aboutsummaryrefslogtreecommitdiff
path: root/sys/contrib/openzfs/module/zfs/dsl_dir.c
blob: 5a64e399cf94bccfa4dd3d09ebf308b0bf6a266a (plain) (blame)
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
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
/*
 * CDDL HEADER START
 *
 * The contents of this file are subject to the terms of the
 * Common Development and Distribution License (the "License").
 * You may not use this file except in compliance with the License.
 *
 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
 * or https://opensource.org/licenses/CDDL-1.0.
 * See the License for the specific language governing permissions
 * and limitations under the License.
 *
 * When distributing Covered Code, include this CDDL HEADER in each
 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
 * If applicable, add the following below this CDDL HEADER, with the
 * fields enclosed by brackets "[]" replaced with your own identifying
 * information: Portions Copyright [yyyy] [name of copyright owner]
 *
 * CDDL HEADER END
 */
/*
 * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
 * Copyright (c) 2012, 2018 by Delphix. All rights reserved.
 * Copyright (c) 2013 Martin Matuska. All rights reserved.
 * Copyright (c) 2014 Joyent, Inc. All rights reserved.
 * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved.
 * Copyright (c) 2016 Actifio, Inc. All rights reserved.
 * Copyright (c) 2018, loli10K <ezomori.nozomu@gmail.com>. All rights reserved.
 */

#include <sys/dmu.h>
#include <sys/dmu_objset.h>
#include <sys/dmu_tx.h>
#include <sys/dsl_dataset.h>
#include <sys/dsl_dir.h>
#include <sys/dsl_prop.h>
#include <sys/dsl_synctask.h>
#include <sys/dsl_deleg.h>
#include <sys/dmu_impl.h>
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/metaslab.h>
#include <sys/zap.h>
#include <sys/zio.h>
#include <sys/arc.h>
#include <sys/sunddi.h>
#include <sys/zfeature.h>
#include <sys/policy.h>
#include <sys/zfs_vfsops.h>
#include <sys/zfs_znode.h>
#include <sys/zvol.h>
#include <sys/zthr.h>
#include "zfs_namecheck.h"
#include "zfs_prop.h"

/*
 * Filesystem and Snapshot Limits
 * ------------------------------
 *
 * These limits are used to restrict the number of filesystems and/or snapshots
 * that can be created at a given level in the tree or below. A typical
 * use-case is with a delegated dataset where the administrator wants to ensure
 * that a user within the zone is not creating too many additional filesystems
 * or snapshots, even though they're not exceeding their space quota.
 *
 * The filesystem and snapshot counts are stored as extensible properties. This
 * capability is controlled by a feature flag and must be enabled to be used.
 * Once enabled, the feature is not active until the first limit is set. At
 * that point, future operations to create/destroy filesystems or snapshots
 * will validate and update the counts.
 *
 * Because the count properties will not exist before the feature is active,
 * the counts are updated when a limit is first set on an uninitialized
 * dsl_dir node in the tree (The filesystem/snapshot count on a node includes
 * all of the nested filesystems/snapshots. Thus, a new leaf node has a
 * filesystem count of 0 and a snapshot count of 0. Non-existent filesystem and
 * snapshot count properties on a node indicate uninitialized counts on that
 * node.) When first setting a limit on an uninitialized node, the code starts
 * at the filesystem with the new limit and descends into all sub-filesystems
 * to add the count properties.
 *
 * In practice this is lightweight since a limit is typically set when the
 * filesystem is created and thus has no children. Once valid, changing the
 * limit value won't require a re-traversal since the counts are already valid.
 * When recursively fixing the counts, if a node with a limit is encountered
 * during the descent, the counts are known to be valid and there is no need to
 * descend into that filesystem's children. The counts on filesystems above the
 * one with the new limit will still be uninitialized, unless a limit is
 * eventually set on one of those filesystems. The counts are always recursively
 * updated when a limit is set on a dataset, unless there is already a limit.
 * When a new limit value is set on a filesystem with an existing limit, it is
 * possible for the new limit to be less than the current count at that level
 * since a user who can change the limit is also allowed to exceed the limit.
 *
 * Once the feature is active, then whenever a filesystem or snapshot is
 * created, the code recurses up the tree, validating the new count against the
 * limit at each initialized level. In practice, most levels will not have a
 * limit set. If there is a limit at any initialized level up the tree, the
 * check must pass or the creation will fail. Likewise, when a filesystem or
 * snapshot is destroyed, the counts are recursively adjusted all the way up
 * the initialized nodes in the tree. Renaming a filesystem into different point
 * in the tree will first validate, then update the counts on each branch up to
 * the common ancestor. A receive will also validate the counts and then update
 * them.
 *
 * An exception to the above behavior is that the limit is not enforced if the
 * user has permission to modify the limit. This is primarily so that
 * recursive snapshots in the global zone always work. We want to prevent a
 * denial-of-service in which a lower level delegated dataset could max out its
 * limit and thus block recursive snapshots from being taken in the global zone.
 * Because of this, it is possible for the snapshot count to be over the limit
 * and snapshots taken in the global zone could cause a lower level dataset to
 * hit or exceed its limit. The administrator taking the global zone recursive
 * snapshot should be aware of this side-effect and behave accordingly.
 * For consistency, the filesystem limit is also not enforced if the user can
 * modify the limit.
 *
 * The filesystem and snapshot limits are validated by dsl_fs_ss_limit_check()
 * and updated by dsl_fs_ss_count_adjust(). A new limit value is setup in
 * dsl_dir_activate_fs_ss_limit() and the counts are adjusted, if necessary, by
 * dsl_dir_init_fs_ss_count().
 */

