summaryrefslogtreecommitdiff
path: root/fs/xfs/quota/xfs_dquot.c
blob: 4d6d051ee56e4a8ac397850927640782ebaef5c6 (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
/*
 * Copyright (c) 2000-2003 Silicon Graphics, Inc.
 * 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.
 *
 * This program is distributed in the hope that it would 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 the Free Software Foundation,
 * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 */
#include "xfs.h"
#include "xfs_fs.h"
#include "xfs_bit.h"
#include "xfs_log.h"
#include "xfs_inum.h"
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
#include "xfs_alloc.h"
#include "xfs_dmapi.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
#include "xfs_dir2_sf.h"
#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_btree.h"
#include "xfs_ialloc.h"
#include "xfs_bmap.h"
#include "xfs_rtalloc.h"
#include "xfs_error.h"
#include "xfs_itable.h"
#include "xfs_rw.h"
#include "xfs_acl.h"
#include "xfs_attr.h"
#include "xfs_buf_item.h"
#include "xfs_trans_space.h"
#include "xfs_trans_priv.h"
#include "xfs_qm.h"


/*
   LOCK ORDER

   inode lock		    (ilock)
   dquot hash-chain lock    (hashlock)
   xqm dquot freelist lock  (freelistlock
   mount's dquot list lock  (mplistlock)
   user dquot lock - lock ordering among dquots is based on the uid or gid
   group dquot lock - similar to udquots. Between the two dquots, the udquot
		      has to be locked first.
   pin lock - the dquot lock must be held to take this lock.
   flush lock - ditto.
*/

STATIC void		xfs_qm_dqflush_done(xfs_buf_t *, xfs_dq_logitem_t *);

#ifdef DEBUG
xfs_buftarg_t *xfs_dqerror_target;
int xfs_do_dqerror;
int xfs_dqreq_num;
int xfs_dqerror_mod = 33;
#endif

static struct lock_class_key xfs_dquot_other_class;

/*
 * Allocate and initialize a dquot. We don't always allocate fresh memory;
 * we try to reclaim a free dquot if the number of incore dquots are above
 * a threshold.
 * The only field inside the core that gets initialized at this point
 * is the d_id field. The idea is to fill in the entire q_core
 * when we read in the on disk dquot.
 */
STATIC xfs_dquot_t *
xfs_qm_dqinit(
	xfs_mount_t  *mp,
	xfs_dqid_t   id,
	uint	     type)
{
	xfs_dquot_t	*dqp;
	boolean_t	brandnewdquot;

	brandnewdquot = xfs_qm_dqalloc_incore(&dqp);
	dqp->dq_flags = type;
	dqp->q_core.d_id = cpu_to_be32(id);
	dqp->q_mount = mp;

	/*
	 * No need to re-initialize these if this is a reclaimed dquot.
	 */
	if (brandnewdquot) {
		dqp->dq_flnext = dqp->dq_flprev = dqp;
		mutex_init(&dqp->q_qlock);
		init_waitqueue_head(&dqp->q_pinwait);

		/*
		 * Because we want to use a counting completion, complete
		 * the flush completion once to allow a single access to
		 * the flush completion without blocking.
		 */
		init_completion(&dqp->q_flush);
		complete(&dqp->q_flush);

#ifdef XFS_DQUOT_TRACE
		dqp->q_trace = ktrace_alloc(DQUOT_TRACE_SIZE, KM_NOFS);
		xfs_dqtrace_entry(dqp, "DQINIT");
#endif
	} else {
		/*
		 * Only the q_core portion was zeroed in dqreclaim_one().
		 * So, we need to reset others.
		 */
		 dqp->q_nrefs = 0;
		 dqp->q_blkno = 0;
		 dqp->MPL_NEXT = dqp->HL_NEXT = NULL;
		 dqp->HL_PREVP = dqp->MPL_PREVP = NULL;
		 dqp->q_bufoffset = 0;
		 dqp->q_fileoffset = 0;
		 dqp->q_transp = NULL;
		 dqp->q_gdquot = NULL;
		 dqp->q_res_bcount = 0;
		 dqp->q_res_icount = 0;
		 dqp->q_res_rtbcount = 0;
		 atomic_set(&dqp->q_pincount, 0);
		 dqp->q_hash = NULL;
		 ASSERT(dqp->dq_flnext == dqp->dq_flprev);

#ifdef XFS_DQUOT_TRACE
		 ASSERT(dqp->q_trace);
		 xfs_dqtrace_entry(dqp, "DQRECLAIMED_INIT");
#endif
	}

	/*
	 * In either case we need to make sure group quotas have a different
	 * lock class than user quotas, to make sure lockdep knows we can
	 * locks of one of each at the same time.
	 */
	if (!(type & XFS_DQ_USER))
		lockdep_set_class(&dqp->q_qlock, &xfs_dquot_other_class);

	/*
	 * log item gets initialized later
	 */
	return (dqp);
}

/*
 * This is called to free all the memory associated with a dquot
 */
void
xfs_qm_dqdestroy(
	xfs_dquot_t	*dqp)
{
	ASSERT(! XFS_DQ_IS_ON_FREELIST(dqp));

	mutex_destroy(&dqp->q_qlock);
	sv_destroy(&dqp->q_pinwait);

#ifdef XFS_DQUOT_TRACE
	if (dqp->q_trace)
	     ktrace_free(dqp->q_trace);
	dqp->q_trace = NULL;
#endif
	kmem_zone_free(xfs_Gqm->qm_dqzone, dqp);
	atomic_dec(&xfs_Gqm->qm_totaldquots);
}

/*
 * This is what a 'fresh' dquot inside a dquot chunk looks like on disk.
 */
STATIC void
xfs_qm_dqinit_core(
	xfs_dqid_t	id,
	uint		type,
	xfs_dqblk_t	*d)
{
	/*
	 * Caller has zero'd the entire dquot 'chunk' already.
	 */
	d->dd_diskdq.d_magic = cpu_to_be16(XFS_DQUOT_MAGIC);
	d->dd_diskdq.d_version = XFS_DQUOT_VERSION;
	d->dd_diskdq.d_id = cpu_to_be32(id);
	d->dd_diskdq.d_flags = type;
}


#ifdef XFS_DQUOT_TRACE
/*
 * Dquot tracing for debugging.
 */
/* ARGSUSED */
void
__xfs_dqtrace_entry(
	xfs_dquot_t	*dqp,
	char		*func,
	void		*retaddr,
	xfs_inode_t	*ip)
{
	xfs_dquot_t	*udqp = NULL;
	xfs_ino_t	ino = 0;

	ASSERT(dqp->q_trace);
	if (ip) {
		ino = ip->i_ino;
		udqp = ip->i_udquot;
	}
	ktrace_enter(dqp->q_trace,
		     (void *)(__psint_t)DQUOT_KTRACE_ENTRY,
		     (void *)func,
		     (void *)(__psint_t)dqp->q_nrefs,
		     (void *)(__psint_t)dqp->dq_flags,
		     (void *)(__psint_t)dqp->q_res_bcount,
		     (void *)(__psint_t)be64_to_cpu(dqp->q_core.d_bcount),
		     (void *)(__psint_t)be64_to_cpu(dqp->q_core.d_icount),
		     (void *)(__psint_t)be64_to_cpu(dqp->q_core.d_blk_hardlimit),
		     (void *)(__psint_t)be64_to_cpu(dqp->q_core.d_blk_softlimit),
		     (void *)(__psint_t)be64_to_cpu(dqp->q_core.d_ino_hardlimit),
		     (void *)(__psint_t)be64_to_cpu(dqp->q_core.d_ino_softlimit),
		     (void *)(__psint_t)be32_to_cpu(dqp->q_core.d_id),
		     (void *)(__psint_t)current_pid(),
		     (void *)(__psint_t)ino,
		     (void *)(__psint_t)retaddr,
		     (void *)(__psint_t)udqp);
	return;
}
#endif


