summaryrefslogtreecommitdiff
path: root/lib/tdb/common/tdb.c
blob: b78f74cc693e9f4fd3b26ba2e06b078f31bb0965 (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
 /* 
   Unix SMB/CIFS implementation.

   trivial database library

   Copyright (C) Andrew Tridgell              1999-2005
   Copyright (C) Paul `Rusty' Russell		   2000
   Copyright (C) Jeremy Allison			   2000-2003
   
     ** NOTE! The following LGPL license applies to the tdb
     ** library. This does NOT imply that all of Samba is released
     ** under the LGPL
   
   This library is free software; you can redistribute it and/or
   modify it under the terms of the GNU Lesser General Public
   License as published by the Free Software Foundation; either
   version 3 of the License, or (at your option) any later version.

   This library is distributed in the hope that it will be useful,
   but WITHOUT ANY WARRANTY; without even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
   Lesser General Public License for more details.

   You should have received a copy of the GNU Lesser General Public
   License along with this library; if not, see <http://www.gnu.org/licenses/>.
*/

#include "tdb_private.h"

TDB_DATA tdb_null;

/*
  non-blocking increment of the tdb sequence number if the tdb has been opened using
  the TDB_SEQNUM flag
*/
void tdb_increment_seqnum_nonblock(struct tdb_context *tdb)
{
	tdb_off_t seqnum=0;
	
	if (!(tdb->flags & TDB_SEQNUM)) {
		return;
	}

	/* we ignore errors from this, as we have no sane way of
	   dealing with them.
	*/
	tdb_ofs_read(tdb, TDB_SEQNUM_OFS, &seqnum);
	seqnum++;
	tdb_ofs_write(tdb, TDB_SEQNUM_OFS, &seqnum);
}

/*
  increment the tdb sequence number if the tdb has been opened using
  the TDB_SEQNUM flag
*/
static void tdb_increment_seqnum(struct tdb_context *tdb)
{
	if (!(tdb->flags & TDB_SEQNUM)) {
		return;
	}

	if (tdb_brlock(tdb, TDB_SEQNUM_OFS, F_WRLCK, F_SETLKW, 1, 1) != 0) {
		return;
	}

	tdb_increment_seqnum_nonblock(tdb);

	tdb_brlock(tdb, TDB_SEQNUM_OFS, F_UNLCK, F_SETLKW, 1, 1);
}

static int tdb_key_compare(TDB_DATA key, TDB_DATA data, void *private_data)
{
	return memcmp(data.dptr, key.dptr, data.dsize);
}

/* Returns 0 on fail.  On success, return offset of record, and fills
   in rec */
static tdb_off_t tdb_find(struct tdb_context *tdb, TDB_DATA key, uint32_t hash,
			struct list_struct *r)
{
	tdb_off_t rec_ptr;
	
	/* read in the hash top */
	if (tdb_ofs_read(tdb, TDB_HASH_TOP(hash), &rec_ptr) == -1)
		return 0;

	/* keep looking until we find the right record */
	while (rec_ptr) {
		if (tdb_rec_read(tdb, rec_ptr, r) == -1)
			return 0;

		if (!TDB_DEAD(r) && hash==r->full_hash
		    && key.dsize==r->key_len
		    && tdb_parse_data(tdb, key, rec_ptr + sizeof(*r),
				      r->key_len, tdb_key_compare,
				      NULL) == 0) {
			return rec_ptr;
		}
		/* detect tight infinite loop */
		if (rec_ptr == r->next) {
			TDB_LOG((tdb, TDB_DEBUG_FATAL, "tdb_find: loop detected.\n"));
			return TDB_ERRCODE(TDB_ERR_CORRUPT, 0);
		}
		rec_ptr = r->next;
	}
	return TDB_ERRCODE(TDB_ERR_NOEXIST, 0);
}

/* As tdb_find, but if you succeed, keep the lock */
tdb_off_t tdb_find_lock_hash(struct tdb_context *tdb, TDB_DATA key, uint32_t hash, int locktype,
			   struct list_struct *rec)
{
	uint32_t rec_ptr;

	if (tdb_lock(tdb, BUCKET(hash), locktype) == -1)
		return 0;
	if (!(rec_ptr = tdb_find(tdb, key, hash, rec)))
		tdb_unlock(tdb, BUCKET(hash), locktype);
	return rec_ptr;
}


