aboutsummaryrefslogtreecommitdiffstats
path: root/fs/ocfs2/journal.c
blob: 54c16b66327e7bcf4ac7d057963bace3f34b9273 (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
/* -*- mode: c; c-basic-offset: 8; -*-
 * vim: noexpandtab sw=8 ts=8 sts=0:
 *
 * journal.c
 *
 * Defines functions of journalling api
 *
 * Copyright (C) 2003, 2004 Oracle.  All rights reserved.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public
 * License as published by the Free Software Foundation; either
 * version 2 of the License, or (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public
 * License along with this program; if not, write to the
 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 * Boston, MA 021110-1307, USA.
 */

#include <linux/fs.h>
#include <linux/types.h>
#include <linux/slab.h>
#include <linux/highmem.h>
#include <linux/kthread.h>
#include <linux/time.h>
#include <linux/random.h>

#define MLOG_MASK_PREFIX ML_JOURNAL
#include <cluster/masklog.h>

#include "ocfs2.h"

#include "alloc.h"
#include "blockcheck.h"
#include "dir.h"
#include "dlmglue.h"
#include "extent_map.h"
#include "heartbeat.h"
#include "inode.h"
#include "journal.h"
#include "localalloc.h"
#include "slot_map.h"
#include "super.h"
#include "sysfile.h"
#include "uptodate.h"
#include "quota.h"

#include "buffer_head_io.h"

DEFINE_SPINLOCK(trans_inc_lock);

#define ORPHAN_SCAN_SCHEDULE_TIMEOUT 300000

static int ocfs2_force_read_journal(struct inode *inode);
static int ocfs2_recover_node(struct ocfs2_super *osb,
			      int node_num, int slot_num);
static int __ocfs2_recovery_thread(void *arg);
static int ocfs2_commit_cache(struct ocfs2_super *osb);
static int __ocfs2_wait_on_mount(struct ocfs2_super *osb, int quota);
static int ocfs2_journal_toggle_dirty(struct ocfs2_super *osb,
				      int dirty, int replayed);
static int ocfs2_trylock_journal(struct ocfs2_super *osb,
				 int slot_num);
static int ocfs2_recover_orphans(struct ocfs2_super *osb,
				 int slot);
static int ocfs2_commit_thread(void *arg);
static void ocfs2_queue_recovery_completion(struct ocfs2_journal *journal,
					    int slot_num,
					    struct ocfs2_dinode *la_dinode,
					    struct ocfs2_dinode *tl_dinode,
					    struct ocfs2_quota_recovery *qrec);

static inline int ocfs2_wait_on_mount(struct ocfs2_super *osb)
{
	return __ocfs2_wait_on_mount(osb, 0);
}

static inline int ocfs2_wait_on_quotas(struct ocfs2_super *osb)
{
	return __ocfs2_wait_on_mount(osb, 1);
}

/*
 * This replay_map is to track online/offline slots, so we could recover
 * offline slots during recovery and mount
 */

enum ocfs2_replay_state {
	REPLAY_UNNEEDED = 0,	/* Replay is not needed, so ignore this map */
	REPLAY_NEEDED, 		/* Replay slots marked in rm_replay_slots */
	REPLAY_DONE 		/* Replay was already queued */
};

struct ocfs2_replay_map {
	unsigned int rm_slots;
	enum ocfs2_replay_state rm_state;
	unsigned char rm_replay_slots[0];
};

void ocfs2_replay_map_set_state(struct ocfs2_super *osb, int state)
{
	if (!osb->replay_map)
		return;

	/* If we've already queued the replay, we don't have any more to do */
	if (osb->replay_map->rm_state == REPLAY_DONE)
		return;

	osb->replay_map->rm_state = state;
}

int ocfs2_compute_replay_slots(struct ocfs2_super *osb)
{
	struct ocfs2_replay_map *replay_map;
	int i, node_num;

	/* If replay map is already set, we don't do it again */
	if (osb->replay_map)
		return 0;

	replay_map = kzalloc(sizeof(struct ocfs2_replay_map) +
			     (osb->max_slots * sizeof(char)), GFP_KERNEL);

	if (!replay_map) {
		mlog_errno(-ENOMEM);
		return -ENOMEM;
	}

	spin_lock(&osb->osb_lock);

	replay_map->rm_slots = osb->max_slots;
	replay_map->rm_state = REPLAY_UNNEEDED;

	/* set rm_replay_slots for offline slot(s) */
	for (i = 0; i < replay_map->rm_slots; i++) {
		if (ocfs2_slot_to_node_num_locked(osb, i, &node_num) == -ENOENT)
			replay_map->rm_replay_slots[i] = 1;
	}

	osb->replay_map = replay_map;
	spin_unlock(&osb->osb_lock);
	return 0;
}

void ocfs2_queue_replay_slots(struct ocfs2_super *osb)
{
	struct ocfs2_replay_map *replay_map = osb->replay_map;
	int i;

	if (!replay_map)
		return;

	if (replay_map->rm_state != REPLAY_NEEDED)
		return;

	for (i = 0; i < replay_map->rm_slots; i++)
		if (replay_map->rm_replay_slots[i])
			ocfs2_queue_recovery_completion(osb->journal, i, NULL,
							NULL, NULL);
	replay_map->rm_state = REPLAY_DONE;
}

void ocfs2_free_replay_slots(struct ocfs2_super *osb)
{
	struct ocfs2_replay_map *replay_map = osb->replay_map;

	if (!osb->replay_map)
		return;

	kfree(replay_map);
	osb->replay_map = NULL;
}

int ocfs2_recovery_init(struct ocfs2_super *osb)
{
	struct ocfs2_recovery_map *rm;

	mutex_init(&osb->recovery_lock);
	osb->disable_recovery = 0;
	osb->recovery_thread_task = NULL;
	init_waitqueue_head(&osb->recovery_event);

	rm = kzalloc(sizeof(struct ocfs2_recovery_map) +
		     osb->max_slots * sizeof(unsigned int),
		     GFP_KERNEL);
	if (!rm) {
		mlog_errno(-ENOMEM);
		return -ENOMEM;
	}

	rm->rm_entries = (unsigned int *)((char *)rm +
					  sizeof(struct ocfs2_recovery_map));
	osb->recovery_map = rm;

	return 0;
}

/* we can't grab the goofy sem lock from inside wait_event, so we use
 * memory barriers to make sure that we'll see the null task before
 * being woken up */
static int ocfs2_recovery_thread_running(struct ocfs2_super *osb)
{
	mb();
	return osb->recovery_thread_task != NULL;
}

void ocfs2_recovery_exit(struct ocfs2_super *osb)
{
	struct ocfs2_recovery_map *rm;

	/* disable any new recovery threads and wait for any currently
	 * running ones to exit. Do this before setting the vol_state. */
	mutex_lock(&osb->recovery_lock);
	osb->disable_recovery = 1;
	mutex_unlock(&osb->recovery_lock);
	wait_event(osb->recovery_event, !ocfs2_recovery_thread_running(osb));

	/* At this point, we know that no more recovery threads can be
	 * launched, so wait for any recovery completion work to
	 * complete. */
	flush_workqueue(ocfs2_wq);

	/*
	 * Now that recovery is shut down, and the osb is about to be
	 * freed,  the osb_lock is not taken here.
	 */
	rm = osb->recovery_map;
	/* XXX: Should we bug if there are dirty entries? */

	kfree(rm);
}

static int __ocfs2_recovery_map_test(struct ocfs2_super *osb,
				     unsigned int node_num)
{
	int i;
	struct ocfs2_recovery_map *rm = osb->recovery_map;

	assert_spin_locked(&osb->osb_lock);

	for (i = 0; i < rm->rm_used; i++) {
		if (rm->rm_entries[i] == node_num)
			return 1;
	}

	return 0;
}

/* Behaves like test-and-set.  Returns the previous value */
static int ocfs2_recovery_map_set(struct ocfs2_super *osb,
				  unsigned int node_num)
{
	struct ocfs2_recovery_map *rm = osb->recovery_map;

	spin_lock(&osb->osb_lock);
	if (__ocfs2_recovery_map_test(osb, node_num)) {
		spin_unlock(&osb->osb_lock);
		return 1;
	}

	/* XXX: Can this be exploited? Not from o2dlm... */
	BUG_ON(rm->rm_used >= osb->max_slots);

	rm->rm_entries[rm->rm_used] = node_num;
	rm->rm_used++;
	spin_unlock(&osb->osb_lock);

	return 0;
}

static void ocfs2_recovery_map_clear(struct ocfs2_super *osb,
				     unsigned int node_num)
{
	int i;
	struct ocfs2_recovery_map *rm = osb->recovery_map;

	spin_lock(&osb->osb_lock);

	for (i = 0; i < rm->rm_used; i++) {
		if (rm->rm_entries[i] == node_num)
			break;
	}

	if (i < rm->rm_used) {
		/* XXX: be careful with the pointer math */
		memmove(&(rm->rm_entries[i]), &(rm->rm_entries[i + 1]),
			(rm->rm_used - i - 1) * sizeof(unsigned int));
		rm->rm_used--;
	}

	spin_unlock(&osb->osb_lock);
}

static int ocfs2_commit_cache(struct ocfs2_super *osb)
{
	int status = 0;
	unsigned int flushed;
	unsigned long old_id;
	struct ocfs2_journal *journal = NULL;

	mlog_entry_void();

	journal = osb->journal;

