + DR_size_offset, DR_size_size);
535 file->mtime = isodate7(isodirrec + DR_date_offset);
536 file->ctime = file->atime = file->mtime;
537 name_len = (size_t)*(const unsigned char *)(isodirrec + DR_name_len_offset);
538 file->name = (char *)malloc(name_len + 1);
539 if (file->name == NULL) {
540 free(file);
541 return (NULL);
543 memcpy(file->name, isodirrec + DR_name_offset, name_len);
544 file->name[name_len] = '\0';
545 flags = *(isodirrec + DR_flags_offset);
546 if (flags & 0x02)
547 file->mode = AE_IFDIR | 0700;
548 else
549 file->mode = AE_IFREG | 0400;
551 /* Rockridge extensions overwrite information from above. */
553 const unsigned char *rr_start, *rr_end;
554 rr_end = (const unsigned char *)isodirrec
555 + *(isodirrec + DR_length_offset);
556 rr_start = (const unsigned char *)(isodirrec + DR_name_offset
557 + name_len);
558 if ((name_len & 1) == 0)
559 rr_start++;
560 rr_start += iso9660->suspOffset;
561 parse_rockridge(iso9660, file, rr_start, rr_end);
564 /* DEBUGGING: Warn about attributes I don't yet fully support. */
565 if ((flags & ~0x02) != 0) {
566 fprintf(stderr, "\n ** Unrecognized flag: ");
567 dump_isodirrec(stderr, isodirrec);
568 fprintf(stderr, "\n");
569 } else if (toi(isodirrec + DR_volume_sequence_number_offset, 2) != 1) {
570 fprintf(stderr, "\n ** Unrecognized sequence number: ");
571 dump_isodirrec(stderr, isodirrec);
572 fprintf(stderr, "\n");
573 } else if (*(isodirrec + DR_file_unit_size_offset) != 0) {
574 fprintf(stderr, "\n ** Unexpected file unit size: ");
575 dump_isodirrec(stderr, isodirrec);
576 fprintf(stderr, "\n");
577 } else if (*(isodirrec + DR_interleave_offset) != 0) {
578 fprintf(stderr, "\n ** Unexpected interleave: ");
579 dump_isodirrec(stderr, isodirrec);
580 fprintf(stderr, "\n");
581 } else if (*(isodirrec + DR_ext_attr_length_offset) != 0) {
582 fprintf(stderr, "\n ** Unexpected extended attribute length: ");
583 dump_isodirrec(stderr, isodirrec);
584 fprintf(stderr, "\n");
587 return (file);
590 static void
591 add_entry(struct iso9660 *iso9660, struct file_info *file)
593 /* Expand our pending files list as necessary. */
594 if (iso9660->pending_files_used >= iso9660->pending_files_allocated) {
595 struct file_info **new_pending_files;
596 int new_size = iso9660->pending_files_allocated * 2;
598 if (iso9660->pending_files_allocated < 1024)
599 new_size = 1024;
600 /* Overflow might keep us from growing the list. */
601 if (new_size <= iso9660->pending_files_allocated)
602 __archive_errx(1, "Out of memory");
603 new_pending_files = (struct file_info **)malloc(new_size * sizeof(new_pending_files[0]));
604 if (new_pending_files == NULL)
605 __archive_errx(1, "Out of memory");
606 memcpy(new_pending_files, iso9660->pending_files,
607 iso9660->pending_files_allocated * sizeof(new_pending_files[0]));
608 if (iso9660->pending_files != NULL)
609 free(iso9660->pending_files);
610 iso9660->pending_files = new_pending_files;
611 iso9660->pending_files_allocated = new_size;
614 iso9660->pending_files[iso9660->pending_files_used++] = file;
617 static void
618 parse_rockridge(struct iso9660 *iso9660, struct file_info *file,
619 const unsigned char *p, const unsigned char *end)
621 (void)iso9660; /* UNUSED */
623 while (p + 4 < end /* Enough space for another entry. */
624 && p[0] >= 'A' && p[0] <= 'Z' /* Sanity-check 1st char of name. */
625 && p[1] >= 'A' && p[1] <= 'Z' /* Sanity-check 2nd char of name. */
626 && p + p[2] <= end) { /* Sanity-check length. */
627 const unsigned char *data = p + 4;
628 int data_length = p[2] - 4;
629 int version = p[3];
632 * Yes, each 'if' here does test p[0] again.
