| 1 | /* An expandable hash tables datatype. |
| 2 | Copyright (C) 1999, 2000, 2001 Free Software Foundation, Inc. |
| 3 | Contributed by Vladimir Makarov (vmakarov@cygnus.com). |
| 4 | |
| 5 | This file is part of the libiberty library. |
| 6 | Libiberty is free software; you can redistribute it and/or |
| 7 | modify it under the terms of the GNU Library General Public |
| 8 | License as published by the Free Software Foundation; either |
| 9 | version 2 of the License, or (at your option) any later version. |
| 10 | |
| 11 | Libiberty 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 | Library General Public License for more details. |
| 15 | |
| 16 | You should have received a copy of the GNU Library General Public |
| 17 | License along with libiberty; see the file COPYING.LIB. If |
| 18 | not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, |
| 19 | Boston, MA 02111-1307, USA. */ |
| 20 | |
| 21 | /* This package implements basic hash table functionality. It is possible |
| 22 | to search for an entry, create an entry and destroy an entry. |
| 23 | |
| 24 | Elements in the table are generic pointers. |
| 25 | |
| 26 | The size of the table is not fixed; if the occupancy of the table |
| 27 | grows too high the hash table will be expanded. |
| 28 | |
| 29 | The abstract data implementation is based on generalized Algorithm D |
| 30 | from Knuth's book "The art of computer programming". Hash table is |
| 31 | expanded by creation of new hash table and transferring elements from |
| 32 | the old table to the new table. */ |
| 33 | |
| 34 | #ifdef HAVE_CONFIG_H |
| 35 | #include "config.h" |
| 36 | #endif |
| 37 | |
| 38 | #include <sys/types.h> |
| 39 | |
| 40 | #ifdef HAVE_STDLIB_H |
| 41 | #include <stdlib.h> |
| 42 | #endif |
| 43 | |
| 44 | #ifdef HAVE_STRING_H |
| 45 | #include <string.h> |
| 46 | #endif |
| 47 | |
| 48 | #include <stdio.h> |
| 49 | |
| 50 | #include "libiberty.h" |
| 51 | #include "hashtab.h" |
| 52 | |
| 53 | /* This macro defines reserved value for empty table entry. */ |
| 54 | |
| 55 | #define EMPTY_ENTRY ((PTR) 0) |
| 56 | |
| 57 | /* This macro defines reserved value for table entry which contained |
| 58 | a deleted element. */ |
| 59 | |
| 60 | #define DELETED_ENTRY ((PTR) 1) |
| 61 | |
| 62 | static unsigned long higher_prime_number PARAMS ((unsigned long)); |
| 63 | static hashval_t hash_pointer PARAMS ((const void *)); |
| 64 | static int eq_pointer PARAMS ((const void *, const void *)); |
| 65 | static int htab_expand PARAMS ((htab_t)); |
| 66 | static PTR *find_empty_slot_for_expand PARAMS ((htab_t, hashval_t)); |
| 67 | |
| 68 | /* At some point, we could make these be NULL, and modify the |
| 69 | hash-table routines to handle NULL specially; that would avoid |
| 70 | function-call overhead for the common case of hashing pointers. */ |
| 71 | htab_hash htab_hash_pointer = hash_pointer; |
| 72 | htab_eq htab_eq_pointer = eq_pointer; |
| 73 | |
| 74 | /* The following function returns a nearest prime number which is |
| 75 | greater than N, and near a power of two. */ |
| 76 | |
| 77 | static unsigned long |
| 78 | higher_prime_number (n) |
| 79 | unsigned long n; |
| 80 | { |
| 81 | /* These are primes that are near, but slightly smaller than, a |