	/* Flush all pending commits and checkpoint the journal. */
	down_write(&journal->j_trans_barrier);

	if (atomic_read(&journal->j_num_trans) == 0) {
		up_write(&journal->j_trans_barrier);
		mlog(0, "No transactions for me to flush!\n");
		goto finally;
	}

	jbd2_journal_lock_updates(journal->j_journal);
	status = jbd2_journal_flush(journal->j_journal);
	jbd2_journal_unlock_updates(journal->j_journal);
	if (status < 0) {
		up_write(&journal->j_trans_barrier);
		mlog_errno(status);
		goto finally;
	}

	old_id = ocfs2_inc_trans_id(journal);

	flushed = atomic_read(&journal->j_num_trans);
	atomic_set(&journal->j_num_trans, 0);
	up_write(&journal->j_trans_barrier);

	mlog(0, "commit_thread: flushed transaction %lu (%u handles)\n",
	     journal->j_trans_id, flushed);

	ocfs2_wake_downconvert_thread(osb);
	wake_up(&journal->j_checkpointed);
finally:
	mlog_exit(status);
	return status;
}

/* pass it NULL and it will allocate a new handle object for you.  If
 * you pass it a handle however, it may still return error, in which
 * case it has free'd the passed handle for you. */
handle_t *ocfs2_start_trans(struct ocfs2_super *osb, int max_buffs)
{
	journal_t *journal = osb->journal->j_journal;
	handle_t *handle;

	BUG_ON(!osb || !osb->journal->j_journal);

	if (ocfs2_is_hard_readonly(osb))
		return ERR_PTR(-EROFS);

	BUG_ON(osb->journal->j_state == OCFS2_JOURNAL_FREE);
	BUG_ON(max_buffs <= 0);

	/* Nested transaction? Just return the handle... */
	if (journal_current_handle())
		return jbd2_journal_start(journal, max_buffs);

	down_read(&osb->journal->j_trans_barrier);

	handle = jbd2_journal_start(journal, max_buffs);
	if (IS_ERR(handle)) {
		up_read(&osb->journal->j_trans_barrier);

		mlog_errno(PTR_ERR(handle));

		if (is_journal_aborted(journal)) {
			ocfs2_abort(osb->sb, "Detected aborted journal");
			handle = ERR_PTR(-EROFS);
		}
	} else {
		if (!ocfs2_mount_local(osb))
			atomic_inc(&(osb->journal->j_num_trans));
	}

	return handle;
}

int ocfs2_commit_trans(struct ocfs2_super *osb,
		       handle_t *handle)
{
	int ret, nested;
	struct ocfs2_journal *journal = osb->journal;

	BUG_ON(!handle);

	nested = handle->h_ref > 1;
	ret = jbd2_journal_stop(handle);
	if (ret < 0)
		mlog_errno(ret);

	if (!nested)
		up_read(&journal->j_trans_barrier);

	return ret;
}

/*
 * 'nblocks' is what you want to add to the current
 * transaction. extend_trans will either extend the current handle by
 * nblocks, or commit it and start a new one with nblocks credits.
 *
 * This might call jbd2_journal_restart() which will commit dirty buffers
 * and then restart the transaction. Before calling
 * ocfs2_extend_trans(), any changed blocks should have been
 * dirtied. After calling it, all blocks which need to be changed must
 * go through another set of journal_access/journal_dirty calls.
 *
 * WARNING: This will not release any semaphores or disk locks taken
 * during the transaction, so make sure they were taken *before*
 * start_trans or we'll have ordering deadlocks.
 *
 * WARNING2: Note that we do *not* drop j_trans_barrier here. This is
 * good because transaction ids haven't yet been recorded on the
 * cluster locks associated with this handle.
 */
int ocfs2_extend_trans(handle_t *handle, int nblocks)
{
	int status;

	BUG_ON(!handle);
	BUG_ON(!nblocks);

	mlog_entry_void();

	mlog(0, "Trying to extend transaction by %d blocks\n", nblocks);

#ifdef CONFIG_OCFS2_DEBUG_FS
	status = 1;
#else
	status = jbd2_journal_extend(handle, nblocks);
	if (status < 0) {
		mlog_errno(status);
		goto bail;
	}
#endif

	if (status > 0) {
		mlog(0,
		     "jbd2_journal_extend failed, trying "
		     "jbd2_journal_restart\n");
		status = jbd2_journal_restart(handle, nblocks);
		if (status < 0) {
			mlog_errno(status);
			goto bail;
		}
	}

	status = 0;
bail:

	mlog_exit(status);
	return status;
}

struct ocfs2_triggers {
	struct jbd2_buffer_trigger_type	ot_triggers;
	int				ot_offset;
};

static inline struct ocfs2_triggers *to_ocfs2_trigger(struct jbd2_buffer_trigger_type *triggers)
{
	return container_of(triggers, struct ocfs2_triggers, ot_triggers);
}

static void ocfs2_commit_trigger(struct jbd2_buffer_trigger_type *triggers,
				 struct buffer_head *bh,
				 void *data, size_t size)
{
	struct ocfs2_triggers *ot = to_ocfs2_trigger(triggers);

	/*
	 * We aren't guaranteed to have the superblock here, so we
	 * must unconditionally compute the ecc data.
	 * __ocfs2_journal_access() will only set the triggers if
	 * metaecc is enabled.
	 */
	ocfs2_block_check_compute(data, size, data + ot->ot_offset);
}

/*
 * Quota blocks have their own trigger because the struct ocfs2_block_check
 * offset depends on the blocksize.
 */
static void ocfs2_dq_commit_trigger(struct jbd2_buffer_trigger_type *triggers,
				 struct buffer_head *bh,
				 void *data, size_t size)
{
	struct ocfs2_disk_dqtrailer *dqt =
		ocfs2_block_dqtrailer(size, data);

	/*
	 * We aren't guaranteed to have the superblock here, so we
	 * must unconditionally compute the ecc data.
	 * __ocfs2_journal_access() will only set the triggers if
	 * metaecc is enabled.
	 */
	ocfs2_block_check_compute(data, size, &dqt->dq_check);
}

/*
 * Directory blocks also have their own trigger because the
 * struct ocfs2_block_check offset depends on the blocksize.
 */
static void ocfs2_db_commit_trigger(struct jbd2_buffer_trigger_type *triggers,
				 struct buffer_head *bh,
				 void *data, size_t size)
{
	struct ocfs2_dir_block_trailer *trailer =
		ocfs2_dir_trailer_from_size(size, data);

	/*
	 * We aren't guaranteed to have the superblock here, so we
	 * must unconditionally compute the ecc data.
	 * __ocfs2_journal_access() will only set the triggers if
	 * metaecc is enabled.
	 */
	ocfs2_block_check_compute(data, size, &trailer->db_check);
}

static void ocfs2_abort_trigger(struct jbd2_buffer_trigger_type *triggers,
				struct buffer_head *bh)
{
	mlog(ML_ERROR,
	     "ocfs2_abort_trigger called by JBD2.  bh = 0x%lx, "
	     "bh->b_blocknr = %llu\n",
	     (unsigned long)bh,
	     (unsigned long long)bh->b_blocknr);

	/* We aren't guaranteed to have the superblock here - but if we
	 * don't, it'll just crash. */
	ocfs2_error(bh->b_assoc_map->host->i_sb,
		    "JBD2 has aborted our journal, ocfs2 cannot continue\n");
}

static struct ocfs2_triggers di_triggers = {
	.ot_triggers = {
		.t_commit = ocfs2_commit_trigger,
		.t_abort = ocfs2_abort_trigger,
	},
	.ot_offset	= offsetof(struct ocfs2_dinode, i_check),
};

static struct ocfs2_triggers eb_triggers = {
	.ot_triggers = {
		.t_commit = ocfs2_commit_trigger,
		.t_abort = ocfs2_abort_trigger,
	},
	.ot_offset	= offsetof(struct ocfs2_extent_block, h_check),
};

static struct ocfs2_triggers rb_triggers = {
	.ot_triggers = {
		.t_commit = ocfs2_commit_trigger,
		.t_abort = ocfs2_abort_trigger,
	},
	.ot_offset	= offsetof(struct ocfs2_refcount_block, rf_check),
};

static struct ocfs2_triggers gd_triggers = {
	.ot_triggers = {
		.t_commit = ocfs2_commit_trigger,
		.t_abort = ocfs2_abort_trigger,
	},
	.ot_offset	= offsetof(struct ocfs2_group_desc, bg_check),
};

static struct ocfs2_triggers db_triggers = {
	.ot_triggers = {
		.t_commit = ocfs2_db_commit_trigger,
		.t_abort = ocfs2_abort_trigger,
	},
};

static struct ocfs2_triggers xb_triggers = {
	.ot_triggers = {
		.t_commit = ocfs2_commit_trigger,
		.t_abort = ocfs2_abort_trigger,
	},
	.ot_offset	= offsetof(struct ocfs2_xattr_block, xb_check),
};

static struct ocfs2_triggers dq_triggers = {
	.ot_triggers = {
		.t_commit = ocfs2_dq_commit_trigger,
		.t_abort = ocfs2_abort_trigger,
	},
};

static struct ocfs2_triggers dr_triggers = {
	.ot_triggers = {
		.t_commit = ocfs2_commit_trigger,
		.t_abort = ocfs2_abort_trigger,
	},
	.ot_offset	= offsetof(struct ocfs2_dx_root_block, dr_check),
};

static struct ocfs2_triggers dl_triggers = {
	.ot_triggers = {
		.t_commit = ocfs2_commit_trigger,
		.t_abort = ocfs2_abort_trigger,
	},
	.ot_offset	= offsetof(struct ocfs2_dx_leaf, dl_check),
};

static int __ocfs2_journal_access(handle_t *handle,
				  struct ocfs2_caching_info *ci,
				  struct buffer_head *bh,
				  struct ocfs2_triggers *triggers,
				  int type)
{
	int status;
	struct ocfs2_super *osb =
		OCFS2_SB(ocfs2_metadata_cache_get_super(ci));

