Import libarchive and bsdtar version 2.0.25
[dragonfly/port-amd64.git] / contrib / libarchive-2.0 / libarchive / archive_read_support_format_iso9660.c
blobd471b9246ccee5d27f7ae4c9d31e05313046da30
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 1.20 2007/03/03 07:37:36 kientzle Exp $");
29 #ifdef HAVE_SYS_STAT_H
30 #include <sys/stat.h>
31 #endif
33 #ifdef HAVE_ERRNO_H
34 #include <errno.h>
35 #endif
36 /* #include <stdint.h> */ /* See archive_platform.h */
37 #include <stdio.h>
38 #ifdef HAVE_STDLIB_H
39 #include <stdlib.h>
40 #endif
41 #ifdef HAVE_STRING_H
42 #include <string.h>
43 #endif
44 #include <time.h>
45 #ifdef HAVE_UNISTD_H
46 #include <unistd.h>
47 #endif
49 #include "archive.h"
50 #include "archive_entry.h"
51 #include "archive_private.h"
52 #include "archive_read_private.h"
53 #include "archive_string.h"
56 * An overview of ISO 9660 format:
58 * Each disk is laid out as follows:
59 * * 32k reserved for private use
60 * * Volume descriptor table. Each volume descriptor
61 * is 2k and specifies basic format information.
62 * The "Primary Volume Descriptor" (PVD) is defined by the
63 * standard and should always be present; other volume
64 * descriptors include various vendor-specific extensions.
65 * * Files and directories. Each file/dir is specified by
66 * an "extent" (starting sector and length in bytes).
67 * Dirs are just files with directory records packed one
68 * after another. The PVD contains a single dir entry
69 * specifying the location of the root directory. Everything
70 * else follows from there.
72 * This module works by first reading the volume descriptors, then
73 * building a list of directory entries, sorted by starting
74 * sector. At each step, I look for the earliest dir entry that
75 * hasn't yet been read, seek forward to that location and read
76 * that entry. If it's a dir, I slurp in the new dir entries and
77 * add them to the heap; if it's a regular file, I return the
78 * corresponding archive_entry and wait for the client to request
79 * the file body. This strategy allows us to read most compliant
80 * CDs with a single pass through the data, as required by libarchive.
83 /* Structure of on-disk primary volume descriptor. */
84 #define PVD_type_offset 0
85 #define PVD_type_size 1
86 #define PVD_id_offset (PVD_type_offset + PVD_type_size)
87 #define PVD_id_size 5
88 #define PVD_version_offset (PVD_id_offset + PVD_id_size)
89 #define PVD_version_size 1
90 #define PVD_reserved1_offset (PVD_version_offset + PVD_version_size)
91 #define PVD_reserved1_size 1
92 #define PVD_system_id_offset (PVD_reserved1_offset + PVD_reserved1_size)
93 #define PVD_system_id_size 32
94 #define PVD_volume_id_offset (PVD_system_id_offset + PVD_system_id_size)
95 #define PVD_volume_id_size 32
96 #define PVD_reserved2_offset (PVD_volume_id_offset + PVD_volume_id_size)
97 #define PVD_reserved2_size 8
98 #define PVD_volume_space_size_offset (PVD_reserved2_offset + PVD_reserved2_size)
99 #define PVD_volume_space_size_size 8
100 #define PVD_reserved3_offset (PVD_volume_space_size_offset + PVD_volume_space_size_size)
101 #define PVD_reserved3_size 32
102 #define PVD_volume_set_size_offset (PVD_reserved3_offset + PVD_reserved3_size)
103 #define PVD_volume_set_size_size 4
104 #define PVD_volume_sequence_number_offset (PVD_volume_set_size_offset + PVD_volume_set_size_size)
105 #define PVD_volume_sequence_number_size 4
106 #define PVD_logical_block_size_offset (PVD_volume_sequence_number_offset + PVD_volume_sequence_number_size)
107 #define PVD_logical_block_size_size 4
108 #define PVD_path_table_size_offset (PVD_logical_block_size_offset + PVD_logical_block_size_size)
109 #define PVD_path_table_size_size 8
110 #define PVD_type_1_path_table_offset (PVD_path_table_size_offset + PVD_path_table_size_size)
111 #define PVD_type_1_path_table_size 4
112 #define PVD_opt_type_1_path_table_offset (PVD_type_1_path_table_offset + PVD_type_1_path_table_size)
113 #define PVD_opt_type_1_path_table_size 4
114 #define PVD_type_m_path_table_offset (PVD_opt_type_1_path_table_offset + PVD_opt_type_1_path_table_size)
115 #define PVD_type_m_path_table_size 4
116 #define PVD_opt_type_m_path_table_offset (PVD_type_m_path_table_offset + PVD_type_m_path_table_size)
117 #define PVD_opt_type_m_path_table_size 4
118 #define PVD_root_directory_record_offset (PVD_opt_type_m_path_table_offset + PVD_opt_type_m_path_table_size)
119 #define PVD_root_directory_record_size 34
120 #define PVD_volume_set_id_offset (PVD_root_directory_record_offset + PVD_root_directory_record_size)
121 #define PVD_volume_set_id_size 128
122 #define PVD_publisher_id_offset (PVD_volume_set_id_offset + PVD_volume_set_id_size)