/*
 * If default limits are in force, push them into the dquot now.
 * We overwrite the dquot limits only if they are zero and this
 * is not the root dquot.
 */
void
xfs_qm_adjust_dqlimits(
	xfs_mount_t		*mp,
	xfs_disk_dquot_t	*d)
{
	xfs_quotainfo_t		*q = mp->m_quotainfo;

	ASSERT(d->d_id);

	if (q->qi_bsoftlimit && !d->d_blk_softlimit)
		d->d_blk_softlimit = cpu_to_be64(q->qi_bsoftlimit);
	if (q->qi_bhardlimit && !d->d_blk_hardlimit)
		d->d_blk_hardlimit = cpu_to_be64(q->qi_bhardlimit);
	if (q->qi_isoftlimit && !d->d_ino_softlimit)
		d->d_ino_softlimit = cpu_to_be64(q->qi_isoftlimit);
	if (q->qi_ihardlimit && !d->d_ino_hardlimit)
		d->d_ino_hardlimit = cpu_to_be64(q->qi_ihardlimit);
	if (q->qi_rtbsoftlimit && !d->d_rtb_softlimit)
		d->d_rtb_softlimit = cpu_to_be64(q->qi_rtbsoftlimit);
	if (q->qi_rtbhardlimit && !d->d_rtb_hardlimit)
		d->d_rtb_hardlimit = cpu_to_be64(q->qi_rtbhardlimit);
}

/*
 * Check the limits and timers of a dquot and start or reset timers
 * if necessary.
 * This gets called even when quota enforcement is OFF, which makes our
 * life a little less complicated. (We just don't reject any quota
 * reservations in that case, when enforcement is off).
 * We also return 0 as the values of the timers in Q_GETQUOTA calls, when
 * enforcement's off.
 * In contrast, warnings are a little different in that they don't
 * 'automatically' get started when limits get exceeded.  They do
 * get reset to zero, however, when we find the count to be under
 * the soft limit (they are only ever set non-zero via userspace).
 */
void
xfs_qm_adjust_dqtimers(
	xfs_mount_t		*mp,
	xfs_disk_dquot_t	*d)
{
	ASSERT(d->d_id);

#ifdef QUOTADEBUG
	if (d->d_blk_hardlimit)
		ASSERT(be64_to_cpu(d->d_blk_softlimit) <=
		       be64_to_cpu(d->d_blk_hardlimit));
	if (d->d_ino_hardlimit)
		ASSERT(be64_to_cpu(d->d_ino_softlimit) <=
		       be64_to_cpu(d->d_ino_hardlimit));
	if (d->d_rtb_hardlimit)
		ASSERT(be64_to_cpu(d->d_rtb_softlimit) <=
		       be64_to_cpu(d->d_rtb_hardlimit));
#endif
	if (!d->d_btimer) {
		if ((d->d_blk_softlimit &&
		     (be64_to_cpu(d->d_bcount) >=
		      be64_to_cpu(d->d_blk_softlimit))) ||
		    (d->d_blk_hardlimit &&
		     (be64_to_cpu(d->d_bcount) >=
		      be64_to_cpu(d->d_blk_hardlimit)))) {
			d->d_btimer = cpu_to_be32(get_seconds() +
					XFS_QI_BTIMELIMIT(mp));
		} else {
			d->d_bwarns = 0;
		}
	} else {
		if ((!d->d_blk_softlimit ||
		     (be64_to_cpu(d->d_bcount) <
		      be64_to_cpu(d->d_blk_softlimit))) &&
		    (!d->d_blk_hardlimit ||
		    (be64_to_cpu(d->d_bcount) <
		     be64_to_cpu(d->d_blk_hardlimit)))) {
			d->d_btimer = 0;
		}
	}

	if (!d->d_itimer) {
		if ((d->d_ino_softlimit &&
		     (be64_to_cpu(d->d_icount) >=
		      be64_to_cpu(d->d_ino_softlimit))) ||
		    (d->d_ino_hardlimit &&
		     (be64_to_cpu(d->d_icount) >=
		      be64_to_cpu(d->d_ino_hardlimit)))) {
			d->d_itimer = cpu_to_be32(get_seconds() +
					XFS_QI_ITIMELIMIT(mp));
		} else {
			d->d_iwarns = 0;
		}
	} else {
		if ((!d->d_ino_softlimit ||
		     (be64_to_cpu(d->d_icount) <
		      be64_to_cpu(d->d_ino_softlimit)))  &&
		    (!d->d_ino_hardlimit ||
		     (be64_to_cpu(d->d_icount) <
		      be64_to_cpu(d->d_ino_hardlimit)))) {
			d->d_itimer = 0;
		}
	}

	if (!d->d_rtbtimer) {
		if ((d->d_rtb_softlimit &&
		     (be64_to_cpu(d->d_rtbcount) >=
		      be64_to_cpu(d->d_rtb_softlimit))) ||
		    (d->d_rtb_hardlimit &&
		     (be64_to_cpu(d->d_rtbcount) >=
		      be64_to_cpu(d->d_rtb_hardlimit)))) {
			d->d_rtbtimer = cpu_to_be32(get_seconds() +
					XFS_QI_RTBTIMELIMIT(mp));
		} else {
			d->d_rtbwarns = 0;
		}
	} else {
		if ((!d->d_rtb_softlimit ||
		     (be64_to_cpu(d->d_rtbcount) <
		      be64_to_cpu(d->d_rtb_softlimit))) &&
		    (!d->d_rtb_hardlimit ||
		     (be64_to_cpu(d->d_rtbcount) <
		      be64_to_cpu(d->d_rtb_hardlimit)))) {
			d->d_rtbtimer = 0;
		}
	}
}

/*
 * initialize a buffer full of dquots and log the whole thing
 */
STATIC void
xfs_qm_init_dquot_blk(
	xfs_trans_t	*tp,
	xfs_mount_t	*mp,
	xfs_dqid_t	id,
	uint		type,
	xfs_buf_t	*bp)
{
	xfs_dqblk_t	*d;
	int		curid, i;

	ASSERT(tp);
	ASSERT(XFS_BUF_ISBUSY(bp));
	ASSERT(XFS_BUF_VALUSEMA(bp) <= 0);

	d = (xfs_dqblk_t *)XFS_BUF_PTR(bp);

	/*
	 * ID of the first dquot in the block - id's are zero based.
	 */
	curid = id - (id % XFS_QM_DQPERBLK(mp));
	ASSERT(curid >= 0);
	memset(d, 0, BBTOB(XFS_QI_DQCHUNKLEN(mp)));
	for (i = 0; i < XFS_QM_DQPERBLK(mp); i++, d++, curid++)
		xfs_qm_dqinit_core(curid, type, d);
	xfs_trans_dquot_buf(tp, bp,
			    (type & XFS_DQ_USER ? XFS_BLI_UDQUOT_BUF :
			    ((type & XFS_DQ_PROJ) ? XFS_BLI_PDQUOT_BUF :
			     XFS_BLI_GDQUOT_BUF)));
	xfs_trans_log_buf(tp, bp, 0, BBTOB(XFS_QI_DQCHUNKLEN(mp)) - 1);
}