	BUG_ON(!ci || !ci->ci_ops);
	BUG_ON(!handle);
	BUG_ON(!bh);

	mlog_entry("bh->b_blocknr=%llu, type=%d (\"%s\"), bh->b_size = %zu\n",
		   (unsigned long long)bh->b_blocknr, type,
		   (type == OCFS2_JOURNAL_ACCESS_CREATE) ?
		   "OCFS2_JOURNAL_ACCESS_CREATE" :
		   "OCFS2_JOURNAL_ACCESS_WRITE",
		   bh->b_size);

	/* we can safely remove this assertion after testing. */
	if (!buffer_uptodate(bh)) {
		mlog(ML_ERROR, "giving me a buffer that's not uptodate!\n");
		mlog(ML_ERROR, "b_blocknr=%llu\n",
		     (unsigned long long)bh->b_blocknr);
		BUG();
	}

	/* Set the current transaction information on the ci so
	 * that the locking code knows whether it can drop it's locks
	 * on this ci or not. We're protected from the commit
	 * thread updating the current transaction id until
	 * ocfs2_commit_trans() because ocfs2_start_trans() took
	 * j_trans_barrier for us. */
	ocfs2_set_ci_lock_trans(osb->journal, ci);

	ocfs2_metadata_cache_io_lock(ci);
	switch (type) {
	case OCFS2_JOURNAL_ACCESS_CREATE:
	case OCFS2_JOURNAL_ACCESS_WRITE:
		status = jbd2_journal_get_write_access(handle, bh);
		break;

	case OCFS2_JOURNAL_ACCESS_UNDO:
		status = jbd2_journal_get_undo_access(handle, bh);
		break;

	default:
		status = -EINVAL;
		mlog(ML_ERROR, "Uknown access type!\n");
	}
	if (!status && ocfs2_meta_ecc(osb) && triggers)
		jbd2_journal_set_triggers(bh, &triggers->ot_triggers);
	ocfs2_metadata_cache_io_unlock(ci);

	if (status < 0)
		mlog(ML_ERROR, "Error %d getting %d access to buffer!\n",
		     status, type);

	mlog_exit(status);
	return status;
}

int ocfs2_journal_access_di(handle_t *handle, struct ocfs2_caching_info *ci,
			    struct buffer_head *bh, int type)
{
	return __ocfs2_journal_access(handle, ci, bh, &di_triggers, type);
}

int ocfs2_journal_access_eb(handle_t *handle, struct ocfs2_caching_info *ci,
			    struct buffer_head *bh, int type)
{
	return __ocfs2_journal_access(handle, ci, bh, &eb_triggers, type);
}

int ocfs2_journal_access_rb(handle_t *handle, struct ocfs2_caching_info *ci,
			    struct buffer_head *bh, int type)
{
	return __ocfs2_journal_access(handle, ci, bh, &rb_triggers,
				      type);
}

int ocfs2_journal_access_gd(handle_t *handle, struct ocfs2_caching_info *ci,
			    struct buffer_head *bh, int type)
{
	return __ocfs2_journal_access(handle, ci, bh, &gd_triggers, type);
}

int ocfs2_journal_access_db(handle_t *handle, struct ocfs2_caching_info *ci,
			    struct buffer_head *bh, int type)
{
	return __ocfs2_journal_access(handle, ci, bh, &db_triggers, type);
}

int ocfs2_journal_access_xb(handle_t *handle, struct ocfs2_caching_info *ci,
			    struct buffer_head *bh, int type)
{
	return __ocfs2_journal_access(handle, ci, bh, &xb_triggers, type);
}

int ocfs2_journal_access_dq(handle_t *handle, struct ocfs2_caching_info *ci,
			    struct buffer_head *bh, int type)
{
	return __ocfs2_journal_access(handle, ci, bh, &dq_triggers, type);
}

int ocfs2_journal_access_dr(handle_t *handle, struct ocfs2_caching_info *ci,
			    struct buffer_head *bh, int type)
{
	return __ocfs2_journal_access(handle, ci, bh, &dr_triggers, type);
}

int ocfs2_journal_access_dl(handle_t *handle, struct ocfs2_caching_info *ci,
			    struct buffer_head *bh, int type)
{
	return __ocfs2_journal_access(handle, ci, bh, &dl_triggers, type);
}

int ocfs2_journal_access(handle_t *handle, struct ocfs2_caching_info *ci,
			 struct buffer_head *bh, int type)
{
	return __ocfs2_journal_access(handle, ci, bh, NULL, type);
}

int ocfs2_journal_dirty(handle_t *handle,
			struct buffer_head *bh)
{
	int status;

	mlog_entry("(bh->b_blocknr=%llu)\n",
		   (unsigned long long)bh->b_blocknr);

	status = jbd2_journal_dirty_metadata(handle, bh);
	if (status < 0)
		mlog(ML_ERROR, "Could not dirty metadata buffer. "
		     "(bh->b_blocknr=%llu)\n",
		     (unsigned long long)bh->b_blocknr);

	mlog_exit(status);
	return status;
}

#define OCFS2_DEFAULT_COMMIT_INTERVAL	(HZ * JBD2_DEFAULT_MAX_COMMIT_AGE)

void ocfs2_set_journal_params(struct ocfs2_super *osb)
{
	journal_t *journal = osb->journal->j_journal;
	unsigned long commit_interval = OCFS2_DEFAULT_COMMIT_INTERVAL;

	if (osb->osb_commit_interval)
		commit_interval = osb->osb_commit_interval;

	spin_lock(&journal->j_state_lock);
	journal->j_commit_interval = commit_interval;
	if (osb->s_mount_opt & OCFS2_MOUNT_BARRIER)
		journal->j_flags |= JBD2_BARRIER;
	else
		journal->j_flags &= ~JBD2_BARRIER;
	spin_unlock(&journal->j_state_lock);
}

int ocfs2_journal_init(struct ocfs2_journal *journal, int *dirty)
{
	int status = -1;
	struct inode *inode = NULL; /* the journal inode */
	journal_t *j_journal = NULL;
	struct ocfs2_dinode *di = NULL;
	struct buffer_head *bh = NULL;
	struct ocfs2_super *osb;
	int inode_lock = 0;

	mlog_entry_void();

	BUG_ON(!journal);

	osb = journal->j_osb;

	/* already have the inode for our journal */
	inode = ocfs2_get_system_file_inode(osb, JOURNAL_SYSTEM_INODE,
					    osb->slot_num);
	if (inode == NULL) {
		status = -EACCES;
		mlog_errno(status);
		goto done;
	}
	if (is_bad_inode(inode)) {
		mlog(ML_ERROR, "access error (bad inode)\n");
		iput(inode);
		inode = NULL;
		status = -EACCES;
		goto done;
	}

	SET_INODE_JOURNAL(inode);
	OCFS2_I(inode)->ip_open_count++;

	/* Skip recovery waits here - journal inode metadata never
	 * changes in a live cluster so it can be considered an
	 * exception to the rule. */
	status = ocfs2_inode_lock_full(inode, &bh, 1, OCFS2_META_LOCK_RECOVERY);
	if (status < 0) {
		if (status != -ERESTARTSYS)
			mlog(ML_ERROR, "Could not get lock on journal!\n");
		goto done;
	}

	inode_lock = 1;
	di = (struct ocfs2_dinode *)bh->b_data;

	if (inode->i_size <  OCFS2_MIN_JOURNAL_SIZE) {
		mlog(ML_ERROR, "Journal file size (%lld) is too small!\n",
		     inode->i_size);
		status = -EINVAL;
		goto done;
	}

	mlog(0, "inode->i_size = %lld\n", inode->i_size);
	mlog(0, "inode->i_blocks = %llu\n",
			(unsigned long long)inode->i_blocks);
	mlog(0, "inode->ip_clusters = %u\n", OCFS2_I(inode)->ip_clusters);

	/* call the kernels journal init function now */
	j_journal = jbd2_journal_init_inode(inode);
	if (j_journal == NULL) {
		mlog(ML_ERROR, "Linux journal layer error\n");
		status = -EINVAL;
		goto done;
	}

	mlog(0, "Returned from jbd2_journal_init_inode\n");
	mlog(0, "j_journal->j_maxlen = %u\n", j_journal->j_maxlen);