| 82 | power of two. */ |
| 83 | static const unsigned long primes[] = { |
| 84 | (unsigned long) 2, |
| 85 | (unsigned long) 7, |
| 86 | (unsigned long) 13, |
| 87 | (unsigned long) 31, |
| 88 | (unsigned long) 61, |
| 89 | (unsigned long) 127, |
| 90 | (unsigned long) 251, |
| 91 | (unsigned long) 509, |
| 92 | (unsigned long) 1021, |
| 93 | (unsigned long) 2039, |
| 94 | (unsigned long) 4093, |
| 95 | (unsigned long) 8191, |
| 96 | (unsigned long) 16381, |
| 97 | (unsigned long) 32749, |
| 98 | (unsigned long) 65521, |
| 99 | (unsigned long) 131071, |
| 100 | (unsigned long) 262139, |
| 101 | (unsigned long) 524287, |
| 102 | (unsigned long) 1048573, |
| 103 | (unsigned long) 2097143, |
| 104 | (unsigned long) 4194301, |
| 105 | (unsigned long) 8388593, |
| 106 | (unsigned long) 16777213, |
| 107 | (unsigned long) 33554393, |
| 108 | (unsigned long) 67108859, |
| 109 | (unsigned long) 134217689, |
| 110 | (unsigned long) 268435399, |
| 111 | (unsigned long) 536870909, |
| 112 | (unsigned long) 1073741789, |
| 113 | (unsigned long) 2147483647, |
| 114 | /* 4294967291L */ |
| 115 | ((unsigned long) 2147483647) + ((unsigned long) 2147483644), |
| 116 | }; |
| 117 | |
| 118 | const unsigned long *low = &primes[0]; |
| 119 | const unsigned long *high = &primes[sizeof(primes) / sizeof(primes[0])]; |
| 120 | |
| 121 | while (low != high) |
| 122 | { |
| 123 | const unsigned long *mid = low + (high - low) / 2; |
| 124 | if (n > *mid) |
| 125 | low = mid + 1; |
| 126 | else |
| 127 | high = mid; |
| 128 | } |
| 129 | |
| 130 | /* If we've run out of primes, abort. */ |
| 131 | if (n > *low) |
| 132 | { |
| 133 | fprintf (stderr, "Cannot find prime bigger than %lu\n", n); |
| 134 | abort (); |
| 135 | } |
| 136 | |
| 137 | return *low; |
| 138 | } |
| 139 | |
| 140 | /* Returns a hash code for P. */ |
| 141 | |
| 142 | static hashval_t |
| 143 | hash_pointer (p) |
| 144 | const PTR p; |
| 145 | { |
| 146 | return (hashval_t) ((long)p >> 3); |
| 147 | } |
| 148 | |
| 149 | /* Returns non-zero if P1 and P2 are equal. */ |
| 150 | |
| 151 | static int |
| 152 | eq_pointer (p1, p2) |
| 153 | const PTR p1; |
| 154 | const PTR p2; |
| 155 | { |
| 156 | return p1 == p2; |
| 157 | } |
| 158 | |
| 159 | /* This function creates table with length slightly longer than given |
| 160 | source length. Created hash table is initiated as empty (all the |
| 161 | hash table entries are EMPTY_ENTRY). The function returns the |
| 162 | created hash table. Memory allocation must not fail. */ |
| 163 | |
| 164 | htab_t |
| 165 | htab_create (size, hash_f, eq_f, del_f) |
| 166 | size_t size; |
| 167 | htab_hash hash_f; |
| 168 | htab_eq eq_f; |
| 169 | htab_del del_f; |
| 170 | { |
| 171 | htab_t result; |
| 172 | |
| 173 | size = higher_prime_number (size); |
| 174 | result = (htab_t) xcalloc (1, sizeof (struct htab)); |
| 175 | result->entries = (PTR *) xcalloc (size, sizeof (PTR)); |
| 176 | result->size = size; |
| 177 | result->hash_f = hash_f; |
| 178 | result->eq_f = eq_f; |
| 179 | result->del_f = del_f; |
| 180 | result->return_allocation_failure = 0; |
| 181 | return result; |
| 182 | } |
| 183 | |
| 184 | /* This function creates table with length slightly longer than given |