/* update an entry in place - this only works if the new data size
   is <= the old data size and the key exists.
   on failure return -1.
*/
static int tdb_update_hash(struct tdb_context *tdb, TDB_DATA key, uint32_t hash, TDB_DATA dbuf)
{
	struct list_struct rec;
	tdb_off_t rec_ptr;

	/* find entry */
	if (!(rec_ptr = tdb_find(tdb, key, hash, &rec)))
		return -1;

	/* must be long enough key, data and tailer */
	if (rec.rec_len < key.dsize + dbuf.dsize + sizeof(tdb_off_t)) {
		tdb->ecode = TDB_SUCCESS; /* Not really an error */
		return -1;
	}

	if (tdb->methods->tdb_write(tdb, rec_ptr + sizeof(rec) + rec.key_len,
		      dbuf.dptr, dbuf.dsize) == -1)
		return -1;

	if (dbuf.dsize != rec.data_len) {
		/* update size */
		rec.data_len = dbuf.dsize;
		return tdb_rec_write(tdb, rec_ptr, &rec);
	}
 
	return 0;
}

/* find an entry in the database given a key */
/* If an entry doesn't exist tdb_err will be set to
 * TDB_ERR_NOEXIST. If a key has no data attached
 * then the TDB_DATA will have zero length but
 * a non-zero pointer
 */
TDB_DATA tdb_fetch(struct tdb_context *tdb, TDB_DATA key)
{
	tdb_off_t rec_ptr;
	struct list_struct rec;
	TDB_DATA ret;
	uint32_t hash;

	/* find which hash bucket it is in */
	hash = tdb->hash_fn(&key);
	if (!(rec_ptr = tdb_find_lock_hash(tdb,key,hash,F_RDLCK,&rec)))
		return tdb_null;

	ret.dptr = tdb_alloc_read(tdb, rec_ptr + sizeof(rec) + rec.key_len,
				  rec.data_len);
	ret.dsize = rec.data_len;
	tdb_unlock(tdb, BUCKET(rec.full_hash), F_RDLCK);
	return ret;
}

/*
 * Find an entry in the database and hand the record's data to a parsing
 * function. The parsing function is executed under the chain read lock, so it
 * should be fast and should not block on other syscalls.
 *
 * DONT CALL OTHER TDB CALLS FROM THE PARSER, THIS MIGHT LEAD TO SEGFAULTS.
 *
 * For mmapped tdb's that do not have a transaction open it points the parsing
 * function directly at the mmap area, it avoids the malloc/memcpy in this
 * case. If a transaction is open or no mmap is available, it has to do
 * malloc/read/parse/free.
 *
 * This is interesting for all readers of potentially large data structures in
 * the tdb records, ldb indexes being one example.
 */

int tdb_parse_record(struct tdb_context *tdb, TDB_DATA key,
		     int (*parser)(TDB_DATA key, TDB_DATA data,
				   void *private_data),
		     void *private_data)
{
	tdb_off_t rec_ptr;
	struct list_struct rec;
	int ret;
	uint32_t hash;

	/* find which hash bucket it is in */
	hash = tdb->hash_fn(&key);

	if (!(rec_ptr = tdb_find_lock_hash(tdb,key,hash,F_RDLCK,&rec))) {
		return TDB_ERRCODE(TDB_ERR_NOEXIST, 0);
	}

	ret = tdb_parse_data(tdb, key, rec_ptr + sizeof(rec) + rec.key_len,
			     rec.data_len, parser, private_data);

	tdb_unlock(tdb, BUCKET(rec.full_hash), F_RDLCK);

	return ret;
}

/* check if an entry in the database exists 

   note that 1 is returned if the key is found and 0 is returned if not found
   this doesn't match the conventions in the rest of this module, but is
   compatible with gdbm
*/
static int tdb_exists_hash(struct tdb_context *tdb, TDB_DATA key, uint32_t hash)
{
	struct list_struct rec;
	
	if (tdb_find_lock_hash(tdb, key, hash, F_RDLCK, &rec) == 0)
		return 0;
	tdb_unlock(tdb, BUCKET(rec.full_hash), F_RDLCK);
	return 1;
}

int tdb_exists(struct tdb_context *tdb, TDB_DATA key)
{
	uint32_t hash = tdb->hash_fn(&key);
	return tdb_exists_hash(tdb, key, hash);
}

