| 1 | /* An expandable hash tables datatype. |
| 2 | Copyright (C) 1999, 2000, 2001, 2002, 2003 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 | #ifdef HAVE_MALLOC_H |
| 49 | #include <malloc.h> |
| 50 | #endif |
| 51 | |
| 52 | #include <stdio.h> |
| 53 | |
| 54 | #include "libiberty.h" |
| 55 | #include "hashtab.h" |
| 56 | |
| 57 | /* This macro defines reserved value for empty table entry. */ |
| 58 | |
| 59 | #define EMPTY_ENTRY ((PTR) 0) |
| 60 | |
| 61 | /* This macro defines reserved value for table entry which contained |
| 62 | a deleted element. */ |
| 63 | |
| 64 | #define DELETED_ENTRY ((PTR) 1) |
| 65 | |
| 66 | static unsigned long higher_prime_number PARAMS ((unsigned long)); |
| 67 | static hashval_t hash_pointer PARAMS ((const void *)); |
| 68 | static int eq_pointer PARAMS ((const void *, const void *)); |
| 69 | static int htab_expand PARAMS ((htab_t)); |
| 70 | static PTR *find_empty_slot_for_expand PARAMS ((htab_t, hashval_t)); |
| 71 | |
| 72 | /* At some point, we could make these be NULL, and modify the |
| 73 | hash-table routines to handle NULL specially; that would avoid |
| 74 | function-call overhead for the common case of hashing pointers. */ |
| 75 | htab_hash htab_hash_pointer = hash_pointer; |
| 76 | htab_eq htab_eq_pointer = eq_pointer; |
| 77 | |
| 78 | /* The following function returns a nearest prime number which is |
| 79 | greater than N, and near a power of two. */ |
| 80 | |
| 81 | static unsigned long |
| 82 | higher_prime_number (n) |
| 83 | unsigned long n; |
| 84 | { |
| 85 | /* These are primes that are near, but slightly smaller than, a |
| 86 | power of two. */ |
| 87 | static const unsigned long primes[] = { |
| 88 | (unsigned long) 7, |
| 89 | (unsigned long) 13, |
| 90 | (unsigned long) 31, |
| 91 | (unsigned long) 61, |
| 92 | (unsigned long) 127, |
| 93 | (unsigned long) 251, |
| 94 | (unsigned long) 509, |
| 95 | (unsigned long) 1021, |
| 96 | (unsigned long) 2039, |
| 97 | (unsigned long) 4093, |
| 98 | (unsigned long) 8191, |
| 99 | (unsigned long) 16381, |
| 100 | (unsigned long) 32749, |
| 101 | (unsigned long) 65521, |
| 102 | (unsigned long) 131071, |
| 103 | (unsigned long) 262139, |
| 104 | (unsigned long) 524287, |
| 105 | (unsigned long) 1048573, |
| 106 | (unsigned long) 2097143, |
| 107 | (unsigned long) 4194301, |
| 108 | (unsigned long) 8388593, |
| 109 | (unsigned long) 16777213, |
| 110 | (unsigned long) 33554393, |
| 111 | (unsigned long) 67108859, |
| 112 | (unsigned long) 134217689, |
| 113 | (unsigned long) 268435399, |
| 114 | (unsigned long) 536870909, |
| 115 | (unsigned long) 1073741789, |
| 116 | (unsigned long) 2147483647, |
| 117 | /* 4294967291L */ |
| 118 | ((unsigned long) 2147483647) + ((unsigned long) 2147483644), |
| 119 | }; |
| 120 | |
| 121 | const unsigned long *low = &primes[0]; |
| 122 | const unsigned long *high = &primes[sizeof(primes) / sizeof(primes[0])]; |
| 123 | |
| 124 | while (low != high) |
| 125 | { |
| 126 | const unsigned long *mid = low + (high - low) / 2; |
| 127 | if (n > *mid) |
| 128 | low = mid + 1; |
| 129 | else |
| 130 | high = mid; |
| 131 | } |
| 132 | |
| 133 | /* If we've run out of primes, abort. */ |
| 134 | if (n > *low) |
| 135 | { |
| 136 | fprintf (stderr, "Cannot find prime bigger than %lu\n", n); |
| 137 | abort (); |
| 138 | } |
| 139 | |
| 140 | return *low; |
| 141 | } |
| 142 | |
| 143 | /* Returns a hash code for P. */ |
| 144 | |
| 145 | static hashval_t |
| 146 | hash_pointer (p) |
| 147 | const PTR p; |