/*
 * Allocate a block and fill it with dquots.
 * This is called when the bmapi finds a hole.
 */
STATIC int
xfs_qm_dqalloc(
	xfs_trans_t	**tpp,
	xfs_mount_t	*mp,
	xfs_dquot_t	*dqp,
	xfs_inode_t	*quotip,
	xfs_fileoff_t	offset_fsb,
	xfs_buf_t	**O_bpp)
{
	xfs_fsblock_t	firstblock;
	xfs_bmap_free_t flist;
	xfs_bmbt_irec_t map;
	int		nmaps, error, committed;
	xfs_buf_t	*bp;
	xfs_trans_t	*tp = *tpp;

	ASSERT(tp != NULL);
	xfs_dqtrace_entry(dqp, "DQALLOC");

	/*
	 * Initialize the bmap freelist prior to calling bmapi code.
	 */
	xfs_bmap_init(&flist, &firstblock);
	xfs_ilock(quotip, XFS_ILOCK_EXCL);
	/*
	 * Return if this type of quotas is turned off while we didn't
	 * have an inode lock
	 */
	if (XFS_IS_THIS_QUOTA_OFF(dqp)) {
		xfs_iunlock(quotip, XFS_ILOCK_EXCL);
		return (ESRCH);
	}

	/*
	 * xfs_trans_commit normally decrements the vnode ref count
	 * when it unlocks the inode. Since we want to keep the quota
	 * inode around, we bump the vnode ref count now.
	 */
	IHOLD(quotip);

	xfs_trans_ijoin(tp, quotip, XFS_ILOCK_EXCL);
	nmaps = 1;
	if ((error = xfs_bmapi(tp, quotip,
			      offset_fsb, XFS_DQUOT_CLUSTER_SIZE_FSB,
			      XFS_BMAPI_METADATA | XFS_BMAPI_WRITE,
			      &firstblock,
			      XFS_QM_DQALLOC_SPACE_RES(mp),
			      &map, &nmaps, &flist, NULL))) {
		goto error0;
	}
	ASSERT(map.br_blockcount == XFS_DQUOT_CLUSTER_SIZE_FSB);
	ASSERT(nmaps == 1);
	ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
	       (map.br_startblock != HOLESTARTBLOCK));

	/*
	 * Keep track of the blkno to save a lookup later
	 */
	dqp->q_blkno = XFS_FSB_TO_DADDR(mp, map.br_startblock);

	/* now we can just get the buffer (there's nothing to read yet) */
	bp = xfs_trans_get_buf(tp, mp->m_ddev_targp,
			       dqp->q_blkno,
			       XFS_QI_DQCHUNKLEN(mp),
			       0);
	if (!bp || (error = XFS_BUF_GETERROR(bp)))
		goto error1;
	/*
	 * Make a chunk of dquots out of this buffer and log
	 * the entire thing.
	 */
	xfs_qm_init_dquot_blk(tp, mp, be32_to_cpu(dqp->q_core.d_id),
			      dqp->dq_flags & XFS_DQ_ALLTYPES, bp);

	/*
	 * xfs_bmap_finish() may commit the current transaction and
	 * start a second transaction if the freelist is not empty.
	 *
	 * Since we still want to modify this buffer, we need to
	 * ensure that the buffer is not released on commit of
	 * the first transaction and ensure the buffer is added to the
	 * second transaction.
	 *
	 * If there is only one transaction then don't stop the buffer
	 * from being released when it commits later on.
	 */

	xfs_trans_bhold(tp, bp);

	if ((error = xfs_bmap_finish(tpp, &flist, &committed))) {
		goto error1;
	}

	if (committed) {
		tp = *tpp;
		xfs_trans_bjoin(tp, bp);
	} else {
		xfs_trans_bhold_release(tp, bp);
	}

	*O_bpp = bp;
	return 0;

      error1:
	xfs_bmap_cancel(&flist);
      error0:
	xfs_iunlock(quotip, XFS_ILOCK_EXCL);

	return (error);
}

/*
 * Maps a dquot to the buffer containing its on-disk version.
 * This returns a ptr to the buffer containing the on-disk dquot
 * in the bpp param, and a ptr to the on-disk dquot within that buffer
 */
STATIC int
xfs_qm_dqtobp(
	xfs_trans_t		**tpp,
	xfs_dquot_t		*dqp,
	xfs_disk_dquot_t	**O_ddpp,
	xfs_buf_t		**O_bpp,
	uint			flags)
{
	xfs_bmbt_irec_t map;
	int		nmaps, error;
	xfs_buf_t	*bp;
	xfs_inode_t	*quotip;
	xfs_mount_t	*mp;
	xfs_disk_dquot_t *ddq;
	xfs_dqid_t	id;
	boolean_t	newdquot;
	xfs_trans_t	*tp = (tpp ? *tpp : NULL);

	mp = dqp->q_mount;
	id = be32_to_cpu(dqp->q_core.d_id);
	nmaps = 1;
	newdquot = B_FALSE;

	/*
	 * If we don't know where the dquot lives, find out.
	 */
	if (dqp->q_blkno == (xfs_daddr_t) 0) {
		/* We use the id as an index */
		dqp->q_fileoffset = (xfs_fileoff_t)id / XFS_QM_DQPERBLK(mp);
		nmaps = 1;
		quotip = XFS_DQ_TO_QIP(dqp);
		xfs_ilock(quotip, XFS_ILOCK_SHARED);
		/*
		 * Return if this type of quotas is turned off while we didn't
		 * have an inode lock
		 */
		if (XFS_IS_THIS_QUOTA_OFF(dqp)) {
			xfs_iunlock(quotip, XFS_ILOCK_SHARED);
			return (ESRCH);
		}
		/*
		 * Find the block map; no allocations yet
		 */
		error = xfs_bmapi(NULL, quotip, dqp->q_fileoffset,
				  XFS_DQUOT_CLUSTER_SIZE_FSB,
				  XFS_BMAPI_METADATA,
				  NULL, 0, &map, &nmaps, NULL, NULL);

		xfs_iunlock(quotip, XFS_ILOCK_SHARED);
		if (error)
			return (error);
		ASSERT(nmaps == 1);
		ASSERT(map.br_blockcount == 1);

		/*
		 * offset of dquot in the (fixed sized) dquot chunk.
		 */
		dqp->q_bufoffset = (id % XFS_QM_DQPERBLK(mp)) *
			sizeof(xfs_dqblk_t);
		if (map.br_startblock == HOLESTARTBLOCK) {
			/*
			 * We don't allocate unless we're asked to
			 */
			if (!(flags & XFS_QMOPT_DQALLOC))
				return (ENOENT);

			ASSERT(tp);
			if ((error = xfs_qm_dqalloc(tpp, mp, dqp, quotip,
						dqp->q_fileoffset, &bp)))
				return (error);
			tp = *tpp;
			newdquot = B_TRUE;
		} else {
			/*
			 * store the blkno etc so that we don't have to do the
			 * mapping all the time
			 */
			dqp->q_blkno = XFS_FSB_TO_DADDR(mp, map.br_startblock);
		}
	}
	ASSERT(dqp->q_blkno != DELAYSTARTBLOCK);
	ASSERT(dqp->q_blkno != HOLESTARTBLOCK);