static uint64_t dsl_dir_space_towrite(dsl_dir_t *dd);

typedef struct ddulrt_arg {
	dsl_dir_t	*ddulrta_dd;
	uint64_t	ddlrta_txg;
} ddulrt_arg_t;

static void
dsl_dir_evict_async(void *dbu)
{
	dsl_dir_t *dd = dbu;
	int t;
	dsl_pool_t *dp __maybe_unused = dd->dd_pool;

	dd->dd_dbuf = NULL;

	for (t = 0; t < TXG_SIZE; t++) {
		ASSERT(!txg_list_member(&dp->dp_dirty_dirs, dd, t));
		ASSERT(dd->dd_tempreserved[t] == 0);
		ASSERT(dd->dd_space_towrite[t] == 0);
	}

	if (dd->dd_parent)
		dsl_dir_async_rele(dd->dd_parent, dd);

	spa_async_close(dd->dd_pool->dp_spa, dd);

	if (dsl_deadlist_is_open(&dd->dd_livelist))
		dsl_dir_livelist_close(dd);

	dsl_prop_fini(dd);
	cv_destroy(&dd->dd_activity_cv);
	mutex_destroy(&dd->dd_activity_lock);
	mutex_destroy(&dd->dd_lock);
	kmem_free(dd, sizeof (dsl_dir_t));
}

int
dsl_dir_hold_obj(dsl_pool_t *dp, uint64_t ddobj,
    const char *tail, const void *tag, dsl_dir_t **ddp)
{
	dmu_buf_t *dbuf;
	dsl_dir_t *dd;
	dmu_object_info_t doi;
	int err;

	ASSERT(dsl_pool_config_held(dp));

	err = dmu_bonus_hold(dp->dp_meta_objset, ddobj, tag, &dbuf);
	if (err != 0)
		return (err);
	dd = dmu_buf_get_user(dbuf);

	dmu_object_info_from_db(dbuf, &doi);
	ASSERT3U(doi.doi_bonus_type, ==, DMU_OT_DSL_DIR);
	ASSERT3U(doi.doi_bonus_size, >=, sizeof (dsl_dir_phys_t));

	if (dd == NULL) {
		dsl_dir_t *winner;

		dd = kmem_zalloc(sizeof (dsl_dir_t), KM_SLEEP);
		dd->dd_object = ddobj;
		dd->dd_dbuf = dbuf;
		dd->dd_pool = dp;

		mutex_init(&dd->dd_lock, NULL, MUTEX_DEFAULT, NULL);
		mutex_init(&dd->dd_activity_lock, NULL, MUTEX_DEFAULT, NULL);
		cv_init(&dd->dd_activity_cv, NULL, CV_DEFAULT, NULL);
		dsl_prop_init(dd);

		if (dsl_dir_is_zapified(dd)) {
			err = zap_lookup(dp->dp_meta_objset,
			    ddobj, DD_FIELD_CRYPTO_KEY_OBJ,
			    sizeof (uint64_t), 1, &dd->dd_crypto_obj);
			if (err == 0) {
				/* check for on-disk format errata */
				if (dsl_dir_incompatible_encryption_version(
				    dd)) {
					dp->dp_spa->spa_errata =
					    ZPOOL_ERRATA_ZOL_6845_ENCRYPTION;
				}
			} else if (err != ENOENT) {
				goto errout;
			}
		}

