summaryrefslogtreecommitdiff
path: root/source3/locking/posix.c
blob: 463aa21fd82ecb310b5b109c81026cc894c359f3 (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
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
/* 
   Unix SMB/CIFS implementation.
   Locking functions
   Copyright (C) Jeremy Allison 1992-2006
   
   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 3 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, see <http://www.gnu.org/licenses/>.

   Revision History:

   POSIX locking support. Jeremy Allison (jeremy@valinux.com), Apr. 2000.
*/

#include "includes.h"
#include "system/filesys.h"
#include "locking/proto.h"
#include "dbwrap/dbwrap.h"
#include "dbwrap/dbwrap_rbt.h"
#include "util_tdb.h"

#undef DBGC_CLASS
#define DBGC_CLASS DBGC_LOCKING

/*
 * The pending close database handle.
 */

static struct db_context *posix_pending_close_db;

/****************************************************************************
 First - the functions that deal with the underlying system locks - these
 functions are used no matter if we're mapping CIFS Windows locks or CIFS
 POSIX locks onto POSIX.
****************************************************************************/

/****************************************************************************
 Utility function to map a lock type correctly depending on the open
 mode of a file.
****************************************************************************/

static int map_posix_lock_type( files_struct *fsp, enum brl_type lock_type)
{
	if((lock_type == WRITE_LOCK) && !fsp->can_write) {
		/*
		 * Many UNIX's cannot get a write lock on a file opened read-only.
		 * Win32 locking semantics allow this.
		 * Do the best we can and attempt a read-only lock.
		 */
		DEBUG(10,("map_posix_lock_type: Downgrading write lock to read due to read-only file.\n"));
		return F_RDLCK;
	}

	/*
	 * This return should be the most normal, as we attempt
	 * to always open files read/write.
	 */

	return (lock_type == READ_LOCK) ? F_RDLCK : F_WRLCK;
}

/****************************************************************************
 Debugging aid :-).
****************************************************************************/

static const char *posix_lock_type_name(int lock_type)
{
	return (lock_type == F_RDLCK) ? "READ" : "WRITE";
}

/****************************************************************************
 Check to see if the given unsigned lock range is within the possible POSIX
 range. Modifies the given args to be in range if possible, just returns
 False if not.
****************************************************************************/

static bool posix_lock_in_range(off_t *offset_out, off_t *count_out,
				uint64_t u_offset, uint64_t u_count)
{
	off_t offset = (off_t)u_offset;
	off_t count = (off_t)u_count;

	/*
	 * For the type of system we are, attempt to
	 * find the maximum positive lock offset as an off_t.
	 */

#if defined(MAX_POSITIVE_LOCK_OFFSET) /* Some systems have arbitrary limits. */

	off_t max_positive_lock_offset = (MAX_POSITIVE_LOCK_OFFSET);
#else
	/*
	 * In this case off_t is 64 bits,
	 * and the underlying system can handle 64 bit signed locks.
	 */

	off_t mask2 = ((off_t)0x4) << (SMB_OFF_T_BITS-4);
	off_t mask = (mask2<<1);
	off_t max_positive_lock_offset = ~mask;

#endif
	/*
	 * POSIX locks of length zero mean lock to end-of-file.
	 * Win32 locks of length zero are point probes. Ignore
	 * any Win32 locks of length zero. JRA.
	 */

	if (count == (off_t)0) {
		DEBUG(10,("posix_lock_in_range: count = 0, ignoring.\n"));
		return False;
	}

	/*
	 * If the given offset was > max_positive_lock_offset then we cannot map this at all
	 * ignore this lock.
	 */

	if (u_offset & ~((uint64_t)max_positive_lock_offset)) {
		DEBUG(10,("posix_lock_in_range: (offset = %.0f) offset > %.0f and we cannot handle this. Ignoring lock.\n",
				(double)u_offset, (double)((uint64_t)max_positive_lock_offset) ));
		return False;
	}

	/*
	 * We must truncate the count to less than max_positive_lock_offset.
	 */

	if (u_count & ~((uint64_t)max_positive_lock_offset)) {
		count = max_positive_lock_offset;
	}

	/*
	 * Truncate count to end at max lock offset.
	 */

	if (offset + count < 0 || offset + count > max_positive_lock_offset) {
		count = max_positive_lock_offset - offset;
	}

	/*
	 * If we ate all the count, ignore this lock.
	 */

	if (count == 0) {
		DEBUG(10,("posix_lock_in_range: Count = 0. Ignoring lock u_offset = %.0f, u_count = %.0f\n",
				(double)u_offset, (double)u_count ));
		return False;
	}

	/*
	 * The mapping was successful.
	 */

	DEBUG(10,("posix_lock_in_range: offset_out = %.0f, count_out = %.0f\n",
			(double)offset, (double)count ));

	*offset_out = offset;
	*count_out = count;
	
	return True;
}

bool smb_vfs_call_lock(struct vfs_handle_struct *handle,
		       struct files_struct *fsp, int op, off_t offset,
		       off_t count, int type)
{
	VFS_FIND(lock);
	return handle->fns->lock_fn(handle, fsp, op, offset, count, type);
}

/****************************************************************************
 Actual function that does POSIX locks. Copes with 64 -> 32 bit cruft and
 broken NFS implementations.
****************************************************************************/

static bool posix_fcntl_lock(files_struct *fsp, int op, off_t offset, off_t count, int type)
{
	bool ret;

