hash.c revision 1.2 1 /* $NetBSD: hash.c,v 1.2 2002/06/30 14:17:44 lukem Exp $ */
2
3 /*
4 * Copyright (c) 1988, 1989, 1990 The Regents of the University of California.
5 * Copyright (c) 1988, 1989 by Adam de Boor
6 * Copyright (c) 1989 by Berkeley Softworks
7 * All rights reserved.
8 *
9 * This code is derived from software contributed to Berkeley by
10 * Adam de Boor.
11 *
12 * Redistribution and use in source and binary forms, with or without
13 * modification, are permitted provided that the following conditions
14 * are met:
15 * 1. Redistributions of source code must retain the above copyright
16 * notice, this list of conditions and the following disclaimer.
17 * 2. Redistributions in binary form must reproduce the above copyright
18 * notice, this list of conditions and the following disclaimer in the
19 * documentation and/or other materials provided with the distribution.
20 * 3. All advertising materials mentioning features or use of this software
21 * must display the following acknowledgement:
22 * This product includes software developed by the University of
23 * California, Berkeley and its contributors.
24 * 4. Neither the name of the University nor the names of its contributors
25 * may be used to endorse or promote products derived from this software
26 * without specific prior written permission.
27 *
28 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
29 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
30 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
31 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
32 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
33 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
34 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
35 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
36 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
37 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
38 * SUCH DAMAGE.
39 */
40
41 #ifdef MAKE_BOOTSTRAP
42 static char rcsid[] = "$NetBSD: hash.c,v 1.2 2002/06/30 14:17:44 lukem Exp $";
43 #else
44 #include <sys/cdefs.h>
45 #ifndef lint
46 #if 0
47 static char sccsid[] = "@(#)hash.c 8.1 (Berkeley) 6/6/93";
48 #else
49 __RCSID("$NetBSD: hash.c,v 1.2 2002/06/30 14:17:44 lukem Exp $");
50 #endif
51 #endif /* not lint */
52 #endif
53
54 #include <sys/types.h>
55
56 #include <stdlib.h>
57 #include <string.h>
58 #include <unistd.h>
59
60 /* hash.c --
61 *
62 * This module contains routines to manipulate a hash table.
63 * See hash.h for a definition of the structure of the hash
64 * table. Hash tables grow automatically as the amount of
65 * information increases.
66 */
67 #include "hash.h"
68 #include "ealloc.h"
69
70 /*
71 * Forward references to local procedures that are used before they're
72 * defined:
73 */
74
75 static void RebuildTable(Hash_Table *);
76
77 /*
78 * The following defines the ratio of # entries to # buckets
79 * at which we rebuild the table to make it larger.
80 */
81
82 #define rebuildLimit 8
83
84 /*
85 *---------------------------------------------------------
86 *
87 * Hash_InitTable --
88 *
89 * This routine just sets up the hash table.
90 *
91 * Input:
92 * t Structure to use to hold table.
93 * numBuckets How many buckets to create for starters. This number
94 * is rounded up to a power of two. If <= 0, a reasonable
95 * default is chosen. The table will grow in size later
96 * as needed.
97 *
98 * Results:
99 * None.
100 *
101 * Side Effects:
102 * Memory is allocated for the initial bucket area.
103 *
104 *---------------------------------------------------------
105 */
106
107 void
108 Hash_InitTable(Hash_Table *t, int numBuckets)
109 {
110 int i;
111 struct Hash_Entry **hp;
112
113 /*
114 * Round up the size to a power of two.
115 */
116 if (numBuckets <= 0)
117 i = 16;
118 else {
119 for (i = 2; i < numBuckets; i <<= 1)
120 continue;
121 }
122 t->numEntries = 0;
123 t->size = i;
124 t->mask = i - 1;
125 t->bucketPtr = hp = (struct Hash_Entry **)emalloc(sizeof(*hp) * i);
126 while (--i >= 0)
127 *hp++ = NULL;
128 }
129
130 /*
131 *---------------------------------------------------------
132 *
133 * Hash_DeleteTable --
134 *
135 * This routine removes everything from a hash table
136 * and frees up the memory space it occupied (except for
137 * the space in the Hash_Table structure).
138 *
139 * Results:
140 * None.