| 185 | source length. The created hash table is initiated as empty (all the |
| 186 | hash table entries are EMPTY_ENTRY). The function returns the created |
| 187 | hash table. Memory allocation may fail; it may return NULL. */ |
| 188 | |
| 189 | htab_t |
| 190 | htab_try_create (size, hash_f, eq_f, del_f) |
| 191 | size_t size; |
| 192 | htab_hash hash_f; |
| 193 | htab_eq eq_f; |
| 194 | htab_del del_f; |
| 195 | { |
| 196 | htab_t result; |
| 197 | |
| 198 | size = higher_prime_number (size); |
| 199 | result = (htab_t) calloc (1, sizeof (struct htab)); |
| 200 | if (result == NULL) |
| 201 | return NULL; |
| 202 | |
| 203 | result->entries = (PTR *) calloc (size, sizeof (PTR)); |
| 204 | if (result->entries == NULL) |
| 205 | { |
| 206 | free (result); |
| 207 | return NULL; |
| 208 | } |
| 209 | |
| 210 | result->size = size; |
| 211 | result->hash_f = hash_f; |
| 212 | result->eq_f = eq_f; |
| 213 | result->del_f = del_f; |
| 214 | result->return_allocation_failure = 1; |
| 215 | return result; |
| 216 | } |
| 217 | |
| 218 | /* This function frees all memory allocated for given hash table. |
| 219 | Naturally the hash table must already exist. */ |
| 220 | |
| 221 | void |
| 222 | htab_delete (htab) |
| 223 | htab_t htab; |
| 224 | { |
| 225 | int i; |
| 226 | |
| 227 | if (htab->del_f) |
| 228 | for (i = htab->size - 1; i >= 0; i--) |
| 229 | if (htab->entries[i] != EMPTY_ENTRY |
| 230 | && htab->entries[i] != DELETED_ENTRY) |
| 231 | (*htab->del_f) (htab->entries[i]); |
| 232 | |
| 233 | free (htab->entries); |
| 234 | free (htab); |
| 235 | } |
| 236 | |
| 237 | /* This function clears all entries in the given hash table. */ |
| 238 | |
| 239 | void |
| 240 | htab_empty (htab) |
| 241 | htab_t htab; |
| 242 | { |
| 243 | int i; |
| 244 | |
| 245 | if (htab->del_f) |
| 246 | for (i = htab->size - 1; i >= 0; i--) |
| 247 | if (htab->entries[i] != EMPTY_ENTRY |
| 248 | && htab->entries[i] != DELETED_ENTRY) |
| 249 | (*htab->del_f) (htab->entries[i]); |
| 250 | |
| 251 | memset (htab->entries, 0, htab->size * sizeof (PTR)); |
| 252 | } |
| 253 | |
| 254 | /* Similar to htab_find_slot, but without several unwanted side effects: |
| 255 | - Does not call htab->eq_f when it finds an existing entry. |
| 256 | - Does not change the count of elements/searches/collisions in the |
| 257 | hash table. |
| 258 | This function also assumes there are no deleted entries in the table. |
| 259 | HASH is the hash value for the element to be inserted. */ |
| 260 | |
| 261 | static PTR * |
| 262 | find_empty_slot_for_expand (htab, hash) |
| 263 | htab_t htab; |
| 264 | hashval_t hash; |
| 265 | { |
| 266 | size_t size = htab->size; |
| 267 | hashval_t hash2 = 1 + hash % (size - 2); |
| 268 | unsigned int index = hash % size; |
| 269 | |
| 270 | for (;;) |
| 271 | { |
| 272 | PTR *slot = htab->entries + index; |
| 273 | |
| 274 | if (*slot == EMPTY_ENTRY) |
| 275 | return slot; |
| 276 | else if (*slot == DELETED_ENTRY) |
| 277 | abort (); |
| 278 | |
| 279 | index += hash2; |
| 280 | if (index >= size) |
| 281 | index -= size; |
| 282 | } |
| 283 | } |
| 284 | |
| 285 | /* The following function changes size of memory allocated for the |
| 286 | entries and repeatedly inserts the table elements. The occupancy |
