summaryrefslogtreecommitdiff
path: root/lib/tdb2/io.c
blob: 3648505f12c37d688337e86ae70883558d7e01e5 (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
 /*
   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 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 "private.h"
#include <assert.h>
#include <ccan/likely/likely.h>

void tdb_munmap(struct tdb_file *file)
{
	if (file->fd == -1)
		return;

	if (file->map_ptr) {
		munmap(file->map_ptr, file->map_size);
		file->map_ptr = NULL;
	}
}

void tdb_mmap(struct tdb_context *tdb)
{
	int mmap_flags;

	if (tdb->flags & TDB_INTERNAL)
		return;

	if (tdb->flags & TDB_NOMMAP)
		return;

	if ((tdb->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)tdb->file->map_size == tdb->file->map_size) {
		tdb->file->map_ptr = mmap(NULL, tdb->file->map_size,
					  mmap_flags,
					  MAP_SHARED, tdb->file->fd, 0);
	} else
		tdb->file->map_ptr = MAP_FAILED;

	/*
	 * NB. When mmap fails it returns MAP_FAILED *NOT* NULL !!!!
	 */
	if (tdb->file->map_ptr == MAP_FAILED) {
		tdb->file->map_ptr = NULL;
		tdb_logerr(tdb, TDB_SUCCESS, TDB_LOG_WARNING,
			   "tdb_mmap failed for size %lld (%s)",
			   (long long)tdb->file->map_size, strerror(errno));
	}
}

/* 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 TDB_ERROR tdb_oob(struct tdb_context *tdb, tdb_off_t len,
			      bool probe)
{
	struct stat st;
	enum TDB_ERROR ecode;

	/* We can't hold pointers during this: we could unmap! */
	assert(!tdb->tdb2.direct_access
	       || (tdb->flags & TDB_NOLOCK)
	       || tdb_has_expansion_lock(tdb));

	if (len <= tdb->file->map_size)
		return TDB_SUCCESS;
	if (tdb->flags & TDB_INTERNAL) {
		if (probe)
			return TDB_SUCCESS;

		tdb_logerr(tdb, TDB_ERR_IO, TDB_LOG_ERROR,
			   "tdb_oob len %lld beyond internal"
			   " malloc size %lld",
			   (long long)len,
			   (long long)tdb->file->map_size);
		return TDB_ERR_IO;
	}

	ecode = tdb_lock_expand(tdb, F_RDLCK);
	if (ecode != TDB_SUCCESS) {
		return ecode;
	}

	if (fstat(tdb->file->fd, &st) != 0) {
		tdb_logerr(tdb, TDB_ERR_IO, TDB_LOG_ERROR,
			   "Failed to fstat file: %s", strerror(errno));
		tdb_unlock_expand(tdb, F_RDLCK);
		return TDB_ERR_IO;
	}

	tdb_unlock_expand(tdb, F_RDLCK);

	if (st.st_size < (size_t)len) {
		if (probe)
			return TDB_SUCCESS;

		tdb_logerr(tdb, TDB_ERR_IO, TDB_LOG_ERROR,
			   "tdb_oob len %zu beyond eof at %zu",
			   (size_t)len, st.st_size);
		return TDB_ERR_IO;
	}

	/* Unmap, update size, remap */
	tdb_munmap(tdb->file);

	tdb->file->map_size = st.st_size;
	tdb_mmap(tdb);
	return TDB_SUCCESS;
}

/* Endian conversion: we only ever deal with 8 byte quantities */
void *tdb_convert(const struct tdb_context *tdb, void *buf, tdb_len_t size)
{
	assert(size % 8 == 0);
	if (unlikely((tdb->flags & TDB_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 tdb_find_nonzero_off(struct tdb_context *tdb,
			      tdb_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 = tdb_access_read(tdb, base + start * sizeof(tdb_off_t),
			      (end - start) * sizeof(tdb_off_t), false);
	if (TDB_PTR_IS_ERR(val)) {
		return TDB_PTR_ERR(val);
	}

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

/* Return first zero offset in num offset array, or num, or -ve error. */
uint64_t tdb_find_zero_off(struct tdb_context *tdb, tdb_off_t off,
			   uint64_t num)
{
	uint64_t i;
	const uint64_t *val;

	/* Zero vs non-zero is the same unconverted: minor optimization. */
	val = tdb_access_read(tdb, off, num * sizeof(tdb_off_t), false);
	if (TDB_PTR_IS_ERR(val)) {
		return TDB_PTR_ERR(val);
	}