141 *
142 * Side Effects:
143 * Lots of memory is freed up.
144 *
145 *---------------------------------------------------------
146 */
147
148 void
149 Hash_DeleteTable(Hash_Table *t)
150 {
151 struct Hash_Entry **hp, *h, *nexth;
152 int i;
153
154 nexth = NULL;
155 for (hp = t->bucketPtr, i = t->size; --i >= 0;) {
156 for (h = *hp++; h != NULL; h = nexth) {
157 nexth = h->next;
158 free((char *)h);
159 }
160 }
161 free((char *)t->bucketPtr);
162
163 /*
164 * Set up the hash table to cause memory faults on any future access
165 * attempts until re-initialization.
166 */
167 t->bucketPtr = NULL;
168 }
169
170 /*
171 *---------------------------------------------------------
172 *
173 * Hash_FindEntry --
174 *
175 * Searches a hash table for an entry corresponding to key.
176 *
177 * Input:
178 * t Hash table to search.
179 * key A hash key.
180 *
181 * Results:
182 * The return value is a pointer to the entry for key,
183 * if key was present in the table. If key was not
184 * present, NULL is returned.
185 *
186 * Side Effects:
187 * None.
188 *
189 *---------------------------------------------------------
190 */
191
192 Hash_Entry *
193 Hash_FindEntry(Hash_Table *t, char *key)
194 {
195 Hash_Entry *e;
196 unsigned h;
197 char *p;
198
199 for (h = 0, p = key; *p;)
200 h = (h << 5) - h + *p++;
201 p = key;
202 for (e = t->bucketPtr[h & t->mask]; e != NULL; e = e->next)
203 if (e->namehash == h && strcmp(e->name, p) == 0)
204 return (e);
205 return (NULL);
206 }
207
208 /*
209 *---------------------------------------------------------
210 *
211 * Hash_CreateEntry --
212 *
213 * Searches a hash table for an entry corresponding to
214 * key. If no entry is found, then one is created.
215 *
216 * Input:
217 * t Hash table to search.
218 * key A hash key.
219 * newPtr Filled in with 1 if new entry created, 0 otherwise.
220 *
221 * Results:
222 * The return value is a pointer to the entry. If *newPtr
223 * isn't NULL, then *newPtr is filled in with TRUE if a
224 * new entry was created, and FALSE if an entry already existed
225 * with the given key.
226 *
227 * Side Effects:
228 * Memory may be allocated, and the hash buckets may be modified.
229 *---------------------------------------------------------
230 */
231
232 Hash_Entry *
233 Hash_CreateEntry(Hash_Table *t, char *key, int *newPtr)
234 {
235 Hash_Entry *e;
236 unsigned h;
237 char *p;
238 int keylen;
239 struct Hash_Entry **hp;
240
241 /*
242 * Hash the key. As a side effect, save the length (strlen) of the
243 * key in case we need to create the entry.
244 */
245 for (h = 0, p = key; *p;)
246 h = (h << 5) - h + *p++;
247 keylen = p - key;
248 p = key;
249 for (e = t->bucketPtr[h & t->mask]; e != NULL; e = e->next) {
250 if (e->namehash == h && strcmp(e->name, p) == 0) {
251 if (newPtr != NULL)
252 *newPtr = 0;
253 return (e);
254 }
255 }
256
257 /*
258 * The desired entry isn't there. Before allocating a new entry,
259 * expand the table if necessary (and this changes the resulting
260 * bucket chain).
261 */
262 if (t->numEntries >= rebuildLimit * t->size)
263 RebuildTable(t);
264 e = (Hash_Entry *) emalloc(sizeof(*e) + keylen);
265 hp = &t->bucketPtr[h & t->mask];
266 e->next = *hp;
267 *hp = e;
268 e->clientData = NULL;
269 e->namehash = h;
270 (void) strcpy(e->name, p);
271 t->numEntries++;
272
273 if (newPtr != NULL)
274 *newPtr = 1;
275 return (e);
276 }
277
278 /*
279 *---------------------------------------------------------
280 *
281 * Hash_DeleteEntry --
282 *
283 * Delete the given hash table entry and free memory associated with
284 * it.
285 *
286 * Results:
287 * None.
288 *
289 * Side Effects:
290 * Hash chain that entry lives in is modified and memory is freed.