| 148 | { |
| 149 | return (hashval_t) ((long)p >> 3); |
| 150 | } |
| 151 | |
| 152 | /* Returns non-zero if P1 and P2 are equal. */ |
| 153 | |
| 154 | static int |
| 155 | eq_pointer (p1, p2) |
| 156 | const PTR p1; |
| 157 | const PTR p2; |
| 158 | { |
| 159 | return p1 == p2; |
| 160 | } |
| 161 | |
| 162 | /* This function creates table with length slightly longer than given |
| 163 | source length. Created hash table is initiated as empty (all the |
| 164 | hash table entries are EMPTY_ENTRY). The function returns the |
| 165 | created hash table, or NULL if memory allocation fails. */ |
| 166 | |
| 167 | htab_t |
| 168 | htab_create_alloc (size, hash_f, eq_f, del_f, alloc_f, free_f) |
| 169 | size_t size; |
| 170 | htab_hash hash_f; |
| 171 | htab_eq eq_f; |
| 172 | htab_del del_f; |
| 173 | htab_alloc alloc_f; |
| 174 | htab_free free_f; |
| 175 | { |
| 176 | htab_t result; |
| 177 | |
| 178 | size = higher_prime_number (size); |
| 179 | result = (htab_t) (*alloc_f) (1, sizeof (struct htab)); |
| 180 | if (result == NULL) |
| 181 | return NULL; |
| 182 | result->entries = (PTR *) (*alloc_f) (size, sizeof (PTR)); |
| 183 | if (result->entries == NULL) |
| 184 | { |
| 185 | if (free_f != NULL) |
| 186 | (*free_f) (result); |
| 187 | return NULL; |
| 188 | } |
| 189 | result->size = size; |
| 190 | result->hash_f = hash_f; |
| 191 | result->eq_f = eq_f; |
| 192 | result->del_f = del_f; |
| 193 | result->alloc_f = alloc_f; |
| 194 | result->free_f = free_f; |
| 195 | return result; |
| 196 | } |
| 197 | |
| 198 | /* As above, but use the variants of alloc_f and free_f which accept |
| 199 | an extra argument. */ |
| 200 | |
| 201 | htab_t |
| 202 | htab_create_alloc_ex (size, hash_f, eq_f, del_f, alloc_arg, alloc_f, |
| 203 | free_f) |
| 204 | size_t size; |
| 205 | htab_hash hash_f; |
| 206 | htab_eq eq_f; |
| 207 | htab_del del_f; |
| 208 | PTR alloc_arg; |
| 209 | htab_alloc_with_arg alloc_f; |
| 210 | htab_free_with_arg free_f; |
| 211 | { |
| 212 | htab_t result; |
| 213 | |
| 214 | size = higher_prime_number (size); |
| 215 | result = (htab_t) (*alloc_f) (alloc_arg, 1, sizeof (struct htab)); |
| 216 | if (result == NULL) |
| 217 | return NULL; |
| 218 | result->entries = (PTR *) (*alloc_f) (alloc_arg, size, sizeof (PTR)); |
| 219 | if (result->entries == NULL) |
| 220 | { |
| 221 | if (free_f != NULL) |
| 222 | (*free_f) (alloc_arg, result); |
| 223 | return NULL; |
| 224 | } |
| 225 | result->size = size; |
| 226 | result->hash_f = hash_f; |
| 227 | result->eq_f = eq_f; |
| 228 | result->del_f = del_f; |
| 229 | result->alloc_arg = alloc_arg; |
| 230 | result->alloc_with_arg_f = alloc_f; |
| 231 | result->free_with_arg_f = free_f; |
| 232 | return result; |
| 233 | } |
| 234 | |
| 235 | /* Update the function pointers and allocation parameter in the htab_t. */ |
| 236 | |
| 237 | void |
| 238 | htab_set_functions_ex (htab, hash_f, eq_f, del_f, alloc_arg, alloc_f, free_f) |
| 239 | htab_t htab; |
| 240 | htab_hash hash_f; |
| 241 | htab_eq eq_f; |
| 242 | htab_del del_f; |
| 243 | PTR alloc_arg; |
| 244 | htab_alloc_with_arg alloc_f; |
| 245 | htab_free_with_arg free_f; |
| 246 | { |
| 247 | htab->hash_f = hash_f; |
| 248 | htab->eq_f = eq_f; |
| 249 | htab->del_f = del_f; |
| 250 | htab->alloc_arg = alloc_arg; |
| 251 | htab->alloc_with_arg_f = alloc_f; |
| 252 | htab->free_with_arg_f = free_f; |
| 253 | } |
| 254 | |
| 255 | /* These functions exist solely for backward compatibility. */ |
| 256 | |
| 257 | #undef htab_create |
| 258 | htab_t |
