1 /* $NetBSD: hash.h,v 1.14 2007/02/03 23:46:09 christos Exp $ */
4 * Copyright (c) 1990, 1993, 1994
5 * The Regents of the University of California. All rights reserved.
7 * This code is derived from software contributed to Berkeley by
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
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.
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
34 * @(#)hash.h 8.3 (Berkeley) 5/31/94
37 #if HAVE_NBTOOL_CONFIG_H
38 #include "nbtool_config.h"
43 HASH_GET
, HASH_PUT
, HASH_PUTNEW
, HASH_DELETE
, HASH_FIRST
, HASH_NEXT
46 /* Buffer Management structures */
47 typedef struct _bufhead BUFHEAD
;
50 BUFHEAD
*prev
; /* LRU links */
51 BUFHEAD
*next
; /* LRU links */
52 BUFHEAD
*ovfl
; /* Overflow page buffer header */
53 uint32_t addr
; /* Address of this page */
54 char *page
; /* Actual page data */
56 #define BUF_MOD 0x0001
57 #define BUF_DISK 0x0002
58 #define BUF_BUCKET 0x0004
59 #define BUF_PIN 0x0008
62 #define IS_BUCKET(X) ((X) & BUF_BUCKET)
64 typedef BUFHEAD
**SEGMENT
;
66 /* Hash Table Information */
67 typedef struct hashhdr
{ /* Disk resident portion */
68 int32_t magic
; /* Magic NO for hash tables */
69 int32_t version
; /* Version ID */
70 uint32_t lorder
; /* Byte Order */
71 int32_t bsize
; /* Bucket/Page Size */
72 int32_t bshift
; /* Bucket shift */
73 int32_t dsize
; /* Directory Size */
74 int32_t ssize
; /* Segment Size */
75 int32_t sshift
; /* Segment shift */
76 int32_t ovfl_point
; /* Where overflow pages are being
78 int32_t last_freed
; /* Last overflow page freed */
79 int32_t max_bucket
; /* ID of Maximum bucket in use */
80 int32_t high_mask
; /* Mask to modulo into entire table */
81 int32_t low_mask
; /* Mask to modulo into lower half of
83 int32_t ffactor
; /* Fill factor */
84 int32_t nkeys
; /* Number of keys in hash table */
85 int32_t hdrpages
; /* Size of table header */
86 int32_t h_charkey
; /* value of hash(CHARKEY) */
87 #define NCACHED 32 /* number of bit maps and spare
89 int32_t spares
[NCACHED
];/* spare pages for overflow */
90 uint16_t bitmaps
[NCACHED
]; /* address of overflow page
94 typedef struct htab
{ /* Memory resident data structure */
95 HASHHDR hdr
; /* Header */
96 int nsegs
; /* Number of allocated segments */
97 int exsegs
; /* Number of extra allocated
99 uint32_t (*hash
)(const void *, size_t); /* Hash function */
100 int flags
; /* Flag values */
101 int fp
; /* File pointer */
102 char *tmp_buf
; /* Temporary Buffer for BIG data */
103 char *tmp_key
; /* Temporary Buffer for BIG keys */
104 BUFHEAD
*cpage
; /* Current page */
105 int cbucket
; /* Current bucket */
106 int cndx
; /* Index of next item on cpage */
107 int err
; /* Error Number -- for DBM
109 int new_file
; /* Indicates if fd is backing store
111 int save_file
; /* Indicates whether we need to flush
114 uint32_t *mapp
[NCACHED
]; /* Pointers to page maps */
115 int nmaps
; /* Initial number of bitmaps */
116 int nbufs
; /* Number of buffers left to
118 BUFHEAD bufhead
; /* Header of buffer lru list */
119 SEGMENT
*dir
; /* Hash Bucket directory */
125 #define MAX_BSIZE 65536 /* 2^16 */
126 #define MIN_BUFFERS 6
127 #define MINHDRSIZE 512
128 #define DEF_BUFSIZE 65536 /* 64 K */
129 #define DEF_BUCKET_SIZE 4096
130 #define DEF_BUCKET_SHIFT 12 /* log2(BUCKET) */
131 #define DEF_SEGSIZE 256
132 #define DEF_SEGSIZE_SHIFT 8 /* log2(SEGSIZE) */
133 #define DEF_DIRSIZE 256
134 #define DEF_FFACTOR 65536
135 #define MIN_FFACTOR 4
137 #define CHARKEY "%$sniglet^&"
138 #define NUMKEY 1038583
140 #define INT_TO_BYTE 2
141 #define INT_BYTE_SHIFT 5
142 #define ALL_SET ((uint32_t)0xFFFFFFFF)
145 #define PTROF(X) ((BUFHEAD *)(void *)((u_long)(X)&~0x3))
146 #define ISMOD(X) ((uint32_t)(u_long)(X)&0x1)
147 #define DOMOD(X) ((X) = (char *)(void *)((u_long)(X)|0x1))
148 #define ISDISK(X) ((uint32_t)(u_long)(X)&0x2)
149 #define DODISK(X) ((X) = (char *)(void *)((u_long)(X)|0x2))
151 #define BITS_PER_MAP 32
153 /* Given the address of the beginning of a big map, clear/set the nth bit */
154 #define CLRBIT(A, N) ((A)[(N)/BITS_PER_MAP] &= ~(1<<((N)%BITS_PER_MAP)))
155 #define SETBIT(A, N) ((A)[(N)/BITS_PER_MAP] |= (1<<((N)%BITS_PER_MAP)))
156 #define ISSET(A, N) ((A)[(N)/BITS_PER_MAP] & (1<<((N)%BITS_PER_MAP)))
158 /* Overflow management */
160 * Overflow page numbers are allocated per split point. At each doubling of
161 * the table, we can allocate extra pages. So, an overflow page number has
162 * the top 5 bits indicate which split point and the lower 11 bits indicate
163 * which page at that split point is indicated (pages within split points are
164 * numberered starting with 1).