		if (dsl_dir_phys(dd)->dd_parent_obj) {
			err = dsl_dir_hold_obj(dp,
			    dsl_dir_phys(dd)->dd_parent_obj, NULL, dd,
			    &dd->dd_parent);
			if (err != 0)
				goto errout;
			if (tail) {
#ifdef ZFS_DEBUG
				uint64_t foundobj;

				err = zap_lookup(dp->dp_meta_objset,
				    dsl_dir_phys(dd->dd_parent)->
				    dd_child_dir_zapobj, tail,
				    sizeof (foundobj), 1, &foundobj);
				ASSERT(err || foundobj == ddobj);
#endif
				(void) strlcpy(dd->dd_myname, tail,
				    sizeof (dd->dd_myname));
			} else {
				err = zap_value_search(dp->dp_meta_objset,
				    dsl_dir_phys(dd->dd_parent)->
				    dd_child_dir_zapobj,
				    ddobj, 0, dd->dd_myname);
			}
			if (err != 0)
				goto errout;
		} else {
			(void) strlcpy(dd->dd_myname, spa_name(dp->dp_spa),
			    sizeof (dd->dd_myname));
		}

		if (dsl_dir_is_clone(dd)) {
			dmu_buf_t *origin_bonus;
			dsl_dataset_phys_t *origin_phys;

			/*
			 * We can't open the origin dataset, because
			 * that would require opening this dsl_dir.
			 * Just look at its phys directly instead.
			 */
			err = dmu_bonus_hold(dp->dp_meta_objset,
			    dsl_dir_phys(dd)->dd_origin_obj, FTAG,
			    &origin_bonus);
			if (err != 0)
				goto errout;
			origin_phys = origin_bonus->db_data;
			dd->dd_origin_txg =
			    origin_phys->ds_creation_txg;
			dmu_buf_rele(origin_bonus, FTAG);
			if (dsl_dir_is_zapified(dd)) {
				uint64_t obj;
				err = zap_lookup(dp->dp_meta_objset,
				    dd->dd_object, DD_FIELD_LIVELIST,
				    sizeof (uint64_t), 1, &obj);
				if (err == 0)
					dsl_dir_livelist_open(dd, obj);
				else if (err != ENOENT)
					goto errout;
			}
		}

		inode_timespec_t t = {0};
		zap_lookup(dd->dd_pool->dp_meta_objset,
		    dsl_dir_phys(dd)->dd_props_zapobj,
		    zfs_prop_to_name(ZFS_PROP_SNAPSHOTS_CHANGED),
		    sizeof (uint64_t),
		    sizeof (inode_timespec_t) / sizeof (uint64_t), &t);
		dd->dd_snap_cmtime = t;

		dmu_buf_init_user(&dd->dd_dbu, NULL, dsl_dir_evict_async,
		    &dd->dd_dbuf);
		winner = dmu_buf_set_user_ie(dbuf, &dd->dd_dbu);
		if (winner != NULL) {
			if (dd->dd_parent)
				dsl_dir_rele(dd->dd_parent, dd);
			if (dsl_deadlist_is_open(&dd->dd_livelist))
				dsl_dir_livelist_close(dd);
			dsl_prop_fini(dd);
			cv_destroy(&dd->dd_activity_cv);
			mutex_destroy(&dd->dd_activity_lock);
			mutex_destroy(&dd->dd_lock);
			kmem_free(dd, sizeof (dsl_dir_t));
			dd = winner;
		} else {
			spa_open_ref(dp->dp_spa, dd);
		}
	}

