aboutsummaryrefslogtreecommitdiff
path: root/db/hash/hash.c
blob: 08f2a7e3c8e3fc6c50140757fa9c0987ccf7babc (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
/*-
 * Copyright (c) 1990, 1993, 1994
 *	The Regents of the University of California.  All rights reserved.
 *
 * This code is derived from software contributed to Berkeley by
 * Margo Seltzer.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *	This product includes software developed by the University of
 *	California, Berkeley and its contributors.
 * 4. Neither the name of the University nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 */

#if defined(LIBC_SCCS) && !defined(lint)
static char sccsid[] = "@(#)hash.c	8.9 (Berkeley) 6/16/94";
#endif /* LIBC_SCCS and not lint */

#include <sys/param.h>
#include <sys/stat.h>

#include <errno.h>
#include <fcntl.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#ifdef DEBUG
#include <assert.h>
#endif

#include <db.h>
#include "hash.h"
#include "page.h"
#include "extern.h"

static int   alloc_segs __P((HTAB *, int));
static int   flush_meta __P((HTAB *));
static int   hash_access __P((HTAB *, ACTION, DBT *, DBT *));
static int   hash_close __P((DB *));
static int   hash_delete __P((const DB *, const DBT *, u_int32_t));
static int   hash_fd __P((const DB *));
static int   hash_get __P((const DB *, const DBT *, DBT *, u_int32_t));
static int   hash_put __P((const DB *, DBT *, const DBT *, u_int32_t));
static void *hash_realloc __P((SEGMENT **, int, int));
static int   hash_seq __P((const DB *, DBT *, DBT *, u_int32_t));
static int   hash_sync __P((const DB *, u_int32_t));
static int   hdestroy __P((HTAB *));
static HTAB *init_hash __P((HTAB *, const char *, HASHINFO *));
static int   init_htab __P((HTAB *, int));
#if BYTE_ORDER == LITTLE_ENDIAN
static void  swap_header __P((HTAB *));
static void  swap_header_copy __P((HASHHDR *, HASHHDR *));
#endif

/* Fast arithmetic, relying on powers of 2, */
#define MOD(x, y)		((x) & ((y) - 1))

#define RETURN_ERROR(ERR, LOC)	{ save_errno = ERR; goto LOC; }

/* Return values */
#define	SUCCESS	 (0)
#define	ERROR	(-1)
#define	ABNORMAL (1)

#ifdef HASH_STATISTICS
int hash_accesses, hash_collisions, hash_expansions, hash_overflows;
#endif

/************************** INTERFACE ROUTINES ***************************/
/* OPEN/CLOSE */

extern DB *
__hash_open(file, flags, mode, info, dflags)
	const char *file;
	int flags, mode, dflags;
	const HASHINFO *info;	/* Special directives for create */
{
	HTAB *hashp;
	struct stat statbuf;
	DB *dbp;
	int bpages, hdrsize, new_table, nsegs, save_errno;

	if ((flags & O_ACCMODE) == O_WRONLY) {
		errno = EINVAL;
		return (NULL);
	}

	if (!(hashp = (HTAB *)calloc(1, sizeof(HTAB))))
		return (NULL);
	hashp->fp = -1;

	/*
	 * Even if user wants write only, we need to be able to read
	 * the actual file, so we need to open it read/write. But, the
	 * field in the hashp structure needs to be accurate so that
	 * we can check accesses.
	 */
	hashp->flags = flags;

	new_table = 0;
	if (!file || (flags & O_TRUNC) ||
	    (stat(file, &statbuf) && (errno == ENOENT))) {
		if (errno == ENOENT)
			errno = 0; /* Just in case someone looks at errno */
		new_table = 1;
	}
	if (file) {
		if ((hashp->fp = open(file, flags, mode)) == -1)
			RETURN_ERROR(errno, error0);
		(void)fcntl(hashp->fp, F_SETFD, 1);
	}
	if (new_table) {
		if (!(hashp = init_hash(hashp, file, (HASHINFO *)info)))
			RETURN_ERROR(errno, error1);
	} else {
		/* Table already exists */
		if (info && info->hash)
			hashp->hash = info->hash;
		else
			hashp->hash = __default_hash;

