summaryrefslogtreecommitdiff
path: root/source3/locking/brlock.c
blob: b5eebc8e0401984d1d618c99e20585b08690ad2f (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
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
/*
   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 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/>.
*/

/* 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"
#include "system/filesys.h"
#include "locking/proto.h"
#include "smbd/globals.h"
#include "dbwrap/dbwrap.h"
#include "dbwrap/dbwrap_open.h"
#include "serverid.h"
#include "messages.h"
#include "util_tdb.h"

#undef DBGC_CLASS
#define DBGC_CLASS DBGC_LOCKING

#define ZERO_ZERO 0

/* The open brlock.tdb database. */

static struct db_context *brlock_db;

struct byte_range_lock {
	struct files_struct *fsp;
	unsigned int num_locks;
	bool modified;
	bool have_read_oplocks;
	struct lock_struct *lock_data;
	struct db_record *record;
};

/****************************************************************************
 Debug info at level 10 for lock struct.
****************************************************************************/

static void print_lock_struct(unsigned int i, const struct lock_struct *pls)
{
	DEBUG(10,("[%u]: smblctx = %llu, tid = %u, pid = %s, ",
			i,
			(unsigned long long)pls->context.smblctx,
			(unsigned int)pls->context.tid,
			server_id_str(talloc_tos(), &pls->context.pid) ));

	DEBUG(10,("start = %.0f, size = %.0f, fnum = %llu, %s %s\n",
		(double)pls->start,
		(double)pls->size,
		(unsigned long long)pls->fnum,
		lock_type_name(pls->lock_type),
		lock_flav_name(pls->lock_flav) ));
}

unsigned int brl_num_locks(const struct byte_range_lock *brl)
{
	return brl->num_locks;
}

struct files_struct *brl_fsp(struct byte_range_lock *brl)
{
	return brl->fsp;
}

bool brl_have_read_oplocks(const struct byte_range_lock *brl)
{
	return brl->have_read_oplocks;
}

void brl_set_have_read_oplocks(struct byte_range_lock *brl,
			       bool have_read_oplocks)
{
	DEBUG(10, ("Setting have_read_oplocks to %s\n",
		   have_read_oplocks ? "true" : "false"));
	SMB_ASSERT(brl->record != NULL); /* otherwise we're readonly */
	brl->have_read_oplocks = have_read_oplocks;
	brl->modified = true;
}

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

static bool brl_same_context(const struct lock_context *ctx1,
			     const struct lock_context *ctx2)
{
	return (serverid_equal(&ctx1->pid, &ctx2->pid) &&
		(ctx1->smblctx == ctx2->smblctx) &&
		(ctx1->tid == ctx2->tid));
}

/****************************************************************************
 See if lck1 and lck2 overlap.
****************************************************************************/

static bool brl_overlap(const struct lock_struct *lck1,
                        const struct lock_struct *lck2)
{
	/* XXX Remove for Win7 compatibility. */
	/* this extra check is not redundant - it copes with locks
	   that go beyond the end of 64 bit file space */
	if (lck1->size != 0 &&
	    lck1->start == lck2->start &&
	    lck1->size == lck2->size) {
		return True;
	}

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

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

static bool brl_conflict(const struct lock_struct *lck1,
			 const struct lock_struct *lck2)
{
	/* Ignore PENDING locks. */
	if (IS_PENDING_LOCK(lck1->lock_type) || IS_PENDING_LOCK(lck2->lock_type))
		return False;

	/* Read locks never conflict. */
	if (lck1->lock_type == READ_LOCK && lck2->lock_type == READ_LOCK) {
		return False;
	}

	/* A READ lock can stack on top of a WRITE lock if they have the same
	 * context & fnum. */
	if (lck1->lock_type == WRITE_LOCK && lck2->lock_type == READ_LOCK &&
	    brl_same_context(&lck1->context, &lck2->context) &&
	    lck1->fnum == lck2->fnum) {
		return False;
	}

	return brl_overlap(lck1, lck2);
}

/****************************************************************************
 See if lock2 can be added when lock1 is in place - when both locks are POSIX
 flavour. POSIX locks ignore fnum - they only care about dev/ino which we
 know already match.
****************************************************************************/

static bool brl_conflict_posix(const struct lock_struct *lck1,
			 	const struct lock_struct *lck2)
{
#if defined(DEVELOPER)
	SMB_ASSERT(lck1->lock_flav == POSIX_LOCK);
	SMB_ASSERT(lck2->lock_flav == POSIX_LOCK);
#endif

	/* Ignore PENDING locks. */
	if (IS_PENDING_LOCK(lck1->lock_type) || IS_PENDING_LOCK(lck2->lock_type))
		return False;

	/* Read locks never conflict. */
	if (lck1->lock_type == READ_LOCK && lck2->lock_type == READ_LOCK) {
		return False;
	}

	/* Locks on the same context con't conflict. Ignore fnum. */
	if (brl_same_context(&lck1->context, &lck2->context)) {
		return False;
	}

	/* One is read, the other write, or the context is different,
	   do they overlap ? */
	return brl_overlap(lck1, lck2);
}

#if ZERO_ZERO
static bool brl_conflict1(const struct lock_struct *lck1,
			 const struct lock_struct *lck2)
{
	if (IS_PENDING_LOCK(lck1->lock_type) || IS_PENDING_LOCK(lck2->lock_type))
		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. This is the read/write lock check code path.
 This is never used in the POSIX lock case.
****************************************************************************/

static bool brl_conflict_other(const struct lock_struct *lck1, const struct lock_struct *lck2)
{
	if (IS_PENDING_LOCK(lck1->lock_type) || IS_PENDING_LOCK(lck2->lock_type))
		return False;

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

	/* POSIX flavour locks never conflict here - this is only called
	   in the read/write path. */

	if (lck1->lock_flav == POSIX_LOCK && lck2->lock_flav == POSIX_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;
	}

	return brl_overlap(lck1, lck2);
}

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

static bool brl_pending_overlap(const struct lock_struct *lock, const 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;
}

/****************************************************************************
 Amazingly enough, w2k3 "remembers" whether the last lock failure on a fnum
 is the same as this one and changes its error code. I wonder if any
 app depends on this ?
****************************************************************************/

static NTSTATUS brl_lock_failed(files_struct *fsp,
				const struct lock_struct *lock,
				bool blocking_lock)
{
	if (lock->start >= 0xEF000000 && (lock->start >> 63) == 0) {
		/* amazing the little things you learn with a test
		   suite. Locks beyond this offset (as a 64 bit
		   number!) always generate the conflict error code,
		   unless the top bit is set */
		if (!blocking_lock) {
			fsp->last_lock_failure = *lock;
		}
		return NT_STATUS_FILE_LOCK_CONFLICT;
	}

	if (serverid_equal(&lock->context.pid, &fsp->last_lock_failure.context.pid) &&
			lock->context.tid == fsp->last_lock_failure.context.tid &&
			lock->fnum == fsp->last_lock_failure.fnum &&
			lock->start == fsp->last_lock_failure.start) {
		return NT_STATUS_FILE_LOCK_CONFLICT;
	}

	if (!blocking_lock) {
		fsp->last_lock_failure = *lock;
	}
	return NT_STATUS_LOCK_NOT_GRANTED;
}

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

void brl_init(bool read_only)
{
	int tdb_flags;

	if (brlock_db) {
		return;
	}

	tdb_flags = TDB_DEFAULT|TDB_VOLATILE|TDB_CLEAR_IF_FIRST|TDB_INCOMPATIBLE_HASH;

	if (!lp_clustering()) {
		/*
		 * We can't use the SEQNUM trick to cache brlock
		 * entries in the clustering case because ctdb seqnum
		 * propagation has a delay.
		 */
		tdb_flags |= TDB_SEQNUM;
	}

	brlock_db = db_open(NULL, lock_path("brlock.tdb"),
			    lp_open_files_db_hash_size(), tdb_flags,
			    read_only?O_RDONLY:(O_RDWR|O_CREAT), 0644,
			    DBWRAP_LOCK_ORDER_2);
	if (!brlock_db) {
		DEBUG(0,("Failed to open byte range locking database %s\n",
			lock_path("brlock.tdb")));
		return;
	}
}

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

void brl_shutdown(void)
{
	TALLOC_FREE(brlock_db);
}

#if ZERO_ZERO
/****************************************************************************
 Compare two locks for sorting.
****************************************************************************/

static int lock_compare(const struct lock_struct *lck1,
			 const 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 - Windows lock semantics.
****************************************************************************/