	/*
	 * The dsl_dir_t has both open-to-close and instantiate-to-evict
	 * holds on the spa.  We need the open-to-close holds because
	 * otherwise the spa_refcnt wouldn't change when we open a
	 * dir which the spa also has open, so we could incorrectly
	 * think it was OK to unload/export/destroy the pool.  We need
	 * the instantiate-to-evict hold because the dsl_dir_t has a
	 * pointer to the dd_pool, which has a pointer to the spa_t.
	 */
	spa_open_ref(dp->dp_spa, tag);
	ASSERT3P(dd->dd_pool, ==, dp);
	ASSERT3U(dd->dd_object, ==, ddobj);
	ASSERT3P(dd->dd_dbuf, ==, dbuf);
	*ddp = dd;
	return (0);

errout:
	if (dd->dd_parent)
		dsl_dir_rele(dd->dd_parent, dd);
	if (dsl_deadlist_is_open(&dd->dd_livelist))
		dsl_dir_livelist_close(dd);
	dsl_prop_fini(dd);
	cv_destroy(&dd->dd_activity_cv);
	mutex_destroy(&dd->dd_activity_lock);
	mutex_destroy(&dd->dd_lock);
	kmem_free(dd, sizeof (dsl_dir_t));
	dmu_buf_rele(dbuf, tag);
	return (err);
}

void
dsl_dir_rele(dsl_dir_t *dd, const void *tag)
{
	dprintf_dd(dd, "%s\n", "");
	spa_close(dd->dd_pool->dp_spa, tag);
	dmu_buf_rele(dd->dd_dbuf, tag);
}

/*
 * Remove a reference to the given dsl dir that is being asynchronously
 * released.  Async releases occur from a taskq performing eviction of
 * dsl datasets and dirs.  This process is identical to a normal release
 * with the exception of using the async API for releasing the reference on
 * the spa.
 */
void
dsl_dir_async_rele(dsl_dir_t *dd, const void *tag)
{
	dprintf_dd(dd, "%s\n", "");
	spa_async_close(dd->dd_pool->dp_spa, tag);
	dmu_buf_rele(dd->dd_dbuf, tag);
}

/* buf must be at least ZFS_MAX_DATASET_NAME_LEN bytes */
void
dsl_dir_name(dsl_dir_t *dd, char *buf)
{
	if (dd->dd_parent) {
		dsl_dir_name(dd->dd_parent, buf);
		VERIFY3U(strlcat(buf, "/", ZFS_MAX_DATASET_NAME_LEN), <,
		    ZFS_MAX_DATASET_NAME_LEN);
	} else {
		buf[0] = '\0';
	}
	if (!MUTEX_HELD(&dd->dd_lock)) {
		/*
		 * recursive mutex so that we can use
		 * dprintf_dd() with dd_lock held
		 */
		mutex_enter(&dd->dd_lock);
		VERIFY3U(strlcat(buf, dd->dd_myname, ZFS_MAX_DATASET_NAME_LEN),
		    <, ZFS_MAX_DATASET_NAME_LEN);
		mutex_exit(&dd->dd_lock);
	} else {
		VERIFY3U(strlcat(buf, dd->dd_myname, ZFS_MAX_DATASET_NAME_LEN),
		    <, ZFS_MAX_DATASET_NAME_LEN);
	}
}

/* Calculate name length, avoiding all the strcat calls of dsl_dir_name */
int
dsl_dir_namelen(dsl_dir_t *dd)
{
	int result = 0;

	if (dd->dd_parent) {
		/* parent's name + 1 for the "/" */
		result = dsl_dir_namelen(dd->dd_parent) + 1;
	}

	if (!MUTEX_HELD(&dd->dd_lock)) {
		/* see dsl_dir_name */
		mutex_enter(&dd->dd_lock);
		result += strlen(dd->dd_myname);
		mutex_exit(&dd->dd_lock);
	} else {
		result += strlen(dd->dd_myname);
	}

	return (result);
}

static int
getcomponent(const char *path, char *component, const char **nextp)
{
	char *p;