123 #define PVD_publisher_id_size 128
124 #define PVD_preparer_id_offset (PVD_publisher_id_offset + PVD_publisher_id_size)
125 #define PVD_preparer_id_size 128
126 #define PVD_application_id_offset (PVD_preparer_id_offset + PVD_preparer_id_size)
127 #define PVD_application_id_size 128
128 #define PVD_copyright_file_id_offset (PVD_application_id_offset + PVD_application_id_size)
129 #define PVD_copyright_file_id_size 37
130 #define PVD_abstract_file_id_offset (PVD_copyright_file_id_offset + PVD_copyright_file_id_size)
131 #define PVD_abstract_file_id_size 37
132 #define PVD_bibliographic_file_id_offset (PVD_abstract_file_id_offset + PVD_abstract_file_id_size)
133 #define PVD_bibliographic_file_id_size 37
134 #define PVD_creation_date_offset (PVD_bibliographic_file_id_offset + PVD_bibliographic_file_id_size)
135 #define PVD_creation_date_size 17
136 #define PVD_modification_date_offset (PVD_creation_date_offset + PVD_creation_date_size)
137 #define PVD_modification_date_size 17
138 #define PVD_expiration_date_offset (PVD_modification_date_offset + PVD_modification_date_size)
139 #define PVD_expiration_date_size 17
140 #define PVD_effective_date_offset (PVD_expiration_date_offset + PVD_expiration_date_size)
141 #define PVD_effective_date_size 17
142 #define PVD_file_structure_version_offset (PVD_effective_date_offset + PVD_effective_date_size)
143 #define PVD_file_structure_version_size 1
144 #define PVD_reserved4_offset (PVD_file_structure_version_offset + PVD_file_structure_version_size)
145 #define PVD_reserved4_size 1
146 #define PVD_application_data_offset (PVD_reserved4_offset + PVD_reserved4_size)
147 #define PVD_application_data_size 512
149 /* Structure of an on-disk directory record. */
150 /* Note: ISO9660 stores each multi-byte integer twice, once in
151 * each byte order. The sizes here are the size of just one
152 * of the two integers. (This is why the offset of a field isn't
153 * the same as the offset+size of the previous field.) */
154 #define DR_length_offset 0
155 #define DR_length_size 1
156 #define DR_ext_attr_length_offset 1
157 #define DR_ext_attr_length_size 1
158 #define DR_extent_offset 2
159 #define DR_extent_size 4
160 #define DR_size_offset 10
161 #define DR_size_size 4
162 #define DR_date_offset 18
163 #define DR_date_size 7
164 #define DR_flags_offset 25
165 #define DR_flags_size 1
166 #define DR_file_unit_size_offset 26
167 #define DR_file_unit_size_size 1
168 #define DR_interleave_offset 27
169 #define DR_interleave_size 1
170 #define DR_volume_sequence_number_offset 28
171 #define DR_volume_sequence_number_size 2
172 #define DR_name_len_offset 32
173 #define DR_name_len_size 1
174 #define DR_name_offset 33
177 * Our private data.
180 /* In-memory storage for a directory record. */
181 struct file_info {
182 struct file_info *parent;
183 int refcount;
184 uint64_t offset; /* Offset on disk. */
185 uint64_t size; /* File size in bytes. */
186 uint64_t ce_offset; /* Offset of CE */
187 uint64_t ce_size; /* Size of CE */
188 time_t mtime; /* File last modified time. */
189 time_t atime; /* File last accessed time. */
190 time_t ctime; /* File creation time. */
191 mode_t mode;
192 uid_t uid;
193 gid_t gid;
194 ino_t inode;
195 int nlinks;
196 char *name; /* Null-terminated filename. */
197 struct archive_string symlink;
201 struct iso9660 {
202 int magic;
203 #define ISO9660_MAGIC 0x96609660
204 int bid; /* If non-zero, return this as our bid. */
205 struct archive_string pathname;
206 char seenRockridge; /* Set true if RR extensions are used. */
207 unsigned char suspOffset;
209 uint64_t previous_offset;
210 uint64_t previous_size;
211 struct archive_string previous_pathname;
213 /* TODO: Make this a heap for fast inserts and deletions. */
214 struct file_info **pending_files;
215 int pending_files_allocated;
216 int pending_files_used;
218 uint64_t current_position;
219 ssize_t logical_block_size;
221 off_t entry_sparse_offset;
222 int64_t entry_bytes_remaining;
225 static void add_entry(struct iso9660 *iso9660, struct file_info *file);
226 static int archive_read_format_iso9660_bid(struct archive_read *);
227 static int archive_read_format_iso9660_cleanup(struct archive_read *);
228 static int archive_read_format_iso9660_read_data(struct archive_read *,
229 const void **, size_t *, off_t *);
230 static int archive_read_format_iso9660_read_data_skip(struct archive_read *);
231 static int archive_read_format_iso9660_read_header(struct archive_read *,
232 struct archive_entry *);
233 static const char *build_pathname(struct archive_string *, struct file_info *);
234 static void dump_isodirrec(FILE *, const unsigned char *isodirrec);
