summaryrefslogtreecommitdiff
path: root/lib/ntdb/io.c
blob: 7645cddc373f2f8228e734f6fa82769a76e58ad2 (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
 /*
   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
   Copyright (C) Rusty Russell			   2010

     ** NOTE! The following LGPL license applies to the ntdb
     ** 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 "private.h"
#include <ccan/likely/likely.h>

static void free_old_mmaps(struct ntdb_context *ntdb)
{
	struct ntdb_old_mmap *i;

	assert(ntdb->file->direct_count == 0);

	while ((i = ntdb->file->old_mmaps) != NULL) {
		ntdb->file->old_mmaps = i->next;
		if (ntdb->flags & NTDB_INTERNAL) {
			ntdb->free_fn(i->map_ptr, ntdb->alloc_data);
		} else {
			munmap(i->map_ptr, i->map_size);
		}
		ntdb->free_fn(i, ntdb->alloc_data);
	}
}

static enum NTDB_ERROR save_old_map(struct ntdb_context *ntdb)
{
	struct ntdb_old_mmap *old;

	assert(ntdb->file->direct_count);

	old = ntdb->alloc_fn(ntdb->file, sizeof(*old), ntdb->alloc_data);
	if (!old) {
		return ntdb_logerr(ntdb, NTDB_ERR_OOM, NTDB_LOG_ERROR,
				   "save_old_map alloc failed");
	}
	old->next = ntdb->file->old_mmaps;
	old->map_ptr = ntdb->file->map_ptr;
	old->map_size = ntdb->file->map_size;
	ntdb->file->old_mmaps = old;

	return NTDB_SUCCESS;
}

enum NTDB_ERROR ntdb_munmap(struct ntdb_context *ntdb)
{
	if (ntdb->file->fd == -1) {
		return NTDB_SUCCESS;
	}

	if (!ntdb->file->map_ptr) {
		return NTDB_SUCCESS;
	}

	/* We can't unmap now if there are accessors. */
	if (ntdb->file->direct_count) {
		return save_old_map(ntdb);
	} else {
		munmap(ntdb->file->map_ptr, ntdb->file->map_size);
		ntdb->file->map_ptr = NULL;
	}
	return NTDB_SUCCESS;
}

enum NTDB_ERROR ntdb_mmap(struct ntdb_context *ntdb)
{
	int mmap_flags;

	if (ntdb->flags & NTDB_INTERNAL)
		return NTDB_SUCCESS;

#ifndef HAVE_INCOHERENT_MMAP
	if (ntdb->flags & NTDB_NOMMAP)
		return NTDB_SUCCESS;
#endif

	if ((ntdb->open_flags & O_ACCMODE) == O_RDONLY)
		mmap_flags = PROT_READ;
	else
		mmap_flags = PROT_READ | PROT_WRITE;

	/* size_t can be smaller than off_t. */
	if ((size_t)ntdb->file->map_size == ntdb->file->map_size) {
		ntdb->file->map_ptr = mmap(NULL, ntdb->file->map_size,
					  mmap_flags,
					  MAP_SHARED, ntdb->file->fd, 0);
	} else
		ntdb->file->map_ptr = MAP_FAILED;

	/*
	 * NB. When mmap fails it returns MAP_FAILED *NOT* NULL !!!!
	 */
	if (ntdb->file->map_ptr == MAP_FAILED) {
		ntdb->file->map_ptr = NULL;
#ifdef HAVE_INCOHERENT_MMAP
		/* Incoherent mmap means everyone must mmap! */
		return ntdb_logerr(ntdb, NTDB_ERR_IO, NTDB_LOG_ERROR,
				  "ntdb_mmap failed for size %lld (%s)",
				  (long long)ntdb->file->map_size,
				  strerror(errno));
#else
		ntdb_logerr(ntdb, NTDB_SUCCESS, NTDB_LOG_WARNING,
			   "ntdb_mmap failed for size %lld (%s)",
			   (long long)ntdb->file->map_size, strerror(errno));
#endif
	}
	return NTDB_SUCCESS;
}

/* check for an out of bounds access - if it is out of bounds then
   see if the database has been expanded by someone else and expand
   if necessary
   note that "len" is the minimum length needed for the db.

