summaryrefslogtreecommitdiff
path: root/fs/xfs/xfs_qm.c
blob: 5cff443f6cdb782f4d89cfed6e767bdb24a32e06 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
/*
 * Copyright (c) 2000-2005 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_alloc.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_ialloc_btree.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_ialloc.h"
#include "xfs_itable.h"
#include "xfs_rtalloc.h"
#include "xfs_error.h"
#include "xfs_bmap.h"
#include "xfs_attr.h"
#include "xfs_buf_item.h"
#include "xfs_trans_space.h"
#include "xfs_utils.h"
#include "xfs_qm.h"
#include "xfs_trace.h"

/*
 * The global quota manager. There is only one of these for the entire
 * system, _not_ one per file system. XQM keeps track of the overall
 * quota functionality, including maintaining the freelist and hash
 * tables of dquots.
 */
struct mutex	xfs_Gqm_lock;
struct xfs_qm	*xfs_Gqm;
uint		ndquot;

kmem_zone_t	*qm_dqzone;
kmem_zone_t	*qm_dqtrxzone;

STATIC void	xfs_qm_list_init(xfs_dqlist_t *, char *, int);
STATIC void	xfs_qm_list_destroy(xfs_dqlist_t *);

STATIC int	xfs_qm_init_quotainos(xfs_mount_t *);
STATIC int	xfs_qm_init_quotainfo(xfs_mount_t *);
STATIC int	xfs_qm_shake(struct shrinker *, struct shrink_control *);

static struct shrinker xfs_qm_shaker = {
	.shrink = xfs_qm_shake,
	.seeks = DEFAULT_SEEKS,
};

/*
 * Initialize the XQM structure.
 * Note that there is not one quota manager per file system.
 */
STATIC struct xfs_qm *
xfs_Gqm_init(void)
{
	xfs_dqhash_t	*udqhash, *gdqhash;
	xfs_qm_t	*xqm;
	size_t		hsize;
	uint		i;

	/*
	 * Initialize the dquot hash tables.
	 */
	udqhash = kmem_zalloc_greedy(&hsize,
				     XFS_QM_HASHSIZE_LOW * sizeof(xfs_dqhash_t),
				     XFS_QM_HASHSIZE_HIGH * sizeof(xfs_dqhash_t));
	if (!udqhash)
		goto out;

	gdqhash = kmem_zalloc_large(hsize);
	if (!gdqhash)
		goto out_free_udqhash;

	hsize /= sizeof(xfs_dqhash_t);
	ndquot = hsize << 8;

	xqm = kmem_zalloc(sizeof(xfs_qm_t), KM_SLEEP);
	xqm->qm_dqhashmask = hsize - 1;
	xqm->qm_usr_dqhtable = udqhash;
	xqm->qm_grp_dqhtable = gdqhash;
	ASSERT(xqm->qm_usr_dqhtable != NULL);
	ASSERT(xqm->qm_grp_dqhtable != NULL);

	for (i = 0; i < hsize; i++) {
		xfs_qm_list_init(&(xqm->qm_usr_dqhtable[i]), "uxdqh", i);
		xfs_qm_list_init(&(xqm->qm_grp_dqhtable[i]), "gxdqh", i);
	}

	/*
	 * Freelist of all dquots of all file systems
	 */
	INIT_LIST_HEAD(&xqm->qm_dqfrlist);
	xqm->qm_dqfrlist_cnt = 0;
	mutex_init(&xqm->qm_dqfrlist_lock);

	/*
	 * dquot zone. we register our own low-memory callback.
	 */
	if (!qm_dqzone) {
		xqm->qm_dqzone = kmem_zone_init(sizeof(xfs_dquot_t),
						"xfs_dquots");
		qm_dqzone = xqm->qm_dqzone;
	} else
		xqm->qm_dqzone = qm_dqzone;

	register_shrinker(&xfs_qm_shaker);

	/*
	 * The t_dqinfo portion of transactions.
	 */
	if (!qm_dqtrxzone) {
		xqm->qm_dqtrxzone = kmem_zone_init(sizeof(xfs_dquot_acct_t),
						   "xfs_dqtrx");
		qm_dqtrxzone = xqm->qm_dqtrxzone;
	} else
		xqm->qm_dqtrxzone = qm_dqtrxzone;

	atomic_set(&xqm->qm_totaldquots, 0);
	xqm->qm_dqfree_ratio = XFS_QM_DQFREE_RATIO;
	xqm->qm_nrefs = 0;
	return xqm;

 out_free_udqhash:
	kmem_free_large(udqhash);
 out:
	return NULL;
}

/*
 * Destroy the global quota manager when its reference count goes to zero.
 */
STATIC void
xfs_qm_destroy(
	struct xfs_qm	*xqm)
{
	struct xfs_dquot *dqp, *n;
	int		hsize, i;

	ASSERT(xqm != NULL);
	ASSERT(xqm->qm_nrefs == 0);
	unregister_shrinker(&xfs_qm_shaker);
	hsize = xqm->qm_dqhashmask + 1;
	for (i = 0; i < hsize; i++) {
		xfs_qm_list_destroy(&(xqm->qm_usr_dqhtable[i]));
		xfs_qm_list_destroy(&(xqm->qm_grp_dqhtable[i]));
	}
	kmem_free_large(xqm->qm_usr_dqhtable);
	kmem_free_large(xqm->qm_grp_dqhtable);
	xqm->qm_usr_dqhtable = NULL;
	xqm->qm_grp_dqhtable = NULL;
	xqm->qm_dqhashmask = 0;

	/* frlist cleanup */
	mutex_lock(&xqm->qm_dqfrlist_lock);
	list_for_each_entry_safe(dqp, n, &xqm->qm_dqfrlist, q_freelist) {
		xfs_dqlock(dqp);
		list_del_init(&dqp->q_freelist);
		xfs_Gqm->qm_dqfrlist_cnt--;
		xfs_dqunlock(dqp);
		xfs_qm_dqdestroy(dqp);
	}
	mutex_unlock(&xqm->qm_dqfrlist_lock);
	mutex_destroy(&xqm->qm_dqfrlist_lock);
	kmem_free(xqm);
}

/*
 * Called at mount time to let XQM know that another file system is
 * starting quotas. This isn't crucial information as the individual mount
 * structures are pretty independent, but it helps the XQM keep a
 * global view of what's going on.
 */
/* ARGSUSED */
STATIC int
xfs_qm_hold_quotafs_ref(
	struct xfs_mount *mp)
{
	/*
	 * Need to lock the xfs_Gqm structure for things like this. For example,
	 * the structure could disappear between the entry to this routine and
	 * a HOLD operation if not locked.
	 */
	mutex_lock(&xfs_Gqm_lock);

	if (!xfs_Gqm) {
		xfs_Gqm = xfs_Gqm_init();
		if (!xfs_Gqm) {
			mutex_unlock(&xfs_Gqm_lock);
			return ENOMEM;
		}
	}

	/*
	 * We can keep a list of all filesystems with quotas mounted for
	 * debugging and statistical purposes, but ...
	 * Just take a reference and get out.
	 */
	xfs_Gqm->qm_nrefs++;
	mutex_unlock(&xfs_Gqm_lock);

	return 0;
}


/*
 * Release the reference that a filesystem took at mount time,
 * so that we know when we need to destroy the entire quota manager.
 */
/* ARGSUSED */
STATIC void
xfs_qm_rele_quotafs_ref(
	struct xfs_mount *mp)
{
	xfs_dquot_t	*dqp, *n;

	ASSERT(xfs_Gqm);
	ASSERT(xfs_Gqm->qm_nrefs > 0);

	/*
	 * Go thru the freelist and destroy all inactive dquots.
	 */
	mutex_lock(&xfs_Gqm->qm_dqfrlist_lock);

	list_for_each_entry_safe(dqp, n, &xfs_Gqm->qm_dqfrlist, q_freelist) {
		xfs_dqlock(dqp);
		if (dqp->dq_flags & XFS_DQ_INACTIVE) {
			ASSERT(dqp->q_mount == NULL);
			ASSERT(! XFS_DQ_IS_DIRTY(dqp));
			ASSERT(list_empty(&dqp->q_hashlist));
			ASSERT(list_empty(&dqp->q_mplist));
			list_del_init(&dqp->q_freelist);
			xfs_Gqm->qm_dqfrlist_cnt--;
			xfs_dqunlock(dqp);
			xfs_qm_dqdestroy(dqp);
		} else {
			xfs_dqunlock(dqp);
		}
	}
	mutex_unlock(&xfs_Gqm->qm_dqfrlist_lock);

	/*
	 * Destroy the entire XQM. If somebody mounts with quotaon, this'll
	 * be restarted.
	 */
	mutex_lock(&xfs_Gqm_lock);
	if (--xfs_Gqm->qm_nrefs == 0) {
		xfs_qm_destroy(xfs_Gqm);
		xfs_Gqm = NULL;
	}
	mutex_unlock(&xfs_Gqm_lock);
}

/*
 * Just destroy the quotainfo structure.
 */
void
xfs_qm_unmount(
	struct xfs_mount	*mp)
{
	if (mp->m_quotainfo) {
		xfs_qm_dqpurge_all(mp, XFS_QMOPT_QUOTALL);
		xfs_qm_destroy_quotainfo(mp);
	}
}


/*
 * This is called from xfs_mountfs to start quotas and initialize all
 * necessary data structures like quotainfo.  This is also responsible for
 * running a quotacheck as necessary.  We are guaranteed that the superblock
 * is consistently read in at this point.
 *
 * If we fail here, the mount will continue with quota turned off. We don't
 * need to inidicate success or failure at all.
 */
void
xfs_qm_mount_quotas(
	xfs_mount_t	*mp)
{
	int		error = 0;
	uint		sbf;

	/*
	 * If quotas on realtime volumes is not supported, we disable
	 * quotas immediately.
	 */
	if (mp->m_sb.sb_rextents) {
		xfs_notice(mp, "Cannot turn on quotas for realtime filesystem");
		mp->m_qflags = 0;
		goto write_changes;
	}

	ASSERT(XFS_IS_QUOTA_RUNNING(mp));

