1 /*-
2 * Copyright (c) 2003-2007 Tim Kientzle
3 * Copyright (c) 2011-2012 Michihiro NAKAJIMA
4 * All rights reserved.
5 *
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following conditions
8 * are met:
9 * 1. Redistributions of source code must retain the above copyright
10 * notice, this list of conditions and the following disclaimer.
11 * 2. Redistributions in binary form must reproduce the above copyright
12 * notice, this list of conditions and the following disclaimer in the
13 * documentation and/or other materials provided with the distribution.
14 *
15 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
16 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
17 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
18 * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
19 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
20 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
21 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
22 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
23 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
24 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
25 */
26
27 #include "archive_platform.h"
28
29 #ifdef HAVE_ERRNO_H
30 #include <errno.h>
31 #endif
32 #include <stdio.h>
33 #ifdef HAVE_STDLIB_H
34 #include <stdlib.h>
35 #endif
36 #ifdef HAVE_STRING_H
37 #include <string.h>
38 #endif
39
40 #include "archive.h"
41 #include "archive_entry.h"
42 #include "archive_entry_locale.h"
43 #include "archive_private.h"
44 #include "archive_write_private.h"
45 #include "archive_write_set_format_private.h"
46
47 static ssize_t archive_write_binary_data(struct archive_write *,
48 const void *buff, size_t s);
49 static int archive_write_binary_close(struct archive_write *);
50 static int archive_write_binary_free(struct archive_write *);
51 static int archive_write_binary_finish_entry(struct archive_write *);
52 static int archive_write_binary_header(struct archive_write *,
53 struct archive_entry *);
54 static int archive_write_binary_options(struct archive_write *,
55 const char *, const char *);
56 static int write_header(struct archive_write *, struct archive_entry *);
57
58 struct cpio {
59 uint64_t entry_bytes_remaining;
60
61 int64_t ino_next;
62
63 struct { int64_t old; int new;} *ino_list;
64 size_t ino_list_size;
65 size_t ino_list_next;
66
67 struct archive_string_conv *opt_sconv;
68 struct archive_string_conv *sconv_default;
69 int init_default_conversion;
70 };
71
72 /* This struct needs to be packed to get the header right */
73
74 #if defined(__GNUC__)
75 #define PACKED(x) x __attribute__((packed))
76 #elif defined(_MSC_VER)
77 #define PACKED(x) __pragma(pack(push, 1)) x __pragma(pack(pop))
78 #else
79 #define PACKED(x) x
80 #endif
81
82 #define HSIZE 26
83
84 PACKED(struct cpio_binary_header {
85 uint16_t h_magic;
86 uint16_t h_dev;
87 uint16_t h_ino;
88 uint16_t h_mode;
89 uint16_t h_uid;
90 uint16_t h_gid;
91 uint16_t h_nlink;
92 uint16_t h_majmin;
93 uint32_t h_mtime;
94 uint16_t h_namesize;
95 uint32_t h_filesize;
96 });
97
98 /* Back in the day, the 7th Edition cpio.c had this, to
99 * adapt to, as the comment said, "VAX, Interdata, ...":
100 *
101 * union { long l; short s[2]; char c[4]; } U;
102 * #define MKSHORT(v,lv) {U.l=1L;if(U.c[0]) U.l=lv,v[0]=U.s[1],v[1]=U.s[0]; else U.l=lv,v[0]=U.s[0],v[1]=U.s[1];}
103 * long mklong(v)
104 * short v[];
105 * {
106 * U.l = 1;
107 * if(U.c[0])
108 * U.s[0] = v[1], U.s[1] = v[0];
109 * else
110 * U.s[0] = v[0], U.s[1] = v[1];
111 * return U.l;
112 * }
113 *
114 * Of course, that assumes that all machines have little-endian shorts,
115 * and just adapts the others to the special endianness of the PDP-11.