	DEBUG(8,("posix_fcntl_lock %d %d %.0f %.0f %d\n",fsp->fh->fd,op,(double)offset,(double)count,type));

	ret = SMB_VFS_LOCK(fsp, op, offset, count, type);

	if (!ret && ((errno == EFBIG) || (errno == ENOLCK) || (errno ==  EINVAL))) {

		DEBUG(0,("posix_fcntl_lock: WARNING: lock request at offset %.0f, length %.0f returned\n",
					(double)offset,(double)count));
		DEBUGADD(0,("an %s error. This can happen when using 64 bit lock offsets\n", strerror(errno)));
		DEBUGADD(0,("on 32 bit NFS mounted file systems.\n"));

		/*
		 * If the offset is > 0x7FFFFFFF then this will cause problems on
		 * 32 bit NFS mounted filesystems. Just ignore it.
		 */

		if (offset & ~((off_t)0x7fffffff)) {
			DEBUG(0,("Offset greater than 31 bits. Returning success.\n"));
			return True;
		}

		if (count & ~((off_t)0x7fffffff)) {
			/* 32 bit NFS file system, retry with smaller offset */
			DEBUG(0,("Count greater than 31 bits - retrying with 31 bit truncated length.\n"));
			errno = 0;
			count &= 0x7fffffff;
			ret = SMB_VFS_LOCK(fsp, op, offset, count, type);
		}
	}

	DEBUG(8,("posix_fcntl_lock: Lock call %s\n", ret ? "successful" : "failed"));
	return ret;
}

bool smb_vfs_call_getlock(struct vfs_handle_struct *handle,
			  struct files_struct *fsp, off_t *poffset,
			  off_t *pcount, int *ptype, pid_t *ppid)
{
	VFS_FIND(getlock);
	return handle->fns->getlock_fn(handle, fsp, poffset, pcount, ptype, 
				       ppid);
}

/****************************************************************************
 Actual function that gets POSIX locks. Copes with 64 -> 32 bit cruft and
 broken NFS implementations.
****************************************************************************/

static bool posix_fcntl_getlock(files_struct *fsp, off_t *poffset, off_t *pcount, int *ptype)
{
	pid_t pid;
	bool ret;

	DEBUG(8,("posix_fcntl_getlock %d %.0f %.0f %d\n",
		fsp->fh->fd,(double)*poffset,(double)*pcount,*ptype));

	ret = SMB_VFS_GETLOCK(fsp, poffset, pcount, ptype, &pid);

	if (!ret && ((errno == EFBIG) || (errno == ENOLCK) || (errno ==  EINVAL))) {

		DEBUG(0,("posix_fcntl_getlock: WARNING: lock request at offset %.0f, length %.0f returned\n",
					(double)*poffset,(double)*pcount));
		DEBUGADD(0,("an %s error. This can happen when using 64 bit lock offsets\n", strerror(errno)));
		DEBUGADD(0,("on 32 bit NFS mounted file systems.\n"));

		/*
		 * If the offset is > 0x7FFFFFFF then this will cause problems on
		 * 32 bit NFS mounted filesystems. Just ignore it.
		 */

		if (*poffset & ~((off_t)0x7fffffff)) {
			DEBUG(0,("Offset greater than 31 bits. Returning success.\n"));
			return True;
		}

		if (*pcount & ~((off_t)0x7fffffff)) {
			/* 32 bit NFS file system, retry with smaller offset */
			DEBUG(0,("Count greater than 31 bits - retrying with 31 bit truncated length.\n"));
			errno = 0;
			*pcount &= 0x7fffffff;
			ret = SMB_VFS_GETLOCK(fsp,poffset,pcount,ptype,&pid);
		}
	}

	DEBUG(8,("posix_fcntl_getlock: Lock query call %s\n", ret ? "successful" : "failed"));
	return ret;
}

/****************************************************************************
 POSIX function to see if a file region is locked. Returns True if the
 region is locked, False otherwise.
****************************************************************************/

bool is_posix_locked(files_struct *fsp,
			uint64_t *pu_offset,
			uint64_t *pu_count,
			enum brl_type *plock_type,
			enum brl_flavour lock_flav)
{
	off_t offset;
	off_t count;
	int posix_lock_type = map_posix_lock_type(fsp,*plock_type);

	DEBUG(10,("is_posix_locked: File %s, offset = %.0f, count = %.0f, "
		  "type = %s\n", fsp_str_dbg(fsp), (double)*pu_offset,
		  (double)*pu_count,  posix_lock_type_name(*plock_type)));

	/*
	 * If the requested lock won't fit in the POSIX range, we will
	 * never set it, so presume it is not locked.
	 */

	if(!posix_lock_in_range(&offset, &count, *pu_offset, *pu_count)) {
		return False;
	}

	if (!posix_fcntl_getlock(fsp,&offset,&count,&posix_lock_type)) {
		return False;
	}

	if (posix_lock_type == F_UNLCK) {
		return False;
	}

	if (lock_flav == POSIX_LOCK) {
		/* Only POSIX lock queries need to know the details. */
		*pu_offset = (uint64_t)offset;
		*pu_count = (uint64_t)count;
		*plock_type = (posix_lock_type == F_RDLCK) ? READ_LOCK : WRITE_LOCK;
	}
	return True;
}

