summaryrefslogtreecommitdiffstats
path: root/source/locking/brlock.c
blob: 990a6a25d2fba201bc09e296662bb9c6cfbcbe4e (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
/* 
   Unix SMB/CIFS implementation.
   byte range locking code
   Updated to handle range splits/merges.

   Copyright (C) Andrew Tridgell 1992-2000
   Copyright (C) Jeremy Allison 1992-2000
   
   This program is free software; you can redistribute it and/or modify
   it under the terms of the GNU General Public License as published by
   the Free Software Foundation; either version 2 of the License, or
   (at your option) any later version.
   
   This program 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 General Public License for more details.
   
   You should have received a copy of the GNU General Public License
   along with this program; if not, write to the Free Software
   Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
*/

/* This module implements a tdb based byte range locking service,
   replacing the fcntl() based byte range locking previously
   used. This allows us to provide the same semantics as NT */

#include "includes.h"

#define ZERO_ZERO 0

/* This contains elements that differentiate locks. The smbpid is a
   client supplied pid, and is essentially the locking context for
   this client */

struct lock_context {
	uint16 smbpid;
	uint16 tid;
	pid_t pid;
};

/* The data in brlock records is an unsorted linear array of these
   records.  It is unnecessary to store the count as tdb provides the
   size of the record */

struct lock_struct {
	struct lock_context context;
	br_off start;
	br_off size;
	int fnum;
	enum brl_type lock_type;
};

/* The key used in the brlock database. */

struct lock_key {
	SMB_DEV_T device;
	SMB_INO_T inode;
};

/* The open brlock.tdb database. */

static TDB_CONTEXT *tdb;

/****************************************************************************
 Create a locking key - ensuring zero filled for pad purposes.
****************************************************************************/

static TDB_DATA locking_key(SMB_DEV_T dev, SMB_INO_T inode)
{
        static struct lock_key key;
        TDB_DATA kbuf;

        memset(&key, '\0', sizeof(key));
        key.device = dev;
        key.inode = inode;
        kbuf.dptr = (char *)&key;
        kbuf.dsize = sizeof(key);
        return kbuf;
}

/****************************************************************************
 See if two locking contexts are equal.
****************************************************************************/

static BOOL brl_same_context(struct lock_context *ctx1, 
			     struct lock_context *ctx2)
{
	return (ctx1->pid == ctx2->pid) &&
		(ctx1->smbpid == ctx2->smbpid) &&
		(ctx1->tid == ctx2->tid);
}

/****************************************************************************
 See if lock2 can be added when lock1 is in place.
****************************************************************************/

static BOOL brl_conflict(struct lock_struct *lck1, 
			 struct lock_struct *lck2)
{
	if (lck1->lock_type == PENDING_LOCK || lck2->lock_type == PENDING_LOCK )
		return False;

	if (lck1->lock_type == READ_LOCK && lck2->lock_type == READ_LOCK) {
		return False;
	}

	if (brl_same_context(&lck1->context, &lck2->context) &&
	    lck2->lock_type == READ_LOCK && lck1->fnum == lck2->fnum) {
		return False;
	}

	if (lck1->start >= (lck2->start + lck2->size) ||
	    lck2->start >= (lck1->start + lck1->size)) {
		return False;
	}
	    
	return True;
} 

#if ZERO_ZERO
static BOOL brl_conflict1(struct lock_struct *lck1, 
			 struct lock_struct *lck2)
{
	if (lck1->lock_type == PENDING_LOCK || lck2->lock_type == PENDING_LOCK )
		return False;

	if (lck1->lock_type == READ_LOCK && lck2->lock_type == READ_LOCK) {
		return False;
	}

	if (brl_same_context(&lck1->context, &lck2->context) &&
	    lck2->lock_type == READ_LOCK && lck1->fnum == lck2->fnum) {
		return False;
	}

	if (lck2->start == 0 && lck2->size == 0 && lck1->size != 0) {
		return True;
	}

	if (lck1->start >= (lck2->start + lck2->size) ||
	    lck2->start >= (lck1->start + lck1->size)) {
		return False;
	}
	    
	return True;
} 
#endif

/****************************************************************************
 Check to see if this lock conflicts, but ignore our own locks on the
 same fnum only.
****************************************************************************/

static BOOL brl_conflict_other(struct lock_struct *lck1, struct lock_struct *lck2)
{
	if (lck1->lock_type == PENDING_LOCK || lck2->lock_type == PENDING_LOCK )
		return False;

	if (lck1->lock_type == READ_LOCK && lck2->lock_type == READ_LOCK) 
		return False;

