2 * Copyright (c) 1990, 1993, 1994
3 * The Regents of the University of California. All rights reserved.
5 * This code is derived from software contributed to Berkeley by
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
11 * 1. Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer.
13 * 2. Redistributions in binary form must reproduce the above copyright
14 * notice, this list of conditions and the following disclaimer in the
15 * documentation and/or other materials provided with the distribution.
16 * 4. Neither the name of the University nor the names of its contributors
17 * may be used to endorse or promote products derived from this software
18 * without specific prior written permission.
20 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
21 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
24 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32 * @(#)hash.h 8.3 (Berkeley) 5/31/94
38 HASH_GET, HASH_PUT, HASH_PUTNEW, HASH_DELETE, HASH_FIRST, HASH_NEXT
41 /* Buffer Management structures */
42 typedef struct _bufhead BUFHEAD;
45 BUFHEAD *prev; /* LRU links */
46 BUFHEAD *next; /* LRU links */
47 BUFHEAD *ovfl; /* Overflow page buffer header */
48 u_int32_t addr; /* Address of this page */
49 char *page; /* Actual page data */
51 #define BUF_MOD 0x0001
52 #define BUF_DISK 0x0002
53 #define BUF_BUCKET 0x0004
54 #define BUF_PIN 0x0008
57 #define IS_BUCKET(X) ((X) & BUF_BUCKET)
59 typedef BUFHEAD **SEGMENT;
61 /* Hash Table Information */
62 typedef struct hashhdr { /* Disk resident portion */
63 int32_t magic; /* Magic NO for hash tables */
64 int32_t version; /* Version ID */
65 u_int32_t lorder; /* Byte Order */
66 int32_t bsize; /* Bucket/Page Size */
67 int32_t bshift; /* Bucket shift */
68 int32_t dsize; /* Directory Size */
69 int32_t ssize; /* Segment Size */
70 int32_t sshift; /* Segment shift */
71 int32_t ovfl_point; /* Where overflow pages are being
73 int32_t last_freed; /* Last overflow page freed */
74 u_int32_t max_bucket; /* ID of Maximum bucket in use */
75 u_int32_t high_mask; /* Mask to modulo into entire table */
76 u_int32_t low_mask; /* Mask to modulo into lower half of
78 u_int32_t ffactor; /* Fill factor */
79 int32_t nkeys; /* Number of keys in hash table */
80 int32_t hdrpages; /* Size of table header */
81 int32_t h_charkey; /* value of hash(CHARKEY) */
82 #define NCACHED 32 /* number of bit maps and spare
84 int32_t spares[NCACHED];/* spare pages for overflow */
85 u_int16_t bitmaps[NCACHED]; /* address of overflow page
89 typedef struct htab { /* Memory resident data structure */
90 HASHHDR hdr; /* Header */
91 int nsegs; /* Number of allocated segments */
92 int exsegs; /* Number of extra allocated
94 u_int32_t /* Hash function */
95 (*hash)(const void *, size_t);
96 int flags; /* Flag values */
97 int fp; /* File pointer */
98 char *tmp_buf; /* Temporary Buffer for BIG data */
99 char *tmp_key; /* Temporary Buffer for BIG keys */
100 BUFHEAD *cpage; /* Current page */
101 int cbucket; /* Current bucket */
102 int cndx; /* Index of next item on cpage */
103 int error; /* Error Number -- for DBM
105 int new_file; /* Indicates if fd is backing store
107 int save_file; /* Indicates whether we need to flush
110 u_int32_t *mapp[NCACHED]; /* Pointers to page maps */
111 int nmaps; /* Initial number of bitmaps */
112 int nbufs; /* Number of buffers left to
114 BUFHEAD bufhead; /* Header of buffer lru list */
115 SEGMENT *dir; /* Hash Bucket directory */
121 #define MAX_BSIZE 32768 /* 2^15 but should be 65536 */
122 #define MIN_BUFFERS 6
123 #define MINHDRSIZE 512
124 #define DEF_BUFSIZE 65536 /* 64 K */
125 #define DEF_BUCKET_SIZE 4096
126 #define DEF_BUCKET_SHIFT 12 /* log2(BUCKET) */
127 #define DEF_SEGSIZE 256
128 #define DEF_SEGSIZE_SHIFT 8 /* log2(SEGSIZE) */
129 #define DEF_DIRSIZE 256
130 #define DEF_FFACTOR 65536
131 #define MIN_FFACTOR 4
133 #define CHARKEY "%$sniglet^&"
134 #define NUMKEY 1038583
136 #define INT_TO_BYTE 2
137 #define INT_BYTE_SHIFT 5
138 #define ALL_SET ((u_int32_t)0xFFFFFFFF)
141 #define PTROF(X) ((BUFHEAD *)((ptrdiff_t)(X)&~0x3))
142 #define ISMOD(X) ((u_int32_t)(ptrdiff_t)(X)&0x1)
143 #define DOMOD(X) ((X) = (char *)((ptrdiff_t)(X)|0x1))
144 #define ISDISK(X) ((u_int32_t)(ptrdiff_t)(X)&0x2)
145 #define DODISK(X) ((X) = (char *)((ptrdiff_t)(X)|0x2))
147 #define BITS_PER_MAP 32
149 /* Given the address of the beginning of a big map, clear/set the nth bit */
150 #define CLRBIT(A, N) ((A)[(N)/BITS_PER_MAP] &= ~(1<<((N)%BITS_PER_MAP)))
151 #define SETBIT(A, N) ((A)[(N)/BITS_PER_MAP] |= (1<<((N)%BITS_PER_MAP)))
152 #define ISSET(A, N) ((A)[(N)/BITS_PER_MAP] & (1<<((N)%BITS_PER_MAP)))
154 /* Overflow management */
156 * Overflow page numbers are allocated per split point. At each doubling of
157 * the table, we can allocate extra pages. So, an overflow page number has
158 * the top 5 bits indicate which split point and the lower 11 bits indicate
159 * which page at that split point is indicated (pages within split points are
160 * numberered starting with 1).