		hdrsize = read(hashp->fp, &hashp->hdr, sizeof(HASHHDR));
#if BYTE_ORDER == LITTLE_ENDIAN
		swap_header(hashp);
#endif
		if (hdrsize == -1)
			RETURN_ERROR(errno, error1);
		if (hdrsize != sizeof(HASHHDR))
			RETURN_ERROR(EFTYPE, error1);
		/* Verify file type, versions and hash function */
		if (hashp->MAGIC != HASHMAGIC)
			RETURN_ERROR(EFTYPE, error1);
#define	OLDHASHVERSION	1
		if (hashp->VERSION != HASHVERSION &&
		    hashp->VERSION != OLDHASHVERSION)
			RETURN_ERROR(EFTYPE, error1);
		if (hashp->hash(CHARKEY, sizeof(CHARKEY))
		    != (u_int32_t) hashp->H_CHARKEY)
			RETURN_ERROR(EFTYPE, error1);
		/*
		 * Figure out how many segments we need.  Max_Bucket is the
		 * maximum bucket number, so the number of buckets is
		 * max_bucket + 1.
		 */
		nsegs = (hashp->MAX_BUCKET + 1 + hashp->SGSIZE - 1) /
			 hashp->SGSIZE;
		hashp->nsegs = 0;
		if (alloc_segs(hashp, nsegs))
			/*
			 * If alloc_segs fails, table will have been destroyed
			 * and errno will have been set.
			 */
			return (NULL);
		/* Read in bitmaps */
		bpages = (hashp->SPARES[hashp->OVFL_POINT] +
		    (hashp->BSIZE << BYTE_SHIFT) - 1) >>
		    (hashp->BSHIFT + BYTE_SHIFT);

		hashp->nmaps = bpages;
		(void)memset(&hashp->mapp[0], 0, bpages * sizeof(u_int32_t *));
	}

	/* Initialize Buffer Manager */
	if (info && info->cachesize)
		__buf_init(hashp, info->cachesize);
	else
		__buf_init(hashp, DEF_BUFSIZE);

	hashp->new_file = new_table;
	hashp->save_file = file && (hashp->flags & O_ACCMODE) != O_RDONLY;
	hashp->cbucket = -1;
	if (!(dbp = (DB *)malloc(sizeof(DB)))) {
		save_errno = errno;
		hdestroy(hashp);
		errno = save_errno;
		return (NULL);
	}
	dbp->internal = hashp;
	dbp->close = hash_close;
	dbp->del = hash_delete;
	dbp->fd = hash_fd;
	dbp->get = hash_get;
	dbp->put = hash_put;
	dbp->seq = hash_seq;
	dbp->sync = hash_sync;
	dbp->type = DB_HASH;

#ifdef DEBUG
	(void)fprintf(stderr,
"%s\n%s%x\n%s%d\n%s%d\n%s%d\n%s%d\n%s%d\n%s%d\n%s%d\n%s%d\n%s%d\n%s%x\n%s%x\n%s%d\n%s%d\n",
	    "init_htab:",
	    "TABLE POINTER   ", hashp,
	    "BUCKET SIZE     ", hashp->BSIZE,
	    "BUCKET SHIFT    ", hashp->BSHIFT,
	    "DIRECTORY SIZE  ", hashp->DSIZE,
	    "SEGMENT SIZE    ", hashp->SGSIZE,
	    "SEGMENT SHIFT   ", hashp->SSHIFT,
	    "FILL FACTOR     ", hashp->FFACTOR,
	    "MAX BUCKET      ", hashp->MAX_BUCKET,
	    "OVFL POINT	     ", hashp->OVFL_POINT,
	    "LAST FREED      ", hashp->LAST_FREED,
	    "HIGH MASK       ", hashp->HIGH_MASK,
	    "LOW  MASK       ", hashp->LOW_MASK,
	    "NSEGS           ", hashp->nsegs,
	    "NKEYS           ", hashp->NKEYS);
#endif
#ifdef HASH_STATISTICS
	hash_overflows = hash_accesses = hash_collisions = hash_expansions = 0;
#endif
	return (dbp);