633 * Otherwise, the fall-through handling to catch
634 * unsupported extensions doesn't work.
636 switch(p[0]) {
637 case 'C':
638 if (p[0] == 'C' && p[1] == 'E' && version == 1) {
640 * CE extension comprises:
641 * 8 byte sector containing extension
642 * 8 byte offset w/in above sector
643 * 8 byte length of continuation
645 file->ce_offset = toi(data, 4)
646 * iso9660->logical_block_size
647 + toi(data + 8, 4);
648 file->ce_size = toi(data + 16, 4);
649 break;
651 /* FALLTHROUGH */
652 case 'N':
653 if (p[0] == 'N' && p[1] == 'M' && version == 1
654 && *data == 0) {
655 /* NM extension with flag byte == 0 */
657 * NM extension comprises:
658 * one byte flag
659 * rest is long name
661 /* TODO: Obey flags. */
662 char *old_name = file->name;
664 data++; /* Skip flag byte. */
665 data_length--;
666 file->name = (char *)malloc(data_length + 1);
667 if (file->name != NULL) {
668 free(old_name);
669 memcpy(file->name, data, data_length);
670 file->name[data_length] = '\0';
671 } else
672 file->name = old_name;
673 break;
675 /* FALLTHROUGH */
676 case 'P':
677 if (p[0] == 'P' && p[1] == 'D' && version == 1) {
679 * PD extension is padding;
680 * contents are always ignored.
682 break;
684 if (p[0] == 'P' && p[1] == 'N' && version == 1) {
685 if (data_length == 16) {
686 file->rdev = toi(data,4);
687 file->rdev <<= 32;
688 file->rdev |= toi(data + 8, 4);
690 break;
692 if (p[0] == 'P' && p[1] == 'X' && version == 1) {
694 * PX extension comprises:
695 * 8 bytes for mode,
696 * 8 bytes for nlinks,
697 * 8 bytes for uid,
698 * 8 bytes for gid,
699 * 8 bytes for inode.
701 if (data_length == 32) {
702 file->mode = toi(data, 4);
703 file->nlinks = toi(data + 8, 4);
704 file->uid = toi(data + 16, 4);
705 file->gid = toi(data + 24, 4);
706 file->inode = toi(data + 32, 4);
708 break;
710 /* FALLTHROUGH */
711 case 'R':
712 if (p[0] == 'R' && p[1] == 'R' && version == 1) {
713 iso9660->seenRockridge = 1;
715 * RR extension comprises:
716 * one byte flag value
718 /* TODO: Handle RR extension. */
719 break;
721 /* FALLTHROUGH */
722 case 'S':
723 if (p[0] == 'S' && p[1] == 'L' && version == 1
724 && *data == 0) {
725 int cont = 1;
726 /* SL extension with flags == 0 */
727 /* TODO: handle non-zero flag values. */
728 data++; /* Skip flag byte. */
729 data_length--;
730 while (data_length > 0) {
731 unsigned char flag = *data++;
732 unsigned char nlen = *data++;
733 data_length -= 2;
735 if (cont == 0)
736 archive_strcat(&file->symlink, "/");
737 cont = 0;
739 switch(flag) {
740 case 0x01: /* Continue */
741 archive_strncat(&file->symlink,
742 (const char *)data, nlen);
743 cont = 1;
744 break;
745 case 0x02: /* Current */
746 archive_strcat(&file->symlink, ".");
747 break;
748 case 0x04: /* Parent */
749 archive_strcat(&file->symlink, "..");
750 break;
751 case 0x08: /* Root */
752 case 0x10: /* Volume root */
753 archive_string_empty(&file->symlink);
754 break;
755 case 0x20: /* Hostname */
756 archive_strcat(&file->symlink, "hostname");
757 break;
758 case 0:
759 archive_strncat(&file->symlink,
760 (const char *)data, nlen);
761 break;
762 default:
763 /* TODO: issue a warning ? */
764 break;
766 data += nlen;
767 data_length -= nlen;
769 break;
771 if (p[0] == 'S' && p[1] == 'P'
772 && version == 1 && data_length == 7
773 && data[0] == (unsigned char)'\xbe'
774 && data[1] == (unsigned char)'\xef') {
776 * SP extension stores the suspOffset
777 * (Number of bytes to skip between
778 * filename and SUSP records.)