	/*
	 * Read in the buffer, unless we've just done the allocation
	 * (in which case we already have the buf).
	 */
	if (! newdquot) {
		xfs_dqtrace_entry(dqp, "DQTOBP READBUF");
		if ((error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp,
					       dqp->q_blkno,
					       XFS_QI_DQCHUNKLEN(mp),
					       0, &bp))) {
			return (error);
		}
		if (error || !bp)
			return XFS_ERROR(error);
	}
	ASSERT(XFS_BUF_ISBUSY(bp));
	ASSERT(XFS_BUF_VALUSEMA(bp) <= 0);

	/*
	 * calculate the location of the dquot inside the buffer.
	 */
	ddq = (xfs_disk_dquot_t *)((char *)XFS_BUF_PTR(bp) + dqp->q_bufoffset);

	/*
	 * A simple sanity check in case we got a corrupted dquot...
	 */
	if (xfs_qm_dqcheck(ddq, id, dqp->dq_flags & XFS_DQ_ALLTYPES,
			   flags & (XFS_QMOPT_DQREPAIR|XFS_QMOPT_DOWARN),
			   "dqtobp")) {
		if (!(flags & XFS_QMOPT_DQREPAIR)) {
			xfs_trans_brelse(tp, bp);
			return XFS_ERROR(EIO);
		}
		XFS_BUF_BUSY(bp); /* We dirtied this */
	}

	*O_bpp = bp;
	*O_ddpp = ddq;

	return (0);
}


/*
 * Read in the ondisk dquot using dqtobp() then copy it to an incore version,
 * and release the buffer immediately.
 *
 */
/* ARGSUSED */
STATIC int
xfs_qm_dqread(
	xfs_trans_t	**tpp,
	xfs_dqid_t	id,
	xfs_dquot_t	*dqp,	/* dquot to get filled in */
	uint		flags)
{
	xfs_disk_dquot_t *ddqp;
	xfs_buf_t	 *bp;
	int		 error;
	xfs_trans_t	 *tp;

	ASSERT(tpp);

	/*
	 * get a pointer to the on-disk dquot and the buffer containing it
	 * dqp already knows its own type (GROUP/USER).
	 */
	xfs_dqtrace_entry(dqp, "DQREAD");
	if ((error = xfs_qm_dqtobp(tpp, dqp, &ddqp, &bp, flags))) {
		return (error);
	}
	tp = *tpp;

	/* copy everything from disk dquot to the incore dquot */
	memcpy(&dqp->q_core, ddqp, sizeof(xfs_disk_dquot_t));
	ASSERT(be32_to_cpu(dqp->q_core.d_id) == id);
	xfs_qm_dquot_logitem_init(dqp);

	/*
	 * Reservation counters are defined as reservation plus current usage
	 * to avoid having to add everytime.
	 */
	dqp->q_res_bcount = be64_to_cpu(ddqp->d_bcount);
	dqp->q_res_icount = be64_to_cpu(ddqp->d_icount);
	dqp->q_res_rtbcount = be64_to_cpu(ddqp->d_rtbcount);

	/* Mark the buf so that this will stay incore a little longer */
	XFS_BUF_SET_VTYPE_REF(bp, B_FS_DQUOT, XFS_DQUOT_REF);

	/*
	 * We got the buffer with a xfs_trans_read_buf() (in dqtobp())
	 * So we need to release with xfs_trans_brelse().
	 * The strategy here is identical to that of inodes; we lock
	 * the dquot in xfs_qm_dqget() before making it accessible to
	 * others. This is because dquots, like inodes, need a good level of
	 * concurrency, and we don't want to take locks on the entire buffers
	 * for dquot accesses.
	 * Note also that the dquot buffer may even be dirty at this point, if
	 * this particular dquot was repaired. We still aren't afraid to
	 * brelse it because we have the changes incore.
	 */
	ASSERT(XFS_BUF_ISBUSY(bp));
	ASSERT(XFS_BUF_VALUSEMA(bp) <= 0);
	xfs_trans_brelse(tp, bp);

	return (error);
}


/*
 * allocate an incore dquot from the kernel heap,
 * and fill its core with quota information kept on disk.
 * If XFS_QMOPT_DQALLOC is set, it'll allocate a dquot on disk
 * if it wasn't already allocated.
 */
STATIC int
xfs_qm_idtodq(
	xfs_mount_t	*mp,
	xfs_dqid_t	id,	 /* gid or uid, depending on type */
	uint		type,	 /* UDQUOT or GDQUOT */
	uint		flags,	 /* DQALLOC, DQREPAIR */
	xfs_dquot_t	**O_dqpp)/* OUT : incore dquot, not locked */
{
	xfs_dquot_t	*dqp;
	int		error;
	xfs_trans_t	*tp;
	int		cancelflags=0;

	dqp = xfs_qm_dqinit(mp, id, type);
	tp = NULL;
	if (flags & XFS_QMOPT_DQALLOC) {
		tp = xfs_trans_alloc(mp, XFS_TRANS_QM_DQALLOC);
		if ((error = xfs_trans_reserve(tp,
				       XFS_QM_DQALLOC_SPACE_RES(mp),
				       XFS_WRITE_LOG_RES(mp) +
					      BBTOB(XFS_QI_DQCHUNKLEN(mp)) - 1 +
					      128,
				       0,
				       XFS_TRANS_PERM_LOG_RES,
				       XFS_WRITE_LOG_COUNT))) {
			cancelflags = 0;
			goto error0;
		}
		cancelflags = XFS_TRANS_RELEASE_LOG_RES;
	}

	/*
	 * Read it from disk; xfs_dqread() takes care of
	 * all the necessary initialization of dquot's fields (locks, etc)
	 */
	if ((error = xfs_qm_dqread(&tp, id, dqp, flags))) {
		/*
		 * This can happen if quotas got turned off (ESRCH),
		 * or if the dquot didn't exist on disk and we ask to
		 * allocate (ENOENT).
		 */
		xfs_dqtrace_entry(dqp, "DQREAD FAIL");
		cancelflags |= XFS_TRANS_ABORT;
		goto error0;
	}
	if (tp) {
		if ((error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES)))
			goto error1;
	}

	*O_dqpp = dqp;
	return (0);

 error0:
	ASSERT(error);
	if (tp)
		xfs_trans_cancel(tp, cancelflags);
 error1:
	xfs_qm_dqdestroy(dqp);
	*O_dqpp = NULL;
	return (error);
}

/*
 * Lookup a dquot in the incore dquot hashtable. We keep two separate
 * hashtables for user and group dquots; and, these are global tables
 * inside the XQM, not per-filesystem tables.
 * The hash chain must be locked by caller, and it is left locked
 * on return. Returning dquot is locked.
 */
STATIC int
xfs_qm_dqlookup(
	xfs_mount_t		*mp,
	xfs_dqid_t		id,
	xfs_dqhash_t		*qh,
	xfs_dquot_t		**O_dqpp)
{
	xfs_dquot_t		*dqp;
	uint			flist_locked;
	xfs_dquot_t		*d;

	ASSERT(mutex_is_locked(&qh->qh_lock));

	flist_locked = B_FALSE;

