archive_write_set_format_cpio.c revision 228753 
1/*-
2 * Copyright (c) 2003-2007 Tim Kientzle
3 * All rights reserved.
4 *
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.
13 *
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.
24 */
25
26#include "archive_platform.h"
27__FBSDID("$FreeBSD: head/lib/libarchive/archive_write_set_format_cpio.c 201170 2009-12-29 06:34:23Z kientzle $");
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_private.h"
43#include "archive_write_private.h"
44
45static ssize_t	archive_write_cpio_data(struct archive_write *,
46		    const void *buff, size_t s);
47static int	archive_write_cpio_finish(struct archive_write *);
48static int	archive_write_cpio_destroy(struct archive_write *);
49static int	archive_write_cpio_finish_entry(struct archive_write *);
50static int	archive_write_cpio_header(struct archive_write *,
51		    struct archive_entry *);
52static int	format_octal(int64_t, void *, int);
53static int64_t	format_octal_recursive(int64_t, char *, int);
54
55struct cpio {
56	uint64_t	  entry_bytes_remaining;
57
58	int64_t		  ino_next;
59
60	struct		 { int64_t old; int new;} *ino_list;
61	size_t		  ino_list_size;
62	size_t		  ino_list_next;
63};
64
65struct cpio_header {
66	char	c_magic[6];
67	char	c_dev[6];
68	char	c_ino[6];
69	char	c_mode[6];
70	char	c_uid[6];
71	char	c_gid[6];
72	char	c_nlink[6];
73	char	c_rdev[6];
74	char	c_mtime[11];
75	char	c_namesize[6];
76	char	c_filesize[11];
77};
78
79/*
80 * Set output format to 'cpio' format.
81 */
82int
83archive_write_set_format_cpio(struct archive *_a)
84{
85	struct archive_write *a = (struct archive_write *)_a;
86	struct cpio *cpio;
87
88	/* If someone else was already registered, unregister them. */
89	if (a->format_destroy != NULL)
90		(a->format_destroy)(a);
91
92	cpio = (struct cpio *)malloc(sizeof(*cpio));
93	if (cpio == NULL) {
94		archive_set_error(&a->archive, ENOMEM, "Can't allocate cpio data");
95		return (ARCHIVE_FATAL);
96	}
97	memset(cpio, 0, sizeof(*cpio));
98	a->format_data = cpio;
99
100	a->pad_uncompressed = 1;
101	a->format_name = "cpio";
102	a->format_write_header = archive_write_cpio_header;
103	a->format_write_data = archive_write_cpio_data;
104	a->format_finish_entry = archive_write_cpio_finish_entry;
105	a->format_finish = archive_write_cpio_finish;
106	a->format_destroy = archive_write_cpio_destroy;
107	a->archive.archive_format = ARCHIVE_FORMAT_CPIO_POSIX;
108	a->archive.archive_format_name = "POSIX cpio";
109	return (ARCHIVE_OK);
110}
111
112/*
113 * Ino values are as long as 64 bits on some systems; cpio format
114 * only allows 18 bits and relies on the ino values to identify hardlinked
115 * files.  So, we can't merely "hash" the ino numbers since collisions
116 * would corrupt the archive.  Instead, we generate synthetic ino values
117 * to store in the archive and maintain a map of original ino values to
118 * synthetic ones so we can preserve hardlink information.
119 *
120 * TODO: Make this more efficient.  It's not as bad as it looks (most
121 * files don't have any hardlinks and we don't do any work here for those),
122 * but it wouldn't be hard to do better.
123 *
124 * TODO: Work with dev/ino pairs here instead of just ino values.
125 */
126static int
127synthesize_ino_value(struct cpio *cpio, struct archive_entry *entry)
128{
129	int64_t ino = archive_entry_ino64(entry);
130	int ino_new;
131	size_t i;
132
133	/*
134	 * If no index number was given, don't assign one.  In
135	 * particular, this handles the end-of-archive marker
136	 * correctly by giving it a zero index value.  (This is also
137	 * why we start our synthetic index numbers with one below.)