NTSTATUS brl_lock_windows_default(struct byte_range_lock *br_lck,
    struct lock_struct *plock, bool blocking_lock)
{
	unsigned int i;
	files_struct *fsp = br_lck->fsp;
	struct lock_struct *locks = br_lck->lock_data;
	NTSTATUS status;

	SMB_ASSERT(plock->lock_type != UNLOCK_LOCK);

	if ((plock->start + plock->size - 1 < plock->start) &&
			plock->size != 0) {
		return NT_STATUS_INVALID_LOCK_RANGE;
	}

	for (i=0; i < br_lck->num_locks; i++) {
		/* Do any Windows or POSIX locks conflict ? */
		if (brl_conflict(&locks[i], plock)) {
			/* Remember who blocked us. */
			plock->context.smblctx = locks[i].context.smblctx;
			return brl_lock_failed(fsp,plock,blocking_lock);
		}
#if ZERO_ZERO
		if (plock->start == 0 && plock->size == 0 &&
				locks[i].size == 0) {
			break;
		}
#endif
	}

	if (!IS_PENDING_LOCK(plock->lock_type)) {
		contend_level2_oplocks_begin(fsp, LEVEL2_CONTEND_WINDOWS_BRL);
	}

	/* We can get the Windows lock, now see if it needs to
	   be mapped into a lower level POSIX one, and if so can
	   we get it ? */

	if (!IS_PENDING_LOCK(plock->lock_type) && lp_posix_locking(fsp->conn->params)) {
		int errno_ret;
		if (!set_posix_lock_windows_flavour(fsp,
				plock->start,
				plock->size,
				plock->lock_type,
				&plock->context,
				locks,
				br_lck->num_locks,
				&errno_ret)) {

			/* We don't know who blocked us. */
			plock->context.smblctx = 0xFFFFFFFFFFFFFFFFLL;

			if (errno_ret == EACCES || errno_ret == EAGAIN) {
				status = NT_STATUS_FILE_LOCK_CONFLICT;
				goto fail;
			} else {
				status = map_nt_error_from_unix(errno);
				goto fail;
			}
		}
	}

	/* no conflicts - add it to the list of locks */
	locks = talloc_realloc(br_lck, locks, struct lock_struct,
			       (br_lck->num_locks + 1));
	if (!locks) {
		status = NT_STATUS_NO_MEMORY;
		goto fail;
	}

	memcpy(&locks[br_lck->num_locks], plock, sizeof(struct lock_struct));
	br_lck->num_locks += 1;
	br_lck->lock_data = locks;
	br_lck->modified = True;

	return NT_STATUS_OK;
 fail:
	if (!IS_PENDING_LOCK(plock->lock_type)) {
		contend_level2_oplocks_end(fsp, LEVEL2_CONTEND_WINDOWS_BRL);
	}
	return status;
}

/****************************************************************************
 Cope with POSIX range splits and merges.
****************************************************************************/

static unsigned int brlock_posix_split_merge(struct lock_struct *lck_arr,	/* Output array. */
						struct lock_struct *ex,		/* existing lock. */
						struct lock_struct *plock)	/* proposed lock. */
{
	bool lock_types_differ = (ex->lock_type != plock->lock_type);

	/* We can't merge non-conflicting locks on different context - ignore fnum. */

	if (!brl_same_context(&ex->context, &plock->context)) {
		/* Just copy. */
		memcpy(&lck_arr[0], ex, sizeof(struct lock_struct));
		return 1;
	}

	/* We now know we have the same context. */

	/* Did we overlap ? */

/*********************************************
                                        +---------+
                                        | ex      |
                                        +---------+
                         +-------+
                         | plock |
                         +-------+
OR....
        +---------+
        |  ex     |
        +---------+
**********************************************/

	if ( (ex->start > (plock->start + plock->size)) ||
		(plock->start > (ex->start + ex->size))) {

		/* No overlap with this lock - copy existing. */

		memcpy(&lck_arr[0], ex, sizeof(struct lock_struct));
		return 1;
	}

/*********************************************
        +---------------------------+
        |          ex               |
        +---------------------------+
        +---------------------------+
        |       plock               | -> replace with plock.
        +---------------------------+
OR
             +---------------+
             |       ex      |
             +---------------+
        +---------------------------+
        |       plock               | -> replace with plock.
        +---------------------------+

**********************************************/

	if ( (ex->start >= plock->start) &&
		(ex->start + ex->size <= plock->start + plock->size) ) {

		/* Replace - discard existing lock. */

		return 0;
	}

/*********************************************
Adjacent after.
                        +-------+
                        |  ex   |
                        +-------+
        +---------------+
        |   plock       |
        +---------------+

BECOMES....
        +---------------+-------+
        |   plock       | ex    | - different lock types.
        +---------------+-------+
OR.... (merge)
        +-----------------------+
        |   plock               | - same lock type.
        +-----------------------+
**********************************************/

	if (plock->start + plock->size == ex->start) {

		/* If the lock types are the same, we merge, if different, we
		   add the remainder of the old lock. */

		if (lock_types_differ) {
			/* Add existing. */
			memcpy(&lck_arr[0], ex, sizeof(struct lock_struct));
			return 1;
		} else {
			/* Merge - adjust incoming lock as we may have more
			 * merging to come. */
			plock->size += ex->size;
			return 0;
		}
	}

/*********************************************
Adjacent before.
        +-------+
        |  ex   |
        +-------+
                +---------------+
                |   plock       |
                +---------------+
BECOMES....
        +-------+---------------+
        | ex    |   plock       | - different lock types
        +-------+---------------+

OR.... (merge)
        +-----------------------+
        |      plock            | - same lock type.
        +-----------------------+

**********************************************/

	if (ex->start + ex->size == plock->start) {

		/* If the lock types are the same, we merge, if different, we
		   add the existing lock. */

		if (lock_types_differ) {
			memcpy(&lck_arr[0], ex, sizeof(struct lock_struct));
			return 1;
		} else {
			/* Merge - adjust incoming lock as we may have more
			 * merging to come. */
			plock->start = ex->start;
			plock->size += ex->size;
			return 0;
		}
	}

/*********************************************
Overlap after.
        +-----------------------+
        |          ex           |
        +-----------------------+
        +---------------+
        |   plock       |
        +---------------+
OR
               +----------------+
               |       ex       |
               +----------------+
        +---------------+
        |   plock       |
        +---------------+

BECOMES....
        +---------------+-------+
        |   plock       | ex    | - different lock types.
        +---------------+-------+
OR.... (merge)
        +-----------------------+
        |   plock               | - same lock type.
        +-----------------------+
**********************************************/

	if ( (ex->start >= plock->start) &&
		(ex->start <= plock->start + plock->size) &&
		(ex->start + ex->size > plock->start + plock->size) ) {

		/* If the lock types are the same, we merge, if different, we
		   add the remainder of the old lock. */

		if (lock_types_differ) {
			/* Add remaining existing. */
			memcpy(&lck_arr[0], ex, sizeof(struct lock_struct));
			/* Adjust existing start and size. */
			lck_arr[0].start = plock->start + plock->size;
			lck_arr[0].size = (ex->start + ex->size) - (plock->start + plock->size);
			return 1;
		} else {
			/* Merge - adjust incoming lock as we may have more
			 * merging to come. */
			plock->size += (ex->start + ex->size) - (plock->start + plock->size);
			return 0;
		}
	}

/*********************************************
Overlap before.
        +-----------------------+
        |  ex                   |
        +-----------------------+
                +---------------+
                |   plock       |
                +---------------+
OR
        +-------------+
        |  ex         |
        +-------------+
                +---------------+
                |   plock       |
                +---------------+

BECOMES....
        +-------+---------------+
        | ex    |   plock       | - different lock types
        +-------+---------------+

OR.... (merge)
        +-----------------------+
        |      plock            | - same lock type.
        +-----------------------+

**********************************************/

	if ( (ex->start < plock->start) &&
			(ex->start + ex->size >= plock->start) &&
			(ex->start + ex->size <= plock->start + plock->size) ) {

