xref: /freebsd/contrib/bmake/hash.c (revision 25ecdc7d52770caf1c9b44b5ec11f468f6b636f3)
1 /*	$NetBSD: hash.c,v 1.61 2021/02/01 17:32:10 rillig Exp $	*/
2 
3 /*
4  * Copyright (c) 1988, 1989, 1990 The Regents of the University of California.
5  * All rights reserved.
6  *
7  * This code is derived from software contributed to Berkeley by
8  * Adam de Boor.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  * 1. Redistributions of source code must retain the above copyright
14  *    notice, this list of conditions and the following disclaimer.
15  * 2. Redistributions in binary form must reproduce the above copyright
16  *    notice, this list of conditions and the following disclaimer in the
17  *    documentation and/or other materials provided with the distribution.
18  * 3. Neither the name of the University nor the names of its contributors
19  *    may be used to endorse or promote products derived from this software
20  *    without specific prior written permission.
21  *
22  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32  * SUCH DAMAGE.
33  */
34 
35 /*
36  * Copyright (c) 1988, 1989 by Adam de Boor
37  * Copyright (c) 1989 by Berkeley Softworks
38  * All rights reserved.
39  *
40  * This code is derived from software contributed to Berkeley by
41  * Adam de Boor.
42  *
43  * Redistribution and use in source and binary forms, with or without
44  * modification, are permitted provided that the following conditions
45  * are met:
46  * 1. Redistributions of source code must retain the above copyright
47  *    notice, this list of conditions and the following disclaimer.
48  * 2. Redistributions in binary form must reproduce the above copyright
49  *    notice, this list of conditions and the following disclaimer in the
50  *    documentation and/or other materials provided with the distribution.
51  * 3. All advertising materials mentioning features or use of this software
52  *    must display the following acknowledgement:
53  *	This product includes software developed by the University of
54  *	California, Berkeley and its contributors.
55  * 4. Neither the name of the University nor the names of its contributors
56  *    may be used to endorse or promote products derived from this software
57  *    without specific prior written permission.
58  *
59  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
60  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
61  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
62  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
63  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
64  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
65  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
66  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
67  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
68  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
69  * SUCH DAMAGE.
70  */
71 
72 /* Hash tables with string keys. */
73 
74 #include "make.h"
75 
76 /*	"@(#)hash.c	8.1 (Berkeley) 6/6/93"	*/
77 MAKE_RCSID("$NetBSD: hash.c,v 1.61 2021/02/01 17:32:10 rillig Exp $");
78 
79 /*
80  * The ratio of # entries to # buckets at which we rebuild the table to
81  * make it larger.
82  */
83 #define rebuildLimit 3
84 
85 /* This hash function matches Gosling's Emacs and java.lang.String. */
86 static unsigned int
87 hash(const char *key, size_t *out_keylen)
88 {
89 	unsigned int h;
90 	const char *p;
91 
92 	h = 0;
93 	for (p = key; *p != '\0'; p++)
94 		h = 31 * h + (unsigned char)*p;
95 
96 	if (out_keylen != NULL)
97 		*out_keylen = (size_t)(p - key);
98 	return h;
99 }
100 
101 unsigned int
102 Hash_Hash(const char *key)
103 {
104 	return hash(key, NULL);
105 }
106 
107 static HashEntry *
108 HashTable_Find(HashTable *t, unsigned int h, const char *key)
109 {
110 	HashEntry *e;
111 	unsigned int chainlen = 0;
112 
113 #ifdef DEBUG_HASH_LOOKUP
114 	DEBUG4(HASH, "%s: %p h=%08x key=%s\n", __func__, t, h, key);
115 #endif
116 
117 	for (e = t->buckets[h & t->bucketsMask]; e != NULL; e = e->next) {
118 		chainlen++;
119 		if (e->key_hash == h && strcmp(e->key, key) == 0)
120 			break;
121 	}
122 
123 	if (chainlen > t->maxchain)
124 		t->maxchain = chainlen;
125 
126 	return e;
127 }
128 
129 /* Set up the hash table. */
130 void
131 HashTable_Init(HashTable *t)
132 {
133 	unsigned int n = 16, i;
134 	HashEntry **buckets = bmake_malloc(sizeof *buckets * n);
135 	for (i = 0; i < n; i++)
136 		buckets[i] = NULL;
137 
138 	t->buckets = buckets;
139 	t->bucketsSize = n;
140 	t->numEntries = 0;
141 	t->bucketsMask = n - 1;
142 	t->maxchain = 0;
143 }
144 
145 /*
146  * Remove everything from the hash table and free up the memory for the keys
147  * of the hash table, but not for the values associated to these keys.
148  */
149 void
150 HashTable_Done(HashTable *t)
151 {
152 	HashEntry **buckets = t->buckets;
153 	size_t i, n = t->bucketsSize;
154 
155 	for (i = 0; i < n; i++) {
156 		HashEntry *he = buckets[i];
157 		while (he != NULL) {
158 			HashEntry *next = he->next;
159 			free(he);
160 			he = next;
161 		}
162 	}
163 
164 	free(t->buckets);
165 #ifdef CLEANUP
166 	t->buckets = NULL;
167 #endif
168 }
169 
170 /* Find the entry corresponding to the key, or return NULL. */
171 HashEntry *
172 HashTable_FindEntry(HashTable *t, const char *key)
173 {
174 	unsigned int h = hash(key, NULL);
175 	return HashTable_Find(t, h, key);
176 }
177 
178 /* Find the value corresponding to the key, or return NULL. */
179 void *
180 HashTable_FindValue(HashTable *t, const char *key)
181 {
182 	HashEntry *he = HashTable_FindEntry(t, key);
183 	return he != NULL ? he->value : NULL;
184 }
185 
186 /*
187  * Find the value corresponding to the key and the precomputed hash,
188  * or return NULL.