	*dirty = (le32_to_cpu(di->id1.journal1.ij_flags) &
		  OCFS2_JOURNAL_DIRTY_FL);

	journal->j_journal = j_journal;
	journal->j_inode = inode;
	journal->j_bh = bh;

	ocfs2_set_journal_params(osb);

	journal->j_state = OCFS2_JOURNAL_LOADED;

	status = 0;
done:
	if (status < 0) {
		if (inode_lock)
			ocfs2_inode_unlock(inode, 1);
		brelse(bh);
		if (inode) {
			OCFS2_I(inode)->ip_open_count--;
			iput(inode);
		}
	}

	mlog_exit(status);
	return status;
}

static void ocfs2_bump_recovery_generation(struct ocfs2_dinode *di)
{
	le32_add_cpu(&(di->id1.journal1.ij_recovery_generation), 1);
}

static u32 ocfs2_get_recovery_generation(struct ocfs2_dinode *di)
{
	return le32_to_cpu(di->id1.journal1.ij_recovery_generation);
}

static int ocfs2_journal_toggle_dirty(struct ocfs2_super *osb,
				      int dirty, int replayed)
{
	int status;
	unsigned int flags;
	struct ocfs2_journal *journal = osb->journal;
	struct buffer_head *bh = journal->j_bh;
	struct ocfs2_dinode *fe;

	mlog_entry_void();

	fe = (struct ocfs2_dinode *)bh->b_data;

	/* The journal bh on the osb always comes from ocfs2_journal_init()
	 * and was validated there inside ocfs2_inode_lock_full().  It's a
	 * code bug if we mess it up. */
	BUG_ON(!OCFS2_IS_VALID_DINODE(fe));

	flags = le32_to_cpu(fe->id1.journal1.ij_flags);
	if (dirty)
		flags |= OCFS2_JOURNAL_DIRTY_FL;
	else
		flags &= ~OCFS2_JOURNAL_DIRTY_FL;
	fe->id1.journal1.ij_flags = cpu_to_le32(flags);

	if (replayed)
		ocfs2_bump_recovery_generation(fe);

	ocfs2_compute_meta_ecc(osb->sb, bh->b_data, &fe->i_check);
	status = ocfs2_write_block(osb, bh, INODE_CACHE(journal->j_inode));
	if (status < 0)
		mlog_errno(status);

	mlog_exit(status);
	return status;
}

/*
 * If the journal has been kmalloc'd it needs to be freed after this
 * call.
 */
void ocfs2_journal_shutdown(struct ocfs2_super *osb)
{
	struct ocfs2_journal *journal = NULL;
	int status = 0;
	struct inode *inode = NULL;
	int num_running_trans = 0;

	mlog_entry_void();

	BUG_ON(!osb);

	journal = osb->journal;
	if (!journal)
		goto done;

	inode = journal->j_inode;

	if (journal->j_state != OCFS2_JOURNAL_LOADED)
		goto done;

	/* need to inc inode use count - jbd2_journal_destroy will iput. */
	if (!igrab(inode))
		BUG();

	num_running_trans = atomic_read(&(osb->journal->j_num_trans));
	if (num_running_trans > 0)
		mlog(0, "Shutting down journal: must wait on %d "
		     "running transactions!\n",
		     num_running_trans);

	/* Do a commit_cache here. It will flush our journal, *and*
	 * release any locks that are still held.
	 * set the SHUTDOWN flag and release the trans lock.
	 * the commit thread will take the trans lock for us below. */
	journal->j_state = OCFS2_JOURNAL_IN_SHUTDOWN;

	/* The OCFS2_JOURNAL_IN_SHUTDOWN will signal to commit_cache to not
	 * drop the trans_lock (which we want to hold until we
	 * completely destroy the journal. */
	if (osb->commit_task) {
		/* Wait for the commit thread */
		mlog(0, "Waiting for ocfs2commit to exit....\n");
		kthread_stop(osb->commit_task);
		osb->commit_task = NULL;
	}

	BUG_ON(atomic_read(&(osb->journal->j_num_trans)) != 0);

	if (ocfs2_mount_local(osb)) {
		jbd2_journal_lock_updates(journal->j_journal);
		status = jbd2_journal_flush(journal->j_journal);
		jbd2_journal_unlock_updates(journal->j_journal);
		if (status < 0)
			mlog_errno(status);
	}

	if (status == 0) {
		/*
		 * Do not toggle if flush was unsuccessful otherwise
		 * will leave dirty metadata in a "clean" journal
		 */
		status = ocfs2_journal_toggle_dirty(osb, 0, 0);
		if (status < 0)
			mlog_errno(status);
	}

	/* Shutdown the kernel journal system */
	jbd2_journal_destroy(journal->j_journal);
	journal->j_journal = NULL;

	OCFS2_I(inode)->ip_open_count--;

	/* unlock our journal */
	ocfs2_inode_unlock(inode, 1);

	brelse(journal->j_bh);
	journal->j_bh = NULL;

	journal->j_state = OCFS2_JOURNAL_FREE;

//	up_write(&journal->j_trans_barrier);
done:
	if (inode)
		iput(inode);
	mlog_exit_void();
}

static void ocfs2_clear_journal_error(struct super_block *sb,
				      journal_t *journal,
				      int slot)
{
	int olderr;

	olderr = jbd2_journal_errno(journal);
	if (olderr) {
		mlog(ML_ERROR, "File system error %d recorded in "
		     "journal %u.\n", olderr, slot);
		mlog(ML_ERROR, "File system on device %s needs checking.\n",
		     sb->s_id);

		jbd2_journal_ack_err(journal);
		jbd2_journal_clear_err(journal);
	}
}

int ocfs2_journal_load(struct ocfs2_journal *journal, int local, int replayed)
{
	int status = 0;
	struct ocfs2_super *osb;

	mlog_entry_void();

	BUG_ON(!journal);

	osb = journal->j_osb;

	status = jbd2_journal_load(journal->j_journal);
	if (status < 0) {
		mlog(ML_ERROR, "Failed to load journal!\n");
		goto done;
	}

	ocfs2_clear_journal_error(osb->sb, journal->j_journal, osb->slot_num);

	status = ocfs2_journal_toggle_dirty(osb, 1, replayed);
	if (status < 0) {
		mlog_errno(status);
		goto done;
	}

	/* Launch the commit thread */
	if (!local) {
		osb->commit_task = kthread_run(ocfs2_commit_thread, osb,
					       "ocfs2cmt");
		if (IS_ERR(osb->commit_task)) {
			status = PTR_ERR(osb->commit_task);
			osb->commit_task = NULL;
			mlog(ML_ERROR, "unable to launch ocfs2commit thread, "
			     "error=%d", status);
			goto done;
		}
	} else
		osb->commit_task = NULL;

done:
	mlog_exit(status);
	return status;
}


/* 'full' flag tells us whether we clear out all blocks or if we just
 * mark the journal clean */
int ocfs2_journal_wipe(struct ocfs2_journal *journal, int full)
{
	int status;

	mlog_entry_void();

	BUG_ON(!journal);

	status = jbd2_journal_wipe(journal->j_journal, full);
	if (status < 0) {
		mlog_errno(status);
		goto bail;
	}

	status = ocfs2_journal_toggle_dirty(journal->j_osb, 0, 0);
	if (status < 0)
		mlog_errno(status);

bail:
	mlog_exit(status);
	return status;
}

static int ocfs2_recovery_completed(struct ocfs2_super *osb)
{
	int empty;
	struct ocfs2_recovery_map *rm = osb->recovery_map;

	spin_lock(&osb->osb_lock);
	empty = (rm->rm_used == 0);
	spin_unlock(&osb->osb_lock);

	return empty;
}

void ocfs2_wait_for_recovery(struct ocfs2_super *osb)
{
	wait_event(osb->recovery_event, ocfs2_recovery_completed(osb));
}

/*
 * JBD Might read a cached version of another nodes journal file. We
 * don't want this as this file changes often and we get no
 * notification on those changes. The only way to be sure that we've
 * got the most up to date version of those blocks then is to force
 * read them off disk. Just searching through the buffer cache won't
 * work as there may be pages backing this file which are still marked
 * up to date. We know things can't change on this file underneath us
 * as we have the lock by now :)
 */
static int ocfs2_force_read_journal(struct inode *inode)
{
	int status = 0;
	int i;
	u64 v_blkno, p_blkno, p_blocks, num_blocks;
#define CONCURRENT_JOURNAL_FILL 32ULL
	struct buffer_head *bhs[CONCURRENT_JOURNAL_FILL];

	mlog_entry_void();

	memset(bhs, 0, sizeof(struct buffer_head *) * CONCURRENT_JOURNAL_FILL);

	num_blocks = ocfs2_blocks_for_bytes(inode->i_sb, inode->i_size);
	v_blkno = 0;
	while (v_blkno < num_blocks) {
		status = ocfs2_extent_map_get_blocks(inode, v_blkno,
						     &p_blkno, &p_blocks, NULL);
		if (status < 0) {
			mlog_errno(status);
			goto bail;
		}

		if (p_blocks > CONCURRENT_JOURNAL_FILL)
			p_blocks = CONCURRENT_JOURNAL_FILL;