/****************************************************************************
 Next - the functions that deal with in memory database storing representations
 of either Windows CIFS locks or POSIX CIFS locks.
****************************************************************************/

/* The key used in the in-memory POSIX databases. */

struct lock_ref_count_key {
	struct file_id id;
	char r;
}; 

/*******************************************************************
 Form a static locking key for a dev/inode pair for the lock ref count
******************************************************************/

static TDB_DATA locking_ref_count_key_fsp(files_struct *fsp,
					  struct lock_ref_count_key *tmp)
{
	ZERO_STRUCTP(tmp);
	tmp->id = fsp->file_id;
	tmp->r = 'r';
	return make_tdb_data((uint8_t *)tmp, sizeof(*tmp));
}

/*******************************************************************
 Convenience function to get an fd_array key from an fsp.
******************************************************************/

static TDB_DATA fd_array_key_fsp(files_struct *fsp)
{
	return make_tdb_data((uint8 *)&fsp->file_id, sizeof(fsp->file_id));
}

/*******************************************************************
 Create the in-memory POSIX lock databases.
********************************************************************/

bool posix_locking_init(bool read_only)
{
	if (posix_pending_close_db != NULL) {
		return true;
	}

	posix_pending_close_db = db_open_rbt(NULL);

	if (posix_pending_close_db == NULL) {
		DEBUG(0,("Failed to open POSIX pending close database.\n"));
		return false;
	}

	return true;
}

/*******************************************************************
 Delete the in-memory POSIX lock databases.
********************************************************************/

bool posix_locking_end(void)
{
	/*
	 * Shouldn't we close all fd's here?
	 */
	TALLOC_FREE(posix_pending_close_db);
	return true;
}

/****************************************************************************
 Next - the functions that deal with storing fd's that have outstanding
 POSIX locks when closed.
****************************************************************************/

/****************************************************************************
 The records in posix_pending_close_tdb are composed of an array of ints
 keyed by dev/ino pair.
 The first int is a reference count of the number of outstanding locks on
 all open fd's on this dev/ino pair. Any subsequent ints are the fd's that
 were open on this dev/ino pair that should have been closed, but can't as
 the lock ref count is non zero.
****************************************************************************/

/****************************************************************************
 Keep a reference count of the number of Windows locks open on this dev/ino
 pair. Creates entry if it doesn't exist.
****************************************************************************/

static void increment_windows_lock_ref_count(files_struct *fsp)
{
	struct lock_ref_count_key tmp;
	struct db_record *rec;
	int lock_ref_count = 0;
	NTSTATUS status;
	TDB_DATA value;

	rec = dbwrap_fetch_locked(
		posix_pending_close_db, talloc_tos(),
		locking_ref_count_key_fsp(fsp, &tmp));

	SMB_ASSERT(rec != NULL);

	value = dbwrap_record_get_value(rec);

	if (value.dptr != NULL) {
		SMB_ASSERT(value.dsize == sizeof(lock_ref_count));
		memcpy(&lock_ref_count, value.dptr,
		       sizeof(lock_ref_count));
	}

	lock_ref_count++;

	status = dbwrap_record_store(rec,
				     make_tdb_data((uint8 *)&lock_ref_count,
				     sizeof(lock_ref_count)), 0);

	SMB_ASSERT(NT_STATUS_IS_OK(status));

	TALLOC_FREE(rec);

	DEBUG(10,("increment_windows_lock_ref_count for file now %s = %d\n",
		  fsp_str_dbg(fsp), lock_ref_count));
}

/****************************************************************************
 Bulk delete - subtract as many locks as we've just deleted.
****************************************************************************/

void reduce_windows_lock_ref_count(files_struct *fsp, unsigned int dcount)
{
	struct lock_ref_count_key tmp;
	struct db_record *rec;
	int lock_ref_count = 0;
	NTSTATUS status;
	TDB_DATA value;

	rec = dbwrap_fetch_locked(
		posix_pending_close_db, talloc_tos(),
		locking_ref_count_key_fsp(fsp, &tmp));

	value = dbwrap_record_get_value(rec);

	SMB_ASSERT((rec != NULL)
		   && (value.dptr != NULL)
		   && (value.dsize == sizeof(lock_ref_count)));

	memcpy(&lock_ref_count, value.dptr, sizeof(lock_ref_count));

	SMB_ASSERT(lock_ref_count > 0);

	lock_ref_count -= dcount;

	status = dbwrap_record_store(rec,
				     make_tdb_data((uint8 *)&lock_ref_count,
				     sizeof(lock_ref_count)), 0);

	SMB_ASSERT(NT_STATUS_IS_OK(status));

	TALLOC_FREE(rec);

	DEBUG(10,("reduce_windows_lock_ref_count for file now %s = %d\n",
		  fsp_str_dbg(fsp), lock_ref_count));
}

static void decrement_windows_lock_ref_count(files_struct *fsp)
{
	reduce_windows_lock_ref_count(fsp, 1);
}