779 * It is mandatory by the SUSP standard
780 * (IEEE 1281).
782 * It allows SUSP to coexist with
783 * non-SUSP uses of the System
784 * Use Area by placing non-SUSP data
785 * before SUSP data.
787 * TODO: Add a check for 'SP' in
788 * first directory entry, disable all SUSP
789 * processing if not found.
791 iso9660->suspOffset = data[2];
792 break;
794 if (p[0] == 'S' && p[1] == 'T'
795 && data_length == 0 && version == 1) {
797 * ST extension marks end of this
798 * block of SUSP entries.
800 * It allows SUSP to coexist with
801 * non-SUSP uses of the System
802 * Use Area by placing non-SUSP data
803 * after SUSP data.
805 return;
807 case 'T':
808 if (p[0] == 'T' && p[1] == 'F' && version == 1) {
809 char flag = data[0];
811 * TF extension comprises:
812 * one byte flag
813 * create time (optional)
814 * modify time (optional)
815 * access time (optional)
816 * attribute time (optional)
817 * Time format and presence of fields
818 * is controlled by flag bits.
820 data++;
821 if (flag & 0x80) {
822 /* Use 17-byte time format. */
823 if (flag & 1) /* Create time. */
824 data += 17;
825 if (flag & 2) { /* Modify time. */
826 file->mtime = isodate17(data);
827 data += 17;
829 if (flag & 4) { /* Access time. */
830 file->atime = isodate17(data);
831 data += 17;
833 if (flag & 8) { /* Attribute time. */
834 file->ctime = isodate17(data);
835 data += 17;
837 } else {
838 /* Use 7-byte time format. */
839 if (flag & 1) /* Create time. */
840 data += 7;
841 if (flag & 2) { /* Modify time. */
842 file->mtime = isodate7(data);
843 data += 7;
845 if (flag & 4) { /* Access time. */
846 file->atime = isodate7(data);
847 data += 7;
849 if (flag & 8) { /* Attribute time. */
850 file->ctime = isodate7(data);
851 data += 7;
854 break;
856 /* FALLTHROUGH */
857 default:
858 /* The FALLTHROUGHs above leave us here for
859 * any unsupported extension. */
861 const unsigned char *t;
862 fprintf(stderr, "\nUnsupported RRIP extension for %s\n", file->name);
863 fprintf(stderr, " %c%c(%d):", p[0], p[1], data_length);
864 for (t = data; t < data + data_length && t < data + 16; t++)
865 fprintf(stderr, " %02x", *t);
866 fprintf(stderr, "\n");
872 p += p[2];
876 static void
877 release_file(struct iso9660 *iso9660, struct file_info *file)
879 struct file_info *parent;
881 if (file->refcount == 0) {
882 parent = file->parent;
883 if (file->name)
884 free(file->name);
885 archive_string_free(&file->symlink);
886 free(file);
887 if (parent != NULL) {
888 parent->refcount--;
889 release_file(iso9660, parent);
894 static int
895 next_entry_seek(struct archive_read *a, struct iso9660 *iso9660,
896 struct file_info **pfile)
898 struct file_info *file;
899 uint64_t offset;
901 *pfile = NULL;
902 for (;;) {
903 *pfile = file = next_entry(iso9660);
904 if (file == NULL)
905 return (ARCHIVE_EOF);
907 /* CE area precedes actual file data? Ignore it. */
908 if (file->ce_offset > file->offset) {
909 fprintf(stderr, " *** Discarding CE data.\n");
910 file->ce_offset = 0;