| 287 | of the table after the call will be about 50%. Naturally the hash |
| 288 | table must already exist. Remember also that the place of the |
| 289 | table entries is changed. If memory allocation failures are allowed, |
| 290 | this function will return zero, indicating that the table could not be |
| 291 | expanded. If all goes well, it will return a non-zero value. */ |
| 292 | |
| 293 | static int |
| 294 | htab_expand (htab) |
| 295 | htab_t htab; |
| 296 | { |
| 297 | PTR *oentries; |
| 298 | PTR *olimit; |
| 299 | PTR *p; |
| 300 | |
| 301 | oentries = htab->entries; |
| 302 | olimit = oentries + htab->size; |
| 303 | |
| 304 | htab->size = higher_prime_number (htab->size * 2); |
| 305 | |
| 306 | if (htab->return_allocation_failure) |
| 307 | { |
| 308 | PTR *nentries = (PTR *) calloc (htab->size, sizeof (PTR *)); |
| 309 | if (nentries == NULL) |
| 310 | return 0; |
| 311 | htab->entries = nentries; |
| 312 | } |
| 313 | else |
| 314 | htab->entries = (PTR *) xcalloc (htab->size, sizeof (PTR *)); |
| 315 | |
| 316 | htab->n_elements -= htab->n_deleted; |
| 317 | htab->n_deleted = 0; |
| 318 | |
| 319 | p = oentries; |
| 320 | do |
| 321 | { |
| 322 | PTR x = *p; |
| 323 | |
| 324 | if (x != EMPTY_ENTRY && x != DELETED_ENTRY) |
| 325 | { |
| 326 | PTR *q = find_empty_slot_for_expand (htab, (*htab->hash_f) (x)); |
| 327 | |
| 328 | *q = x; |
| 329 | } |
| 330 | |
| 331 | p++; |
| 332 | } |
| 333 | while (p < olimit); |
| 334 | |
| 335 | free (oentries); |
| 336 | return 1; |
| 337 | } |
| 338 | |
| 339 | /* This function searches for a hash table entry equal to the given |
| 340 | element. It cannot be used to insert or delete an element. */ |
| 341 | |
| 342 | PTR |
| 343 | htab_find_with_hash (htab, element, hash) |
| 344 | htab_t htab; |
| 345 | const PTR element; |
| 346 | hashval_t hash; |
| 347 | { |
| 348 | unsigned int index; |
| 349 | hashval_t hash2; |
| 350 | size_t size; |
| 351 | PTR entry; |
| 352 | |
| 353 | htab->searches++; |
| 354 | size = htab->size; |
| 355 | index = hash % size; |
| 356 | |
| 357 | entry = htab->entries[index]; |
| 358 | if (entry == EMPTY_ENTRY |
| 359 | || (entry != DELETED_ENTRY && (*htab->eq_f) (entry, element))) |
| 360 | return entry; |
| 361 | |
| 362 | hash2 = 1 + hash % (size - 2); |
| 363 | |
| 364 | for (;;) |
| 365 | { |
| 366 | htab->collisions++; |
| 367 | index += hash2; |
| 368 | if (index >= size) |
| 369 | index -= size; |
| 370 | |
| 371 | entry = htab->entries[index]; |
| 372 | if (entry == EMPTY_ENTRY |
| 373 | || (entry != DELETED_ENTRY && (*htab->eq_f) (entry, element))) |
| 374 | return entry; |
| 375 | } |
| 376 | } |
| 377 | |
| 378 | /* Like htab_find_slot_with_hash, but compute the hash value from the |
| 379 | element. */ |
| 380 | |
| 381 | PTR |
| 382 | htab_find (htab, element) |
| 383 | htab_t htab; |
| 384 | const PTR element; |
| 385 | { |
| 386 | return htab_find_with_hash (htab, element, (*htab->hash_f) (element)); |
| 387 | } |
| 388 | |
| 389 | /* This function searches for a hash table slot containing an entry |
| 390 | equal to the given element. To delete an entry, call this with |
| 391 | INSERT = 0, then call htab_clear_slot on the slot returned (possibly |
| 392 | after doing some checks). To insert an entry, call this with |