167 #define SPLITSHIFT 11
168 #define SPLITMASK 0x7FF
169 #define SPLITNUM(N) (((uint32_t)(N)) >> SPLITSHIFT)
170 #define OPAGENUM(N) ((N) & SPLITMASK)
171 #define OADDR_OF(S,O) ((uint32_t)((uint32_t)(S) << SPLITSHIFT) + (O))
173 #define BUCKET_TO_PAGE(B) \
174 (B) + hashp->HDRPAGES + \
175 ((B) ? hashp->SPARES[__log2((uint32_t)((B)+1))-1] : 0)
176 #define OADDR_TO_PAGE(B) \
177 BUCKET_TO_PAGE ( (1 << SPLITNUM((B))) -1 ) + OPAGENUM((B));
180 * page.h contains a detailed description of the page format.
182 * Normally, keys and data are accessed from offset tables in the top of
183 * each page which point to the beginning of the key and data. There are
184 * four flag values which may be stored in these offset tables which indicate
188 * OVFLPAGE Rather than a key data pair, this pair contains
189 * the address of an overflow page. The format of
191 * OVERFLOW_PAGE_NUMBER OVFLPAGE
193 * PARTIAL_KEY This must be the first key/data pair on a page
194 * and implies that page contains only a partial key.
195 * That is, the key is too big to fit on a single page
196 * so it starts on this page and continues on the next.
197 * The format of the page is:
198 * KEY_OFF PARTIAL_KEY OVFL_PAGENO OVFLPAGE
200 * KEY_OFF -- offset of the beginning of the key
202 * OVFL_PAGENO - page number of the next overflow page
205 * FULL_KEY This must be the first key/data pair on the page. It
206 * is used in two cases.
209 * There is a complete key on the page but no data
210 * (because it wouldn't fit). The next page contains
214 * KEY_OFF FULL_KEY OVFL_PAGENO OVFL_PAGE
216 * KEY_OFF -- offset of the beginning of the key
218 * OVFL_PAGENO - page number of the next overflow page
222 * This page contains no key, but part of a large
223 * data field, which is continued on the next page.
226 * DATA_OFF FULL_KEY OVFL_PAGENO OVFL_PAGE
228 * KEY_OFF -- offset of the beginning of the data on
231 * OVFL_PAGENO - page number of the next overflow page
235 * This must be the first key/data pair on the page.
236 * There are two cases:
239 * This page contains a key and the beginning of the
240 * data field, but the data field is continued on the
244 * KEY_OFF FULL_KEY_DATA OVFL_PAGENO DATA_OFF
246 * KEY_OFF -- offset of the beginning of the key
248 * OVFL_PAGENO - page number of the next overflow page
249 * DATA_OFF -- offset of the beginning of the data
252 * This page contains the last page of a big data pair.
253 * There is no key, only the tail end of the data
257 * DATA_OFF FULL_KEY_DATA <OVFL_PAGENO> <OVFLPAGE>
259 * DATA_OFF -- offset of the beginning of the data on
262 * OVFL_PAGENO - page number of the next overflow page
265 * OVFL_PAGENO and OVFLPAGE are optional (they are
266 * not present if there is no next page).
270 #define PARTIAL_KEY 1
272 #define FULL_KEY_DATA 3
275 /* Short hands for accessing structure */
276 #define BSIZE hdr.bsize
277 #define BSHIFT hdr.bshift
278 #define DSIZE hdr.dsize
279 #define SGSIZE hdr.ssize
280 #define SSHIFT hdr.sshift
281 #define LORDER hdr.lorder
282 #define OVFL_POINT hdr.ovfl_point
283 #define LAST_FREED hdr.last_freed
284 #define MAX_BUCKET hdr.max_bucket
285 #define FFACTOR hdr.ffactor
286 #define HIGH_MASK hdr.high_mask
287 #define LOW_MASK hdr.low_mask
288 #define NKEYS hdr.nkeys
289 #define HDRPAGES hdr.hdrpages
290 #define SPARES hdr.spares
291 #define BITMAPS hdr.bitmaps
292 #define VERSION hdr.version
293 #define MAGIC hdr.magic
294 #define NEXT_FREE hdr.next_free
295 #define H_CHARKEY hdr.h_charkey