error1:
	if (hashp != NULL)
		(void)close(hashp->fp);

error0:
	free(hashp);
	errno = save_errno;
	return (NULL);
}

static int
hash_close(dbp)
	DB *dbp;
{
	HTAB *hashp;
	int retval;

	if (!dbp)
		return (ERROR);

	hashp = (HTAB *)dbp->internal;
	retval = hdestroy(hashp);
	free(dbp);
	return (retval);
}

static int
hash_fd(dbp)
	const DB *dbp;
{
	HTAB *hashp;

	if (!dbp)
		return (ERROR);

	hashp = (HTAB *)dbp->internal;
	if (hashp->fp == -1) {
		errno = ENOENT;
		return (-1);
	}
	return (hashp->fp);
}

/************************** LOCAL CREATION ROUTINES **********************/
static HTAB *
init_hash(hashp, file, info)
	HTAB *hashp;
	const char *file;
	HASHINFO *info;
{
	struct stat statbuf;
	int nelem;

	nelem = 1;
	hashp->NKEYS = 0;
	hashp->LORDER = BYTE_ORDER;
	hashp->BSIZE = DEF_BUCKET_SIZE;
	hashp->BSHIFT = DEF_BUCKET_SHIFT;
	hashp->SGSIZE = DEF_SEGSIZE;
	hashp->SSHIFT = DEF_SEGSIZE_SHIFT;
	hashp->DSIZE = DEF_DIRSIZE;
	hashp->FFACTOR = DEF_FFACTOR;
	hashp->hash = __default_hash;
	memset(hashp->SPARES, 0, sizeof(hashp->SPARES));
	memset(hashp->BITMAPS, 0, sizeof (hashp->BITMAPS));

	/* Fix bucket size to be optimal for file system */
	if (file != NULL) {
		if (stat(file, &statbuf))
			return (NULL);
		hashp->BSIZE = statbuf.st_blksize;
		hashp->BSHIFT = __hash_log2(hashp->BSIZE);
	}

	if (info) {
		if (info->bsize) {
			/* Round pagesize up to power of 2 */
			hashp->BSHIFT = __hash_log2(info->bsize);
			hashp->BSIZE = 1 << hashp->BSHIFT;
			if (hashp->BSIZE > MAX_BSIZE) {
				errno = EINVAL;
				return (NULL);
			}
		}
		if (info->ffactor)
			hashp->FFACTOR = info->ffactor;
		if (info->hash)
			hashp->hash = info->hash;
		if (info->nelem)
			nelem = info->nelem;
		if (info->lorder) {
			if (info->lorder != BIG_ENDIAN &&
			    info->lorder != LITTLE_ENDIAN) {
				errno = EINVAL;
				return (NULL);
			}
			hashp->LORDER = info->lorder;
		}
	}
	/* init_htab should destroy the table and set errno if it fails */
	if (init_htab(hashp, nelem))
		return (NULL);
	else
		return (hashp);
}
/*
 * This calls alloc_segs which may run out of memory.  Alloc_segs will destroy
 * the table and set errno, so we just pass the error information along.
 *
 * Returns 0 on No Error
 */
static int
init_htab(hashp, nelem)
	HTAB *hashp;
	int nelem;
{
	register int nbuckets, nsegs;
	int l2;

	/*
	 * Divide number of elements by the fill factor and determine a
	 * desired number of buckets.  Allocate space for the next greater
	 * power of two number of buckets.
	 */
	nelem = (nelem - 1) / hashp->FFACTOR + 1;

	l2 = __hash_log2(MAX(nelem, 2));
	nbuckets = 1 << l2;

	hashp->SPARES[l2] = l2 + 1;
	hashp->SPARES[l2 + 1] = l2 + 1;
	hashp->OVFL_POINT = l2;
	hashp->LAST_FREED = 2;

