xref: /openbsd-src/usr.sbin/config/hash.c (revision 2b0358df1d88d06ef4139321dd05bd5e05d91eaf)
1 /*	$OpenBSD: hash.c,v 1.14 2004/01/04 18:30:05 deraadt Exp $	*/
2 /*	$NetBSD: hash.c,v 1.4 1996/11/07 22:59:43 gwr Exp $	*/
3 
4 /*
5  * Copyright (c) 1992, 1993
6  *	The Regents of the University of California.  All rights reserved.
7  *
8  * This software was developed by the Computer Systems Engineering group
9  * at Lawrence Berkeley Laboratory under DARPA contract BG 91-66 and
10  * contributed to Berkeley.
11  *
12  * All advertising materials mentioning features or use of this software
13  * must display the following acknowledgement:
14  *	This product includes software developed by the University of
15  *	California, Lawrence Berkeley Laboratories.
16  *
17  * Redistribution and use in source and binary forms, with or without
18  * modification, are permitted provided that the following conditions
19  * are met:
20  * 1. Redistributions of source code must retain the above copyright
21  *    notice, this list of conditions and the following disclaimer.
22  * 2. Redistributions in binary form must reproduce the above copyright
23  *    notice, this list of conditions and the following disclaimer in the
24  *    documentation and/or other materials provided with the distribution.
25  * 3. Neither the name of the University nor the names of its contributors
26  *    may be used to endorse or promote products derived from this software
27  *    without specific prior written permission.
28  *
29  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
30  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
31  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
32  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
33  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
34  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
35  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
36  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
37  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
38  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
39  * SUCH DAMAGE.
40  *
41  *	from: @(#)hash.c	8.1 (Berkeley) 6/6/93
42  */
43 
44 #include <sys/param.h>
45 #include <stdlib.h>
46 #include <string.h>
47 #include "config.h"
48 
49 /*
50  * These are really for MAKE_BOOTSTRAP but harmless.
51  * XXX - Why not just use malloc in here, anyway?
52  */
53 #ifndef	ALIGNBYTES
54 #define	ALIGNBYTES 3
55 #endif
56 #ifndef ALIGN
57 #define	ALIGN(p)	(((long)(p) + ALIGNBYTES) &~ ALIGNBYTES)
58 #endif
59 
60 /*
61  * Interned strings are kept in a hash table.  By making each string
62  * unique, the program can compare strings by comparing pointers.
63  */
64 struct hashent {
65 	struct	hashent *h_next;	/* hash buckets are chained */
66 	const char *h_name;		/* the string */
67 	u_int	h_hash;			/* its hash value */
68 	void	*h_value;		/* other values (for name=value) */
69 };
70 struct hashtab {
71 	size_t	ht_size;		/* size (power of 2) */
72 	u_int	ht_mask;		/* == ht_size - 1 */
73 	u_int	ht_used;		/* number of entries used */
74 	u_int	ht_lim;			/* when to expand */
75 	struct	hashent **ht_tab;	/* base of table */
76 };
77 static struct hashtab strings;
78 
79 /*
80  * HASHFRACTION controls ht_lim, which in turn controls the average chain
81  * length.  We allow a few entries, on average, as comparing them is usually
82  * cheap (the h_hash values prevent a strcmp).
83  */
84 #define	HASHFRACTION(sz) ((sz) * 3 / 2)
85 
86 /* round up to next multiple of y, where y is a power of 2 */
87 #define	ROUND(x, y) (((x) + (y) - 1) & ~((y) - 1))
88 
89 static void *poolalloc(size_t);
90 static void ht_init(struct hashtab *, size_t);
91 static void ht_expand(struct hashtab *);
92 /*
93  * Allocate space that will never be freed.
94  */
95 static void *
96 poolalloc(size_t size)
97 {
98 	char *p;
99 	size_t alloc;
100 	static char *pool;
101 	static size_t nleft;
102 
103 	if (nleft < size) {
104 		/*
105 		 * Compute a `good' size to allocate via malloc.
106 		 * 16384 is a guess at a good page size for malloc;
107 		 * 32 is a guess at malloc's overhead.
108 		 */
109 		alloc = ROUND(size + 32, 16384) - 32;
110 		p = emalloc(alloc);
111 		nleft = alloc - size;
112 	} else {
113 		p = pool;
114 		nleft -= size;
115 	}
116 	pool = p + size;
117 	return (p);
118 }
119 
120 /*
121  * Initialize a new hash table.  The size must be a power of 2.
122  */
123 static void
124 ht_init(struct hashtab *ht, size_t sz)
125 {
126 	struct hashent **h;
127 	u_int n;
128 
129 	h = emalloc(sz * sizeof *h);
130 	ht->ht_tab = h;
131 	ht->ht_size = sz;
132 	ht->ht_mask = sz - 1;
133 	for (n = 0; n < sz; n++)
134 		*h++ = NULL;
135 	ht->ht_used = 0;
136 	ht->ht_lim = HASHFRACTION(sz);
137 }
138 
139 /*
140  * Expand an existing hash table.
