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_integer.h"
44 #include "archive_private.h"
45 #include "archive_write_private.h"
46 #include "archive_write_set_format_private.h"
47
48 static ssize_t archive_write_binary_data(struct archive_write *,
49 const void *buff, size_t s);
50 static int archive_write_binary_close(struct archive_write *);
51 static int archive_write_binary_free(struct archive_write *);
52 static int archive_write_binary_finish_entry(struct archive_write *);
53 static int archive_write_binary_header(struct archive_write *,
54 struct archive_entry *);
55 static int archive_write_binary_options(struct archive_write *,
56 const char *, const char *);
57 static int write_header(struct archive_write *, struct archive_entry *);
58
59 struct cpio {
60 uint64_t entry_bytes_remaining;
61
62 int64_t ino_next;
63
64 struct { int64_t old; int new;} *ino_list;
65 size_t ino_list_size;
66 size_t ino_list_next;
67
68 struct archive_string_conv *opt_sconv;
69 struct archive_string_conv *sconv_default;
70 int init_default_conversion;
71 };
72
73 /* This struct needs to be packed to get the header right */
74
75 #if defined(__GNUC__)
76 #define PACKED(x) x __attribute__((packed))
77 #elif defined(_MSC_VER)
78 #define PACKED(x) __pragma(pack(push, 1)) x __pragma(pack(pop))
79 #else
80 #define PACKED(x) x
81 #endif
82
83 #define HSIZE 26
84
85 PACKED(struct cpio_binary_header {
86 uint16_t h_magic;
87 uint16_t h_dev;
88 uint16_t h_ino;
89 uint16_t h_mode;
90 uint16_t h_uid;
91 uint16_t h_gid;
92 uint16_t h_nlink;
93 uint16_t h_majmin;
94 uint32_t h_mtime;
95 uint16_t h_namesize;
96 uint32_t h_filesize;
97 });
98
99 /* Back in the day, the 7th Edition cpio.c had this, to
100 * adapt to, as the comment said, "VAX, Interdata, ...":
101 *
102 * union { long l; short s[2]; char c[4]; } U;
103 * #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];}
104 * long mklong(v)
105 * short v[];
106 * {
107 * U.l = 1;
108 * if(U.c[0])
109 * U.s[0] = v[1], U.s[1] = v[0];
110 * else
111 * U.s[0] = v[0], U.s[1] = v[1];
112 * return U.l;
113 * }
114 *
115 * Of course, that assumes that all machines have little-endian shorts,
116 * and just adapts the others to the special endianness of the PDP-11.
117 *
118 * Now, we could do this:
119 *
120 * union { uint32_t l; uint16_t s[2]; uint8_t c[4]; } U;
121 * #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];}
122 * #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];}
123 *
124 * ...but it feels a little better to do it like this:
125 */
126
la_swap16(uint16_t in)127 static uint16_t la_swap16(uint16_t in) {
128 union {
129 uint16_t s[2];
130 uint8_t c[4];
131 } U;
132 U.s[0] = 1;
133 if (U.c[0])
134 return in;
135 else {
136 U.s[0] = in;
137 U.c[2] = U.c[1];
138 U.c[3] = U.c[0];
139 return U.s[1];
140 }
141 /* NOTREACHED */
142 }
143
la_swap32(uint32_t in)144 static uint32_t la_swap32(uint32_t in) {
145 union {
146 uint32_t l;
147 uint16_t s[2];
148 uint8_t c[4];
149 } U;
150 U.l = 1;
151 if (U.c[0]) { /* Little-endian */
152 uint16_t t;
153 U.l = in;
154 t = U.s[0];
155 U.s[0] = U.s[1];
156 U.s[1] = t;
157 } else if (U.c[3]) { /* Big-endian */
158 U.l = in;
159 U.s[0] = la_swap16(U.s[0]);
160 U.s[1] = la_swap16(U.s[1]);
161 } else { /* PDP-endian */
162 U.l = in;
163 }
164 return U.l;
165 }
166
167 /*
168 * Set output format to the selected binary variant
169 */
170 static int
archive_write_set_format_cpio_binary(struct archive * _a,int format)171 archive_write_set_format_cpio_binary(struct archive *_a, int format)