	/* First bitmap page is at: splitpoint l2 page offset 1 */
	if (__ibitmap(hashp, OADDR_OF(l2, 1), l2 + 1, 0))
		return (-1);

	hashp->MAX_BUCKET = hashp->LOW_MASK = nbuckets - 1;
	hashp->HIGH_MASK = (nbuckets << 1) - 1;
	hashp->HDRPAGES = ((MAX(sizeof(HASHHDR), MINHDRSIZE) - 1) >>
	    hashp->BSHIFT) + 1;

	nsegs = (nbuckets - 1) / hashp->SGSIZE + 1;
	nsegs = 1 << __hash_log2(nsegs);

	if (nsegs > hashp->DSIZE)
		hashp->DSIZE = nsegs;
	return (alloc_segs(hashp, nsegs));
}

/********************** DESTROY/CLOSE ROUTINES ************************/

/*
 * Flushes any changes to the file if necessary and destroys the hashp
 * structure, freeing all allocated space.
 */
static int
hdestroy(hashp)
	HTAB *hashp;
{
	int i, save_errno;

	save_errno = 0;

#ifdef HASH_STATISTICS
	(void)fprintf(stderr, "hdestroy: accesses %ld collisions %ld\n",
	    hash_accesses, hash_collisions);
	(void)fprintf(stderr, "hdestroy: expansions %ld\n",
	    hash_expansions);
	(void)fprintf(stderr, "hdestroy: overflows %ld\n",
	    hash_overflows);
	(void)fprintf(stderr, "keys %ld maxp %d segmentcount %d\n",
	    hashp->NKEYS, hashp->MAX_BUCKET, hashp->nsegs);

	for (i = 0; i < NCACHED; i++)
		(void)fprintf(stderr,
		    "spares[%d] = %d\n", i, hashp->SPARES[i]);
#endif
	/*
	 * Call on buffer manager to free buffers, and if required,
	 * write them to disk.
	 */
	if (__buf_free(hashp, 1, hashp->save_file))
		save_errno = errno;
	if (hashp->dir) {
		free(*hashp->dir);	/* Free initial segments */
		/* Free extra segments */
		while (hashp->exsegs--)
			free(hashp->dir[--hashp->nsegs]);
		free(hashp->dir);
	}
	if (flush_meta(hashp) && !save_errno)
		save_errno = errno;
	/* Free Bigmaps */
	for (i = 0; i < hashp->nmaps; i++)
		if (hashp->mapp[i])
			free(hashp->mapp[i]);

	if (hashp->fp != -1)
		(void)close(hashp->fp);

	free(hashp);

	if (save_errno) {
		errno = save_errno;
		return (ERROR);
	}
	return (SUCCESS);
}
/*
 * Write modified pages to disk
 *
 * Returns:
 *	 0 == OK
 *	-1 ERROR
 */
static int
hash_sync(dbp, flags)
	const DB *dbp;
	u_int32_t flags;
{
	HTAB *hashp;

	if (flags != 0) {
		errno = EINVAL;
		return (ERROR);
	}

	if (!dbp)
		return (ERROR);

	hashp = (HTAB *)dbp->internal;
	if (!hashp->save_file)
		return (0);
	if (__buf_free(hashp, 0, 1) || flush_meta(hashp))
		return (ERROR);
	hashp->new_file = 0;
	return (0);
}

/*
 * Returns:
 *	 0 == OK
 *	-1 indicates that errno should be set
 */
static int
flush_meta(hashp)
	HTAB *hashp;
{
	HASHHDR *whdrp;
#if BYTE_ORDER == LITTLE_ENDIAN
	HASHHDR whdr;
#endif
	int fp, i, wsize;

	if (!hashp->save_file)
		return (0);
	hashp->MAGIC = HASHMAGIC;
	hashp->VERSION = HASHVERSION;
	hashp->H_CHARKEY = hashp->hash(CHARKEY, sizeof(CHARKEY));

	fp = hashp->fp;
	whdrp = &hashp->hdr;
#if BYTE_ORDER == LITTLE_ENDIAN
	whdrp = &whdr;
	swap_header_copy(&hashp->hdr, whdrp);
#endif
	if ((lseek(fp, (off_t)0, SEEK_SET) == -1) ||
	    ((wsize = write(fp, whdrp, sizeof(HASHHDR))) == -1))
		return (-1);
	else
		if (wsize != sizeof(HASHHDR)) {
			errno = EFTYPE;
			hashp->errnum = errno;
			return (-1);
		}
	for (i = 0; i < NCACHED; i++)
		if (hashp->mapp[i])
			if (__put_page(hashp, (char *)hashp->mapp[i],
				hashp->BITMAPS[i], 0, 1))
				return (-1);
	return (0);
}