235 static time_t time_from_tm(struct tm *);
236 static time_t isodate17(const unsigned char *);
237 static time_t isodate7(const unsigned char *);
238 static int isPVD(struct iso9660 *, const unsigned char *);
239 static struct file_info *next_entry(struct iso9660 *);
240 static int next_entry_seek(struct archive_read *a, struct iso9660 *iso9660,
241 struct file_info **pfile);
242 static struct file_info *
243 parse_file_info(struct iso9660 *iso9660,
244 struct file_info *parent, const unsigned char *isodirrec);
245 static void parse_rockridge(struct iso9660 *iso9660,
246 struct file_info *file, const unsigned char *start,
247 const unsigned char *end);
248 static void release_file(struct iso9660 *, struct file_info *);
249 static unsigned toi(const void *p, int n);
252 archive_read_support_format_iso9660(struct archive *_a)
254 struct archive_read *a = (struct archive_read *)_a;
255 struct iso9660 *iso9660;
256 int r;
258 iso9660 = (struct iso9660 *)malloc(sizeof(*iso9660));
259 if (iso9660 == NULL) {
260 archive_set_error(&a->archive, ENOMEM, "Can't allocate iso9660 data");
261 return (ARCHIVE_FATAL);
263 memset(iso9660, 0, sizeof(*iso9660));
264 iso9660->magic = ISO9660_MAGIC;
265 iso9660->bid = -1; /* We haven't yet bid. */
267 r = __archive_read_register_format(a,
268 iso9660,
269 archive_read_format_iso9660_bid,
270 archive_read_format_iso9660_read_header,
271 archive_read_format_iso9660_read_data,
272 archive_read_format_iso9660_read_data_skip,
273 archive_read_format_iso9660_cleanup);
275 if (r != ARCHIVE_OK) {
276 free(iso9660);
277 return (r);
279 return (ARCHIVE_OK);
283 static int
284 archive_read_format_iso9660_bid(struct archive_read *a)
286 struct iso9660 *iso9660;
287 ssize_t bytes_read;
288 const void *h;
289 const char *p;
291 iso9660 = (struct iso9660 *)*(a->pformat_data);
293 if (iso9660->bid >= 0)
294 return (iso9660->bid);
297 * Skip the first 32k (reserved area) and get the first
298 * 8 sectors of the volume descriptor table. Of course,
299 * if the I/O layer gives us more, we'll take it.
301 bytes_read = (a->compression_read_ahead)(a, &h, 32768 + 8*2048);
302 if (bytes_read < 32768 + 8*2048)
303 return (iso9660->bid = -1);
304 p = (const char *)h;
306 /* Skip the reserved area. */
307 bytes_read -= 32768;
308 p += 32768;
310 /* Check each volume descriptor to locate the PVD. */
311 for (; bytes_read > 2048; bytes_read -= 2048, p += 2048) {
312 iso9660->bid = isPVD(iso9660, p);
313 if (iso9660->bid > 0)
314 return (iso9660->bid);
315 if (*p == '\xff') /* End-of-volume-descriptor marker. */
316 break;
319 /* We didn't find a valid PVD; return a bid of zero. */
320 iso9660->bid = 0;
321 return (iso9660->bid);
324 static int
325 isPVD(struct iso9660 *iso9660, const unsigned char *h)
327 struct file_info *file;
329 if (h[0] != 1)
330 return (0);
331 if (memcmp(h+1, "CD001", 5) != 0)
332 return (0);
334 iso9660->logical_block_size = toi(h + PVD_logical_block_size_offset, 2);
336 /* Store the root directory in the pending list. */
337 file = parse_file_info(iso9660, NULL, h + PVD_root_directory_record_offset);
338 add_entry(iso9660, file);
339 return (48);
342 static int
343 archive_read_format_iso9660_read_header(struct archive_read *a,
344 struct archive_entry *entry)
346 struct stat st;
347 struct iso9660 *iso9660;
348 struct file_info *file;
349 ssize_t bytes_read;
350 int r;
352 iso9660 = (struct iso9660 *)*(a->pformat_data);
354 if (!a->archive.archive_format) {
355 a->archive.archive_format = ARCHIVE_FORMAT_ISO9660;
356 a->archive.archive_format_name = "ISO9660";
359 /* Get the next entry that appears after the current offset. */
360 r = next_entry_seek(a, iso9660, &file);
361 if (r != ARCHIVE_OK)
362 return (r);
364 iso9660->entry_bytes_remaining = file->size;
365 iso9660->entry_sparse_offset = 0; /* Offset for sparse-file-aware clients. */
367 /* Set up the entry structure with information about this entry. */
368 memset(&st, 0, sizeof(st));
369 st.st_mode = file->mode;
370 st.st_uid = file->uid;
371 st.st_gid = file->gid;
372 st.st_nlink = file->nlinks;
373 st.st_ino = file->inode;
374 st.st_mtime = file->mtime;
375 st.st_ctime = file->ctime;
376 st.st_atime = file->atime;
377 st.st_size = iso9660->entry_bytes_remaining;
378 archive_entry_copy_stat(entry, &st);
379 archive_string_empty(&iso9660->pathname);
380 archive_entry_set_pathname(entry,
381 build_pathname(&iso9660->pathname, file));
382 if (file->symlink.s != NULL)
383 archive_entry_copy_symlink(entry, file->symlink.s);
385 /* If this entry points to the same data as the previous
386 * entry, convert this into a hardlink to that entry.