		/* We are reading journal data which should not
		 * be put in the uptodate cache */
		status = ocfs2_read_blocks_sync(OCFS2_SB(inode->i_sb),
						p_blkno, p_blocks, bhs);
		if (status < 0) {
			mlog_errno(status);
			goto bail;
		}

		for(i = 0; i < p_blocks; i++) {
			brelse(bhs[i]);
			bhs[i] = NULL;
		}

		v_blkno += p_blocks;
	}

bail:
	for(i = 0; i < CONCURRENT_JOURNAL_FILL; i++)
		brelse(bhs[i]);
	mlog_exit(status);
	return status;
}

struct ocfs2_la_recovery_item {
	struct list_head	lri_list;
	int			lri_slot;
	struct ocfs2_dinode	*lri_la_dinode;
	struct ocfs2_dinode	*lri_tl_dinode;
	struct ocfs2_quota_recovery *lri_qrec;
};

/* Does the second half of the recovery process. By this point, the
 * node is marked clean and can actually be considered recovered,
 * hence it's no longer in the recovery map, but there's still some
 * cleanup we can do which shouldn't happen within the recovery thread
 * as locking in that context becomes very difficult if we are to take
 * recovering nodes into account.
 *
 * NOTE: This function can and will sleep on recovery of other nodes
 * during cluster locking, just like any other ocfs2 process.
 */
void ocfs2_complete_recovery(struct work_struct *work)
{
	int ret;
	struct ocfs2_journal *journal =
		container_of(work, struct ocfs2_journal, j_recovery_work);
	struct ocfs2_super *osb = journal->j_osb;
	struct ocfs2_dinode *la_dinode, *tl_dinode;
	struct ocfs2_la_recovery_item *item, *n;
	struct ocfs2_quota_recovery *qrec;
	LIST_HEAD(tmp_la_list);

	mlog_entry_void();

	mlog(0, "completing recovery from keventd\n");

	spin_lock(&journal->j_lock);
	list_splice_init(&journal->j_la_cleanups, &tmp_la_list);
	spin_unlock(&journal->j_lock);

	list_for_each_entry_safe(item, n, &tmp_la_list, lri_list) {
		list_del_init(&item->lri_list);

		mlog(0, "Complete recovery for slot %d\n", item->lri_slot);

		ocfs2_wait_on_quotas(osb);

		la_dinode = item->lri_la_dinode;
		if (la_dinode) {
			mlog(0, "Clean up local alloc %llu\n",
			     (unsigned long long)le64_to_cpu(la_dinode->i_blkno));

			ret = ocfs2_complete_local_alloc_recovery(osb,
								  la_dinode);
			if (ret < 0)
				mlog_errno(ret);

			kfree(la_dinode);
		}

		tl_dinode = item->lri_tl_dinode;
		if (tl_dinode) {
			mlog(0, "Clean up truncate log %llu\n",
			     (unsigned long long)le64_to_cpu(tl_dinode->i_blkno));

			ret = ocfs2_complete_truncate_log_recovery(osb,
								   tl_dinode);
			if (ret < 0)
				mlog_errno(ret);

			kfree(tl_dinode);
		}

		ret = ocfs2_recover_orphans(osb, item->lri_slot);
		if (ret < 0)
			mlog_errno(ret);

		qrec = item->lri_qrec;
		if (qrec) {
			mlog(0, "Recovering quota files");
			ret = ocfs2_finish_quota_recovery(osb, qrec,
							  item->lri_slot);
			if (ret < 0)
				mlog_errno(ret);
			/* Recovery info is already freed now */
		}

		kfree(item);
	}

	mlog(0, "Recovery completion\n");
	mlog_exit_void();
}

/* NOTE: This function always eats your references to la_dinode and
 * tl_dinode, either manually on error, or by passing them to
 * ocfs2_complete_recovery */
static void ocfs2_queue_recovery_completion(struct ocfs2_journal *journal,
					    int slot_num,
					    struct ocfs2_dinode *la_dinode,
					    struct ocfs2_dinode *tl_dinode,
					    struct ocfs2_quota_recovery *qrec)
{
	struct ocfs2_la_recovery_item *item;

	item = kmalloc(sizeof(struct ocfs2_la_recovery_item), GFP_NOFS);
	if (!item) {
		/* Though we wish to avoid it, we are in fact safe in
		 * skipping local alloc cleanup as fsck.ocfs2 is more
		 * than capable of reclaiming unused space. */
		if (la_dinode)
			kfree(la_dinode);

		if (tl_dinode)
			kfree(tl_dinode);

		if (qrec)
			ocfs2_free_quota_recovery(qrec);

		mlog_errno(-ENOMEM);
		return;
	}

	INIT_LIST_HEAD(&item->lri_list);
	item->lri_la_dinode = la_dinode;
	item->lri_slot = slot_num;
	item->lri_tl_dinode = tl_dinode;
	item->lri_qrec = qrec;

	spin_lock(&journal->j_lock);
	list_add_tail(&item->lri_list, &journal->j_la_cleanups);
	queue_work(ocfs2_wq, &journal->j_recovery_work);
	spin_unlock(&journal->j_lock);
}

/* Called by the mount code to queue recovery the last part of
 * recovery for it's own and offline slot(s). */
void ocfs2_complete_mount_recovery(struct ocfs2_super *osb)
{
	struct ocfs2_journal *journal = osb->journal;

	/* No need to queue up our truncate_log as regular cleanup will catch
	 * that */
	ocfs2_queue_recovery_completion(journal, osb->slot_num,
					osb->local_alloc_copy, NULL, NULL);
	ocfs2_schedule_truncate_log_flush(osb, 0);

	osb->local_alloc_copy = NULL;
	osb->dirty = 0;

	/* queue to recover orphan slots for all offline slots */
	ocfs2_replay_map_set_state(osb, REPLAY_NEEDED);
	ocfs2_queue_replay_slots(osb);
	ocfs2_free_replay_slots(osb);
}

void ocfs2_complete_quota_recovery(struct ocfs2_super *osb)
{
	if (osb->quota_rec) {
		ocfs2_queue_recovery_completion(osb->journal,
						osb->slot_num,
						NULL,
						NULL,
						osb->quota_rec);
		osb->quota_rec = NULL;
	}
}

static int __ocfs2_recovery_thread(void *arg)
{
	int status, node_num, slot_num;
	struct ocfs2_super *osb = arg;
	struct ocfs2_recovery_map *rm = osb->recovery_map;
	int *rm_quota = NULL;
	int rm_quota_used = 0, i;
	struct ocfs2_quota_recovery *qrec;

	mlog_entry_void();

	status = ocfs2_wait_on_mount(osb);
	if (status < 0) {
		goto bail;
	}

	rm_quota = kzalloc(osb->max_slots * sizeof(int), GFP_NOFS);
	if (!rm_quota) {
		status = -ENOMEM;
		goto bail;
	}
restart:
	status = ocfs2_super_lock(osb, 1);
	if (status < 0) {
		mlog_errno(status);
		goto bail;
	}

	status = ocfs2_compute_replay_slots(osb);
	if (status < 0)
		mlog_errno(status);

	/* queue recovery for our own slot */
	ocfs2_queue_recovery_completion(osb->journal, osb->slot_num, NULL,
					NULL, NULL);

	spin_lock(&osb->osb_lock);
	while (rm->rm_used) {
		/* It's always safe to remove entry zero, as we won't
		 * clear it until ocfs2_recover_node() has succeeded. */
		node_num = rm->rm_entries[0];
		spin_unlock(&osb->osb_lock);
		mlog(0, "checking node %d\n", node_num);
		slot_num = ocfs2_node_num_to_slot(osb, node_num);
		if (slot_num == -ENOENT) {
			status = 0;
			mlog(0, "no slot for this node, so no recovery"
			     "required.\n");
			goto skip_recovery;
		}
		mlog(0, "node %d was using slot %d\n", node_num, slot_num);

		/* It is a bit subtle with quota recovery. We cannot do it
		 * immediately because we have to obtain cluster locks from
		 * quota files and we also don't want to just skip it because
		 * then quota usage would be out of sync until some node takes
		 * the slot. So we remember which nodes need quota recovery
		 * and when everything else is done, we recover quotas. */
		for (i = 0; i < rm_quota_used && rm_quota[i] != slot_num; i++);
		if (i == rm_quota_used)
			rm_quota[rm_quota_used++] = slot_num;

		status = ocfs2_recover_node(osb, node_num, slot_num);
skip_recovery:
		if (!status) {
			ocfs2_recovery_map_clear(osb, node_num);
		} else {
			mlog(ML_ERROR,
			     "Error %d recovering node %d on device (%u,%u)!\n",
			     status, node_num,
			     MAJOR(osb->sb->s_dev), MINOR(osb->sb->s_dev));
			mlog(ML_ERROR, "Volume requires unmount.\n");
		}

		spin_lock(&osb->osb_lock);
	}
	spin_unlock(&osb->osb_lock);
	mlog(0, "All nodes recovered\n");