	/*
	 * Allocate the quotainfo structure inside the mount struct, and
	 * create quotainode(s), and change/rev superblock if necessary.
	 */
	error = xfs_qm_init_quotainfo(mp);
	if (error) {
		/*
		 * We must turn off quotas.
		 */
		ASSERT(mp->m_quotainfo == NULL);
		mp->m_qflags = 0;
		goto write_changes;
	}
	/*
	 * If any of the quotas are not consistent, do a quotacheck.
	 */
	if (XFS_QM_NEED_QUOTACHECK(mp)) {
		error = xfs_qm_quotacheck(mp);
		if (error) {
			/* Quotacheck failed and disabled quotas. */
			return;
		}
	}
	/* 
	 * If one type of quotas is off, then it will lose its
	 * quotachecked status, since we won't be doing accounting for
	 * that type anymore.
	 */
	if (!XFS_IS_UQUOTA_ON(mp))
		mp->m_qflags &= ~XFS_UQUOTA_CHKD;
	if (!(XFS_IS_GQUOTA_ON(mp) || XFS_IS_PQUOTA_ON(mp)))
		mp->m_qflags &= ~XFS_OQUOTA_CHKD;

 write_changes:
	/*
	 * We actually don't have to acquire the m_sb_lock at all.
	 * This can only be called from mount, and that's single threaded. XXX
	 */
	spin_lock(&mp->m_sb_lock);
	sbf = mp->m_sb.sb_qflags;
	mp->m_sb.sb_qflags = mp->m_qflags & XFS_MOUNT_QUOTA_ALL;
	spin_unlock(&mp->m_sb_lock);

	if (sbf != (mp->m_qflags & XFS_MOUNT_QUOTA_ALL)) {
		if (xfs_qm_write_sb_changes(mp, XFS_SB_QFLAGS)) {
			/*
			 * We could only have been turning quotas off.
			 * We aren't in very good shape actually because
			 * the incore structures are convinced that quotas are
			 * off, but the on disk superblock doesn't know that !
			 */
			ASSERT(!(XFS_IS_QUOTA_RUNNING(mp)));
			xfs_alert(mp, "%s: Superblock update failed!",
				__func__);
		}
	}

	if (error) {
		xfs_warn(mp, "Failed to initialize disk quotas.");
		return;
	}
}

/*
 * Called from the vfsops layer.
 */
void
xfs_qm_unmount_quotas(
	xfs_mount_t	*mp)
{
	/*
	 * Release the dquots that root inode, et al might be holding,
	 * before we flush quotas and blow away the quotainfo structure.
	 */
	ASSERT(mp->m_rootip);
	xfs_qm_dqdetach(mp->m_rootip);
	if (mp->m_rbmip)
		xfs_qm_dqdetach(mp->m_rbmip);
	if (mp->m_rsumip)
		xfs_qm_dqdetach(mp->m_rsumip);

	/*
	 * Release the quota inodes.
	 */
	if (mp->m_quotainfo) {
		if (mp->m_quotainfo->qi_uquotaip) {
			IRELE(mp->m_quotainfo->qi_uquotaip);
			mp->m_quotainfo->qi_uquotaip = NULL;
		}
		if (mp->m_quotainfo->qi_gquotaip) {
			IRELE(mp->m_quotainfo->qi_gquotaip);
			mp->m_quotainfo->qi_gquotaip = NULL;
		}
	}
}

/*
 * Flush all dquots of the given file system to disk. The dquots are
 * _not_ purged from memory here, just their data written to disk.
 */
STATIC int
xfs_qm_dqflush_all(
	struct xfs_mount	*mp,
	int			sync_mode)
{
	struct xfs_quotainfo	*q = mp->m_quotainfo;
	int			recl;
	struct xfs_dquot	*dqp;
	int			error;

	if (!q)
		return 0;
again:
	mutex_lock(&q->qi_dqlist_lock);
	list_for_each_entry(dqp, &q->qi_dqlist, q_mplist) {
		xfs_dqlock(dqp);
		if (! XFS_DQ_IS_DIRTY(dqp)) {
			xfs_dqunlock(dqp);
			continue;
		}

		/* XXX a sentinel would be better */
		recl = q->qi_dqreclaims;
		if (!xfs_dqflock_nowait(dqp)) {
			/*
			 * If we can't grab the flush lock then 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.
			 */
			xfs_qm_dqflock_pushbuf_wait(dqp);
		}
		/*
		 * Let go of the mplist lock. We don't want to hold it
		 * across a disk write.
		 */
		mutex_unlock(&q->qi_dqlist_lock);
		error = xfs_qm_dqflush(dqp, sync_mode);
		xfs_dqunlock(dqp);
		if (error)
			return error;

		mutex_lock(&q->qi_dqlist_lock);
		if (recl != q->qi_dqreclaims) {
			mutex_unlock(&q->qi_dqlist_lock);
			/* XXX restart limit */
			goto again;
		}
	}

	mutex_unlock(&q->qi_dqlist_lock);
	/* return ! busy */
	return 0;
}
/*
 * Release the group dquot pointers the user dquots may be
 * carrying around as a hint. mplist is locked on entry and exit.
 */
STATIC void
xfs_qm_detach_gdquots(
	struct xfs_mount	*mp)
{
	struct xfs_quotainfo	*q = mp->m_quotainfo;
	struct xfs_dquot	*dqp, *gdqp;
	int			nrecl;

 again:
	ASSERT(mutex_is_locked(&q->qi_dqlist_lock));
	list_for_each_entry(dqp, &q->qi_dqlist, q_mplist) {
		xfs_dqlock(dqp);
		if ((gdqp = dqp->q_gdquot)) {
			xfs_dqlock(gdqp);
			dqp->q_gdquot = NULL;
		}
		xfs_dqunlock(dqp);

		if (gdqp) {
			/*
			 * Can't hold the mplist lock across a dqput.
			 * XXXmust convert to marker based iterations here.
			 */
			nrecl = q->qi_dqreclaims;
			mutex_unlock(&q->qi_dqlist_lock);
			xfs_qm_dqput(gdqp);

			mutex_lock(&q->qi_dqlist_lock);
			if (nrecl != q->qi_dqreclaims)
				goto again;
		}
	}
}

/*
 * Go through all the incore dquots of this file system and take them
 * off the mplist and hashlist, if the dquot type matches the dqtype
 * parameter. This is used when turning off quota accounting for
 * users and/or groups, as well as when the filesystem is unmounting.
 */
STATIC int
xfs_qm_dqpurge_int(
	struct xfs_mount	*mp,
	uint			flags)
{
	struct xfs_quotainfo	*q = mp->m_quotainfo;
	struct xfs_dquot	*dqp, *n;
	uint			dqtype;
	int			nrecl;
	int			nmisses;

	if (!q)
		return 0;

	dqtype = (flags & XFS_QMOPT_UQUOTA) ? XFS_DQ_USER : 0;
	dqtype |= (flags & XFS_QMOPT_PQUOTA) ? XFS_DQ_PROJ : 0;
	dqtype |= (flags & XFS_QMOPT_GQUOTA) ? XFS_DQ_GROUP : 0;

	mutex_lock(&q->qi_dqlist_lock);

	/*
	 * In the first pass through all incore dquots of this filesystem,
	 * we release the group dquot pointers the user dquots may be
	 * carrying around as a hint. We need to do this irrespective of
	 * what's being turned off.
	 */
	xfs_qm_detach_gdquots(mp);

      again:
	nmisses = 0;
	ASSERT(mutex_is_locked(&q->qi_dqlist_lock));
	/*
	 * Try to get rid of all of the unwanted dquots. The idea is to
	 * get them off mplist and hashlist, but leave them on freelist.
	 */
	list_for_each_entry_safe(dqp, n, &q->qi_dqlist, q_mplist) {
		/*
		 * It's OK to look at the type without taking dqlock here.
		 * We're holding the mplist lock here, and that's needed for
		 * a dqreclaim.
		 */
		if ((dqp->dq_flags & dqtype) == 0)
			continue;

		if (!mutex_trylock(&dqp->q_hash->qh_lock)) {
			nrecl = q->qi_dqreclaims;
			mutex_unlock(&q->qi_dqlist_lock);
			mutex_lock(&dqp->q_hash->qh_lock);
			mutex_lock(&q->qi_dqlist_lock);

			/*
			 * XXXTheoretically, we can get into a very long
			 * ping pong game here.
			 * No one can be adding dquots to the mplist at
			 * this point, but somebody might be taking things off.
			 */
			if (nrecl != q->qi_dqreclaims) {
				mutex_unlock(&dqp->q_hash->qh_lock);
				goto again;
			}
		}

		/*
		 * Take the dquot off the mplist and hashlist. It may remain on
		 * freelist in INACTIVE state.
		 */
		nmisses += xfs_qm_dqpurge(dqp);
	}
	mutex_unlock(&q->qi_dqlist_lock);
	return nmisses;
}

int
xfs_qm_dqpurge_all(
	xfs_mount_t	*mp,
	uint		flags)
{
	int		ndquots;

	/*
	 * Purge the dquot cache.
	 * None of the dquots should really be busy at this point.
	 */
	if (mp->m_quotainfo) {
		while ((ndquots = xfs_qm_dqpurge_int(mp, flags))) {
			delay(ndquots * 10);
		}
	}
	return 0;
}

STATIC int
xfs_qm_dqattach_one(
	xfs_inode_t	*ip,
	xfs_dqid_t	id,
	uint		type,
	uint		doalloc,
	xfs_dquot_t	*udqhint, /* hint */
	xfs_dquot_t	**IO_idqpp)
{
	xfs_dquot_t	*dqp;
	int		error;

	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
	error = 0;

	/*
	 * See if we already have it in the inode itself. IO_idqpp is
	 * &i_udquot or &i_gdquot. This made the code look weird, but
	 * made the logic a lot simpler.
	 */
	dqp = *IO_idqpp;
	if (dqp) {
		trace_xfs_dqattach_found(dqp);
		return 0;
	}

	/*
	 * udqhint is the i_udquot field in inode, and is non-NULL only
	 * when the type arg is group/project. Its purpose is to save a
	 * lookup by dqid (xfs_qm_dqget) by caching a group dquot inside
	 * the user dquot.
	 */
	if (udqhint) {
		ASSERT(type == XFS_DQ_GROUP || type == XFS_DQ_PROJ);
		xfs_dqlock(udqhint);