387 * But don't bother for zero-length files. */
388 if (file->offset == iso9660->previous_offset
389 && file->size == iso9660->previous_size
390 && file->size > 0) {
391 archive_entry_set_hardlink(entry,
392 iso9660->previous_pathname.s);
393 iso9660->entry_bytes_remaining = 0;
394 iso9660->entry_sparse_offset = 0;
395 release_file(iso9660, file);
396 return (ARCHIVE_OK);
399 /* If the offset is before our current position, we can't
400 * seek backwards to extract it, so issue a warning. */
401 if (file->offset < iso9660->current_position) {
402 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC,
403 "Ignoring out-of-order file");
404 iso9660->entry_bytes_remaining = 0;
405 iso9660->entry_sparse_offset = 0;
406 release_file(iso9660, file);
407 return (ARCHIVE_WARN);
410 iso9660->previous_size = file->size;
411 iso9660->previous_offset = file->offset;
412 archive_strcpy(&iso9660->previous_pathname, iso9660->pathname.s);
414 /* If this is a directory, read in all of the entries right now. */
415 if (S_ISDIR(st.st_mode)) {
416 while (iso9660->entry_bytes_remaining > 0) {
417 const void *block;
418 const unsigned char *p;
419 ssize_t step = iso9660->logical_block_size;
420 if (step > iso9660->entry_bytes_remaining)
421 step = iso9660->entry_bytes_remaining;
422 bytes_read = (a->compression_read_ahead)(a, &block, step);
423 if (bytes_read < step) {
424 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC,
425 "Failed to read full block when scanning ISO9660 directory list");
426 release_file(iso9660, file);
427 return (ARCHIVE_FATAL);
429 if (bytes_read > step)
430 bytes_read = step;
431 (a->compression_read_consume)(a, bytes_read);
432 iso9660->current_position += bytes_read;
433 iso9660->entry_bytes_remaining -= bytes_read;
434 for (p = (const unsigned char *)block;
435 *p != 0 && p < (const unsigned char *)block + bytes_read;
436 p += *p) {
437 struct file_info *child;
439 /* Skip '.' entry. */
440 if (*(p + DR_name_len_offset) == 1
441 && *(p + DR_name_offset) == '\0')
442 continue;
443 /* Skip '..' entry. */
444 if (*(p + DR_name_len_offset) == 1
445 && *(p + DR_name_offset) == '\001')
446 continue;
447 child = parse_file_info(iso9660, file, p);
448 add_entry(iso9660, child);
449 if (iso9660->seenRockridge) {
450 a->archive.archive_format =
451 ARCHIVE_FORMAT_ISO9660_ROCKRIDGE;
452 a->archive.archive_format_name =
453 "ISO9660 with Rockridge extensions";
459 release_file(iso9660, file);
460 return (ARCHIVE_OK);
463 static int
464 archive_read_format_iso9660_read_data_skip(struct archive_read *a)
466 /* Because read_next_header always does an explicit skip
467 * to the next entry, we don't need to do anything here. */
468 (void)a; /* UNUSED */
469 return (ARCHIVE_OK);
472 static int
473 archive_read_format_iso9660_read_data(struct archive_read *a,
474 const void **buff, size_t *size, off_t *offset)
476 ssize_t bytes_read;
477 struct iso9660 *iso9660;
479 iso9660 = (struct iso9660 *)*(a->pformat_data);
480 if (iso9660->entry_bytes_remaining <= 0) {
481 *buff = NULL;
482 *size = 0;
483 *offset = iso9660->entry_sparse_offset;
484 return (ARCHIVE_EOF);
487 bytes_read = (a->compression_read_ahead)(a, buff, 1);
488 if (bytes_read == 0)
489 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC,
490 "Truncated input file");
491 if (bytes_read <= 0)
492 return (ARCHIVE_FATAL);
493 if (bytes_read > iso9660->entry_bytes_remaining)
494 bytes_read = iso9660->entry_bytes_remaining;
495 *size = bytes_read;
496 *offset = iso9660->entry_sparse_offset;
497 iso9660->entry_sparse_offset += bytes_read;
498 iso9660->entry_bytes_remaining -= bytes_read;
499 iso9660->current_position += bytes_read;
500 (a->compression_read_consume)(a, bytes_read);
501 return (ARCHIVE_OK);
504 static int
505 archive_read_format_iso9660_cleanup(struct archive_read *a)
507 struct iso9660 *iso9660;
508 struct file_info *file;
510 iso9660 = (struct iso9660 *)*(a->pformat_data);
511 while ((file = next_entry(iso9660)) != NULL)
512 release_file(iso9660, file);
513 archive_string_free(&iso9660->pathname);
514 archive_string_free(&iso9660->previous_pathname);
515 free(iso9660);
516 *(a->pformat_data) = NULL;
517 return (ARCHIVE_OK);
521 * This routine parses a single ISO directory record, makes sense
522 * of any extensions, and stores the result in memory.