189  */
190 void *
191 HashTable_FindValueHash(HashTable *t, const char *key, unsigned int h)
192 {
193 	HashEntry *he = HashTable_Find(t, h, key);
194 	return he != NULL ? he->value : NULL;
195 }
196 
197 /*
198  * Make the hash table larger. Any bucket numbers from the old table become
199  * invalid; the hash codes stay valid though.
200  */
201 static void
202 HashTable_Enlarge(HashTable *t)
203 {
204 	unsigned int oldSize = t->bucketsSize;
205 	HashEntry **oldBuckets = t->buckets;
206 	unsigned int newSize = 2 * oldSize;
207 	unsigned int newMask = newSize - 1;
208 	HashEntry **newBuckets = bmake_malloc(sizeof *newBuckets * newSize);
209 	size_t i;
210 
211 	for (i = 0; i < newSize; i++)
212 		newBuckets[i] = NULL;
213 
214 	for (i = 0; i < oldSize; i++) {
215 		HashEntry *he = oldBuckets[i];
216 		while (he != NULL) {
217 			HashEntry *next = he->next;
218 			he->next = newBuckets[he->key_hash & newMask];
219 			newBuckets[he->key_hash & newMask] = he;
220 			he = next;
221 		}
222 	}
223 
224 	free(oldBuckets);
225 
226 	t->bucketsSize = newSize;
227 	t->bucketsMask = newMask;
228 	t->buckets = newBuckets;
229 	DEBUG5(HASH, "%s: %p size=%d entries=%d maxchain=%d\n",
230 	       __func__, t, t->bucketsSize, t->numEntries, t->maxchain);
231 	t->maxchain = 0;
232 }
233 
234 /*
235  * Find or create an entry corresponding to the key.
236  * Return in out_isNew whether a new entry has been created.
237  */
238 HashEntry *
239 HashTable_CreateEntry(HashTable *t, const char *key, Boolean *out_isNew)
240 {
241 	size_t keylen;
242 	unsigned int h = hash(key, &keylen);
243 	HashEntry *he = HashTable_Find(t, h, key);
244 
245 	if (he != NULL) {
246 		if (out_isNew != NULL)
247 			*out_isNew = FALSE;
248 		return he;
249 	}
250 
251 	if (t->numEntries >= rebuildLimit * t->bucketsSize)
252 		HashTable_Enlarge(t);
253 
254 	he = bmake_malloc(sizeof *he + keylen);
255 	he->value = NULL;
256 	he->key_hash = h;
257 	memcpy(he->key, key, keylen + 1);
258 
259 	he->next = t->buckets[h & t->bucketsMask];
260 	t->buckets[h & t->bucketsMask] = he;
261 	t->numEntries++;
262 
263 	if (out_isNew != NULL)
264 		*out_isNew = TRUE;
265 	return he;
266 }
267 
268 HashEntry *
269 HashTable_Set(HashTable *t, const char *key, void *value)
270 {
271 	HashEntry *he = HashTable_CreateEntry(t, key, NULL);
272 	HashEntry_Set(he, value);
273 	return he;
274 }
275 
276 /* Delete the entry from the table and free the associated memory. */
277 void
278 HashTable_DeleteEntry(HashTable *t, HashEntry *he)
279 {
280 	HashEntry **ref = &t->buckets[he->key_hash & t->bucketsMask];
281 	HashEntry *p;
282 
283 	for (; (p = *ref) != NULL; ref = &p->next) {
284 		if (p == he) {
285 			*ref = p->next;
286 			free(p);
287 			t->numEntries--;
288 			return;
289 		}
290 	}
291 	abort();
292 }
293 
294 /* Set things up for iterating over all entries in the hash table. */
295 void
296 HashIter_Init(HashIter *hi, HashTable *t)
297 {
298 	hi->table = t;
299 	hi->nextBucket = 0;
300 	hi->entry = NULL;
301 }
302 
303 /*
304  * Return the next entry in the hash table, or NULL if the end of the table
305  * is reached.
306  */
307 HashEntry *
308 HashIter_Next(HashIter *hi)
309 {
310 	HashTable *t = hi->table;
311 	HashEntry *he = hi->entry;
312 	HashEntry **buckets = t->buckets;
313 	unsigned int bucketsSize = t->bucketsSize;
314 
315 	if (he != NULL)
316 		he = he->next;	/* skip the most recently returned entry */
317 
318 	while (he == NULL) {	/* find the next nonempty chain */
319 		if (hi->nextBucket >= bucketsSize)
320 			return NULL;
321 		he = buckets[hi->nextBucket++];
322 	}
323 	hi->entry = he;
324 	return he;
325 }
326 
327 void
328 HashTable_DebugStats(HashTable *t, const char *name)
329 {
330 	DEBUG4(HASH, "HashTable %s: size=%u numEntries=%u maxchain=%u\n",
331 	       name, t->bucketsSize, t->numEntries, t->maxchain);
332 }
333