172 {
173 struct archive_write *a = (struct archive_write *)_a;
174 struct cpio *cpio;
175
176 if (sizeof(struct cpio_binary_header) != HSIZE) {
177 archive_set_error(&a->archive, EINVAL,
178 "Binary cpio format not supported on this platform");
179 return (ARCHIVE_FATAL);
180 }
181
182 archive_check_magic(_a, ARCHIVE_WRITE_MAGIC,
183 ARCHIVE_STATE_NEW, "archive_write_set_format_cpio_binary");
184
185 /* If someone else was already registered, unregister them. */
186 (void)__archive_write_unregister_format(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 = 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, size;
311 if (cpio->ino_list_size < 512)
312 newsize = 512;
313 else if (archive_ckd_mul_size(&newsize,
314 cpio->ino_list_size, 2))
315 return (-1);
316 if (archive_ckd_mul_size(&size,
317 newsize, sizeof(cpio->ino_list[0])))
318 return (-1);
319 void *newlist = realloc(cpio->ino_list, size);
320 if (newlist == NULL)
321 return (-1);
322
323 cpio->ino_list_size = newsize;
324 cpio->ino_list = newlist;
325 }
326
327 /* Record and return the new value. */
328 cpio->ino_list[cpio->ino_list_next].old = ino;
329 cpio->ino_list[cpio->ino_list_next].new = ino_new;
330 ++cpio->ino_list_next;
331 return (ino_new);
332 }
333
334
335 static struct archive_string_conv *
get_sconv(struct archive_write * a)336 get_sconv(struct archive_write *a)
337 {
338 struct cpio *cpio = a->format_data;
339 struct archive_string_conv *sconv;
340
341 sconv = cpio->opt_sconv;
342 if (sconv == NULL) {
343 if (!cpio->init_default_conversion) {
344 cpio->sconv_default =
345 archive_string_default_conversion_for_write(
346 &(a->archive));
347 cpio->init_default_conversion = 1;
348 }
349 sconv = cpio->sconv_default;
350 }
351 return (sconv);
352 }
353
354 static int
archive_write_binary_header(struct archive_write * a,struct archive_entry * entry)355 archive_write_binary_header(struct archive_write *a, struct archive_entry *entry)
356 {
357 const char *path;
358 size_t len;
359
360 if (archive_entry_filetype(entry) == 0 && archive_entry_hardlink(entry) == NULL) {
361 archive_set_error(&a->archive, -1, "Filetype required");
362 return (ARCHIVE_FAILED);
363 }
364
365 if (archive_entry_pathname_l(entry, &path, &len, get_sconv(a)) != 0
366 && errno == ENOMEM) {
367 archive_set_error(&a->archive, ENOMEM,
368 "Can't allocate memory for Pathname");
369 return (ARCHIVE_FATAL);
370 }
371 if (len == 0 || path == NULL || path[0] == '\0') {
372 archive_set_error(&a->archive, -1, "Pathname required");
373 return (ARCHIVE_FAILED);
374 }
375
376 if (!archive_entry_size_is_set(entry) || archive_entry_size(entry) < 0) {
377 archive_set_error(&a->archive, -1, "Size required");
378 return (ARCHIVE_FAILED);
379 }
380 return write_header(a, entry);
381 }
382
383 static int
write_header(struct archive_write * a,struct archive_entry * entry)384 write_header(struct archive_write *a, struct archive_entry *entry)
385 {
386 struct cpio *cpio = a->format_data;
387 const char *p, *path;
388 int ret, ret_final;
389 int64_t ino;
390 struct cpio_binary_header h;
391 struct archive_string_conv *sconv;
392 struct archive_entry *entry_main;
393 size_t len, pathlength;
394
395 ret_final = ARCHIVE_OK;
396 sconv = get_sconv(a);
397
398 #if defined(_WIN32) && !defined(__CYGWIN__)
399 /* Make sure the path separators in pathname, hardlink and symlink
400 * are all slash '/', not the Windows path separator '\'. */
401 entry_main = __la_win_entry_in_posix_pathseparator(entry);