524 static struct file_info *
525 parse_file_info(struct iso9660 *iso9660, struct file_info *parent,
526 const unsigned char *isodirrec)
528 struct file_info *file;
529 size_t name_len;
530 int flags;
532 /* TODO: Sanity check that name_len doesn't exceed length, etc. */
534 /* Create a new file entry and copy data from the ISO dir record. */
535 file = (struct file_info *)malloc(sizeof(*file));
536 if (file == NULL)
537 return (NULL);
538 memset(file, 0, sizeof(*file));
539 file->parent = parent;
540 if (parent != NULL)
541 parent->refcount++;
542 file->offset = toi(isodirrec + DR_extent_offset, DR_extent_size)
543 * iso9660->logical_block_size;
544 file->size = toi(isodirrec + DR_size_offset, DR_size_size);
545 file->mtime = isodate7(isodirrec + DR_date_offset);
546 file->ctime = file->atime = file->mtime;
547 name_len = (size_t)*(const unsigned char *)(isodirrec + DR_name_len_offset);
548 file->name = (char *)malloc(name_len + 1);
549 if (file->name == NULL) {
550 free(file);
551 return (NULL);
553 memcpy(file->name, isodirrec + DR_name_offset, name_len);
554 file->name[name_len] = '\0';
555 flags = *(isodirrec + DR_flags_offset);
556 if (flags & 0x02)
557 file->mode = S_IFDIR | 0700;
558 else
559 file->mode = S_IFREG | 0400;
561 /* Rockridge extensions overwrite information from above. */
563 const unsigned char *rr_start, *rr_end;
564 rr_end = (const unsigned char *)isodirrec
565 + *(isodirrec + DR_length_offset);
566 rr_start = (const unsigned char *)(isodirrec + DR_name_offset
567 + name_len);
568 if ((name_len & 1) == 0)
569 rr_start++;
570 rr_start += iso9660->suspOffset;
571 parse_rockridge(iso9660, file, rr_start, rr_end);
574 /* DEBUGGING: Warn about attributes I don't yet fully support. */
575 if ((flags & ~0x02) != 0) {
576 fprintf(stderr, "\n ** Unrecognized flag: ");
577 dump_isodirrec(stderr, isodirrec);
578 fprintf(stderr, "\n");
579 } else if (toi(isodirrec + DR_volume_sequence_number_offset, 2) != 1) {
580 fprintf(stderr, "\n ** Unrecognized sequence number: ");
581 dump_isodirrec(stderr, isodirrec);
582 fprintf(stderr, "\n");
583 } else if (*(isodirrec + DR_file_unit_size_offset) != 0) {
584 fprintf(stderr, "\n ** Unexpected file unit size: ");
585 dump_isodirrec(stderr, isodirrec);
586 fprintf(stderr, "\n");
587 } else if (*(isodirrec + DR_interleave_offset) != 0) {
588 fprintf(stderr, "\n ** Unexpected interleave: ");
589 dump_isodirrec(stderr, isodirrec);
590 fprintf(stderr, "\n");
591 } else if (*(isodirrec + DR_ext_attr_length_offset) != 0) {
592 fprintf(stderr, "\n ** Unexpected extended attribute length: ");
593 dump_isodirrec(stderr, isodirrec);
594 fprintf(stderr, "\n");
597 return (file);
600 static void
601 add_entry(struct iso9660 *iso9660, struct file_info *file)
603 /* Expand our pending files list as necessary. */
604 if (iso9660->pending_files_used >= iso9660->pending_files_allocated) {
605 struct file_info **new_pending_files;
606 int new_size = iso9660->pending_files_allocated * 2;
608 if (new_size < 1024)
609 new_size = 1024;
610 new_pending_files = (struct file_info **)malloc(new_size * sizeof(new_pending_files[0]));
611 if (new_pending_files == NULL)
612 __archive_errx(1, "Out of memory");
613 memcpy(new_pending_files, iso9660->pending_files,
614 iso9660->pending_files_allocated * sizeof(new_pending_files[0]));
615 if (iso9660->pending_files != NULL)
616 free(iso9660->pending_files);
617 iso9660->pending_files = new_pending_files;
618 iso9660->pending_files_allocated = new_size;
621 iso9660->pending_files[iso9660->pending_files_used++] = file;
624 static void
625 parse_rockridge(struct iso9660 *iso9660, struct file_info *file,
626 const unsigned char *p, const unsigned char *end)
628 (void)iso9660; /* UNUSED */
630 while (p + 4 < end /* Enough space for another entry. */
631 && p[0] >= 'A' && p[0] <= 'Z' /* Sanity-check 1st char of name. */
632 && p[1] >= 'A' && p[1] <= 'Z' /* Sanity-check 2nd char of name. */
633 && p + p[2] <= end) { /* Sanity-check length. */
634 const unsigned char *data = p + 4;
635 int data_length = p[2] - 4;
636 int version = p[3];
639 * Yes, each 'if' here does test p[0] again.
640 * Otherwise, the fall-through handling to catch
641 * unsupported extensions doesn't work.