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

enum TDB_ERROR zero_out(struct tdb_context *tdb, tdb_off_t off, tdb_len_t len)
{
	char buf[8192] = { 0 };
	void *p = tdb->tdb2.io->direct(tdb, off, len, true);
	enum TDB_ERROR ecode = TDB_SUCCESS;

	assert(!(tdb->flags & TDB_RDONLY));
	if (TDB_PTR_IS_ERR(p)) {
		return TDB_PTR_ERR(p);
	}
	if (p) {
		memset(p, 0, len);
		return ecode;
	}
	while (len) {
		unsigned todo = len < sizeof(buf) ? len : sizeof(buf);
		ecode = tdb->tdb2.io->twrite(tdb, off, buf, todo);
		if (ecode != TDB_SUCCESS) {
			break;
		}
		len -= todo;
		off += todo;
	}
	return ecode;
}

tdb_off_t tdb_read_off(struct tdb_context *tdb, tdb_off_t off)
{
	tdb_off_t ret;
	enum TDB_ERROR ecode;

	if (likely(!(tdb->flags & TDB_CONVERT))) {
		tdb_off_t *p = tdb->tdb2.io->direct(tdb, off, sizeof(*p),
						    false);
		if (TDB_PTR_IS_ERR(p)) {
			return TDB_PTR_ERR(p);
		}
		if (p)
			return *p;
	}

	ecode = tdb_read_convert(tdb, off, &ret, sizeof(ret));
	if (ecode != TDB_SUCCESS) {
		return ecode;
	}
	return ret;
}

/* write a lump of data at a specified offset */
static enum TDB_ERROR tdb_write(struct tdb_context *tdb, tdb_off_t off,
				const void *buf, tdb_len_t len)
{
	enum TDB_ERROR ecode;

	if (tdb->flags & TDB_RDONLY) {
		return tdb_logerr(tdb, TDB_ERR_RDONLY, TDB_LOG_USE_ERROR,
				  "Write to read-only database");
	}

	ecode = tdb->tdb2.io->oob(tdb, off + len, false);
	if (ecode != TDB_SUCCESS) {
		return ecode;
	}

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

			return tdb_logerr(tdb, TDB_ERR_IO, TDB_LOG_ERROR,
					  "tdb_write: %zi at %zu len=%zu (%s)",
					  ret, (size_t)off, (size_t)len,
					  strerror(errno));
		}
	}
	return TDB_SUCCESS;
}

/* read a lump of data at a specified offset */
static enum TDB_ERROR tdb_read(struct tdb_context *tdb, tdb_off_t off,
			       void *buf, tdb_len_t len)
{
	enum TDB_ERROR ecode;

	ecode = tdb->tdb2.io->oob(tdb, off + len, false);
	if (ecode != TDB_SUCCESS) {
		return ecode;
	}

	if (tdb->file->map_ptr) {
		memcpy(buf, off + (char *)tdb->file->map_ptr, len);
	} else {
		ssize_t r = pread(tdb->file->fd, buf, len, off);
		if (r != len) {
			return tdb_logerr(tdb, TDB_ERR_IO, TDB_LOG_ERROR,
					  "tdb_read failed with %zi at %zu "
					  "len=%zu (%s) map_size=%zu",
					  r, (size_t)off, (size_t)len,
					  strerror(errno),
					  (size_t)tdb->file->map_size);
		}
	}
	return TDB_SUCCESS;
}

enum TDB_ERROR tdb_write_convert(struct tdb_context *tdb, tdb_off_t off,
				 const void *rec, size_t len)
{
	enum TDB_ERROR ecode;

	if (unlikely((tdb->flags & TDB_CONVERT))) {
		void *conv = malloc(len);
		if (!conv) {
			return tdb_logerr(tdb, TDB_ERR_OOM, TDB_LOG_ERROR,
					  "tdb_write: no memory converting"
					  " %zu bytes", len);
		}
		memcpy(conv, rec, len);
		ecode = tdb->tdb2.io->twrite(tdb, off,
					   tdb_convert(tdb, conv, len), len);
		free(conv);
	} else {
		ecode = tdb->tdb2.io->twrite(tdb, off, rec, len);
	}
	return ecode;
}

enum TDB_ERROR tdb_read_convert(struct tdb_context *tdb, tdb_off_t off,
				void *rec, size_t len)
{
	enum TDB_ERROR ecode = tdb->tdb2.io->tread(tdb, off, rec, len);
	tdb_convert(tdb, rec, len);
	return ecode;
}