| 393 | INSERT = 1, then write the value you want into the returned slot. |
| 394 | When inserting an entry, NULL may be returned if memory allocation |
| 395 | fails. */ |
| 396 | |
| 397 | PTR * |
| 398 | htab_find_slot_with_hash (htab, element, hash, insert) |
| 399 | htab_t htab; |
| 400 | const PTR element; |
| 401 | hashval_t hash; |
| 402 | enum insert_option insert; |
| 403 | { |
| 404 | PTR *first_deleted_slot; |
| 405 | unsigned int index; |
| 406 | hashval_t hash2; |
| 407 | size_t size; |
| 408 | |
| 409 | if (insert == INSERT && htab->size * 3 <= htab->n_elements * 4 |
| 410 | && htab_expand (htab) == 0) |
| 411 | return NULL; |
| 412 | |
| 413 | size = htab->size; |
| 414 | hash2 = 1 + hash % (size - 2); |
| 415 | index = hash % size; |
| 416 | |
| 417 | htab->searches++; |
| 418 | first_deleted_slot = NULL; |
| 419 | |
| 420 | for (;;) |
| 421 | { |
| 422 | PTR entry = htab->entries[index]; |
| 423 | if (entry == EMPTY_ENTRY) |
| 424 | { |
| 425 | if (insert == NO_INSERT) |
| 426 | return NULL; |
| 427 | |
| 428 | htab->n_elements++; |
| 429 | |
| 430 | if (first_deleted_slot) |
| 431 | { |
| 432 | *first_deleted_slot = EMPTY_ENTRY; |
| 433 | return first_deleted_slot; |
| 434 | } |
| 435 | |
| 436 | return &htab->entries[index]; |
| 437 | } |
| 438 | |
| 439 | if (entry == DELETED_ENTRY) |
| 440 | { |
| 441 | if (!first_deleted_slot) |
| 442 | first_deleted_slot = &htab->entries[index]; |
| 443 | } |
| 444 | else if ((*htab->eq_f) (entry, element)) |
| 445 | return &htab->entries[index]; |
| 446 | |
| 447 | htab->collisions++; |
| 448 | index += hash2; |
| 449 | if (index >= size) |
| 450 | index -= size; |
| 451 | } |
| 452 | } |
| 453 | |
| 454 | /* Like htab_find_slot_with_hash, but compute the hash value from the |
| 455 | element. */ |
| 456 | |
| 457 | PTR * |
| 458 | htab_find_slot (htab, element, insert) |
| 459 | htab_t htab; |
| 460 | const PTR element; |
| 461 | enum insert_option insert; |
| 462 | { |
| 463 | return htab_find_slot_with_hash (htab, element, (*htab->hash_f) (element), |
| 464 | insert); |
| 465 | } |
| 466 | |
| 467 | /* This function deletes an element with the given value from hash |
| 468 | table. If there is no matching element in the hash table, this |
| 469 | function does nothing. */ |
| 470 | |
| 471 | void |
| 472 | htab_remove_elt (htab, element) |
| 473 | htab_t htab; |
| 474 | PTR element; |
| 475 | { |
| 476 | PTR *slot; |
| 477 | |
| 478 | slot = htab_find_slot (htab, element, NO_INSERT); |
| 479 | if (*slot == EMPTY_ENTRY) |
| 480 | return; |
| 481 | |
| 482 | if (htab->del_f) |
| 483 | (*htab->del_f) (*slot); |
| 484 | |
| 485 | *slot = DELETED_ENTRY; |
| 486 | htab->n_deleted++; |
| 487 | } |
| 488 | |
| 489 | /* This function clears a specified slot in a hash table. It is |
| 490 | useful when you've already done the lookup and don't want to do it |
| 491 | again. */ |
| 492 | |
| 493 | void |
| 494 | htab_clear_slot (htab, slot) |
| 495 | htab_t htab; |
| 496 | PTR *slot; |
| 497 | { |
| 498 | if (slot < htab->entries || slot >= htab->entries + htab->size |
| 499 | || *slot == EMPTY_ENTRY || *slot == DELETED_ENTRY) |
| 500 | abort (); |
| 501 | |
| 502 | if (htab->del_f) |
| 503 | (*htab->del_f) (*slot); |
| 504 | |
