iconv/gconv_cache.c

   1 /* Cache handling for iconv modules.
   2    Copyright (C) 2001-2024 Free Software Foundation, Inc.
   3    This file is part of the GNU C Library.
   4
   5    The GNU C Library is free software; you can redistribute it and/or
   6    modify it under the terms of the GNU Lesser General Public
   7    License as published by the Free Software Foundation; either
   8    version 2.1 of the License, or (at your option) any later version.
   9
  10    The GNU C Library is distributed in the hope that it will be useful,
  11    but WITHOUT ANY WARRANTY; without even the implied warranty of
  12    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13    Lesser General Public License for more details.
  14
  15    You should have received a copy of the GNU Lesser General Public
  16    License along with the GNU C Library; if not, see
  17    <https://www.gnu.org/licenses/>.  */
  18
  19 #include <dlfcn.h>
  20 #include <errno.h>
  21 #include <fcntl.h>
  22 #include <stdlib.h>
  23 #include <string.h>
  24 #include <unistd.h>
  25 #include <sys/mman.h>
  26 #include <sys/stat.h>
  27
  28 #include <gconv_int.h>
  29 #include <iconvconfig.h>
  30 #include <not-cancel.h>
  31 #include <pointer_guard.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 __gconv_load_cache (void)
  49 {
  50   int fd;
  51   struct __stat64_t64 st;
  52   struct gconvcache_header *header;
  53
  54   /* We cannot use the cache if the GCONV_PATH environment variable is
  55      set.  */
  56   __gconv_path_envvar = getenv ("GCONV_PATH");
  57   if (__gconv_path_envvar != NULL)
  58     return -1;
  59
  60   /* See whether the cache file exists.  */
  61   fd = __open_nocancel (GCONV_MODULES_CACHE, O_RDONLY | O_CLOEXEC, 0);
  62   if (__builtin_expect (fd, 0) == -1)
  63     /* Not available.  */
  64     return -1;
  65
  66   /* Get information about the file.  */
  67   if (__glibc_unlikely (__fstat64_time64 (fd, &st) < 0)
  68       /* We do not have to start looking at the file if it cannot contain
  69          at least the cache header.  */
  70       || (size_t) st.st_size < sizeof (struct gconvcache_header))
  71     {
  72     close_and_exit:
  73       __close_nocancel_nostatus (fd);
  74       return -1;
  75     }
  76
  77   /* Make the file content available.  */
  78   cache_size = st.st_size;
  79 #ifdef _POSIX_MAPPED_FILES
  80   gconv_cache = __mmap (NULL, cache_size, PROT_READ, MAP_SHARED, fd, 0);
  81   if (__glibc_unlikely (gconv_cache == MAP_FAILED))
  82 #endif
  83     {
  84       size_t already_read;
  85
  86       gconv_cache = malloc (cache_size);
  87       if (gconv_cache == NULL)
  88         goto close_and_exit;
  89
  90       already_read = 0;
  91       do
  92         {
  93           ssize_t n = __read (fd, (char *) gconv_cache + already_read,
  94                               cache_size - already_read);
  95           if (__builtin_expect (n, 0) == -1)
  96             {
  97               free (gconv_cache);
  98               gconv_cache = NULL;
  99               goto close_and_exit;
 100             }
 101
 102           already_read += n;
 103         }
 104       while (already_read < cache_size);
 105
 106       cache_malloced = 1;
 107     }
 108
 109   /* We don't need the file descriptor anymore.  */
 110   __close_nocancel_nostatus (fd);
 111
 112   /* Check the consistency.  */
 113   header = (struct gconvcache_header *) gconv_cache;
 114   if (__builtin_expect (header->magic, GCONVCACHE_MAGIC) != GCONVCACHE_MAGIC
 115       || __builtin_expect (header->string_offset >= cache_size, 0)
 116       || __builtin_expect (header->hash_offset >= cache_size, 0)
 117       || __builtin_expect (header->hash_size == 0, 0)
 118       || __builtin_expect ((header->hash_offset
 119                             + header->hash_size * sizeof (struct hash_entry))
 120                            > cache_size, 0)
 121       || __builtin_expect (header->module_offset >= cache_size, 0)
