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
28#ifdef HAVE_SYS_TYPES_H
29#include <sys/types.h>
30#endif
31#ifdef HAVE_ERRNO_H
32#include <errno.h>
33#endif
34#ifdef HAVE_GRP_H
35#include <grp.h>
36#endif
37#ifdef HAVE_PWD_H
38#include <pwd.h>
39#endif
40#ifdef HAVE_STDLIB_H
41#include <stdlib.h>
42#endif
43#ifdef HAVE_STRING_H
44#include <string.h>
45#endif
46
47#include "archive.h"
48#include "archive_private.h"
49#include "archive_read_private.h"
50#include "archive_write_disk_private.h"
51
52struct bucket {
53	char	*name;
54	int	 hash;
55	id_t	 id;
56};
57
58static const size_t cache_size = 127;
59static unsigned int	hash(const char *);
60static int64_t	lookup_gid(void *, const char *uname, int64_t);
61static int64_t	lookup_uid(void *, const char *uname, int64_t);
62static void	cleanup(void *);
63
64/*
65 * Installs functions that use getpwnam()/getgrnam()---along with
66 * a simple cache to accelerate such lookups---into the archive_write_disk
67 * object.  This is in a separate file because getpwnam()/getgrnam()
68 * can pull in a LOT of library code (including NIS/LDAP functions, which
69 * pull in DNS resolvers, etc).  This can easily top 500kB, which makes
70 * it inappropriate for some space-constrained applications.
71 *
72 * Applications that are size-sensitive may want to just use the
73 * real default functions (defined in archive_write_disk.c) that just
74 * use the uid/gid without the lookup.  Or define your own custom functions
75 * if you prefer.
76 *
77 * TODO: Replace these hash tables with simpler move-to-front LRU
78 * lists with a bounded size (128 items?).  The hash is a bit faster,
79 * but has a bad pathology in which it thrashes a single bucket.  Even
80 * walking a list of 128 items is a lot faster than calling
81 * getpwnam()!
82 */
83int
84archive_write_disk_set_standard_lookup(struct archive *a)
85{
86	struct bucket *ucache = calloc(cache_size, sizeof(struct bucket));
87	struct bucket *gcache = calloc(cache_size, sizeof(struct bucket));
88	if (ucache == NULL || gcache == NULL) {
89		free(ucache);
90		free(gcache);
91		return (ARCHIVE_FATAL);
92	}
93	archive_write_disk_set_group_lookup(a, gcache, lookup_gid, cleanup);
94	archive_write_disk_set_user_lookup(a, ucache, lookup_uid, cleanup);
95	return (ARCHIVE_OK);
96}
97
98static int64_t
99lookup_gid(void *private_data, const char *gname, int64_t gid)
100{
101	int h;
102	struct bucket *b;
103	struct bucket *gcache = (struct bucket *)private_data;
104
105	/* If no gname, just use the gid provided. */
106	if (gname == NULL || *gname == '\0')
107		return (gid);
108
109	/* Try to find gname in the cache. */
110	h = hash(gname);
111	b = &gcache[h % cache_size ];
112	if (b->name != NULL && b->hash == h && strcmp(gname, b->name) == 0)
113		return ((gid_t)b->id);
114
115	/* Free the cache slot for a new entry. */
116	free(b->name);
117	b->name = strdup(gname);
118	/* Note: If strdup fails, that's okay; we just won't cache. */
119	b->hash = h;
120#if HAVE_GRP_H
121#  if HAVE_GETGRNAM_R
122	{
123		char _buffer[128];
124		size_t bufsize = 128;
125		char *buffer = _buffer;
126		char *allocated = NULL;
127		struct group	grent, *result;
128		int r;
129
130		for (;;) {
131			result = &grent; /* Old getgrnam_r ignores last arg. */
132			r = getgrnam_r(gname, &grent, buffer, bufsize, &result);
133			if (r == 0)
134				break;
135			if (r != ERANGE)
136				break;
137			bufsize *= 2;
138			free(allocated);
139			allocated = malloc(bufsize);
140			if (allocated == NULL)
141				break;
142			buffer = allocated;
143		}
144		if (result != NULL)
145			gid = result->gr_gid;
146		free(allocated);
147	}
148#  else /* HAVE_GETGRNAM_R */
149	{
150		struct group *result;
151
152		result = getgrnam(gname);
153		if (result != NULL)
154			gid = result->gr_gid;
155	}
156#  endif /* HAVE_GETGRNAM_R */
157#elif defined(_WIN32) && !defined(__CYGWIN__)
158	/* TODO: do a gname->gid lookup for Windows. */
159#else
160	#error No way to perform gid lookups on this platform
161#endif
162	b->id = (gid_t)gid;
163
164	return (gid);
165}
166
167static int64_t
168lookup_uid(void *private_data, const char *uname, int64_t uid)
169{
170	int h;
171	struct bucket *b;
172	struct bucket *ucache = (struct bucket *)private_data;
173
174	/* If no uname, just use the uid provided. */
175	if (uname == NULL || *uname == '\0')
176		return (uid);
177
178	/* Try to find uname in the cache. */
179	h = hash(uname);
180	b = &ucache[h % cache_size ];
181	if (b->name != NULL && b->hash == h && strcmp(uname, b->name) == 0)
182		return ((uid_t)b->id);
183
184	/* Free the cache slot for a new entry. */
185	free(b->name);
186	b->name = strdup(uname);
187	/* Note: If strdup fails, that's okay; we just won't cache. */
188	b->hash = h;
189#if HAVE_PWD_H
190#  if HAVE_GETPWNAM_R
191	{
192		char _buffer[128];
193		size_t bufsize = 128;
194		char *buffer = _buffer;
195		char *allocated = NULL;
196		struct passwd	pwent, *result;
197		int r;
198
199		for (;;) {
200			result = &pwent; /* Old getpwnam_r ignores last arg. */
201			r = getpwnam_r(uname, &pwent, buffer, bufsize, &result);
202			if (r == 0)
203				break;
204			if (r != ERANGE)
205				break;
206			bufsize *= 2;
207			free(allocated);
208			allocated = malloc(bufsize);
209			if (allocated == NULL)
210				break;
211			buffer = allocated;
212		}
213		if (result != NULL)
214			uid = result->pw_uid;
215		free(allocated);
216	}
217#  else /* HAVE_GETPWNAM_R */
218	{
219		struct passwd *result;
220
221		result = getpwnam(uname);
222		if (result != NULL)
223			uid = result->pw_uid;
224	}
225#endif	/* HAVE_GETPWNAM_R */
226#elif defined(_WIN32) && !defined(__CYGWIN__)
227	/* TODO: do a uname->uid lookup for Windows. */
228#else
229	#error No way to look up uids on this platform
230#endif
231	b->id = (uid_t)uid;
232
233	return (uid);
234}
235
236static void
237cleanup(void *private)
238{
239	size_t i;
240	struct bucket *cache = (struct bucket *)private;
241
242	for (i = 0; i < cache_size; i++)
243		free(cache[i].name);
244	free(cache);
245}
246
247
248static unsigned int
249hash(const char *p)
250{
251	/* A 32-bit version of Peter Weinberger's (PJW) hash algorithm,
252	   as used by ELF for hashing function names. */
253	unsigned g, h = 0;
254	while (*p != '\0') {
255		h = (h << 4) + *p++;
256		if ((g = h & 0xF0000000) != 0) {
257			h ^= g >> 24;
258			h &= 0x0FFFFFFF;
259		}
260	}
261	return h;
262}
263