/*******************************SEARCH ROUTINES *****************************/
/*
 * All the access routines return
 *
 * Returns:
 *	 0 on SUCCESS
 *	 1 to indicate an external ERROR (i.e. key not found, etc)
 *	-1 to indicate an internal ERROR (i.e. out of memory, etc)
 */
static int
hash_get(dbp, key, data, flag)
	const DB *dbp;
	const DBT *key;
	DBT *data;
	u_int32_t flag;
{
	HTAB *hashp;

	hashp = (HTAB *)dbp->internal;
	if (flag) {
		hashp->errnum = errno = EINVAL;
		return (ERROR);
	}
	return (hash_access(hashp, HASH_GET, (DBT *)key, data));
}

static int
hash_put(dbp, key, data, flag)
	const DB *dbp;
	DBT *key;
	const DBT *data;
	u_int32_t flag;
{
	HTAB *hashp;

	hashp = (HTAB *)dbp->internal;
	if (flag && flag != R_NOOVERWRITE) {
		hashp->errnum = errno = EINVAL;
		return (ERROR);
	}
	if ((hashp->flags & O_ACCMODE) == O_RDONLY) {
		hashp->errnum = errno = EPERM;
		return (ERROR);
	}
	return (hash_access(hashp, flag == R_NOOVERWRITE ?
	    HASH_PUTNEW : HASH_PUT, (DBT *)key, (DBT *)data));
}

static int
hash_delete(dbp, key, flag)
	const DB *dbp;
	const DBT *key;
	u_int32_t flag;		/* Ignored */
{
	HTAB *hashp;

	hashp = (HTAB *)dbp->internal;
	if (flag && flag != R_CURSOR) {
		hashp->errnum = errno = EINVAL;
		return (ERROR);
	}
	if ((hashp->flags & O_ACCMODE) == O_RDONLY) {
		hashp->errnum = errno = EPERM;
		return (ERROR);
	}
	return (hash_access(hashp, HASH_DELETE, (DBT *)key, NULL));
}

/*
 * Assume that hashp has been set in wrapper routine.
 */
static int
hash_access(hashp, action, key, val)
	HTAB *hashp;
	ACTION action;
	DBT *key, *val;
{
	register BUFHEAD *rbufp;
	BUFHEAD *bufp, *save_bufp;
	register u_int16_t *bp;
	register int n, ndx, off, size;
	register char *kp;
	u_int16_t pageno;

#ifdef HASH_STATISTICS
	hash_accesses++;
#endif

	off = hashp->BSIZE;
	size = key->size;
	kp = (char *)key->data;
	rbufp = __get_buf(hashp, __call_hash(hashp, kp, size), NULL, 0);
	if (!rbufp)
		return (ERROR);
	save_bufp = rbufp;