	/* Refresh all journal recovery generations from disk */
	status = ocfs2_check_journals_nolocks(osb);
	status = (status == -EROFS) ? 0 : status;
	if (status < 0)
		mlog_errno(status);

	/* Now it is right time to recover quotas... We have to do this under
	 * superblock lock so that noone can start using the slot (and crash)
	 * before we recover it */
	for (i = 0; i < rm_quota_used; i++) {
		qrec = ocfs2_begin_quota_recovery(osb, rm_quota[i]);
		if (IS_ERR(qrec)) {
			status = PTR_ERR(qrec);
			mlog_errno(status);
			continue;
		}
		ocfs2_queue_recovery_completion(osb->journal, rm_quota[i],
						NULL, NULL, qrec);
	}

	ocfs2_super_unlock(osb, 1);

	/* queue recovery for offline slots */
	ocfs2_queue_replay_slots(osb);

bail:
	mutex_lock(&osb->recovery_lock);
	if (!status && !ocfs2_recovery_completed(osb)) {
		mutex_unlock(&osb->recovery_lock);
		goto restart;
	}

	ocfs2_free_replay_slots(osb);
	osb->recovery_thread_task = NULL;
	mb(); /* sync with ocfs2_recovery_thread_running */
	wake_up(&osb->recovery_event);

	mutex_unlock(&osb->recovery_lock);

	if (rm_quota)
		kfree(rm_quota);

	mlog_exit(status);
	/* no one is callint kthread_stop() for us so the kthread() api
	 * requires that we call do_exit().  And it isn't exported, but
	 * complete_and_exit() seems to be a minimal wrapper around it. */
	complete_and_exit(NULL, status);
	return status;
}

void ocfs2_recovery_thread(struct ocfs2_super *osb, int node_num)
{
	mlog_entry("(node_num=%d, osb->node_num = %d)\n",
		   node_num, osb->node_num);

	mutex_lock(&osb->recovery_lock);
	if (osb->disable_recovery)
		goto out;

	/* People waiting on recovery will wait on
	 * the recovery map to empty. */
	if (ocfs2_recovery_map_set(osb, node_num))
		mlog(0, "node %d already in recovery map.\n", node_num);

	mlog(0, "starting recovery thread...\n");

	if (osb->recovery_thread_task)
		goto out;

	osb->recovery_thread_task =  kthread_run(__ocfs2_recovery_thread, osb,
						 "ocfs2rec");
	if (IS_ERR(osb->recovery_thread_task)) {
		mlog_errno((int)PTR_ERR(osb->recovery_thread_task));
		osb->recovery_thread_task = NULL;
	}

out:
	mutex_unlock(&osb->recovery_lock);
	wake_up(&osb->recovery_event);

	mlog_exit_void();
}

static int ocfs2_read_journal_inode(struct ocfs2_super *osb,
				    int slot_num,
				    struct buffer_head **bh,
				    struct inode **ret_inode)
{
	int status = -EACCES;
	struct inode *inode = NULL;

	BUG_ON(slot_num >= osb->max_slots);

	inode = ocfs2_get_system_file_inode(osb, JOURNAL_SYSTEM_INODE,
					    slot_num);
	if (!inode || is_bad_inode(inode)) {
		mlog_errno(status);
		goto bail;
	}
	SET_INODE_JOURNAL(inode);

	status = ocfs2_read_inode_block_full(inode, bh, OCFS2_BH_IGNORE_CACHE);
	if (status < 0) {
		mlog_errno(status);
		goto bail;
	}

	status = 0;

bail:
	if (inode) {
		if (status || !ret_inode)
			iput(inode);
		else
			*ret_inode = inode;
	}
	return status;
}

/* Does the actual journal replay and marks the journal inode as
 * clean. Will only replay if the journal inode is marked dirty. */
static int ocfs2_replay_journal(struct ocfs2_super *osb,
				int node_num,
				int slot_num)
{
	int status;
	int got_lock = 0;
	unsigned int flags;
	struct inode *inode = NULL;
	struct ocfs2_dinode *fe;
	journal_t *journal = NULL;
	struct buffer_head *bh = NULL;
	u32 slot_reco_gen;

	status = ocfs2_read_journal_inode(osb, slot_num, &bh, &inode);
	if (status) {
		mlog_errno(status);
		goto done;
	}

	fe = (struct ocfs2_dinode *)bh->b_data;
	slot_reco_gen = ocfs2_get_recovery_generation(fe);
	brelse(bh);
	bh = NULL;

	/*
	 * As the fs recovery is asynchronous, there is a small chance that
	 * another node mounted (and recovered) the slot before the recovery
	 * thread could get the lock. To handle that, we dirty read the journal
	 * inode for that slot to get the recovery generation. If it is
	 * different than what we expected, the slot has been recovered.
	 * If not, it needs recovery.
	 */
	if (osb->slot_recovery_generations[slot_num] != slot_reco_gen) {
		mlog(0, "Slot %u already recovered (old/new=%u/%u)\n", slot_num,
		     osb->slot_recovery_generations[slot_num], slot_reco_gen);
		osb->slot_recovery_generations[slot_num] = slot_reco_gen;
		status = -EBUSY;
		goto done;
	}

	/* Continue with recovery as the journal has not yet been recovered */

	status = ocfs2_inode_lock_full(inode, &bh, 1, OCFS2_META_LOCK_RECOVERY);
	if (status < 0) {
		mlog(0, "status returned from ocfs2_inode_lock=%d\n", status);
		if (status != -ERESTARTSYS)
			mlog(ML_ERROR, "Could not lock journal!\n");
		goto done;
	}
	got_lock = 1;

	fe = (struct ocfs2_dinode *) bh->b_data;

	flags = le32_to_cpu(fe->id1.journal1.ij_flags);
	slot_reco_gen = ocfs2_get_recovery_generation(fe);

	if (!(flags & OCFS2_JOURNAL_DIRTY_FL)) {
		mlog(0, "No recovery required for node %d\n", node_num);
		/* Refresh recovery generation for the slot */
		osb->slot_recovery_generations[slot_num] = slot_reco_gen;
		goto done;
	}

	/* we need to run complete recovery for offline orphan slots */
	ocfs2_replay_map_set_state(osb, REPLAY_NEEDED);

	mlog(ML_NOTICE, "Recovering node %d from slot %d on device (%u,%u)\n",
	     node_num, slot_num,
	     MAJOR(osb->sb->s_dev), MINOR(osb->sb->s_dev));

	OCFS2_I(inode)->ip_clusters = le32_to_cpu(fe->i_clusters);

	status = ocfs2_force_read_journal(inode);
	if (status < 0) {
		mlog_errno(status);
		goto done;
	}

	mlog(0, "calling journal_init_inode\n");
	journal = jbd2_journal_init_inode(inode);
	if (journal == NULL) {
		mlog(ML_ERROR, "Linux journal layer error\n");
		status = -EIO;
		goto done;
	}

	status = jbd2_journal_load(journal);
	if (status < 0) {
		mlog_errno(status);
		if (!igrab(inode))
			BUG();
		jbd2_journal_destroy(journal);
		goto done;
	}

	ocfs2_clear_journal_error(osb->sb, journal, slot_num);

	/* wipe the journal */
	mlog(0, "flushing the journal.\n");
	jbd2_journal_lock_updates(journal);
	status = jbd2_journal_flush(journal);
	jbd2_journal_unlock_updates(journal);
	if (status < 0)
		mlog_errno(status);

	/* This will mark the node clean */
	flags = le32_to_cpu(fe->id1.journal1.ij_flags);
	flags &= ~OCFS2_JOURNAL_DIRTY_FL;
	fe->id1.journal1.ij_flags = cpu_to_le32(flags);

	/* Increment recovery generation to indicate successful recovery */
	ocfs2_bump_recovery_generation(fe);
	osb->slot_recovery_generations[slot_num] =
					ocfs2_get_recovery_generation(fe);

	ocfs2_compute_meta_ecc(osb->sb, bh->b_data, &fe->i_check);
	status = ocfs2_write_block(osb, bh, INODE_CACHE(inode));
	if (status < 0)
		mlog_errno(status);

	if (!igrab(inode))
		BUG();

	jbd2_journal_destroy(journal);

done:
	/* drop the lock on this nodes journal */
	if (got_lock)
		ocfs2_inode_unlock(inode, 1);

	if (inode)
		iput(inode);

	brelse(bh);

	mlog_exit(status);
	return status;
}

/*
 * Do the most important parts of node recovery:
 *  - Replay it's journal
 *  - Stamp a clean local allocator file
 *  - Stamp a clean truncate log
 *  - Mark the node clean
 *
 * If this function completes without error, a node in OCFS2 can be
 * said to have been safely recovered. As a result, failure during the
 * second part of a nodes recovery process (local alloc recovery) is
 * far less concerning.
 */
static int ocfs2_recover_node(struct ocfs2_super *osb,
			      int node_num, int slot_num)
{
	int status = 0;
	struct ocfs2_dinode *la_copy = NULL;
	struct ocfs2_dinode *tl_copy = NULL;

	mlog_entry("(node_num=%d, slot_num=%d, osb->node_num = %d)\n",
		   node_num, slot_num, osb->node_num);