		/* If the lock types are the same, we merge, if different, we
		   add the truncated old lock. */

		if (lock_types_differ) {
			memcpy(&lck_arr[0], ex, sizeof(struct lock_struct));
			/* Adjust existing size. */
			lck_arr[0].size = plock->start - ex->start;
			return 1;
		} else {
			/* Merge - adjust incoming lock as we may have more
			 * merging to come. MUST ADJUST plock SIZE FIRST ! */
			plock->size += (plock->start - ex->start);
			plock->start = ex->start;
			return 0;
		}
	}

/*********************************************
Complete overlap.
        +---------------------------+
        |        ex                 |
        +---------------------------+
                +---------+
                |  plock  |
                +---------+
BECOMES.....
        +-------+---------+---------+
        | ex    |  plock  | ex      | - different lock types.
        +-------+---------+---------+
OR
        +---------------------------+
        |        plock              | - same lock type.
        +---------------------------+
**********************************************/

	if ( (ex->start < plock->start) && (ex->start + ex->size > plock->start + plock->size) ) {

		if (lock_types_differ) {

			/* We have to split ex into two locks here. */

			memcpy(&lck_arr[0], ex, sizeof(struct lock_struct));
			memcpy(&lck_arr[1], ex, sizeof(struct lock_struct));

			/* Adjust first existing size. */
			lck_arr[0].size = plock->start - ex->start;

			/* Adjust second existing start and size. */
			lck_arr[1].start = plock->start + plock->size;
			lck_arr[1].size = (ex->start + ex->size) - (plock->start + plock->size);
			return 2;
		} else {
			/* Just eat the existing locks, merge them into plock. */
			plock->start = ex->start;
			plock->size = ex->size;
			return 0;
		}
	}

	/* Never get here. */
	smb_panic("brlock_posix_split_merge");
	/* Notreached. */

	/* Keep some compilers happy. */
	return 0;
}

/****************************************************************************
 Lock a range of bytes - POSIX lock semantics.
 We must cope with range splits and merges.
****************************************************************************/

static NTSTATUS brl_lock_posix(struct messaging_context *msg_ctx,
			       struct byte_range_lock *br_lck,
			       struct lock_struct *plock)
{
	unsigned int i, count, posix_count;
	struct lock_struct *locks = br_lck->lock_data;
	struct lock_struct *tp;
	bool signal_pending_read = False;
	bool break_oplocks = false;
	NTSTATUS status;

	/* No zero-zero locks for POSIX. */
	if (plock->start == 0 && plock->size == 0) {
		return NT_STATUS_INVALID_PARAMETER;
	}

	/* Don't allow 64-bit lock wrap. */
	if (plock->start + plock->size - 1 < plock->start) {
		return NT_STATUS_INVALID_PARAMETER;
	}

	/* The worst case scenario here is we have to split an
	   existing POSIX lock range into two, and add our lock,
	   so we need at most 2 more entries. */

	tp = talloc_array(br_lck, struct lock_struct, br_lck->num_locks + 2);
	if (!tp) {
		return NT_STATUS_NO_MEMORY;
	}

	count = posix_count = 0;

	for (i=0; i < br_lck->num_locks; i++) {
		struct lock_struct *curr_lock = &locks[i];

		/* If we have a pending read lock, a lock downgrade should
		   trigger a lock re-evaluation. */
		if (curr_lock->lock_type == PENDING_READ_LOCK &&
				brl_pending_overlap(plock, curr_lock)) {
			signal_pending_read = True;
		}

		if (curr_lock->lock_flav == WINDOWS_LOCK) {
			/* Do any Windows flavour locks conflict ? */
			if (brl_conflict(curr_lock, plock)) {
				/* No games with error messages. */
				TALLOC_FREE(tp);
				/* Remember who blocked us. */
				plock->context.smblctx = curr_lock->context.smblctx;
				return NT_STATUS_FILE_LOCK_CONFLICT;
			}
			/* Just copy the Windows lock into the new array. */
			memcpy(&tp[count], curr_lock, sizeof(struct lock_struct));
			count++;
		} else {
			unsigned int tmp_count = 0;

			/* POSIX conflict semantics are different. */
			if (brl_conflict_posix(curr_lock, plock)) {
				/* Can't block ourselves with POSIX locks. */
				/* No games with error messages. */
				TALLOC_FREE(tp);
				/* Remember who blocked us. */
				plock->context.smblctx = curr_lock->context.smblctx;
				return NT_STATUS_FILE_LOCK_CONFLICT;
			}

			/* Work out overlaps. */
			tmp_count += brlock_posix_split_merge(&tp[count], curr_lock, plock);
			posix_count += tmp_count;
			count += tmp_count;
		}
	}

	/*
	 * Break oplocks while we hold a brl. Since lock() and unlock() calls
	 * are not symetric with POSIX semantics, we cannot guarantee our
	 * contend_level2_oplocks_begin/end calls will be acquired and
	 * released one-for-one as with Windows semantics. Therefore we only
	 * call contend_level2_oplocks_begin if this is the first POSIX brl on
	 * the file.
	 */
	break_oplocks = (!IS_PENDING_LOCK(plock->lock_type) &&
			 posix_count == 0);
	if (break_oplocks) {
		contend_level2_oplocks_begin(br_lck->fsp,
					     LEVEL2_CONTEND_POSIX_BRL);
	}

	/* Try and add the lock in order, sorted by lock start. */
	for (i=0; i < count; i++) {
		struct lock_struct *curr_lock = &tp[i];

		if (curr_lock->start <= plock->start) {
			continue;
		}
	}

	if (i < count) {
		memmove(&tp[i+1], &tp[i],
			(count - i)*sizeof(struct lock_struct));
	}
	memcpy(&tp[i], plock, sizeof(struct lock_struct));
	count++;

	/* We can get the POSIX lock, now see if it needs to
	   be mapped into a lower level POSIX one, and if so can
	   we get it ? */

	if (!IS_PENDING_LOCK(plock->lock_type) && lp_posix_locking(br_lck->fsp->conn->params)) {
		int errno_ret;

		/* The lower layer just needs to attempt to
		   get the system POSIX lock. We've weeded out
		   any conflicts above. */

		if (!set_posix_lock_posix_flavour(br_lck->fsp,
				plock->start,
				plock->size,
				plock->lock_type,
				&errno_ret)) {

			/* We don't know who blocked us. */
			plock->context.smblctx = 0xFFFFFFFFFFFFFFFFLL;

			if (errno_ret == EACCES || errno_ret == EAGAIN) {
				TALLOC_FREE(tp);
				status = NT_STATUS_FILE_LOCK_CONFLICT;
				goto fail;
			} else {
				TALLOC_FREE(tp);
				status = map_nt_error_from_unix(errno);
				goto fail;
			}
		}
	}

	/* If we didn't use all the allocated size,
	 * Realloc so we don't leak entries per lock call. */
	if (count < br_lck->num_locks + 2) {
		tp = talloc_realloc(br_lck, tp, struct lock_struct, count);
		if (!tp) {
			status = NT_STATUS_NO_MEMORY;
			goto fail;
		}
	}

	br_lck->num_locks = count;
	TALLOC_FREE(br_lck->lock_data);
	br_lck->lock_data = tp;
	locks = tp;
	br_lck->modified = True;

	/* A successful downgrade from write to read lock can trigger a lock
	   re-evalutation where waiting readers can now proceed. */

	if (signal_pending_read) {
		/* Send unlock messages to any pending read waiters that overlap. */
		for (i=0; i < br_lck->num_locks; i++) {
			struct lock_struct *pend_lock = &locks[i];

			/* Ignore non-pending locks. */
			if (!IS_PENDING_LOCK(pend_lock->lock_type)) {
				continue;
			}

			if (pend_lock->lock_type == PENDING_READ_LOCK &&
					brl_pending_overlap(plock, pend_lock)) {
				DEBUG(10,("brl_lock_posix: sending unlock message to pid %s\n",
					procid_str_static(&pend_lock->context.pid )));

				messaging_send(msg_ctx, pend_lock->context.pid,
					       MSG_SMB_UNLOCK, &data_blob_null);
			}
		}
	}

	return NT_STATUS_OK;
 fail:
	if (break_oplocks) {
		contend_level2_oplocks_end(br_lck->fsp,
					   LEVEL2_CONTEND_POSIX_BRL);
	}
	return status;
}

NTSTATUS smb_vfs_call_brl_lock_windows(struct vfs_handle_struct *handle,
				       struct byte_range_lock *br_lck,
				       struct lock_struct *plock,
				       bool blocking_lock,
				       struct blocking_lock_record *blr)
{
	VFS_FIND(brl_lock_windows);
	return handle->fns->brl_lock_windows_fn(handle, br_lck, plock,
						blocking_lock, blr);
}