643 switch(p[0]) {
644 case 'C':
645 if (p[0] == 'C' && p[1] == 'E' && version == 1) {
647 * CE extension comprises:
648 * 8 byte sector containing extension
649 * 8 byte offset w/in above sector
650 * 8 byte length of continuation
652 file->ce_offset = toi(data, 4)
653 * iso9660->logical_block_size
654 + toi(data + 8, 4);
655 file->ce_size = toi(data + 16, 4);
656 break;
658 /* FALLTHROUGH */
659 case 'N':
660 if (p[0] == 'N' && p[1] == 'M' && version == 1
661 && *data == 0) {
662 /* NM extension with flag byte == 0 */
664 * NM extension comprises:
665 * one byte flag
666 * rest is long name
668 /* TODO: Obey flags. */
669 char *old_name = file->name;
671 data++; /* Skip flag byte. */
672 data_length--;
673 file->name = (char *)malloc(data_length + 1);
674 if (file->name != NULL) {
675 free(old_name);
676 memcpy(file->name, data, data_length);
677 file->name[data_length] = '\0';
678 } else
679 file->name = old_name;
680 break;
682 /* FALLTHROUGH */
683 case 'P':
684 if (p[0] == 'P' && p[1] == 'D' && version == 1) {
686 * PD extension is padding;
687 * contents are always ignored.
689 break;
691 if (p[0] == 'P' && p[1] == 'X' && version == 1) {
693 * PX extension comprises:
694 * 8 bytes for mode,
695 * 8 bytes for nlinks,
696 * 8 bytes for uid,
697 * 8 bytes for gid,
698 * 8 bytes for inode.
700 if (data_length == 32) {
701 file->mode = toi(data, 4);
702 file->nlinks = toi(data + 8, 4);
703 file->uid = toi(data + 16, 4);
704 file->gid = toi(data + 24, 4);
705 file->inode = toi(data + 32, 4);
707 break;
709 /* FALLTHROUGH */
710 case 'R':
711 if (p[0] == 'R' && p[1] == 'R' && version == 1) {
712 iso9660->seenRockridge = 1;
714 * RR extension comprises:
715 * one byte flag value
717 /* TODO: Handle RR extension. */
718 break;
720 /* FALLTHROUGH */
721 case 'S':
722 if (p[0] == 'S' && p[1] == 'L' && version == 1
723 && *data == 0) {
724 int cont = 1;
725 /* SL extension with flags == 0 */
726 /* TODO: handle non-zero flag values. */
727 data++; /* Skip flag byte. */
728 data_length--;
729 while (data_length > 0) {
730 unsigned char flag = *data++;
731 unsigned char nlen = *data++;
732 data_length -= 2;
734 if (cont == 0)
735 archive_strcat(&file->symlink, "/");
736 cont = 0;
738 switch(flag) {
739 case 0x01: /* Continue */
740 archive_strncat(&file->symlink,
741 (const char *)data, nlen);
742 cont = 1;
743 break;
744 case 0x02: /* Current */
745 archive_strcat(&file->symlink, ".");
746 break;
747 case 0x04: /* Parent */
748 archive_strcat(&file->symlink, "..");
749 break;
750 case 0x08: /* Root */
751 case 0x10: /* Volume root */
752 archive_string_empty(&file->symlink);
753 break;
754 case 0x20: /* Hostname */
755 archive_strcat(&file->symlink, "hostname");
756 break;
757 case 0:
758 archive_strncat(&file->symlink,
759 (const char *)data, nlen);
760 break;
761 default:
762 /* TODO: issue a warning ? */
763 break;
765 data += nlen;
766 data_length -= nlen;
768 break;
770 if (p[0] == 'S' && p[1] == 'P'
771 && version == 1 && data_length == 7
772 && data[0] == (unsigned char)'\xbe'
773 && data[1] == (unsigned char)'\xef') {
775 * SP extension stores the suspOffset
776 * (Number of bytes to skip between
777 * filename and SUSP records.)
778 * It is mandatory by the SUSP standard
779 * (IEEE 1281).
781 * It allows SUSP to coexist with
782 * non-SUSP uses of the System
783 * Use Area by placing non-SUSP data
784 * before SUSP data.
786 * TODO: Add a check for 'SP' in
787 * first directory entry, disable all SUSP
788 * processing if not found.
790 iso9660->suspOffset = data[2];
791 break;
793 if (p[0] == 'S' && p[1] == 'T'
794 && data_length == 0 && version == 1) {
796 * ST extension marks end of this
797 * block of SUSP entries.
799 * It allows SUSP to coexist with
800 * non-SUSP uses of the System
801 * Use Area by placing non-SUSP data
802 * after SUSP data.
804 return;
806 case 'T':
807 if (p[0] == 'T' && p[1] == 'F' && version == 1) {
808 char flag = data[0];
810 * TF extension comprises:
811 * one byte flag
812 * create time (optional)
813 * modify time (optional)
814 * access time (optional)
815 * attribute time (optional)
816 * Time format and presence of fields
817 * is controlled by flag bits.