	/*
	 * Incoming WRITE locks conflict with existing READ locks even
	 * if the context is the same. JRA. See LOCKTEST7 in smbtorture.
	 */

	if (!(lck2->lock_type == WRITE_LOCK && lck1->lock_type == READ_LOCK)) {
		if (brl_same_context(&lck1->context, &lck2->context) &&
					lck1->fnum == lck2->fnum)
			return False;
	}

	if (lck1->start >= (lck2->start + lck2->size) ||
	    lck2->start >= (lck1->start + lck1->size))
		return False;
	    
	return True;
} 


#if DONT_DO_THIS
	/* doing this traversal could kill solaris machines under high load (tridge) */
	/* delete any dead locks */

/****************************************************************************
 Delete a record if it is for a dead process, if check_self is true, then
 delete any records belonging to this pid also (there shouldn't be any).
****************************************************************************/

static int delete_fn(TDB_CONTEXT *ttdb, TDB_DATA kbuf, TDB_DATA dbuf, void *state)
{
	struct lock_struct *locks;
	int count, i;
	BOOL check_self = *(BOOL *)state;
	pid_t mypid = sys_getpid();

	tdb_chainlock(tdb, kbuf);

	locks = (struct lock_struct *)dbuf.dptr;

	count = dbuf.dsize / sizeof(*locks);
	for (i=0; i<count; i++) {
		struct lock_struct *lock = &locks[i];

		/* If check_self is true we want to remove our own records. */
		if (check_self && (mypid == lock->context.pid)) {

			DEBUG(0,("brlock : delete_fn. LOGIC ERROR ! Shutting down and a record for my pid (%u) exists !\n",
					(unsigned int)lock->context.pid ));

		} else if (process_exists(lock->context.pid)) {

			DEBUG(10,("brlock : delete_fn. pid %u exists.\n", (unsigned int)lock->context.pid ));
			continue;
		}

		DEBUG(10,("brlock : delete_fn. Deleting record for process %u\n",
				(unsigned int)lock->context.pid ));

		if (count > 1 && i < count-1) {
			memmove(&locks[i], &locks[i+1], 
				sizeof(*locks)*((count-1) - i));
		}
		count--;
		i--;
	}

	if (count == 0) {
		tdb_delete(tdb, kbuf);
	} else if (count < (dbuf.dsize / sizeof(*locks))) {
		dbuf.dsize = count * sizeof(*locks);
		tdb_store(tdb, kbuf, dbuf, TDB_REPLACE);
	}

	tdb_chainunlock(tdb, kbuf);
	return 0;
}
#endif

/****************************************************************************
 Open up the brlock.tdb database.
****************************************************************************/

void brl_init(int read_only)
{
	if (tdb)
		return;
	tdb = tdb_open_log(lock_path("brlock.tdb"), 0,  TDB_DEFAULT|(read_only?0x0:TDB_CLEAR_IF_FIRST),
		       read_only?O_RDONLY:(O_RDWR|O_CREAT), 0644 );
	if (!tdb) {
		DEBUG(0,("Failed to open byte range locking database\n"));
		return;
	}

#if DONT_DO_THIS
	/* doing this traversal could kill solaris machines under high load (tridge) */
	/* delete any dead locks */
	if (!read_only) {
		BOOL check_self = False;
		tdb_traverse(tdb, delete_fn, &check_self);
	}
#endif
}

/****************************************************************************
 Close down the brlock.tdb database.
****************************************************************************/

void brl_shutdown(int read_only)
{
	if (!tdb)
		return;

#if DONT_DO_THIS
	/* doing this traversal could kill solaris machines under high load (tridge) */
	/* delete any dead locks */
	if (!read_only) {
		BOOL check_self = True;
		tdb_traverse(tdb, delete_fn, &check_self);
	}
#endif

	tdb_close(tdb);
}

#if ZERO_ZERO
/****************************************************************************
compare two locks for sorting
****************************************************************************/
static int lock_compare(struct lock_struct *lck1, 
			 struct lock_struct *lck2)
{
	if (lck1->start != lck2->start) return (lck1->start - lck2->start);
	if (lck2->size != lck1->size) {
		return ((int)lck1->size - (int)lck2->size);
	}
	return 0;
}
#endif