116 *
117 * Now, we could do this:
118 *
119 * union { uint32_t l; uint16_t s[2]; uint8_t c[4]; } U;
120 * #define PUTI16(v,sv) {U.s[0]=1;if(U.c[0]) v=sv; else U.s[0]=sv,U.c[2]=U.c[1],U.c[3]=U.c[0],v=U.s[1];}
121 * #define PUTI32(v,lv) {char_t Ut;U.l=1;if(U.c[0]) U.l=lv,v[0]=U.s[1],v[1]=U.s[0]; else U.l=lv,Ut=U.c[0],U.c[0]=U.c[1],U.c[1]=Ut,Ut=U.c[2],U.c[2]=U.c[3],U.c[3]=Ut,v[0]=U.s[0],v[1]=U.s[1];}
122 *
123 * ...but it feels a little better to do it like this:
124 */
125
la_swap16(uint16_t in)126 static uint16_t la_swap16(uint16_t in) {
127 union {
128 uint16_t s[2];
129 uint8_t c[4];
130 } U;
131 U.s[0] = 1;
132 if (U.c[0])
133 return in;
134 else {
135 U.s[0] = in;
136 U.c[2] = U.c[1];
137 U.c[3] = U.c[0];
138 return U.s[1];
139 }
140 /* NOTREACHED */
141 }
142
la_swap32(uint32_t in)143 static uint32_t la_swap32(uint32_t in) {
144 union {
145 uint32_t l;
146 uint16_t s[2];
147 uint8_t c[4];
148 } U;
149 U.l = 1;
150 if (U.c[0]) { /* Little-endian */
151 uint16_t t;
152 U.l = in;
153 t = U.s[0];
154 U.s[0] = U.s[1];
155 U.s[1] = t;
156 } else if (U.c[3]) { /* Big-endian */
157 U.l = in;
158 U.s[0] = la_swap16(U.s[0]);
159 U.s[1] = la_swap16(U.s[1]);
160 } else { /* PDP-endian */
161 U.l = in;
162 }
163 return U.l;
164 }
165
166 /*
167 * Set output format to the selected binary variant
168 */
169 static int
archive_write_set_format_cpio_binary(struct archive * _a,int format)170 archive_write_set_format_cpio_binary(struct archive *_a, int format)
171 {
172 struct archive_write *a = (struct archive_write *)_a;
173 struct cpio *cpio;
174
175 if (sizeof(struct cpio_binary_header) != HSIZE) {
176 archive_set_error(&a->archive, EINVAL,
177 "Binary cpio format not supported on this platform");
178 return (ARCHIVE_FATAL);
179 }
180
181 archive_check_magic(_a, ARCHIVE_WRITE_MAGIC,
182 ARCHIVE_STATE_NEW, "archive_write_set_format_cpio_binary");
183
184 /* If someone else was already registered, unregister them. */
185 if (a->format_free != NULL)
186 (a->format_free)(a);
187
188 cpio = calloc(1, sizeof(*cpio));
189 if (cpio == NULL) {
190 archive_set_error(&a->archive, ENOMEM, "Can't allocate cpio data");
191 return (ARCHIVE_FATAL);
192 }
193 a->format_data = cpio;
194 a->format_name = "cpio";
195 a->format_options = archive_write_binary_options;
196 a->format_write_header = archive_write_binary_header;
197 a->format_write_data = archive_write_binary_data;
198 a->format_finish_entry = archive_write_binary_finish_entry;
199 a->format_close = archive_write_binary_close;
200 a->format_free = archive_write_binary_free;
201 a->archive.archive_format = format;
202 switch (format) {
203 case ARCHIVE_FORMAT_CPIO_PWB:
204 a->archive.archive_format_name = "PWB cpio";
205 break;
206 case ARCHIVE_FORMAT_CPIO_BIN_LE:
207 a->archive.archive_format_name = "7th Edition cpio";
208 break;
209 default:
210 archive_set_error(&a->archive, EINVAL, "binary format must be 'pwb' or 'bin'");
211 return (ARCHIVE_FATAL);
212 }
213 return (ARCHIVE_OK);
214 }
215
216 /*
217 * Set output format to PWB (6th Edition) binary format
218 */
219 int
archive_write_set_format_cpio_pwb(struct archive * _a)220 archive_write_set_format_cpio_pwb(struct archive *_a)
221 {
222 return archive_write_set_format_cpio_binary(_a, ARCHIVE_FORMAT_CPIO_PWB);
223 }
224
225 /*
226 * Set output format to 7th Edition binary format
227 */
228 int
archive_write_set_format_cpio_bin(struct archive * _a)229 archive_write_set_format_cpio_bin(struct archive *_a)
230 {
231 return archive_write_set_format_cpio_binary(_a, ARCHIVE_FORMAT_CPIO_BIN_LE);
232 }
233
234 static int
archive_write_binary_options(struct archive_write * a,const char * key,const char * val)235 archive_write_binary_options(struct archive_write *a, const char *key,
236 const char *val)
237 {
238 struct cpio *cpio = (struct cpio *)a->format_data;
239 int ret = ARCHIVE_FAILED;
240
241 if (strcmp(key, "hdrcharset") == 0) {
242 if (val == NULL || val[0] == 0)
243 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC,
244 "%s: hdrcharset option needs a character-set name",
245 a->format_name);
246 else {
247 cpio->opt_sconv = archive_string_conversion_to_charset(
248 &a->archive, val, 0);
249 if (cpio->opt_sconv != NULL)
250 ret = ARCHIVE_OK;
251 else
252 ret = ARCHIVE_FATAL;
253 }
254 return (ret);
255 }
256
257 /* Note: The "warn" return is just to inform the options
258 * supervisor that we didn't handle it. It will generate
259 * a suitable error if no one used this option. */
260 return (ARCHIVE_WARN);
261 }
262
263 /*
264 * Ino values are as long as 64 bits on some systems; cpio format
265 * only allows 16 bits and relies on the ino values to identify hardlinked
266 * files. So, we can't merely "hash" the ino numbers since collisions
267 * would corrupt the archive. Instead, we generate synthetic ino values
268 * to store in the archive and maintain a map of original ino values to
269 * synthetic ones so we can preserve hardlink information.
270 *
271 * TODO: Make this more efficient. It's not as bad as it looks (most
272 * files don't have any hardlinks and we don't do any work here for those),
273 * but it wouldn't be hard to do better.
274 *
275 * TODO: Work with dev/ino pairs here instead of just ino values.
276 */
277 static int
synthesize_ino_value(struct cpio * cpio,struct archive_entry * entry)278 synthesize_ino_value(struct cpio *cpio, struct archive_entry *entry)
279 {
280 int64_t ino = archive_entry_ino64(entry);
281 int ino_new;
282 size_t i;
283
284 /*
285 * If no index number was given, don't assign one. In
286 * particular, this handles the end-of-archive marker
287 * correctly by giving it a zero index value. (This is also
288 * why we start our synthetic index numbers with one below.)