		/*
		 * No need to take dqlock to look at the id.
		 *
		 * The ID can't change until it gets reclaimed, and it won't
		 * be reclaimed as long as we have a ref from inode and we
		 * hold the ilock.
		 */
		dqp = udqhint->q_gdquot;
		if (dqp && be32_to_cpu(dqp->q_core.d_id) == id) {
			xfs_dqlock(dqp);
			XFS_DQHOLD(dqp);
			ASSERT(*IO_idqpp == NULL);
			*IO_idqpp = dqp;

			xfs_dqunlock(dqp);
			xfs_dqunlock(udqhint);
			return 0;
		}

		/*
		 * We can't hold a dquot lock when we call the dqget code.
		 * We'll deadlock in no time, because of (not conforming to)
		 * lock ordering - the inodelock comes before any dquot lock,
		 * and we may drop and reacquire the ilock in xfs_qm_dqget().
		 */
		xfs_dqunlock(udqhint);
	}

	/*
	 * Find the dquot from somewhere. This bumps the
	 * reference count of dquot and returns it locked.
	 * 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.
	 */
	error = xfs_qm_dqget(ip->i_mount, ip, id, type, XFS_QMOPT_DOWARN, &dqp);
	if (error)
		return error;

	trace_xfs_dqattach_get(dqp);

	/*
	 * dqget may have dropped and re-acquired the ilock, but it guarantees
	 * that the dquot returned is the one that should go in the inode.
	 */
	*IO_idqpp = dqp;
	xfs_dqunlock(dqp);
	return 0;
}


/*
 * Given a udquot and gdquot, attach a ptr to the group dquot in the
 * udquot as a hint for future lookups. The idea sounds simple, but the
 * execution isn't, because the udquot might have a group dquot attached
 * already and getting rid of that gets us into lock ordering constraints.
 * The process is complicated more by the fact that the dquots may or may not
 * be locked on entry.
 */
STATIC void
xfs_qm_dqattach_grouphint(
	xfs_dquot_t	*udq,
	xfs_dquot_t	*gdq)
{
	xfs_dquot_t	*tmp;

	xfs_dqlock(udq);

	if ((tmp = udq->q_gdquot)) {
		if (tmp == gdq) {
			xfs_dqunlock(udq);
			return;
		}

		udq->q_gdquot = NULL;
		/*
		 * We can't keep any dqlocks when calling dqrele,
		 * because the freelist lock comes before dqlocks.
		 */
		xfs_dqunlock(udq);
		/*
		 * we took a hard reference once upon a time in dqget,
		 * so give it back when the udquot no longer points at it
		 * dqput() does the unlocking of the dquot.
		 */
		xfs_qm_dqrele(tmp);

		xfs_dqlock(udq);
		xfs_dqlock(gdq);

	} else {
		ASSERT(XFS_DQ_IS_LOCKED(udq));
		xfs_dqlock(gdq);
	}

	ASSERT(XFS_DQ_IS_LOCKED(udq));
	ASSERT(XFS_DQ_IS_LOCKED(gdq));
	/*
	 * Somebody could have attached a gdquot here,
	 * when we dropped the uqlock. If so, just do nothing.
	 */
	if (udq->q_gdquot == NULL) {
		XFS_DQHOLD(gdq);
		udq->q_gdquot = gdq;
	}

	xfs_dqunlock(gdq);
	xfs_dqunlock(udq);
}


/*
 * Given a locked inode, attach dquot(s) to it, taking U/G/P-QUOTAON
 * into account.
 * If XFS_QMOPT_DQALLOC, the dquot(s) will be allocated if needed.
 * Inode may get unlocked and relocked in here, and the caller must deal with
 * the consequences.
 */
int
xfs_qm_dqattach_locked(
	xfs_inode_t	*ip,
	uint		flags)
{
	xfs_mount_t	*mp = ip->i_mount;
	uint		nquotas = 0;
	int		error = 0;

	if (!XFS_IS_QUOTA_RUNNING(mp) ||
	    !XFS_IS_QUOTA_ON(mp) ||
	    !XFS_NOT_DQATTACHED(mp, ip) ||
	    ip->i_ino == mp->m_sb.sb_uquotino ||
	    ip->i_ino == mp->m_sb.sb_gquotino)
		return 0;

	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));

	if (XFS_IS_UQUOTA_ON(mp)) {
		error = xfs_qm_dqattach_one(ip, ip->i_d.di_uid, XFS_DQ_USER,
						flags & XFS_QMOPT_DQALLOC,
						NULL, &ip->i_udquot);
		if (error)
			goto done;
		nquotas++;
	}

	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
	if (XFS_IS_OQUOTA_ON(mp)) {
		error = XFS_IS_GQUOTA_ON(mp) ?
			xfs_qm_dqattach_one(ip, ip->i_d.di_gid, XFS_DQ_GROUP,
						flags & XFS_QMOPT_DQALLOC,
						ip->i_udquot, &ip->i_gdquot) :
			xfs_qm_dqattach_one(ip, xfs_get_projid(ip), XFS_DQ_PROJ,
						flags & XFS_QMOPT_DQALLOC,
						ip->i_udquot, &ip->i_gdquot);
		/*
		 * Don't worry about the udquot that we may have
		 * attached above. It'll get detached, if not already.
		 */
		if (error)
			goto done;
		nquotas++;
	}

	/*
	 * Attach this group quota to the user quota as a hint.
	 * This WON'T, in general, result in a thrash.
	 */
	if (nquotas == 2) {
		ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
		ASSERT(ip->i_udquot);
		ASSERT(ip->i_gdquot);

		/*
		 * We may or may not have the i_udquot locked at this point,
		 * but this check is OK since we don't depend on the i_gdquot to
		 * be accurate 100% all the time. It is just a hint, and this
		 * will succeed in general.
		 */
		if (ip->i_udquot->q_gdquot == ip->i_gdquot)
			goto done;
		/*
		 * Attach i_gdquot to the gdquot hint inside the i_udquot.
		 */
		xfs_qm_dqattach_grouphint(ip->i_udquot, ip->i_gdquot);
	}

 done:
#ifdef DEBUG
	if (!error) {
		if (XFS_IS_UQUOTA_ON(mp))
			ASSERT(ip->i_udquot);
		if (XFS_IS_OQUOTA_ON(mp))
			ASSERT(ip->i_gdquot);
	}
	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
#endif
	return error;
}

int
xfs_qm_dqattach(
	struct xfs_inode	*ip,
	uint			flags)
{
	int			error;

	xfs_ilock(ip, XFS_ILOCK_EXCL);
	error = xfs_qm_dqattach_locked(ip, flags);
	xfs_iunlock(ip, XFS_ILOCK_EXCL);

	return error;
}

/*
 * Release dquots (and their references) if any.
 * The inode should be locked EXCL except when this's called by
 * xfs_ireclaim.
 */
void
xfs_qm_dqdetach(
	xfs_inode_t	*ip)
{
	if (!(ip->i_udquot || ip->i_gdquot))
		return;

	trace_xfs_dquot_dqdetach(ip);

	ASSERT(ip->i_ino != ip->i_mount->m_sb.sb_uquotino);
	ASSERT(ip->i_ino != ip->i_mount->m_sb.sb_gquotino);
	if (ip->i_udquot) {
		xfs_qm_dqrele(ip->i_udquot);
		ip->i_udquot = NULL;
	}
	if (ip->i_gdquot) {
		xfs_qm_dqrele(ip->i_gdquot);
		ip->i_gdquot = NULL;
	}
}

int
xfs_qm_sync(
	struct xfs_mount	*mp,
	int			flags)
{
	struct xfs_quotainfo	*q = mp->m_quotainfo;
	int			recl, restarts;
	struct xfs_dquot	*dqp;
	int			error;

	if (!XFS_IS_QUOTA_RUNNING(mp) || !XFS_IS_QUOTA_ON(mp))
		return 0;

	restarts = 0;

  again:
	mutex_lock(&q->qi_dqlist_lock);
	/*
	 * dqpurge_all() also takes the mplist lock and iterate thru all dquots
	 * in quotaoff. However, if the QUOTA_ACTIVE bits are not cleared
	 * when we have the mplist lock, we know that dquots will be consistent
	 * as long as we have it locked.
	 */
	if (!XFS_IS_QUOTA_ON(mp)) {
		mutex_unlock(&q->qi_dqlist_lock);
		return 0;
	}
	ASSERT(mutex_is_locked(&q->qi_dqlist_lock));
	list_for_each_entry(dqp, &q->qi_dqlist, q_mplist) {
		/*
		 * If this is vfs_sync calling, then skip the dquots that
		 * don't 'seem' to be dirty. ie. don't acquire dqlock.
		 * This is very similar to what xfs_sync does with inodes.
		 */
		if (flags & SYNC_TRYLOCK) {
			if (!XFS_DQ_IS_DIRTY(dqp))
				continue;
			if (!xfs_qm_dqlock_nowait(dqp))
				continue;
		} else {
			xfs_dqlock(dqp);
		}

		/*
		 * Now, find out for sure if this dquot is dirty or not.
		 */
		if (! XFS_DQ_IS_DIRTY(dqp)) {
			xfs_dqunlock(dqp);
			continue;
		}

		/* XXX a sentinel would be better */
		recl = q->qi_dqreclaims;
		if (!xfs_dqflock_nowait(dqp)) {
			if (flags & SYNC_TRYLOCK) {
				xfs_dqunlock(dqp);
				continue;
			}
			/*
			 * If we can't grab the flush lock then if the caller
			 * really wanted us to give this our best shot, so
			 * see if we can give a push to the buffer before we wait
			 * on the flush lock. At this point, we know that
			 * even though the dquot is being flushed,
			 * it has (new) dirty data.
			 */
			xfs_qm_dqflock_pushbuf_wait(dqp);
		}
		/*
		 * Let go of the mplist lock. We don't want to hold it
		 * across a disk write
		 */
		mutex_unlock(&q->qi_dqlist_lock);
		error = xfs_qm_dqflush(dqp, flags);
		xfs_dqunlock(dqp);
		if (error && XFS_FORCED_SHUTDOWN(mp))
			return 0;	/* Need to prevent umount failure */
		else if (error)
			return error;

		mutex_lock(&q->qi_dqlist_lock);
		if (recl != q->qi_dqreclaims) {
			if (++restarts >= XFS_QM_SYNC_MAX_RESTARTS)
				break;

			mutex_unlock(&q->qi_dqlist_lock);
			goto again;
		}
	}

	mutex_unlock(&q->qi_dqlist_lock);
	return 0;
}

/*
 * The hash chains and the mplist use the same xfs_dqhash structure as
 * their list head, but we can take the mplist qh_lock and one of the
 * hash qh_locks at the same time without any problem as they aren't
 * related.
 */
static struct lock_class_key xfs_quota_mplist_class;