	/* Pin the bucket chain */
	rbufp->flags |= BUF_PIN;
	for (bp = (u_int16_t *)rbufp->page, n = *bp++, ndx = 1; ndx < n;)
		if (bp[1] >= REAL_KEY) {
			/* Real key/data pair */
			if (size == off - *bp &&
			    memcmp(kp, rbufp->page + *bp, size) == 0)
				goto found;
			off = bp[1];
#ifdef HASH_STATISTICS
			hash_collisions++;
#endif
			bp += 2;
			ndx += 2;
		} else if (bp[1] == OVFLPAGE) {
			rbufp = __get_buf(hashp, *bp, rbufp, 0);
			if (!rbufp) {
				save_bufp->flags &= ~BUF_PIN;
				return (ERROR);
			}
			/* FOR LOOP INIT */
			bp = (u_int16_t *)rbufp->page;
			n = *bp++;
			ndx = 1;
			off = hashp->BSIZE;
		} else if (bp[1] < REAL_KEY) {
			if ((ndx =
			    __find_bigpair(hashp, rbufp, ndx, kp, size)) > 0)
				goto found;
			if (ndx == -2) {
				bufp = rbufp;
				if (!(pageno =
				    __find_last_page(hashp, &bufp))) {
					ndx = 0;
					rbufp = bufp;
					break;	/* FOR */
				}
				rbufp = __get_buf(hashp, pageno, bufp, 0);
				if (!rbufp) {
					save_bufp->flags &= ~BUF_PIN;
					return (ERROR);
				}
				/* FOR LOOP INIT */
				bp = (u_int16_t *)rbufp->page;
				n = *bp++;
				ndx = 1;
				off = hashp->BSIZE;
			} else {
				save_bufp->flags &= ~BUF_PIN;
				return (ERROR);
			}
		}

	/* Not found */
	switch (action) {
	case HASH_PUT:
	case HASH_PUTNEW:
		if (__addel(hashp, rbufp, key, val)) {
			save_bufp->flags &= ~BUF_PIN;
			return (ERROR);
		} else {
			save_bufp->flags &= ~BUF_PIN;
			return (SUCCESS);
		}
	case HASH_GET:
	case HASH_DELETE:
	default:
		save_bufp->flags &= ~BUF_PIN;
		return (ABNORMAL);
	}

found:
	switch (action) {
	case HASH_PUTNEW:
		save_bufp->flags &= ~BUF_PIN;
		return (ABNORMAL);
	case HASH_GET:
		bp = (u_int16_t *)rbufp->page;
		if (bp[ndx + 1] < REAL_KEY) {
			if (__big_return(hashp, rbufp, ndx, val, 0))
				return (ERROR);
		} else {
			val->data = (u_char *)rbufp->page + (int)bp[ndx + 1];
			val->size = bp[ndx] - bp[ndx + 1];
		}
		break;
	case HASH_PUT:
		if ((__delpair(hashp, rbufp, ndx)) ||
		    (__addel(hashp, rbufp, key, val))) {
			save_bufp->flags &= ~BUF_PIN;
			return (ERROR);
		}
		break;
	case HASH_DELETE:
		if (__delpair(hashp, rbufp, ndx))
			return (ERROR);
		break;
	default:
		abort();
	}
	save_bufp->flags &= ~BUF_PIN;
	return (SUCCESS);
}

static int
hash_seq(dbp, key, data, flag)
	const DB *dbp;
	DBT *key, *data;
	u_int32_t flag;
{
	register u_int32_t bucket;
	register BUFHEAD *bufp;
	HTAB *hashp;
	u_int16_t *bp, ndx;

	hashp = (HTAB *)dbp->internal;
	if (flag && flag != R_FIRST && flag != R_NEXT) {
		hashp->errnum = errno = EINVAL;
		return (ERROR);
	}
#ifdef HASH_STATISTICS
	hash_accesses++;
#endif
	if ((hashp->cbucket < 0) || (flag == R_FIRST)) {
		hashp->cbucket = 0;
		hashp->cndx = 1;
		hashp->cpage = NULL;
	}

	for (bp = NULL; !bp || !bp[0]; ) {
		if (!(bufp = hashp->cpage)) {
			for (bucket = hashp->cbucket;
			    bucket <= (u_int32_t) hashp->MAX_BUCKET;
			    bucket++, hashp->cndx = 1) {
				bufp = __get_buf(hashp, bucket, NULL, 0);
				if (!bufp)
					return (ERROR);
				hashp->cpage = bufp;
				bp = (u_int16_t *)bufp->page;
				if (bp[0])
					break;
			}
			hashp->cbucket = bucket;
			if (hashp->cbucket > hashp->MAX_BUCKET) {
				hashp->cbucket = -1;
				return (ABNORMAL);
			}
		} else
			bp = (u_int16_t *)hashp->cpage->page;