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

NTSTATUS brl_lock(struct messaging_context *msg_ctx,
		struct byte_range_lock *br_lck,
		uint64_t smblctx,
		struct server_id pid,
		br_off start,
		br_off size,
		enum brl_type lock_type,
		enum brl_flavour lock_flav,
		bool blocking_lock,
		uint64_t *psmblctx,
		struct blocking_lock_record *blr)
{
	NTSTATUS ret;
	struct lock_struct lock;

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

#ifdef DEVELOPER
	/* Quieten valgrind on test. */
	ZERO_STRUCT(lock);
#endif

	lock.context.smblctx = smblctx;
	lock.context.pid = pid;
	lock.context.tid = br_lck->fsp->conn->cnum;
	lock.start = start;
	lock.size = size;
	lock.fnum = br_lck->fsp->fnum;
	lock.lock_type = lock_type;
	lock.lock_flav = lock_flav;

	if (lock_flav == WINDOWS_LOCK) {
		ret = SMB_VFS_BRL_LOCK_WINDOWS(br_lck->fsp->conn, br_lck,
		    &lock, blocking_lock, blr);
	} else {
		ret = brl_lock_posix(msg_ctx, br_lck, &lock);
	}

#if ZERO_ZERO
	/* sort the lock list */
	TYPESAFE_QSORT(br_lck->lock_data, (size_t)br_lck->num_locks, lock_compare);
#endif

	/* If we're returning an error, return who blocked us. */
	if (!NT_STATUS_IS_OK(ret) && psmblctx) {
		*psmblctx = lock.context.smblctx;
	}
	return ret;
}

/****************************************************************************
 Unlock a range of bytes - Windows semantics.
****************************************************************************/

bool brl_unlock_windows_default(struct messaging_context *msg_ctx,
			       struct byte_range_lock *br_lck,
			       const struct lock_struct *plock)
{
	unsigned int i, j;
	struct lock_struct *locks = br_lck->lock_data;
	enum brl_type deleted_lock_type = READ_LOCK; /* shut the compiler up.... */

	SMB_ASSERT(plock->lock_type == UNLOCK_LOCK);

#if ZERO_ZERO
	/* Delete write locks by preference... The lock list
	   is sorted in the zero zero case. */

	for (i = 0; i < br_lck->num_locks; i++) {
		struct lock_struct *lock = &locks[i];

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

			/* found it - delete it */
			deleted_lock_type = lock->lock_type;
			break;
		}
	}

	if (i != br_lck->num_locks) {
		/* We found it - don't search again. */
		goto unlock_continue;
	}
#endif

	for (i = 0; i < br_lck->num_locks; i++) {
		struct lock_struct *lock = &locks[i];

		if (IS_PENDING_LOCK(lock->lock_type)) {
			continue;
		}

		/* Only remove our own locks that match in start, size, and flavour. */
		if (brl_same_context(&lock->context, &plock->context) &&
					lock->fnum == plock->fnum &&
					lock->lock_flav == WINDOWS_LOCK &&
					lock->start == plock->start &&
					lock->size == plock->size ) {
			deleted_lock_type = lock->lock_type;
			break;
		}
	}

	if (i == br_lck->num_locks) {
		/* we didn't find it */
		return False;
	}

#if ZERO_ZERO
  unlock_continue:
#endif

	/* Actually delete the lock. */
	if (i < br_lck->num_locks - 1) {
		memmove(&locks[i], &locks[i+1],
			sizeof(*locks)*((br_lck->num_locks-1) - i));
	}

	br_lck->num_locks -= 1;
	br_lck->modified = True;

	/* Unlock the underlying POSIX regions. */
	if(lp_posix_locking(br_lck->fsp->conn->params)) {
		release_posix_lock_windows_flavour(br_lck->fsp,
				plock->start,
				plock->size,
				deleted_lock_type,
				&plock->context,
				locks,
				br_lck->num_locks);
	}

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

		/* Ignore non-pending locks. */
		if (!IS_PENDING_LOCK(pend_lock->lock_type)) {
			continue;
		}

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

			messaging_send(msg_ctx, pend_lock->context.pid,
				       MSG_SMB_UNLOCK, &data_blob_null);
		}
	}

	contend_level2_oplocks_end(br_lck->fsp, LEVEL2_CONTEND_WINDOWS_BRL);
	return True;
}

/****************************************************************************
 Unlock a range of bytes - POSIX semantics.
****************************************************************************/

static bool brl_unlock_posix(struct messaging_context *msg_ctx,
			     struct byte_range_lock *br_lck,
			     struct lock_struct *plock)
{
	unsigned int i, j, count;
	struct lock_struct *tp;
	struct lock_struct *locks = br_lck->lock_data;
	bool overlap_found = False;

	/* No zero-zero locks for POSIX. */
	if (plock->start == 0 && plock->size == 0) {
		return False;
	}

	/* Don't allow 64-bit lock wrap. */
	if (plock->start + plock->size < plock->start ||
			plock->start + plock->size < plock->size) {
		DEBUG(10,("brl_unlock_posix: lock wrap\n"));
		return False;
	}

	/* The worst case scenario here is we have to split an
	   existing POSIX lock range into two, so we need at most
	   1 more entry. */

	tp = talloc_array(br_lck, struct lock_struct, br_lck->num_locks + 1);
	if (!tp) {
		DEBUG(10,("brl_unlock_posix: malloc fail\n"));
		return False;
	}

	count = 0;
	for (i = 0; i < br_lck->num_locks; i++) {
		struct lock_struct *lock = &locks[i];
		unsigned int tmp_count;

		/* Only remove our own locks - ignore fnum. */
		if (IS_PENDING_LOCK(lock->lock_type) ||
				!brl_same_context(&lock->context, &plock->context)) {
			memcpy(&tp[count], lock, sizeof(struct lock_struct));
			count++;
			continue;
		}

		if (lock->lock_flav == WINDOWS_LOCK) {
			/* Do any Windows flavour locks conflict ? */
			if (brl_conflict(lock, plock)) {
				TALLOC_FREE(tp);
				return false;
			}
			/* Just copy the Windows lock into the new array. */
			memcpy(&tp[count], lock, sizeof(struct lock_struct));
			count++;
			continue;
		}

		/* Work out overlaps. */
		tmp_count = brlock_posix_split_merge(&tp[count], lock, plock);

		if (tmp_count == 0) {
			/* plock overlapped the existing lock completely,
			   or replaced it. Don't copy the existing lock. */
			overlap_found = true;
		} else if (tmp_count == 1) {
			/* Either no overlap, (simple copy of existing lock) or
			 * an overlap of an existing lock. */
			/* If the lock changed size, we had an overlap. */
			if (tp[count].size != lock->size) {
				overlap_found = true;
			}
			count += tmp_count;
		} else if (tmp_count == 2) {
			/* We split a lock range in two. */
			overlap_found = true;
			count += tmp_count;

			/* Optimisation... */
			/* We know we're finished here as we can't overlap any
			   more POSIX locks. Copy the rest of the lock array. */

			if (i < br_lck->num_locks - 1) {
				memcpy(&tp[count], &locks[i+1],
					sizeof(*locks)*((br_lck->num_locks-1) - i));
				count += ((br_lck->num_locks-1) - i);
			}
			break;
		}

	}

	if (!overlap_found) {
		/* Just ignore - no change. */
		TALLOC_FREE(tp);
		DEBUG(10,("brl_unlock_posix: No overlap - unlocked.\n"));
		return True;
	}

	/* Unlock any POSIX regions. */
	if(lp_posix_locking(br_lck->fsp->conn->params)) {
		release_posix_lock_posix_flavour(br_lck->fsp,
						plock->start,
						plock->size,
						&plock->context,
						tp,
						count);
	}

	/* Realloc so we don't leak entries per unlock call. */
	if (count) {
		tp = talloc_realloc(br_lck, tp, struct lock_struct, count);
		if (!tp) {
			DEBUG(10,("brl_unlock_posix: realloc fail\n"));
			return False;
		}
	} else {
		/* We deleted the last lock. */
		TALLOC_FREE(tp);
		tp = NULL;
	}

	contend_level2_oplocks_end(br_lck->fsp,
				   LEVEL2_CONTEND_POSIX_BRL);

	br_lck->num_locks = count;
	TALLOC_FREE(br_lck->lock_data);
	locks = tp;
	br_lck->lock_data = tp;
	br_lck->modified = True;