/****************************************************************************
 Lock a range of bytes.
****************************************************************************/

NTSTATUS brl_lock(SMB_DEV_T dev, SMB_INO_T ino, int fnum,
		  uint16 smbpid, pid_t pid, uint16 tid,
		  br_off start, br_off size, 
		  enum brl_type lock_type, BOOL *my_lock_ctx)
{
	TDB_DATA kbuf, dbuf;
	int count, i;
	struct lock_struct lock, *locks;
	char *tp;
	NTSTATUS status = NT_STATUS_OK;
	static int last_failed = -1;
	static br_off last_failed_start;

	*my_lock_ctx = False;
	kbuf = locking_key(dev,ino);

	dbuf.dptr = NULL;

#if !ZERO_ZERO
	if (start == 0 && size == 0) {
		DEBUG(0,("client sent 0/0 lock - please report this\n"));
	}
#endif

	tdb_chainlock(tdb, kbuf);
	dbuf = tdb_fetch(tdb, kbuf);

	lock.context.smbpid = smbpid;
	lock.context.pid = pid;
	lock.context.tid = tid;
	lock.start = start;
	lock.size = size;
	lock.fnum = fnum;
	lock.lock_type = lock_type;

	if (dbuf.dptr) {
		/* there are existing locks - make sure they don't conflict */
		locks = (struct lock_struct *)dbuf.dptr;
		count = dbuf.dsize / sizeof(*locks);
		for (i=0; i<count; i++) {
			if (brl_conflict(&locks[i], &lock)) {
				status = NT_STATUS_LOCK_NOT_GRANTED;
				/* Did we block ourselves ? */
				if (brl_same_context(&locks[i].context, &lock.context))
					*my_lock_ctx = True;
				goto fail;
			}
#if ZERO_ZERO
			if (lock.start == 0 && lock.size == 0 && 
			    locks[i].size == 0) {
				break;
			}
#endif
		}
	}

	/* no conflicts - add it to the list of locks */
	tp = Realloc(dbuf.dptr, dbuf.dsize + sizeof(*locks));
	if (!tp) {
		status = NT_STATUS_NO_MEMORY;
		goto fail;
	} else {
		dbuf.dptr = tp;
	}
	memcpy(dbuf.dptr + dbuf.dsize, &lock, sizeof(lock));
	dbuf.dsize += sizeof(lock);

#if ZERO_ZERO
	/* sort the lock list */
	qsort(dbuf.dptr, dbuf.dsize/sizeof(lock), sizeof(lock), lock_compare);
#endif

	tdb_store(tdb, kbuf, dbuf, TDB_REPLACE);

	SAFE_FREE(dbuf.dptr);
	tdb_chainunlock(tdb, kbuf);
	return NT_STATUS_OK;

 fail:
	/* this is a nasty hack to try to simulate the lock result cache code in w2k.
	   It isn't completely accurate as I haven't yet worked out the correct
	   semantics (tridge)
	*/
	if (last_failed == fnum &&
	    last_failed_start == start &&
	    NT_STATUS_EQUAL(status, NT_STATUS_LOCK_NOT_GRANTED)) {
		status = NT_STATUS_FILE_LOCK_CONFLICT;
	}
	last_failed = fnum;
	last_failed_start = start;

	SAFE_FREE(dbuf.dptr);
	tdb_chainunlock(tdb, kbuf);
	return status;
}

/****************************************************************************
 Check if an unlock overlaps a pending lock.
****************************************************************************/

static BOOL brl_pending_overlap(struct lock_struct *lock, struct lock_struct *pend_lock)
{
	if ((lock->start <= pend_lock->start) && (lock->start + lock->size > pend_lock->start))
		return True;
	if ((lock->start >= pend_lock->start) && (lock->start <= pend_lock->start + pend_lock->size))
		return True;
	return False;
}

/****************************************************************************
 Unlock a range of bytes.
****************************************************************************/