 122       || __builtin_expect (header->otherconv_offset > cache_size, 0))
 123     {
 124       if (cache_malloced)
 125         {
 126           free (gconv_cache);
 127           cache_malloced = 0;
 128         }
 129 #ifdef _POSIX_MAPPED_FILES
 130       else
 131         __munmap (gconv_cache, cache_size);
 132 #endif
 133       gconv_cache = NULL;
 134
 135       return -1;
 136     }
 137
 138   /* That worked.  */
 139   return 0;
 140 }
 141
 142
 143 static int
 144 find_module_idx (const char *str, size_t *idxp)
 145 {
 146   unsigned int idx;
 147   unsigned int hval;
 148   unsigned int hval2;
 149   const struct gconvcache_header *header;
 150   const char *strtab;
 151   const struct hash_entry *hashtab;
 152   unsigned int limit;
 153
 154   header = (const struct gconvcache_header *) gconv_cache;
 155   strtab = (char *) gconv_cache + header->string_offset;
 156   hashtab = (struct hash_entry *) ((char *) gconv_cache
 157                                    + header->hash_offset);
 158
 159   hval = __hash_string (str);
 160   idx = hval % header->hash_size;
 161   hval2 = 1 + hval % (header->hash_size - 2);
 162
 163   limit = cache_size - header->string_offset;
 164   while (hashtab[idx].string_offset != 0)
 165     if (hashtab[idx].string_offset < limit
 166         && strcmp (str, strtab + hashtab[idx].string_offset) == 0)
 167       {
 168         *idxp = hashtab[idx].module_idx;
 169         return 0;
 170       }
 171     else
 172       if ((idx += hval2) >= header->hash_size)
 173         idx -= header->hash_size;
 174
 175   /* Nothing found.  */
 176   return -1;
 177 }
 178
 179
 180 #ifndef STATIC_GCONV
 181 static int
 182 find_module (const char *directory, const char *filename,
 183              struct __gconv_step *result)
 184 {
 185   size_t dirlen = strlen (directory);
 186   size_t fnamelen = strlen (filename) + 1;
 187   char fullname[dirlen + fnamelen];
 188   int status = __GCONV_NOCONV;
 189
 190   memcpy (__mempcpy (fullname, directory, dirlen), filename, fnamelen);
 191
 192   result->__shlib_handle = __gconv_find_shlib (fullname);
 193   if (result->__shlib_handle != NULL)
 194     {
 195       status = __GCONV_OK;
 196
 197       result->__modname = NULL;
 198       result->__fct = result->__shlib_handle->fct;
 199       result->__init_fct = result->__shlib_handle->init_fct;
 200       result->__end_fct = result->__shlib_handle->end_fct;
 201
 202       /* These settings can be overridden by the init function.  */
 203       result->__btowc_fct = NULL;
 204       result->__data = NULL;
 205
 206       /* Call the init function.  */
 207       __gconv_init_fct init_fct = result->__init_fct;
 208       PTR_DEMANGLE (init_fct);
 209       if (init_fct != NULL)
 210         {
 211           status = DL_CALL_FCT (init_fct, (result));
 212           PTR_MANGLE (result->__btowc_fct);
 213         }
 214     }
 215
 216   return status;
 217 }
 218 #endif
 219
 220
 221 int
 222 __gconv_compare_alias_cache (const char *name1, const char *name2, int *result)
 223 {
 224   size_t name1_idx;
 225   size_t name2_idx;
 226
 227   if (gconv_cache == NULL)
 228     return -1;
 229
 230   if (find_module_idx (name1, &name1_idx) != 0
 231       || find_module_idx (name2, &name2_idx) != 0)
 232     *result = strcmp (name1, name2);
 233   else
 234     *result = (int) (name1_idx - name2_idx);
 235
 236   return 0;
 237 }
 238
 239
 240 int
 241 __gconv_lookup_cache (const char *toset, const char *fromset,
 242                       struct __gconv_step **handle, size_t *nsteps, int flags)
 243 {
 244   const struct gconvcache_header *header;
 245   const char *strtab;
 246   size_t fromidx;
 247   size_t toidx;
 248   const struct module_entry *modtab;
 249   const struct module_entry *from_module;
 250   const struct module_entry *to_module;
 251   struct __gconv_step *result;
 252
 253   if (gconv_cache == NULL)
 254     /* We have no cache available.  */
 255     return __GCONV_NODB;
 256
