lh | 9ed821d | 2023-04-07 01:36:19 -0700 | [diff] [blame^] | 1 | /* Cache handling for iconv modules. |
| 2 | Copyright (C) 2001-2015 Free Software Foundation, Inc. |
| 3 | This file is part of the GNU C Library. |
| 4 | Contributed by Ulrich Drepper <drepper@cygnus.com>, 2001. |
| 5 | |
| 6 | The GNU C Library is free software; you can redistribute it and/or |
| 7 | modify it under the terms of the GNU Lesser General Public |
| 8 | License as published by the Free Software Foundation; either |
| 9 | version 2.1 of the License, or (at your option) any later version. |
| 10 | |
| 11 | The GNU C Library is distributed in the hope that it will be useful, |
| 12 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 13 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 14 | Lesser General Public License for more details. |
| 15 | |
| 16 | You should have received a copy of the GNU Lesser General Public |
| 17 | License along with the GNU C Library; if not, see |
| 18 | <http://www.gnu.org/licenses/>. */ |
| 19 | |
| 20 | #include <dlfcn.h> |
| 21 | #include <errno.h> |
| 22 | #include <fcntl.h> |
| 23 | #include <stdlib.h> |
| 24 | #include <string.h> |
| 25 | #include <unistd.h> |
| 26 | #include <sys/mman.h> |
| 27 | #include <sys/stat.h> |
| 28 | |
| 29 | #include <gconv_int.h> |
| 30 | #include <iconvconfig.h> |
| 31 | #include <not-cancel.h> |
| 32 | |
| 33 | #include "../intl/hash-string.h" |
| 34 | |
| 35 | static void *gconv_cache; |
| 36 | static size_t cache_size; |
| 37 | static int cache_malloced; |
| 38 | |
| 39 | |
| 40 | void * |
| 41 | __gconv_get_cache (void) |
| 42 | { |
| 43 | return gconv_cache; |
| 44 | } |
| 45 | |
| 46 | |
| 47 | int |
| 48 | internal_function |
| 49 | __gconv_load_cache (void) |
| 50 | { |
| 51 | int fd; |
| 52 | struct stat64 st; |
| 53 | struct gconvcache_header *header; |
| 54 | |
| 55 | /* We cannot use the cache if the GCONV_PATH environment variable is |
| 56 | set. */ |
| 57 | __gconv_path_envvar = getenv ("GCONV_PATH"); |
| 58 | if (__gconv_path_envvar != NULL) |
| 59 | return -1; |
| 60 | |
| 61 | /* See whether the cache file exists. */ |
| 62 | fd = open_not_cancel (GCONV_MODULES_CACHE, O_RDONLY, 0); |
| 63 | if (__builtin_expect (fd, 0) == -1) |
| 64 | /* Not available. */ |
| 65 | return -1; |
| 66 | |
| 67 | /* Get information about the file. */ |
| 68 | if (__builtin_expect (__fxstat64 (_STAT_VER, fd, &st), 0) < 0 |
| 69 | /* We do not have to start looking at the file if it cannot contain |
| 70 | at least the cache header. */ |
| 71 | || (size_t) st.st_size < sizeof (struct gconvcache_header)) |
| 72 | { |
| 73 | close_and_exit: |
| 74 | close_not_cancel_no_status (fd); |
| 75 | return -1; |
| 76 | } |
| 77 | |
| 78 | /* Make the file content available. */ |
| 79 | cache_size = st.st_size; |
| 80 | #ifdef _POSIX_MAPPED_FILES |
| 81 | gconv_cache = __mmap (NULL, cache_size, PROT_READ, MAP_SHARED, fd, 0); |
| 82 | if (__glibc_unlikely (gconv_cache == MAP_FAILED)) |
| 83 | #endif |
| 84 | { |
| 85 | size_t already_read; |
| 86 | |
| 87 | gconv_cache = malloc (cache_size); |
| 88 | if (gconv_cache == NULL) |
| 89 | goto close_and_exit; |
| 90 | |
| 91 | already_read = 0; |
| 92 | do |
| 93 | { |