   If probe is true, len being too large isn't a failure.
*/
static enum NTDB_ERROR ntdb_normal_oob(struct ntdb_context *ntdb,
				       ntdb_off_t off, ntdb_len_t len,
				       bool probe)
{
	struct stat st;
	enum NTDB_ERROR ecode;

	if (len + off < len) {
		if (probe)
			return NTDB_SUCCESS;

		return ntdb_logerr(ntdb, NTDB_ERR_IO, NTDB_LOG_ERROR,
				  "ntdb_oob off %llu len %llu wrap\n",
				  (long long)off, (long long)len);
	}

	if (ntdb->flags & NTDB_INTERNAL) {
		if (probe)
			return NTDB_SUCCESS;

		ntdb_logerr(ntdb, NTDB_ERR_IO, NTDB_LOG_ERROR,
			   "ntdb_oob len %lld beyond internal"
			   " alloc size %lld",
			   (long long)(off + len),
			   (long long)ntdb->file->map_size);
		return NTDB_ERR_IO;
	}

	ecode = ntdb_lock_expand(ntdb, F_RDLCK);
	if (ecode != NTDB_SUCCESS) {
		return ecode;
	}

	if (fstat(ntdb->file->fd, &st) != 0) {
		ntdb_logerr(ntdb, NTDB_ERR_IO, NTDB_LOG_ERROR,
			   "Failed to fstat file: %s", strerror(errno));
		ntdb_unlock_expand(ntdb, F_RDLCK);
		return NTDB_ERR_IO;
	}

	ntdb_unlock_expand(ntdb, F_RDLCK);

	if (st.st_size < off + len) {
		if (probe)
			return NTDB_SUCCESS;

		ntdb_logerr(ntdb, NTDB_ERR_IO, NTDB_LOG_ERROR,
			   "ntdb_oob len %llu beyond eof at %llu",
			   (long long)(off + len), (long long)st.st_size);
		return NTDB_ERR_IO;
	}

	/* Unmap, update size, remap */
	ecode = ntdb_munmap(ntdb);
	if (ecode) {
		return ecode;
	}

	ntdb->file->map_size = st.st_size;
	return ntdb_mmap(ntdb);
}

/* Endian conversion: we only ever deal with 8 byte quantities */
void *ntdb_convert(const struct ntdb_context *ntdb, void *buf, ntdb_len_t size)
{
	assert(size % 8 == 0);
	if (unlikely((ntdb->flags & NTDB_CONVERT)) && buf) {
		uint64_t i, *p = (uint64_t *)buf;
		for (i = 0; i < size / 8; i++)
			p[i] = bswap_64(p[i]);
	}
	return buf;
}

/* Return first non-zero offset in offset array, or end, or -ve error. */
/* FIXME: Return the off? */
uint64_t ntdb_find_nonzero_off(struct ntdb_context *ntdb,
			      ntdb_off_t base, uint64_t start, uint64_t end)
{
	uint64_t i;
	const uint64_t *val;

	/* Zero vs non-zero is the same unconverted: minor optimization. */
	val = ntdb_access_read(ntdb, base + start * sizeof(ntdb_off_t),
			      (end - start) * sizeof(ntdb_off_t), false);
	if (NTDB_PTR_IS_ERR(val)) {
		return NTDB_ERR_TO_OFF(NTDB_PTR_ERR(val));
	}

	for (i = 0; i < (end - start); i++) {
		if (val[i])
			break;
	}
	ntdb_access_release(ntdb, val);
	return start + i;
}

/* Return first zero offset in num offset array, or num, or -ve error. */
uint64_t ntdb_find_zero_off(struct ntdb_context *ntdb, ntdb_off_t off,
			   uint64_t num)
{
	uint64_t i;
	const uint64_t *val;

