xref: /freebsd/contrib/libarchive/libarchive/archive_write_set_format_cpio_odc.c (revision 185becb1e1bd2657c156f78aeb52edac05ba5fb5)
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_odc_data(struct archive_write *,
49 		    const void *buff, size_t s);
50 static int	archive_write_odc_close(struct archive_write *);
51 static int	archive_write_odc_free(struct archive_write *);
52 static int	archive_write_odc_finish_entry(struct archive_write *);
53 static int	archive_write_odc_header(struct archive_write *,
54 		    struct archive_entry *);
55 static int	archive_write_odc_options(struct archive_write *,
56 		    const char *, const char *);
57 static int	format_octal(int64_t, void *, int);
58 static int64_t	format_octal_recursive(int64_t, char *, int);
59 static int	write_header(struct archive_write *, struct archive_entry *);
60 
61 struct cpio {
62 	uint64_t	  entry_bytes_remaining;
63 
64 	int64_t		  ino_next;
65 
66 	struct		 { int64_t old; int new;} *ino_list;
67 	size_t		  ino_list_size;
68 	size_t		  ino_list_next;
69 
70 	struct archive_string_conv *opt_sconv;
71 	struct archive_string_conv *sconv_default;
72 	int		  init_default_conversion;
73 };
74 
75 #define	c_magic_offset 0
76 #define	c_magic_size 6
77 #define	c_dev_offset 6
78 #define	c_dev_size 6
79 #define	c_ino_offset 12
80 #define	c_ino_size 6
81 #define	c_mode_offset 18
82 #define	c_mode_size 6
83 #define	c_uid_offset 24
84 #define	c_uid_size 6
85 #define	c_gid_offset 30
86 #define	c_gid_size 6
87 #define	c_nlink_offset 36
88 #define	c_nlink_size 6
89 #define	c_rdev_offset 42
90 #define	c_rdev_size 6
91 #define	c_mtime_offset 48
92 #define	c_mtime_size 11
93 #define	c_namesize_offset 59
94 #define	c_namesize_size 6
95 #define	c_filesize_offset 65
96 #define	c_filesize_size 11
97 
98 /*
99  * Set output format to 'cpio' format.
100  */
101 int
archive_write_set_format_cpio_odc(struct archive * _a)102 archive_write_set_format_cpio_odc(struct archive *_a)
103 {
104 	struct archive_write *a = (struct archive_write *)_a;
105 	struct cpio *cpio;
106 
107 	archive_check_magic(_a, ARCHIVE_WRITE_MAGIC,
108 	    ARCHIVE_STATE_NEW, "archive_write_set_format_cpio_odc");
109 
110 	/* If someone else was already registered, unregister them. */
111 	(void)__archive_write_unregister_format(a);
112 
113 	cpio = calloc(1, sizeof(*cpio));
114 	if (cpio == NULL) {
115 		archive_set_error(&a->archive, ENOMEM, "Can't allocate cpio data");
116 		return (ARCHIVE_FATAL);
117 	}
118 	a->format_data = cpio;
119 	a->format_name = "cpio";
120 	a->format_options = archive_write_odc_options;
121 	a->format_write_header = archive_write_odc_header;
122 	a->format_write_data = archive_write_odc_data;
123 	a->format_finish_entry = archive_write_odc_finish_entry;
124 	a->format_close = archive_write_odc_close;
125 	a->format_free = archive_write_odc_free;
126 	a->archive.archive_format = ARCHIVE_FORMAT_CPIO_POSIX;
127 	a->archive.archive_format_name = "POSIX cpio";
128 	return (ARCHIVE_OK);
129 }
130 
131 static int
archive_write_odc_options(struct archive_write * a,const char * key,const char * val)132 archive_write_odc_options(struct archive_write *a, const char *key,
133     const char *val)
134 {
135 	struct cpio *cpio = a->format_data;
136 	int ret = ARCHIVE_FAILED;
137 
138 	if (strcmp(key, "hdrcharset")  == 0) {
139 		if (val == NULL || val[0] == 0)
140 			archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC,
141 			    "%s: hdrcharset option needs a character-set name",
142 			    a->format_name);
143 		else {
144 			cpio->opt_sconv = archive_string_conversion_to_charset(
145 			    &a->archive, val, 0);
146 			if (cpio->opt_sconv != NULL)
147 				ret = ARCHIVE_OK;
148 			else
149 				ret = ARCHIVE_FATAL;
150 		}
151 		return (ret);
152 	}
153 
154 	/* Note: The "warn" return is just to inform the options
155 	 * supervisor that we didn't handle it.  It will generate
156 	 * a suitable error if no one used this option. */
157 	return (ARCHIVE_WARN);
158 }
159 
160 /*
161  * Ino values are as long as 64 bits on some systems; cpio format
162  * only allows 18 bits and relies on the ino values to identify hardlinked
163  * files.  So, we can't merely "hash" the ino numbers since collisions
164  * would corrupt the archive.  Instead, we generate synthetic ino values
165  * to store in the archive and maintain a map of original ino values to
166  * synthetic ones so we can preserve hardlink information.