819 data++;
820 if (flag & 0x80) {
821 /* Use 17-byte time format. */
822 if (flag & 1) /* Create time. */
823 data += 17;
824 if (flag & 2) { /* Modify time. */
825 file->mtime = isodate17(data);
826 data += 17;
828 if (flag & 4) { /* Access time. */
829 file->atime = isodate17(data);
830 data += 17;
832 if (flag & 8) { /* Attribute time. */
833 file->ctime = isodate17(data);
834 data += 17;
836 } else {
837 /* Use 7-byte time format. */
838 if (flag & 1) /* Create time. */
839 data += 7;
840 if (flag & 2) { /* Modify time. */
841 file->mtime = isodate7(data);
842 data += 7;
844 if (flag & 4) { /* Access time. */
845 file->atime = isodate7(data);
846 data += 7;
848 if (flag & 8) { /* Attribute time. */
849 file->ctime = isodate7(data);
850 data += 7;
853 break;
855 /* FALLTHROUGH */
856 default:
857 /* The FALLTHROUGHs above leave us here for
858 * any unsupported extension. */
860 const unsigned char *t;
861 fprintf(stderr, "\nUnsupported RRIP extension for %s\n", file->name);
862 fprintf(stderr, " %c%c(%d):", p[0], p[1], data_length);
863 for (t = data; t < data + data_length && t < data + 16; t++)
864 fprintf(stderr, " %02x", *t);
865 fprintf(stderr, "\n");
871 p += p[2];
875 static void
876 release_file(struct iso9660 *iso9660, struct file_info *file)
878 struct file_info *parent;
880 if (file->refcount == 0) {
881 parent = file->parent;
882 if (file->name)
883 free(file->name);
884 archive_string_free(&file->symlink);
885 free(file);
886 if (parent != NULL) {
887 parent->refcount--;
888 release_file(iso9660, parent);
893 static int
894 next_entry_seek(struct archive_read *a, struct iso9660 *iso9660,
895 struct file_info **pfile)
897 struct file_info *file;
898 uint64_t offset;
900 *pfile = NULL;
901 for (;;) {
902 *pfile = file = next_entry(iso9660);
903 if (file == NULL)
904 return (ARCHIVE_EOF);
906 /* CE area precedes actual file data? Ignore it. */
907 if (file->ce_offset > file->offset) {
908 fprintf(stderr, " *** Discarding CE data.\n");
909 file->ce_offset = 0;
910 file->ce_size = 0;
913 /* If CE exists, find and read it now. */
914 if (file->ce_offset > 0)
915 offset = file->ce_offset;
916 else
917 offset = file->offset;
919 /* Seek forward to the start of the entry. */
920 /* Use fast compression_skip if it's available. */
921 if (iso9660->current_position < offset
922 && a->compression_skip != NULL) {
923 off_t step = offset - iso9660->current_position;
924 off_t bytes_read;
925 bytes_read = (a->compression_skip)(a, step);
926 iso9660->current_position += bytes_read;
929 /* Use a series of reads if compression_skip didn't
930 * get us all the way there. */
931 while (iso9660->current_position < offset) {
932 ssize_t step = offset - iso9660->current_position;
933 ssize_t bytes_read;
934 const void *buff;
936 if (step > iso9660->logical_block_size)
937 step = iso9660->logical_block_size;
938 bytes_read = (a->compression_read_ahead)(a, &buff, step);
939 if (bytes_read <= 0) {
940 release_file(iso9660, file);
941 return (ARCHIVE_FATAL);
943 if (bytes_read > step)
944 bytes_read = step;
945 iso9660->current_position += bytes_read;
946 (a->compression_read_consume)(a, bytes_read);
949 /* We found body of file; handle it now. */
950 if (offset == file->offset)
951 return (ARCHIVE_OK);
953 /* Found CE? Process it and push the file back onto list. */
954 if (offset == file->ce_offset) {
955 const void *p;
956 ssize_t size = file->ce_size;
957 ssize_t bytes_read;
958 const unsigned char *rr_start;
960 file->ce_offset = 0;
961 file->ce_size = 0;
962 bytes_read = (a->compression_read_ahead)(a, &p, size);
963 if (bytes_read > size)
964 bytes_read = size;
965 rr_start = (const unsigned char *)p;
966 parse_rockridge(iso9660, file, rr_start,
967 rr_start + bytes_read);
968 (a->compression_read_consume)(a, bytes_read);
969 iso9660->current_position += bytes_read;
970 add_entry(iso9660, file);
975 static struct file_info *
976 next_entry(struct iso9660 *iso9660)
978 int least_index;
979 uint64_t least_end_offset;
980 int i;
981 struct file_info *r;
983 if (iso9660->pending_files_used < 1)
984 return (NULL);
986 /* Assume the first file in the list is the earliest on disk. */
987 least_index = 0;
988 least_end_offset = iso9660->pending_files[0]->offset
989 + iso9660->pending_files[0]->size;
991 /* Now, try to find an earlier one. */
992 for (i = 0; i < iso9660->pending_files_used; i++) {
993 /* Use the position of the file *end* as our comparison. */
994 uint64_t end_offset = iso9660->pending_files[i]->offset
995 + iso9660->pending_files[i]->size;
996 if (iso9660->pending_files[i]->ce_offset > 0
997 && iso9660->pending_files[i]->ce_offset < iso9660->pending_files[i]->offset)
998 end_offset = iso9660->pending_files[i]->ce_offset
999 + iso9660->pending_files[i]->ce_size;