141  */
142 static void
143 ht_expand(struct hashtab *ht)
144 {
145 	struct hashent *p, **h, **oldh, *q;
146 	u_int n, i;
147 
148 	n = ht->ht_size * 2;
149 	h = emalloc(n * sizeof *h);
150 	for (i = 0; i < n; i++)
151 		h[i] = NULL;
152 	oldh = ht->ht_tab;
153 	n--;
154 	for (i = ht->ht_size; i != 0; i--) {
155 		for (p = *oldh++; p != NULL; p = q) {
156 			q = p->h_next;
157 			p->h_next = h[p->h_hash & n];
158 			h[p->h_hash & n] = p;
159 		}
160 	}
161 	free(ht->ht_tab);
162 	ht->ht_tab = h;
163 	ht->ht_mask = n;
164 	ht->ht_size = ++n;
165 	ht->ht_lim = HASHFRACTION(n);
166 }
167 
168 /*
169  * Make a new hash entry, setting its h_next to NULL.
170  */
171 static __inline struct hashent *
172 newhashent(const char *name, u_int h)
173 {
174 	struct	hashent *hp;
175 	char	*m;
176 
177 	m = poolalloc(sizeof(*hp) + ALIGNBYTES);
178 	hp = (struct hashent *)ALIGN(m);
179 	hp->h_name = name;
180 	hp->h_hash = h;
181 	hp->h_next = NULL;
182 	return (hp);
183 }
184 
185 /*
186  * Hash a string.
187  */
188 static __inline u_int
189 hash(const char *str)
190 {
191 	u_int h;
192 
193 	for (h = 0; *str;)
194 		h = (h << 5) + h + *str++;
195 	return (h);
196 }
197 
198 void
199 initintern(void)
200 {
201 
202 	ht_init(&strings, 128);
203 }
204 
205 /*
206  * Generate a single unique copy of the given string.  We expect this
207  * function to be used frequently, so it should be fast.
208  */
209 const char *
210 intern(const char *s)
211 {
212 	struct hashtab *ht;
213 	struct hashent *hp, **hpp;
214 	u_int h;
215 	char *p;
216 	size_t l;
217 
218 	ht = &strings;
219 	h = hash(s);
220 	hpp = &ht->ht_tab[h & ht->ht_mask];
221 	for (; (hp = *hpp) != NULL; hpp = &hp->h_next)
222 		if (hp->h_hash == h && strcmp(hp->h_name, s) == 0)
223 			return (hp->h_name);
224 	l = strlen(s) + 1;
225 	p = poolalloc(l);
226 	bcopy(s, p, l);
227 	*hpp = newhashent(p, h);
228 	if (++ht->ht_used > ht->ht_lim)
229 		ht_expand(ht);
230 	return (p);
231 }
232 
233 struct hashtab *
234 ht_new(void)
235 {
236 	struct hashtab *ht;
237 
238 	ht = emalloc(sizeof *ht);
239 	ht_init(ht, 8);
240 	return (ht);
241 }
242 
243 /*
244  * Remove.
245  */
246 int
247 ht_remove(struct hashtab *ht, const char *nam)
248 {
249 	struct hashent *hp, *thp;
250 	u_int h;
251 
252 	h = hash(nam);
253 	hp = ht->ht_tab[h & ht->ht_mask];
254 	while (hp && hp->h_name == nam)	{
255 		ht->ht_tab[h & ht->ht_mask] = hp->h_next;
256 		/* XXX free hp ? */
257 		hp = ht->ht_tab[h & ht->ht_mask];
258 	}
259 
260 	if ((hp = ht->ht_tab[h & ht->ht_mask]) == NULL)
261 		return (0);
262 
263 	for (thp = hp->h_next; thp != NULL; thp = hp->h_next) {
264 		if (thp->h_name == nam) {
265 			hp->h_next = thp->h_next;
266 			/* XXX free thp ? */
267 		} else
268 			hp = thp;
269 	}
270 
271 	return (0);
272 }
273 
274 /*
275  * Insert and/or replace.
276  */
277 int
278 ht_insrep(struct hashtab *ht, const char *nam, void *val, int replace)
279 {
280 	struct hashent *hp, **hpp;
281 	u_int h;
282 
283 	h = hash(nam);
284 	hpp = &ht->ht_tab[h & ht->ht_mask];
285 	for (; (hp = *hpp) != NULL; hpp = &hp->h_next) {
286 		if (hp->h_name == nam) {
287 			if (replace)
288 				hp->h_value = val;
289 			return (1);
290 		}
291 	}
292 	*hpp = hp = newhashent(nam, h);
293 	hp->h_value = val;
294 	if (++ht->ht_used > ht->ht_lim)
295 		ht_expand(ht);
296 	return (0);
297 }
298 
299 void *
300 ht_lookup(struct hashtab *ht, const char *nam)
301 {
302 	struct hashent *hp, **hpp;
303 	u_int h;
304 
305 	h = hash(nam);
306 	hpp = &ht->ht_tab[h & ht->ht_mask];
307 	for (; (hp = *hpp) != NULL; hpp = &hp->h_next)
308 		if (hp->h_name == nam)
309 			return (hp->h_value);
310 	return (NULL);
311 }
312