/* actually delete an entry in the database given the offset */
int tdb_do_delete(struct tdb_context *tdb, tdb_off_t rec_ptr, struct list_struct *rec)
{
	tdb_off_t last_ptr, i;
	struct list_struct lastrec;

	if (tdb->read_only || tdb->traverse_read) return -1;

	if (((tdb->traverse_write != 0) && (!TDB_DEAD(rec))) ||
	    tdb_write_lock_record(tdb, rec_ptr) == -1) {
		/* Someone traversing here: mark it as dead */
		rec->magic = TDB_DEAD_MAGIC;
		return tdb_rec_write(tdb, rec_ptr, rec);
	}
	if (tdb_write_unlock_record(tdb, rec_ptr) != 0)
		return -1;

	/* find previous record in hash chain */
	if (tdb_ofs_read(tdb, TDB_HASH_TOP(rec->full_hash), &i) == -1)
		return -1;
	for (last_ptr = 0; i != rec_ptr; last_ptr = i, i = lastrec.next)
		if (tdb_rec_read(tdb, i, &lastrec) == -1)
			return -1;

	/* unlink it: next ptr is at start of record. */
	if (last_ptr == 0)
		last_ptr = TDB_HASH_TOP(rec->full_hash);
	if (tdb_ofs_write(tdb, last_ptr, &rec->next) == -1)
		return -1;

	/* recover the space */
	if (tdb_free(tdb, rec_ptr, rec) == -1)
		return -1;
	return 0;
}

static int tdb_count_dead(struct tdb_context *tdb, uint32_t hash)
{
	int res = 0;
	tdb_off_t rec_ptr;
	struct list_struct rec;
	
	/* read in the hash top */
	if (tdb_ofs_read(tdb, TDB_HASH_TOP(hash), &rec_ptr) == -1)
		return 0;

	while (rec_ptr) {
		if (tdb_rec_read(tdb, rec_ptr, &rec) == -1)
			return 0;

		if (rec.magic == TDB_DEAD_MAGIC) {
			res += 1;
		}
		rec_ptr = rec.next;
	}
	return res;
}

/*
 * Purge all DEAD records from a hash chain
 */
static int tdb_purge_dead(struct tdb_context *tdb, uint32_t hash)
{
	int res = -1;
	struct list_struct rec;
	tdb_off_t rec_ptr;

	if (tdb_lock(tdb, -1, F_WRLCK) == -1) {
		return -1;
	}
	
	/* read in the hash top */
	if (tdb_ofs_read(tdb, TDB_HASH_TOP(hash), &rec_ptr) == -1)
		goto fail;

	while (rec_ptr) {
		tdb_off_t next;

		if (tdb_rec_read(tdb, rec_ptr, &rec) == -1) {
			goto fail;
		}

		next = rec.next;

		if (rec.magic == TDB_DEAD_MAGIC
		    && tdb_do_delete(tdb, rec_ptr, &rec) == -1) {
			goto fail;
		}
		rec_ptr = next;
	}
	res = 0;
 fail:
	tdb_unlock(tdb, -1, F_WRLCK);
	return res;
}

/* delete an entry in the database given a key */
static int tdb_delete_hash(struct tdb_context *tdb, TDB_DATA key, uint32_t hash)
{
	tdb_off_t rec_ptr;
	struct list_struct rec;
	int ret;

	if (tdb->max_dead_records != 0) {

		/*
		 * Allow for some dead records per hash chain, mainly for
		 * tdb's with a very high create/delete rate like locking.tdb.
		 */

		if (tdb_lock(tdb, BUCKET(hash), F_WRLCK) == -1)
			return -1;

		if (tdb_count_dead(tdb, hash) >= tdb->max_dead_records) {
			/*
			 * Don't let the per-chain freelist grow too large,
			 * delete all existing dead records
			 */
			tdb_purge_dead(tdb, hash);
		}

		if (!(rec_ptr = tdb_find(tdb, key, hash, &rec))) {
			tdb_unlock(tdb, BUCKET(hash), F_WRLCK);
			return -1;
		}