 257   header = (const struct gconvcache_header *) gconv_cache;
 258   strtab = (char *) gconv_cache + header->string_offset;
 259   modtab = (const struct module_entry *) ((char *) gconv_cache
 260                                           + header->module_offset);
 261
 262   if (find_module_idx (fromset, &fromidx) != 0
 263       || (header->module_offset + (fromidx + 1) * sizeof (struct module_entry)
 264           > cache_size))
 265     return __GCONV_NOCONV;
 266   from_module = &modtab[fromidx];
 267
 268   if (find_module_idx (toset, &toidx) != 0
 269       || (header->module_offset + (toidx + 1) * sizeof (struct module_entry)
 270           > cache_size))
 271     return __GCONV_NOCONV;
 272   to_module = &modtab[toidx];
 273
 274   /* Avoid copy-only transformations if the user requests.   */
 275   if (__builtin_expect (flags & GCONV_AVOID_NOCONV, 0) && fromidx == toidx)
 276     return __GCONV_NULCONV;
 277
 278   /* If there are special conversions available examine them first.  */
 279   if (fromidx != 0 && toidx != 0
 280       && __builtin_expect (from_module->extra_offset, 0) != 0)
 281     {
 282       /* Search through the list to see whether there is a module
 283          matching the destination character set.  */
 284       const struct extra_entry *extra;
 285
 286       /* Note the -1.  This is due to the offset added in iconvconfig.
 287          See there for more explanations.  */
 288       extra = (const struct extra_entry *) ((char *) gconv_cache
 289                                             + header->otherconv_offset
 290                                             + from_module->extra_offset - 1);
 291       while (extra->module_cnt != 0
 292              && extra->module[extra->module_cnt - 1].outname_offset != toidx)
 293         extra = (const struct extra_entry *) ((char *) extra
 294                                               + sizeof (struct extra_entry)
 295                                               + (extra->module_cnt
 296                                                  * sizeof (struct extra_entry_module)));
 297
 298       if (extra->module_cnt != 0)
 299         {
 300           /* Use the extra module.  First determine how many steps.  */
 301           char *fromname;
 302           int idx;
 303
 304           *nsteps = extra->module_cnt;
 305           *handle = result =
 306             (struct __gconv_step *) malloc (extra->module_cnt
 307                                             * sizeof (struct __gconv_step));
 308           if (result == NULL)
 309             return __GCONV_NOMEM;
 310
 311           fromname = (char *) strtab + from_module->canonname_offset;
 312           idx = 0;
 313           do
 314             {
 315               result[idx].__from_name = fromname;
 316               fromname = result[idx].__to_name =
 317                 (char *) strtab + modtab[extra->module[idx].outname_offset].canonname_offset;
 318
 319               result[idx].__counter = 1;
 320               result[idx].__data = NULL;
 321
 322 #ifndef STATIC_GCONV
 323               if (strtab[extra->module[idx].dir_offset] != '\0')
 324                 {
 325                   /* Load the module, return handle for it.  */
 326                   int res;
 327
 328                   res = find_module (strtab + extra->module[idx].dir_offset,
 329                                      strtab + extra->module[idx].name_offset,
 330                                      &result[idx]);
 331                   if (__builtin_expect (res, __GCONV_OK) != __GCONV_OK)
 332                     {
 333                       /* Something went wrong.  */
 334                       free (result);
 335                       goto try_internal;
 336                     }
 337                 }
 338               else
 339 #endif
 340                 /* It's a builtin transformation.  */
 341                 __gconv_get_builtin_trans (strtab
 342                                            + extra->module[idx].name_offset,