enum TDB_ERROR tdb_write_off(struct tdb_context *tdb,
			     tdb_off_t off, tdb_off_t val)
{
	if (tdb->flags & TDB_RDONLY) {
		return tdb_logerr(tdb, TDB_ERR_RDONLY, TDB_LOG_USE_ERROR,
				  "Write to read-only database");
	}

	if (likely(!(tdb->flags & TDB_CONVERT))) {
		tdb_off_t *p = tdb->tdb2.io->direct(tdb, off, sizeof(*p),
						    true);
		if (TDB_PTR_IS_ERR(p)) {
			return TDB_PTR_ERR(p);
		}
		if (p) {
			*p = val;
			return TDB_SUCCESS;
		}
	}
	return tdb_write_convert(tdb, off, &val, sizeof(val));
}

static void *_tdb_alloc_read(struct tdb_context *tdb, tdb_off_t offset,
			     tdb_len_t len, unsigned int prefix)
{
	unsigned char *buf;
	enum TDB_ERROR ecode;

	/* some systems don't like zero length malloc */
	buf = malloc(prefix + len ? prefix + len : 1);
	if (!buf) {
		tdb_logerr(tdb, TDB_ERR_OOM, TDB_LOG_USE_ERROR,
			   "tdb_alloc_read malloc failed len=%zu",
			   (size_t)(prefix + len));
		return TDB_ERR_PTR(TDB_ERR_OOM);
	} else {
		ecode = tdb->tdb2.io->tread(tdb, offset, buf+prefix, len);
		if (unlikely(ecode != TDB_SUCCESS)) {
			free(buf);
			return TDB_ERR_PTR(ecode);
		}
	}
	return buf;
}

/* read a lump of data, allocating the space for it */
void *tdb_alloc_read(struct tdb_context *tdb, tdb_off_t offset, tdb_len_t len)
{
	return _tdb_alloc_read(tdb, offset, len, 0);
}

static enum TDB_ERROR fill(struct tdb_context *tdb,
			   const void *buf, size_t size,
			   tdb_off_t off, tdb_len_t len)
{
	while (len) {
		size_t n = len > size ? size : len;
		ssize_t ret = pwrite(tdb->file->fd, buf, n, off);
		if (ret != n) {
			if (ret >= 0)
				errno = ENOSPC;

			return tdb_logerr(tdb, TDB_ERR_IO, TDB_LOG_ERROR,
					  "fill failed:"
					  " %zi at %zu len=%zu (%s)",
					  ret, (size_t)off, (size_t)len,
					  strerror(errno));
		}
		len -= n;
		off += n;
	}
	return TDB_SUCCESS;
}

/* expand a file.  we prefer to use ftruncate, as that is what posix
  says to use for mmap expansion */
static enum TDB_ERROR tdb_expand_file(struct tdb_context *tdb,
				      tdb_len_t addition)
{
	char buf[8192];
	enum TDB_ERROR ecode;

	if (tdb->flags & TDB_RDONLY) {
		return tdb_logerr(tdb, TDB_ERR_RDONLY, TDB_LOG_USE_ERROR,
				  "Expand on read-only database");
	}

	if (tdb->flags & TDB_INTERNAL) {
		char *new = realloc(tdb->file->map_ptr,
				    tdb->file->map_size + addition);
		if (!new) {
			return tdb_logerr(tdb, TDB_ERR_OOM, TDB_LOG_ERROR,
					  "No memory to expand database");
		}
		tdb->file->map_ptr = new;
		tdb->file->map_size += addition;
	} else {
		/* Unmap before trying to write; old TDB claimed OpenBSD had
		 * problem with this otherwise. */
		tdb_munmap(tdb->file);

		/* If this fails, we try to fill anyway. */
		if (ftruncate(tdb->file->fd, tdb->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(tdb, buf, sizeof(buf), tdb->file->map_size,
			     addition);
		if (ecode != TDB_SUCCESS)
			return ecode;
		tdb->file->map_size += addition;
		tdb_mmap(tdb);
	}
	return TDB_SUCCESS;
}

const void *tdb_access_read(struct tdb_context *tdb,
			    tdb_off_t off, tdb_len_t len, bool convert)
{
	void *ret = NULL;

	if (likely(!(tdb->flags & TDB_CONVERT))) {
		ret = tdb->tdb2.io->direct(tdb, off, len, false);