| 94 | ssize_t n = __read (fd, (char *) gconv_cache + already_read, |
| 95 | cache_size - already_read); |
| 96 | if (__builtin_expect (n, 0) == -1) |
| 97 | { |
| 98 | free (gconv_cache); |
| 99 | gconv_cache = NULL; |
| 100 | goto close_and_exit; |
| 101 | } |
| 102 | |
| 103 | already_read += n; |
| 104 | } |
| 105 | while (already_read < cache_size); |
| 106 | |
| 107 | cache_malloced = 1; |
| 108 | } |
| 109 | |
| 110 | /* We don't need the file descriptor anymore. */ |
| 111 | close_not_cancel_no_status (fd); |
| 112 | |
| 113 | /* Check the consistency. */ |
| 114 | header = (struct gconvcache_header *) gconv_cache; |
| 115 | if (__builtin_expect (header->magic, GCONVCACHE_MAGIC) != GCONVCACHE_MAGIC |
| 116 | || __builtin_expect (header->string_offset >= cache_size, 0) |
| 117 | || __builtin_expect (header->hash_offset >= cache_size, 0) |
| 118 | || __builtin_expect (header->hash_size == 0, 0) |
| 119 | || __builtin_expect ((header->hash_offset |
| 120 | + header->hash_size * sizeof (struct hash_entry)) |
| 121 | > cache_size, 0) |
| 122 | || __builtin_expect (header->module_offset >= cache_size, 0) |
| 123 | || __builtin_expect (header->otherconv_offset > cache_size, 0)) |
| 124 | { |
| 125 | if (cache_malloced) |
| 126 | { |
| 127 | free (gconv_cache); |
| 128 | cache_malloced = 0; |
| 129 | } |
| 130 | #ifdef _POSIX_MAPPED_FILES |
| 131 | else |
| 132 | __munmap (gconv_cache, cache_size); |
| 133 | #endif |
| 134 | gconv_cache = NULL; |
| 135 | |
| 136 | return -1; |
| 137 | } |
| 138 | |
| 139 | /* That worked. */ |
| 140 | return 0; |
| 141 | } |
| 142 | |
| 143 | |
| 144 | static int |
| 145 | internal_function |
| 146 | find_module_idx (const char *str, size_t *idxp) |
| 147 | { |
| 148 | unsigned int idx; |
| 149 | unsigned int hval; |
| 150 | unsigned int hval2; |
| 151 | const struct gconvcache_header *header; |
| 152 | const char *strtab; |
| 153 | const struct hash_entry *hashtab; |
| 154 | unsigned int limit; |
| 155 | |
| 156 | header = (const struct gconvcache_header *) gconv_cache; |
| 157 | strtab = (char *) gconv_cache + header->string_offset; |
| 158 | hashtab = (struct hash_entry *) ((char *) gconv_cache |
| 159 | + header->hash_offset); |
| 160 | |
| 161 | hval = __hash_string (str); |
| 162 | idx = hval % header->hash_size; |
| 163 | hval2 = 1 + hval % (header->hash_size - 2); |
| 164 | |
| 165 | limit = cache_size - header->string_offset; |
| 166 | while (hashtab[idx].string_offset != 0) |
| 167 | if (hashtab[idx].string_offset < limit |
| 168 | && strcmp (str, strtab + hashtab[idx].string_offset) == 0) |
| 169 | { |
| 170 | *idxp = hashtab[idx].module_idx; |
| 171 | return 0; |
| 172 | } |
| 173 | else |
| 174 | if ((idx += hval2) >= header->hash_size) |
| 175 | idx -= header->hash_size; |
| 176 | |
| 177 | /* Nothing found. */ |
| 178 | return -1; |
| 179 | } |
| 180 | |
| 181 | |
| 182 | #ifndef STATIC_GCONV |
| 183 | static int |
| 184 | internal_function |
| 185 | find_module (const char *directory, const char *filename, |
| 186 | struct __gconv_step *result) |
| 187 | { |
| 188 | size_t dirlen = strlen (directory); |
| 189 | size_t fnamelen = strlen (filename) + 1; |
| 190 | char fullname[dirlen + fnamelen]; |