167  *
168  * TODO: Make this more efficient.  It's not as bad as it looks (most
169  * files don't have any hardlinks and we don't do any work here for those),
170  * but it wouldn't be hard to do better.
171  *
172  * TODO: Work with dev/ino pairs here instead of just ino values.
173  */
174 static int
synthesize_ino_value(struct cpio * cpio,struct archive_entry * entry)175 synthesize_ino_value(struct cpio *cpio, struct archive_entry *entry)
176 {
177 	int64_t ino = archive_entry_ino64(entry);
178 	int ino_new;
179 	size_t i;
180 
181 	/*
182 	 * If no index number was given, don't assign one.  In
183 	 * particular, this handles the end-of-archive marker
184 	 * correctly by giving it a zero index value.  (This is also
185 	 * why we start our synthetic index numbers with one below.)
186 	 */
187 	if (ino == 0)
188 		return (0);
189 
190 	/* Don't store a mapping if we don't need to. */
191 	if (archive_entry_nlink(entry) < 2) {
192 		return (int)(++cpio->ino_next);
193 	}
194 
195 	/* Look up old ino; if we have it, this is a hardlink
196 	 * and we reuse the same value. */
197 	for (i = 0; i < cpio->ino_list_next; ++i) {
198 		if (cpio->ino_list[i].old == ino)
199 			return (cpio->ino_list[i].new);
200 	}
201 
202 	/* Assign a new index number. */
203 	ino_new = (int)(++cpio->ino_next);
204 
205 	/* Ensure space for the new mapping. */
206 	if (cpio->ino_list_size <= cpio->ino_list_next) {
207 		size_t newsize, size;
208 		if (cpio->ino_list_size < 512)
209 			newsize = 512;
210 		else if (archive_ckd_mul_size(&newsize,
211 		    cpio->ino_list_size, 2))
212 			return (-1);
213 		if (archive_ckd_mul_size(&size,
214 		    newsize, sizeof(cpio->ino_list[0])))
215 			return (-1);
216 		void *newlist = realloc(cpio->ino_list, size);
217 		if (newlist == NULL)
218 			return (-1);
219 
220 		cpio->ino_list_size = newsize;
221 		cpio->ino_list = newlist;
222 	}
223 
224 	/* Record and return the new value. */
225 	cpio->ino_list[cpio->ino_list_next].old = ino;
226 	cpio->ino_list[cpio->ino_list_next].new = ino_new;
227 	++cpio->ino_list_next;
228 	return (ino_new);
229 }
230 
231 
232 static struct archive_string_conv *
get_sconv(struct archive_write * a)233 get_sconv(struct archive_write *a)
234 {
235 	struct cpio *cpio = a->format_data;
236 	struct archive_string_conv *sconv;
237 
238 	sconv = cpio->opt_sconv;
239 	if (sconv == NULL) {
240 		if (!cpio->init_default_conversion) {
241 			cpio->sconv_default =
242 			    archive_string_default_conversion_for_write(
243 			      &(a->archive));
244 			cpio->init_default_conversion = 1;
245 		}
246 		sconv = cpio->sconv_default;
247 	}
248 	return (sconv);
249 }
250 
251 static int
archive_write_odc_header(struct archive_write * a,struct archive_entry * entry)252 archive_write_odc_header(struct archive_write *a, struct archive_entry *entry)
253 {
254 	const char *path;
255 	size_t len;
256 
257 	if (archive_entry_filetype(entry) == 0 && archive_entry_hardlink(entry) == NULL) {
258 		archive_set_error(&a->archive, -1, "Filetype required");
259 		return (ARCHIVE_FAILED);
260 	}
261 
262 	if (archive_entry_pathname_l(entry, &path, &len, get_sconv(a)) != 0
263 	    && errno == ENOMEM) {
264 		archive_set_error(&a->archive, ENOMEM,
265 		    "Can't allocate memory for Pathname");
266 		return (ARCHIVE_FATAL);
267 	}
268 	if (len == 0 || path == NULL || path[0] == '\0') {
269 		archive_set_error(&a->archive, -1, "Pathname required");
270 		return (ARCHIVE_FAILED);
271 	}
272 
273 	if (!archive_entry_size_is_set(entry) || archive_entry_size(entry) < 0) {
274 		archive_set_error(&a->archive, -1, "Size required");
275 		return (ARCHIVE_FAILED);
276 	}
277 	return write_header(a, entry);
278 }
279 
280 static int