	/* Zero vs non-zero is the same unconverted: minor optimization. */
	val = ntdb_access_read(ntdb, off, num * sizeof(ntdb_off_t), false);
	if (NTDB_PTR_IS_ERR(val)) {
		return NTDB_ERR_TO_OFF(NTDB_PTR_ERR(val));
	}

	for (i = 0; i < num; i++) {
		if (!val[i])
			break;
	}
	ntdb_access_release(ntdb, val);
	return i;
}

enum NTDB_ERROR zero_out(struct ntdb_context *ntdb, ntdb_off_t off, ntdb_len_t len)
{
	char buf[8192] = { 0 };
	void *p = ntdb->io->direct(ntdb, off, len, true);
	enum NTDB_ERROR ecode = NTDB_SUCCESS;

	assert(!(ntdb->flags & NTDB_RDONLY));
	if (NTDB_PTR_IS_ERR(p)) {
		return NTDB_PTR_ERR(p);
	}
	if (p) {
		memset(p, 0, len);
		return ecode;
	}
	while (len) {
		unsigned todo = len < sizeof(buf) ? len : sizeof(buf);
		ecode = ntdb->io->twrite(ntdb, off, buf, todo);
		if (ecode != NTDB_SUCCESS) {
			break;
		}
		len -= todo;
		off += todo;
	}
	return ecode;
}

/* write a lump of data at a specified offset */
static enum NTDB_ERROR ntdb_write(struct ntdb_context *ntdb, ntdb_off_t off,
				const void *buf, ntdb_len_t len)
{
	enum NTDB_ERROR ecode;

	if (ntdb->flags & NTDB_RDONLY) {
		return ntdb_logerr(ntdb, NTDB_ERR_RDONLY, NTDB_LOG_USE_ERROR,
				  "Write to read-only database");
	}

	ecode = ntdb_oob(ntdb, off, len, false);
	if (ecode != NTDB_SUCCESS) {
		return ecode;
	}

	if (ntdb->file->map_ptr) {
		memcpy(off + (char *)ntdb->file->map_ptr, buf, len);
	} else {
#ifdef HAVE_INCOHERENT_MMAP
		return NTDB_ERR_IO;
#else
		ssize_t ret;
		ret = pwrite(ntdb->file->fd, buf, len, off);
		if (ret != len) {
			/* This shouldn't happen: we avoid sparse files. */
			if (ret >= 0)
				errno = ENOSPC;

			return ntdb_logerr(ntdb, NTDB_ERR_IO, NTDB_LOG_ERROR,
					  "ntdb_write: %zi at %zu len=%zu (%s)",
					  ret, (size_t)off, (size_t)len,
					  strerror(errno));
		}
#endif
	}
	return NTDB_SUCCESS;
}

/* read a lump of data at a specified offset */
static enum NTDB_ERROR ntdb_read(struct ntdb_context *ntdb, ntdb_off_t off,
			       void *buf, ntdb_len_t len)
{
	enum NTDB_ERROR ecode;

	ecode = ntdb_oob(ntdb, off, len, false);
	if (ecode != NTDB_SUCCESS) {
		return ecode;
	}

	if (ntdb->file->map_ptr) {
		memcpy(buf, off + (char *)ntdb->file->map_ptr, len);
	} else {
#ifdef HAVE_INCOHERENT_MMAP
		return NTDB_ERR_IO;
#else
		ssize_t r = pread(ntdb->file->fd, buf, len, off);
		if (r != len) {
			return ntdb_logerr(ntdb, NTDB_ERR_IO, NTDB_LOG_ERROR,
					  "ntdb_read failed with %zi at %zu "
					  "len=%zu (%s) map_size=%zu",
					  r, (size_t)off, (size_t)len,
					  strerror(errno),
					  (size_t)ntdb->file->map_size);
		}
#endif
	}
	return NTDB_SUCCESS;
}

enum NTDB_ERROR ntdb_write_convert(struct ntdb_context *ntdb, ntdb_off_t off,
				 const void *rec, size_t len)
{
	enum NTDB_ERROR ecode;

	if (unlikely((ntdb->flags & NTDB_CONVERT))) {
		void *conv = ntdb->alloc_fn(ntdb, len, ntdb->alloc_data);
		if (!conv) {
			return ntdb_logerr(ntdb, NTDB_ERR_OOM, NTDB_LOG_ERROR,
					  "ntdb_write: no memory converting"
					  " %zu bytes", len);
		}
		memcpy(conv, rec, len);
		ecode = ntdb->io->twrite(ntdb, off,
					 ntdb_convert(ntdb, conv, len), len);
		ntdb->free_fn(conv, ntdb->alloc_data);
	} else {
		ecode = ntdb->io->twrite(ntdb, off, rec, len);
	}
	return ecode;
}

enum NTDB_ERROR ntdb_read_convert(struct ntdb_context *ntdb, ntdb_off_t off,
				void *rec, size_t len)
{
	enum NTDB_ERROR ecode = ntdb->io->tread(ntdb, off, rec, len);
	ntdb_convert(ntdb, rec, len);
	return ecode;
}

static void *_ntdb_alloc_read(struct ntdb_context *ntdb, ntdb_off_t offset,
			     ntdb_len_t len, unsigned int prefix)
{
	unsigned char *buf;
	enum NTDB_ERROR ecode;

