xref: /freebsd/contrib/bmake/hash.c (revision 22619282)
1 /*	$NetBSD: hash.c,v 1.79 2024/07/07 09:37:00 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 and pointer values. */
73 
74 #include "make.h"
75 
76 /*	"@(#)hash.c	8.1 (Berkeley) 6/6/93"	*/
77 MAKE_RCSID("$NetBSD: hash.c,v 1.79 2024/07/07 09:37:00 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 *he;
117 	size_t keyLen = Substring_Length(key);
118 
119 #ifdef DEBUG_HASH_LOOKUP
120 	DEBUG4(HASH, "HashTable_Find: %p h=%08x key=%.*s\n",
121 	    t, h, (int)keyLen, key.start);
122 #endif
123 
124 	for (he = t->buckets[h & t->bucketsMask]; he != NULL; he = he->next) {
125 		if (he->hash == h &&
126 		    strncmp(he->key, key.start, keyLen) == 0 &&
127 		    he->key[keyLen] == '\0')
128 			break;
129 	}
130 
131 	return he;
132 }
133 
134 /* Set up the hash table. */
135 void
HashTable_Init(HashTable * t)136 HashTable_Init(HashTable *t)
137 {
138 	unsigned int n = 16, i;
139 	HashEntry **buckets = bmake_malloc(sizeof *buckets * n);
140 	for (i = 0; i < n; i++)
141 		buckets[i] = NULL;
142 
143 	t->buckets = buckets;
144 	t->bucketsSize = n;
145 	t->numEntries = 0;
146 	t->bucketsMask = n - 1;
147 }
148 
149 /*
150  * Remove everything from the hash table and free up the memory for the keys
151  * of the hash table, but not for the values associated to these keys.
152  */
153 void
HashTable_Done(HashTable * t)154 HashTable_Done(HashTable *t)
155 {
156 	HashEntry **buckets = t->buckets;
157 	size_t i, n = t->bucketsSize;
158 
159 	for (i = 0; i < n; i++) {
160 		HashEntry *he = buckets[i];
161 		while (he != NULL) {
162 			HashEntry *next = he->next;
163 			free(he);
164 			he = next;
165 		}
166 	}
167 
168 	free(t->buckets);
169 #ifdef CLEANUP
170 	t->buckets = NULL;
171 #endif
172 }
173 
174 /* Find the entry corresponding to the key, or return NULL. */
175 HashEntry *
HashTable_FindEntry(HashTable * t,const char * key)176 HashTable_FindEntry(HashTable *t, const char *key)
177 {
178 	const char *keyEnd;
179 	unsigned int h = Hash_String(key, &keyEnd);
180 	return HashTable_Find(t, Substring_Init(key, keyEnd), h);
181 }
182 
183 /* Find the value corresponding to the key, or return NULL. */
184 void *
HashTable_FindValue(HashTable * t,const char * key)185 HashTable_FindValue(HashTable *t, const char *key)
186 {
187 	HashEntry *he = HashTable_FindEntry(t, key);
188 	return he != NULL ? he->value : NULL;
189 }
190 
191 /*
192  * Find the value corresponding to the key and the precomputed hash,
193  * or return NULL.
194  */
195 void *
HashTable_FindValueBySubstringHash(HashTable * t,Substring key,unsigned int h)196 HashTable_FindValueBySubstringHash(HashTable *t, Substring key, unsigned int h)
197 {
198 	HashEntry *he = HashTable_Find(t, key, h);
199 	return he != NULL ? he->value : NULL;
200 }
201 
202 static unsigned
HashTable_MaxChain(const HashTable * t)203 HashTable_MaxChain(const HashTable *t)
204 {
205 	unsigned b, cl, max_cl = 0;
206 	for (b = 0; b < t->bucketsSize; b++) {
207 		const HashEntry *he = t->buckets[b];
208 		for (cl = 0; he != NULL; he = he->next)
209 			cl++;
210 		if (cl > max_cl)
211 			max_cl = cl;
212 	}
213 	return max_cl;
214 }
215 
216 /*
217  * Make the hash table larger. Any bucket numbers from the old table become
218  * invalid; the hash values stay valid though.