		/*
		 * Just mark the record as dead.
		 */
		rec.magic = TDB_DEAD_MAGIC;
		ret = tdb_rec_write(tdb, rec_ptr, &rec);
	}
	else {
		if (!(rec_ptr = tdb_find_lock_hash(tdb, key, hash, F_WRLCK,
						   &rec)))
			return -1;

		ret = tdb_do_delete(tdb, rec_ptr, &rec);
	}

	if (ret == 0) {
		tdb_increment_seqnum(tdb);
	}

	if (tdb_unlock(tdb, BUCKET(rec.full_hash), F_WRLCK) != 0)
		TDB_LOG((tdb, TDB_DEBUG_WARNING, "tdb_delete: WARNING tdb_unlock failed!\n"));
	return ret;
}

int tdb_delete(struct tdb_context *tdb, TDB_DATA key)
{
	uint32_t hash = tdb->hash_fn(&key);
	return tdb_delete_hash(tdb, key, hash);
}

/*
 * See if we have a dead record around with enough space
 */
static tdb_off_t tdb_find_dead(struct tdb_context *tdb, uint32_t hash,
			       struct list_struct *r, tdb_len_t length)
{
	tdb_off_t rec_ptr;
	
	/* read in the hash top */
	if (tdb_ofs_read(tdb, TDB_HASH_TOP(hash), &rec_ptr) == -1)
		return 0;

	/* keep looking until we find the right record */
	while (rec_ptr) {
		if (tdb_rec_read(tdb, rec_ptr, r) == -1)
			return 0;

		if (TDB_DEAD(r) && r->rec_len >= length) {
			/*
			 * First fit for simple coding, TODO: change to best
			 * fit
			 */
			return rec_ptr;
		}
		rec_ptr = r->next;
	}
	return 0;
}

/* store an element in the database, replacing any existing element
   with the same key 

   return 0 on success, -1 on failure
*/
int tdb_store(struct tdb_context *tdb, TDB_DATA key, TDB_DATA dbuf, int flag)
{
	struct list_struct rec;
	uint32_t hash;
	tdb_off_t rec_ptr;
	char *p = NULL;
	int ret = -1;

	if (tdb->read_only || tdb->traverse_read) {
		tdb->ecode = TDB_ERR_RDONLY;
		return -1;
	}

	/* find which hash bucket it is in */
	hash = tdb->hash_fn(&key);
	if (tdb_lock(tdb, BUCKET(hash), F_WRLCK) == -1)
		return -1;

	/* check for it existing, on insert. */
	if (flag == TDB_INSERT) {
		if (tdb_exists_hash(tdb, key, hash)) {
			tdb->ecode = TDB_ERR_EXISTS;
			goto fail;
		}
	} else {
		/* first try in-place update, on modify or replace. */
		if (tdb_update_hash(tdb, key, hash, dbuf) == 0) {
			goto done;
		}
		if (tdb->ecode == TDB_ERR_NOEXIST &&
		    flag == TDB_MODIFY) {
			/* if the record doesn't exist and we are in TDB_MODIFY mode then
			 we should fail the store */
			goto fail;
		}
	}
	/* reset the error code potentially set by the tdb_update() */
	tdb->ecode = TDB_SUCCESS;

	/* delete any existing record - if it doesn't exist we don't
           care.  Doing this first reduces fragmentation, and avoids
           coalescing with `allocated' block before it's updated. */
	if (flag != TDB_INSERT)
		tdb_delete_hash(tdb, key, hash);

	/* Copy key+value *before* allocating free space in case malloc
	   fails and we are left with a dead spot in the tdb. */

	if (!(p = (char *)malloc(key.dsize + dbuf.dsize))) {
		tdb->ecode = TDB_ERR_OOM;
		goto fail;
	}

	memcpy(p, key.dptr, key.dsize);
	if (dbuf.dsize)
		memcpy(p+key.dsize, dbuf.dptr, dbuf.dsize);

	if (tdb->max_dead_records != 0) {
		/*
		 * Allow for some dead records per hash chain, look if we can
		 * find one that can hold the new record. We need enough space
		 * for key, data and tailer. If we find one, we don't have to
		 * consult the central freelist.
		 */
		rec_ptr = tdb_find_dead(
			tdb, hash, &rec,
			key.dsize + dbuf.dsize + sizeof(tdb_off_t));

		if (rec_ptr != 0) {
			rec.key_len = key.dsize;
			rec.data_len = dbuf.dsize;
			rec.full_hash = hash;
			rec.magic = TDB_MAGIC;
			if (tdb_rec_write(tdb, rec_ptr, &rec) == -1
			    || tdb->methods->tdb_write(
				    tdb, rec_ptr + sizeof(rec),
				    p, key.dsize + dbuf.dsize) == -1) {
				goto fail;
			}
			goto done;
		}
	}

