xref: /freebsd/lib/libc/db/hash/hash_func.c (revision 2008043f386721d58158e37e0d7e50df8095942d)
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/types.h>
39 
40 #include <db.h>
41 #include "hash.h"
42 #include "page.h"
43 #include "extern.h"
44 
45 #ifdef notdef
46 static u_int32_t hash1(const void *, size_t) __unused;
47 static u_int32_t hash2(const void *, size_t) __unused;
48 static u_int32_t hash3(const void *, size_t) __unused;
49 #endif
50 static u_int32_t hash4(const void *, size_t);
51 
52 /* Default hash function. */
53 u_int32_t (*__default_hash)(const void *, size_t) = hash4;
54 
55 #ifdef notdef
56 /*
57  * Assume that we've already split the bucket to which this key hashes,
58  * calculate that bucket, and check that in fact we did already split it.
59  *
60  * EJB's original hsearch hash.
61  */
62 #define PRIME1		37
63 #define PRIME2		1048583
64 
65 u_int32_t
66 hash1(const void *key, size_t len)
67 {
68 	u_int32_t h;
69 	u_int8_t *k;
70 
71 	h = 0;
72 	k = (u_int8_t *)key;
73 	/* Convert string to integer */
74 	while (len--)
75 		h = h * PRIME1 ^ (*k++ - ' ');
76 	h %= PRIME2;
77 	return (h);
78 }
79 
80 /*
81  * Phong Vo's linear congruential hash
82  */
83 #define dcharhash(h, c)	((h) = 0x63c63cd9*(h) + 0x9c39c33d + (c))
84 
85 u_int32_t
86 hash2(const void *key, size_t len)
87 {
88 	u_int32_t h;
89 	u_int8_t *e, c, *k;
90 
91 	k = (u_int8_t *)key;
92 	e = k + len;
93 	for (h = 0; k != e;) {
94 		c = *k++;
95 		if (!c && k > e)
96 			break;
97 		dcharhash(h, c);
98 	}
99 	return (h);
100 }
101 
102 /*
103  * This is INCREDIBLY ugly, but fast.  We break the string up into 8 byte
104  * units.  On the first time through the loop we get the "leftover bytes"
105  * (strlen % 8).  On every other iteration, we perform 8 HASHC's so we handle
106  * all 8 bytes.  Essentially, this saves us 7 cmp & branch instructions.  If
107  * this routine is heavily used enough, it's worth the ugly coding.
108  *
109  * Ozan Yigit's original sdbm hash.
110  */
111 u_int32_t
112 hash3(const void *key, size_t len)
113 {
114 	u_int32_t n, loop;
115 	u_int8_t *k;
116 
117 #define HASHC   n = *k++ + 65599 * n
118 
119 	n = 0;
120 	k = (u_int8_t *)key;
121 	if (len > 0) {
122 		loop = (len + 8 - 1) >> 3;
123 
124 		switch (len & (8 - 1)) {
125 		case 0:
126 			do {	/* All fall throughs */
127 				HASHC;
128 		case 7:
129 				HASHC;
130 		case 6:
131 				HASHC;
132 		case 5:
133 				HASHC;
134 		case 4:
135 				HASHC;
136 		case 3:
137 				HASHC;
138 		case 2:
139 				HASHC;
140 		case 1:
141 				HASHC;
142 			} while (--loop);
143 		}
144 
145 	}
146 	return (n);
147 }
148 #endif /* notdef */
149 
150 /* Chris Torek's hash function. */
151 u_int32_t
152 hash4(const void *key, size_t len)
153 {
154 	u_int32_t h, loop;
155 	const u_int8_t *k;
156 
157 #define HASH4a   h = (h << 5) - h + *k++;
158 #define HASH4b   h = (h << 5) + h + *k++;
159 #define HASH4 HASH4b
160 
161 	h = 0;
162 	k = key;
163 	if (len > 0) {
164 		loop = (len + 8 - 1) >> 3;
165 
166 		switch (len & (8 - 1)) {
167 		case 0:
168 			do {	/* All fall throughs */
169 				HASH4;
170 		case 7:
171 				HASH4;
172 		case 6:
173 				HASH4;
174 		case 5:
175 				HASH4;
176 		case 4:
177 				HASH4;
178 		case 3:
179 				HASH4;
180 		case 2:
181 				HASH4;
182 		case 1:
183 				HASH4;
184 			} while (--loop);
185 		}
186 
187 	}
188 	return (h);
189 }
190