/****************************************************************************
 Fetch the lock ref count.
****************************************************************************/

static int get_windows_lock_ref_count(files_struct *fsp)
{
	struct lock_ref_count_key tmp;
	TDB_DATA dbuf;
	NTSTATUS status;
	int lock_ref_count = 0;

	status = dbwrap_fetch(
		posix_pending_close_db, talloc_tos(),
		locking_ref_count_key_fsp(fsp, &tmp), &dbuf);

	if (NT_STATUS_EQUAL(status, NT_STATUS_NOT_FOUND)) {
		goto done;
	}

	if (!NT_STATUS_IS_OK(status)) {
		DEBUG(0, ("get_windows_lock_ref_count: Error fetching "
			  "lock ref count for file %s: %s\n",
			  fsp_str_dbg(fsp), nt_errstr(status)));
		goto done;
	}

	if (dbuf.dsize != sizeof(lock_ref_count)) {
		DEBUG(0, ("get_windows_lock_ref_count: invalid entry "
			  "in lock ref count record for file %s: "
			  "(invalid data size %u)\n",
			  fsp_str_dbg(fsp), (unsigned int)dbuf.dsize));
		goto done;
	}

	memcpy(&lock_ref_count, dbuf.dptr, sizeof(lock_ref_count));
	TALLOC_FREE(dbuf.dptr);

done:
	DEBUG(10,("get_windows_lock_count for file %s = %d\n",
		  fsp_str_dbg(fsp), lock_ref_count));

	return lock_ref_count;
}

/****************************************************************************
 Delete a lock_ref_count entry.
****************************************************************************/

static void delete_windows_lock_ref_count(files_struct *fsp)
{
	struct lock_ref_count_key tmp;
	struct db_record *rec;

	rec = dbwrap_fetch_locked(
		posix_pending_close_db, talloc_tos(),
		locking_ref_count_key_fsp(fsp, &tmp));

	SMB_ASSERT(rec != NULL);

	/* Not a bug if it doesn't exist - no locks were ever granted. */

	dbwrap_record_delete(rec);
	TALLOC_FREE(rec);

	DEBUG(10,("delete_windows_lock_ref_count for file %s\n",
		  fsp_str_dbg(fsp)));
}

/****************************************************************************
 Add an fd to the pending close tdb.
****************************************************************************/

static void add_fd_to_close_entry(files_struct *fsp)
{
	struct db_record *rec;
	uint8_t *new_data;
	NTSTATUS status;
	TDB_DATA value;

	rec = dbwrap_fetch_locked(
		posix_pending_close_db, talloc_tos(),
		fd_array_key_fsp(fsp));

	SMB_ASSERT(rec != NULL);

	value = dbwrap_record_get_value(rec);

	new_data = talloc_array(rec, uint8_t,
				value.dsize + sizeof(fsp->fh->fd));

	SMB_ASSERT(new_data != NULL);

	memcpy(new_data, value.dptr, value.dsize);
	memcpy(new_data + value.dsize,
	       &fsp->fh->fd, sizeof(fsp->fh->fd));

	status = dbwrap_record_store(
		rec, make_tdb_data(new_data,
				   value.dsize + sizeof(fsp->fh->fd)), 0);

	SMB_ASSERT(NT_STATUS_IS_OK(status));

	TALLOC_FREE(rec);

	DEBUG(10,("add_fd_to_close_entry: added fd %d file %s\n",
		  fsp->fh->fd, fsp_str_dbg(fsp)));
}

/****************************************************************************
 Remove all fd entries for a specific dev/inode pair from the tdb.
****************************************************************************/

static void delete_close_entries(files_struct *fsp)
{
	struct db_record *rec;

	rec = dbwrap_fetch_locked(
		posix_pending_close_db, talloc_tos(),
		fd_array_key_fsp(fsp));

	SMB_ASSERT(rec != NULL);
	dbwrap_record_delete(rec);
	TALLOC_FREE(rec);
}

/****************************************************************************
 Get the array of POSIX pending close records for an open fsp. Returns number
 of entries.
****************************************************************************/

static size_t get_posix_pending_close_entries(TALLOC_CTX *mem_ctx,
					      files_struct *fsp, int **entries)
{
	TDB_DATA dbuf;
	NTSTATUS status;

	status = dbwrap_fetch(
		posix_pending_close_db, mem_ctx, fd_array_key_fsp(fsp),
		&dbuf);

	if (NT_STATUS_EQUAL(status, NT_STATUS_NOT_FOUND)) {
		*entries = NULL;
		return 0;
	}

	SMB_ASSERT(NT_STATUS_IS_OK(status));

	if (dbuf.dsize == 0) {
		*entries = NULL;
		return 0;
	}

	*entries = (int *)dbuf.dptr;
	return (size_t)(dbuf.dsize / sizeof(int));
}

/****************************************************************************
 Deal with pending closes needed by POSIX locking support.
 Note that posix_locking_close_file() is expected to have been called
 to delete all locks on this fsp before this function is called.
****************************************************************************/

int fd_close_posix(struct files_struct *fsp)
{
	int saved_errno = 0;
	int ret;
	int *fd_array = NULL;
	size_t count, i;

	if (!lp_locking(fsp->conn->params) ||
	    !lp_posix_locking(fsp->conn->params))
	{
		/*
		 * No locking or POSIX to worry about or we want POSIX semantics
		 * which will lose all locks on all fd's open on this dev/inode,
		 * just close.
		 */
		return close(fsp->fh->fd);
	}

	if (get_windows_lock_ref_count(fsp)) {

		/*
		 * There are outstanding locks on this dev/inode pair on
		 * other fds. Add our fd to the pending close tdb and set
		 * fsp->fh->fd to -1.
		 */

		add_fd_to_close_entry(fsp);
		return 0;
	}

	/*
	 * No outstanding locks. Get the pending close fd's
	 * from the tdb and close them all.
	 */

	count = get_posix_pending_close_entries(talloc_tos(), fsp, &fd_array);

	if (count) {
		DEBUG(10,("fd_close_posix: doing close on %u fd's.\n",
			  (unsigned int)count));

		for(i = 0; i < count; i++) {
			if (close(fd_array[i]) == -1) {
				saved_errno = errno;
			}
		}