	/*
	 * We have to allocate some space from the freelist, so this means we
	 * have to lock it. Use the chance to purge all the DEAD records from
	 * the hash chain under the freelist lock.
	 */

	if (tdb_lock(tdb, -1, F_WRLCK) == -1) {
		goto fail;
	}

	if ((tdb->max_dead_records != 0)
	    && (tdb_purge_dead(tdb, hash) == -1)) {
		tdb_unlock(tdb, -1, F_WRLCK);
		goto fail;
	}

	/* we have to allocate some space */
	rec_ptr = tdb_allocate(tdb, key.dsize + dbuf.dsize, &rec);

	tdb_unlock(tdb, -1, F_WRLCK);

	if (rec_ptr == 0) {
		goto fail;
	}

	/* Read hash top into next ptr */
	if (tdb_ofs_read(tdb, TDB_HASH_TOP(hash), &rec.next) == -1)
		goto fail;

	rec.key_len = key.dsize;
	rec.data_len = dbuf.dsize;
	rec.full_hash = hash;
	rec.magic = TDB_MAGIC;

	/* write out and point the top of the hash chain at it */
	if (tdb_rec_write(tdb, rec_ptr, &rec) == -1
	    || tdb->methods->tdb_write(tdb, rec_ptr+sizeof(rec), p, key.dsize+dbuf.dsize)==-1
	    || tdb_ofs_write(tdb, TDB_HASH_TOP(hash), &rec_ptr) == -1) {
		/* Need to tdb_unallocate() here */
		goto fail;
	}

 done:
	ret = 0;
 fail:
	if (ret == 0) {
		tdb_increment_seqnum(tdb);
	}

	SAFE_FREE(p); 
	tdb_unlock(tdb, BUCKET(hash), F_WRLCK);
	return ret;
}


/* Append to an entry. Create if not exist. */
int tdb_append(struct tdb_context *tdb, TDB_DATA key, TDB_DATA new_dbuf)
{
	uint32_t hash;
	TDB_DATA dbuf;
	int ret = -1;

	/* find which hash bucket it is in */
	hash = tdb->hash_fn(&key);
	if (tdb_lock(tdb, BUCKET(hash), F_WRLCK) == -1)
		return -1;

	dbuf = tdb_fetch(tdb, key);

	if (dbuf.dptr == NULL) {
		dbuf.dptr = (unsigned char *)malloc(new_dbuf.dsize);
	} else {
		unsigned int new_len = dbuf.dsize + new_dbuf.dsize;
		unsigned char *new_dptr;

		/* realloc '0' is special: don't do that. */
		if (new_len == 0)
			new_len = 1;
		new_dptr = (unsigned char *)realloc(dbuf.dptr, new_len);
		if (new_dptr == NULL) {
			free(dbuf.dptr);
		}
		dbuf.dptr = new_dptr;
	}

	if (dbuf.dptr == NULL) {
		tdb->ecode = TDB_ERR_OOM;
		goto failed;
	}

	memcpy(dbuf.dptr + dbuf.dsize, new_dbuf.dptr, new_dbuf.dsize);
	dbuf.dsize += new_dbuf.dsize;

	ret = tdb_store(tdb, key, dbuf, 0);
	
failed:
	tdb_unlock(tdb, BUCKET(hash), F_WRLCK);
	SAFE_FREE(dbuf.dptr);
	return ret;
}


/*
  return the name of the current tdb file
  useful for external logging functions
*/
const char *tdb_name(struct tdb_context *tdb)
{
	return tdb->name;
}

/*
  return the underlying file descriptor being used by tdb, or -1
  useful for external routines that want to check the device/inode
  of the fd
*/
int tdb_fd(struct tdb_context *tdb)
{
	return tdb->fd;
}

/*
  return the current logging function
  useful for external tdb routines that wish to log tdb errors
*/
tdb_log_func tdb_log_fn(struct tdb_context *tdb)
{
	return tdb->log.log_fn;
}


/*
  get the tdb sequence number. Only makes sense if the writers opened
  with TDB_SEQNUM set. Note that this sequence number will wrap quite
  quickly, so it should only be used for a 'has something changed'
  test, not for code that relies on the count of the number of changes
  made. If you want a counter then use a tdb record.