402 if (entry_main == NULL) {
403 archive_set_error(&a->archive, ENOMEM,
404 "Can't allocate ustar data");
405 return(ARCHIVE_FATAL);
406 }
407 if (entry != entry_main)
408 entry = entry_main;
409 else
410 entry_main = NULL;
411 #else
412 entry_main = NULL;
413 #endif
414
415 ret = archive_entry_pathname_l(entry, &path, &len, sconv);
416 if (ret != 0) {
417 if (errno == ENOMEM) {
418 archive_set_error(&a->archive, ENOMEM,
419 "Can't allocate memory for Pathname");
420 ret_final = ARCHIVE_FATAL;
421 goto exit_write_header;
422 }
423 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT,
424 "Can't translate pathname '%s' to %s",
425 archive_entry_pathname(entry),
426 archive_string_conversion_charset_name(sconv));
427 ret_final = ARCHIVE_WARN;
428 }
429 /* Include trailing null */
430 pathlength = len + 1;
431
432 h.h_magic = la_swap16(070707);
433 h.h_dev = la_swap16(archive_entry_dev(entry));
434
435 ino = synthesize_ino_value(cpio, entry);
436 if (ino < 0) {
437 archive_set_error(&a->archive, ENOMEM,
438 "No memory for ino translation table");
439 ret_final = ARCHIVE_FATAL;
440 goto exit_write_header;
441 } else if (ino > 077777) {
442 archive_set_error(&a->archive, ERANGE,
443 "Too many files for this cpio format");
444 ret_final = ARCHIVE_FATAL;
445 goto exit_write_header;
446 }
447 h.h_ino = la_swap16((uint16_t)ino);
448
449 h.h_mode = archive_entry_mode(entry);
450 if (((h.h_mode & AE_IFMT) == AE_IFSOCK) || ((h.h_mode & AE_IFMT) == AE_IFIFO)) {
451 archive_set_error(&a->archive, EINVAL,
452 "sockets and fifos cannot be represented in the binary cpio formats");
453 ret_final = ARCHIVE_FATAL;
454 goto exit_write_header;
455 }
456 if (a->archive.archive_format == ARCHIVE_FORMAT_CPIO_PWB) {
457 if ((h.h_mode & AE_IFMT) == AE_IFLNK) {
458 archive_set_error(&a->archive, EINVAL,
459 "symbolic links cannot be represented in the PWB cpio format");
460 ret_final = ARCHIVE_FATAL;
461 goto exit_write_header;
462 }
463 /* we could turn off AE_IFREG here, but it does no harm, */
464 /* and allows v7 cpio to read the entry without confusion */
465 }
466 h.h_mode = la_swap16(h.h_mode);
467
468 h.h_uid = la_swap16((uint16_t)archive_entry_uid(entry));
469 h.h_gid = la_swap16((uint16_t)archive_entry_gid(entry));
470 h.h_nlink = la_swap16((uint16_t)archive_entry_nlink(entry));
471
472 if (archive_entry_filetype(entry) == AE_IFBLK
473 || archive_entry_filetype(entry) == AE_IFCHR)
474 h.h_majmin = la_swap16(archive_entry_rdev(entry));
475 else
476 h.h_majmin = 0;
477
478 h.h_mtime = la_swap32((uint32_t)archive_entry_mtime(entry));
479 if (pathlength > 0xffff) {
480 archive_set_error(&a->archive, ERANGE,
481 "Filename is too long for cpio format");
482 ret_final = ARCHIVE_FAILED;
483 goto exit_write_header;
484 }
485 h.h_namesize = la_swap16((uint16_t)pathlength);
486
487 /* Non-regular files don't store bodies. */
488 if (archive_entry_filetype(entry) != AE_IFREG)
489 archive_entry_set_size(entry, 0);
490
491 /* Symlinks get the link written as the body of the entry. */
492 ret = archive_entry_symlink_l(entry, &p, &len, sconv);
493 if (ret != 0) {
494 if (errno == ENOMEM) {
495 archive_set_error(&a->archive, ENOMEM,
496 "Can't allocate memory for Linkname");
497 ret_final = ARCHIVE_FATAL;
498 goto exit_write_header;
499 }
500 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT,
501 "Can't translate linkname '%s' to %s",
502 archive_entry_symlink(entry),
503 archive_string_conversion_charset_name(sconv));
504 ret_final = ARCHIVE_WARN;