| 259 | htab_create (size, hash_f, eq_f, del_f) |
| 260 | size_t size; |
| 261 | htab_hash hash_f; |
| 262 | htab_eq eq_f; |
| 263 | htab_del del_f; |
| 264 | { |
| 265 | return htab_create_alloc (size, hash_f, eq_f, del_f, xcalloc, free); |
| 266 | } |
| 267 | |
| 268 | htab_t |
| 269 | htab_try_create (size, hash_f, eq_f, del_f) |
| 270 | size_t size; |
| 271 | htab_hash hash_f; |
| 272 | htab_eq eq_f; |
| 273 | htab_del del_f; |
| 274 | { |
| 275 | return htab_create_alloc (size, hash_f, eq_f, del_f, calloc, free); |
| 276 | } |
| 277 | |
| 278 | /* This function frees all memory allocated for given hash table. |
| 279 | Naturally the hash table must already exist. */ |
| 280 | |
| 281 | void |
| 282 | htab_delete (htab) |
| 283 | htab_t htab; |
| 284 | { |
| 285 | int i; |
| 286 | |
| 287 | if (htab->del_f) |
| 288 | for (i = htab->size - 1; i >= 0; i--) |
| 289 | if (htab->entries[i] != EMPTY_ENTRY |
| 290 | && htab->entries[i] != DELETED_ENTRY) |
| 291 | (*htab->del_f) (htab->entries[i]); |
| 292 | |
| 293 | if (htab->free_f != NULL) |
| 294 | { |
| 295 | (*htab->free_f) (htab->entries); |
| 296 | (*htab->free_f) (htab); |
| 297 | } |
| 298 | else if (htab->free_with_arg_f != NULL) |
| 299 | { |
| 300 | (*htab->free_with_arg_f) (htab->alloc_arg, htab->entries); |
| 301 | (*htab->free_with_arg_f) (htab->alloc_arg, htab); |
| 302 | } |
| 303 | } |
| 304 | |
| 305 | /* This function clears all entries in the given hash table. */ |
| 306 | |
| 307 | void |
| 308 | htab_empty (htab) |
| 309 | htab_t htab; |
| 310 | { |
| 311 | int i; |
| 312 | |
| 313 | if (htab->del_f) |
| 314 | for (i = htab->size - 1; i >= 0; i--) |
| 315 | if (htab->entries[i] != EMPTY_ENTRY |
| 316 | && htab->entries[i] != DELETED_ENTRY) |
| 317 | (*htab->del_f) (htab->entries[i]); |
| 318 | |
| 319 | memset (htab->entries, 0, htab->size * sizeof (PTR)); |
| 320 | } |
| 321 | |
| 322 | /* Similar to htab_find_slot, but without several unwanted side effects: |
| 323 | - Does not call htab->eq_f when it finds an existing entry. |
| 324 | - Does not change the count of elements/searches/collisions in the |
| 325 | hash table. |
| 326 | This function also assumes there are no deleted entries in the table. |
| 327 | HASH is the hash value for the element to be inserted. */ |
| 328 | |
| 329 | static PTR * |
| 330 | find_empty_slot_for_expand (htab, hash) |
| 331 | htab_t htab; |
| 332 | hashval_t hash; |
| 333 | { |
| 334 | size_t size = htab->size; |
| 335 | unsigned int index = hash % size; |
| 336 | PTR *slot = htab->entries + index; |
| 337 | hashval_t hash2; |
| 338 | |
| 339 | if (*slot == EMPTY_ENTRY) |
| 340 | return slot; |
| 341 | else if (*slot == DELETED_ENTRY) |
| 342 | abort (); |
| 343 | |
| 344 | hash2 = 1 + hash % (size - 2); |
| 345 | for (;;) |
| 346 | { |
| 347 | index += hash2; |
| 348 | if (index >= size) |
| 349 | index -= size; |
| 350 | |
| 351 | slot = htab->entries + index; |
| 352 | if (*slot == EMPTY_ENTRY) |
| 353 | return slot; |
| 354 | else if (*slot == DELETED_ENTRY) |
| 355 | abort (); |
| 356 | } |
| 357 | } |
| 358 | |
| 359 | /* The following function changes size of memory allocated for the |
| 360 | entries and repeatedly inserts the table elements. The occupancy |
| 361 | of the table after the call will be about 50%. Naturally the hash |
| 362 | table must already exist. Remember also that the place of the |
| 363 | table entries is changed. If memory allocation failures are allowed, |
| 364 | this function will return zero, indicating that the table could not be |