911 file->ce_size = 0;
914 /* Don't waste time seeking for zero-length bodies. */
915 if (file->size == 0) {
916 file->offset = iso9660->current_position;
919 /* If CE exists, find and read it now. */
920 if (file->ce_offset > 0)
921 offset = file->ce_offset;
922 else
923 offset = file->offset;
925 /* Seek forward to the start of the entry. */
926 if (iso9660->current_position < offset) {
927 off_t step = offset - iso9660->current_position;
928 off_t bytes_read;
929 bytes_read = (a->decompressor->skip)(a, step);
930 if (bytes_read < 0)
931 return (bytes_read);
932 iso9660->current_position = offset;
935 /* We found body of file; handle it now. */
936 if (offset == file->offset)
937 return (ARCHIVE_OK);
939 /* Found CE? Process it and push the file back onto list. */
940 if (offset == file->ce_offset) {
941 const void *p;
942 ssize_t size = file->ce_size;
943 ssize_t bytes_read;
944 const unsigned char *rr_start;
946 file->ce_offset = 0;
947 file->ce_size = 0;
948 bytes_read = (a->decompressor->read_ahead)(a, &p, size);
949 if (bytes_read > size)
950 bytes_read = size;
951 rr_start = (const unsigned char *)p;
952 parse_rockridge(iso9660, file, rr_start,
953 rr_start + bytes_read);
954 (a->decompressor->consume)(a, bytes_read);
955 iso9660->current_position += bytes_read;
956 add_entry(iso9660, file);
961 static struct file_info *
962 next_entry(struct iso9660 *iso9660)
964 int least_index;
965 uint64_t least_end_offset;
966 int i;
967 struct file_info *r;
969 if (iso9660->pending_files_used < 1)
970 return (NULL);
972 /* Assume the first file in the list is the earliest on disk. */
973 least_index = 0;
974 least_end_offset = iso9660->pending_files[0]->offset
975 + iso9660->pending_files[0]->size;
977 /* Now, try to find an earlier one. */
978 for (i = 0; i < iso9660->pending_files_used; i++) {
979 /* Use the position of the file *end* as our comparison. */
980 uint64_t end_offset = iso9660->pending_files[i]->offset
981 + iso9660->pending_files[i]->size;
982 if (iso9660->pending_files[i]->ce_offset > 0
983 && iso9660->pending_files[i]->ce_offset < iso9660->pending_files[i]->offset)
984 end_offset = iso9660->pending_files[i]->ce_offset
985 + iso9660->pending_files[i]->ce_size;
986 if (least_end_offset > end_offset) {
987 least_index = i;
988 least_end_offset = end_offset;
991 r = iso9660->pending_files[least_index];
992 iso9660->pending_files[least_index]
993 = iso9660->pending_files[--iso9660->pending_files_used];
994 return (r);
997 static unsigned int
998 toi(const void *p, int n)
1000 const unsigned char *v = (const unsigned char *)p;
1001 if (n > 1)
1002 return v[0] + 256 * toi(v + 1, n - 1);
1003 if (n == 1)
1004 return v[0];
1005 return (0);
1008 static time_t
1009 isodate7(const unsigned char *v)
1011 struct tm tm;
1012 int offset;
1013 memset(&tm, 0, sizeof(tm));
1014 tm.tm_year = v[0];
1015 tm.tm_mon = v[1] - 1;
1016 tm.tm_mday = v[2];
1017 tm.tm_hour = v[3];
1018 tm.tm_min = v[4];
1019 tm.tm_sec = v[5];
1020 /* v[6] is the signed timezone offset, in 1/4-hour increments. */
