Home | History | Annotate | Line # | Download | only in hash
      1 /*	$NetBSD: hash.h,v 1.17 2020/02/21 22:04:06 kamil Exp $	*/
      2 
      3 /*-
      4  * Copyright (c) 1990, 1993, 1994
      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  *	@(#)hash.h	8.3 (Berkeley) 5/31/94
     35  */
     36 
     37 #if HAVE_NBTOOL_CONFIG_H
     38 #include "nbtool_config.h"
     39 #endif
     40 
     41 /* Operations */
     42 typedef enum {
     43 	HASH_GET, HASH_PUT, HASH_PUTNEW, HASH_DELETE, HASH_FIRST, HASH_NEXT
     44 } ACTION;
     45 
     46 /* Buffer Management structures */
     47 typedef struct _bufhead BUFHEAD;
     48 
     49 struct _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 */
     55 	char	 	flags;
     56 #define	BUF_MOD		0x0001
     57 #define BUF_DISK	0x0002
     58 #define	BUF_BUCKET	0x0004
     59 #define	BUF_PIN		0x0008
     60 };
     61 
     62 #define IS_BUCKET(X)	((X) & BUF_BUCKET)
     63 
     64 typedef BUFHEAD **SEGMENT;
     65 
     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
     77 					 * allocated */
     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
     82 					 * table */
     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
     88 					 * points */
     89 	int32_t		spares[NCACHED];/* spare pages for overflow */
     90 	uint16_t	bitmaps[NCACHED];	/* address of overflow page
     91 						 * bitmaps */
     92 } HASHHDR;
     93 
     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
     98 					 * segments */
     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
    108 					 * compatibility */
    109 	int		new_file;	/* Indicates if fd is backing store
    110 					 * or no */
    111 	int		save_file;	/* Indicates whether we need to flush
    112 					 * file at
    113 					 * exit */
    114 	uint32_t	*mapp[NCACHED];	/* Pointers to page maps */
    115 	int		nmaps;		/* Initial number of bitmaps */
    116 	int		nbufs;		/* Number of buffers left to
    117 					 * allocate */
    118 	BUFHEAD 	bufhead;	/* Header of buffer lru list */
    119 	SEGMENT 	*dir;		/* Hash Bucket directory */
    120 } HTAB;
    121 
    122 /*
    123  * Constants
    124  */
    125 #define	MAX_BSIZE		65536		/* 2^16 */
    126 /*
    127  * Make it fit in uint16_t; a better way would be to store size - 1, but
    128  * then we'd need to bump the version.
    129  */
    130 #define HASH_BSIZE(hp)	((hp)->BSIZE == MAX_BSIZE ? MAX_BSIZE - 1 : (hp)->BSIZE)
    131 #define MIN_BUFFERS		6
    132 #define MINHDRSIZE		512
    133 #define DEF_BUFSIZE		65536		/* 64 K */
    134 #define DEF_BUCKET_SIZE		4096
    135 #define DEF_BUCKET_SHIFT	12		/* log2(BUCKET) */
    136 #define DEF_SEGSIZE		256
    137 #define DEF_SEGSIZE_SHIFT	8		/* log2(SEGSIZE)	 */
    138 #define DEF_DIRSIZE		256
    139 #define DEF_FFACTOR		65536
    140 #define MIN_FFACTOR		4
    141 #define SPLTMAX			8
    142 #define CHARKEY			"%$sniglet^&"
    143 #define NUMKEY			1038583
    144 #define BYTE_SHIFT		3
    145 #define INT_TO_BYTE		2
    146 #define INT_BYTE_SHIFT		5
    147 #define ALL_SET			((uint32_t)0xFFFFFFFF)
    148 #define ALL_CLEAR		0
    149 
    150 #define PTROF(X)	((BUFHEAD *)(void *)((u_long)(X)&~0x3))
    151 #define ISMOD(X)	((uint32_t)(u_long)(X)&0x1)
    152 #define DOMOD(X)	((X) = (char *)(void *)((u_long)(X)|0x1))
    153 #define ISDISK(X)	((uint32_t)(u_long)(X)&0x2)
    154 #define DODISK(X)	((X) = (char *)(void *)((u_long)(X)|0x2))
    155 
    156 #define BITS_PER_MAP	32
    157 
    158 /* Given the address of the beginning of a big map, clear/set the nth bit */
    159 #define CLRBIT(A, N)	((A)[(N)/BITS_PER_MAP] &= ~(1U<<((N)%BITS_PER_MAP)))
    160 #define SETBIT(A, N)	((A)[(N)/BITS_PER_MAP] |= (1U<<((N)%BITS_PER_MAP)))
    161 #define ISSET(A, N)	((A)[(N)/BITS_PER_MAP] & (1U<<((N)%BITS_PER_MAP)))
    162 
    163 /* Overflow management */
    164 /*
    165  * Overflow page numbers are allocated per split point.  At each doubling of
    166  * the table, we can allocate extra pages.  So, an overflow page number has
    167  * the top 5 bits indicate which split point and the lower 11 bits indicate
    168  * which page at that split point is indicated (pages within split points are
    169  * numberered starting with 1).