163 #define SPLITSHIFT 11
164 #define SPLITMASK 0x7FF
165 #define SPLITNUM(N) (((u_int32_t)(N)) >> SPLITSHIFT)
166 #define OPAGENUM(N) ((N) & SPLITMASK)
167 #define OADDR_OF(S,O) ((u_int32_t)((u_int32_t)(S) << SPLITSHIFT) + (O))
169 #define BUCKET_TO_PAGE(B) \
170 (B) + hashp->HDRPAGES + ((B) ? hashp->SPARES[__log2((B)+1)-1] : 0)
171 #define OADDR_TO_PAGE(B) \
172 BUCKET_TO_PAGE ( (1 << SPLITNUM((B))) -1 ) + OPAGENUM((B));
175 * page.h contains a detailed description of the page format.
177 * Normally, keys and data are accessed from offset tables in the top of
178 * each page which point to the beginning of the key and data. There are
179 * four flag values which may be stored in these offset tables which indicate
183 * OVFLPAGE Rather than a key data pair, this pair contains
184 * the address of an overflow page. The format of
186 * OVERFLOW_PAGE_NUMBER OVFLPAGE
188 * PARTIAL_KEY This must be the first key/data pair on a page
189 * and implies that page contains only a partial key.
190 * That is, the key is too big to fit on a single page
191 * so it starts on this page and continues on the next.
192 * The format of the page is:
193 * KEY_OFF PARTIAL_KEY OVFL_PAGENO OVFLPAGE
195 * KEY_OFF -- offset of the beginning of the key
197 * OVFL_PAGENO - page number of the next overflow page
200 * FULL_KEY This must be the first key/data pair on the page. It
201 * is used in two cases.
204 * There is a complete key on the page but no data
205 * (because it wouldn't fit). The next page contains
209 * KEY_OFF FULL_KEY OVFL_PAGENO OVFL_PAGE
211 * KEY_OFF -- offset of the beginning of the key
213 * OVFL_PAGENO - page number of the next overflow page
217 * This page contains no key, but part of a large
218 * data field, which is continued on the next page.
221 * DATA_OFF FULL_KEY OVFL_PAGENO OVFL_PAGE
223 * KEY_OFF -- offset of the beginning of the data on
226 * OVFL_PAGENO - page number of the next overflow page
230 * This must be the first key/data pair on the page.
231 * There are two cases:
234 * This page contains a key and the beginning of the
235 * data field, but the data field is continued on the
239 * KEY_OFF FULL_KEY_DATA OVFL_PAGENO DATA_OFF
241 * KEY_OFF -- offset of the beginning of the key
243 * OVFL_PAGENO - page number of the next overflow page
244 * DATA_OFF -- offset of the beginning of the data
247 * This page contains the last page of a big data pair.
248 * There is no key, only the tail end of the data
252 * DATA_OFF FULL_KEY_DATA <OVFL_PAGENO> <OVFLPAGE>
254 * DATA_OFF -- offset of the beginning of the data on
257 * OVFL_PAGENO - page number of the next overflow page
260 * OVFL_PAGENO and OVFLPAGE are optional (they are
261 * not present if there is no next page).
265 #define PARTIAL_KEY 1
267 #define FULL_KEY_DATA 3
270 /* Short hands for accessing structure */
271 #define BSIZE hdr.bsize
272 #define BSHIFT hdr.bshift
273 #define DSIZE hdr.dsize
274 #define SGSIZE hdr.ssize
275 #define SSHIFT hdr.sshift
276 #define LORDER hdr.lorder
277 #define OVFL_POINT hdr.ovfl_point
278 #define LAST_FREED hdr.last_freed
279 #define MAX_BUCKET hdr.max_bucket
280 #define FFACTOR hdr.ffactor
281 #define HIGH_MASK hdr.high_mask
282 #define LOW_MASK hdr.low_mask
283 #define NKEYS hdr.nkeys
284 #define HDRPAGES hdr.hdrpages
285 #define SPARES hdr.spares
286 #define BITMAPS hdr.bitmaps
287 #define VERSION hdr.version
288 #define MAGIC hdr.magic
289 #define NEXT_FREE hdr.next_free
290 #define H_CHARKEY hdr.h_charkey