write_header(struct archive_write * a,struct archive_entry * entry)281 write_header(struct archive_write *a, struct archive_entry *entry)
282 {
283 	struct cpio *cpio = a->format_data;
284 	const char *p, *path;
285 	int ret, ret_final;
286 	int64_t	ino;
287 	char h[76];
288 	struct archive_string_conv *sconv;
289 	struct archive_entry *entry_main;
290 	size_t len, pathlength;
291 
292 	ret_final = ARCHIVE_OK;
293 	sconv = get_sconv(a);
294 
295 #if defined(_WIN32) && !defined(__CYGWIN__)
296 	/* Make sure the path separators in pathname, hardlink and symlink
297 	 * are all slash '/', not the Windows path separator '\'. */
298 	entry_main = __la_win_entry_in_posix_pathseparator(entry);
299 	if (entry_main == NULL) {
300 		archive_set_error(&a->archive, ENOMEM,
301 		    "Can't allocate ustar data");
302 		return(ARCHIVE_FATAL);
303 	}
304 	if (entry != entry_main)
305 		entry = entry_main;
306 	else
307 		entry_main = NULL;
308 #else
309 	entry_main = NULL;
310 #endif
311 
312 	ret = archive_entry_pathname_l(entry, &path, &len, sconv);
313 	if (ret != 0) {
314 		if (errno == ENOMEM) {
315 			archive_set_error(&a->archive, ENOMEM,
316 			    "Can't allocate memory for Pathname");
317 			ret_final = ARCHIVE_FATAL;
318 			goto exit_write_header;
319 		}
320 		archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT,
321 		    "Can't translate pathname '%s' to %s",
322 		    archive_entry_pathname(entry),
323 		    archive_string_conversion_charset_name(sconv));
324 		ret_final = ARCHIVE_WARN;
325 	}
326 	/* Include trailing null. */
327 	pathlength = len + 1;
328 
329 	memset(h, 0, sizeof(h));
330 	format_octal(070707, h + c_magic_offset, c_magic_size);
331 	format_octal(archive_entry_dev(entry), h + c_dev_offset, c_dev_size);
332 
333 	ino = synthesize_ino_value(cpio, entry);
334 	if (ino < 0) {
335 		archive_set_error(&a->archive, ENOMEM,
336 		    "No memory for ino translation table");
337 		ret_final = ARCHIVE_FATAL;
338 		goto exit_write_header;
339 	} else if (ino > 0777777) {
340 		archive_set_error(&a->archive, ERANGE,
341 		    "Too many files for this cpio format");
342 		ret_final = ARCHIVE_FATAL;
343 		goto exit_write_header;
344 	}
345 	format_octal(ino & 0777777, h + c_ino_offset, c_ino_size);
346 
347 	/* TODO: Set ret_final to ARCHIVE_WARN if any of these overflow. */
348 	format_octal(archive_entry_mode(entry), h + c_mode_offset, c_mode_size);
349 	format_octal(archive_entry_uid(entry), h + c_uid_offset, c_uid_size);
350 	format_octal(archive_entry_gid(entry), h + c_gid_offset, c_gid_size);
351 	format_octal(archive_entry_nlink(entry), h + c_nlink_offset, c_nlink_size);
352 	if (archive_entry_filetype(entry) == AE_IFBLK
353 	    || archive_entry_filetype(entry) == AE_IFCHR)
354 	    format_octal(archive_entry_rdev(entry), h + c_rdev_offset, c_rdev_size);
355 	else
356 	    format_octal(0, h + c_rdev_offset, c_rdev_size);
357 	format_octal(archive_entry_mtime(entry), h + c_mtime_offset, c_mtime_size);
358 	if (format_octal((int64_t)pathlength, h + c_namesize_offset, c_namesize_size)) {
359 		archive_set_error(&a->archive, ERANGE,
360 		    "Filename is too long for cpio format");
361 		ret_final = ARCHIVE_FAILED;
362 		goto exit_write_header;
363 	}
364 
365 	/* Non-regular files don't store bodies. */
366 	if (archive_entry_filetype(entry) != AE_IFREG)
367 		archive_entry_set_size(entry, 0);
368 
369 	/* Symlinks get the link written as the body of the entry. */
370 	ret = archive_entry_symlink_l(entry, &p, &len, sconv);
371 	if (ret != 0) {
372 		if (errno == ENOMEM) {
373 			archive_set_error(&a->archive, ENOMEM,
374 			    "Can't allocate memory for Linkname");
375 			ret_final = ARCHIVE_FATAL;
376 			goto exit_write_header;
377 		}
378 		archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT,