		/*
		 * Delete all fd's stored in the tdb
		 * for this dev/inode pair.
		 */

		delete_close_entries(fsp);
	}

	TALLOC_FREE(fd_array);

	/* Don't need a lock ref count on this dev/ino anymore. */
	delete_windows_lock_ref_count(fsp);

	/*
	 * Finally close the fd associated with this fsp.
	 */

	ret = close(fsp->fh->fd);

	if (ret == 0 && saved_errno != 0) {
		errno = saved_errno;
		ret = -1;
	}

	return ret;
}

/****************************************************************************
 Next - the functions that deal with the mapping CIFS Windows locks onto
 the underlying system POSIX locks.
****************************************************************************/

/*
 * Structure used when splitting a lock range
 * into a POSIX lock range. Doubly linked list.
 */

struct lock_list {
	struct lock_list *next;
	struct lock_list *prev;
	off_t start;
	off_t size;
};

/****************************************************************************
 Create a list of lock ranges that don't overlap a given range. Used in calculating
 POSIX locks and unlocks. This is a difficult function that requires ASCII art to
 understand it :-).
****************************************************************************/

static struct lock_list *posix_lock_list(TALLOC_CTX *ctx,
						struct lock_list *lhead,
						const struct lock_context *lock_ctx, /* Lock context lhead belongs to. */
						files_struct *fsp,
						const struct lock_struct *plocks,
						int num_locks)
{
	int i;

	/*
	 * Check the current lock list on this dev/inode pair.
	 * Quit if the list is deleted.
	 */

	DEBUG(10,("posix_lock_list: curr: start=%.0f,size=%.0f\n",
		(double)lhead->start, (double)lhead->size ));

	for (i=0; i<num_locks && lhead; i++) {
		const struct lock_struct *lock = &plocks[i];
		struct lock_list *l_curr;

		/* Ignore all but read/write locks. */
		if (lock->lock_type != READ_LOCK && lock->lock_type != WRITE_LOCK) {
			continue;
		}

		/* Ignore locks not owned by this process. */
		if (!procid_equal(&lock->context.pid, &lock_ctx->pid)) {
			continue;
		}

		/*
		 * Walk the lock list, checking for overlaps. Note that
		 * the lock list can expand within this loop if the current
		 * range being examined needs to be split.
		 */

		for (l_curr = lhead; l_curr;) {

			DEBUG(10,("posix_lock_list: lock: fnum=%d: start=%.0f,size=%.0f:type=%s", lock->fnum,
				(double)lock->start, (double)lock->size, posix_lock_type_name(lock->lock_type) ));

			if ( (l_curr->start >= (lock->start + lock->size)) ||
				 (lock->start >= (l_curr->start + l_curr->size))) {

				/* No overlap with existing lock - leave this range alone. */
/*********************************************
                                             +---------+
                                             | l_curr  |
                                             +---------+
                                +-------+
                                | lock  |
                                +-------+
OR....
             +---------+
             |  l_curr |
             +---------+
**********************************************/

				DEBUG(10,(" no overlap case.\n" ));

				l_curr = l_curr->next;

			} else if ( (l_curr->start >= lock->start) &&
						(l_curr->start + l_curr->size <= lock->start + lock->size) ) {

				/*
				 * This range is completely overlapped by this existing lock range
				 * and thus should have no effect. Delete it from the list.
				 */
/*********************************************
                +---------+
                |  l_curr |
                +---------+
        +---------------------------+
        |       lock                |
        +---------------------------+
**********************************************/
				/* Save the next pointer */
				struct lock_list *ul_next = l_curr->next;

				DEBUG(10,(" delete case.\n" ));

				DLIST_REMOVE(lhead, l_curr);
				if(lhead == NULL) {
					break; /* No more list... */
				}

				l_curr = ul_next;
				
			} else if ( (l_curr->start >= lock->start) &&
						(l_curr->start < lock->start + lock->size) &&
						(l_curr->start + l_curr->size > lock->start + lock->size) ) {

				/*
				 * This range overlaps the existing lock range at the high end.
				 * Truncate by moving start to existing range end and reducing size.
				 */
/*********************************************
                +---------------+
                |  l_curr       |
                +---------------+
        +---------------+
        |    lock       |
        +---------------+
BECOMES....
                        +-------+
                        | l_curr|
                        +-------+
**********************************************/

				l_curr->size = (l_curr->start + l_curr->size) - (lock->start + lock->size);
				l_curr->start = lock->start + lock->size;

				DEBUG(10,(" truncate high case: start=%.0f,size=%.0f\n",
								(double)l_curr->start, (double)l_curr->size ));

				l_curr = l_curr->next;