	/* some systems don't like zero length malloc */
	buf = ntdb->alloc_fn(ntdb, prefix + len ? prefix + len : 1,
			  ntdb->alloc_data);
	if (!buf) {
		ntdb_logerr(ntdb, NTDB_ERR_OOM, NTDB_LOG_ERROR,
			   "ntdb_alloc_read alloc failed len=%zu",
			   (size_t)(prefix + len));
		return NTDB_ERR_PTR(NTDB_ERR_OOM);
	} else {
		ecode = ntdb->io->tread(ntdb, offset, buf+prefix, len);
		if (unlikely(ecode != NTDB_SUCCESS)) {
			ntdb->free_fn(buf, ntdb->alloc_data);
			return NTDB_ERR_PTR(ecode);
		}
	}
	return buf;
}

/* read a lump of data, allocating the space for it */
void *ntdb_alloc_read(struct ntdb_context *ntdb, ntdb_off_t offset, ntdb_len_t len)
{
	return _ntdb_alloc_read(ntdb, offset, len, 0);
}

static enum NTDB_ERROR fill(struct ntdb_context *ntdb,
			   const void *buf, size_t size,
			   ntdb_off_t off, ntdb_len_t len)
{
	while (len) {
		size_t n = len > size ? size : len;
		ssize_t ret = pwrite(ntdb->file->fd, buf, n, off);
		if (ret != n) {
			if (ret >= 0)
				errno = ENOSPC;

			return ntdb_logerr(ntdb, NTDB_ERR_IO, NTDB_LOG_ERROR,
					  "fill failed:"
					  " %zi at %zu len=%zu (%s)",
					  ret, (size_t)off, (size_t)len,
					  strerror(errno));
		}
		len -= n;
		off += n;
	}
	return NTDB_SUCCESS;
}

/* expand a file.  we prefer to use ftruncate, as that is what posix
  says to use for mmap expansion */
static enum NTDB_ERROR ntdb_expand_file(struct ntdb_context *ntdb,
				      ntdb_len_t addition)
{
	char buf[8192];
	enum NTDB_ERROR ecode;

	assert((ntdb->file->map_size + addition) % NTDB_PGSIZE == 0);
	if (ntdb->flags & NTDB_RDONLY) {
		return ntdb_logerr(ntdb, NTDB_ERR_RDONLY, NTDB_LOG_USE_ERROR,
				  "Expand on read-only database");
	}

	if (ntdb->flags & NTDB_INTERNAL) {
		char *new;

		/* Can't free it if we have direct accesses. */
		if (ntdb->file->direct_count) {
			ecode = save_old_map(ntdb);
			if (ecode) {
				return ecode;
			}
			new = ntdb->alloc_fn(ntdb->file,
					     ntdb->file->map_size + addition,
					     ntdb->alloc_data);
			if (new) {
				memcpy(new, ntdb->file->map_ptr,
				       ntdb->file->map_size);
			}
		} else {
			new = ntdb->expand_fn(ntdb->file->map_ptr,
					      ntdb->file->map_size + addition,
					      ntdb->alloc_data);
		}
		if (!new) {
			return ntdb_logerr(ntdb, NTDB_ERR_OOM, NTDB_LOG_ERROR,
					  "No memory to expand database");
		}
		ntdb->file->map_ptr = new;
		ntdb->file->map_size += addition;
		return NTDB_SUCCESS;
	} else {
		/* Unmap before trying to write; old NTDB claimed OpenBSD had
		 * problem with this otherwise. */
		ecode = ntdb_munmap(ntdb);
		if (ecode) {
			return ecode;
		}