#ifdef DEBUG
		assert(bp);
		assert(bufp);
#endif
		while (bp[hashp->cndx + 1] == OVFLPAGE) {
			bufp = hashp->cpage =
			    __get_buf(hashp, bp[hashp->cndx], bufp, 0);
			if (!bufp)
				return (ERROR);
			bp = (u_int16_t *)(bufp->page);
			hashp->cndx = 1;
		}
		if (!bp[0]) {
			hashp->cpage = NULL;
			++hashp->cbucket;
		}
	}
	ndx = hashp->cndx;
	if (bp[ndx + 1] < REAL_KEY) {
		if (__big_keydata(hashp, bufp, key, data, 1))
			return (ERROR);
	} else {
		key->data = (u_char *)hashp->cpage->page + bp[ndx];
		key->size = (ndx > 1 ? bp[ndx - 1] : hashp->BSIZE) - bp[ndx];
		data->data = (u_char *)hashp->cpage->page + bp[ndx + 1];
		data->size = bp[ndx] - bp[ndx + 1];
		ndx += 2;
		if (ndx > bp[0]) {
			hashp->cpage = NULL;
			hashp->cbucket++;
			hashp->cndx = 1;
		} else
			hashp->cndx = ndx;
	}
	return (SUCCESS);
}

/********************************* UTILITIES ************************/

/*
 * Returns:
 *	 0 ==> OK
 *	-1 ==> Error
 */
extern int
__expand_table(hashp)
	HTAB *hashp;
{
	u_int32_t old_bucket, new_bucket;
	int dirsize, new_segnum, spare_ndx;

#ifdef HASH_STATISTICS
	hash_expansions++;
#endif
	new_bucket = ++hashp->MAX_BUCKET;
	old_bucket = (hashp->MAX_BUCKET & hashp->LOW_MASK);

	new_segnum = new_bucket >> hashp->SSHIFT;

	/* Check if we need a new segment */
	if (new_segnum >= hashp->nsegs) {
		/* Check if we need to expand directory */
		if (new_segnum >= hashp->DSIZE) {
			/* Reallocate directory */
			dirsize = hashp->DSIZE * sizeof(SEGMENT *);
			if (!hash_realloc(&hashp->dir, dirsize, dirsize << 1))
				return (-1);
			hashp->DSIZE = dirsize << 1;
		}
		if ((hashp->dir[new_segnum] =
		    (SEGMENT)calloc(hashp->SGSIZE, sizeof(SEGMENT))) == NULL)
			return (-1);
		hashp->exsegs++;
		hashp->nsegs++;
	}
	/*
	 * If the split point is increasing (MAX_BUCKET's log base 2
	 * * increases), we need to copy the current contents of the spare
	 * split bucket to the next bucket.
	 */
	spare_ndx = __hash_log2(hashp->MAX_BUCKET + 1);
	if (spare_ndx > hashp->OVFL_POINT) {
		hashp->SPARES[spare_ndx] = hashp->SPARES[hashp->OVFL_POINT];
		hashp->OVFL_POINT = spare_ndx;
	}

	if (new_bucket > (u_int32_t) hashp->HIGH_MASK) {
		/* Starting a new doubling */
		hashp->LOW_MASK = hashp->HIGH_MASK;
		hashp->HIGH_MASK = new_bucket | hashp->LOW_MASK;
	}
	/* Relocate records to the new bucket */
	return (__split_page(hashp, old_bucket, new_bucket));
}

/*
 * If realloc guarantees that the pointer is not destroyed if the realloc
 * fails, then this routine can go away.
 */
static void *
hash_realloc(p_ptr, oldsize, newsize)
	SEGMENT **p_ptr;
	int oldsize, newsize;
{
	register void *p;

	if (p = malloc(newsize)) {
		memmove(p, *p_ptr, oldsize);
		memset((char *)p + oldsize, 0, newsize - oldsize);
		free(*p_ptr);
		*p_ptr = p;
	}
	return (p);
}

extern u_int32_t
__call_hash(hashp, k, len)
	HTAB *hashp;
	char *k;
	int len;
{
	int n, bucket;

	n = hashp->hash(k, len);
	bucket = n & hashp->HIGH_MASK;
	if (bucket > hashp->MAX_BUCKET)
		bucket = bucket & hashp->LOW_MASK;
	return (bucket);
}