			} else if ( (l_curr->start < lock->start) &&
						(l_curr->start + l_curr->size > lock->start) &&
						(l_curr->start + l_curr->size <= lock->start + lock->size) ) {

				/*
				 * This range overlaps the existing lock range at the low end.
				 * Truncate by reducing size.
				 */
/*********************************************
   +---------------+
   |  l_curr       |
   +---------------+
           +---------------+
           |    lock       |
           +---------------+
BECOMES....
   +-------+
   | l_curr|
   +-------+
**********************************************/

				l_curr->size = lock->start - l_curr->start;

				DEBUG(10,(" truncate low case: start=%.0f,size=%.0f\n",
								(double)l_curr->start, (double)l_curr->size ));

				l_curr = l_curr->next;
		
			} else if ( (l_curr->start < lock->start) &&
						(l_curr->start + l_curr->size > lock->start + lock->size) ) {
				/*
				 * Worst case scenario. Range completely overlaps an existing
				 * lock range. Split the request into two, push the new (upper) request
				 * into the dlink list, and continue with the entry after l_new (as we
				 * know that l_new will not overlap with this lock).
				 */
/*********************************************
        +---------------------------+
        |        l_curr             |
        +---------------------------+
                +---------+
                | lock    |
                +---------+
BECOMES.....
        +-------+         +---------+
        | l_curr|         | l_new   |
        +-------+         +---------+
**********************************************/
				struct lock_list *l_new = talloc(ctx, struct lock_list);

				if(l_new == NULL) {
					DEBUG(0,("posix_lock_list: talloc fail.\n"));
					return NULL; /* The talloc_destroy takes care of cleanup. */
				}

				ZERO_STRUCTP(l_new);
				l_new->start = lock->start + lock->size;
				l_new->size = l_curr->start + l_curr->size - l_new->start;

				/* Truncate the l_curr. */
				l_curr->size = lock->start - l_curr->start;

				DEBUG(10,(" split case: curr: start=%.0f,size=%.0f \
new: start=%.0f,size=%.0f\n", (double)l_curr->start, (double)l_curr->size,
								(double)l_new->start, (double)l_new->size ));

				/*
				 * Add into the dlink list after the l_curr point - NOT at lhead. 
				 */
				DLIST_ADD_AFTER(lhead, l_new, l_curr);

				/* And move after the link we added. */
				l_curr = l_new->next;

			} else {

				/*
				 * This logic case should never happen. Ensure this is the
				 * case by forcing an abort.... Remove in production.
				 */
				char *msg = NULL;

				if (asprintf(&msg, "logic flaw in cases: l_curr: start = %.0f, size = %.0f : \
lock: start = %.0f, size = %.0f", (double)l_curr->start, (double)l_curr->size, (double)lock->start, (double)lock->size ) != -1) {
					smb_panic(msg);
				} else {
					smb_panic("posix_lock_list");
				}
			}
		} /* end for ( l_curr = lhead; l_curr;) */
	} /* end for (i=0; i<num_locks && ul_head; i++) */

	return lhead;
}

/****************************************************************************
 POSIX function to acquire a lock. Returns True if the
 lock could be granted, False if not.
****************************************************************************/

bool set_posix_lock_windows_flavour(files_struct *fsp,
			uint64_t u_offset,
			uint64_t u_count,
			enum brl_type lock_type,
			const struct lock_context *lock_ctx,
			const struct lock_struct *plocks,
			int num_locks,
			int *errno_ret)
{
	off_t offset;
	off_t count;
	int posix_lock_type = map_posix_lock_type(fsp,lock_type);
	bool ret = True;
	size_t lock_count;
	TALLOC_CTX *l_ctx = NULL;
	struct lock_list *llist = NULL;
	struct lock_list *ll = NULL;

	DEBUG(5,("set_posix_lock_windows_flavour: File %s, offset = %.0f, "
		 "count = %.0f, type = %s\n", fsp_str_dbg(fsp),
		 (double)u_offset, (double)u_count,
		 posix_lock_type_name(lock_type)));

	/*
	 * If the requested lock won't fit in the POSIX range, we will
	 * pretend it was successful.
	 */

	if(!posix_lock_in_range(&offset, &count, u_offset, u_count)) {
		increment_windows_lock_ref_count(fsp);
		return True;
	}

	/*
	 * Windows is very strange. It allows read locks to be overlayed
	 * (even over a write lock), but leaves the write lock in force until the first
	 * unlock. It also reference counts the locks. This means the following sequence :
	 *
	 * process1                                      process2
	 * ------------------------------------------------------------------------
	 * WRITE LOCK : start = 2, len = 10
	 *                                            READ LOCK: start =0, len = 10 - FAIL
	 * READ LOCK : start = 0, len = 14 
	 *                                            READ LOCK: start =0, len = 10 - FAIL
	 * UNLOCK : start = 2, len = 10
	 *                                            READ LOCK: start =0, len = 10 - OK
	 *
	 * Under POSIX, the same sequence in steps 1 and 2 would not be reference counted, but
	 * would leave a single read lock over the 0-14 region.
	 */
	
	if ((l_ctx = talloc_init("set_posix_lock")) == NULL) {
		DEBUG(0,("set_posix_lock_windows_flavour: unable to init talloc context.\n"));
		return False;
	}

	if ((ll = talloc(l_ctx, struct lock_list)) == NULL) {
		DEBUG(0,("set_posix_lock_windows_flavour: unable to talloc unlock list.\n"));
		talloc_destroy(l_ctx);
		return False;
	}