		/* If this fails, we try to fill anyway. */
		if (ftruncate(ntdb->file->fd, ntdb->file->map_size + addition))
			;

		/* now fill the file with something. This ensures that the
		   file isn't sparse, which would be very bad if we ran out of
		   disk. This must be done with write, not via mmap */
		memset(buf, 0x43, sizeof(buf));
		ecode = fill(ntdb, buf, sizeof(buf), ntdb->file->map_size,
			     addition);
		if (ecode != NTDB_SUCCESS)
			return ecode;
		ntdb->file->map_size += addition;
		return ntdb_mmap(ntdb);
	}
}

const void *ntdb_access_read(struct ntdb_context *ntdb,
			    ntdb_off_t off, ntdb_len_t len, bool convert)
{
	void *ret = NULL;

	if (likely(!(ntdb->flags & NTDB_CONVERT))) {
		ret = ntdb->io->direct(ntdb, off, len, false);

		if (NTDB_PTR_IS_ERR(ret)) {
			return ret;
		}
	}
	if (!ret) {
		struct ntdb_access_hdr *hdr;
		hdr = _ntdb_alloc_read(ntdb, off, len, sizeof(*hdr));
		if (NTDB_PTR_IS_ERR(hdr)) {
			return hdr;
		}
		hdr->next = ntdb->access;
		ntdb->access = hdr;
		ret = hdr + 1;
		if (convert) {
			ntdb_convert(ntdb, (void *)ret, len);
		}
	} else {
		ntdb->file->direct_count++;
	}

	return ret;
}

void *ntdb_access_write(struct ntdb_context *ntdb,
		       ntdb_off_t off, ntdb_len_t len, bool convert)
{
	void *ret = NULL;

	if (ntdb->flags & NTDB_RDONLY) {
		ntdb_logerr(ntdb, NTDB_ERR_RDONLY, NTDB_LOG_USE_ERROR,
			   "Write to read-only database");
		return NTDB_ERR_PTR(NTDB_ERR_RDONLY);
	}

	if (likely(!(ntdb->flags & NTDB_CONVERT))) {
		ret = ntdb->io->direct(ntdb, off, len, true);

		if (NTDB_PTR_IS_ERR(ret)) {
			return ret;
		}
	}

	if (!ret) {
		struct ntdb_access_hdr *hdr;
		hdr = _ntdb_alloc_read(ntdb, off, len, sizeof(*hdr));
		if (NTDB_PTR_IS_ERR(hdr)) {
			return hdr;
		}
		hdr->next = ntdb->access;
		ntdb->access = hdr;
		hdr->off = off;
		hdr->len = len;
		hdr->convert = convert;
		ret = hdr + 1;
		if (convert)
			ntdb_convert(ntdb, (void *)ret, len);
	} else {
		ntdb->file->direct_count++;
	}
	return ret;
}

static struct ntdb_access_hdr **find_hdr(struct ntdb_context *ntdb, const void *p)
{
	struct ntdb_access_hdr **hp;

	for (hp = &ntdb->access; *hp; hp = &(*hp)->next) {
		if (*hp + 1 == p)
			return hp;
	}
	return NULL;
}

void ntdb_access_release(struct ntdb_context *ntdb, const void *p)
{
	struct ntdb_access_hdr *hdr, **hp = find_hdr(ntdb, p);

	if (hp) {
		hdr = *hp;
		*hp = hdr->next;
		ntdb->free_fn(hdr, ntdb->alloc_data);
	} else {
		if (--ntdb->file->direct_count == 0) {
			free_old_mmaps(ntdb);
		}
	}
}

enum NTDB_ERROR ntdb_access_commit(struct ntdb_context *ntdb, void *p)
{
	struct ntdb_access_hdr *hdr, **hp = find_hdr(ntdb, p);
	enum NTDB_ERROR ecode;

	if (hp) {
		hdr = *hp;
		if (hdr->convert)
			ecode = ntdb_write_convert(ntdb, hdr->off, p, hdr->len);
		else
			ecode = ntdb_write(ntdb, hdr->off, p, hdr->len);
		*hp = hdr->next;
		ntdb->free_fn(hdr, ntdb->alloc_data);
	} else {
		if (--ntdb->file->direct_count == 0) {
			free_old_mmaps(ntdb);
		}
		ecode = NTDB_SUCCESS;
	}