| 365 | expanded. If all goes well, it will return a non-zero value. */ |
| 366 | |
| 367 | static int |
| 368 | htab_expand (htab) |
| 369 | htab_t htab; |
| 370 | { |
| 371 | PTR *oentries; |
| 372 | PTR *olimit; |
| 373 | PTR *p; |
| 374 | PTR *nentries; |
| 375 | size_t nsize; |
| 376 | |
| 377 | oentries = htab->entries; |
| 378 | olimit = oentries + htab->size; |
| 379 | |
| 380 | /* Resize only when table after removal of unused elements is either |
| 381 | too full or too empty. */ |
| 382 | if ((htab->n_elements - htab->n_deleted) * 2 > htab->size |
| 383 | || ((htab->n_elements - htab->n_deleted) * 8 < htab->size |
| 384 | && htab->size > 32)) |
| 385 | nsize = higher_prime_number ((htab->n_elements - htab->n_deleted) * 2); |
| 386 | else |
| 387 | nsize = htab->size; |
| 388 | |
| 389 | if (htab->alloc_with_arg_f != NULL) |
| 390 | nentries = (PTR *) (*htab->alloc_with_arg_f) (htab->alloc_arg, nsize, |
| 391 | sizeof (PTR *)); |
| 392 | else |
| 393 | nentries = (PTR *) (*htab->alloc_f) (nsize, sizeof (PTR *)); |
| 394 | if (nentries == NULL) |
| 395 | return 0; |
| 396 | htab->entries = nentries; |
| 397 | htab->size = nsize; |
| 398 | |
| 399 | htab->n_elements -= htab->n_deleted; |
| 400 | htab->n_deleted = 0; |
| 401 | |
| 402 | p = oentries; |
| 403 | do |
| 404 | { |
| 405 | PTR x = *p; |
| 406 | |
| 407 | if (x != EMPTY_ENTRY && x != DELETED_ENTRY) |
| 408 | { |
| 409 | PTR *q = find_empty_slot_for_expand (htab, (*htab->hash_f) (x)); |
| 410 | |
| 411 | *q = x; |
| 412 | } |
| 413 | |
| 414 | p++; |
| 415 | } |
| 416 | while (p < olimit); |
| 417 | |
| 418 | if (htab->free_f != NULL) |
| 419 | (*htab->free_f) (oentries); |
| 420 | else if (htab->free_with_arg_f != NULL) |
| 421 | (*htab->free_with_arg_f) (htab->alloc_arg, oentries); |
| 422 | return 1; |
| 423 | } |
| 424 | |
| 425 | /* This function searches for a hash table entry equal to the given |
| 426 | element. It cannot be used to insert or delete an element. */ |
| 427 | |
| 428 | PTR |
| 429 | htab_find_with_hash (htab, element, hash) |
| 430 | htab_t htab; |
| 431 | const PTR element; |
| 432 | hashval_t hash; |
| 433 | { |
| 434 | unsigned int index; |
| 435 | hashval_t hash2; |
| 436 | size_t size; |
| 437 | PTR entry; |
| 438 | |
| 439 | htab->searches++; |
| 440 | size = htab->size; |
| 441 | index = hash % size; |
| 442 | |
| 443 | entry = htab->entries[index]; |
| 444 | if (entry == EMPTY_ENTRY |
| 445 | || (entry != DELETED_ENTRY && (*htab->eq_f) (entry, element))) |
| 446 | return entry; |
| 447 | |
| 448 | hash2 = 1 + hash % (size - 2); |
| 449 | |
| 450 | for (;;) |
| 451 | { |
| 452 | htab->collisions++; |
| 453 | index += hash2; |
| 454 | if (index >= size) |
| 455 | index -= size; |
| 456 | |
| 457 | entry = htab->entries[index]; |
| 458 | if (entry == EMPTY_ENTRY |
| 459 | || (entry != DELETED_ENTRY && (*htab->eq_f) (entry, element))) |
| 460 | return entry; |
| 461 | } |
| 462 | } |
| 463 | |
| 464 | /* Like htab_find_slot_with_hash, but compute the hash value from the |
| 465 | element. */ |
| 466 | |
| 467 | PTR |
| 468 | htab_find (htab, element) |
| 469 | htab_t htab; |
| 470 | const PTR element; |
| 471 | { |
| 472 | return htab_find_with_hash (htab, element, (*htab->hash_f) (element)); |
| 473 | } |
| 474 | |
| 475 | /* This function searches for a hash table slot containing an entry |
| 476 | equal to the given element. To delete an entry, call this with |
| 477 | INSERT = 0, then call htab_clear_slot on the slot returned (possibly |