| 191 | int status = __GCONV_NOCONV; |
| 192 | |
| 193 | memcpy (__mempcpy (fullname, directory, dirlen), filename, fnamelen); |
| 194 | |
| 195 | result->__shlib_handle = __gconv_find_shlib (fullname); |
| 196 | if (result->__shlib_handle != NULL) |
| 197 | { |
| 198 | status = __GCONV_OK; |
| 199 | |
| 200 | result->__modname = NULL; |
| 201 | result->__fct = result->__shlib_handle->fct; |
| 202 | result->__init_fct = result->__shlib_handle->init_fct; |
| 203 | result->__end_fct = result->__shlib_handle->end_fct; |
| 204 | |
| 205 | /* These settings can be overridden by the init function. */ |
| 206 | result->__btowc_fct = NULL; |
| 207 | result->__data = NULL; |
| 208 | |
| 209 | /* Call the init function. */ |
| 210 | if (result->__init_fct != NULL) |
| 211 | { |
| 212 | __gconv_init_fct init_fct = result->__init_fct; |
| 213 | #ifdef PTR_DEMANGLE |
| 214 | PTR_DEMANGLE (init_fct); |
| 215 | #endif |
| 216 | status = DL_CALL_FCT (init_fct, (result)); |
| 217 | |
| 218 | #ifdef PTR_MANGLE |
| 219 | if (result->__btowc_fct != NULL) |
| 220 | PTR_MANGLE (result->__btowc_fct); |
| 221 | #endif |
| 222 | } |
| 223 | } |
| 224 | |
| 225 | return status; |
| 226 | } |
| 227 | #endif |
| 228 | |
| 229 | |
| 230 | int |
| 231 | internal_function |
| 232 | __gconv_compare_alias_cache (const char *name1, const char *name2, int *result) |
| 233 | { |
| 234 | size_t name1_idx; |
| 235 | size_t name2_idx; |
| 236 | |
| 237 | if (gconv_cache == NULL) |
| 238 | return -1; |
| 239 | |
| 240 | if (find_module_idx (name1, &name1_idx) != 0 |
| 241 | || find_module_idx (name2, &name2_idx) != 0) |
| 242 | *result = strcmp (name1, name2); |
| 243 | else |
| 244 | *result = (int) (name1_idx - name2_idx); |
| 245 | |
| 246 | return 0; |
| 247 | } |
| 248 | |
| 249 | |
| 250 | int |
| 251 | internal_function |
| 252 | __gconv_lookup_cache (const char *toset, const char *fromset, |
| 253 | struct __gconv_step **handle, size_t *nsteps, int flags) |
| 254 | { |
| 255 | const struct gconvcache_header *header; |
| 256 | const char *strtab; |
| 257 | size_t fromidx; |
| 258 | size_t toidx; |
| 259 | const struct module_entry *modtab; |
| 260 | const struct module_entry *from_module; |
| 261 | const struct module_entry *to_module; |
| 262 | struct __gconv_step *result; |
| 263 | |
| 264 | if (gconv_cache == NULL) |
| 265 | /* We have no cache available. */ |
| 266 | return __GCONV_NODB; |
| 267 | |
| 268 | header = (const struct gconvcache_header *) gconv_cache; |
| 269 | strtab = (char *) gconv_cache + header->string_offset; |
| 270 | modtab = (const struct module_entry *) ((char *) gconv_cache |
| 271 | + header->module_offset); |
| 272 | |
| 273 | if (find_module_idx (fromset, &fromidx) != 0 |
| 274 | || (header->module_offset + (fromidx + 1) * sizeof (struct module_entry) |
| 275 | > cache_size)) |
| 276 | return __GCONV_NOCONV; |
| 277 | from_module = &modtab[fromidx]; |
| 278 | |
| 279 | if (find_module_idx (toset, &toidx) != 0 |
| 280 | || (header->module_offset + (toidx + 1) * sizeof (struct module_entry) |
| 281 | > cache_size)) |
| 282 | return __GCONV_NOCONV; |
| 283 | to_module = &modtab[toidx]; |
| 284 | |
| 285 | /* Avoid copy-only transformations if the user requests. */ |