/*
 * This initializes all the quota information that's kept in the
 * mount structure
 */
STATIC int
xfs_qm_init_quotainfo(
	xfs_mount_t	*mp)
{
	xfs_quotainfo_t *qinf;
	int		error;
	xfs_dquot_t	*dqp;

	ASSERT(XFS_IS_QUOTA_RUNNING(mp));

	/*
	 * Tell XQM that we exist as soon as possible.
	 */
	if ((error = xfs_qm_hold_quotafs_ref(mp))) {
		return error;
	}

	qinf = mp->m_quotainfo = kmem_zalloc(sizeof(xfs_quotainfo_t), KM_SLEEP);

	/*
	 * See if quotainodes are setup, and if not, allocate them,
	 * and change the superblock accordingly.
	 */
	if ((error = xfs_qm_init_quotainos(mp))) {
		kmem_free(qinf);
		mp->m_quotainfo = NULL;
		return error;
	}

	INIT_LIST_HEAD(&qinf->qi_dqlist);
	mutex_init(&qinf->qi_dqlist_lock);
	lockdep_set_class(&qinf->qi_dqlist_lock, &xfs_quota_mplist_class);

	qinf->qi_dqreclaims = 0;

	/* mutex used to serialize quotaoffs */
	mutex_init(&qinf->qi_quotaofflock);

	/* Precalc some constants */
	qinf->qi_dqchunklen = XFS_FSB_TO_BB(mp, XFS_DQUOT_CLUSTER_SIZE_FSB);
	ASSERT(qinf->qi_dqchunklen);
	qinf->qi_dqperchunk = BBTOB(qinf->qi_dqchunklen);
	do_div(qinf->qi_dqperchunk, sizeof(xfs_dqblk_t));

	mp->m_qflags |= (mp->m_sb.sb_qflags & XFS_ALL_QUOTA_CHKD);

	/*
	 * We try to get the limits from the superuser's limits fields.
	 * This is quite hacky, but it is standard quota practice.
	 * We look at the USR dquot with id == 0 first, but if user quotas
	 * are not enabled we goto the GRP dquot with id == 0.
	 * We don't really care to keep separate default limits for user
	 * and group quotas, at least not at this point.
	 */
	error = xfs_qm_dqget(mp, NULL, (xfs_dqid_t)0,
			     XFS_IS_UQUOTA_RUNNING(mp) ? XFS_DQ_USER : 
			     (XFS_IS_GQUOTA_RUNNING(mp) ? XFS_DQ_GROUP :
				XFS_DQ_PROJ),
			     XFS_QMOPT_DQSUSER|XFS_QMOPT_DOWARN,
			     &dqp);
	if (! error) {
		xfs_disk_dquot_t	*ddqp = &dqp->q_core;

		/*
		 * The warnings and timers set the grace period given to
		 * a user or group before he or she can not perform any
		 * more writing. If it is zero, a default is used.
		 */
		qinf->qi_btimelimit = ddqp->d_btimer ?
			be32_to_cpu(ddqp->d_btimer) : XFS_QM_BTIMELIMIT;
		qinf->qi_itimelimit = ddqp->d_itimer ?
			be32_to_cpu(ddqp->d_itimer) : XFS_QM_ITIMELIMIT;
		qinf->qi_rtbtimelimit = ddqp->d_rtbtimer ?
			be32_to_cpu(ddqp->d_rtbtimer) : XFS_QM_RTBTIMELIMIT;
		qinf->qi_bwarnlimit = ddqp->d_bwarns ?
			be16_to_cpu(ddqp->d_bwarns) : XFS_QM_BWARNLIMIT;
		qinf->qi_iwarnlimit = ddqp->d_iwarns ?
			be16_to_cpu(ddqp->d_iwarns) : XFS_QM_IWARNLIMIT;
		qinf->qi_rtbwarnlimit = ddqp->d_rtbwarns ?
			be16_to_cpu(ddqp->d_rtbwarns) : XFS_QM_RTBWARNLIMIT;
		qinf->qi_bhardlimit = be64_to_cpu(ddqp->d_blk_hardlimit);
		qinf->qi_bsoftlimit = be64_to_cpu(ddqp->d_blk_softlimit);
		qinf->qi_ihardlimit = be64_to_cpu(ddqp->d_ino_hardlimit);
		qinf->qi_isoftlimit = be64_to_cpu(ddqp->d_ino_softlimit);
		qinf->qi_rtbhardlimit = be64_to_cpu(ddqp->d_rtb_hardlimit);
		qinf->qi_rtbsoftlimit = be64_to_cpu(ddqp->d_rtb_softlimit);
 
		/*
		 * We sent the XFS_QMOPT_DQSUSER flag to dqget because
		 * we don't want this dquot cached. We haven't done a
		 * quotacheck yet, and quotacheck doesn't like incore dquots.
		 */
		xfs_qm_dqdestroy(dqp);
	} else {
		qinf->qi_btimelimit = XFS_QM_BTIMELIMIT;
		qinf->qi_itimelimit = XFS_QM_ITIMELIMIT;
		qinf->qi_rtbtimelimit = XFS_QM_RTBTIMELIMIT;
		qinf->qi_bwarnlimit = XFS_QM_BWARNLIMIT;
		qinf->qi_iwarnlimit = XFS_QM_IWARNLIMIT;
		qinf->qi_rtbwarnlimit = XFS_QM_RTBWARNLIMIT;
	}

	return 0;
}


/*
 * Gets called when unmounting a filesystem or when all quotas get
 * turned off.
 * This purges the quota inodes, destroys locks and frees itself.
 */
void
xfs_qm_destroy_quotainfo(
	xfs_mount_t	*mp)
{
	xfs_quotainfo_t *qi;

	qi = mp->m_quotainfo;
	ASSERT(qi != NULL);
	ASSERT(xfs_Gqm != NULL);

	/*
	 * Release the reference that XQM kept, so that we know
	 * when the XQM structure should be freed. We cannot assume
	 * that xfs_Gqm is non-null after this point.
	 */
	xfs_qm_rele_quotafs_ref(mp);

	ASSERT(list_empty(&qi->qi_dqlist));
	mutex_destroy(&qi->qi_dqlist_lock);

	if (qi->qi_uquotaip) {
		IRELE(qi->qi_uquotaip);
		qi->qi_uquotaip = NULL; /* paranoia */
	}
	if (qi->qi_gquotaip) {
		IRELE(qi->qi_gquotaip);
		qi->qi_gquotaip = NULL;
	}
	mutex_destroy(&qi->qi_quotaofflock);
	kmem_free(qi);
	mp->m_quotainfo = NULL;
}



/* ------------------- PRIVATE STATIC FUNCTIONS ----------------------- */

/* ARGSUSED */
STATIC void
xfs_qm_list_init(
	xfs_dqlist_t	*list,
	char		*str,
	int		n)
{
	mutex_init(&list->qh_lock);
	INIT_LIST_HEAD(&list->qh_list);
	list->qh_version = 0;
	list->qh_nelems = 0;
}

STATIC void
xfs_qm_list_destroy(
	xfs_dqlist_t	*list)
{
	mutex_destroy(&(list->qh_lock));
}

/*
 * Create an inode and return with a reference already taken, but unlocked
 * This is how we create quota inodes
 */
STATIC int
xfs_qm_qino_alloc(
	xfs_mount_t	*mp,
	xfs_inode_t	**ip,
	__int64_t	sbfields,
	uint		flags)
{
	xfs_trans_t	*tp;
	int		error;
	int		committed;

	tp = xfs_trans_alloc(mp, XFS_TRANS_QM_QINOCREATE);
	if ((error = xfs_trans_reserve(tp,
				      XFS_QM_QINOCREATE_SPACE_RES(mp),
				      XFS_CREATE_LOG_RES(mp), 0,
				      XFS_TRANS_PERM_LOG_RES,
				      XFS_CREATE_LOG_COUNT))) {
		xfs_trans_cancel(tp, 0);
		return error;
	}

	error = xfs_dir_ialloc(&tp, NULL, S_IFREG, 1, 0, 0, 1, ip, &committed);
	if (error) {
		xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES |
				 XFS_TRANS_ABORT);
		return error;
	}

	/*
	 * Make the changes in the superblock, and log those too.
	 * sbfields arg may contain fields other than *QUOTINO;
	 * VERSIONNUM for example.
	 */
	spin_lock(&mp->m_sb_lock);
	if (flags & XFS_QMOPT_SBVERSION) {
		ASSERT(!xfs_sb_version_hasquota(&mp->m_sb));
		ASSERT((sbfields & (XFS_SB_VERSIONNUM | XFS_SB_UQUOTINO |
				   XFS_SB_GQUOTINO | XFS_SB_QFLAGS)) ==
		       (XFS_SB_VERSIONNUM | XFS_SB_UQUOTINO |
			XFS_SB_GQUOTINO | XFS_SB_QFLAGS));

		xfs_sb_version_addquota(&mp->m_sb);
		mp->m_sb.sb_uquotino = NULLFSINO;
		mp->m_sb.sb_gquotino = NULLFSINO;

		/* qflags will get updated _after_ quotacheck */
		mp->m_sb.sb_qflags = 0;
	}
	if (flags & XFS_QMOPT_UQUOTA)
		mp->m_sb.sb_uquotino = (*ip)->i_ino;
	else
		mp->m_sb.sb_gquotino = (*ip)->i_ino;
	spin_unlock(&mp->m_sb_lock);
	xfs_mod_sb(tp, sbfields);

	if ((error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES))) {
		xfs_alert(mp, "%s failed (error %d)!", __func__, error);
		return error;
	}
	return 0;
}