289 */
290 if (ino == 0)
291 return (0);
292
293 /* Don't store a mapping if we don't need to. */
294 if (archive_entry_nlink(entry) < 2) {
295 return (int)(++cpio->ino_next);
296 }
297
298 /* Look up old ino; if we have it, this is a hardlink
299 * and we reuse the same value. */
300 for (i = 0; i < cpio->ino_list_next; ++i) {
301 if (cpio->ino_list[i].old == ino)
302 return (cpio->ino_list[i].new);
303 }
304
305 /* Assign a new index number. */
306 ino_new = (int)(++cpio->ino_next);
307
308 /* Ensure space for the new mapping. */
309 if (cpio->ino_list_size <= cpio->ino_list_next) {
310 size_t newsize = cpio->ino_list_size < 512
311 ? 512 : cpio->ino_list_size * 2;
312 void *newlist = realloc(cpio->ino_list,
313 sizeof(cpio->ino_list[0]) * newsize);
314 if (newlist == NULL)
315 return (-1);
316
317 cpio->ino_list_size = newsize;
318 cpio->ino_list = newlist;
319 }
320
321 /* Record and return the new value. */
322 cpio->ino_list[cpio->ino_list_next].old = ino;
323 cpio->ino_list[cpio->ino_list_next].new = ino_new;
324 ++cpio->ino_list_next;
325 return (ino_new);
326 }
327
328
329 static struct archive_string_conv *
get_sconv(struct archive_write * a)330 get_sconv(struct archive_write *a)
331 {
332 struct cpio *cpio;
333 struct archive_string_conv *sconv;
334
335 cpio = (struct cpio *)a->format_data;
336 sconv = cpio->opt_sconv;
337 if (sconv == NULL) {
338 if (!cpio->init_default_conversion) {
339 cpio->sconv_default =
340 archive_string_default_conversion_for_write(
341 &(a->archive));
342 cpio->init_default_conversion = 1;
343 }
344 sconv = cpio->sconv_default;
345 }
346 return (sconv);
347 }
348
349 static int
archive_write_binary_header(struct archive_write * a,struct archive_entry * entry)350 archive_write_binary_header(struct archive_write *a, struct archive_entry *entry)
351 {
352 const char *path;
353 size_t len;
354
355 if (archive_entry_filetype(entry) == 0 && archive_entry_hardlink(entry) == NULL) {
356 archive_set_error(&a->archive, -1, "Filetype required");
357 return (ARCHIVE_FAILED);
358 }
359
360 if (archive_entry_pathname_l(entry, &path, &len, get_sconv(a)) != 0
361 && errno == ENOMEM) {
362 archive_set_error(&a->archive, ENOMEM,
363 "Can't allocate memory for Pathname");
364 return (ARCHIVE_FATAL);
365 }
366 if (len == 0 || path == NULL || path[0] == '\0') {
367 archive_set_error(&a->archive, -1, "Pathname required");
368 return (ARCHIVE_FAILED);
369 }
370
371 if (!archive_entry_size_is_set(entry) || archive_entry_size(entry) < 0) {
372 archive_set_error(&a->archive, -1, "Size required");
373 return (ARCHIVE_FAILED);
374 }
375 return write_header(a, entry);
376 }
377
378 static int
write_header(struct archive_write * a,struct archive_entry * entry)379 write_header(struct archive_write *a, struct archive_entry *entry)
380 {
381 struct cpio *cpio;
382 const char *p, *path;
383 int pathlength, ret, ret_final;
384 int64_t ino;
385 struct cpio_binary_header h;
386 struct archive_string_conv *sconv;
387 struct archive_entry *entry_main;
388 size_t len;
389
390 cpio = (struct cpio *)a->format_data;
391 ret_final = ARCHIVE_OK;
392 sconv = get_sconv(a);
393
394 #if defined(_WIN32) && !defined(__CYGWIN__)
395 /* Make sure the path separators in pathname, hardlink and symlink
396 * are all slash '/', not the Windows path separator '\'. */
397 entry_main = __la_win_entry_in_posix_pathseparator(entry);
398 if (entry_main == NULL) {