	/*
	 * Traverse the hashchain looking for a match
	 */
	for (dqp = qh->qh_next; dqp != NULL; dqp = dqp->HL_NEXT) {
		/*
		 * We already have the hashlock. We don't need the
		 * dqlock to look at the id field of the dquot, since the
		 * id can't be modified without the hashlock anyway.
		 */
		if (be32_to_cpu(dqp->q_core.d_id) == id && dqp->q_mount == mp) {
			xfs_dqtrace_entry(dqp, "DQFOUND BY LOOKUP");
			/*
			 * All in core dquots must be on the dqlist of mp
			 */
			ASSERT(dqp->MPL_PREVP != NULL);

			xfs_dqlock(dqp);
			if (dqp->q_nrefs == 0) {
				ASSERT (XFS_DQ_IS_ON_FREELIST(dqp));
				if (! xfs_qm_freelist_lock_nowait(xfs_Gqm)) {
					xfs_dqtrace_entry(dqp, "DQLOOKUP: WANT");

					/*
					 * We may have raced with dqreclaim_one()
					 * (and lost). So, flag that we don't
					 * want the dquot to be reclaimed.
					 */
					dqp->dq_flags |= XFS_DQ_WANT;
					xfs_dqunlock(dqp);
					xfs_qm_freelist_lock(xfs_Gqm);
					xfs_dqlock(dqp);
					dqp->dq_flags &= ~(XFS_DQ_WANT);
				}
				flist_locked = B_TRUE;
			}

			/*
			 * id couldn't have changed; we had the hashlock all
			 * along
			 */
			ASSERT(be32_to_cpu(dqp->q_core.d_id) == id);

			if (flist_locked) {
				if (dqp->q_nrefs != 0) {
					xfs_qm_freelist_unlock(xfs_Gqm);
					flist_locked = B_FALSE;
				} else {
					/*
					 * take it off the freelist
					 */
					xfs_dqtrace_entry(dqp,
							"DQLOOKUP: TAKEOFF FL");
					XQM_FREELIST_REMOVE(dqp);
					/* xfs_qm_freelist_print(&(xfs_Gqm->
							qm_dqfreelist),
							"after removal"); */
				}
			}

			/*
			 * grab a reference
			 */
			XFS_DQHOLD(dqp);

			if (flist_locked)
				xfs_qm_freelist_unlock(xfs_Gqm);
			/*
			 * move the dquot to the front of the hashchain
			 */
			ASSERT(mutex_is_locked(&qh->qh_lock));
			if (dqp->HL_PREVP != &qh->qh_next) {
				xfs_dqtrace_entry(dqp,
						  "DQLOOKUP: HASH MOVETOFRONT");
				if ((d = dqp->HL_NEXT))
					d->HL_PREVP = dqp->HL_PREVP;
				*(dqp->HL_PREVP) = d;
				d = qh->qh_next;
				d->HL_PREVP = &dqp->HL_NEXT;
				dqp->HL_NEXT = d;
				dqp->HL_PREVP = &qh->qh_next;
				qh->qh_next = dqp;
			}
			xfs_dqtrace_entry(dqp, "LOOKUP END");
			*O_dqpp = dqp;
			ASSERT(mutex_is_locked(&qh->qh_lock));
			return (0);
		}
	}

	*O_dqpp = NULL;
	ASSERT(mutex_is_locked(&qh->qh_lock));
	return (1);
}

/*
 * Given the file system, inode OR id, and type (UDQUOT/GDQUOT), return a
 * a locked dquot, doing an allocation (if requested) as needed.
 * When both an inode and an id are given, the inode's id takes precedence.
 * That is, if the id changes while we don't hold the ilock inside this
 * function, the new dquot is returned, not necessarily the one requested
 * in the id argument.
 */
int
xfs_qm_dqget(
	xfs_mount_t	*mp,
	xfs_inode_t	*ip,	  /* locked inode (optional) */
	xfs_dqid_t	id,	  /* uid/projid/gid depending on type */
	uint		type,	  /* XFS_DQ_USER/XFS_DQ_PROJ/XFS_DQ_GROUP */
	uint		flags,	  /* DQALLOC, DQSUSER, DQREPAIR, DOWARN */
	xfs_dquot_t	**O_dqpp) /* OUT : locked incore dquot */
{
	xfs_dquot_t	*dqp;
	xfs_dqhash_t	*h;
	uint		version;
	int		error;

	ASSERT(XFS_IS_QUOTA_RUNNING(mp));
	if ((! XFS_IS_UQUOTA_ON(mp) && type == XFS_DQ_USER) ||
	    (! XFS_IS_PQUOTA_ON(mp) && type == XFS_DQ_PROJ) ||
	    (! XFS_IS_GQUOTA_ON(mp) && type == XFS_DQ_GROUP)) {
		return (ESRCH);
	}
	h = XFS_DQ_HASH(mp, id, type);

#ifdef DEBUG
	if (xfs_do_dqerror) {
		if ((xfs_dqerror_target == mp->m_ddev_targp) &&
		    (xfs_dqreq_num++ % xfs_dqerror_mod) == 0) {
			cmn_err(CE_DEBUG, "Returning error in dqget");
			return (EIO);
		}
	}
#endif

 again:

#ifdef DEBUG
	ASSERT(type == XFS_DQ_USER ||
	       type == XFS_DQ_PROJ ||
	       type == XFS_DQ_GROUP);
	if (ip) {
		ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
		if (type == XFS_DQ_USER)
			ASSERT(ip->i_udquot == NULL);
		else
			ASSERT(ip->i_gdquot == NULL);
	}
#endif
	mutex_lock(&h->qh_lock);

	/*
	 * Look in the cache (hashtable).
	 * The chain is kept locked during lookup.
	 */
	if (xfs_qm_dqlookup(mp, id, h, O_dqpp) == 0) {
		XQM_STATS_INC(xqmstats.xs_qm_dqcachehits);
		/*
		 * The dquot was found, moved to the front of the chain,
		 * taken off the freelist if it was on it, and locked
		 * at this point. Just unlock the hashchain and return.
		 */
		ASSERT(*O_dqpp);
		ASSERT(XFS_DQ_IS_LOCKED(*O_dqpp));
		mutex_unlock(&h->qh_lock);
		xfs_dqtrace_entry(*O_dqpp, "DQGET DONE (FROM CACHE)");
		return (0);	/* success */
	}
	XQM_STATS_INC(xqmstats.xs_qm_dqcachemisses);

	/*
	 * Dquot cache miss. We don't want to keep the inode lock across
	 * a (potential) disk read. Also we don't want to deal with the lock
	 * ordering between quotainode and this inode. OTOH, dropping the inode
	 * lock here means dealing with a chown that can happen before
	 * we re-acquire the lock.
	 */
	if (ip)
		xfs_iunlock(ip, XFS_ILOCK_EXCL);
	/*
	 * Save the hashchain version stamp, and unlock the chain, so that
	 * we don't keep the lock across a disk read
	 */
	version = h->qh_version;
	mutex_unlock(&h->qh_lock);

	/*
	 * Allocate the dquot on the kernel heap, and read the ondisk
	 * portion off the disk. Also, do all the necessary initialization
	 * This can return ENOENT if dquot didn't exist on disk and we didn't
	 * ask it to allocate; ESRCH if quotas got turned off suddenly.
	 */
	if ((error = xfs_qm_idtodq(mp, id, type,
				  flags & (XFS_QMOPT_DQALLOC|XFS_QMOPT_DQREPAIR|
					   XFS_QMOPT_DOWARN),
				  &dqp))) {
		if (ip)
			xfs_ilock(ip, XFS_ILOCK_EXCL);
		return (error);
	}