STATIC void
xfs_qm_reset_dqcounts(
	xfs_mount_t	*mp,
	xfs_buf_t	*bp,
	xfs_dqid_t	id,
	uint		type)
{
	xfs_disk_dquot_t	*ddq;
	int			j;

	trace_xfs_reset_dqcounts(bp, _RET_IP_);

	/*
	 * Reset all counters and timers. They'll be
	 * started afresh by xfs_qm_quotacheck.
	 */
#ifdef DEBUG
	j = XFS_FSB_TO_B(mp, XFS_DQUOT_CLUSTER_SIZE_FSB);
	do_div(j, sizeof(xfs_dqblk_t));
	ASSERT(mp->m_quotainfo->qi_dqperchunk == j);
#endif
	ddq = bp->b_addr;
	for (j = 0; j < mp->m_quotainfo->qi_dqperchunk; j++) {
		/*
		 * Do a sanity check, and if needed, repair the dqblk. Don't
		 * output any warnings because it's perfectly possible to
		 * find uninitialised dquot blks. See comment in xfs_qm_dqcheck.
		 */
		(void) xfs_qm_dqcheck(mp, ddq, id+j, type, XFS_QMOPT_DQREPAIR,
				      "xfs_quotacheck");
		ddq->d_bcount = 0;
		ddq->d_icount = 0;
		ddq->d_rtbcount = 0;
		ddq->d_btimer = 0;
		ddq->d_itimer = 0;
		ddq->d_rtbtimer = 0;
		ddq->d_bwarns = 0;
		ddq->d_iwarns = 0;
		ddq->d_rtbwarns = 0;
		ddq = (xfs_disk_dquot_t *) ((xfs_dqblk_t *)ddq + 1);
	}
}

STATIC int
xfs_qm_dqiter_bufs(
	xfs_mount_t	*mp,
	xfs_dqid_t	firstid,
	xfs_fsblock_t	bno,
	xfs_filblks_t	blkcnt,
	uint		flags)
{
	xfs_buf_t	*bp;
	int		error;
	int		type;

	ASSERT(blkcnt > 0);
	type = flags & XFS_QMOPT_UQUOTA ? XFS_DQ_USER :
		(flags & XFS_QMOPT_PQUOTA ? XFS_DQ_PROJ : XFS_DQ_GROUP);
	error = 0;

	/*
	 * Blkcnt arg can be a very big number, and might even be
	 * larger than the log itself. So, we have to break it up into
	 * manageable-sized transactions.
	 * Note that we don't start a permanent transaction here; we might
	 * not be able to get a log reservation for the whole thing up front,
	 * and we don't really care to either, because we just discard
	 * everything if we were to crash in the middle of this loop.
	 */
	while (blkcnt--) {
		error = xfs_trans_read_buf(mp, NULL, mp->m_ddev_targp,
			      XFS_FSB_TO_DADDR(mp, bno),
			      mp->m_quotainfo->qi_dqchunklen, 0, &bp);
		if (error)
			break;

		xfs_qm_reset_dqcounts(mp, bp, firstid, type);
		xfs_buf_delwri_queue(bp);
		xfs_buf_relse(bp);
		/*
		 * goto the next block.
		 */
		bno++;
		firstid += mp->m_quotainfo->qi_dqperchunk;
	}
	return error;
}

/*
 * Iterate over all allocated USR/GRP/PRJ dquots in the system, calling a
 * caller supplied function for every chunk of dquots that we find.
 */
STATIC int
xfs_qm_dqiterate(
	xfs_mount_t	*mp,
	xfs_inode_t	*qip,
	uint		flags)
{
	xfs_bmbt_irec_t		*map;
	int			i, nmaps;	/* number of map entries */
	int			error;		/* return value */
	xfs_fileoff_t		lblkno;
	xfs_filblks_t		maxlblkcnt;
	xfs_dqid_t		firstid;
	xfs_fsblock_t		rablkno;
	xfs_filblks_t		rablkcnt;

	error = 0;
	/*
	 * This looks racy, but we can't keep an inode lock across a
	 * trans_reserve. But, this gets called during quotacheck, and that
	 * happens only at mount time which is single threaded.
	 */
	if (qip->i_d.di_nblocks == 0)
		return 0;

	map = kmem_alloc(XFS_DQITER_MAP_SIZE * sizeof(*map), KM_SLEEP);

	lblkno = 0;
	maxlblkcnt = XFS_B_TO_FSB(mp, (xfs_ufsize_t)XFS_MAXIOFFSET(mp));
	do {
		nmaps = XFS_DQITER_MAP_SIZE;
		/*
		 * We aren't changing the inode itself. Just changing
		 * some of its data. No new blocks are added here, and
		 * the inode is never added to the transaction.
		 */
		xfs_ilock(qip, XFS_ILOCK_SHARED);
		error = xfs_bmapi_read(qip, lblkno, maxlblkcnt - lblkno,
				       map, &nmaps, 0);
		xfs_iunlock(qip, XFS_ILOCK_SHARED);
		if (error)
			break;

		ASSERT(nmaps <= XFS_DQITER_MAP_SIZE);
		for (i = 0; i < nmaps; i++) {
			ASSERT(map[i].br_startblock != DELAYSTARTBLOCK);
			ASSERT(map[i].br_blockcount);


			lblkno += map[i].br_blockcount;

			if (map[i].br_startblock == HOLESTARTBLOCK)
				continue;

			firstid = (xfs_dqid_t) map[i].br_startoff *
				mp->m_quotainfo->qi_dqperchunk;
			/*
			 * Do a read-ahead on the next extent.
			 */
			if ((i+1 < nmaps) &&
			    (map[i+1].br_startblock != HOLESTARTBLOCK)) {
				rablkcnt =  map[i+1].br_blockcount;
				rablkno = map[i+1].br_startblock;
				while (rablkcnt--) {
					xfs_buf_readahead(mp->m_ddev_targp,
					       XFS_FSB_TO_DADDR(mp, rablkno),
					       mp->m_quotainfo->qi_dqchunklen);
					rablkno++;
				}
			}
			/*
			 * Iterate thru all the blks in the extent and
			 * reset the counters of all the dquots inside them.
			 */
			if ((error = xfs_qm_dqiter_bufs(mp,
						       firstid,
						       map[i].br_startblock,
						       map[i].br_blockcount,
						       flags))) {
				break;
			}
		}

		if (error)
			break;
	} while (nmaps > 0);

	kmem_free(map);

	return error;
}

/*
 * Called by dqusage_adjust in doing a quotacheck.
 *
 * Given the inode, and a dquot id this updates both the incore dqout as well
 * as the buffer copy. This is so that once the quotacheck is done, we can
 * just log all the buffers, as opposed to logging numerous updates to
 * individual dquots.
 */
STATIC int
xfs_qm_quotacheck_dqadjust(
	struct xfs_inode	*ip,
	xfs_dqid_t		id,
	uint			type,
	xfs_qcnt_t		nblks,
	xfs_qcnt_t		rtblks)
{
	struct xfs_mount	*mp = ip->i_mount;
	struct xfs_dquot	*dqp;
	int			error;

	error = xfs_qm_dqget(mp, ip, id, type,
			     XFS_QMOPT_DQALLOC | XFS_QMOPT_DOWARN, &dqp);
	if (error) {
		/*
		 * Shouldn't be able to turn off quotas here.
		 */
		ASSERT(error != ESRCH);
		ASSERT(error != ENOENT);
		return error;
	}

	trace_xfs_dqadjust(dqp);

	/*
	 * Adjust the inode count and the block count to reflect this inode's
	 * resource usage.
	 */
	be64_add_cpu(&dqp->q_core.d_icount, 1);
	dqp->q_res_icount++;
	if (nblks) {
		be64_add_cpu(&dqp->q_core.d_bcount, nblks);
		dqp->q_res_bcount += nblks;
	}
	if (rtblks) {
		be64_add_cpu(&dqp->q_core.d_rtbcount, rtblks);
		dqp->q_res_rtbcount += rtblks;
	}

	/*
	 * Set default limits, adjust timers (since we changed usages)
	 *
	 * There are no timers for the default values set in the root dquot.
	 */
	if (dqp->q_core.d_id) {
		xfs_qm_adjust_dqlimits(mp, &dqp->q_core);
		xfs_qm_adjust_dqtimers(mp, &dqp->q_core);
	}

	dqp->dq_flags |= XFS_DQ_DIRTY;
	xfs_qm_dqput(dqp);
	return 0;
}

STATIC int
xfs_qm_get_rtblks(
	xfs_inode_t	*ip,
	xfs_qcnt_t	*O_rtblks)
{
	xfs_filblks_t	rtblks;			/* total rt blks */
	xfs_extnum_t	idx;			/* extent record index */
	xfs_ifork_t	*ifp;			/* inode fork pointer */
	xfs_extnum_t	nextents;		/* number of extent entries */
	int		error;

	ASSERT(XFS_IS_REALTIME_INODE(ip));
	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
		if ((error = xfs_iread_extents(NULL, ip, XFS_DATA_FORK)))
			return error;
	}
	rtblks = 0;
	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
	for (idx = 0; idx < nextents; idx++)
		rtblks += xfs_bmbt_get_blockcount(xfs_iext_get_ext(ifp, idx));
	*O_rtblks = (xfs_qcnt_t)rtblks;
	return 0;
}

/*
 * callback routine supplied to bulkstat(). Given an inumber, find its
 * dquots and update them to account for resources taken by that inode.
 */
/* ARGSUSED */
STATIC int
xfs_qm_dqusage_adjust(
	xfs_mount_t	*mp,		/* mount point for filesystem */
	xfs_ino_t	ino,		/* inode number to get data for */
	void		__user *buffer,	/* not used */
	int		ubsize,		/* not used */
	int		*ubused,	/* not used */
	int		*res)		/* result code value */
{
	xfs_inode_t	*ip;
	xfs_qcnt_t	nblks, rtblks = 0;
	int		error;

	ASSERT(XFS_IS_QUOTA_RUNNING(mp));

	/*
	 * rootino must have its resources accounted for, not so with the quota
	 * inodes.
	 */
	if (ino == mp->m_sb.sb_uquotino || ino == mp->m_sb.sb_gquotino) {
		*res = BULKSTAT_RV_NOTHING;
		return XFS_ERROR(EINVAL);
	}