| 478 | after doing some checks). To insert an entry, call this with |
| 479 | INSERT = 1, then write the value you want into the returned slot. |
| 480 | When inserting an entry, NULL may be returned if memory allocation |
| 481 | fails. */ |
| 482 | |
| 483 | PTR * |
| 484 | htab_find_slot_with_hash (htab, element, hash, insert) |
| 485 | htab_t htab; |
| 486 | const PTR element; |
| 487 | hashval_t hash; |
| 488 | enum insert_option insert; |
| 489 | { |
| 490 | PTR *first_deleted_slot; |
| 491 | unsigned int index; |
| 492 | hashval_t hash2; |
| 493 | size_t size; |
| 494 | PTR entry; |
| 495 | |
| 496 | if (insert == INSERT && htab->size * 3 <= htab->n_elements * 4 |
| 497 | && htab_expand (htab) == 0) |
| 498 | return NULL; |
| 499 | |
| 500 | size = htab->size; |
| 501 | index = hash % size; |
| 502 | |
| 503 | htab->searches++; |
| 504 | first_deleted_slot = NULL; |
| 505 | |
| 506 | entry = htab->entries[index]; |
| 507 | if (entry == EMPTY_ENTRY) |
| 508 | goto empty_entry; |
| 509 | else if (entry == DELETED_ENTRY) |
| 510 | first_deleted_slot = &htab->entries[index]; |
| 511 | else if ((*htab->eq_f) (entry, element)) |
| 512 | return &htab->entries[index]; |
| 513 | |
| 514 | hash2 = 1 + hash % (size - 2); |
| 515 | for (;;) |
| 516 | { |
| 517 | htab->collisions++; |
| 518 | index += hash2; |
| 519 | if (index >= size) |
| 520 | index -= size; |
| 521 | |
| 522 | entry = htab->entries[index]; |
| 523 | if (entry == EMPTY_ENTRY) |
| 524 | goto empty_entry; |
| 525 | else if (entry == DELETED_ENTRY) |
| 526 | { |
| 527 | if (!first_deleted_slot) |
| 528 | first_deleted_slot = &htab->entries[index]; |
| 529 | } |
| 530 | else if ((*htab->eq_f) (entry, element)) |
| 531 | return &htab->entries[index]; |
| 532 | } |
| 533 | |
| 534 | empty_entry: |
| 535 | if (insert == NO_INSERT) |
| 536 | return NULL; |
| 537 | |
| 538 | if (first_deleted_slot) |
| 539 | { |
| 540 | htab->n_deleted--; |
| 541 | *first_deleted_slot = EMPTY_ENTRY; |
| 542 | return first_deleted_slot; |
| 543 | } |
| 544 | |
| 545 | htab->n_elements++; |
| 546 | return &htab->entries[index]; |
| 547 | } |
| 548 | |
| 549 | /* Like htab_find_slot_with_hash, but compute the hash value from the |
| 550 | element. */ |
| 551 | |
| 552 | PTR * |
| 553 | htab_find_slot (htab, element, insert) |
| 554 | htab_t htab; |
| 555 | const PTR element; |
| 556 | enum insert_option insert; |
| 557 | { |
| 558 | return htab_find_slot_with_hash (htab, element, (*htab->hash_f) (element), |
| 559 | insert); |
| 560 | } |
| 561 | |
| 562 | /* This function deletes an element with the given value from hash |
| 563 | table. If there is no matching element in the hash table, this |
| 564 | function does nothing. */ |
| 565 | |
| 566 | void |
| 567 | htab_remove_elt (htab, element) |
| 568 | htab_t htab; |
| 569 | PTR element; |
| 570 | { |
| 571 | PTR *slot; |
| 572 | |
| 573 | slot = htab_find_slot (htab, element, NO_INSERT); |
| 574 | if (*slot == EMPTY_ENTRY) |
| 575 | return; |
| 576 | |
| 577 | if (htab->del_f) |
| 578 | (*htab->del_f) (*slot); |
| 579 | |
| 580 | *slot = DELETED_ENTRY; |
| 581 | htab->n_deleted++; |
| 582 | } |
| 583 | |
| 584 | /* This function clears a specified slot in a hash table. It is |
| 585 | useful when you've already done the lookup and don't want to do it |
| 586 | again. */ |
| 587 | |
| 588 | void |
| 589 | htab_clear_slot (htab, slot) |
| 590 | htab_t htab; |
| 591 | PTR *slot; |
| 592 | { |
| 593 | if (slot < htab->entries || slot >= htab->entries + htab->size |