| 286 | if (__builtin_expect (flags & GCONV_AVOID_NOCONV, 0) && fromidx == toidx) |
| 287 | return __GCONV_NULCONV; |
| 288 | |
| 289 | /* If there are special conversions available examine them first. */ |
| 290 | if (fromidx != 0 && toidx != 0 |
| 291 | && __builtin_expect (from_module->extra_offset, 0) != 0) |
| 292 | { |
| 293 | /* Search through the list to see whether there is a module |
| 294 | matching the destination character set. */ |
| 295 | const struct extra_entry *extra; |
| 296 | |
| 297 | /* Note the -1. This is due to the offset added in iconvconfig. |
| 298 | See there for more explanations. */ |
| 299 | extra = (const struct extra_entry *) ((char *) gconv_cache |
| 300 | + header->otherconv_offset |
| 301 | + from_module->extra_offset - 1); |
| 302 | while (extra->module_cnt != 0 |
| 303 | && extra->module[extra->module_cnt - 1].outname_offset != toidx) |
| 304 | extra = (const struct extra_entry *) ((char *) extra |
| 305 | + sizeof (struct extra_entry) |
| 306 | + (extra->module_cnt |
| 307 | * sizeof (struct extra_entry_module))); |
| 308 | |
| 309 | if (extra->module_cnt != 0) |
| 310 | { |
| 311 | /* Use the extra module. First determine how many steps. */ |
| 312 | char *fromname; |
| 313 | int idx; |
| 314 | |
| 315 | *nsteps = extra->module_cnt; |
| 316 | *handle = result = |
| 317 | (struct __gconv_step *) malloc (extra->module_cnt |
| 318 | * sizeof (struct __gconv_step)); |
| 319 | if (result == NULL) |
| 320 | return __GCONV_NOMEM; |
| 321 | |
| 322 | fromname = (char *) strtab + from_module->canonname_offset; |
| 323 | idx = 0; |
| 324 | do |
| 325 | { |
| 326 | result[idx].__from_name = fromname; |
| 327 | fromname = result[idx].__to_name = |
| 328 | (char *) strtab + modtab[extra->module[idx].outname_offset].canonname_offset; |
| 329 | |
| 330 | result[idx].__counter = 1; |
| 331 | result[idx].__data = NULL; |
| 332 | |
| 333 | #ifndef STATIC_GCONV |
| 334 | if (strtab[extra->module[idx].dir_offset] != '\0') |
| 335 | { |
| 336 | /* Load the module, return handle for it. */ |
| 337 | int res; |
| 338 | |
| 339 | res = find_module (strtab + extra->module[idx].dir_offset, |
| 340 | strtab + extra->module[idx].name_offset, |
| 341 | &result[idx]); |
| 342 | if (__builtin_expect (res, __GCONV_OK) != __GCONV_OK) |
| 343 | { |
| 344 | /* Something went wrong. */ |
| 345 | free (result); |
| 346 | goto try_internal; |
| 347 | } |
| 348 | } |
| 349 | else |
| 350 | #endif |
| 351 | /* It's a builtin transformation. */ |
| 352 | __gconv_get_builtin_trans (strtab |
| 353 | + extra->module[idx].name_offset, |
| 354 | &result[idx]); |
| 355 | |
| 356 | } |
| 357 | while (++idx < extra->module_cnt); |
| 358 | |
| 359 | return __GCONV_OK; |
| 360 | } |
| 361 | } |
| 362 | |
| 363 | try_internal: |
| 364 | /* See whether we can convert via the INTERNAL charset. */ |
| 365 | if ((fromidx != 0 && __builtin_expect (from_module->fromname_offset, 1) == 0) |
| 366 | || (toidx != 0 && __builtin_expect (to_module->toname_offset, 1) == 0) |
| 367 | || (fromidx == 0 && toidx == 0)) |
| 368 | /* Not possible. Nothing we can do. */ |
| 369 | return __GCONV_NOCONV; |
| 370 | |
| 371 | /* We will use up to two modules. Always allocate room for two. */ |