	/*
	 * We don't _need_ to take the ilock EXCL. However, the xfs_qm_dqget
	 * interface expects the inode to be exclusively locked because that's
	 * the case in all other instances. It's OK that we do this because
	 * quotacheck is done only at mount time.
	 */
	error = xfs_iget(mp, NULL, ino, 0, XFS_ILOCK_EXCL, &ip);
	if (error) {
		*res = BULKSTAT_RV_NOTHING;
		return error;
	}

	ASSERT(ip->i_delayed_blks == 0);

	if (XFS_IS_REALTIME_INODE(ip)) {
		/*
		 * Walk thru the extent list and count the realtime blocks.
		 */
		error = xfs_qm_get_rtblks(ip, &rtblks);
		if (error)
			goto error0;
	}

	nblks = (xfs_qcnt_t)ip->i_d.di_nblocks - rtblks;

	/*
	 * Add the (disk blocks and inode) resources occupied by this
	 * inode to its dquots. We do this adjustment in the incore dquot,
	 * and also copy the changes to its buffer.
	 * We don't care about putting these changes in a transaction
	 * envelope because if we crash in the middle of a 'quotacheck'
	 * we have to start from the beginning anyway.
	 * Once we're done, we'll log all the dquot bufs.
	 *
	 * The *QUOTA_ON checks below may look pretty racy, but quotachecks
	 * and quotaoffs don't race. (Quotachecks happen at mount time only).
	 */
	if (XFS_IS_UQUOTA_ON(mp)) {
		error = xfs_qm_quotacheck_dqadjust(ip, ip->i_d.di_uid,
						   XFS_DQ_USER, nblks, rtblks);
		if (error)
			goto error0;
	}

	if (XFS_IS_GQUOTA_ON(mp)) {
		error = xfs_qm_quotacheck_dqadjust(ip, ip->i_d.di_gid,
						   XFS_DQ_GROUP, nblks, rtblks);
		if (error)
			goto error0;
	}

	if (XFS_IS_PQUOTA_ON(mp)) {
		error = xfs_qm_quotacheck_dqadjust(ip, xfs_get_projid(ip),
						   XFS_DQ_PROJ, nblks, rtblks);
		if (error)
			goto error0;
	}

	xfs_iunlock(ip, XFS_ILOCK_EXCL);
	IRELE(ip);
	*res = BULKSTAT_RV_DIDONE;
	return 0;

error0:
	xfs_iunlock(ip, XFS_ILOCK_EXCL);
	IRELE(ip);
	*res = BULKSTAT_RV_GIVEUP;
	return error;
}

/*
 * Walk thru all the filesystem inodes and construct a consistent view
 * of the disk quota world. If the quotacheck fails, disable quotas.
 */
int
xfs_qm_quotacheck(
	xfs_mount_t	*mp)
{
	int		done, count, error;
	xfs_ino_t	lastino;
	size_t		structsz;
	xfs_inode_t	*uip, *gip;
	uint		flags;

	count = INT_MAX;
	structsz = 1;
	lastino = 0;
	flags = 0;

	ASSERT(mp->m_quotainfo->qi_uquotaip || mp->m_quotainfo->qi_gquotaip);
	ASSERT(XFS_IS_QUOTA_RUNNING(mp));

	/*
	 * There should be no cached dquots. The (simplistic) quotacheck
	 * algorithm doesn't like that.
	 */
	ASSERT(list_empty(&mp->m_quotainfo->qi_dqlist));

	xfs_notice(mp, "Quotacheck needed: Please wait.");

	/*
	 * First we go thru all the dquots on disk, USR and GRP/PRJ, and reset
	 * their counters to zero. We need a clean slate.
	 * We don't log our changes till later.
	 */
	uip = mp->m_quotainfo->qi_uquotaip;
	if (uip) {
		error = xfs_qm_dqiterate(mp, uip, XFS_QMOPT_UQUOTA);
		if (error)
			goto error_return;
		flags |= XFS_UQUOTA_CHKD;
	}

	gip = mp->m_quotainfo->qi_gquotaip;
	if (gip) {
		error = xfs_qm_dqiterate(mp, gip, XFS_IS_GQUOTA_ON(mp) ?
					XFS_QMOPT_GQUOTA : XFS_QMOPT_PQUOTA);
		if (error)
			goto error_return;
		flags |= XFS_OQUOTA_CHKD;
	}

	do {
		/*
		 * Iterate thru all the inodes in the file system,
		 * adjusting the corresponding dquot counters in core.
		 */
		error = xfs_bulkstat(mp, &lastino, &count,
				     xfs_qm_dqusage_adjust,
				     structsz, NULL, &done);
		if (error)
			break;

	} while (!done);

	/*
	 * We've made all the changes that we need to make incore.
	 * Flush them down to disk buffers if everything was updated
	 * successfully.
	 */
	if (!error)
		error = xfs_qm_dqflush_all(mp, 0);

	/*
	 * We can get this error if we couldn't do a dquot allocation inside
	 * xfs_qm_dqusage_adjust (via bulkstat). We don't care about the
	 * dirty dquots that might be cached, we just want to get rid of them
	 * and turn quotaoff. The dquots won't be attached to any of the inodes
	 * at this point (because we intentionally didn't in dqget_noattach).
	 */
	if (error) {
		xfs_qm_dqpurge_all(mp, XFS_QMOPT_QUOTALL);
		goto error_return;
	}

	/*
	 * We didn't log anything, because if we crashed, we'll have to
	 * start the quotacheck from scratch anyway. However, we must make
	 * sure that our dquot changes are secure before we put the
	 * quotacheck'd stamp on the superblock. So, here we do a synchronous
	 * flush.
	 */
	xfs_flush_buftarg(mp->m_ddev_targp, 1);

	/*
	 * If one type of quotas is off, then it will lose its
	 * quotachecked status, since we won't be doing accounting for
	 * that type anymore.
	 */
	mp->m_qflags &= ~(XFS_OQUOTA_CHKD | XFS_UQUOTA_CHKD);
	mp->m_qflags |= flags;

 error_return:
	if (error) {
		xfs_warn(mp,
	"Quotacheck: Unsuccessful (Error %d): Disabling quotas.",
			error);
		/*
		 * We must turn off quotas.
		 */
		ASSERT(mp->m_quotainfo != NULL);
		ASSERT(xfs_Gqm != NULL);
		xfs_qm_destroy_quotainfo(mp);
		if (xfs_mount_reset_sbqflags(mp)) {
			xfs_warn(mp,
				"Quotacheck: Failed to reset quota flags.");
		}
	} else
		xfs_notice(mp, "Quotacheck: Done.");
	return (error);
}

/*
 * This is called after the superblock has been read in and we're ready to
 * iget the quota inodes.
 */
STATIC int
xfs_qm_init_quotainos(
	xfs_mount_t	*mp)
{
	xfs_inode_t	*uip, *gip;
	int		error;
	__int64_t	sbflags;
	uint		flags;

	ASSERT(mp->m_quotainfo);
	uip = gip = NULL;
	sbflags = 0;
	flags = 0;

	/*
	 * Get the uquota and gquota inodes
	 */
	if (xfs_sb_version_hasquota(&mp->m_sb)) {
		if (XFS_IS_UQUOTA_ON(mp) &&
		    mp->m_sb.sb_uquotino != NULLFSINO) {
			ASSERT(mp->m_sb.sb_uquotino > 0);
			if ((error = xfs_iget(mp, NULL, mp->m_sb.sb_uquotino,
					     0, 0, &uip)))
				return XFS_ERROR(error);
		}
		if (XFS_IS_OQUOTA_ON(mp) &&
		    mp->m_sb.sb_gquotino != NULLFSINO) {
			ASSERT(mp->m_sb.sb_gquotino > 0);
			if ((error = xfs_iget(mp, NULL, mp->m_sb.sb_gquotino,
					     0, 0, &gip))) {
				if (uip)
					IRELE(uip);
				return XFS_ERROR(error);
			}
		}
	} else {
		flags |= XFS_QMOPT_SBVERSION;
		sbflags |= (XFS_SB_VERSIONNUM | XFS_SB_UQUOTINO |
			    XFS_SB_GQUOTINO | XFS_SB_QFLAGS);
	}

	/*
	 * Create the two inodes, if they don't exist already. The changes
	 * made above will get added to a transaction and logged in one of
	 * the qino_alloc calls below.  If the device is readonly,
	 * temporarily switch to read-write to do this.
	 */
	if (XFS_IS_UQUOTA_ON(mp) && uip == NULL) {
		if ((error = xfs_qm_qino_alloc(mp, &uip,
					      sbflags | XFS_SB_UQUOTINO,
					      flags | XFS_QMOPT_UQUOTA)))
			return XFS_ERROR(error);

		flags &= ~XFS_QMOPT_SBVERSION;
	}
	if (XFS_IS_OQUOTA_ON(mp) && gip == NULL) {
		flags |= (XFS_IS_GQUOTA_ON(mp) ?
				XFS_QMOPT_GQUOTA : XFS_QMOPT_PQUOTA);
		error = xfs_qm_qino_alloc(mp, &gip,
					  sbflags | XFS_SB_GQUOTINO, flags);
		if (error) {
			if (uip)
				IRELE(uip);

			return XFS_ERROR(error);
		}
	}

	mp->m_quotainfo->qi_uquotaip = uip;
	mp->m_quotainfo->qi_gquotaip = gip;

	return 0;
}



/*
 * Just pop the least recently used dquot off the freelist and
 * recycle it. The returned dquot is locked.
 */
STATIC xfs_dquot_t *
xfs_qm_dqreclaim_one(void)
{
	xfs_dquot_t	*dqpout;
	xfs_dquot_t	*dqp;
	int		restarts;
	int		startagain;

	restarts = 0;
	dqpout = NULL;

	/* lockorder: hashchainlock, freelistlock, mplistlock, dqlock, dqflock */
again:
	startagain = 0;
	mutex_lock(&xfs_Gqm->qm_dqfrlist_lock);

	list_for_each_entry(dqp, &xfs_Gqm->qm_dqfrlist, q_freelist) {
		struct xfs_mount *mp = dqp->q_mount;
		xfs_dqlock(dqp);