	/* Send unlock messages to any pending waiters that overlap. */

	for (j=0; j < br_lck->num_locks; j++) {
		struct lock_struct *pend_lock = &locks[j];

		/* Ignore non-pending locks. */
		if (!IS_PENDING_LOCK(pend_lock->lock_type)) {
			continue;
		}

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

			messaging_send(msg_ctx, pend_lock->context.pid,
				       MSG_SMB_UNLOCK, &data_blob_null);
		}
	}

	return True;
}

bool smb_vfs_call_brl_unlock_windows(struct vfs_handle_struct *handle,
				     struct messaging_context *msg_ctx,
				     struct byte_range_lock *br_lck,
				     const struct lock_struct *plock)
{
	VFS_FIND(brl_unlock_windows);
	return handle->fns->brl_unlock_windows_fn(handle, msg_ctx, br_lck,
						  plock);
}

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

bool brl_unlock(struct messaging_context *msg_ctx,
		struct byte_range_lock *br_lck,
		uint64_t smblctx,
		struct server_id pid,
		br_off start,
		br_off size,
		enum brl_flavour lock_flav)
{
	struct lock_struct lock;

	lock.context.smblctx = smblctx;
	lock.context.pid = pid;
	lock.context.tid = br_lck->fsp->conn->cnum;
	lock.start = start;
	lock.size = size;
	lock.fnum = br_lck->fsp->fnum;
	lock.lock_type = UNLOCK_LOCK;
	lock.lock_flav = lock_flav;

	if (lock_flav == WINDOWS_LOCK) {
		return SMB_VFS_BRL_UNLOCK_WINDOWS(br_lck->fsp->conn, msg_ctx,
		    br_lck, &lock);
	} else {
		return brl_unlock_posix(msg_ctx, br_lck, &lock);
	}
}

/****************************************************************************
 Test if we could add a lock if we wanted to.
 Returns True if the region required is currently unlocked, False if locked.
****************************************************************************/

bool brl_locktest(struct byte_range_lock *br_lck,
		uint64_t smblctx,
		struct server_id pid,
		br_off start,
		br_off size,
		enum brl_type lock_type,
		enum brl_flavour lock_flav)
{
	bool ret = True;
	unsigned int i;
	struct lock_struct lock;
	const struct lock_struct *locks = br_lck->lock_data;
	files_struct *fsp = br_lck->fsp;

	lock.context.smblctx = smblctx;
	lock.context.pid = pid;
	lock.context.tid = br_lck->fsp->conn->cnum;
	lock.start = start;
	lock.size = size;
	lock.fnum = fsp->fnum;
	lock.lock_type = lock_type;
	lock.lock_flav = lock_flav;

	/* Make sure existing locks don't conflict */
	for (i=0; i < br_lck->num_locks; i++) {
		/*
		 * Our own locks don't conflict.
		 */
		if (brl_conflict_other(&locks[i], &lock)) {
			return False;
		}
	}

	/*
	 * There is no lock held by an SMB daemon, check to
	 * see if there is a POSIX lock from a UNIX or NFS process.
	 * This only conflicts with Windows locks, not POSIX locks.
	 */

	if(lp_posix_locking(fsp->conn->params) && (lock_flav == WINDOWS_LOCK)) {
		ret = is_posix_locked(fsp, &start, &size, &lock_type, WINDOWS_LOCK);

		DEBUG(10,("brl_locktest: posix start=%.0f len=%.0f %s for %s file %s\n",
			(double)start, (double)size, ret ? "locked" : "unlocked",
			fsp_fnum_dbg(fsp), fsp_str_dbg(fsp)));

		/* We need to return the inverse of is_posix_locked. */
		ret = !ret;
        }

	/* no conflicts - we could have added it */
	return ret;
}

/****************************************************************************
 Query for existing locks.
****************************************************************************/

NTSTATUS brl_lockquery(struct byte_range_lock *br_lck,
		uint64_t *psmblctx,
		struct server_id pid,
		br_off *pstart,
		br_off *psize,
		enum brl_type *plock_type,
		enum brl_flavour lock_flav)
{
	unsigned int i;
	struct lock_struct lock;
	const struct lock_struct *locks = br_lck->lock_data;
	files_struct *fsp = br_lck->fsp;

	lock.context.smblctx = *psmblctx;
	lock.context.pid = pid;
	lock.context.tid = br_lck->fsp->conn->cnum;
	lock.start = *pstart;
	lock.size = *psize;
	lock.fnum = fsp->fnum;
	lock.lock_type = *plock_type;
	lock.lock_flav = lock_flav;

	/* Make sure existing locks don't conflict */
	for (i=0; i < br_lck->num_locks; i++) {
		const struct lock_struct *exlock = &locks[i];
		bool conflict = False;

		if (exlock->lock_flav == WINDOWS_LOCK) {
			conflict = brl_conflict(exlock, &lock);
		} else {
			conflict = brl_conflict_posix(exlock, &lock);
		}

		if (conflict) {
			*psmblctx = exlock->context.smblctx;
        		*pstart = exlock->start;
		        *psize = exlock->size;
        		*plock_type = exlock->lock_type;
			return NT_STATUS_LOCK_NOT_GRANTED;
		}
	}

	/*
	 * There is no lock held by an SMB daemon, check to
	 * see if there is a POSIX lock from a UNIX or NFS process.
	 */

	if(lp_posix_locking(fsp->conn->params)) {
		bool ret = is_posix_locked(fsp, pstart, psize, plock_type, POSIX_LOCK);

		DEBUG(10,("brl_lockquery: posix start=%.0f len=%.0f %s for %s file %s\n",
			(double)*pstart, (double)*psize, ret ? "locked" : "unlocked",
			fsp_fnum_dbg(fsp), fsp_str_dbg(fsp)));

		if (ret) {
			/* Hmmm. No clue what to set smblctx to - use -1. */
			*psmblctx = 0xFFFFFFFFFFFFFFFFLL;
			return NT_STATUS_LOCK_NOT_GRANTED;
		}
        }

	return NT_STATUS_OK;
}


bool smb_vfs_call_brl_cancel_windows(struct vfs_handle_struct *handle,
				     struct byte_range_lock *br_lck,
				     struct lock_struct *plock,
				     struct blocking_lock_record *blr)
{
	VFS_FIND(brl_cancel_windows);
	return handle->fns->brl_cancel_windows_fn(handle, br_lck, plock, blr);
}

/****************************************************************************
 Remove a particular pending lock.
****************************************************************************/
bool brl_lock_cancel(struct byte_range_lock *br_lck,
		uint64_t smblctx,
		struct server_id pid,
		br_off start,
		br_off size,
		enum brl_flavour lock_flav,
		struct blocking_lock_record *blr)
{
	bool ret;
	struct lock_struct lock;

	lock.context.smblctx = smblctx;
	lock.context.pid = pid;
	lock.context.tid = br_lck->fsp->conn->cnum;
	lock.start = start;
	lock.size = size;
	lock.fnum = br_lck->fsp->fnum;
	lock.lock_flav = lock_flav;
	/* lock.lock_type doesn't matter */

	if (lock_flav == WINDOWS_LOCK) {
		ret = SMB_VFS_BRL_CANCEL_WINDOWS(br_lck->fsp->conn, br_lck,
		    &lock, blr);
	} else {
		ret = brl_lock_cancel_default(br_lck, &lock);
	}

	return ret;
}

bool brl_lock_cancel_default(struct byte_range_lock *br_lck,
		struct lock_struct *plock)
{
	unsigned int i;
	struct lock_struct *locks = br_lck->lock_data;

	SMB_ASSERT(plock);

	for (i = 0; i < br_lck->num_locks; i++) {
		struct lock_struct *lock = &locks[i];