1021 offset = ((const signed char *)v)[6];
1022 if (offset > -48 && offset < 52) {
1023 tm.tm_hour -= offset / 4;
1024 tm.tm_min -= (offset % 4) * 15;
1026 return (time_from_tm(&tm));
1029 static time_t
1030 isodate17(const unsigned char *v)
1032 struct tm tm;
1033 int offset;
1034 memset(&tm, 0, sizeof(tm));
1035 tm.tm_year = (v[0] - '0') * 1000 + (v[1] - '0') * 100
1036 + (v[2] - '0') * 10 + (v[3] - '0')
1037 - 1900;
1038 tm.tm_mon = (v[4] - '0') * 10 + (v[5] - '0');
1039 tm.tm_mday = (v[6] - '0') * 10 + (v[7] - '0');
1040 tm.tm_hour = (v[8] - '0') * 10 + (v[9] - '0');
1041 tm.tm_min = (v[10] - '0') * 10 + (v[11] - '0');
1042 tm.tm_sec = (v[12] - '0') * 10 + (v[13] - '0');
1043 /* v[16] is the signed timezone offset, in 1/4-hour increments. */
1044 offset = ((const signed char *)v)[16];
1045 if (offset > -48 && offset < 52) {
1046 tm.tm_hour -= offset / 4;
1047 tm.tm_min -= (offset % 4) * 15;
1049 return (time_from_tm(&tm));
1052 static time_t
1053 time_from_tm(struct tm *t)
1055 #if HAVE_TIMEGM
1056 /* Use platform timegm() if available. */
1057 return (timegm(t));
1058 #else
1059 /* Else use direct calculation using POSIX assumptions. */
1060 /* First, fix up tm_yday based on the year/month/day. */
1061 mktime(t);
1062 /* Then we can compute timegm() from first principles. */
1063 return (t->tm_sec + t->tm_min * 60 + t->tm_hour * 3600
1064 + t->tm_yday * 86400 + (t->tm_year - 70) * 31536000
1065 + ((t->tm_year - 69) / 4) * 86400 -
1066 ((t->tm_year - 1) / 100) * 86400
1067 + ((t->tm_year + 299) / 400) * 86400);
1068 #endif
1071 static const char *
1072 build_pathname(struct archive_string *as, struct file_info *file)
1074 if (file->parent != NULL && file->parent->name[0] != '\0') {
1075 build_pathname(as, file->parent);
1076 archive_strcat(as, "/");
1078 if (file->name[0] == '\0')
1079 archive_strcat(as, ".");
1080 else
1081 archive_strcat(as, file->name);
1082 return (as->s);
1085 static void
1086 dump_isodirrec(FILE *out, const unsigned char *isodirrec)
1088 fprintf(out, " l %d,",
1089 toi(isodirrec + DR_length_offset, DR_length_size));
1090 fprintf(out, " a %d,",
1091 toi(isodirrec + DR_ext_attr_length_offset, DR_ext_attr_length_size));
1092 fprintf(out, " ext 0x%x,",
1093 toi(isodirrec + DR_extent_offset, DR_extent_size));
1094 fprintf(out, " s %d,",
1095 toi(isodirrec + DR_size_offset, DR_extent_size));
1096 fprintf(out, " f 0x%02x,",
1097 toi(isodirrec + DR_flags_offset, DR_flags_size));
1098 fprintf(out, " u %d,",
1099 toi(isodirrec + DR_file_unit_size_offset, DR_file_unit_size_size));
1100 fprintf(out, " ilv %d,",
1101 toi(isodirrec + DR_interleave_offset, DR_interleave_size));
1102 fprintf(out, " seq %d,",
1103 toi(isodirrec + DR_volume_sequence_number_offset, DR_volume_sequence_number_size));
1104 fprintf(out, " nl %d:",
1105 toi(isodirrec + DR_name_len_offset, DR_name_len_size));
1106 fprintf(out, " `%.*s'",
1107 toi(isodirrec + DR_name_len_offset, DR_name_len_size), isodirrec + DR_name_offset);