		/*
		 * We are racing with dqlookup here. Naturally we don't
		 * want to reclaim a dquot that lookup wants. We release the
		 * freelist lock and start over, so that lookup will grab
		 * both the dquot and the freelistlock.
		 */
		if (dqp->dq_flags & XFS_DQ_WANT) {
			ASSERT(! (dqp->dq_flags & XFS_DQ_INACTIVE));

			trace_xfs_dqreclaim_want(dqp);
			XQM_STATS_INC(xqmstats.xs_qm_dqwants);
			restarts++;
			startagain = 1;
			goto dqunlock;
		}

		/*
		 * If the dquot is inactive, we are assured that it is
		 * not on the mplist or the hashlist, and that makes our
		 * life easier.
		 */
		if (dqp->dq_flags & XFS_DQ_INACTIVE) {
			ASSERT(mp == NULL);
			ASSERT(! XFS_DQ_IS_DIRTY(dqp));
			ASSERT(list_empty(&dqp->q_hashlist));
			ASSERT(list_empty(&dqp->q_mplist));
			list_del_init(&dqp->q_freelist);
			xfs_Gqm->qm_dqfrlist_cnt--;
			dqpout = dqp;
			XQM_STATS_INC(xqmstats.xs_qm_dqinact_reclaims);
			goto dqunlock;
		}

		ASSERT(dqp->q_hash);
		ASSERT(!list_empty(&dqp->q_mplist));

		/*
		 * Try to grab the flush lock. If this dquot is in the process
		 * of getting flushed to disk, we don't want to reclaim it.
		 */
		if (!xfs_dqflock_nowait(dqp))
			goto dqunlock;

		/*
		 * We have the flush lock so we know that this is not in the
		 * process of being flushed. So, if this is dirty, flush it
		 * DELWRI so that we don't get a freelist infested with
		 * dirty dquots.
		 */
		if (XFS_DQ_IS_DIRTY(dqp)) {
			int	error;

			trace_xfs_dqreclaim_dirty(dqp);

			/*
			 * We flush it delayed write, so don't bother
			 * releasing the freelist lock.
			 */
			error = xfs_qm_dqflush(dqp, 0);
			if (error) {
				xfs_warn(mp, "%s: dquot %p flush failed",
					__func__, dqp);
			}
			goto dqunlock;
		}

		/*
		 * We're trying to get the hashlock out of order. This races
		 * with dqlookup; so, we giveup and goto the next dquot if
		 * we couldn't get the hashlock. This way, we won't starve
		 * a dqlookup process that holds the hashlock that is
		 * waiting for the freelist lock.
		 */
		if (!mutex_trylock(&dqp->q_hash->qh_lock)) {
			restarts++;
			goto dqfunlock;
		}

		/*
		 * This races with dquot allocation code as well as dqflush_all
		 * and reclaim code. So, if we failed to grab the mplist lock,
		 * giveup everything and start over.
		 */
		if (!mutex_trylock(&mp->m_quotainfo->qi_dqlist_lock)) {
			restarts++;
			startagain = 1;
			goto qhunlock;
		}

		ASSERT(dqp->q_nrefs == 0);
		list_del_init(&dqp->q_mplist);
		mp->m_quotainfo->qi_dquots--;
		mp->m_quotainfo->qi_dqreclaims++;
		list_del_init(&dqp->q_hashlist);
		dqp->q_hash->qh_version++;
		list_del_init(&dqp->q_freelist);
		xfs_Gqm->qm_dqfrlist_cnt--;
		dqpout = dqp;
		mutex_unlock(&mp->m_quotainfo->qi_dqlist_lock);
qhunlock:
		mutex_unlock(&dqp->q_hash->qh_lock);
dqfunlock:
		xfs_dqfunlock(dqp);
dqunlock:
		xfs_dqunlock(dqp);
		if (dqpout)
			break;
		if (restarts >= XFS_QM_RECLAIM_MAX_RESTARTS)
			break;
		if (startagain) {
			mutex_unlock(&xfs_Gqm->qm_dqfrlist_lock);
			goto again;
		}
	}
	mutex_unlock(&xfs_Gqm->qm_dqfrlist_lock);
	return dqpout;
}

/*
 * Traverse the freelist of dquots and attempt to reclaim a maximum of
 * 'howmany' dquots. This operation races with dqlookup(), and attempts to
 * favor the lookup function ...
 */
STATIC int
xfs_qm_shake_freelist(
	int	howmany)
{
	int		nreclaimed = 0;
	xfs_dquot_t	*dqp;

	if (howmany <= 0)
		return 0;

	while (nreclaimed < howmany) {
		dqp = xfs_qm_dqreclaim_one();
		if (!dqp)
			return nreclaimed;
		xfs_qm_dqdestroy(dqp);
		nreclaimed++;
	}
	return nreclaimed;
}

/*
 * The kmem_shake interface is invoked when memory is running low.
 */
/* ARGSUSED */
STATIC int
xfs_qm_shake(
	struct shrinker	*shrink,
	struct shrink_control *sc)
{
	int	ndqused, nfree, n;
	gfp_t gfp_mask = sc->gfp_mask;

	if (!kmem_shake_allow(gfp_mask))
		return 0;
	if (!xfs_Gqm)
		return 0;

	nfree = xfs_Gqm->qm_dqfrlist_cnt; /* free dquots */
	/* incore dquots in all f/s's */
	ndqused = atomic_read(&xfs_Gqm->qm_totaldquots) - nfree;

	ASSERT(ndqused >= 0);

	if (nfree <= ndqused && nfree < ndquot)
		return 0;

	ndqused *= xfs_Gqm->qm_dqfree_ratio;	/* target # of free dquots */
	n = nfree - ndqused - ndquot;		/* # over target */

	return xfs_qm_shake_freelist(MAX(nfree, n));
}


/*------------------------------------------------------------------*/

/*
 * Return a new incore dquot. Depending on the number of
 * dquots in the system, we either allocate a new one on the kernel heap,
 * or reclaim a free one.
 * Return value is B_TRUE if we allocated a new dquot, B_FALSE if we managed
 * to reclaim an existing one from the freelist.
 */
boolean_t
xfs_qm_dqalloc_incore(
	xfs_dquot_t **O_dqpp)
{
	xfs_dquot_t	*dqp;

	/*
	 * Check against high water mark to see if we want to pop
	 * a nincompoop dquot off the freelist.
	 */
	if (atomic_read(&xfs_Gqm->qm_totaldquots) >= ndquot) {
		/*
		 * Try to recycle a dquot from the freelist.
		 */
		if ((dqp = xfs_qm_dqreclaim_one())) {
			XQM_STATS_INC(xqmstats.xs_qm_dqreclaims);
			/*
			 * Just zero the core here. The rest will get
			 * reinitialized by caller. XXX we shouldn't even
			 * do this zero ...
			 */
			memset(&dqp->q_core, 0, sizeof(dqp->q_core));
			*O_dqpp = dqp;
			return B_FALSE;
		}
		XQM_STATS_INC(xqmstats.xs_qm_dqreclaim_misses);
	}

	/*
	 * Allocate a brand new dquot on the kernel heap and return it
	 * to the caller to initialize.
	 */
	ASSERT(xfs_Gqm->qm_dqzone != NULL);
	*O_dqpp = kmem_zone_zalloc(xfs_Gqm->qm_dqzone, KM_SLEEP);
	atomic_inc(&xfs_Gqm->qm_totaldquots);

	return B_TRUE;
}


/*
 * Start a transaction and write the incore superblock changes to
 * disk. flags parameter indicates which fields have changed.
 */
int
xfs_qm_write_sb_changes(
	xfs_mount_t	*mp,
	__int64_t	flags)
{
	xfs_trans_t	*tp;
	int		error;

	tp = xfs_trans_alloc(mp, XFS_TRANS_QM_SBCHANGE);
	if ((error = xfs_trans_reserve(tp, 0,
				      mp->m_sb.sb_sectsize + 128, 0,
				      0,
				      XFS_DEFAULT_LOG_COUNT))) {
		xfs_trans_cancel(tp, 0);
		return error;
	}

	xfs_mod_sb(tp, flags);
	error = xfs_trans_commit(tp, 0);

	return error;
}


/* --------------- utility functions for vnodeops ---------------- */


/*
 * Given an inode, a uid, gid and prid make sure that we have
 * allocated relevant dquot(s) on disk, and that we won't exceed inode
 * quotas by creating this file.
 * This also attaches dquot(s) to the given inode after locking it,
 * and returns the dquots corresponding to the uid and/or gid.
 *
 * in	: inode (unlocked)
 * out	: udquot, gdquot with references taken and unlocked
 */
int
xfs_qm_vop_dqalloc(
	struct xfs_inode	*ip,
	uid_t			uid,
	gid_t			gid,
	prid_t			prid,
	uint			flags,
	struct xfs_dquot	**O_udqpp,
	struct xfs_dquot	**O_gdqpp)
{
	struct xfs_mount	*mp = ip->i_mount;
	struct xfs_dquot	*uq, *gq;
	int			error;
	uint			lockflags;

	if (!XFS_IS_QUOTA_RUNNING(mp) || !XFS_IS_QUOTA_ON(mp))
		return 0;

	lockflags = XFS_ILOCK_EXCL;
	xfs_ilock(ip, lockflags);

	if ((flags & XFS_QMOPT_INHERIT) && XFS_INHERIT_GID(ip))
		gid = ip->i_d.di_gid;