    170  */
    171 
    172 #define SPLITSHIFT	11
    173 #define SPLITMASK	0x7FF
    174 #define SPLITNUM(N)	(((uint32_t)(N)) >> SPLITSHIFT)
    175 #define OPAGENUM(N)	((N) & SPLITMASK)
    176 #define	OADDR_OF(S,O)	((uint32_t)((uint32_t)(S) << SPLITSHIFT) + (O))
    177 
    178 #define BUCKET_TO_PAGE(B) \
    179 	(B) + hashp->HDRPAGES + \
    180 	((B) ? hashp->SPARES[__log2((uint32_t)((B)+1))-1] : 0)
    181 #define OADDR_TO_PAGE(B) 	\
    182 	BUCKET_TO_PAGE ( (1 << SPLITNUM((B))) -1 ) + OPAGENUM((B));
    183 
    184 /*
    185  * page.h contains a detailed description of the page format.
    186  *
    187  * Normally, keys and data are accessed from offset tables in the top of
    188  * each page which point to the beginning of the key and data.  There are
    189  * four flag values which may be stored in these offset tables which indicate
    190  * the following:
    191  *
    192  *
    193  * OVFLPAGE	Rather than a key data pair, this pair contains
    194  *		the address of an overflow page.  The format of
    195  *		the pair is:
    196  *		    OVERFLOW_PAGE_NUMBER OVFLPAGE
    197  *
    198  * PARTIAL_KEY	This must be the first key/data pair on a page
    199  *		and implies that page contains only a partial key.
    200  *		That is, the key is too big to fit on a single page
    201  *		so it starts on this page and continues on the next.
    202  *		The format of the page is:
    203  *		    KEY_OFF PARTIAL_KEY OVFL_PAGENO OVFLPAGE
    204  *
    205  *		    KEY_OFF -- offset of the beginning of the key
    206  *		    PARTIAL_KEY -- 1
    207  *		    OVFL_PAGENO - page number of the next overflow page
    208  *		    OVFLPAGE -- 0
    209  *
    210  * FULL_KEY	This must be the first key/data pair on the page.  It
    211  *		is used in two cases.
    212  *
    213  *		Case 1:
    214  *		    There is a complete key on the page but no data
    215  *		    (because it wouldn't fit).  The next page contains
    216  *		    the data.
    217  *
    218  *		    Page format it:
    219  *		    KEY_OFF FULL_KEY OVFL_PAGENO OVFL_PAGE
    220  *
    221  *		    KEY_OFF -- offset of the beginning of the key
    222  *		    FULL_KEY -- 2
    223  *		    OVFL_PAGENO - page number of the next overflow page
    224  *		    OVFLPAGE -- 0
    225  *
    226  *		Case 2:
    227  *		    This page contains no key, but part of a large
    228  *		    data field, which is continued on the next page.
    229  *
    230  *		    Page format it:
    231  *		    DATA_OFF FULL_KEY OVFL_PAGENO OVFL_PAGE
    232  *
    233  *		    KEY_OFF -- offset of the beginning of the data on
    234  *				this page
    235  *		    FULL_KEY -- 2
    236  *		    OVFL_PAGENO - page number of the next overflow page
    237  *		    OVFLPAGE -- 0
    238  *
    239  * FULL_KEY_DATA
    240  *		This must be the first key/data pair on the page.
    241  *		There are two cases:
    242  *
    243  *		Case 1:
    244  *		    This page contains a key and the beginning of the
    245  *		    data field, but the data field is continued on the
    246  *		    next page.
    247  *
    248  *		    Page format is:
    249  *		    KEY_OFF FULL_KEY_DATA OVFL_PAGENO DATA_OFF
    250  *
    251  *		    KEY_OFF -- offset of the beginning of the key
    252  *		    FULL_KEY_DATA -- 3
    253  *		    OVFL_PAGENO - page number of the next overflow page
    254  *		    DATA_OFF -- offset of the beginning of the data
    255  *
    256  *		Case 2:
    257  *		    This page contains the last page of a big data pair.
    258  *		    There is no key, only the  tail end of the data
    259  *		    on this page.
    260  *
    261  *		    Page format is:
    262  *		    DATA_OFF FULL_KEY_DATA <OVFL_PAGENO> <OVFLPAGE>
    263  *
    264  *		    DATA_OFF -- offset of the beginning of the data on
    265  *				this page
    266  *		    FULL_KEY_DATA -- 3
    267  *		    OVFL_PAGENO - page number of the next overflow page
    268  *		    OVFLPAGE -- 0
    269  *
    270  *		    OVFL_PAGENO and OVFLPAGE are optional (they are
    271  *		    not present if there is no next page).
    272  */
    273 
    274 #define OVFLPAGE	0
    275 #define PARTIAL_KEY	1
    276 #define FULL_KEY	2
    277 #define FULL_KEY_DATA	3
    278 #define	REAL_KEY	4
    279 
    280 /* Short hands for accessing structure */
    281 #define BSIZE		hdr.bsize
    282 #define BSHIFT		hdr.bshift
    283 #define DSIZE		hdr.dsize
    284 #define SGSIZE		hdr.ssize
    285 #define SSHIFT		hdr.sshift
    286 #define LORDER		hdr.lorder
    287 #define OVFL_POINT	hdr.ovfl_point
    288 #define	LAST_FREED	hdr.last_freed
    289 #define MAX_BUCKET	hdr.max_bucket
    290 #define FFACTOR		hdr.ffactor
    291 #define HIGH_MASK	hdr.high_mask
    292 #define LOW_MASK	hdr.low_mask
    293 #define NKEYS		hdr.nkeys
    294 #define HDRPAGES	hdr.hdrpages
    295 #define SPARES		hdr.spares
    296 #define BITMAPS		hdr.bitmaps
    297 #define VERSION		hdr.version
    298 #define MAGIC		hdr.magic
    299 #define NEXT_FREE	hdr.next_free
    300 #define H_CHARKEY	hdr.h_charkey
    301