/*
 * Allocate segment table.  On error, destroy the table and set errno.
 *
 * Returns 0 on success
 */
static int
alloc_segs(hashp, nsegs)
	HTAB *hashp;
	int nsegs;
{
	register int i;
	register SEGMENT store;

	int save_errno;

	if ((hashp->dir =
	    (SEGMENT *)calloc(hashp->DSIZE, sizeof(SEGMENT *))) == NULL) {
		save_errno = errno;
		(void)hdestroy(hashp);
		errno = save_errno;
		return (-1);
	}
	/* Allocate segments */
	if ((store =
	    (SEGMENT)calloc(nsegs << hashp->SSHIFT, sizeof(SEGMENT))) == NULL) {
		save_errno = errno;
		(void)hdestroy(hashp);
		errno = save_errno;
		return (-1);
	}
	for (i = 0; i < nsegs; i++, hashp->nsegs++)
		hashp->dir[i] = &store[i << hashp->SSHIFT];
	return (0);
}

#if BYTE_ORDER == LITTLE_ENDIAN
/*
 * Hashp->hdr needs to be byteswapped.
 */
static void
swap_header_copy(srcp, destp)
	HASHHDR *srcp, *destp;
{
	int i;

	P_32_COPY(srcp->magic, destp->magic);
	P_32_COPY(srcp->version, destp->version);
	P_32_COPY(srcp->lorder, destp->lorder);
	P_32_COPY(srcp->bsize, destp->bsize);
	P_32_COPY(srcp->bshift, destp->bshift);
	P_32_COPY(srcp->dsize, destp->dsize);
	P_32_COPY(srcp->ssize, destp->ssize);
	P_32_COPY(srcp->sshift, destp->sshift);
	P_32_COPY(srcp->ovfl_point, destp->ovfl_point);
	P_32_COPY(srcp->last_freed, destp->last_freed);
	P_32_COPY(srcp->max_bucket, destp->max_bucket);
	P_32_COPY(srcp->high_mask, destp->high_mask);
	P_32_COPY(srcp->low_mask, destp->low_mask);
	P_32_COPY(srcp->ffactor, destp->ffactor);
	P_32_COPY(srcp->nkeys, destp->nkeys);
	P_32_COPY(srcp->hdrpages, destp->hdrpages);
	P_32_COPY(srcp->h_charkey, destp->h_charkey);
	for (i = 0; i < NCACHED; i++) {
		P_32_COPY(srcp->spares[i], destp->spares[i]);
		P_16_COPY(srcp->bitmaps[i], destp->bitmaps[i]);
	}
}

static void
swap_header(hashp)
	HTAB *hashp;
{
	HASHHDR *hdrp;
	int i;

	hdrp = &hashp->hdr;

	M_32_SWAP(hdrp->magic);
	M_32_SWAP(hdrp->version);
	M_32_SWAP(hdrp->lorder);
	M_32_SWAP(hdrp->bsize);
	M_32_SWAP(hdrp->bshift);
	M_32_SWAP(hdrp->dsize);
	M_32_SWAP(hdrp->ssize);
	M_32_SWAP(hdrp->sshift);
	M_32_SWAP(hdrp->ovfl_point);
	M_32_SWAP(hdrp->last_freed);
	M_32_SWAP(hdrp->max_bucket);
	M_32_SWAP(hdrp->high_mask);
	M_32_SWAP(hdrp->low_mask);
	M_32_SWAP(hdrp->ffactor);
	M_32_SWAP(hdrp->nkeys);
	M_32_SWAP(hdrp->hdrpages);
	M_32_SWAP(hdrp->h_charkey);
	for (i = 0; i < NCACHED; i++) {
		M_32_SWAP(hdrp->spares[i]);
		M_16_SWAP(hdrp->bitmaps[i]);
	}
}
#endif