BOOL brl_unlock(SMB_DEV_T dev, SMB_INO_T ino, int fnum,
		uint16 smbpid, pid_t pid, uint16 tid,
		br_off start, br_off size,
		BOOL remove_pending_locks_only,
		void (*pre_unlock_fn)(void *),
		void *pre_unlock_data)
{
	TDB_DATA kbuf, dbuf;
	int count, i, j;
	struct lock_struct *locks;
	struct lock_context context;

	kbuf = locking_key(dev,ino);

	dbuf.dptr = NULL;

	tdb_chainlock(tdb, kbuf);
	dbuf = tdb_fetch(tdb, kbuf);

	if (!dbuf.dptr) {
		DEBUG(10,("brl_unlock: tdb_fetch failed !\n"));
		goto fail;
	}

	context.smbpid = smbpid;
	context.pid = pid;
	context.tid = tid;

	/* there are existing locks - find a match */
	locks = (struct lock_struct *)dbuf.dptr;
	count = dbuf.dsize / sizeof(*locks);

#if ZERO_ZERO
	for (i=0; i<count; i++) {
		struct lock_struct *lock = &locks[i];

		if (lock->lock_type == WRITE_LOCK &&
		    brl_same_context(&lock->context, &context) &&
		    lock->fnum == fnum &&
		    lock->start == start &&
		    lock->size == size) {

			if (pre_unlock_fn)
				(*pre_unlock_fn)(pre_unlock_data);

			/* found it - delete it */
			if (count == 1) {
				tdb_delete(tdb, kbuf);
			} else {
				if (i < count-1) {
					memmove(&locks[i], &locks[i+1], 
						sizeof(*locks)*((count-1) - i));
				}
				dbuf.dsize -= sizeof(*locks);
				tdb_store(tdb, kbuf, dbuf, TDB_REPLACE);
			}

			SAFE_FREE(dbuf.dptr);
			tdb_chainunlock(tdb, kbuf);
			return True;
		}
	}
#endif

	locks = (struct lock_struct *)dbuf.dptr;
	count = dbuf.dsize / sizeof(*locks);
	for (i=0; i<count; i++) {
		struct lock_struct *lock = &locks[i];

		if (brl_same_context(&lock->context, &context) &&
				lock->fnum == fnum &&
				lock->start == start &&
				lock->size == size) {

			if (remove_pending_locks_only && lock->lock_type != PENDING_LOCK)
				continue;

			if (lock->lock_type != PENDING_LOCK) {

				/* Do any POSIX unlocks needed. */
				if (pre_unlock_fn)
					(*pre_unlock_fn)(pre_unlock_data);

				/* Send unlock messages to any pending waiters that overlap. */
				for (j=0; j<count; j++) {
					struct lock_struct *pend_lock = &locks[j];

					/* Ignore non-pending locks. */
					if (pend_lock->lock_type != PENDING_LOCK)
						continue;

					/* We could send specific lock info here... */
					if (brl_pending_overlap(lock, pend_lock)) {
						DEBUG(10,("brl_unlock: sending unlock message to pid %u\n",
									(unsigned int)pend_lock->context.pid ));

						message_send_pid(pend_lock->context.pid,
								MSG_SMB_UNLOCK,
								NULL, 0, True);
					}
				}
			}

			/* found it - delete it */
			if (count == 1) {
				tdb_delete(tdb, kbuf);
			} else {
				if (i < count-1) {
					memmove(&locks[i], &locks[i+1], 
						sizeof(*locks)*((count-1) - i));
				}
				dbuf.dsize -= sizeof(*locks);
				tdb_store(tdb, kbuf, dbuf, TDB_REPLACE);
			}

			SAFE_FREE(dbuf.dptr);
			tdb_chainunlock(tdb, kbuf);
			return True;
		}
	}

	/* we didn't find it */

 fail:
	SAFE_FREE(dbuf.dptr);
	tdb_chainunlock(tdb, kbuf);
	return False;
}


/****************************************************************************
 Test if we could add a lock if we wanted to.
****************************************************************************/