	/*
	 * Attach the dquot(s) to this inode, doing a dquot allocation
	 * if necessary. The dquot(s) will not be locked.
	 */
	if (XFS_NOT_DQATTACHED(mp, ip)) {
		error = xfs_qm_dqattach_locked(ip, XFS_QMOPT_DQALLOC);
		if (error) {
			xfs_iunlock(ip, lockflags);
			return error;
		}
	}

	uq = gq = NULL;
	if ((flags & XFS_QMOPT_UQUOTA) && XFS_IS_UQUOTA_ON(mp)) {
		if (ip->i_d.di_uid != uid) {
			/*
			 * What we need is the dquot that has this uid, and
			 * if we send the inode to dqget, the uid of the inode
			 * takes priority over what's sent in the uid argument.
			 * We must unlock inode here before calling dqget if
			 * we're not sending the inode, because otherwise
			 * we'll deadlock by doing trans_reserve while
			 * holding ilock.
			 */
			xfs_iunlock(ip, lockflags);
			if ((error = xfs_qm_dqget(mp, NULL, (xfs_dqid_t) uid,
						 XFS_DQ_USER,
						 XFS_QMOPT_DQALLOC |
						 XFS_QMOPT_DOWARN,
						 &uq))) {
				ASSERT(error != ENOENT);
				return error;
			}
			/*
			 * Get the ilock in the right order.
			 */
			xfs_dqunlock(uq);
			lockflags = XFS_ILOCK_SHARED;
			xfs_ilock(ip, lockflags);
		} else {
			/*
			 * Take an extra reference, because we'll return
			 * this to caller
			 */
			ASSERT(ip->i_udquot);
			uq = ip->i_udquot;
			xfs_dqlock(uq);
			XFS_DQHOLD(uq);
			xfs_dqunlock(uq);
		}
	}
	if ((flags & XFS_QMOPT_GQUOTA) && XFS_IS_GQUOTA_ON(mp)) {
		if (ip->i_d.di_gid != gid) {
			xfs_iunlock(ip, lockflags);
			if ((error = xfs_qm_dqget(mp, NULL, (xfs_dqid_t)gid,
						 XFS_DQ_GROUP,
						 XFS_QMOPT_DQALLOC |
						 XFS_QMOPT_DOWARN,
						 &gq))) {
				if (uq)
					xfs_qm_dqrele(uq);
				ASSERT(error != ENOENT);
				return error;
			}
			xfs_dqunlock(gq);
			lockflags = XFS_ILOCK_SHARED;
			xfs_ilock(ip, lockflags);
		} else {
			ASSERT(ip->i_gdquot);
			gq = ip->i_gdquot;
			xfs_dqlock(gq);
			XFS_DQHOLD(gq);
			xfs_dqunlock(gq);
		}
	} else if ((flags & XFS_QMOPT_PQUOTA) && XFS_IS_PQUOTA_ON(mp)) {
		if (xfs_get_projid(ip) != prid) {
			xfs_iunlock(ip, lockflags);
			if ((error = xfs_qm_dqget(mp, NULL, (xfs_dqid_t)prid,
						 XFS_DQ_PROJ,
						 XFS_QMOPT_DQALLOC |
						 XFS_QMOPT_DOWARN,
						 &gq))) {
				if (uq)
					xfs_qm_dqrele(uq);
				ASSERT(error != ENOENT);
				return (error);
			}
			xfs_dqunlock(gq);
			lockflags = XFS_ILOCK_SHARED;
			xfs_ilock(ip, lockflags);
		} else {
			ASSERT(ip->i_gdquot);
			gq = ip->i_gdquot;
			xfs_dqlock(gq);
			XFS_DQHOLD(gq);
			xfs_dqunlock(gq);
		}
	}
	if (uq)
		trace_xfs_dquot_dqalloc(ip);

	xfs_iunlock(ip, lockflags);
	if (O_udqpp)
		*O_udqpp = uq;
	else if (uq)
		xfs_qm_dqrele(uq);
	if (O_gdqpp)
		*O_gdqpp = gq;
	else if (gq)
		xfs_qm_dqrele(gq);
	return 0;
}

/*
 * Actually transfer ownership, and do dquot modifications.
 * These were already reserved.
 */
xfs_dquot_t *
xfs_qm_vop_chown(
	xfs_trans_t	*tp,
	xfs_inode_t	*ip,
	xfs_dquot_t	**IO_olddq,
	xfs_dquot_t	*newdq)
{
	xfs_dquot_t	*prevdq;
	uint		bfield = XFS_IS_REALTIME_INODE(ip) ?
				 XFS_TRANS_DQ_RTBCOUNT : XFS_TRANS_DQ_BCOUNT;


	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
	ASSERT(XFS_IS_QUOTA_RUNNING(ip->i_mount));

	/* old dquot */
	prevdq = *IO_olddq;
	ASSERT(prevdq);
	ASSERT(prevdq != newdq);

	xfs_trans_mod_dquot(tp, prevdq, bfield, -(ip->i_d.di_nblocks));
	xfs_trans_mod_dquot(tp, prevdq, XFS_TRANS_DQ_ICOUNT, -1);

	/* the sparkling new dquot */
	xfs_trans_mod_dquot(tp, newdq, bfield, ip->i_d.di_nblocks);
	xfs_trans_mod_dquot(tp, newdq, XFS_TRANS_DQ_ICOUNT, 1);

	/*
	 * Take an extra reference, because the inode
	 * is going to keep this dquot pointer even
	 * after the trans_commit.
	 */
	xfs_dqlock(newdq);
	XFS_DQHOLD(newdq);
	xfs_dqunlock(newdq);
	*IO_olddq = newdq;

	return prevdq;
}

/*
 * Quota reservations for setattr(AT_UID|AT_GID|AT_PROJID).
 */
int
xfs_qm_vop_chown_reserve(
	xfs_trans_t	*tp,
	xfs_inode_t	*ip,
	xfs_dquot_t	*udqp,
	xfs_dquot_t	*gdqp,
	uint		flags)
{
	xfs_mount_t	*mp = ip->i_mount;
	uint		delblks, blkflags, prjflags = 0;
	xfs_dquot_t	*unresudq, *unresgdq, *delblksudq, *delblksgdq;
	int		error;


	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED));
	ASSERT(XFS_IS_QUOTA_RUNNING(mp));

	delblks = ip->i_delayed_blks;
	delblksudq = delblksgdq = unresudq = unresgdq = NULL;
	blkflags = XFS_IS_REALTIME_INODE(ip) ?
			XFS_QMOPT_RES_RTBLKS : XFS_QMOPT_RES_REGBLKS;

	if (XFS_IS_UQUOTA_ON(mp) && udqp &&
	    ip->i_d.di_uid != (uid_t)be32_to_cpu(udqp->q_core.d_id)) {
		delblksudq = udqp;
		/*
		 * If there are delayed allocation blocks, then we have to
		 * unreserve those from the old dquot, and add them to the
		 * new dquot.
		 */
		if (delblks) {
			ASSERT(ip->i_udquot);
			unresudq = ip->i_udquot;
		}
	}
	if (XFS_IS_OQUOTA_ON(ip->i_mount) && gdqp) {
		if (XFS_IS_PQUOTA_ON(ip->i_mount) &&
		     xfs_get_projid(ip) != be32_to_cpu(gdqp->q_core.d_id))
			prjflags = XFS_QMOPT_ENOSPC;

		if (prjflags ||
		    (XFS_IS_GQUOTA_ON(ip->i_mount) &&
		     ip->i_d.di_gid != be32_to_cpu(gdqp->q_core.d_id))) {
			delblksgdq = gdqp;
			if (delblks) {
				ASSERT(ip->i_gdquot);
				unresgdq = ip->i_gdquot;
			}
		}
	}

	if ((error = xfs_trans_reserve_quota_bydquots(tp, ip->i_mount,
				delblksudq, delblksgdq, ip->i_d.di_nblocks, 1,
				flags | blkflags | prjflags)))
		return (error);

	/*
	 * Do the delayed blks reservations/unreservations now. Since, these
	 * are done without the help of a transaction, if a reservation fails
	 * its previous reservations won't be automatically undone by trans
	 * code. So, we have to do it manually here.
	 */
	if (delblks) {
		/*
		 * Do the reservations first. Unreservation can't fail.
		 */
		ASSERT(delblksudq || delblksgdq);
		ASSERT(unresudq || unresgdq);
		if ((error = xfs_trans_reserve_quota_bydquots(NULL, ip->i_mount,
				delblksudq, delblksgdq, (xfs_qcnt_t)delblks, 0,
				flags | blkflags | prjflags)))
			return (error);
		xfs_trans_reserve_quota_bydquots(NULL, ip->i_mount,
				unresudq, unresgdq, -((xfs_qcnt_t)delblks), 0,
				blkflags);
	}

	return (0);
}

int
xfs_qm_vop_rename_dqattach(
	struct xfs_inode	**i_tab)
{
	struct xfs_mount	*mp = i_tab[0]->i_mount;
	int			i;

	if (!XFS_IS_QUOTA_RUNNING(mp) || !XFS_IS_QUOTA_ON(mp))
		return 0;

	for (i = 0; (i < 4 && i_tab[i]); i++) {
		struct xfs_inode	*ip = i_tab[i];
		int			error;

		/*
		 * Watch out for duplicate entries in the table.
		 */
		if (i == 0 || ip != i_tab[i-1]) {
			if (XFS_NOT_DQATTACHED(mp, ip)) {
				error = xfs_qm_dqattach(ip, 0);
				if (error)
					return error;
			}
		}
	}
	return 0;
}

void
xfs_qm_vop_create_dqattach(
	struct xfs_trans	*tp,
	struct xfs_inode	*ip,
	struct xfs_dquot	*udqp,
	struct xfs_dquot	*gdqp)
{
	struct xfs_mount	*mp = tp->t_mountp;

	if (!XFS_IS_QUOTA_RUNNING(mp) || !XFS_IS_QUOTA_ON(mp))
		return;

	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
	ASSERT(XFS_IS_QUOTA_RUNNING(mp));

	if (udqp) {
		xfs_dqlock(udqp);
		XFS_DQHOLD(udqp);
		xfs_dqunlock(udqp);
		ASSERT(ip->i_udquot == NULL);
		ip->i_udquot = udqp;
		ASSERT(XFS_IS_UQUOTA_ON(mp));
		ASSERT(ip->i_d.di_uid == be32_to_cpu(udqp->q_core.d_id));
		xfs_trans_mod_dquot(tp, udqp, XFS_TRANS_DQ_ICOUNT, 1);
	}
	if (gdqp) {
		xfs_dqlock(gdqp);
		XFS_DQHOLD(gdqp);
		xfs_dqunlock(gdqp);
		ASSERT(ip->i_gdquot == NULL);
		ip->i_gdquot = gdqp;
		ASSERT(XFS_IS_OQUOTA_ON(mp));
		ASSERT((XFS_IS_GQUOTA_ON(mp) ?
			ip->i_d.di_gid : xfs_get_projid(ip)) ==
				be32_to_cpu(gdqp->q_core.d_id));
		xfs_trans_mod_dquot(tp, gdqp, XFS_TRANS_DQ_ICOUNT, 1);
	}
}