379 		    "Can't translate linkname '%s' to %s",
380 		    archive_entry_symlink(entry),
381 		    archive_string_conversion_charset_name(sconv));
382 		ret_final = ARCHIVE_WARN;
383 	}
384 	if (len > 0 && p != NULL  &&  *p != '\0')
385 		ret = format_octal(strlen(p), h + c_filesize_offset,
386 		    c_filesize_size);
387 	else
388 		ret = format_octal(archive_entry_size(entry),
389 		    h + c_filesize_offset, c_filesize_size);
390 	if (ret) {
391 		archive_set_error(&a->archive, ERANGE,
392 		    "File is too large for cpio format");
393 		ret_final = ARCHIVE_FAILED;
394 		goto exit_write_header;
395 	}
396 
397 	ret = __archive_write_output(a, h, sizeof(h));
398 	if (ret != ARCHIVE_OK) {
399 		ret_final = ARCHIVE_FATAL;
400 		goto exit_write_header;
401 	}
402 
403 	ret = __archive_write_output(a, path, pathlength);
404 	if (ret != ARCHIVE_OK) {
405 		ret_final = ARCHIVE_FATAL;
406 		goto exit_write_header;
407 	}
408 
409 	cpio->entry_bytes_remaining = archive_entry_size(entry);
410 
411 	/* Write the symlink now. */
412 	if (p != NULL  &&  *p != '\0') {
413 		ret = __archive_write_output(a, p, strlen(p));
414 		if (ret != ARCHIVE_OK) {
415 			ret_final = ARCHIVE_FATAL;
416 			goto exit_write_header;
417 		}
418 	}
419 exit_write_header:
420 	archive_entry_free(entry_main);
421 	return (ret_final);
422 }
423 
424 static ssize_t
archive_write_odc_data(struct archive_write * a,const void * buff,size_t s)425 archive_write_odc_data(struct archive_write *a, const void *buff, size_t s)
426 {
427 	struct cpio *cpio = a->format_data;
428 	int ret;
429 
430 	if (s > cpio->entry_bytes_remaining)
431 		s = (size_t)cpio->entry_bytes_remaining;
432 
433 	ret = __archive_write_output(a, buff, s);
434 	cpio->entry_bytes_remaining -= s;
435 	if (ret >= 0)
436 		return (s);
437 	else
438 		return (ret);
439 }
440 
441 /*
442  * Format a number into the specified field.
443  */
444 static int
format_octal(int64_t v,void * p,int digits)445 format_octal(int64_t v, void *p, int digits)
446 {
447 	int64_t	max;
448 	int	ret;
449 
450 	max = (((int64_t)1) << (digits * 3)) - 1;
451 	if (v >= 0  &&  v <= max) {
452 	    format_octal_recursive(v, (char *)p, digits);
453 	    ret = 0;
454 	} else {
455 	    format_octal_recursive(max, (char *)p, digits);
456 	    ret = -1;
457 	}
458 	return (ret);
459 }
460 
461 static int64_t
format_octal_recursive(int64_t v,char * p,int s)462 format_octal_recursive(int64_t v, char *p, int s)
463 {
464 	if (s == 0)
465 		return (v);
466 	v = format_octal_recursive(v, p+1, s-1);
467 	*p = '0' + ((char)v & 7);
468 	return (v >> 3);
469 }
470 
471 static int
archive_write_odc_close(struct archive_write * a)472 archive_write_odc_close(struct archive_write *a)
473 {
474 	int er;
475 	struct archive_entry *trailer;
476 
477 	trailer = archive_entry_new2(NULL);
478 	if (trailer == NULL) {
479 		return ARCHIVE_FATAL;
480 	}
481 	/* nlink = 1 here for GNU cpio compat. */
482 	archive_entry_set_nlink(trailer, 1);
483 	archive_entry_set_size(trailer, 0);
484 	archive_entry_set_pathname(trailer, "TRAILER!!!");
485 	er = write_header(a, trailer);
486 	archive_entry_free(trailer);
487 	return (er);
488 }
489 
490 static int
archive_write_odc_free(struct archive_write * a)491 archive_write_odc_free(struct archive_write *a)
492 {
493 	struct cpio *cpio = a->format_data;
494 
495 	free(cpio->ino_list);
496 	free(cpio);
497 	a->format_data = NULL;
498 	return (ARCHIVE_OK);
499 }
500 
501 static int
archive_write_odc_finish_entry(struct archive_write * a)502 archive_write_odc_finish_entry(struct archive_write *a)
503 {
504 	struct cpio *cpio = a->format_data;
505 
506 	return (__archive_write_nulls(a,
507 	    cpio->entry_bytes_remaining));
508 }
509