	/* Should not ever be called to recover ourselves -- in that
	 * case we should've called ocfs2_journal_load instead. */
	BUG_ON(osb->node_num == node_num);

	status = ocfs2_replay_journal(osb, node_num, slot_num);
	if (status < 0) {
		if (status == -EBUSY) {
			mlog(0, "Skipping recovery for slot %u (node %u) "
			     "as another node has recovered it\n", slot_num,
			     node_num);
			status = 0;
			goto done;
		}
		mlog_errno(status);
		goto done;
	}

	/* Stamp a clean local alloc file AFTER recovering the journal... */
	status = ocfs2_begin_local_alloc_recovery(osb, slot_num, &la_copy);
	if (status < 0) {
		mlog_errno(status);
		goto done;
	}

	/* An error from begin_truncate_log_recovery is not
	 * serious enough to warrant halting the rest of
	 * recovery. */
	status = ocfs2_begin_truncate_log_recovery(osb, slot_num, &tl_copy);
	if (status < 0)
		mlog_errno(status);

	/* Likewise, this would be a strange but ultimately not so
	 * harmful place to get an error... */
	status = ocfs2_clear_slot(osb, slot_num);
	if (status < 0)
		mlog_errno(status);

	/* This will kfree the memory pointed to by la_copy and tl_copy */
	ocfs2_queue_recovery_completion(osb->journal, slot_num, la_copy,
					tl_copy, NULL);

	status = 0;
done:

	mlog_exit(status);
	return status;
}

/* Test node liveness by trylocking his journal. If we get the lock,
 * we drop it here. Return 0 if we got the lock, -EAGAIN if node is
 * still alive (we couldn't get the lock) and < 0 on error. */
static int ocfs2_trylock_journal(struct ocfs2_super *osb,
				 int slot_num)
{
	int status, flags;
	struct inode *inode = NULL;

	inode = ocfs2_get_system_file_inode(osb, JOURNAL_SYSTEM_INODE,
					    slot_num);
	if (inode == NULL) {
		mlog(ML_ERROR, "access error\n");
		status = -EACCES;
		goto bail;
	}
	if (is_bad_inode(inode)) {
		mlog(ML_ERROR, "access error (bad inode)\n");
		iput(inode);
		inode = NULL;
		status = -EACCES;
		goto bail;
	}
	SET_INODE_JOURNAL(inode);

	flags = OCFS2_META_LOCK_RECOVERY | OCFS2_META_LOCK_NOQUEUE;
	status = ocfs2_inode_lock_full(inode, NULL, 1, flags);
	if (status < 0) {
		if (status != -EAGAIN)
			mlog_errno(status);
		goto bail;
	}

	ocfs2_inode_unlock(inode, 1);
bail:
	if (inode)
		iput(inode);

	return status;
}

/* Call this underneath ocfs2_super_lock. It also assumes that the
 * slot info struct has been updated from disk. */
int ocfs2_mark_dead_nodes(struct ocfs2_super *osb)
{
	unsigned int node_num;
	int status, i;
	u32 gen;
	struct buffer_head *bh = NULL;
	struct ocfs2_dinode *di;

	/* This is called with the super block cluster lock, so we
	 * know that the slot map can't change underneath us. */

	for (i = 0; i < osb->max_slots; i++) {
		/* Read journal inode to get the recovery generation */
		status = ocfs2_read_journal_inode(osb, i, &bh, NULL);
		if (status) {
			mlog_errno(status);
			goto bail;
		}
		di = (struct ocfs2_dinode *)bh->b_data;
		gen = ocfs2_get_recovery_generation(di);
		brelse(bh);
		bh = NULL;

		spin_lock(&osb->osb_lock);
		osb->slot_recovery_generations[i] = gen;

		mlog(0, "Slot %u recovery generation is %u\n", i,
		     osb->slot_recovery_generations[i]);

		if (i == osb->slot_num) {
			spin_unlock(&osb->osb_lock);
			continue;
		}

		status = ocfs2_slot_to_node_num_locked(osb, i, &node_num);
		if (status == -ENOENT) {
			spin_unlock(&osb->osb_lock);
			continue;
		}

		if (__ocfs2_recovery_map_test(osb, node_num)) {
			spin_unlock(&osb->osb_lock);
			continue;
		}
		spin_unlock(&osb->osb_lock);

		/* Ok, we have a slot occupied by another node which
		 * is not in the recovery map. We trylock his journal
		 * file here to test if he's alive. */
		status = ocfs2_trylock_journal(osb, i);
		if (!status) {
			/* Since we're called from mount, we know that
			 * the recovery thread can't race us on
			 * setting / checking the recovery bits. */
			ocfs2_recovery_thread(osb, node_num);
		} else if ((status < 0) && (status != -EAGAIN)) {
			mlog_errno(status);
			goto bail;
		}
	}

	status = 0;
bail:
	mlog_exit(status);
	return status;
}

/*
 * Scan timer should get fired every ORPHAN_SCAN_SCHEDULE_TIMEOUT. Add some
 * randomness to the timeout to minimize multple nodes firing the timer at the
 * same time.
 */
static inline unsigned long ocfs2_orphan_scan_timeout(void)
{
	unsigned long time;

	get_random_bytes(&time, sizeof(time));
	time = ORPHAN_SCAN_SCHEDULE_TIMEOUT + (time % 5000);
	return msecs_to_jiffies(time);
}

/*
 * ocfs2_queue_orphan_scan calls ocfs2_queue_recovery_completion for
 * every slot, queuing a recovery of the slot on the ocfs2_wq thread. This
 * is done to catch any orphans that are left over in orphan directories.
 *
 * ocfs2_queue_orphan_scan gets called every ORPHAN_SCAN_SCHEDULE_TIMEOUT
 * seconds.  It gets an EX lock on os_lockres and checks sequence number
 * stored in LVB. If the sequence number has changed, it means some other
 * node has done the scan.  This node skips the scan and tracks the
 * sequence number.  If the sequence number didn't change, it means a scan
 * hasn't happened.  The node queues a scan and increments the
 * sequence number in the LVB.
 */
void ocfs2_queue_orphan_scan(struct ocfs2_super *osb)
{
	struct ocfs2_orphan_scan *os;
	int status, i;
	u32 seqno = 0;

	os = &osb->osb_orphan_scan;

	if (atomic_read(&os->os_state) == ORPHAN_SCAN_INACTIVE)
		goto out;

	status = ocfs2_orphan_scan_lock(osb, &seqno);
	if (status < 0) {
		if (status != -EAGAIN)
			mlog_errno(status);
		goto out;
	}

	/* Do no queue the tasks if the volume is being umounted */
	if (atomic_read(&os->os_state) == ORPHAN_SCAN_INACTIVE)
		goto unlock;

	if (os->os_seqno != seqno) {
		os->os_seqno = seqno;
		goto unlock;
	}

	for (i = 0; i < osb->max_slots; i++)
		ocfs2_queue_recovery_completion(osb->journal, i, NULL, NULL,
						NULL);
	/*
	 * We queued a recovery on orphan slots, increment the sequence
	 * number and update LVB so other node will skip the scan for a while
	 */
	seqno++;
	os->os_count++;
	os->os_scantime = CURRENT_TIME;
unlock:
	ocfs2_orphan_scan_unlock(osb, seqno);
out:
	return;
}

/* Worker task that gets fired every ORPHAN_SCAN_SCHEDULE_TIMEOUT millsec */
void ocfs2_orphan_scan_work(struct work_struct *work)
{
	struct ocfs2_orphan_scan *os;
	struct ocfs2_super *osb;

	os = container_of(work, struct ocfs2_orphan_scan,
			  os_orphan_scan_work.work);
	osb = os->os_osb;

	mutex_lock(&os->os_lock);
	ocfs2_queue_orphan_scan(osb);
	if (atomic_read(&os->os_state) == ORPHAN_SCAN_ACTIVE)
		schedule_delayed_work(&os->os_orphan_scan_work,
				      ocfs2_orphan_scan_timeout());
	mutex_unlock(&os->os_lock);
}

void ocfs2_orphan_scan_stop(struct ocfs2_super *osb)
{
	struct ocfs2_orphan_scan *os;

	os = &osb->osb_orphan_scan;
	if (atomic_read(&os->os_state) == ORPHAN_SCAN_ACTIVE) {
		atomic_set(&os->os_state, ORPHAN_SCAN_INACTIVE);
		mutex_lock(&os->os_lock);
		cancel_delayed_work(&os->os_orphan_scan_work);
		mutex_unlock(&os->os_lock);
	}
}

void ocfs2_orphan_scan_init(struct ocfs2_super *osb)
{
	struct ocfs2_orphan_scan *os;

	os = &osb->osb_orphan_scan;
	os->os_osb = osb;
	os->os_count = 0;
	os->os_seqno = 0;
	mutex_init(&os->os_lock);
	INIT_DELAYED_WORK(&os->os_orphan_scan_work, ocfs2_orphan_scan_work);
}

void ocfs2_orphan_scan_start(struct ocfs2_super *osb)
{
	struct ocfs2_orphan_scan *os;

	os = &osb->osb_orphan_scan;
	os->os_scantime = CURRENT_TIME;
	if (ocfs2_is_hard_readonly(osb) || ocfs2_mount_local(osb))
		atomic_set(&os->os_state, ORPHAN_SCAN_INACTIVE);
	else {
		atomic_set(&os->os_state, ORPHAN_SCAN_ACTIVE);
		schedule_delayed_work(&os->os_orphan_scan_work,
				      ocfs2_orphan_scan_timeout());
	}
}

struct ocfs2_orphan_filldir_priv {
	struct inode		*head;
	struct ocfs2_super	*osb;
};

static int ocfs2_orphan_filldir(void *priv, const char *name, int name_len,
				loff_t pos, u64 ino, unsigned type)
{
	struct ocfs2_orphan_filldir_priv *p = priv;
	struct inode *iter;

	if (name_len == 1 && !strncmp(".", name, 1))
		return 0;
	if (name_len == 2 && !strncmp("..", name, 2))
		return 0;