  The aim of this sequence number is to allow for a very lightweight
  test of a possible tdb change.
*/
int tdb_get_seqnum(struct tdb_context *tdb)
{
	tdb_off_t seqnum=0;

	tdb_ofs_read(tdb, TDB_SEQNUM_OFS, &seqnum);
	return seqnum;
}

int tdb_hash_size(struct tdb_context *tdb)
{
	return tdb->header.hash_size;
}

size_t tdb_map_size(struct tdb_context *tdb)
{
	return tdb->map_size;
}

int tdb_get_flags(struct tdb_context *tdb)
{
	return tdb->flags;
}

void tdb_add_flags(struct tdb_context *tdb, unsigned flags)
{
	tdb->flags |= flags;
}

void tdb_remove_flags(struct tdb_context *tdb, unsigned flags)
{
	tdb->flags &= ~flags;
}


/*
  enable sequence number handling on an open tdb
*/
void tdb_enable_seqnum(struct tdb_context *tdb)
{
	tdb->flags |= TDB_SEQNUM;
}


/*
  add a region of the file to the freelist. Length is the size of the region in bytes, 
  which includes the free list header that needs to be added
 */
static int tdb_free_region(struct tdb_context *tdb, tdb_off_t offset, ssize_t length)
{
	struct list_struct rec;
	if (length <= sizeof(rec)) {
		/* the region is not worth adding */
		return 0;
	}
	if (length + offset > tdb->map_size) {
		TDB_LOG((tdb, TDB_DEBUG_FATAL,"tdb_free_region: adding region beyond end of file\n"));
		return -1;		
	}
	memset(&rec,'\0',sizeof(rec));
	rec.rec_len = length - sizeof(rec);
	if (tdb_free(tdb, offset, &rec) == -1) {
		TDB_LOG((tdb, TDB_DEBUG_FATAL,"tdb_free_region: failed to add free record\n"));
		return -1;
	}
	return 0;
}

/*
  wipe the entire database, deleting all records. This can be done
  very fast by using a global lock. The entire data portion of the
  file becomes a single entry in the freelist.

  This code carefully steps around the recovery area, leaving it alone
 */
int tdb_wipe_all(struct tdb_context *tdb)
{
	int i;
	tdb_off_t offset = 0;
	ssize_t data_len;
	tdb_off_t recovery_head;
	tdb_len_t recovery_size = 0;

	if (tdb_lockall(tdb) != 0) {
		return -1;
	}

	/* see if the tdb has a recovery area, and remember its size
	   if so. We don't want to lose this as otherwise each
	   tdb_wipe_all() in a transaction will increase the size of
	   the tdb by the size of the recovery area */
	if (tdb_ofs_read(tdb, TDB_RECOVERY_HEAD, &recovery_head) == -1) {
		TDB_LOG((tdb, TDB_DEBUG_FATAL, "tdb_wipe_all: failed to read recovery head\n"));
		goto failed;
	}

	if (recovery_head != 0) {
		struct list_struct rec;
		if (tdb->methods->tdb_read(tdb, recovery_head, &rec, sizeof(rec), DOCONV()) == -1) {
			TDB_LOG((tdb, TDB_DEBUG_FATAL, "tdb_wipe_all: failed to read recovery record\n"));
			return -1;
		}	
		recovery_size = rec.rec_len + sizeof(rec);
	}

	/* wipe the hashes */
	for (i=0;i<tdb->header.hash_size;i++) {
		if (tdb_ofs_write(tdb, TDB_HASH_TOP(i), &offset) == -1) {
			TDB_LOG((tdb, TDB_DEBUG_FATAL,"tdb_wipe_all: failed to write hash %d\n", i));
			goto failed;
		}
	}

	/* wipe the freelist */
	if (tdb_ofs_write(tdb, FREELIST_TOP, &offset) == -1) {
		TDB_LOG((tdb, TDB_DEBUG_FATAL,"tdb_wipe_all: failed to write freelist\n"));
		goto failed;
	}