| 594 | || *slot == EMPTY_ENTRY || *slot == DELETED_ENTRY) |
| 595 | abort (); |
| 596 | |
| 597 | if (htab->del_f) |
| 598 | (*htab->del_f) (*slot); |
| 599 | |
| 600 | *slot = DELETED_ENTRY; |
| 601 | htab->n_deleted++; |
| 602 | } |
| 603 | |
| 604 | /* This function scans over the entire hash table calling |
| 605 | CALLBACK for each live entry. If CALLBACK returns false, |
| 606 | the iteration stops. INFO is passed as CALLBACK's second |
| 607 | argument. */ |
| 608 | |
| 609 | void |
| 610 | htab_traverse_noresize (htab, callback, info) |
| 611 | htab_t htab; |
| 612 | htab_trav callback; |
| 613 | PTR info; |
| 614 | { |
| 615 | PTR *slot; |
| 616 | PTR *limit; |
| 617 | |
| 618 | slot = htab->entries; |
| 619 | limit = slot + htab->size; |
| 620 | |
| 621 | do |
| 622 | { |
| 623 | PTR x = *slot; |
| 624 | |
| 625 | if (x != EMPTY_ENTRY && x != DELETED_ENTRY) |
| 626 | if (!(*callback) (slot, info)) |
| 627 | break; |
| 628 | } |
| 629 | while (++slot < limit); |
| 630 | } |
| 631 | |
| 632 | /* Like htab_traverse_noresize, but does resize the table when it is |
| 633 | too empty to improve effectivity of subsequent calls. */ |
| 634 | |
| 635 | void |
| 636 | htab_traverse (htab, callback, info) |
| 637 | htab_t htab; |
| 638 | htab_trav callback; |
| 639 | PTR info; |
| 640 | { |
| 641 | if ((htab->n_elements - htab->n_deleted) * 8 < htab->size) |
| 642 | htab_expand (htab); |
| 643 | |
| 644 | htab_traverse_noresize (htab, callback, info); |
| 645 | } |
| 646 | |
| 647 | /* Return the current size of given hash table. */ |
| 648 | |
| 649 | size_t |
| 650 | htab_size (htab) |
| 651 | htab_t htab; |
| 652 | { |
| 653 | return htab->size; |
| 654 | } |
| 655 | |
| 656 | /* Return the current number of elements in given hash table. */ |
| 657 | |
| 658 | size_t |
| 659 | htab_elements (htab) |
| 660 | htab_t htab; |
| 661 | { |
| 662 | return htab->n_elements - htab->n_deleted; |
| 663 | } |
| 664 | |
| 665 | /* Return the fraction of fixed collisions during all work with given |
| 666 | hash table. */ |
| 667 | |
| 668 | double |
| 669 | htab_collisions (htab) |
| 670 | htab_t htab; |
| 671 | { |
| 672 | if (htab->searches == 0) |
| 673 | return 0.0; |
| 674 | |
| 675 | return (double) htab->collisions / (double) htab->searches; |
| 676 | } |
| 677 | |
| 678 | /* Hash P as a null-terminated string. |
| 679 | |
| 680 | Copied from gcc/hashtable.c. Zack had the following to say with respect |
| 681 | to applicability, though note that unlike hashtable.c, this hash table |
| 682 | implementation re-hashes rather than chain buckets. |
| 683 | |
| 684 | http://gcc.gnu.org/ml/gcc-patches/2001-08/msg01021.html |
| 685 | From: Zack Weinberg <zackw@panix.com> |
| 686 | Date: Fri, 17 Aug 2001 02:15:56 -0400 |
| 687 | |
| 688 | I got it by extracting all the identifiers from all the source code |
| 689 | I had lying around in mid-1999, and testing many recurrences of |
| 690 | the form "H_n = H_{n-1} * K + c_n * L + M" where K, L, M were either |
| 691 | prime numbers or the appropriate identity. This was the best one. |
| 692 | I don't remember exactly what constituted "best", except I was |
| 693 | looking at bucket-length distributions mostly. |
| 694 | |
| 695 | So it should be very good at hashing identifiers, but might not be |
| 696 | as good at arbitrary strings. |
| 697 | |
| 698 | I'll add that it thoroughly trounces the hash functions recommended |
| 699 | for this use at http://burtleburtle.net/bob/hash/index.html, both |