	/* Skip bad inodes so that recovery can continue */
	iter = ocfs2_iget(p->osb, ino,
			  OCFS2_FI_FLAG_ORPHAN_RECOVERY, 0);
	if (IS_ERR(iter))
		return 0;

	mlog(0, "queue orphan %llu\n",
	     (unsigned long long)OCFS2_I(iter)->ip_blkno);
	/* No locking is required for the next_orphan queue as there
	 * is only ever a single process doing orphan recovery. */
	OCFS2_I(iter)->ip_next_orphan = p->head;
	p->head = iter;

	return 0;
}

static int ocfs2_queue_orphans(struct ocfs2_super *osb,
			       int slot,
			       struct inode **head)
{
	int status;
	struct inode *orphan_dir_inode = NULL;
	struct ocfs2_orphan_filldir_priv priv;
	loff_t pos = 0;

	priv.osb = osb;
	priv.head = *head;

	orphan_dir_inode = ocfs2_get_system_file_inode(osb,
						       ORPHAN_DIR_SYSTEM_INODE,
						       slot);
	if  (!orphan_dir_inode) {
		status = -ENOENT;
		mlog_errno(status);
		return status;
	}	

	mutex_lock(&orphan_dir_inode->i_mutex);
	status = ocfs2_inode_lock(orphan_dir_inode, NULL, 0);
	if (status < 0) {
		mlog_errno(status);
		goto out;
	}

	status = ocfs2_dir_foreach(orphan_dir_inode, &pos, &priv,
				   ocfs2_orphan_filldir);
	if (status) {
		mlog_errno(status);
		goto out_cluster;
	}

	*head = priv.head;

out_cluster:
	ocfs2_inode_unlock(orphan_dir_inode, 0);
out:
	mutex_unlock(&orphan_dir_inode->i_mutex);
	iput(orphan_dir_inode);
	return status;
}

static int ocfs2_orphan_recovery_can_continue(struct ocfs2_super *osb,
					      int slot)
{
	int ret;

	spin_lock(&osb->osb_lock);
	ret = !osb->osb_orphan_wipes[slot];
	spin_unlock(&osb->osb_lock);
	return ret;
}

static void ocfs2_mark_recovering_orphan_dir(struct ocfs2_super *osb,
					     int slot)
{
	spin_lock(&osb->osb_lock);
	/* Mark ourselves such that new processes in delete_inode()
	 * know to quit early. */
	ocfs2_node_map_set_bit(osb, &osb->osb_recovering_orphan_dirs, slot);
	while (osb->osb_orphan_wipes[slot]) {
		/* If any processes are already in the middle of an
		 * orphan wipe on this dir, then we need to wait for
		 * them. */
		spin_unlock(&osb->osb_lock);
		wait_event_interruptible(osb->osb_wipe_event,
					 ocfs2_orphan_recovery_can_continue(osb, slot));
		spin_lock(&osb->osb_lock);
	}
	spin_unlock(&osb->osb_lock);
}

static void ocfs2_clear_recovering_orphan_dir(struct ocfs2_super *osb,
					      int slot)
{
	ocfs2_node_map_clear_bit(osb, &osb->osb_recovering_orphan_dirs, slot);
}

/*
 * Orphan recovery. Each mounted node has it's own orphan dir which we
 * must run during recovery. Our strategy here is to build a list of
 * the inodes in the orphan dir and iget/iput them. The VFS does
 * (most) of the rest of the work.
 *
 * Orphan recovery can happen at any time, not just mount so we have a
 * couple of extra considerations.
 *
 * - We grab as many inodes as we can under the orphan dir lock -
 *   doing iget() outside the orphan dir risks getting a reference on
 *   an invalid inode.
 * - We must be sure not to deadlock with other processes on the
 *   system wanting to run delete_inode(). This can happen when they go
 *   to lock the orphan dir and the orphan recovery process attempts to
 *   iget() inside the orphan dir lock. This can be avoided by
 *   advertising our state to ocfs2_delete_inode().
 */
static int ocfs2_recover_orphans(struct ocfs2_super *osb,
				 int slot)
{
	int ret = 0;
	struct inode *inode = NULL;
	struct inode *iter;
	struct ocfs2_inode_info *oi;

	mlog(0, "Recover inodes from orphan dir in slot %d\n", slot);

	ocfs2_mark_recovering_orphan_dir(osb, slot);
	ret = ocfs2_queue_orphans(osb, slot, &inode);
	ocfs2_clear_recovering_orphan_dir(osb, slot);

	/* Error here should be noted, but we want to continue with as
	 * many queued inodes as we've got. */
	if (ret)
		mlog_errno(ret);

	while (inode) {
		oi = OCFS2_I(inode);
		mlog(0, "iput orphan %llu\n", (unsigned long long)oi->ip_blkno);

		iter = oi->ip_next_orphan;

		spin_lock(&oi->ip_lock);
		/* The remote delete code may have set these on the
		 * assumption that the other node would wipe them
		 * successfully.  If they are still in the node's
		 * orphan dir, we need to reset that state. */
		oi->ip_flags &= ~(OCFS2_INODE_DELETED|OCFS2_INODE_SKIP_DELETE);

		/* Set the proper information to get us going into
		 * ocfs2_delete_inode. */
		oi->ip_flags |= OCFS2_INODE_MAYBE_ORPHANED;
		spin_unlock(&oi->ip_lock);

		iput(inode);

		inode = iter;
	}

	return ret;
}

static int __ocfs2_wait_on_mount(struct ocfs2_super *osb, int quota)
{
	/* This check is good because ocfs2 will wait on our recovery
	 * thread before changing it to something other than MOUNTED
	 * or DISABLED. */
	wait_event(osb->osb_mount_event,
		  (!quota && atomic_read(&osb->vol_state) == VOLUME_MOUNTED) ||
		   atomic_read(&osb->vol_state) == VOLUME_MOUNTED_QUOTAS ||
		   atomic_read(&osb->vol_state) == VOLUME_DISABLED);

	/* If there's an error on mount, then we may never get to the
	 * MOUNTED flag, but this is set right before
	 * dismount_volume() so we can trust it. */
	if (atomic_read(&osb->vol_state) == VOLUME_DISABLED) {
		mlog(0, "mount error, exiting!\n");
		return -EBUSY;
	}

	return 0;
}

static int ocfs2_commit_thread(void *arg)
{
	int status;
	struct ocfs2_super *osb = arg;
	struct ocfs2_journal *journal = osb->journal;

	/* we can trust j_num_trans here because _should_stop() is only set in
	 * shutdown and nobody other than ourselves should be able to start
	 * transactions.  committing on shutdown might take a few iterations
	 * as final transactions put deleted inodes on the list */
	while (!(kthread_should_stop() &&
		 atomic_read(&journal->j_num_trans) == 0)) {

		wait_event_interruptible(osb->checkpoint_event,
					 atomic_read(&journal->j_num_trans)
					 || kthread_should_stop());

		status = ocfs2_commit_cache(osb);
		if (status < 0)
			mlog_errno(status);

		if (kthread_should_stop() && atomic_read(&journal->j_num_trans)){
			mlog(ML_KTHREAD,
			     "commit_thread: %u transactions pending on "
			     "shutdown\n",
			     atomic_read(&journal->j_num_trans));
		}
	}

	return 0;
}

/* Reads all the journal inodes without taking any cluster locks. Used
 * for hard readonly access to determine whether any journal requires
 * recovery. Also used to refresh the recovery generation numbers after
 * a journal has been recovered by another node.
 */
int ocfs2_check_journals_nolocks(struct ocfs2_super *osb)
{
	int ret = 0;
	unsigned int slot;
	struct buffer_head *di_bh = NULL;
	struct ocfs2_dinode *di;
	int journal_dirty = 0;

	for(slot = 0; slot < osb->max_slots; slot++) {
		ret = ocfs2_read_journal_inode(osb, slot, &di_bh, NULL);
		if (ret) {
			mlog_errno(ret);
			goto out;
		}

		di = (struct ocfs2_dinode *) di_bh->b_data;

		osb->slot_recovery_generations[slot] =
					ocfs2_get_recovery_generation(di);

		if (le32_to_cpu(di->id1.journal1.ij_flags) &
		    OCFS2_JOURNAL_DIRTY_FL)
			journal_dirty = 1;

		brelse(di_bh);
		di_bh = NULL;
	}

out:
	if (journal_dirty)
		ret = -EROFS;
	return ret;
}