138	 */
139	if (ino == 0)
140		return (0);
141
142	/* Don't store a mapping if we don't need to. */
143	if (archive_entry_nlink(entry) < 2) {
144		return ++cpio->ino_next;
145	}
146
147	/* Look up old ino; if we have it, this is a hardlink
148	 * and we reuse the same value. */
149	for (i = 0; i < cpio->ino_list_next; ++i) {
150		if (cpio->ino_list[i].old == ino)
151			return (cpio->ino_list[i].new);
152	}
153
154	/* Assign a new index number. */
155	ino_new = ++cpio->ino_next;
156
157	/* Ensure space for the new mapping. */
158	if (cpio->ino_list_size <= cpio->ino_list_next) {
159		size_t newsize = cpio->ino_list_size < 512
160		    ? 512 : cpio->ino_list_size * 2;
161		void *newlist = realloc(cpio->ino_list,
162		    sizeof(cpio->ino_list[0]) * newsize);
163		if (newlist == NULL)
164			return (-1);
165
166		cpio->ino_list_size = newsize;
167		cpio->ino_list = newlist;
168	}
169
170	/* Record and return the new value. */
171	cpio->ino_list[cpio->ino_list_next].old = ino;
172	cpio->ino_list[cpio->ino_list_next].new = ino_new;
173	++cpio->ino_list_next;
174	return (ino_new);
175}
176
177static int
178archive_write_cpio_header(struct archive_write *a, struct archive_entry *entry)
179{
180	struct cpio *cpio;
181	const char *p, *path;
182	int pathlength, ret, ret2;
183	int64_t	ino;
184	struct cpio_header	 h;
185
186	cpio = (struct cpio *)a->format_data;
187	ret2 = ARCHIVE_OK;
188
189	path = archive_entry_pathname(entry);
190	pathlength = (int)strlen(path) + 1; /* Include trailing null. */
191
192	memset(&h, 0, sizeof(h));
193	format_octal(070707, &h.c_magic, sizeof(h.c_magic));
194	format_octal(archive_entry_dev(entry), &h.c_dev, sizeof(h.c_dev));
195
196	ino = synthesize_ino_value(cpio, entry);
197	if (ino < 0) {
198		archive_set_error(&a->archive, ENOMEM,
199		    "No memory for ino translation table");
200		return (ARCHIVE_FATAL);
201	} else if (ino > 0777777) {
202		archive_set_error(&a->archive, ERANGE,
203		    "Too many files for this cpio format");
204		return (ARCHIVE_FATAL);
205	}
206	format_octal(ino & 0777777, &h.c_ino, sizeof(h.c_ino));
207
208	format_octal(archive_entry_mode(entry), &h.c_mode, sizeof(h.c_mode));
209	format_octal(archive_entry_uid(entry), &h.c_uid, sizeof(h.c_uid));
210	format_octal(archive_entry_gid(entry), &h.c_gid, sizeof(h.c_gid));
211	format_octal(archive_entry_nlink(entry), &h.c_nlink, sizeof(h.c_nlink));
212	if (archive_entry_filetype(entry) == AE_IFBLK
213	    || archive_entry_filetype(entry) == AE_IFCHR)
214	    format_octal(archive_entry_dev(entry), &h.c_rdev, sizeof(h.c_rdev));
215	else
216	    format_octal(0, &h.c_rdev, sizeof(h.c_rdev));
217	format_octal(archive_entry_mtime(entry), &h.c_mtime, sizeof(h.c_mtime));
218	format_octal(pathlength, &h.c_namesize, sizeof(h.c_namesize));
219
220	/* Non-regular files don't store bodies. */
221	if (archive_entry_filetype(entry) != AE_IFREG)
222		archive_entry_set_size(entry, 0);
223
224	/* Symlinks get the link written as the body of the entry. */