291 *
292 *---------------------------------------------------------
293 */
294
295 void
296 Hash_DeleteEntry(Hash_Table *t, Hash_Entry *e)
297 {
298 Hash_Entry **hp, *p;
299
300 if (e == NULL)
301 return;
302 for (hp = &t->bucketPtr[e->namehash & t->mask];
303 (p = *hp) != NULL; hp = &p->next) {
304 if (p == e) {
305 *hp = p->next;
306 free((char *)p);
307 t->numEntries--;
308 return;
309 }
310 }
311 (void)write(2, "bad call to Hash_DeleteEntry\n", 29);
312 abort();
313 }
314
315 /*
316 *---------------------------------------------------------
317 *
318 * Hash_EnumFirst --
319 * This procedure sets things up for a complete search
320 * of all entries recorded in the hash table.
321 *
322 * Input:
323 * t Table to be searched.
324 * searchPtr Area in which to keep state about search.
325 *
326 * Results:
327 * The return value is the address of the first entry in
328 * the hash table, or NULL if the table is empty.
329 *
330 * Side Effects:
331 * The information in searchPtr is initialized so that successive
332 * calls to Hash_Next will return successive HashEntry's
333 * from the table.
334 *
335 *---------------------------------------------------------
336 */
337
338 Hash_Entry *
339 Hash_EnumFirst(Hash_Table *t, Hash_Search *searchPtr)
340 {
341
342 searchPtr->tablePtr = t;
343 searchPtr->nextIndex = 0;
344 searchPtr->hashEntryPtr = NULL;
345 return Hash_EnumNext(searchPtr);
346 }
347
348 /*
349 *---------------------------------------------------------
350 *
351 * Hash_EnumNext --
352 * This procedure returns successive entries in the hash table.
353 *
354 * Results:
355 * The return value is a pointer to the next HashEntry
356 * in the table, or NULL when the end of the table is
357 * reached.
358 *
359 * Side Effects:
360 * The information in searchPtr is modified to advance to the
361 * next entry.
362 *
363 *---------------------------------------------------------
364 */
365
366 Hash_Entry *
367 Hash_EnumNext(Hash_Search *searchPtr)
368 {
369 Hash_Entry *e;
370 Hash_Table *t = searchPtr->tablePtr;
371
372 /*
373 * The hashEntryPtr field points to the most recently returned
374 * entry, or is nil if we are starting up. If not nil, we have
375 * to start at the next one in the chain.
376 */
377 e = searchPtr->hashEntryPtr;
378 if (e != NULL)
379 e = e->next;
380 /*
381 * If the chain ran out, or if we are starting up, we need to
382 * find the next nonempty chain.
383 */
384 while (e == NULL) {
385 if (searchPtr->nextIndex >= t->size)
386 return (NULL);
387 e = t->bucketPtr[searchPtr->nextIndex++];
388 }
389 searchPtr->hashEntryPtr = e;
390 return (e);
391 }
392
393 /*
394 *---------------------------------------------------------
395 *
396 * RebuildTable --
397 * This local routine makes a new hash table that
398 * is larger than the old one.
399 *
400 * Results:
401 * None.
402 *
403 * Side Effects:
404 * The entire hash table is moved, so any bucket numbers
405 * from the old table are invalid.
406 *
407 *---------------------------------------------------------
408 */
409
410 static void
411 RebuildTable(Hash_Table *t)
412 {
413 Hash_Entry *e, *next, **hp, **xp;
414 int i, mask;
415 Hash_Entry **oldhp;
416 int oldsize;
417
418 next = NULL;
419 oldhp = t->bucketPtr;
420 oldsize = i = t->size;
421 i <<= 1;
422 t->size = i;
423 t->mask = mask = i - 1;
424 t->bucketPtr = hp = (struct Hash_Entry **) emalloc(sizeof(*hp) * i);
425 while (--i >= 0)
426 *hp++ = NULL;
427 for (hp = oldhp, i = oldsize; --i >= 0;) {
428 for (e = *hp++; e != NULL; e = next) {
429 next = e->next;
430 xp = &t->bucketPtr[e->namehash & mask];
431 e->next = *xp;
432 *xp = e;
433 }
434 }
435 free((char *)oldhp);
436 }
437