399 archive_set_error(&a->archive, ENOMEM,
400 "Can't allocate ustar data");
401 return(ARCHIVE_FATAL);
402 }
403 if (entry != entry_main)
404 entry = entry_main;
405 else
406 entry_main = NULL;
407 #else
408 entry_main = NULL;
409 #endif
410
411 ret = archive_entry_pathname_l(entry, &path, &len, sconv);
412 if (ret != 0) {
413 if (errno == ENOMEM) {
414 archive_set_error(&a->archive, ENOMEM,
415 "Can't allocate memory for Pathname");
416 ret_final = ARCHIVE_FATAL;
417 goto exit_write_header;
418 }
419 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT,
420 "Can't translate pathname '%s' to %s",
421 archive_entry_pathname(entry),
422 archive_string_conversion_charset_name(sconv));
423 ret_final = ARCHIVE_WARN;
424 }
425 /* Include trailing null */
426 pathlength = (int)len + 1;
427
428 h.h_magic = la_swap16(070707);
429 h.h_dev = la_swap16(archive_entry_dev(entry));
430
431 ino = synthesize_ino_value(cpio, entry);
432 if (ino < 0) {
433 archive_set_error(&a->archive, ENOMEM,
434 "No memory for ino translation table");
435 ret_final = ARCHIVE_FATAL;
436 goto exit_write_header;
437 } else if (ino > 077777) {
438 archive_set_error(&a->archive, ERANGE,
439 "Too many files for this cpio format");
440 ret_final = ARCHIVE_FATAL;
441 goto exit_write_header;
442 }
443 h.h_ino = la_swap16((uint16_t)ino);
444
445 h.h_mode = archive_entry_mode(entry);
446 if (((h.h_mode & AE_IFMT) == AE_IFSOCK) || ((h.h_mode & AE_IFMT) == AE_IFIFO)) {
447 archive_set_error(&a->archive, EINVAL,
448 "sockets and fifos cannot be represented in the binary cpio formats");
449 ret_final = ARCHIVE_FATAL;
450 goto exit_write_header;
451 }
452 if (a->archive.archive_format == ARCHIVE_FORMAT_CPIO_PWB) {
453 if ((h.h_mode & AE_IFMT) == AE_IFLNK) {
454 archive_set_error(&a->archive, EINVAL,
455 "symbolic links cannot be represented in the PWB cpio format");
456 ret_final = ARCHIVE_FATAL;
457 goto exit_write_header;
458 }
459 /* we could turn off AE_IFREG here, but it does no harm, */
460 /* and allows v7 cpio to read the entry without confusion */
461 }
462 h.h_mode = la_swap16(h.h_mode);
463
464 h.h_uid = la_swap16((uint16_t)archive_entry_uid(entry));
465 h.h_gid = la_swap16((uint16_t)archive_entry_gid(entry));
466 h.h_nlink = la_swap16((uint16_t)archive_entry_nlink(entry));
467
468 if (archive_entry_filetype(entry) == AE_IFBLK
469 || archive_entry_filetype(entry) == AE_IFCHR)
470 h.h_majmin = la_swap16(archive_entry_rdev(entry));
471 else
472 h.h_majmin = 0;
473
474 h.h_mtime = la_swap32((uint32_t)archive_entry_mtime(entry));
475 h.h_namesize = la_swap16(pathlength);
476
477 /* Non-regular files don't store bodies. */
478 if (archive_entry_filetype(entry) != AE_IFREG)
479 archive_entry_set_size(entry, 0);
480
481 /* Symlinks get the link written as the body of the entry. */
482 ret = archive_entry_symlink_l(entry, &p, &len, sconv);
483 if (ret != 0) {
484 if (errno == ENOMEM) {
485 archive_set_error(&a->archive, ENOMEM,
486 "Can't allocate memory for Linkname");
487 ret_final = ARCHIVE_FATAL;
488 goto exit_write_header;
489 }
490 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT,
491 "Can't translate linkname '%s' to %s",
492 archive_entry_symlink(entry),
493 archive_string_conversion_charset_name(sconv));
494 ret_final = ARCHIVE_WARN;
495 }
496
497 if (len > 0 && p != NULL && *p != '\0') {
498 if (a->archive.archive_format == ARCHIVE_FORMAT_CPIO_PWB) {
499 archive_set_error(&a->archive, EINVAL,