| 372 | result = (struct __gconv_step *) malloc (2 * sizeof (struct __gconv_step)); |
| 373 | if (result == NULL) |
| 374 | return __GCONV_NOMEM; |
| 375 | |
| 376 | *handle = result; |
| 377 | *nsteps = 0; |
| 378 | |
| 379 | /* Generate data structure for conversion to INTERNAL. */ |
| 380 | if (fromidx != 0) |
| 381 | { |
| 382 | result[0].__from_name = (char *) strtab + from_module->canonname_offset; |
| 383 | result[0].__to_name = (char *) "INTERNAL"; |
| 384 | |
| 385 | result[0].__counter = 1; |
| 386 | result[0].__data = NULL; |
| 387 | |
| 388 | #ifndef STATIC_GCONV |
| 389 | if (strtab[from_module->todir_offset] != '\0') |
| 390 | { |
| 391 | /* Load the module, return handle for it. */ |
| 392 | int res = find_module (strtab + from_module->todir_offset, |
| 393 | strtab + from_module->toname_offset, |
| 394 | &result[0]); |
| 395 | if (__builtin_expect (res, __GCONV_OK) != __GCONV_OK) |
| 396 | { |
| 397 | /* Something went wrong. */ |
| 398 | free (result); |
| 399 | return res; |
| 400 | } |
| 401 | } |
| 402 | else |
| 403 | #endif |
| 404 | /* It's a builtin transformation. */ |
| 405 | __gconv_get_builtin_trans (strtab + from_module->toname_offset, |
| 406 | &result[0]); |
| 407 | |
| 408 | ++*nsteps; |
| 409 | } |
| 410 | |
| 411 | /* Generate data structure for conversion from INTERNAL. */ |
| 412 | if (toidx != 0) |
| 413 | { |
| 414 | int idx = *nsteps; |
| 415 | |
| 416 | result[idx].__from_name = (char *) "INTERNAL"; |
| 417 | result[idx].__to_name = (char *) strtab + to_module->canonname_offset; |
| 418 | |
| 419 | result[idx].__counter = 1; |
| 420 | result[idx].__data = NULL; |
| 421 | |
| 422 | #ifndef STATIC_GCONV |
| 423 | if (strtab[to_module->fromdir_offset] != '\0') |
| 424 | { |
| 425 | /* Load the module, return handle for it. */ |
| 426 | int res = find_module (strtab + to_module->fromdir_offset, |
| 427 | strtab + to_module->fromname_offset, |
| 428 | &result[idx]); |
| 429 | if (__builtin_expect (res, __GCONV_OK) != __GCONV_OK) |
| 430 | { |
| 431 | /* Something went wrong. */ |
| 432 | if (idx != 0) |
| 433 | __gconv_release_step (&result[0]); |
| 434 | free (result); |
| 435 | return res; |
| 436 | } |
| 437 | } |
| 438 | else |
| 439 | #endif |
| 440 | /* It's a builtin transformation. */ |
| 441 | __gconv_get_builtin_trans (strtab + to_module->fromname_offset, |
| 442 | &result[idx]); |
| 443 | |
| 444 | ++*nsteps; |
| 445 | } |
| 446 | |
| 447 | return __GCONV_OK; |
| 448 | } |
| 449 | |
| 450 | |
| 451 | /* Free memory allocated for the transformation record. */ |
| 452 | void |
| 453 | internal_function |
| 454 | __gconv_release_cache (struct __gconv_step *steps, size_t nsteps) |
| 455 | { |
| 456 | if (gconv_cache != NULL) |
| 457 | /* The only thing we have to deallocate is the record with the |
| 458 | steps. */ |
| 459 | free (steps); |
| 460 | } |
| 461 | |
| 462 | |
| 463 | /* Free all resources if necessary. */ |
| 464 | libc_freeres_fn (free_mem) |
| 465 | { |
| 466 | if (cache_malloced) |
| 467 | free (gconv_cache); |
| 468 | #ifdef _POSIX_MAPPED_FILES |
| 469 | else if (gconv_cache != NULL) |
| 470 | __munmap (gconv_cache, cache_size); |
| 471 | #endif |
| 472 | } |