1000 if (least_end_offset > end_offset) {
1001 least_index = i;
1002 least_end_offset = end_offset;
1005 r = iso9660->pending_files[least_index];
1006 iso9660->pending_files[least_index]
1007 = iso9660->pending_files[--iso9660->pending_files_used];
1008 return (r);
1011 static unsigned int
1012 toi(const void *p, int n)
1014 const unsigned char *v = (const unsigned char *)p;
1015 if (n > 1)
1016 return v[0] + 256 * toi(v + 1, n - 1);
1017 if (n == 1)
1018 return v[0];
1019 return (0);
1022 static time_t
1023 isodate7(const unsigned char *v)
1025 struct tm tm;
1026 int offset;
1027 memset(&tm, 0, sizeof(tm));
1028 tm.tm_year = v[0];
1029 tm.tm_mon = v[1] - 1;
1030 tm.tm_mday = v[2];
1031 tm.tm_hour = v[3];
1032 tm.tm_min = v[4];
1033 tm.tm_sec = v[5];
1034 /* v[6] is the signed timezone offset, in 1/4-hour increments. */
1035 offset = ((const signed char *)v)[6];
1036 if (offset > -48 && offset < 52) {
1037 tm.tm_hour -= offset / 4;
1038 tm.tm_min -= (offset % 4) * 15;
1040 return (time_from_tm(&tm));
1043 static time_t
1044 isodate17(const unsigned char *v)
1046 struct tm tm;
1047 int offset;
1048 memset(&tm, 0, sizeof(tm));
1049 tm.tm_year = (v[0] - '0') * 1000 + (v[1] - '0') * 100
1050 + (v[2] - '0') * 10 + (v[3] - '0')
1051 - 1900;
1052 tm.tm_mon = (v[4] - '0') * 10 + (v[5] - '0');
1053 tm.tm_mday = (v[6] - '0') * 10 + (v[7] - '0');
1054 tm.tm_hour = (v[8] - '0') * 10 + (v[9] - '0');
1055 tm.tm_min = (v[10] - '0') * 10 + (v[11] - '0');
1056 tm.tm_sec = (v[12] - '0') * 10 + (v[13] - '0');
1057 /* v[16] is the signed timezone offset, in 1/4-hour increments. */
1058 offset = ((const signed char *)v)[16];
1059 if (offset > -48 && offset < 52) {
1060 tm.tm_hour -= offset / 4;
1061 tm.tm_min -= (offset % 4) * 15;
1063 return (time_from_tm(&tm));
1067 * timegm() converts a struct tm to a time_t, except it isn't standard,
1068 * so I provide my own function here that (ideally) is just a wrapper
1069 * for timegm().
1071 static time_t
1072 time_from_tm(struct tm *t)
1074 #if HAVE_TIMEGM
1075 return (timegm(t));
1076 #else
1078 * Unfortunately, timegm() isn't standard. The standard
1079 * mktime() function is a close match, except that it uses
1080 * local timezone instead of GMT. Close enough for now.
1081 * Note that it is not possible to emulate timegm() using
1082 * completely standard interfaces:
1083 * * ANSI C90 does not even guarantee that time_t is
1084 * an arithmetic type, so time adjustments can only be
1085 * done by manipulating struct tm elements. You cannot
1086 * portably calculate time_t values.
1087 * * POSIX does promise that time_t is an arithmetic type
1088 * measured in seconds, so you can do time_t calculations
1089 * while remaining POSIX-compliant.
1090 * * Neither ANSI nor POSIX provides an easy way to measure
1091 * the timezone offset, so you can't adjust mktime() to
1092 * work like timegm().
1093 * * POSIX does not promise that the epoch begins in 1970,
1094 * so you can't write a portable timegm() function from
1095 * scratch.
1096 * In practice, of course, mktime() is a reasonable approximation
1097 * and most POSIX systems do use seconds since 1970, so you
1098 * can roll your own and have it work on all but a few pretty
1099 * whacky systems.
1101 time_t result = mktime(t);
1102 /* TODO: Find a way to improve this approximation to timegm(). */
1103 return result;
1104 #endif
1107 static const char *
1108 build_pathname(struct archive_string *as, struct file_info *file)
1110 if (file->parent != NULL && file->parent->name[0] != '\0') {
1111 build_pathname(as, file->parent);
1112 archive_strcat(as, "/");
1114 if (file->name[0] == '\0')
1115 archive_strcat(as, ".");
1116 else
1117 archive_strcat(as, file->name);
1118 return (as->s);
1121 static void
1122 dump_isodirrec(FILE *out, const unsigned char *isodirrec)
1124 fprintf(out, " l %d,",
1125 toi(isodirrec + DR_length_offset, DR_length_size));
1126 fprintf(out, " a %d,",
1127 toi(isodirrec + DR_ext_attr_length_offset, DR_ext_attr_length_size));
1128 fprintf(out, " ext 0x%x,",
1129 toi(isodirrec + DR_extent_offset, DR_extent_size));
1130 fprintf(out, " s %d,",
1131 toi(isodirrec + DR_size_offset, DR_extent_size));
1132 fprintf(out, " f 0x%02x,",
1133 toi(isodirrec + DR_flags_offset, DR_flags_size));
1134 fprintf(out, " u %d,",
1135 toi(isodirrec + DR_file_unit_size_offset, DR_file_unit_size_size));
1136 fprintf(out, " ilv %d,",
1137 toi(isodirrec + DR_interleave_offset, DR_interleave_size));
1138 fprintf(out, " seq %d,",
1139 toi(isodirrec + DR_volume_sequence_number_offset, DR_volume_sequence_number_size));
1140 fprintf(out, " nl %d:",
1141 toi(isodirrec + DR_name_len_offset, DR_name_len_size));
1142 fprintf(out, " `%.*s'",
1143 toi(isodirrec + DR_name_len_offset, DR_name_len_size), isodirrec + DR_name_offset);