| 700 | on speed and bucket distribution. I haven't tried it against the |
| 701 | function they just started using for Perl's hashes. */ |
| 702 | |
| 703 | hashval_t |
| 704 | htab_hash_string (p) |
| 705 | const PTR p; |
| 706 | { |
| 707 | const unsigned char *str = (const unsigned char *) p; |
| 708 | hashval_t r = 0; |
| 709 | unsigned char c; |
| 710 | |
| 711 | while ((c = *str++) != 0) |
| 712 | r = r * 67 + c - 113; |
| 713 | |
| 714 | return r; |
| 715 | } |
| 716 | |
| 717 | /* DERIVED FROM: |
| 718 | -------------------------------------------------------------------- |
| 719 | lookup2.c, by Bob Jenkins, December 1996, Public Domain. |
| 720 | hash(), hash2(), hash3, and mix() are externally useful functions. |
| 721 | Routines to test the hash are included if SELF_TEST is defined. |
| 722 | You can use this free for any purpose. It has no warranty. |
| 723 | -------------------------------------------------------------------- |
| 724 | */ |
| 725 | |
| 726 | /* |
| 727 | -------------------------------------------------------------------- |
| 728 | mix -- mix 3 32-bit values reversibly. |
| 729 | For every delta with one or two bit set, and the deltas of all three |
| 730 | high bits or all three low bits, whether the original value of a,b,c |
| 731 | is almost all zero or is uniformly distributed, |
| 732 | * If mix() is run forward or backward, at least 32 bits in a,b,c |
| 733 | have at least 1/4 probability of changing. |
| 734 | * If mix() is run forward, every bit of c will change between 1/3 and |
| 735 | 2/3 of the time. (Well, 22/100 and 78/100 for some 2-bit deltas.) |
| 736 | mix() was built out of 36 single-cycle latency instructions in a |
| 737 | structure that could supported 2x parallelism, like so: |
| 738 | a -= b; |
| 739 | a -= c; x = (c>>13); |
| 740 | b -= c; a ^= x; |
| 741 | b -= a; x = (a<<8); |
| 742 | c -= a; b ^= x; |
| 743 | c -= b; x = (b>>13); |
| 744 | ... |
| 745 | Unfortunately, superscalar Pentiums and Sparcs can't take advantage |
| 746 | of that parallelism. They've also turned some of those single-cycle |
| 747 | latency instructions into multi-cycle latency instructions. Still, |
| 748 | this is the fastest good hash I could find. There were about 2^^68 |
| 749 | to choose from. I only looked at a billion or so. |
| 750 | -------------------------------------------------------------------- |
| 751 | */ |
| 752 | /* same, but slower, works on systems that might have 8 byte hashval_t's */ |
| 753 | #define mix(a,b,c) \ |
| 754 | { \ |
| 755 | a -= b; a -= c; a ^= (c>>13); \ |
| 756 | b -= c; b -= a; b ^= (a<< 8); \ |
| 757 | c -= a; c -= b; c ^= ((b&0xffffffff)>>13); \ |
| 758 | a -= b; a -= c; a ^= ((c&0xffffffff)>>12); \ |
| 759 | b -= c; b -= a; b = (b ^ (a<<16)) & 0xffffffff; \ |
| 760 | c -= a; c -= b; c = (c ^ (b>> 5)) & 0xffffffff; \ |
| 761 | a -= b; a -= c; a = (a ^ (c>> 3)) & 0xffffffff; \ |
| 762 | b -= c; b -= a; b = (b ^ (a<<10)) & 0xffffffff; \ |
| 763 | c -= a; c -= b; c = (c ^ (b>>15)) & 0xffffffff; \ |
| 764 | } |
| 765 | |
| 766 | /* |
| 767 | -------------------------------------------------------------------- |
| 768 | hash() -- hash a variable-length key into a 32-bit value |
| 769 | k : the key (the unaligned variable-length array of bytes) |
| 770 | len : the length of the key, counting by bytes |
| 771 | level : can be any 4-byte value |
| 772 | Returns a 32-bit value. Every bit of the key affects every bit of |
| 773 | the return value. Every 1-bit and 2-bit delta achieves avalanche. |
