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_disk_set_standard_lookup.c 201083 2009-12-28 02:09:57Z kientzle $"); 28 29 #ifdef HAVE_SYS_TYPES_H 30 #include <sys/types.h> 31 #endif 32 #ifdef HAVE_ERRNO_H 33 #include <errno.h> 34 #endif 35 #ifdef HAVE_GRP_H 36 #include <grp.h> 37 #endif 38 #ifdef HAVE_PWD_H 39 #include <pwd.h> 40 #endif 41 #ifdef HAVE_STDLIB_H 42 #include <stdlib.h> 43 #endif 44 #ifdef HAVE_STRING_H 45 #include <string.h> 46 #endif 47 48 #include "archive.h" 49 #include "archive_private.h" 50 #include "archive_read_private.h" 51 #include "archive_write_disk_private.h" 52 53 struct bucket { 54 char *name; 55 int hash; 56 id_t id; 57 }; 58 59 static const size_t cache_size = 127; 60 static unsigned int hash(const char *); 61 static int64_t lookup_gid(void *, const char *uname, int64_t); 62 static int64_t lookup_uid(void *, const char *uname, int64_t); 63 static void cleanup(void *); 64 65 /* 66 * Installs functions that use getpwnam()/getgrnam()---along with 67 * a simple cache to accelerate such lookups---into the archive_write_disk 68 * object. This is in a separate file because getpwnam()/getgrnam() 69 * can pull in a LOT of library code (including NIS/LDAP functions, which 70 * pull in DNS resolvers, etc). This can easily top 500kB, which makes 71 * it inappropriate for some space-constrained applications. 72 * 73 * Applications that are size-sensitive may want to just use the 74 * real default functions (defined in archive_write_disk.c) that just 75 * use the uid/gid without the lookup. Or define your own custom functions 76 * if you prefer. 77 * 78 * TODO: Replace these hash tables with simpler move-to-front LRU 79 * lists with a bounded size (128 items?). The hash is a bit faster, 80 * but has a bad pathology in which it thrashes a single bucket. Even 81 * walking a list of 128 items is a lot faster than calling 82 * getpwnam()! 83 */ 84 int 85 archive_write_disk_set_standard_lookup(struct archive *a) 86 { 87 struct bucket *ucache = malloc(cache_size * sizeof(struct bucket)); 88 struct bucket *gcache = malloc(cache_size * sizeof(struct bucket)); 89 if (ucache == NULL || gcache == NULL) { 90 free(ucache); 91 free(gcache); 92 return (ARCHIVE_FATAL); 93 } 94 memset(ucache, 0, cache_size * sizeof(struct bucket)); 95 memset(gcache, 0, cache_size * sizeof(struct bucket)); 96 archive_write_disk_set_group_lookup(a, gcache, lookup_gid, cleanup); 97 archive_write_disk_set_user_lookup(a, ucache, lookup_uid, cleanup); 98 return (ARCHIVE_OK); 99 } 100 101 static int64_t 102 lookup_gid(void *private_data, const char *gname, int64_t gid) 103 { 104 int h; 105 struct bucket *b; 106 struct bucket *gcache = (struct bucket *)private_data; 107 108 /* If no gname, just use the gid provided. */ 109 if (gname == NULL || *gname == '\0') 110 return (gid); 111 112 /* Try to find gname in the cache. */ 113 h = hash(gname); 114 b = &gcache[h % cache_size ]; 115 if (b->name != NULL && b->hash == h && strcmp(gname, b->name) == 0) 116 return ((gid_t)b->id); 117 118 /* Free the cache slot for a new entry. */ 119 if (b->name != NULL) 120 free(b->name); 121 b->name = strdup(gname); 122 /* Note: If strdup fails, that's okay; we just won't cache. */ 123 b->hash = h; 124 #if HAVE_GRP_H 125 # if HAVE_GETGRNAM_R 126 { 127 char _buffer[128]; 