	/*
	 * See if this is mount code calling to look at the overall quota limits
	 * which are stored in the id == 0 user or group's dquot.
	 * Since we may not have done a quotacheck by this point, just return
	 * the dquot without attaching it to any hashtables, lists, etc, or even
	 * taking a reference.
	 * The caller must dqdestroy this once done.
	 */
	if (flags & XFS_QMOPT_DQSUSER) {
		ASSERT(id == 0);
		ASSERT(! ip);
		goto dqret;
	}

	/*
	 * Dquot lock comes after hashlock in the lock ordering
	 */
	if (ip) {
		xfs_ilock(ip, XFS_ILOCK_EXCL);
		if (! XFS_IS_DQTYPE_ON(mp, type)) {
			/* inode stays locked on return */
			xfs_qm_dqdestroy(dqp);
			return XFS_ERROR(ESRCH);
		}
		/*
		 * A dquot could be attached to this inode by now, since
		 * we had dropped the ilock.
		 */
		if (type == XFS_DQ_USER) {
			if (ip->i_udquot) {
				xfs_qm_dqdestroy(dqp);
				dqp = ip->i_udquot;
				xfs_dqlock(dqp);
				goto dqret;
			}
		} else {
			if (ip->i_gdquot) {
				xfs_qm_dqdestroy(dqp);
				dqp = ip->i_gdquot;
				xfs_dqlock(dqp);
				goto dqret;
			}
		}
	}

	/*
	 * Hashlock comes after ilock in lock order
	 */
	mutex_lock(&h->qh_lock);
	if (version != h->qh_version) {
		xfs_dquot_t *tmpdqp;
		/*
		 * Now, see if somebody else put the dquot in the
		 * hashtable before us. This can happen because we didn't
		 * keep the hashchain lock. We don't have to worry about
		 * lock order between the two dquots here since dqp isn't
		 * on any findable lists yet.
		 */
		if (xfs_qm_dqlookup(mp, id, h, &tmpdqp) == 0) {
			/*
			 * Duplicate found. Just throw away the new dquot
			 * and start over.
			 */
			xfs_qm_dqput(tmpdqp);
			mutex_unlock(&h->qh_lock);
			xfs_qm_dqdestroy(dqp);
			XQM_STATS_INC(xqmstats.xs_qm_dquot_dups);
			goto again;
		}
	}

	/*
	 * Put the dquot at the beginning of the hash-chain and mp's list
	 * LOCK ORDER: hashlock, freelistlock, mplistlock, udqlock, gdqlock ..
	 */
	ASSERT(mutex_is_locked(&h->qh_lock));
	dqp->q_hash = h;
	XQM_HASHLIST_INSERT(h, dqp);

	/*
	 * Attach this dquot to this filesystem's list of all dquots,
	 * kept inside the mount structure in m_quotainfo field
	 */
	xfs_qm_mplist_lock(mp);

	/*
	 * We return a locked dquot to the caller, with a reference taken
	 */
	xfs_dqlock(dqp);
	dqp->q_nrefs = 1;

	XQM_MPLIST_INSERT(&(XFS_QI_MPL_LIST(mp)), dqp);

	xfs_qm_mplist_unlock(mp);
	mutex_unlock(&h->qh_lock);
 dqret:
	ASSERT((ip == NULL) || xfs_isilocked(ip, XFS_ILOCK_EXCL));
	xfs_dqtrace_entry(dqp, "DQGET DONE");
	*O_dqpp = dqp;
	return (0);
}


/*
 * Release a reference to the dquot (decrement ref-count)
 * and unlock it. If there is a group quota attached to this
 * dquot, carefully release that too without tripping over
 * deadlocks'n'stuff.
 */
void
xfs_qm_dqput(
	xfs_dquot_t	*dqp)
{
	xfs_dquot_t	*gdqp;

	ASSERT(dqp->q_nrefs > 0);
	ASSERT(XFS_DQ_IS_LOCKED(dqp));
	xfs_dqtrace_entry(dqp, "DQPUT");

	if (dqp->q_nrefs != 1) {
		dqp->q_nrefs--;
		xfs_dqunlock(dqp);
		return;
	}

	/*
	 * drop the dqlock and acquire the freelist and dqlock
	 * in the right order; but try to get it out-of-order first
	 */
	if (! xfs_qm_freelist_lock_nowait(xfs_Gqm)) {
		xfs_dqtrace_entry(dqp, "DQPUT: FLLOCK-WAIT");
		xfs_dqunlock(dqp);
		xfs_qm_freelist_lock(xfs_Gqm);
		xfs_dqlock(dqp);
	}

	while (1) {
		gdqp = NULL;

		/* We can't depend on nrefs being == 1 here */
		if (--dqp->q_nrefs == 0) {
			xfs_dqtrace_entry(dqp, "DQPUT: ON FREELIST");
			/*
			 * insert at end of the freelist.
			 */
			XQM_FREELIST_INSERT(&(xfs_Gqm->qm_dqfreelist), dqp);

			/*
			 * If we just added a udquot to the freelist, then
			 * we want to release the gdquot reference that
			 * it (probably) has. Otherwise it'll keep the
			 * gdquot from getting reclaimed.
			 */
			if ((gdqp = dqp->q_gdquot)) {
				/*
				 * Avoid a recursive dqput call
				 */
				xfs_dqlock(gdqp);
				dqp->q_gdquot = NULL;
			}

			/* xfs_qm_freelist_print(&(xfs_Gqm->qm_dqfreelist),
			   "@@@@@++ Free list (after append) @@@@@+");
			   */
		}
		xfs_dqunlock(dqp);

		/*
		 * If we had a group quota inside the user quota as a hint,
		 * release it now.
		 */
		if (! gdqp)
			break;
		dqp = gdqp;
	}
	xfs_qm_freelist_unlock(xfs_Gqm);
}

/*
 * Release a dquot. Flush it if dirty, then dqput() it.
 * dquot must not be locked.
 */
void
xfs_qm_dqrele(
	xfs_dquot_t	*dqp)
{
	if (!dqp)
		return;

	xfs_dqtrace_entry(dqp, "DQRELE");

	xfs_dqlock(dqp);
	/*
	 * We don't care to flush it if the dquot is dirty here.
	 * That will create stutters that we want to avoid.
	 * Instead we do a delayed write when we try to reclaim
	 * a dirty dquot. Also xfs_sync will take part of the burden...
	 */
	xfs_qm_dqput(dqp);
}


/*
 * Write a modified dquot to disk.
 * The dquot must be locked and the flush lock too taken by caller.
 * The flush lock will not be unlocked until the dquot reaches the disk,
 * but the dquot is free to be unlocked and modified by the caller
 * in the interim. Dquot is still locked on return. This behavior is
 * identical to that of inodes.
 */
int
xfs_qm_dqflush(
	xfs_dquot_t		*dqp,
	uint			flags)
{
	xfs_mount_t		*mp;
	xfs_buf_t		*bp;
	xfs_disk_dquot_t	*ddqp;
	int			error;

	ASSERT(XFS_DQ_IS_LOCKED(dqp));
	ASSERT(!completion_done(&dqp->q_flush));
	xfs_dqtrace_entry(dqp, "DQFLUSH");