| 774 | About 36+6len instructions. |
| 775 | |
| 776 | The best hash table sizes are powers of 2. There is no need to do |
| 777 | mod a prime (mod is sooo slow!). If you need less than 32 bits, |
| 778 | use a bitmask. For example, if you need only 10 bits, do |
| 779 | h = (h & hashmask(10)); |
| 780 | In which case, the hash table should have hashsize(10) elements. |
| 781 | |
| 782 | If you are hashing n strings (ub1 **)k, do it like this: |
| 783 | for (i=0, h=0; i<n; ++i) h = hash( k[i], len[i], h); |
| 784 | |
| 785 | By Bob Jenkins, 1996. bob_jenkins@burtleburtle.net. You may use this |
| 786 | code any way you wish, private, educational, or commercial. It's free. |
| 787 | |
| 788 | See http://burtleburtle.net/bob/hash/evahash.html |
| 789 | Use for hash table lookup, or anything where one collision in 2^32 is |
| 790 | acceptable. Do NOT use for cryptographic purposes. |
| 791 | -------------------------------------------------------------------- |
| 792 | */ |
| 793 | |
| 794 | hashval_t iterative_hash (k_in, length, initval) |
| 795 | const PTR k_in; /* the key */ |
| 796 | register size_t length; /* the length of the key */ |
| 797 | register hashval_t initval; /* the previous hash, or an arbitrary value */ |
| 798 | { |
| 799 | register const unsigned char *k = (const unsigned char *)k_in; |
| 800 | register hashval_t a,b,c,len; |
| 801 | |
| 802 | /* Set up the internal state */ |
| 803 | len = length; |
| 804 | a = b = 0x9e3779b9; /* the golden ratio; an arbitrary value */ |
| 805 | c = initval; /* the previous hash value */ |
| 806 | |
| 807 | /*---------------------------------------- handle most of the key */ |
| 808 | #ifndef WORDS_BIGENDIAN |
| 809 | /* On a little-endian machine, if the data is 4-byte aligned we can hash |
| 810 | by word for better speed. This gives nondeterministic results on |
| 811 | big-endian machines. */ |
| 812 | if (sizeof (hashval_t) == 4 && (((size_t)k)&3) == 0) |
| 813 | while (len >= 12) /* aligned */ |
| 814 | { |
| 815 | a += *(hashval_t *)(k+0); |
| 816 | b += *(hashval_t *)(k+4); |
| 817 | c += *(hashval_t *)(k+8); |
| 818 | mix(a,b,c); |
| 819 | k += 12; len -= 12; |
| 820 | } |
| 821 | else /* unaligned */ |
| 822 | #endif |
| 823 | while (len >= 12) |
| 824 | { |
| 825 | a += (k[0] +((hashval_t)k[1]<<8) +((hashval_t)k[2]<<16) +((hashval_t)k[3]<<24)); |
| 826 | b += (k[4] +((hashval_t)k[5]<<8) +((hashval_t)k[6]<<16) +((hashval_t)k[7]<<24)); |
| 827 | c += (k[8] +((hashval_t)k[9]<<8) +((hashval_t)k[10]<<16)+((hashval_t)k[11]<<24)); |
| 828 | mix(a,b,c); |
| 829 | k += 12; len -= 12; |
| 830 | } |
| 831 | |
| 832 | /*------------------------------------- handle the last 11 bytes */ |
| 833 | c += length; |
| 834 | switch(len) /* all the case statements fall through */ |
| 835 | { |
| 836 | case 11: c+=((hashval_t)k[10]<<24); |
| 837 | case 10: c+=((hashval_t)k[9]<<16); |
| 838 | case 9 : c+=((hashval_t)k[8]<<8); |
| 839 | /* the first byte of c is reserved for the length */ |
| 840 | case 8 : b+=((hashval_t)k[7]<<24); |
| 841 | case 7 : b+=((hashval_t)k[6]<<16); |
| 842 | case 6 : b+=((hashval_t)k[5]<<8); |
| 843 | case 5 : b+=k[4]; |
| 844 | case 4 : a+=((hashval_t)k[3]<<24); |
| 845 | case 3 : a+=((hashval_t)k[2]<<16); |
| 846 | case 2 : a+=((hashval_t)k[1]<<8); |
| 847 | case 1 : a+=k[0]; |
| 848 | /* case 0: nothing left to add */ |
| 849 | } |
| 850 | mix(a,b,c); |
| 851 | /*-------------------------------------------- report the result */ |
| 852 | return c; |
| 853 | } |