219  */
220 static void
HashTable_Enlarge(HashTable * t)221 HashTable_Enlarge(HashTable *t)
222 {
223 	unsigned int oldSize = t->bucketsSize;
224 	HashEntry **oldBuckets = t->buckets;
225 	unsigned int newSize = 2 * oldSize;
226 	unsigned int newMask = newSize - 1;
227 	HashEntry **newBuckets = bmake_malloc(sizeof *newBuckets * newSize);
228 	size_t i;
229 
230 	for (i = 0; i < newSize; i++)
231 		newBuckets[i] = NULL;
232 
233 	for (i = 0; i < oldSize; i++) {
234 		HashEntry *he = oldBuckets[i];
235 		while (he != NULL) {
236 			HashEntry *next = he->next;
237 			he->next = newBuckets[he->hash & newMask];
238 			newBuckets[he->hash & newMask] = he;
239 			he = next;
240 		}
241 	}
242 
243 	free(oldBuckets);
244 
245 	t->bucketsSize = newSize;
246 	t->bucketsMask = newMask;
247 	t->buckets = newBuckets;
248 	DEBUG4(HASH, "HashTable_Enlarge: %p size=%d entries=%d maxchain=%d\n",
249 	    (void *)t, t->bucketsSize, t->numEntries, HashTable_MaxChain(t));
250 }
251 
252 /*
253  * Find or create an entry corresponding to the key.
254  * Return in out_isNew whether a new entry has been created.
255  */
256 HashEntry *
HashTable_CreateEntry(HashTable * t,const char * key,bool * out_isNew)257 HashTable_CreateEntry(HashTable *t, const char *key, bool *out_isNew)
258 {
259 	const char *keyEnd;
260 	unsigned int h = Hash_String(key, &keyEnd);
261 	HashEntry *he = HashTable_Find(t, Substring_Init(key, keyEnd), h);
262 
263 	if (he != NULL) {
264 		if (out_isNew != NULL)
265 			*out_isNew = false;
266 		return he;
267 	}
268 
269 	if (t->numEntries >= rebuildLimit * t->bucketsSize)
270 		HashTable_Enlarge(t);
271 
272 	he = bmake_malloc(sizeof *he + (size_t)(keyEnd - key));
273 	he->value = NULL;
274 	he->hash = h;
275 	memcpy(he->key, key, (size_t)(keyEnd - key) + 1);
276 
277 	he->next = t->buckets[h & t->bucketsMask];
278 	t->buckets[h & t->bucketsMask] = he;
279 	t->numEntries++;
280 
281 	if (out_isNew != NULL)
282 		*out_isNew = true;
283 	return he;
284 }
285 
286 void
HashTable_Set(HashTable * t,const char * key,void * value)287 HashTable_Set(HashTable *t, const char *key, void *value)
288 {
289 	HashEntry *he = HashTable_CreateEntry(t, key, NULL);
290 	HashEntry_Set(he, value);
291 }
292 
293 /* Delete the entry from the table, don't free the value of the entry. */
294 void
HashTable_DeleteEntry(HashTable * t,HashEntry * he)295 HashTable_DeleteEntry(HashTable *t, HashEntry *he)
296 {
297 	HashEntry **ref = &t->buckets[he->hash & t->bucketsMask];
298 
299 	for (; *ref != he; ref = &(*ref)->next)
300 		continue;
301 	*ref = he->next;
302 	free(he);
303 	t->numEntries--;
304 }
305 
306 /*
307  * Place the next entry from the hash table in hi->entry, or return false if
308  * the end of the table is reached.
309  */
310 bool
HashIter_Next(HashIter * hi)311 HashIter_Next(HashIter *hi)
312 {
313 	HashTable *t = hi->table;
314 	HashEntry *he = hi->entry;
315 	HashEntry **buckets = t->buckets;
316 	unsigned int bucketsSize = t->bucketsSize;
317 
318 	if (he != NULL)
319 		he = he->next;	/* skip the most recently returned entry */
320 
321 	while (he == NULL) {	/* find the next nonempty chain */
322 		if (hi->nextBucket >= bucketsSize)
323 			return false;
324 		he = buckets[hi->nextBucket++];
325 	}
326 	hi->entry = he;
327 	return true;
328 }
329 
330 void
HashTable_DebugStats(const HashTable * t,const char * name)331 HashTable_DebugStats(const HashTable *t, const char *name)
332 {
333 	DEBUG4(HASH, "HashTable \"%s\": size=%u entries=%u maxchain=%u\n",
334 	    name, t->bucketsSize, t->numEntries, HashTable_MaxChain(t));
335 }
336