	return ecode;
}

static void *ntdb_direct(struct ntdb_context *ntdb, ntdb_off_t off, size_t len,
			bool write_mode)
{
	enum NTDB_ERROR ecode;

	if (unlikely(!ntdb->file->map_ptr))
		return NULL;

	ecode = ntdb_oob(ntdb, off, len, false);
	if (unlikely(ecode != NTDB_SUCCESS))
		return NTDB_ERR_PTR(ecode);
	return (char *)ntdb->file->map_ptr + off;
}

static ntdb_off_t ntdb_read_normal_off(struct ntdb_context *ntdb,
				       ntdb_off_t off)
{
	ntdb_off_t ret;
	enum NTDB_ERROR ecode;
	ntdb_off_t *p;

	p = ntdb_direct(ntdb, off, sizeof(*p), false);
	if (NTDB_PTR_IS_ERR(p)) {
		return NTDB_ERR_TO_OFF(NTDB_PTR_ERR(p));
	}
	if (likely(p)) {
		return *p;
	}

	ecode = ntdb_read(ntdb, off, &ret, sizeof(ret));
	if (ecode != NTDB_SUCCESS) {
		return NTDB_ERR_TO_OFF(ecode);
	}
	return ret;
}

static ntdb_off_t ntdb_read_convert_off(struct ntdb_context *ntdb,
					ntdb_off_t off)
{
	ntdb_off_t ret;
	enum NTDB_ERROR ecode;

	ecode = ntdb_read_convert(ntdb, off, &ret, sizeof(ret));
	if (ecode != NTDB_SUCCESS) {
		return NTDB_ERR_TO_OFF(ecode);
	}
	return ret;
}

static enum NTDB_ERROR ntdb_write_normal_off(struct ntdb_context *ntdb,
					     ntdb_off_t off, ntdb_off_t val)
{
	ntdb_off_t *p;

	p = ntdb_direct(ntdb, off, sizeof(*p), true);
	if (NTDB_PTR_IS_ERR(p)) {
		return NTDB_PTR_ERR(p);
	}
	if (likely(p)) {
		*p = val;
		return NTDB_SUCCESS;
	}
	return ntdb_write(ntdb, off, &val, sizeof(val));
}

static enum NTDB_ERROR ntdb_write_convert_off(struct ntdb_context *ntdb,
					      ntdb_off_t off, ntdb_off_t val)
{
	return ntdb_write_convert(ntdb, off, &val, sizeof(val));
}

void ntdb_inc_seqnum(struct ntdb_context *ntdb)
{
	ntdb_off_t seq;

	if (likely(!(ntdb->flags & NTDB_CONVERT))) {
		int64_t *direct;

		direct = ntdb->io->direct(ntdb,
					 offsetof(struct ntdb_header, seqnum),
					 sizeof(*direct), true);
		if (likely(direct)) {
			/* Don't let it go negative, even briefly */
			if (unlikely((*direct) + 1) < 0)
				*direct = 0;
			(*direct)++;
			return;
		}
	}

	seq = ntdb_read_off(ntdb, offsetof(struct ntdb_header, seqnum));
	if (!NTDB_OFF_IS_ERR(seq)) {
		seq++;
		if (unlikely((int64_t)seq < 0))
			seq = 0;
		ntdb_write_off(ntdb, offsetof(struct ntdb_header, seqnum), seq);
	}
}

static const struct ntdb_methods io_methods = {
	ntdb_read,
	ntdb_write,
	ntdb_normal_oob,
	ntdb_expand_file,
	ntdb_direct,
	ntdb_read_normal_off,
	ntdb_write_normal_off,
};

static const struct ntdb_methods io_convert_methods = {
	ntdb_read,
	ntdb_write,
	ntdb_normal_oob,
	ntdb_expand_file,
	ntdb_direct,
	ntdb_read_convert_off,
	ntdb_write_convert_off,
};

/*
  initialise the default methods table
*/
void ntdb_io_init(struct ntdb_context *ntdb)
{
	if (ntdb->flags & NTDB_CONVERT)
		ntdb->io = &io_convert_methods;
	else
		ntdb->io = &io_methods;
}