		/* For pending locks we *always* care about the fnum. */
		if (brl_same_context(&lock->context, &plock->context) &&
				lock->fnum == plock->fnum &&
				IS_PENDING_LOCK(lock->lock_type) &&
				lock->lock_flav == plock->lock_flav &&
				lock->start == plock->start &&
				lock->size == plock->size) {
			break;
		}
	}

	if (i == br_lck->num_locks) {
		/* Didn't find it. */
		return False;
	}

	if (i < br_lck->num_locks - 1) {
		/* Found this particular pending lock - delete it */
		memmove(&locks[i], &locks[i+1],
			sizeof(*locks)*((br_lck->num_locks-1) - i));
	}

	br_lck->num_locks -= 1;
	br_lck->modified = True;
	return True;
}

/****************************************************************************
 Remove any locks associated with a open file.
 We return True if this process owns any other Windows locks on this
 fd and so we should not immediately close the fd.
****************************************************************************/

void brl_close_fnum(struct messaging_context *msg_ctx,
		    struct byte_range_lock *br_lck)
{
	files_struct *fsp = br_lck->fsp;
	uint32_t tid = fsp->conn->cnum;
	uint64_t fnum = fsp->fnum;
	unsigned int i;
	struct lock_struct *locks = br_lck->lock_data;
	struct server_id pid = messaging_server_id(fsp->conn->sconn->msg_ctx);
	struct lock_struct *locks_copy;
	unsigned int num_locks_copy;

	/* Copy the current lock array. */
	if (br_lck->num_locks) {
		locks_copy = (struct lock_struct *)talloc_memdup(br_lck, locks, br_lck->num_locks * sizeof(struct lock_struct));
		if (!locks_copy) {
			smb_panic("brl_close_fnum: talloc failed");
			}
	} else {
		locks_copy = NULL;
	}

	num_locks_copy = br_lck->num_locks;

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

		if (lock->context.tid == tid && serverid_equal(&lock->context.pid, &pid) &&
				(lock->fnum == fnum)) {
			brl_unlock(msg_ctx,
				br_lck,
				lock->context.smblctx,
				pid,
				lock->start,
				lock->size,
				lock->lock_flav);
		}
	}
}

bool brl_mark_disconnected(struct files_struct *fsp)
{
	uint32_t tid = fsp->conn->cnum;
	uint64_t smblctx = fsp->op->global->open_persistent_id;
	uint64_t fnum = fsp->fnum;
	unsigned int i;
	struct server_id self = messaging_server_id(fsp->conn->sconn->msg_ctx);
	struct byte_range_lock *br_lck = NULL;

	if (!fsp->op->global->durable) {
		return false;
	}

	if (fsp->current_lock_count == 0) {
		return true;
	}

	br_lck = brl_get_locks(talloc_tos(), fsp);
	if (br_lck == NULL) {
		return false;
	}

	for (i=0; i < br_lck->num_locks; i++) {
		struct lock_struct *lock = &br_lck->lock_data[i];

		/*
		 * as this is a durable handle, we only expect locks
		 * of the current file handle!
		 */

		if (lock->context.smblctx != smblctx) {
			TALLOC_FREE(br_lck);
			return false;
		}

		if (lock->context.tid != tid) {
			TALLOC_FREE(br_lck);
			return false;
		}

		if (!serverid_equal(&lock->context.pid, &self)) {
			TALLOC_FREE(br_lck);
			return false;
		}

		if (lock->fnum != fnum) {
			TALLOC_FREE(br_lck);
			return false;
		}

		server_id_set_disconnected(&lock->context.pid);
		lock->context.tid = TID_FIELD_INVALID;
		lock->fnum = FNUM_FIELD_INVALID;
	}

	br_lck->modified = true;
	TALLOC_FREE(br_lck);
	return true;
}

bool brl_reconnect_disconnected(struct files_struct *fsp)
{
	uint32_t tid = fsp->conn->cnum;
	uint64_t smblctx = fsp->op->global->open_persistent_id;
	uint64_t fnum = fsp->fnum;
	unsigned int i;
	struct server_id self = messaging_server_id(fsp->conn->sconn->msg_ctx);
	struct byte_range_lock *br_lck = NULL;

	if (!fsp->op->global->durable) {
		return false;
	}

	/*
	 * When reconnecting, we do not want to validate the brlock entries
	 * and thereby remove our own (disconnected) entries but reactivate
	 * them instead.
	 */
	fsp->lockdb_clean = true;

	br_lck = brl_get_locks(talloc_tos(), fsp);
	if (br_lck == NULL) {
		return false;
	}

	if (br_lck->num_locks == 0) {
		TALLOC_FREE(br_lck);
		return true;
	}

	for (i=0; i < br_lck->num_locks; i++) {
		struct lock_struct *lock = &br_lck->lock_data[i];

		/*
		 * as this is a durable handle we only expect locks
		 * of the current file handle!
		 */

		if (lock->context.smblctx != smblctx) {
			TALLOC_FREE(br_lck);
			return false;
		}

		if (lock->context.tid != TID_FIELD_INVALID) {
			TALLOC_FREE(br_lck);
			return false;
		}

		if (!server_id_is_disconnected(&lock->context.pid)) {
			TALLOC_FREE(br_lck);
			return false;
		}

		if (lock->fnum != FNUM_FIELD_INVALID) {
			TALLOC_FREE(br_lck);
			return false;
		}

		lock->context.pid = self;
		lock->context.tid = tid;
		lock->fnum = fnum;
	}

	fsp->current_lock_count = br_lck->num_locks;
	br_lck->modified = true;
	TALLOC_FREE(br_lck);
	return true;
}

/****************************************************************************
 Ensure this set of lock entries is valid.
****************************************************************************/
static bool validate_lock_entries(TALLOC_CTX *mem_ctx,
				  unsigned int *pnum_entries, struct lock_struct **pplocks,
				  bool keep_disconnected)
{
	unsigned int i;
	unsigned int num_valid_entries = 0;
	struct lock_struct *locks = *pplocks;
	TALLOC_CTX *frame = talloc_stackframe();
	struct server_id *ids;
	bool *exists;

	ids = talloc_array(frame, struct server_id, *pnum_entries);
	if (ids == NULL) {
		DEBUG(0, ("validate_lock_entries: "
			  "talloc_array(struct server_id, %u) failed\n",
			  *pnum_entries));
		talloc_free(frame);
		return false;
	}

	exists = talloc_array(frame, bool, *pnum_entries);
	if (exists == NULL) {
		DEBUG(0, ("validate_lock_entries: "
			  "talloc_array(bool, %u) failed\n",
			  *pnum_entries));
		talloc_free(frame);
		return false;
	}

	for (i = 0; i < *pnum_entries; i++) {
		ids[i] = locks[i].context.pid;
	}

	if (!serverids_exist(ids, *pnum_entries, exists)) {
		DEBUG(3, ("validate_lock_entries: serverids_exists failed\n"));
		talloc_free(frame);
		return false;
	}

	for (i = 0; i < *pnum_entries; i++) {
		if (exists[i]) {
			num_valid_entries++;
			continue;
		}

		if (keep_disconnected &&
		    server_id_is_disconnected(&ids[i]))
		{
			num_valid_entries++;
			continue;
		}

		/* This process no longer exists - mark this
		   entry as invalid by zeroing it. */
		ZERO_STRUCTP(&locks[i]);
	}
	TALLOC_FREE(frame);

	if (num_valid_entries != *pnum_entries) {
		struct lock_struct *new_lock_data = NULL;

		if (num_valid_entries) {
			new_lock_data = talloc_array(
				mem_ctx, struct lock_struct,
				num_valid_entries);
			if (!new_lock_data) {
				DEBUG(3, ("malloc fail\n"));
				return False;
			}

			num_valid_entries = 0;
			for (i = 0; i < *pnum_entries; i++) {
				struct lock_struct *lock_data = &locks[i];
				if (lock_data->context.smblctx &&
						lock_data->context.tid) {
					/* Valid (nonzero) entry - copy it. */
					memcpy(&new_lock_data[num_valid_entries],
						lock_data, sizeof(struct lock_struct));
					num_valid_entries++;
				}
			}
		}

		TALLOC_FREE(*pplocks);
		*pplocks = new_lock_data;
		*pnum_entries = num_valid_entries;
	}

	return True;
}

struct brl_forall_cb {
	void (*fn)(struct file_id id, struct server_id pid,
		   enum brl_type lock_type,
		   enum brl_flavour lock_flav,
		   br_off start, br_off size,
		   void *private_data);
	void *private_data;
};

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

static int brl_traverse_fn(struct db_record *rec, void *state)
{
	struct brl_forall_cb *cb = (struct brl_forall_cb *)state;
	struct lock_struct *locks;
	struct file_id *key;
	unsigned int i;
	unsigned int num_locks = 0;
	unsigned int orig_num_locks = 0;
	TDB_DATA dbkey;
	TDB_DATA value;

	dbkey = dbwrap_record_get_key(rec);
	value = dbwrap_record_get_value(rec);