	/*
	 * If not dirty, or it's pinned and we are not supposed to
	 * block, nada.
	 */
	if (!XFS_DQ_IS_DIRTY(dqp) ||
	    (!(flags & XFS_QMOPT_SYNC) && atomic_read(&dqp->q_pincount) > 0)) {
		xfs_dqfunlock(dqp);
		return 0;
	}
	xfs_qm_dqunpin_wait(dqp);

	/*
	 * This may have been unpinned because the filesystem is shutting
	 * down forcibly. If that's the case we must not write this dquot
	 * to disk, because the log record didn't make it to disk!
	 */
	if (XFS_FORCED_SHUTDOWN(dqp->q_mount)) {
		dqp->dq_flags &= ~(XFS_DQ_DIRTY);
		xfs_dqfunlock(dqp);
		return XFS_ERROR(EIO);
	}

	/*
	 * Get the buffer containing the on-disk dquot
	 * We don't need a transaction envelope because we know that the
	 * the ondisk-dquot has already been allocated for.
	 */
	if ((error = xfs_qm_dqtobp(NULL, dqp, &ddqp, &bp, XFS_QMOPT_DOWARN))) {
		xfs_dqtrace_entry(dqp, "DQTOBP FAIL");
		ASSERT(error != ENOENT);
		/*
		 * Quotas could have gotten turned off (ESRCH)
		 */
		xfs_dqfunlock(dqp);
		return (error);
	}

	if (xfs_qm_dqcheck(&dqp->q_core, be32_to_cpu(ddqp->d_id),
			   0, XFS_QMOPT_DOWARN, "dqflush (incore copy)")) {
		xfs_force_shutdown(dqp->q_mount, SHUTDOWN_CORRUPT_INCORE);
		return XFS_ERROR(EIO);
	}

	/* This is the only portion of data that needs to persist */
	memcpy(ddqp, &(dqp->q_core), sizeof(xfs_disk_dquot_t));

	/*
	 * Clear the dirty field and remember the flush lsn for later use.
	 */
	dqp->dq_flags &= ~(XFS_DQ_DIRTY);
	mp = dqp->q_mount;

	xfs_trans_ail_copy_lsn(mp->m_ail, &dqp->q_logitem.qli_flush_lsn,
					&dqp->q_logitem.qli_item.li_lsn);

	/*
	 * Attach an iodone routine so that we can remove this dquot from the
	 * AIL and release the flush lock once the dquot is synced to disk.
	 */
	xfs_buf_attach_iodone(bp, (void(*)(xfs_buf_t *, xfs_log_item_t *))
			      xfs_qm_dqflush_done, &(dqp->q_logitem.qli_item));
	/*
	 * If the buffer is pinned then push on the log so we won't
	 * get stuck waiting in the write for too long.
	 */
	if (XFS_BUF_ISPINNED(bp)) {
		xfs_dqtrace_entry(dqp, "DQFLUSH LOG FORCE");
		xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE);
	}

	if (flags & XFS_QMOPT_DELWRI) {
		xfs_bdwrite(mp, bp);
	} else if (flags & XFS_QMOPT_ASYNC) {
		error = xfs_bawrite(mp, bp);
	} else {
		error = xfs_bwrite(mp, bp);
	}
	xfs_dqtrace_entry(dqp, "DQFLUSH END");
	/*
	 * dqp is still locked, but caller is free to unlock it now.
	 */
	return (error);

}

/*
 * This is the dquot flushing I/O completion routine.  It is called
 * from interrupt level when the buffer containing the dquot is
 * flushed to disk.  It is responsible for removing the dquot logitem
 * from the AIL if it has not been re-logged, and unlocking the dquot's
 * flush lock. This behavior is very similar to that of inodes..
 */
/*ARGSUSED*/
STATIC void
xfs_qm_dqflush_done(
	xfs_buf_t		*bp,
	xfs_dq_logitem_t	*qip)
{
	xfs_dquot_t		*dqp;
	struct xfs_ail		*ailp;

	dqp = qip->qli_dquot;
	ailp = qip->qli_item.li_ailp;

	/*
	 * We only want to pull the item from the AIL if its
	 * location in the log has not changed since we started the flush.
	 * Thus, we only bother if the dquot's lsn has
	 * not changed. First we check the lsn outside the lock
	 * since it's cheaper, and then we recheck while
	 * holding the lock before removing the dquot from the AIL.
	 */
	if ((qip->qli_item.li_flags & XFS_LI_IN_AIL) &&
	    qip->qli_item.li_lsn == qip->qli_flush_lsn) {

		/* xfs_trans_ail_delete() drops the AIL lock. */
		spin_lock(&ailp->xa_lock);
		if (qip->qli_item.li_lsn == qip->qli_flush_lsn)
			xfs_trans_ail_delete(ailp, (xfs_log_item_t*)qip);
		else
			spin_unlock(&ailp->xa_lock);
	}

	/*
	 * Release the dq's flush lock since we're done with it.
	 */
	xfs_dqfunlock(dqp);
}

int
xfs_qm_dqlock_nowait(
	xfs_dquot_t *dqp)
{
	return mutex_trylock(&dqp->q_qlock);
}

void
xfs_dqlock(
	xfs_dquot_t *dqp)
{
	mutex_lock(&dqp->q_qlock);
}

void
xfs_dqunlock(
	xfs_dquot_t *dqp)
{
	mutex_unlock(&(dqp->q_qlock));
	if (dqp->q_logitem.qli_dquot == dqp) {
		/* Once was dqp->q_mount, but might just have been cleared */
		xfs_trans_unlocked_item(dqp->q_logitem.qli_item.li_ailp,
					(xfs_log_item_t*)&(dqp->q_logitem));
	}
}


void
xfs_dqunlock_nonotify(
	xfs_dquot_t *dqp)
{
	mutex_unlock(&(dqp->q_qlock));
}

/*
 * Lock two xfs_dquot structures.
 *
 * To avoid deadlocks we always lock the quota structure with
 * the lowerd id first.
 */
void
xfs_dqlock2(
	xfs_dquot_t	*d1,
	xfs_dquot_t	*d2)
{
	if (d1 && d2) {
		ASSERT(d1 != d2);
		if (be32_to_cpu(d1->q_core.d_id) >
		    be32_to_cpu(d2->q_core.d_id)) {
			mutex_lock(&d2->q_qlock);
			mutex_lock_nested(&d1->q_qlock, XFS_QLOCK_NESTED);
		} else {
			mutex_lock(&d1->q_qlock);
			mutex_lock_nested(&d2->q_qlock, XFS_QLOCK_NESTED);
		}
	} else if (d1) {
		mutex_lock(&d1->q_qlock);
	} else if (d2) {
		mutex_lock(&d2->q_qlock);
	}
}


/*
 * Take a dquot out of the mount's dqlist as well as the hashlist.
 * This is called via unmount as well as quotaoff, and the purge
 * will always succeed unless there are soft (temp) references
 * outstanding.
 *
 * This returns 0 if it was purged, 1 if it wasn't. It's not an error code
 * that we're returning! XXXsup - not cool.
 */
/* ARGSUSED */
int
xfs_qm_dqpurge(
	xfs_dquot_t	*dqp)
{
	xfs_dqhash_t	*thishash;
	xfs_mount_t	*mp = dqp->q_mount;