		if (TDB_PTR_IS_ERR(ret)) {
			return ret;
		}
	}
	if (!ret) {
		struct tdb_access_hdr *hdr;
		hdr = _tdb_alloc_read(tdb, off, len, sizeof(*hdr));
		if (TDB_PTR_IS_ERR(hdr)) {
			return hdr;
		}
		hdr->next = tdb->tdb2.access;
		tdb->tdb2.access = hdr;
		ret = hdr + 1;
		if (convert) {
			tdb_convert(tdb, (void *)ret, len);
		}
	} else
		tdb->tdb2.direct_access++;

	return ret;
}

void *tdb_access_write(struct tdb_context *tdb,
		       tdb_off_t off, tdb_len_t len, bool convert)
{
	void *ret = NULL;

	if (tdb->flags & TDB_RDONLY) {
		tdb_logerr(tdb, TDB_ERR_RDONLY, TDB_LOG_USE_ERROR,
			   "Write to read-only database");
		return TDB_ERR_PTR(TDB_ERR_RDONLY);
	}

	if (likely(!(tdb->flags & TDB_CONVERT))) {
		ret = tdb->tdb2.io->direct(tdb, off, len, true);

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

	if (!ret) {
		struct tdb_access_hdr *hdr;
		hdr = _tdb_alloc_read(tdb, off, len, sizeof(*hdr));
		if (TDB_PTR_IS_ERR(hdr)) {
			return hdr;
		}
		hdr->next = tdb->tdb2.access;
		tdb->tdb2.access = hdr;
		hdr->off = off;
		hdr->len = len;
		hdr->convert = convert;
		ret = hdr + 1;
		if (convert)
			tdb_convert(tdb, (void *)ret, len);
	} else
		tdb->tdb2.direct_access++;

	return ret;
}

static struct tdb_access_hdr **find_hdr(struct tdb_context *tdb, const void *p)
{
	struct tdb_access_hdr **hp;

	for (hp = &tdb->tdb2.access; *hp; hp = &(*hp)->next) {
		if (*hp + 1 == p)
			return hp;
	}
	return NULL;
}

void tdb_access_release(struct tdb_context *tdb, const void *p)
{
	struct tdb_access_hdr *hdr, **hp = find_hdr(tdb, p);

	if (hp) {
		hdr = *hp;
		*hp = hdr->next;
		free(hdr);
	} else
		tdb->tdb2.direct_access--;
}

enum TDB_ERROR tdb_access_commit(struct tdb_context *tdb, void *p)
{
	struct tdb_access_hdr *hdr, **hp = find_hdr(tdb, p);
	enum TDB_ERROR ecode;

	if (hp) {
		hdr = *hp;
		if (hdr->convert)
			ecode = tdb_write_convert(tdb, hdr->off, p, hdr->len);
		else
			ecode = tdb_write(tdb, hdr->off, p, hdr->len);
		*hp = hdr->next;
		free(hdr);
	} else {
		tdb->tdb2.direct_access--;
		ecode = TDB_SUCCESS;
	}

	return ecode;
}

static void *tdb_direct(struct tdb_context *tdb, tdb_off_t off, size_t len,
			bool write_mode)
{
	enum TDB_ERROR ecode;

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

	ecode = tdb_oob(tdb, off + len, false);
	if (unlikely(ecode != TDB_SUCCESS))
		return TDB_ERR_PTR(ecode);
	return (char *)tdb->file->map_ptr + off;
}

void tdb_inc_seqnum(struct tdb_context *tdb)
{
	tdb_off_t seq;

	if (tdb->flags & TDB_VERSION1) {
		tdb1_increment_seqnum_nonblock(tdb);
		return;
	}

	if (likely(!(tdb->flags & TDB_CONVERT))) {
		int64_t *direct;

		direct = tdb->tdb2.io->direct(tdb,
					      offsetof(struct tdb_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 = tdb_read_off(tdb, offsetof(struct tdb_header, seqnum));
	if (!TDB_OFF_IS_ERR(seq)) {
		seq++;
		if (unlikely((int64_t)seq < 0))
			seq = 0;
		tdb_write_off(tdb, offsetof(struct tdb_header, seqnum), seq);
	}
}

static const struct tdb_methods io_methods = {
	tdb_read,
	tdb_write,
	tdb_oob,
	tdb_expand_file,
	tdb_direct,
};

/*
  initialise the default methods table
*/
void tdb_io_init(struct tdb_context *tdb)
{
	tdb->tdb2.io = &io_methods;
}