	/* In a traverse function we must make a copy of
	   dbuf before modifying it. */

	locks = (struct lock_struct *)talloc_memdup(
		talloc_tos(), value.dptr, value.dsize);
	if (!locks) {
		return -1; /* Terminate traversal. */
	}

	key = (struct file_id *)dbkey.dptr;
	orig_num_locks = num_locks = value.dsize/sizeof(*locks);

	/* Ensure the lock db is clean of entries from invalid processes. */

	if (!validate_lock_entries(talloc_tos(), &num_locks, &locks, true)) {
		TALLOC_FREE(locks);
		return -1; /* Terminate traversal */
	}

	if (orig_num_locks != num_locks) {
		if (num_locks) {
			TDB_DATA data;
			data.dptr = (uint8_t *)locks;
			data.dsize = num_locks*sizeof(struct lock_struct);
			dbwrap_record_store(rec, data, TDB_REPLACE);
		} else {
			dbwrap_record_delete(rec);
		}
	}

	if (cb->fn) {
		for ( i=0; i<num_locks; i++) {
			cb->fn(*key,
				locks[i].context.pid,
				locks[i].lock_type,
				locks[i].lock_flav,
				locks[i].start,
				locks[i].size,
				cb->private_data);
		}
	}

	TALLOC_FREE(locks);
	return 0;
}

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

int brl_forall(void (*fn)(struct file_id id, struct server_id pid,
			  enum brl_type lock_type,
			  enum brl_flavour lock_flav,
			  br_off start, br_off size,
			  void *private_data),
	       void *private_data)
{
	struct brl_forall_cb cb;
	NTSTATUS status;
	int count = 0;

	if (!brlock_db) {
		return 0;
	}
	cb.fn = fn;
	cb.private_data = private_data;
	status = dbwrap_traverse(brlock_db, brl_traverse_fn, &cb, &count);

	if (!NT_STATUS_IS_OK(status)) {
		return -1;
	} else {
		return count;
	}
}

/*******************************************************************
 Store a potentially modified set of byte range lock data back into
 the database.
 Unlock the record.
********************************************************************/

static void byte_range_lock_flush(struct byte_range_lock *br_lck)
{
	size_t data_len;
	if (!br_lck->modified) {
		DEBUG(10, ("br_lck not modified\n"));
		goto done;
	}

	data_len = br_lck->num_locks * sizeof(struct lock_struct);

	if (br_lck->have_read_oplocks) {
		data_len += 1;
	}

	DEBUG(10, ("data_len=%d\n", (int)data_len));

	if (data_len == 0) {
		/* No locks - delete this entry. */
		NTSTATUS status = dbwrap_record_delete(br_lck->record);
		if (!NT_STATUS_IS_OK(status)) {
			DEBUG(0, ("delete_rec returned %s\n",
				  nt_errstr(status)));
			smb_panic("Could not delete byte range lock entry");
		}
	} else {
		TDB_DATA data;
		NTSTATUS status;

		data.dsize = data_len;
		data.dptr = talloc_array(talloc_tos(), uint8_t, data_len);
		SMB_ASSERT(data.dptr != NULL);

		memcpy(data.dptr, br_lck->lock_data,
		       br_lck->num_locks * sizeof(struct lock_struct));

		if (br_lck->have_read_oplocks) {
			data.dptr[data_len-1] = 1;
		}

		status = dbwrap_record_store(br_lck->record, data, TDB_REPLACE);
		TALLOC_FREE(data.dptr);
		if (!NT_STATUS_IS_OK(status)) {
			DEBUG(0, ("store returned %s\n", nt_errstr(status)));
			smb_panic("Could not store byte range mode entry");
		}
	}

	DEBUG(10, ("seqnum=%d\n", dbwrap_get_seqnum(brlock_db)));

 done:
	br_lck->modified = false;
	TALLOC_FREE(br_lck->record);
}

static int byte_range_lock_destructor(struct byte_range_lock *br_lck)
{
	byte_range_lock_flush(br_lck);
	return 0;
}

/*******************************************************************
 Fetch a set of byte range lock data from the database.
 Leave the record locked.
 TALLOC_FREE(brl) will release the lock in the destructor.
********************************************************************/

struct byte_range_lock *brl_get_locks(TALLOC_CTX *mem_ctx, files_struct *fsp)
{
	TDB_DATA key, data;
	struct byte_range_lock *br_lck = talloc(mem_ctx, struct byte_range_lock);

	if (br_lck == NULL) {
		return NULL;
	}

	br_lck->fsp = fsp;
	br_lck->num_locks = 0;
	br_lck->have_read_oplocks = false;
	br_lck->modified = False;

	key.dptr = (uint8 *)&fsp->file_id;
	key.dsize = sizeof(struct file_id);

	br_lck->record = dbwrap_fetch_locked(brlock_db, br_lck, key);

	if (br_lck->record == NULL) {
		DEBUG(3, ("Could not lock byte range lock entry\n"));
		TALLOC_FREE(br_lck);
		return NULL;
	}

	data = dbwrap_record_get_value(br_lck->record);

	br_lck->lock_data = NULL;

	talloc_set_destructor(br_lck, byte_range_lock_destructor);

	br_lck->num_locks = data.dsize / sizeof(struct lock_struct);

	if (br_lck->num_locks != 0) {
		br_lck->lock_data = talloc_array(
			br_lck, struct lock_struct, br_lck->num_locks);
		if (br_lck->lock_data == NULL) {
			DEBUG(0, ("malloc failed\n"));
			TALLOC_FREE(br_lck);
			return NULL;
		}

		memcpy(br_lck->lock_data, data.dptr,
		       talloc_get_size(br_lck->lock_data));
	}

	DEBUG(10, ("data.dsize=%d\n", (int)data.dsize));

	if ((data.dsize % sizeof(struct lock_struct)) == 1) {
		br_lck->have_read_oplocks = (data.dptr[data.dsize-1] == 1);
	}

	if (!fsp->lockdb_clean) {
		int orig_num_locks = br_lck->num_locks;

		/*
		 * This is the first time we access the byte range lock
		 * record with this fsp. Go through and ensure all entries
		 * are valid - remove any that don't.
		 * This makes the lockdb self cleaning at low cost.
		 *
		 * Note: Disconnected entries belong to disconnected
		 * durable handles. So at this point, we have a new
		 * handle on the file and the disconnected durable has
		 * already been closed (we are not a durable reconnect).
		 * So we need to clean the disconnected brl entry.
		 */

		if (!validate_lock_entries(br_lck, &br_lck->num_locks,
					   &br_lck->lock_data, false)) {
			TALLOC_FREE(br_lck);
			return NULL;
		}

		/* Ensure invalid locks are cleaned up in the destructor. */
		if (orig_num_locks != br_lck->num_locks) {
			br_lck->modified = True;
		}

		/* Mark the lockdb as "clean" as seen from this open file. */
		fsp->lockdb_clean = True;
	}

	if (DEBUGLEVEL >= 10) {
		unsigned int i;
		struct lock_struct *locks = br_lck->lock_data;
		DEBUG(10,("brl_get_locks_internal: %u current locks on file_id %s\n",
			br_lck->num_locks,
			  file_id_string_tos(&fsp->file_id)));
		for( i = 0; i < br_lck->num_locks; i++) {
			print_lock_struct(i, &locks[i]);
		}
	}

	return br_lck;
}

struct brl_get_locks_readonly_state {
	TALLOC_CTX *mem_ctx;
	struct byte_range_lock **br_lock;
};

static void brl_get_locks_readonly_parser(TDB_DATA key, TDB_DATA data,
					  void *private_data)
{
	struct brl_get_locks_readonly_state *state =
		(struct brl_get_locks_readonly_state *)private_data;
	struct byte_range_lock *br_lock;

	br_lock = talloc_pooled_object(
		state->mem_ctx, struct byte_range_lock, 1, data.dsize);
	if (br_lock == NULL) {
		*state->br_lock = NULL;
		return;
	}
	br_lock->lock_data = (struct lock_struct *)talloc_memdup(
		br_lock, data.dptr, data.dsize);
	br_lock->num_locks = data.dsize / sizeof(struct lock_struct);

	if ((data.dsize % sizeof(struct lock_struct)) == 1) {
		br_lock->have_read_oplocks = (data.dptr[data.dsize-1] == 1);
	}