BOOL brl_locktest(SMB_DEV_T dev, SMB_INO_T ino, int fnum,
		  uint16 smbpid, pid_t pid, uint16 tid,
		  br_off start, br_off size, 
		  enum brl_type lock_type, int check_self)
{
	TDB_DATA kbuf, dbuf;
	int count, i;
	struct lock_struct lock, *locks;

	kbuf = locking_key(dev,ino);

	dbuf.dptr = NULL;

	tdb_chainlock(tdb, kbuf);
	dbuf = tdb_fetch(tdb, kbuf);

	lock.context.smbpid = smbpid;
	lock.context.pid = pid;
	lock.context.tid = tid;
	lock.start = start;
	lock.size = size;
	lock.fnum = fnum;
	lock.lock_type = lock_type;

	if (dbuf.dptr) {
		/* there are existing locks - make sure they don't conflict */
		locks = (struct lock_struct *)dbuf.dptr;
		count = dbuf.dsize / sizeof(*locks);
		for (i=0; i<count; i++) {
			if (check_self) {
				if (brl_conflict(&locks[i], &lock))
					goto fail;
			} else {
				/*
				 * Our own locks don't conflict.
				 */
				if (brl_conflict_other(&locks[i], &lock))
					goto fail;
			}
		}
	}

	/* no conflicts - we could have added it */
	SAFE_FREE(dbuf.dptr);
	tdb_chainunlock(tdb, kbuf);
	return True;

 fail:
	SAFE_FREE(dbuf.dptr);
	tdb_chainunlock(tdb, kbuf);
	return False;
}

/****************************************************************************
 Remove any locks associated with a open file.
****************************************************************************/

void brl_close(SMB_DEV_T dev, SMB_INO_T ino, pid_t pid, int tid, int fnum)
{
	TDB_DATA kbuf, dbuf;
	int count, i, j, dcount=0;
	struct lock_struct *locks;

	kbuf = locking_key(dev,ino);

	dbuf.dptr = NULL;

	tdb_chainlock(tdb, kbuf);
	dbuf = tdb_fetch(tdb, kbuf);

	if (!dbuf.dptr) goto fail;

	/* there are existing locks - remove any for this fnum */
	locks = (struct lock_struct *)dbuf.dptr;
	count = dbuf.dsize / sizeof(*locks);

	for (i=0; i<count; i++) {
		struct lock_struct *lock = &locks[i];

		if (lock->context.tid == tid &&
		    lock->context.pid == pid &&
		    lock->fnum == fnum) {

			/* Send unlock messages to any pending waiters that overlap. */
			for (j=0; j<count; j++) {
				struct lock_struct *pend_lock = &locks[j];

				/* Ignore our own or non-pending locks. */
				if (pend_lock->lock_type != PENDING_LOCK)
					continue;

				if (pend_lock->context.tid == tid &&
				    pend_lock->context.pid == pid &&
				    pend_lock->fnum == fnum)
					continue;

				/* We could send specific lock info here... */
				if (brl_pending_overlap(lock, pend_lock))
					message_send_pid(pend_lock->context.pid,
							MSG_SMB_UNLOCK,
							NULL, 0, True);
			}

			/* found it - delete it */
			if (count > 1 && i < count-1) {
				memmove(&locks[i], &locks[i+1], 
					sizeof(*locks)*((count-1) - i));
			}
			count--;
			i--;
			dcount++;
		}
	}

	if (count == 0) {
		tdb_delete(tdb, kbuf);
	} else if (count < (dbuf.dsize / sizeof(*locks))) {
		dbuf.dsize -= dcount * sizeof(*locks);
		tdb_store(tdb, kbuf, dbuf, TDB_REPLACE);
	}

	/* we didn't find it */
 fail:
	SAFE_FREE(dbuf.dptr);
	tdb_chainunlock(tdb, kbuf);
}

/****************************************************************************
 Traverse the whole database with this function, calling traverse_callback
 on each lock.
****************************************************************************/

static int traverse_fn(TDB_CONTEXT *ttdb, TDB_DATA kbuf, TDB_DATA dbuf, void *state)
{
	struct lock_struct *locks;
	struct lock_key *key;
	int i;

	BRLOCK_FN(traverse_callback) = (BRLOCK_FN_CAST())state;

	locks = (struct lock_struct *)dbuf.dptr;
	key = (struct lock_key *)kbuf.dptr;

	for (i=0;i<dbuf.dsize/sizeof(*locks);i++) {
		traverse_callback(key->device, key->inode,
				  locks[i].context.pid,
				  locks[i].lock_type,
				  locks[i].start,
				  locks[i].size);
	}
	return 0;
}

/*******************************************************************
 Call the specified function on each lock in the database.
********************************************************************/

int brl_forall(BRLOCK_FN(fn))
{
	if (!tdb) return 0;
	return tdb_traverse(tdb, traverse_fn, (void *)fn);
}