	ASSERT(XFS_QM_IS_MPLIST_LOCKED(mp));
	ASSERT(mutex_is_locked(&dqp->q_hash->qh_lock));

	xfs_dqlock(dqp);
	/*
	 * We really can't afford to purge a dquot that is
	 * referenced, because these are hard refs.
	 * It shouldn't happen in general because we went thru _all_ inodes in
	 * dqrele_all_inodes before calling this and didn't let the mountlock go.
	 * However it is possible that we have dquots with temporary
	 * references that are not attached to an inode. e.g. see xfs_setattr().
	 */
	if (dqp->q_nrefs != 0) {
		xfs_dqunlock(dqp);
		mutex_unlock(&dqp->q_hash->qh_lock);
		return (1);
	}

	ASSERT(XFS_DQ_IS_ON_FREELIST(dqp));

	/*
	 * If we're turning off quotas, we have to make sure that, for
	 * example, we don't delete quota disk blocks while dquots are
	 * in the process of getting written to those disk blocks.
	 * This dquot might well be on AIL, and we can't leave it there
	 * if we're turning off quotas. Basically, we need this flush
	 * lock, and are willing to block on it.
	 */
	if (!xfs_dqflock_nowait(dqp)) {
		/*
		 * Block on the flush lock after nudging dquot buffer,
		 * if it is incore.
		 */
		xfs_qm_dqflock_pushbuf_wait(dqp);
	}

	/*
	 * XXXIf we're turning this type of quotas off, we don't care
	 * about the dirty metadata sitting in this dquot. OTOH, if
	 * we're unmounting, we do care, so we flush it and wait.
	 */
	if (XFS_DQ_IS_DIRTY(dqp)) {
		int	error;
		xfs_dqtrace_entry(dqp, "DQPURGE ->DQFLUSH: DQDIRTY");
		/* dqflush unlocks dqflock */
		/*
		 * Given that dqpurge is a very rare occurrence, it is OK
		 * that we're holding the hashlist and mplist locks
		 * across the disk write. But, ... XXXsup
		 *
		 * We don't care about getting disk errors here. We need
		 * to purge this dquot anyway, so we go ahead regardless.
		 */
		error = xfs_qm_dqflush(dqp, XFS_QMOPT_SYNC);
		if (error)
			xfs_fs_cmn_err(CE_WARN, mp,
				"xfs_qm_dqpurge: dquot %p flush failed", dqp);
		xfs_dqflock(dqp);
	}
	ASSERT(atomic_read(&dqp->q_pincount) == 0);
	ASSERT(XFS_FORCED_SHUTDOWN(mp) ||
	       !(dqp->q_logitem.qli_item.li_flags & XFS_LI_IN_AIL));

	thishash = dqp->q_hash;
	XQM_HASHLIST_REMOVE(thishash, dqp);
	XQM_MPLIST_REMOVE(&(XFS_QI_MPL_LIST(mp)), dqp);
	/*
	 * XXX Move this to the front of the freelist, if we can get the
	 * freelist lock.
	 */
	ASSERT(XFS_DQ_IS_ON_FREELIST(dqp));

	dqp->q_mount = NULL;
	dqp->q_hash = NULL;
	dqp->dq_flags = XFS_DQ_INACTIVE;
	memset(&dqp->q_core, 0, sizeof(dqp->q_core));
	xfs_dqfunlock(dqp);
	xfs_dqunlock(dqp);
	mutex_unlock(&thishash->qh_lock);
	return (0);
}


#ifdef QUOTADEBUG
void
xfs_qm_dqprint(xfs_dquot_t *dqp)
{
	cmn_err(CE_DEBUG, "-----------KERNEL DQUOT----------------");
	cmn_err(CE_DEBUG, "---- dquotID =  %d",
		(int)be32_to_cpu(dqp->q_core.d_id));
	cmn_err(CE_DEBUG, "---- type    =  %s", DQFLAGTO_TYPESTR(dqp));
	cmn_err(CE_DEBUG, "---- fs      =  0x%p", dqp->q_mount);
	cmn_err(CE_DEBUG, "---- blkno   =  0x%x", (int) dqp->q_blkno);
	cmn_err(CE_DEBUG, "---- boffset =  0x%x", (int) dqp->q_bufoffset);
	cmn_err(CE_DEBUG, "---- blkhlimit =  %Lu (0x%x)",
		be64_to_cpu(dqp->q_core.d_blk_hardlimit),
		(int)be64_to_cpu(dqp->q_core.d_blk_hardlimit));
	cmn_err(CE_DEBUG, "---- blkslimit =  %Lu (0x%x)",
		be64_to_cpu(dqp->q_core.d_blk_softlimit),
		(int)be64_to_cpu(dqp->q_core.d_blk_softlimit));
	cmn_err(CE_DEBUG, "---- inohlimit =  %Lu (0x%x)",
		be64_to_cpu(dqp->q_core.d_ino_hardlimit),
		(int)be64_to_cpu(dqp->q_core.d_ino_hardlimit));
	cmn_err(CE_DEBUG, "---- inoslimit =  %Lu (0x%x)",
		be64_to_cpu(dqp->q_core.d_ino_softlimit),
		(int)be64_to_cpu(dqp->q_core.d_ino_softlimit));
	cmn_err(CE_DEBUG, "---- bcount  =  %Lu (0x%x)",
		be64_to_cpu(dqp->q_core.d_bcount),
		(int)be64_to_cpu(dqp->q_core.d_bcount));
	cmn_err(CE_DEBUG, "---- icount  =  %Lu (0x%x)",
		be64_to_cpu(dqp->q_core.d_icount),
		(int)be64_to_cpu(dqp->q_core.d_icount));
	cmn_err(CE_DEBUG, "---- btimer  =  %d",
		(int)be32_to_cpu(dqp->q_core.d_btimer));
	cmn_err(CE_DEBUG, "---- itimer  =  %d",
		(int)be32_to_cpu(dqp->q_core.d_itimer));
	cmn_err(CE_DEBUG, "---------------------------");
}
#endif

/*
 * Give the buffer a little push if it is incore and
 * wait on the flush lock.
 */
void
xfs_qm_dqflock_pushbuf_wait(
	xfs_dquot_t	*dqp)
{
	xfs_buf_t	*bp;

	/*
	 * Check to see if the dquot has been flushed delayed
	 * write.  If so, grab its buffer and send it
	 * out immediately.  We'll be able to acquire
	 * the flush lock when the I/O completes.
	 */
	bp = xfs_incore(dqp->q_mount->m_ddev_targp, dqp->q_blkno,
		    XFS_QI_DQCHUNKLEN(dqp->q_mount),
		    XFS_INCORE_TRYLOCK);
	if (bp != NULL) {
		if (XFS_BUF_ISDELAYWRITE(bp)) {
			int	error;
			if (XFS_BUF_ISPINNED(bp)) {
				xfs_log_force(dqp->q_mount,
					      (xfs_lsn_t)0,
					      XFS_LOG_FORCE);
			}
			error = xfs_bawrite(dqp->q_mount, bp);
			if (error)
				xfs_fs_cmn_err(CE_WARN, dqp->q_mount,
					"xfs_qm_dqflock_pushbuf_wait: "
					"pushbuf error %d on dqp %p, bp %p",
					error, dqp, bp);
		} else {
			xfs_buf_relse(bp);
		}
	}
	xfs_dqflock(dqp);
}