	/* add all the rest of the file to the freelist, possibly leaving a gap 
	   for the recovery area */
	if (recovery_size == 0) {
		/* the simple case - the whole file can be used as a freelist */
		data_len = (tdb->map_size - TDB_DATA_START(tdb->header.hash_size));
		if (tdb_free_region(tdb, TDB_DATA_START(tdb->header.hash_size), data_len) != 0) {
			goto failed;
		}
	} else {
		/* we need to add two freelist entries - one on either
		   side of the recovery area 

		   Note that we cannot shift the recovery area during
		   this operation. Only the transaction.c code may
		   move the recovery area or we risk subtle data
		   corruption
		*/
		data_len = (recovery_head - TDB_DATA_START(tdb->header.hash_size));
		if (tdb_free_region(tdb, TDB_DATA_START(tdb->header.hash_size), data_len) != 0) {
			goto failed;
		}
		/* and the 2nd free list entry after the recovery area - if any */
		data_len = tdb->map_size - (recovery_head+recovery_size);
		if (tdb_free_region(tdb, recovery_head+recovery_size, data_len) != 0) {
			goto failed;
		}
	}

	if (tdb_unlockall(tdb) != 0) {
		TDB_LOG((tdb, TDB_DEBUG_FATAL,"tdb_wipe_all: failed to unlock\n"));
		goto failed;
	}

	return 0;

failed:
	tdb_unlockall(tdb);
	return -1;
}

struct traverse_state {
	bool error;
	struct tdb_context *dest_db;
};

/*
  traverse function for repacking
 */
static int repack_traverse(struct tdb_context *tdb, TDB_DATA key, TDB_DATA data, void *private)
{
	struct traverse_state *state = (struct traverse_state *)private;
	if (tdb_store(state->dest_db, key, data, TDB_INSERT) != 0) {
		state->error = true;
		return -1;
	}
	return 0;
}

/*
  repack a tdb
 */
int tdb_repack(struct tdb_context *tdb)
{
	struct tdb_context *tmp_db;
	struct traverse_state state;

	if (tdb_transaction_start(tdb) != 0) {
		TDB_LOG((tdb, TDB_DEBUG_FATAL, __location__ " Failed to start transaction\n"));
		return -1;
	}

	tmp_db = tdb_open("tmpdb", tdb_hash_size(tdb), TDB_INTERNAL, O_RDWR|O_CREAT, 0);
	if (tmp_db == NULL) {
		TDB_LOG((tdb, TDB_DEBUG_FATAL, __location__ " Failed to create tmp_db\n"));
		tdb_transaction_cancel(tdb);
		return -1;
	}

	state.error = false;
	state.dest_db = tmp_db;

	if (tdb_traverse_read(tdb, repack_traverse, &state) == -1) {
		TDB_LOG((tdb, TDB_DEBUG_FATAL, __location__ " Failed to traverse copying out\n"));
		tdb_transaction_cancel(tdb);
		tdb_close(tmp_db);
		return -1;		
	}

	if (state.error) {
		TDB_LOG((tdb, TDB_DEBUG_FATAL, __location__ " Error during traversal\n"));
		tdb_transaction_cancel(tdb);
		tdb_close(tmp_db);
		return -1;
	}

	if (tdb_wipe_all(tdb) != 0) {
		TDB_LOG((tdb, TDB_DEBUG_FATAL, __location__ " Failed to wipe database\n"));
		tdb_transaction_cancel(tdb);
		tdb_close(tmp_db);
		return -1;
	}

	state.error = false;
	state.dest_db = tdb;

	if (tdb_traverse_read(tmp_db, repack_traverse, &state) == -1) {
		TDB_LOG((tdb, TDB_DEBUG_FATAL, __location__ " Failed to traverse copying back\n"));
		tdb_transaction_cancel(tdb);
		tdb_close(tmp_db);
		return -1;		
	}

	if (state.error) {
		TDB_LOG((tdb, TDB_DEBUG_FATAL, __location__ " Error during second traversal\n"));
		tdb_transaction_cancel(tdb);
		tdb_close(tmp_db);
		return -1;
	}

	tdb_close(tmp_db);

	if (tdb_transaction_commit(tdb) != 0) {
		TDB_LOG((tdb, TDB_DEBUG_FATAL, __location__ " Failed to commit\n"));
		return -1;
	}

	return 0;
}