	/*
	 * Create the initial list entry containing the
	 * lock we want to add.
	 */

	ZERO_STRUCTP(ll);
	ll->start = offset;
	ll->size = count;

	DLIST_ADD(llist, ll);

	/*
	 * The following call calculates if there are any
	 * overlapping locks held by this process on
	 * fd's open on the same file and splits this list
	 * into a list of lock ranges that do not overlap with existing
	 * POSIX locks.
	 */

	llist = posix_lock_list(l_ctx,
				llist,
				lock_ctx, /* Lock context llist belongs to. */
				fsp,
				plocks,
				num_locks);

	/*
	 * Add the POSIX locks on the list of ranges returned.
	 * As the lock is supposed to be added atomically, we need to
	 * back out all the locks if any one of these calls fail.
	 */

	for (lock_count = 0, ll = llist; ll; ll = ll->next, lock_count++) {
		offset = ll->start;
		count = ll->size;

		DEBUG(5,("set_posix_lock_windows_flavour: Real lock: Type = %s: offset = %.0f, count = %.0f\n",
			posix_lock_type_name(posix_lock_type), (double)offset, (double)count ));

		if (!posix_fcntl_lock(fsp,F_SETLK,offset,count,posix_lock_type)) {
			*errno_ret = errno;
			DEBUG(5,("set_posix_lock_windows_flavour: Lock fail !: Type = %s: offset = %.0f, count = %.0f. Errno = %s\n",
				posix_lock_type_name(posix_lock_type), (double)offset, (double)count, strerror(errno) ));
			ret = False;
			break;
		}
	}

	if (!ret) {

		/*
		 * Back out all the POSIX locks we have on fail.
		 */

		for (ll = llist; lock_count; ll = ll->next, lock_count--) {
			offset = ll->start;
			count = ll->size;

			DEBUG(5,("set_posix_lock_windows_flavour: Backing out locks: Type = %s: offset = %.0f, count = %.0f\n",
				posix_lock_type_name(posix_lock_type), (double)offset, (double)count ));

			posix_fcntl_lock(fsp,F_SETLK,offset,count,F_UNLCK);
		}
	} else {
		/* Remember the number of Windows locks we have on this dev/ino pair. */
		increment_windows_lock_ref_count(fsp);
	}

	talloc_destroy(l_ctx);
	return ret;
}

/****************************************************************************
 POSIX function to release a lock. Returns True if the
 lock could be released, False if not.
****************************************************************************/

bool release_posix_lock_windows_flavour(files_struct *fsp,
				uint64_t u_offset,
				uint64_t u_count,
				enum brl_type deleted_lock_type,
				const struct lock_context *lock_ctx,
				const struct lock_struct *plocks,
				int num_locks)
{
	off_t offset;
	off_t count;
	bool ret = True;
	TALLOC_CTX *ul_ctx = NULL;
	struct lock_list *ulist = NULL;
	struct lock_list *ul = NULL;

	DEBUG(5,("release_posix_lock_windows_flavour: File %s, offset = %.0f, "
		 "count = %.0f\n", fsp_str_dbg(fsp),
		 (double)u_offset, (double)u_count));

	/* Remember the number of Windows locks we have on this dev/ino pair. */
	decrement_windows_lock_ref_count(fsp);

	/*
	 * If the requested lock won't fit in the POSIX range, we will
	 * pretend it was successful.
	 */

	if(!posix_lock_in_range(&offset, &count, u_offset, u_count)) {
		return True;
	}

	if ((ul_ctx = talloc_init("release_posix_lock")) == NULL) {
		DEBUG(0,("release_posix_lock_windows_flavour: unable to init talloc context.\n"));
		return False;
	}

	if ((ul = talloc(ul_ctx, struct lock_list)) == NULL) {
		DEBUG(0,("release_posix_lock_windows_flavour: unable to talloc unlock list.\n"));
		talloc_destroy(ul_ctx);
		return False;
	}

	/*
	 * Create the initial list entry containing the
	 * lock we want to remove.
	 */

	ZERO_STRUCTP(ul);
	ul->start = offset;
	ul->size = count;

	DLIST_ADD(ulist, ul);

	/*
	 * The following call calculates if there are any
	 * overlapping locks held by this process on
	 * fd's open on the same file and creates a
	 * list of unlock ranges that will allow
	 * POSIX lock ranges to remain on the file whilst the
	 * unlocks are performed.
	 */

	ulist = posix_lock_list(ul_ctx,
				ulist,
				lock_ctx, /* Lock context ulist belongs to. */
				fsp,
				plocks,
				num_locks);

	/*
	 * If there were any overlapped entries (list is > 1 or size or start have changed),
	 * and the lock_type we just deleted from
	 * the upper layer tdb was a write lock, then before doing the unlock we need to downgrade
	 * the POSIX lock to a read lock. This allows any overlapping read locks
	 * to be atomically maintained.
	 */

	if (deleted_lock_type == WRITE_LOCK &&
			(!ulist || ulist->next != NULL || ulist->start != offset || ulist->size != count)) {

