xref: /netbsd/usr.bin/make/hash.c (revision 16b60d57)
1 /*	$NetBSD: hash.c,v 1.72 2022/02/09 21:09:24 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 1.72 2022/02/09 21:09:24 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
Hash_String(const char * key,const char ** out_keyEnd)87 Hash_String(const char *key, const char **out_keyEnd)
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 	*out_keyEnd = p;
97 	return h;
98 }
99 
100 /* This hash function matches Gosling's Emacs and java.lang.String. */
101 unsigned int
Hash_Substring(Substring key)102 Hash_Substring(Substring key)
103 {
104 	unsigned int h;
105 	const char *p;
106 
107 	h = 0;
108 	for (p = key.start; p != key.end; p++)
109 		h = 31 * h + (unsigned char)*p;
110 	return h;
111 }
112 
113 static HashEntry *
HashTable_Find(HashTable * t,Substring key,unsigned int h)114 HashTable_Find(HashTable *t, Substring key, unsigned int h)
115 {
116 	HashEntry *e;
117 	unsigned int chainlen = 0;
118 	size_t keyLen = Substring_Length(key);
119 
120 #ifdef DEBUG_HASH_LOOKUP
121 	DEBUG4(HASH, "HashTable_Find: %p h=%08x key=%.*s\n",
122 	    t, h, (int)keyLen, key.start);
123 #endif
124 
125 	for (e = t->buckets[h & t->bucketsMask]; e != NULL; e = e->next) {
126 		chainlen++;
127 		if (e->key_hash == h &&
128 		    strncmp(e->key, key.start, keyLen) == 0 &&
129 		    e->key[keyLen] == '\0')
130 			break;
131 	}
132 
133 	if (chainlen > t->maxchain)
134 		t->maxchain = chainlen;
135 
136 	return e;
137 }
138 
139 /* Set up the hash table. */
140 void
HashTable_Init(HashTable * t)141 HashTable_Init(HashTable *t)
142 {
143 	unsigned int n = 16, i;
144 	HashEntry **buckets = bmake_malloc(sizeof *buckets * n);
145 	for (i = 0; i < n; i++)
146 		buckets[i] = NULL;
147 
148 	t->buckets = buckets;
149 	t->bucketsSize = n;
150 	t->numEntries = 0;
151 	t->bucketsMask = n - 1;
152 	t->maxchain = 0;
153 }
154 
155 /*
156  * Remove everything from the hash table and free up the memory for the keys
157  * of the hash table, but not for the values associated to these keys.
158  */
159 void
HashTable_Done(HashTable * t)160 HashTable_Done(HashTable *t)
161 {
162 	HashEntry **buckets = t->buckets;
163 	size_t i, n = t->bucketsSize;
164 
165 	for (i = 0; i < n; i++) {
166 		HashEntry *he = buckets[i];
167 		while (he != NULL) {
168 			HashEntry *next = he->next;
169 			free(he);
170 			he = next;
171 		}
172 	}
173 
174 	free(t->buckets);
175 #ifdef CLEANUP
176 	t->buckets = NULL;
177 #endif
178 }
179 
180 /* Find the entry corresponding to the key, or return NULL. */
181 HashEntry *
HashTable_FindEntry(HashTable * t,const char * key)182 HashTable_FindEntry(HashTable *t, const char *key)
183 {
184 	const char *keyEnd;
185 	unsigned int h = Hash_String(key, &keyEnd);
186 	return HashTable_Find(t, Substring_Init(key, keyEnd), h);
187 }
188 
189 /* Find the value corresponding to the key, or return NULL. */
190 void *
HashTable_FindValue(HashTable * t,const char * key)191 HashTable_FindValue(HashTable *t, const char *key)
192 {
193 	HashEntry *he = HashTable_FindEntry(t, key);
194 	return he != NULL ? he->value : NULL;
195 }
196 
197 /*
198  * Find the value corresponding to the key and the precomputed hash,
199  * or return NULL.
200  */
201 void *
HashTable_FindValueBySubstringHash(HashTable * t,Substring key,unsigned int h)202 HashTable_FindValueBySubstringHash(HashTable *t, Substring key, unsigned int h)
203 {
204 	HashEntry *he = HashTable_Find(t, key, h);
205 	return he != NULL ? he->value : NULL;
206 }
207 
208 /*
209  * Make the hash table larger. Any bucket numbers from the old table become
210  * invalid; the hash codes stay valid though.