 343                                            &result[idx]);
 344
 345             }
 346           while (++idx < extra->module_cnt);
 347
 348           return __GCONV_OK;
 349         }
 350     }
 351
 352  try_internal:
 353   /* See whether we can convert via the INTERNAL charset.  */
 354   if ((fromidx != 0 && __builtin_expect (from_module->fromname_offset, 1) == 0)
 355       || (toidx != 0 && __builtin_expect (to_module->toname_offset, 1) == 0)
 356       || (fromidx == 0 && toidx == 0))
 357     /* Not possible.  Nothing we can do.  */
 358     return __GCONV_NOCONV;
 359
 360   /* We will use up to two modules.  Always allocate room for two.  */
 361   result = (struct __gconv_step *) malloc (2 * sizeof (struct __gconv_step));
 362   if (result == NULL)
 363     return __GCONV_NOMEM;
 364
 365   *handle = result;
 366   *nsteps = 0;
 367
 368   /* Generate data structure for conversion to INTERNAL.  */
 369   if (fromidx != 0)
 370     {
 371       result[0].__from_name = (char *) strtab + from_module->canonname_offset;
 372       result[0].__to_name = (char *) "INTERNAL";
 373
 374       result[0].__counter = 1;
 375       result[0].__data = NULL;
 376
 377 #ifndef STATIC_GCONV
 378       if (strtab[from_module->todir_offset] != '\0')
 379         {
 380           /* Load the module, return handle for it.  */
 381           int res = find_module (strtab + from_module->todir_offset,
 382                                  strtab + from_module->toname_offset,
 383                                  &result[0]);
 384           if (__builtin_expect (res, __GCONV_OK) != __GCONV_OK)
 385             {
 386               /* Something went wrong.  */
 387               free (result);
 388               return res;
 389             }
 390         }
 391       else
 392 #endif
 393         /* It's a builtin transformation.  */
 394         __gconv_get_builtin_trans (strtab + from_module->toname_offset,
 395                                    &result[0]);
 396
 397       ++*nsteps;
 398     }
 399
 400   /* Generate data structure for conversion from INTERNAL.  */
 401   if (toidx != 0)
 402     {
 403       int idx = *nsteps;
 404
 405       result[idx].__from_name = (char *) "INTERNAL";
 406       result[idx].__to_name = (char *) strtab + to_module->canonname_offset;
 407
 408       result[idx].__counter = 1;
 409       result[idx].__data = NULL;
 410
 411 #ifndef STATIC_GCONV
 412       if (strtab[to_module->fromdir_offset] != '\0')
 413         {
 414           /* Load the module, return handle for it.  */
 415           int res = find_module (strtab + to_module->fromdir_offset,
 416                                  strtab + to_module->fromname_offset,
 417                                  &result[idx]);
 418           if (__builtin_expect (res, __GCONV_OK) != __GCONV_OK)
 419             {
 420               /* Something went wrong.  */
 421               if (idx != 0)
 422                 __gconv_release_step (&result[0]);
 423               free (result);
 424               return res;
 425             }
 426         }
 427       else
 428 #endif
 429         /* It's a builtin transformation.  */
 430         __gconv_get_builtin_trans (strtab + to_module->fromname_offset,
 431                                    &result[idx]);
 432
 433       ++*nsteps;
 434     }
 435
 436   return __GCONV_OK;
 437 }
 438
 439
 440 /* Free memory allocated for the transformation record.  */
 441 void
 442 __gconv_release_cache (struct __gconv_step *steps, size_t nsteps)
 443 {
 444   if (gconv_cache != NULL)
 445     /* The only thing we have to deallocate is the record with the
 446        steps.  */
 447     free (steps);
 448 }
 449
 450
 451 /* Free all resources if necessary.  */
 452 void
 453 __gconv_cache_freemem (void)
 454 {
 455   if (cache_malloced)
 456     free (gconv_cache);
 457 #ifdef _POSIX_MAPPED_FILES
 458   else if (gconv_cache != NULL)
 459     __munmap (gconv_cache, cache_size);
 460 #endif
 461 }