505 }
506
507 if (len > 0 && p != NULL && *p != '\0') {
508 if (a->archive.archive_format == ARCHIVE_FORMAT_CPIO_PWB) {
509 archive_set_error(&a->archive, EINVAL,
510 "symlinks are not supported by UNIX V6 or by PWB cpio");
511 ret_final = ARCHIVE_FATAL;
512 goto exit_write_header;
513 }
514 h.h_filesize = la_swap32((uint32_t)strlen(p)); /* symlink */
515 } else {
516 if ((a->archive.archive_format == ARCHIVE_FORMAT_CPIO_PWB) &&
517 (archive_entry_size(entry) > 256*256*256-1)) {
518 archive_set_error(&a->archive, ERANGE,
519 "File is too large for PWB binary cpio format");
520 ret_final = ARCHIVE_FAILED;
521 goto exit_write_header;
522 } else if (archive_entry_size(entry) > INT32_MAX) {
523 archive_set_error(&a->archive, ERANGE,
524 "File is too large for binary cpio format");
525 ret_final = ARCHIVE_FAILED;
526 goto exit_write_header;
527 }
528 h.h_filesize = la_swap32((uint32_t)archive_entry_size(entry)); /* file */
529 }
530
531 ret = __archive_write_output(a, &h, HSIZE);
532 if (ret != ARCHIVE_OK) {
533 ret_final = ARCHIVE_FATAL;
534 goto exit_write_header;
535 }
536
537 ret = __archive_write_output(a, path, pathlength);
538 if ((ret == ARCHIVE_OK) && ((pathlength % 2) != 0))
539 ret = __archive_write_nulls(a, 1);
540 if (ret != ARCHIVE_OK) {
541 ret_final = ARCHIVE_FATAL;
542 goto exit_write_header;
543 }
544
545 cpio->entry_bytes_remaining = archive_entry_size(entry);
546 if ((cpio->entry_bytes_remaining % 2) != 0)
547 cpio->entry_bytes_remaining++;
548
549 /* Write the symlink now. */
550 if (p != NULL && *p != '\0') {
551 ret = __archive_write_output(a, p, strlen(p));
552 if ((ret == ARCHIVE_OK) && ((strlen(p) % 2) != 0))
553 ret = __archive_write_nulls(a, 1);
554 if (ret != ARCHIVE_OK) {
555 ret_final = ARCHIVE_FATAL;
556 goto exit_write_header;
557 }
558 }
559
560 exit_write_header:
561 archive_entry_free(entry_main);
562 return (ret_final);
563 }
564
565 static ssize_t
archive_write_binary_data(struct archive_write * a,const void * buff,size_t s)566 archive_write_binary_data(struct archive_write *a, const void *buff, size_t s)
567 {
568 struct cpio *cpio = a->format_data;
569 int ret;
570
571 if (s > cpio->entry_bytes_remaining)
572 s = (size_t)cpio->entry_bytes_remaining;
573
574 ret = __archive_write_output(a, buff, s);
575 cpio->entry_bytes_remaining -= s;
576 if (ret >= 0)
577 return (s);
578 else
579 return (ret);
580 }
581
582 static int
archive_write_binary_close(struct archive_write * a)583 archive_write_binary_close(struct archive_write *a)
584 {
585 int er;
586 struct archive_entry *trailer;
587
588 trailer = archive_entry_new2(NULL);
589 if (trailer == NULL) {
590 return ARCHIVE_FATAL;
591 }
592 /* nlink = 1 here for GNU cpio compat. */
593 archive_entry_set_nlink(trailer, 1);
594 archive_entry_set_size(trailer, 0);
595 archive_entry_set_pathname(trailer, "TRAILER!!!");
596 er = write_header(a, trailer);
597 archive_entry_free(trailer);
598 return (er);
599 }
600
601 static int
archive_write_binary_free(struct archive_write * a)602 archive_write_binary_free(struct archive_write *a)
603 {
604 struct cpio *cpio = a->format_data;
605
606 free(cpio->ino_list);
607 free(cpio);
608 a->format_data = NULL;
609 return (ARCHIVE_OK);
610 }
611
612 static int
archive_write_binary_finish_entry(struct archive_write * a)613 archive_write_binary_finish_entry(struct archive_write *a)
614 {
615 struct cpio *cpio = a->format_data;
616
617 return (__archive_write_nulls(a, cpio->entry_bytes_remaining));
618 }
619