	DEBUG(10, ("Got %d bytes, have_read_oplocks: %s\n", (int)data.dsize,
		   br_lock->have_read_oplocks ? "true" : "false"));

	*state->br_lock = br_lock;
}

struct byte_range_lock *brl_get_locks_readonly(files_struct *fsp)
{
	struct byte_range_lock *br_lock = NULL;
	struct byte_range_lock *rw = NULL;

	DEBUG(10, ("seqnum=%d, fsp->brlock_seqnum=%d\n",
		   dbwrap_get_seqnum(brlock_db), fsp->brlock_seqnum));

	if ((fsp->brlock_rec != NULL)
	    && (dbwrap_get_seqnum(brlock_db) == fsp->brlock_seqnum)) {
		/*
		 * We have cached the brlock_rec and the database did not
		 * change.
		 */
		return fsp->brlock_rec;
	}

	if (!fsp->lockdb_clean) {
		/*
		 * Fetch the record in R/W mode to give validate_lock_entries
		 * a chance to kick in once.
		 */
		rw = brl_get_locks(talloc_tos(), fsp);
		if (rw == NULL) {
			return NULL;
		}
		fsp->lockdb_clean = true;
	}

	if (rw != NULL) {
		size_t lock_data_size;

		/*
		 * Make a copy of the already retrieved and sanitized rw record
		 */
		lock_data_size = rw->num_locks * sizeof(struct lock_struct);
		br_lock = talloc_pooled_object(
			fsp, struct byte_range_lock, 1, lock_data_size);
		if (br_lock == NULL) {
			goto fail;
		}
		br_lock->have_read_oplocks = rw->have_read_oplocks;
		br_lock->num_locks = rw->num_locks;
		br_lock->lock_data = (struct lock_struct *)talloc_memdup(
			br_lock, rw->lock_data, lock_data_size);
	} else {
		struct brl_get_locks_readonly_state state;
		NTSTATUS status;

		/*
		 * Parse the record fresh from the database
		 */

		state.mem_ctx = fsp;
		state.br_lock = &br_lock;

		status = dbwrap_parse_record(
			brlock_db,
			make_tdb_data((uint8_t *)&fsp->file_id,
				      sizeof(fsp->file_id)),
			brl_get_locks_readonly_parser, &state);
		if (!NT_STATUS_IS_OK(status)) {
			DEBUG(3, ("Could not parse byte range lock record: "
				  "%s\n", nt_errstr(status)));
			goto fail;
		}
		if (br_lock == NULL) {
			goto fail;
		}
	}

	br_lock->fsp = fsp;
	br_lock->modified = false;
	br_lock->record = NULL;

	if (lp_clustering()) {
		/*
		 * In the cluster case we can't cache the brlock struct
		 * because dbwrap_get_seqnum does not work reliably over
		 * ctdb. Thus we have to throw away the brlock struct soon.
		 */
		talloc_steal(talloc_tos(), br_lock);
	} else {
		/*
		 * Cache the brlock struct, invalidated when the dbwrap_seqnum
		 * changes. See beginning of this routine.
		 */
		TALLOC_FREE(fsp->brlock_rec);
		fsp->brlock_rec = br_lock;
		fsp->brlock_seqnum = dbwrap_get_seqnum(brlock_db);
	}

fail:
	TALLOC_FREE(rw);
	return br_lock;
}

struct brl_revalidate_state {
	ssize_t array_size;
	uint32 num_pids;
	struct server_id *pids;
};

/*
 * Collect PIDs of all processes with pending entries
 */

static void brl_revalidate_collect(struct file_id id, struct server_id pid,
				   enum brl_type lock_type,
				   enum brl_flavour lock_flav,
				   br_off start, br_off size,
				   void *private_data)
{
	struct brl_revalidate_state *state =
		(struct brl_revalidate_state *)private_data;

	if (!IS_PENDING_LOCK(lock_type)) {
		return;
	}

	add_to_large_array(state, sizeof(pid), (void *)&pid,
			   &state->pids, &state->num_pids,
			   &state->array_size);
}

/*
 * qsort callback to sort the processes
 */

static int compare_procids(const void *p1, const void *p2)
{
	const struct server_id *i1 = (const struct server_id *)p1;
	const struct server_id *i2 = (const struct server_id *)p2;

	if (i1->pid < i2->pid) return -1;
	if (i2->pid > i2->pid) return 1;
	return 0;
}

/*
 * Send a MSG_SMB_UNLOCK message to all processes with pending byte range
 * locks so that they retry. Mainly used in the cluster code after a node has
 * died.
 *
 * Done in two steps to avoid double-sends: First we collect all entries in an
 * array, then qsort that array and only send to non-dupes.
 */

void brl_revalidate(struct messaging_context *msg_ctx,
		    void *private_data,
		    uint32_t msg_type,
		    struct server_id server_id,
		    DATA_BLOB *data)
{
	struct brl_revalidate_state *state;
	uint32 i;
	struct server_id last_pid;

	if (!(state = talloc_zero(NULL, struct brl_revalidate_state))) {
		DEBUG(0, ("talloc failed\n"));
		return;
	}

	brl_forall(brl_revalidate_collect, state);

	if (state->array_size == -1) {
		DEBUG(0, ("talloc failed\n"));
		goto done;
	}

	if (state->num_pids == 0) {
		goto done;
	}

	TYPESAFE_QSORT(state->pids, state->num_pids, compare_procids);

	ZERO_STRUCT(last_pid);

	for (i=0; i<state->num_pids; i++) {
		if (serverid_equal(&last_pid, &state->pids[i])) {
			/*
			 * We've seen that one already
			 */
			continue;
		}

		messaging_send(msg_ctx, state->pids[i], MSG_SMB_UNLOCK,
			       &data_blob_null);
		last_pid = state->pids[i];
	}

 done:
	TALLOC_FREE(state);
	return;
}

bool brl_cleanup_disconnected(struct file_id fid, uint64_t open_persistent_id)
{
	bool ret = false;
	TALLOC_CTX *frame = talloc_stackframe();
	TDB_DATA key, val;
	struct db_record *rec;
	struct lock_struct *lock;
	unsigned n, num;
	NTSTATUS status;

	key = make_tdb_data((void*)&fid, sizeof(fid));

	rec = dbwrap_fetch_locked(brlock_db, frame, key);
	if (rec == NULL) {
		DEBUG(5, ("brl_cleanup_disconnected: failed to fetch record "
			  "for file %s\n", file_id_string(frame, &fid)));
		goto done;
	}

	val = dbwrap_record_get_value(rec);
	lock = (struct lock_struct*)val.dptr;
	num = val.dsize / sizeof(struct lock_struct);
	if (lock == NULL) {
		DEBUG(10, ("brl_cleanup_disconnected: no byte range locks for "
			   "file %s\n", file_id_string(frame, &fid)));
		ret = true;
		goto done;
	}

	for (n=0; n<num; n++) {
		struct lock_context *ctx = &lock[n].context;

		if (!server_id_is_disconnected(&ctx->pid)) {
			DEBUG(5, ("brl_cleanup_disconnected: byte range lock "
				  "%s used by server %s, do not cleanup\n",
				  file_id_string(frame, &fid),
				  server_id_str(frame, &ctx->pid)));
			goto done;
		}

		if (ctx->smblctx != open_persistent_id)	{
			DEBUG(5, ("brl_cleanup_disconnected: byte range lock "
				  "%s expected smblctx %llu but found %llu"
				  ", do not cleanup\n",
				  file_id_string(frame, &fid),
				  (unsigned long long)open_persistent_id,
				  (unsigned long long)ctx->smblctx));
			goto done;
		}
	}

	status = dbwrap_record_delete(rec);
	if (!NT_STATUS_IS_OK(status)) {
		DEBUG(5, ("brl_cleanup_disconnected: failed to delete record "
			  "for file %s from %s, open %llu: %s\n",
			  file_id_string(frame, &fid), dbwrap_name(brlock_db),
			  (unsigned long long)open_persistent_id,
			  nt_errstr(status)));
		goto done;
	}

	DEBUG(10, ("brl_cleanup_disconnected: "
		   "file %s cleaned up %u entries from open %llu\n",
		   file_id_string(frame, &fid), num,
		   (unsigned long long)open_persistent_id));

	ret = true;
done:
	talloc_free(frame);
	return ret;
}