xref: /freebsd/lib/libc/db/hash/hash_func.c (revision 0957b409)
1 /*-
2  * SPDX-License-Identifier: BSD-3-Clause
3  *
4  * Copyright (c) 1990, 1993
5  *	The Regents of the University of California.  All rights reserved.
6  *
7  * This code is derived from software contributed to Berkeley by
8  * Margo Seltzer.
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 #if defined(LIBC_SCCS) && !defined(lint)
36 static char sccsid[] = "@(#)hash_func.c	8.2 (Berkeley) 2/21/94";
37 #endif /* LIBC_SCCS and not lint */
38 #include <sys/cdefs.h>
39 __FBSDID("$FreeBSD$");
40 
41 #include <sys/types.h>
42 
43 #include <db.h>
44 #include "hash.h"
45 #include "page.h"
46 #include "extern.h"
47 
48 #ifdef notdef
49 static u_int32_t hash1(const void *, size_t) __unused;
50 static u_int32_t hash2(const void *, size_t) __unused;
51 static u_int32_t hash3(const void *, size_t) __unused;
52 #endif
53 static u_int32_t hash4(const void *, size_t);
54 
55 /* Default hash function. */
56 u_int32_t (*__default_hash)(const void *, size_t) = hash4;
57 
58 #ifdef notdef
59 /*
60  * Assume that we've already split the bucket to which this key hashes,
61  * calculate that bucket, and check that in fact we did already split it.
62  *
63  * EJB's original hsearch hash.
64  */
65 #define PRIME1		37
66 #define PRIME2		1048583
67 
68 u_int32_t
69 hash1(const void *key, size_t len)
70 {
71 	u_int32_t h;
72 	u_int8_t *k;
73 
74 	h = 0;
75 	k = (u_int8_t *)key;
76 	/* Convert string to integer */
77 	while (len--)
78 		h = h * PRIME1 ^ (*k++ - ' ');
79 	h %= PRIME2;
80 	return (h);
81 }
82 
83 /*
84  * Phong Vo's linear congruential hash
85  */
86 #define dcharhash(h, c)	((h) = 0x63c63cd9*(h) + 0x9c39c33d + (c))
87 
88 u_int32_t
89 hash2(const void *key, size_t len)
90 {
91 	u_int32_t h;
92 	u_int8_t *e, c, *k;
93 
94 	k = (u_int8_t *)key;
95 	e = k + len;
96 	for (h = 0; k != e;) {
97 		c = *k++;
98 		if (!c && k > e)
99 			break;
100 		dcharhash(h, c);
101 	}
102 	return (h);
103 }
104 
105 /*
106  * This is INCREDIBLY ugly, but fast.  We break the string up into 8 byte
107  * units.  On the first time through the loop we get the "leftover bytes"
108  * (strlen % 8).  On every other iteration, we perform 8 HASHC's so we handle
109  * all 8 bytes.  Essentially, this saves us 7 cmp & branch instructions.  If
110  * this routine is heavily used enough, it's worth the ugly coding.
111  *
112  * Ozan Yigit's original sdbm hash.
113  */
114 u_int32_t
115 hash3(const void *key, size_t len)
116 {
117 	u_int32_t n, loop;
118 	u_int8_t *k;
119 
120 #define HASHC   n = *k++ + 65599 * n
121 
122 	n = 0;
123 	k = (u_int8_t *)key;
124 	if (len > 0) {
125 		loop = (len + 8 - 1) >> 3;
126 
127 		switch (len & (8 - 1)) {
128 		case 0:
129 			do {	/* All fall throughs */
130 				HASHC;
131 		case 7:
132 				HASHC;
133 		case 6:
134 				HASHC;
135 		case 5:
136 				HASHC;
137 		case 4:
138 				HASHC;
139 		case 3:
140 				HASHC;
141 		case 2:
142 				HASHC;
143 		case 1:
144 				HASHC;
145 			} while (--loop);
146 		}
147 
148 	}
149 	return (n);
150 }
151 #endif /* notdef */
152 
153 /* Chris Torek's hash function. */
154 u_int32_t
155 hash4(const void *key, size_t len)
156 {
157 	u_int32_t h, loop;
158 	const u_int8_t *k;
159 
160 #define HASH4a   h = (h << 5) - h + *k++;
161 #define HASH4b   h = (h << 5) + h + *k++;
162 #define HASH4 HASH4b
163 
164 	h = 0;
165 	k = key;
166 	if (len > 0) {
167 		loop = (len + 8 - 1) >> 3;
168 
169 		switch (len & (8 - 1)) {
170 		case 0:
171 			do {	/* All fall throughs */
172 				HASH4;
173 		case 7:
174 				HASH4;
175 		case 6:
176 				HASH4;
177 		case 5:
178 				HASH4;
179 		case 4:
180 				HASH4;
181 		case 3:
182 				HASH4;
183 		case 2:
184 				HASH4;
185 		case 1:
186 				HASH4;
187 			} while (--loop);
188 		}
189 
190 	}
191 	return (h);
192 }
193