211  */
212 static void
HashTable_Enlarge(HashTable * t)213 HashTable_Enlarge(HashTable *t)
214 {
215 	unsigned int oldSize = t->bucketsSize;
216 	HashEntry **oldBuckets = t->buckets;
217 	unsigned int newSize = 2 * oldSize;
218 	unsigned int newMask = newSize - 1;
219 	HashEntry **newBuckets = bmake_malloc(sizeof *newBuckets * newSize);
220 	size_t i;
221 
222 	for (i = 0; i < newSize; i++)
223 		newBuckets[i] = NULL;
224 
225 	for (i = 0; i < oldSize; i++) {
226 		HashEntry *he = oldBuckets[i];
227 		while (he != NULL) {
228 			HashEntry *next = he->next;
229 			he->next = newBuckets[he->key_hash & newMask];
230 			newBuckets[he->key_hash & newMask] = he;
231 			he = next;
232 		}
233 	}
234 
235 	free(oldBuckets);
236 
237 	t->bucketsSize = newSize;
238 	t->bucketsMask = newMask;
239 	t->buckets = newBuckets;
240 	DEBUG4(HASH, "HashTable_Enlarge: %p size=%d entries=%d maxchain=%d\n",
241 	    (void *)t, t->bucketsSize, t->numEntries, t->maxchain);
242 	t->maxchain = 0;
243 }
244 
245 /*
246  * Find or create an entry corresponding to the key.
247  * Return in out_isNew whether a new entry has been created.
248  */
249 HashEntry *
HashTable_CreateEntry(HashTable * t,const char * key,bool * out_isNew)250 HashTable_CreateEntry(HashTable *t, const char *key, bool *out_isNew)
251 {
252 	const char *keyEnd;
253 	unsigned int h = Hash_String(key, &keyEnd);
254 	HashEntry *he = HashTable_Find(t, Substring_Init(key, keyEnd), h);
255 
256 	if (he != NULL) {
257 		if (out_isNew != NULL)
258 			*out_isNew = false;
259 		return he;
260 	}
261 
262 	if (t->numEntries >= rebuildLimit * t->bucketsSize)
263 		HashTable_Enlarge(t);
264 
265 	he = bmake_malloc(sizeof *he + (size_t)(keyEnd - key));
266 	he->value = NULL;
267 	he->key_hash = h;
268 	memcpy(he->key, key, (size_t)(keyEnd - key) + 1);
269 
270 	he->next = t->buckets[h & t->bucketsMask];
271 	t->buckets[h & t->bucketsMask] = he;
272 	t->numEntries++;
273 
274 	if (out_isNew != NULL)
275 		*out_isNew = true;
276 	return he;
277 }
278 
279 void
HashTable_Set(HashTable * t,const char * key,void * value)280 HashTable_Set(HashTable *t, const char *key, void *value)
281 {
282 	HashEntry *he = HashTable_CreateEntry(t, key, NULL);
283 	HashEntry_Set(he, value);
284 }
285 
286 /* Delete the entry from the table, don't free the value of the entry. */
287 void
HashTable_DeleteEntry(HashTable * t,HashEntry * he)288 HashTable_DeleteEntry(HashTable *t, HashEntry *he)
289 {
290 	HashEntry **ref = &t->buckets[he->key_hash & t->bucketsMask];
291 	HashEntry *p;
292 
293 	for (; (p = *ref) != NULL; ref = &p->next) {
294 		if (p == he) {
295 			*ref = p->next;
296 			free(p);
297 			t->numEntries--;
298 			return;
299 		}
300 	}
301 	abort();
302 }
303 
304 /*
305  * Return the next entry in the hash table, or NULL if the end of the table
306  * is reached.
307  */
308 HashEntry *
HashIter_Next(HashIter * hi)309 HashIter_Next(HashIter *hi)
310 {
311 	HashTable *t = hi->table;
312 	HashEntry *he = hi->entry;
313 	HashEntry **buckets = t->buckets;
314 	unsigned int bucketsSize = t->bucketsSize;
315 
316 	if (he != NULL)
317 		he = he->next;	/* skip the most recently returned entry */
318 
319 	while (he == NULL) {	/* find the next nonempty chain */
320 		if (hi->nextBucket >= bucketsSize)
321 			return NULL;
322 		he = buckets[hi->nextBucket++];
323 	}
324 	hi->entry = he;
325 	return he;
326 }
327 
328 void
HashTable_DebugStats(HashTable * t,const char * name)329 HashTable_DebugStats(HashTable *t, const char *name)
330 {
331 	DEBUG4(HASH, "HashTable %s: size=%u numEntries=%u maxchain=%u\n",
332 	    name, t->bucketsSize, t->numEntries, t->maxchain);
333 }
334