500 "symlinks are not supported by UNIX V6 or by PWB cpio");
501 ret_final = ARCHIVE_FATAL;
502 goto exit_write_header;
503 }
504 h.h_filesize = la_swap32((uint32_t)strlen(p)); /* symlink */
505 } else {
506 if ((a->archive.archive_format == ARCHIVE_FORMAT_CPIO_PWB) &&
507 (archive_entry_size(entry) > 256*256*256-1)) {
508 archive_set_error(&a->archive, ERANGE,
509 "File is too large for PWB binary cpio format.");
510 ret_final = ARCHIVE_FAILED;
511 goto exit_write_header;
512 } else if (archive_entry_size(entry) > INT32_MAX) {
513 archive_set_error(&a->archive, ERANGE,
514 "File is too large for binary cpio format.");
515 ret_final = ARCHIVE_FAILED;
516 goto exit_write_header;
517 }
518 h.h_filesize = la_swap32((uint32_t)archive_entry_size(entry)); /* file */
519 }
520
521 ret = __archive_write_output(a, &h, HSIZE);
522 if (ret != ARCHIVE_OK) {
523 ret_final = ARCHIVE_FATAL;
524 goto exit_write_header;
525 }
526
527 ret = __archive_write_output(a, path, pathlength);
528 if ((ret == ARCHIVE_OK) && ((pathlength % 2) != 0))
529 ret = __archive_write_nulls(a, 1);
530 if (ret != ARCHIVE_OK) {
531 ret_final = ARCHIVE_FATAL;
532 goto exit_write_header;
533 }
534
535 cpio->entry_bytes_remaining = archive_entry_size(entry);
536 if ((cpio->entry_bytes_remaining % 2) != 0)
537 cpio->entry_bytes_remaining++;
538
539 /* Write the symlink now. */
540 if (p != NULL && *p != '\0') {
541 ret = __archive_write_output(a, p, strlen(p));
542 if ((ret == ARCHIVE_OK) && ((strlen(p) % 2) != 0))
543 ret = __archive_write_nulls(a, 1);
544 if (ret != ARCHIVE_OK) {
545 ret_final = ARCHIVE_FATAL;
546 goto exit_write_header;
547 }
548 }
549
550 exit_write_header:
551 archive_entry_free(entry_main);
552 return (ret_final);
553 }
554
555 static ssize_t
archive_write_binary_data(struct archive_write * a,const void * buff,size_t s)556 archive_write_binary_data(struct archive_write *a, const void *buff, size_t s)
557 {
558 struct cpio *cpio;
559 int ret;
560
561 cpio = (struct cpio *)a->format_data;
562 if (s > cpio->entry_bytes_remaining)
563 s = (size_t)cpio->entry_bytes_remaining;
564
565 ret = __archive_write_output(a, buff, s);
566 cpio->entry_bytes_remaining -= s;
567 if (ret >= 0)
568 return (s);
569 else
570 return (ret);
571 }
572
573 static int
archive_write_binary_close(struct archive_write * a)574 archive_write_binary_close(struct archive_write *a)
575 {
576 int er;
577 struct archive_entry *trailer;
578
579 trailer = archive_entry_new2(NULL);
580 if (trailer == NULL) {
581 return ARCHIVE_FATAL;
582 }
583 /* nlink = 1 here for GNU cpio compat. */
584 archive_entry_set_nlink(trailer, 1);
585 archive_entry_set_size(trailer, 0);
586 archive_entry_set_pathname(trailer, "TRAILER!!!");
587 er = write_header(a, trailer);
588 archive_entry_free(trailer);
589 return (er);
590 }
591
592 static int
archive_write_binary_free(struct archive_write * a)593 archive_write_binary_free(struct archive_write *a)
594 {
595 struct cpio *cpio;
596
597 cpio = (struct cpio *)a->format_data;
598 free(cpio->ino_list);
599 free(cpio);
600 a->format_data = NULL;
601 return (ARCHIVE_OK);
602 }
603
604 static int
archive_write_binary_finish_entry(struct archive_write * a)605 archive_write_binary_finish_entry(struct archive_write *a)
606 {
607 struct cpio *cpio;
608
609 cpio = (struct cpio *)a->format_data;
610 return (__archive_write_nulls(a,
611 (size_t)cpio->entry_bytes_remaining));
612 }
613