128 size_t bufsize = 128; 129 char *buffer = _buffer; 130 char *allocated = NULL; 131 struct group grent, *result; 132 int r; 133 134 for (;;) { 135 result = &grent; /* Old getgrnam_r ignores last arg. */ 136 r = getgrnam_r(gname, &grent, buffer, bufsize, &result); 137 if (r == 0) 138 break; 139 if (r != ERANGE) 140 break; 141 bufsize *= 2; 142 free(allocated); 143 allocated = malloc(bufsize); 144 if (allocated == NULL) 145 break; 146 buffer = allocated; 147 } 148 if (result != NULL) 149 gid = result->gr_gid; 150 free(allocated); 151 } 152 # else /* HAVE_GETGRNAM_R */ 153 { 154 struct group *result; 155 156 result = getgrnam(gname); 157 if (result != NULL) 158 gid = result->gr_gid; 159 } 160 # endif /* HAVE_GETGRNAM_R */ 161 #elif defined(_WIN32) && !defined(__CYGWIN__) 162 /* TODO: do a gname->gid lookup for Windows. */ 163 #else 164 #error No way to perform gid lookups on this platform 165 #endif 166 b->id = (gid_t)gid; 167 168 return (gid); 169 } 170 171 static int64_t 172 lookup_uid(void *private_data, const char *uname, int64_t uid) 173 { 174 int h; 175 struct bucket *b; 176 struct bucket *ucache = (struct bucket *)private_data; 177 178 /* If no uname, just use the uid provided. */ 179 if (uname == NULL || *uname == '\0') 180 return (uid); 181 182 /* Try to find uname in the cache. */ 183 h = hash(uname); 184 b = &ucache[h % cache_size ]; 185 if (b->name != NULL && b->hash == h && strcmp(uname, b->name) == 0) 186 return ((uid_t)b->id); 187 188 /* Free the cache slot for a new entry. */ 189 if (b->name != NULL) 190 free(b->name); 191 b->name = strdup(uname); 192 /* Note: If strdup fails, that's okay; we just won't cache. */ 193 b->hash = h; 194 #if HAVE_PWD_H 195 # if HAVE_GETPWNAM_R 196 { 197 char _buffer[128]; 198 size_t bufsize = 128; 199 char *buffer = _buffer; 200 char *allocated = NULL; 201 struct passwd pwent, *result; 202 int r; 203 204 for (;;) { 205 result = &pwent; /* Old getpwnam_r ignores last arg. */ 206 r = getpwnam_r(uname, &pwent, buffer, bufsize, &result); 207 if (r == 0) 208 break; 209 if (r != ERANGE) 210 break; 211 bufsize *= 2; 212 free(allocated); 213 allocated = malloc(bufsize); 214 if (allocated == NULL) 215 break; 216 buffer = allocated; 217 } 218 if (result != NULL) 219 uid = result->pw_uid; 220 free(allocated); 221 } 222 # else /* HAVE_GETPWNAM_R */ 223 { 224 struct passwd *result; 225 226 result = getpwnam(uname); 227 if (result != NULL) 228 uid = result->pw_uid; 229 } 230 #endif /* HAVE_GETPWNAM_R */ 231 #elif defined(_WIN32) && !defined(__CYGWIN__) 232 /* TODO: do a uname->uid lookup for Windows. */ 233 #else 234 #error No way to look up uids on this platform 235 #endif 236 b->id = (uid_t)uid; 237 238 return (uid); 239 } 240 241 static void 242 cleanup(void *private) 243 { 244 size_t i; 245 struct bucket *cache = (struct bucket *)private; 246 247 for (i = 0; i < cache_size; i++) 248 free(cache[i].name); 249 free(cache); 250 } 251 252 253 static unsigned int 254 hash(const char *p) 255 { 256 /* A 32-bit version of Peter Weinberger's (PJW) hash algorithm, 257 as used by ELF for hashing function names. */ 258 unsigned g, h = 0; 259 while (*p != '\0') { 260 h = (h << 4) + *p++; 261 if ((g = h & 0xF0000000) != 0) { 262 h ^= g >> 24; 263 h &= 0x0FFFFFFF; 264 } 265 } 266 return h; 267 } 268