| 505 | *slot = DELETED_ENTRY; |
| 506 | htab->n_deleted++; |
| 507 | } |
| 508 | |
| 509 | /* This function scans over the entire hash table calling |
| 510 | CALLBACK for each live entry. If CALLBACK returns false, |
| 511 | the iteration stops. INFO is passed as CALLBACK's second |
| 512 | argument. */ |
| 513 | |
| 514 | void |
| 515 | htab_traverse (htab, callback, info) |
| 516 | htab_t htab; |
| 517 | htab_trav callback; |
| 518 | PTR info; |
| 519 | { |
| 520 | PTR *slot = htab->entries; |
| 521 | PTR *limit = slot + htab->size; |
| 522 | |
| 523 | do |
| 524 | { |
| 525 | PTR x = *slot; |
| 526 | |
| 527 | if (x != EMPTY_ENTRY && x != DELETED_ENTRY) |
| 528 | if (!(*callback) (slot, info)) |
| 529 | break; |
| 530 | } |
| 531 | while (++slot < limit); |
| 532 | } |
| 533 | |
| 534 | /* Return the current size of given hash table. */ |
| 535 | |
| 536 | size_t |
| 537 | htab_size (htab) |
| 538 | htab_t htab; |
| 539 | { |
| 540 | return htab->size; |
| 541 | } |
| 542 | |
| 543 | /* Return the current number of elements in given hash table. */ |
| 544 | |
| 545 | size_t |
| 546 | htab_elements (htab) |
| 547 | htab_t htab; |
| 548 | { |
| 549 | return htab->n_elements - htab->n_deleted; |
| 550 | } |
| 551 | |
| 552 | /* Return the fraction of fixed collisions during all work with given |
| 553 | hash table. */ |
| 554 | |
| 555 | double |
| 556 | htab_collisions (htab) |
| 557 | htab_t htab; |
| 558 | { |
| 559 | if (htab->searches == 0) |
| 560 | return 0.0; |
| 561 | |
| 562 | return (double) htab->collisions / (double) htab->searches; |
| 563 | } |
| 564 | |
| 565 | /* Hash P as a null-terminated string. |
| 566 | |
| 567 | Copied from gcc/hashtable.c. Zack had the following to say with respect |
| 568 | to applicability, though note that unlike hashtable.c, this hash table |
| 569 | implementation re-hashes rather than chain buckets. |
| 570 | |
| 571 | http://gcc.gnu.org/ml/gcc-patches/2001-08/msg01021.html |
| 572 | From: Zack Weinberg <zackw@panix.com> |
| 573 | Date: Fri, 17 Aug 2001 02:15:56 -0400 |
| 574 | |
| 575 | I got it by extracting all the identifiers from all the source code |
| 576 | I had lying around in mid-1999, and testing many recurrences of |
| 577 | the form "H_n = H_{n-1} * K + c_n * L + M" where K, L, M were either |
| 578 | prime numbers or the appropriate identity. This was the best one. |
| 579 | I don't remember exactly what constituted "best", except I was |
| 580 | looking at bucket-length distributions mostly. |
| 581 | |
| 582 | So it should be very good at hashing identifiers, but might not be |
| 583 | as good at arbitrary strings. |
| 584 | |
| 585 | I'll add that it thoroughly trounces the hash functions recommended |
| 586 | for this use at http://burtleburtle.net/bob/hash/index.html, both |
| 587 | on speed and bucket distribution. I haven't tried it against the |
| 588 | function they just started using for Perl's hashes. */ |
| 589 | |
| 590 | hashval_t |
| 591 | htab_hash_string (p) |
| 592 | const PTR p; |
| 593 | { |
| 594 | const unsigned char *str = (const unsigned char *) p; |
| 595 | hashval_t r = 0; |
| 596 | unsigned char c; |
| 597 | |
| 598 | while ((c = *str++) != 0) |
| 599 | r = r * 67 + c - 113; |
| 600 | |
| 601 | return r; |
| 602 | } |