		DEBUG(5,("release_posix_lock_windows_flavour: downgrading lock to READ: offset = %.0f, count = %.0f\n",
			(double)offset, (double)count ));

		if (!posix_fcntl_lock(fsp,F_SETLK,offset,count,F_RDLCK)) {
			DEBUG(0,("release_posix_lock_windows_flavour: downgrade of lock failed with error %s !\n", strerror(errno) ));
			talloc_destroy(ul_ctx);
			return False;
		}
	}

	/*
	 * Release the POSIX locks on the list of ranges returned.
	 */

	for(; ulist; ulist = ulist->next) {
		offset = ulist->start;
		count = ulist->size;

		DEBUG(5,("release_posix_lock_windows_flavour: Real unlock: offset = %.0f, count = %.0f\n",
			(double)offset, (double)count ));

		if (!posix_fcntl_lock(fsp,F_SETLK,offset,count,F_UNLCK)) {
			ret = False;
		}
	}

	talloc_destroy(ul_ctx);
	return ret;
}

/****************************************************************************
 Next - the functions that deal with mapping CIFS POSIX locks onto
 the underlying system POSIX locks.
****************************************************************************/

/****************************************************************************
 POSIX function to acquire a lock. Returns True if the
 lock could be granted, False if not.
 As POSIX locks don't stack or conflict (they just overwrite)
 we can map the requested lock directly onto a system one. We
 know it doesn't conflict with locks on other contexts as the
 upper layer would have refused it.
****************************************************************************/

bool set_posix_lock_posix_flavour(files_struct *fsp,
			uint64_t u_offset,
			uint64_t u_count,
			enum brl_type lock_type,
			int *errno_ret)
{
	off_t offset;
	off_t count;
	int posix_lock_type = map_posix_lock_type(fsp,lock_type);

	DEBUG(5,("set_posix_lock_posix_flavour: File %s, offset = %.0f, count "
		 "= %.0f, type = %s\n", fsp_str_dbg(fsp),
		 (double)u_offset, (double)u_count,
		 posix_lock_type_name(lock_type)));

	/*
	 * If the requested lock won't fit in the POSIX range, we will
	 * pretend it was successful.
	 */

	if(!posix_lock_in_range(&offset, &count, u_offset, u_count)) {
		return True;
	}

	if (!posix_fcntl_lock(fsp,F_SETLK,offset,count,posix_lock_type)) {
		*errno_ret = errno;
		DEBUG(5,("set_posix_lock_posix_flavour: Lock fail !: Type = %s: offset = %.0f, count = %.0f. Errno = %s\n",
			posix_lock_type_name(posix_lock_type), (double)offset, (double)count, strerror(errno) ));
		return False;
	}
	return True;
}

/****************************************************************************
 POSIX function to release a lock. Returns True if the
 lock could be released, False if not.
 We are given a complete lock state from the upper layer which is what the lock
 state should be after the unlock has already been done, so what
 we do is punch out holes in the unlock range where locks owned by this process
 have a different lock context.
****************************************************************************/

bool release_posix_lock_posix_flavour(files_struct *fsp,
				uint64_t u_offset,
				uint64_t u_count,
				const struct lock_context *lock_ctx,
				const struct lock_struct *plocks,
				int num_locks)
{
	bool ret = True;
	off_t offset;
	off_t count;
	TALLOC_CTX *ul_ctx = NULL;
	struct lock_list *ulist = NULL;
	struct lock_list *ul = NULL;

	DEBUG(5,("release_posix_lock_posix_flavour: File %s, offset = %.0f, "
		 "count = %.0f\n", fsp_str_dbg(fsp),
		 (double)u_offset, (double)u_count));

	/*
	 * If the requested lock won't fit in the POSIX range, we will
	 * pretend it was successful.
	 */

	if(!posix_lock_in_range(&offset, &count, u_offset, u_count)) {
		return True;
	}

	if ((ul_ctx = talloc_init("release_posix_lock")) == NULL) {
		DEBUG(0,("release_posix_lock_windows_flavour: unable to init talloc context.\n"));
		return False;
	}

	if ((ul = talloc(ul_ctx, struct lock_list)) == NULL) {
		DEBUG(0,("release_posix_lock_windows_flavour: unable to talloc unlock list.\n"));
		talloc_destroy(ul_ctx);
		return False;
	}

	/*
	 * Create the initial list entry containing the
	 * lock we want to remove.
	 */

	ZERO_STRUCTP(ul);
	ul->start = offset;
	ul->size = count;

	DLIST_ADD(ulist, ul);

	/*
	 * Walk the given array creating a linked list
	 * of unlock requests.
	 */

	ulist = posix_lock_list(ul_ctx,
				ulist,
				lock_ctx, /* Lock context ulist belongs to. */
				fsp,
				plocks,
				num_locks);

	/*
	 * Release the POSIX locks on the list of ranges returned.
	 */

	for(; ulist; ulist = ulist->next) {
		offset = ulist->start;
		count = ulist->size;

		DEBUG(5,("release_posix_lock_posix_flavour: Real unlock: offset = %.0f, count = %.0f\n",
			(double)offset, (double)count ));

		if (!posix_fcntl_lock(fsp,F_SETLK,offset,count,F_UNLCK)) {
			ret = False;
		}
	}

	talloc_destroy(ul_ctx);
	return ret;
}