225	p = archive_entry_symlink(entry);
226	if (p != NULL  &&  *p != '\0')
227		format_octal(strlen(p), &h.c_filesize, sizeof(h.c_filesize));
228	else
229		format_octal(archive_entry_size(entry),
230		    &h.c_filesize, sizeof(h.c_filesize));
231
232	ret = (a->compressor.write)(a, &h, sizeof(h));
233	if (ret != ARCHIVE_OK)
234		return (ARCHIVE_FATAL);
235
236	ret = (a->compressor.write)(a, path, pathlength);
237	if (ret != ARCHIVE_OK)
238		return (ARCHIVE_FATAL);
239
240	cpio->entry_bytes_remaining = archive_entry_size(entry);
241
242	/* Write the symlink now. */
243	if (p != NULL  &&  *p != '\0')
244		ret = (a->compressor.write)(a, p, strlen(p));
245
246	if (ret == ARCHIVE_OK)
247		ret = ret2;
248	return (ret);
249}
250
251static ssize_t
252archive_write_cpio_data(struct archive_write *a, const void *buff, size_t s)
253{
254	struct cpio *cpio;
255	int ret;
256
257	cpio = (struct cpio *)a->format_data;
258	if (s > cpio->entry_bytes_remaining)
259		s = cpio->entry_bytes_remaining;
260
261	ret = (a->compressor.write)(a, buff, s);
262	cpio->entry_bytes_remaining -= s;
263	if (ret >= 0)
264		return (s);
265	else
266		return (ret);
267}
268
269/*
270 * Format a number into the specified field.
271 */
272static int
273format_octal(int64_t v, void *p, int digits)
274{
275	int64_t	max;
276	int	ret;
277
278	max = (((int64_t)1) << (digits * 3)) - 1;
279	if (v >= 0  &&  v <= max) {
280	    format_octal_recursive(v, (char *)p, digits);
281	    ret = 0;
282	} else {
283	    format_octal_recursive(max, (char *)p, digits);
284	    ret = -1;
285	}
286	return (ret);
287}
288
289static int64_t
290format_octal_recursive(int64_t v, char *p, int s)
291{
292	if (s == 0)
293		return (v);
294	v = format_octal_recursive(v, p+1, s-1);
295	*p = '0' + (v & 7);
296	return (v >> 3);
297}
298
299static int
300archive_write_cpio_finish(struct archive_write *a)
301{
302	int er;
303	struct archive_entry *trailer;
304
305	trailer = archive_entry_new();
306	/* nlink = 1 here for GNU cpio compat. */
307	archive_entry_set_nlink(trailer, 1);
308	archive_entry_set_pathname(trailer, "TRAILER!!!");
309	er = archive_write_cpio_header(a, trailer);
310	archive_entry_free(trailer);
311	return (er);
312}
313
314static int
315archive_write_cpio_destroy(struct archive_write *a)
316{
317	struct cpio *cpio;
318
319	cpio = (struct cpio *)a->format_data;
320	free(cpio->ino_list);
321	free(cpio);
322	a->format_data = NULL;
323	return (ARCHIVE_OK);
324}
325
326static int
327archive_write_cpio_finish_entry(struct archive_write *a)
328{
329	struct cpio *cpio;
330	size_t to_write;
331	int ret;
332
333	cpio = (struct cpio *)a->format_data;
334	ret = ARCHIVE_OK;
335	while (cpio->entry_bytes_remaining > 0) {
336		to_write = cpio->entry_bytes_remaining < a->null_length ?
337		    cpio->entry_bytes_remaining : a->null_length;
338		ret = (a->compressor.write)(a, a->nulls, to_write);
339		if (ret != ARCHIVE_OK)
340			return (ret);
341		cpio->entry_bytes_remaining -= to_write;
342	}
343	return (ret);
344}
345