	if ((path == NULL) || (path[0] == '\0'))
		return (SET_ERROR(ENOENT));
	/* This would be a good place to reserve some namespace... */
	p = strpbrk(path, "/@");
	if (p && (p[1] == '/' || p[1] == '@')) {
		/* two separators in a row */
		return (SET_ERROR(EINVAL));
	}
	if (p == NULL || p == path) {
		/*
		 * if the first thing is an @ or /, it had better be an
		 * @ and it had better not have any more ats or slashes,
		 * and it had better have something after the @.
		 */
		if (p != NULL &&
		    (p[0] != '@' || strpbrk(path+1, "/@") || p[1] == '\0'))
			return (SET_ERROR(EINVAL));
		if (strlen(path) >= ZFS_MAX_DATASET_NAME_LEN)
			return (SET_ERROR(ENAMETOOLONG));
		(void) strlcpy(component, path, ZFS_MAX_DATASET_NAME_LEN);
		p = NULL;
	} else if (p[0] == '/') {
		if (p - path >= ZFS_MAX_DATASET_NAME_LEN)
			return (SET_ERROR(ENAMETOOLONG));
		(void) strncpy(component, path, p - path);
		component[p - path] = '\0';
		p++;
	} else if (p[0] == '@') {
		/*
		 * if the next separator is an @, there better not be
		 * any more slashes.
		 */
		if (strchr(path, '/'))
			return (SET_ERROR(EINVAL));
		if (p - path >= ZFS_MAX_DATASET_NAME_LEN)
			return (SET_ERROR(ENAMETOOLONG));
		(void) strncpy(component, path, p - path);
		component[p - path] = '\0';
	} else {
		panic("invalid p=%p", (void *)p);
	}
	*nextp = p;
	return (0);
}

/*
 * Return the dsl_dir_t, and possibly the last component which couldn't
 * be found in *tail.  The name must be in the specified dsl_pool_t.  This
 * thread must hold the dp_config_rwlock for the pool.  Returns NULL if the
 * path is bogus, or if tail==NULL and we couldn't parse the whole name.
 * (*tail)[0] == '@' means that the last component is a snapshot.
 */
int
dsl_dir_hold(dsl_pool_t *dp, const char *name, const void *tag,
    dsl_dir_t **ddp, const char **tailp)
{
	char *buf;
	const char *spaname, *next, *nextnext = NULL;
	int err;
	dsl_dir_t *dd;
	uint64_t ddobj;

	buf = kmem_alloc(ZFS_MAX_DATASET_NAME_LEN, KM_SLEEP);
	err = getcomponent(name, buf, &next);
	if (err != 0)
		goto error;

	/* Make sure the name is in the specified pool. */
	spaname = spa_name(dp->dp_spa);
	if (strcmp(buf, spaname) != 0) {
		err = SET_ERROR(EXDEV);
		goto error;
	}

	ASSERT(dsl_pool_config_held(dp));

	err = dsl_dir_hold_obj(dp, dp->dp_root_dir_obj, NULL, tag, &dd);
	if (err != 0) {
		goto error;
	}

	while (next != NULL) {
		dsl_dir_t *child_dd;
		err = getcomponent(next, buf, &nextnext);
		if (err != 0)
			break;
		ASSERT(next[0] != '\0');
		if (next[0] == '@')
			break;
		dprintf("looking up %s in obj%lld\n",
		    buf, (longlong_t)dsl_dir_phys(dd)->dd_child_dir_zapobj);

		err = zap_lookup(dp->dp_meta_objset,
		    dsl_dir_phys(dd)->dd_child_dir_zapobj,
		    buf, sizeof (ddobj), 1, &ddobj);
		if (err != 0) {
			if (err == ENOENT)
				err = 0;
			break;
		}

		err = dsl_dir_hold_obj(dp, ddobj, buf, tag, &child_dd);
		if (err != 0)
			break;
		dsl_dir_rele(dd, tag);
		dd = child_dd;
		next = nextnext;
	}

	if (err != 0) {
		dsl_dir_rele(dd, tag);
		goto error;
	}

	/*
	 * It's an error if there's more than one component left, or
	 * tailp==NULL and there's any component left.
	 */
	if (next != NULL &&
	    (tailp == NULL || (nextnext && nextnext[0] != '\0'))) {
		/* bad path name */
		dsl_dir_rele(dd, tag);
		dprintf("next=%p (%s) tail=%p\n", next, next?next:"", tailp);
		err = SET_ERROR(ENOENT);
	}
	if (tailp != NULL)
		*tailp = next;
	if (err == 0)
		*ddp = dd;
error:
	kmem_free(buf, ZFS_MAX_DATASET_NAME_LEN);
	return (err);
}

/*
 * If the counts are already initialized for this filesystem and its
 * descendants then do nothing, otherwise initialize the counts.
 *
 * The counts on this filesystem, and those below, may be uninitialized due to
 * either the use of a pre-existing pool which did not support the
 * filesystem/snapshot limit feature, or one in which the feature had not yet
 * been enabled.
 *
 * Recursively descend the filesystem tree and update the filesystem/snapshot
 * counts on each filesystem below, then update the cumulative count on the
 * current filesystem. If the filesystem already has a count set on it,
 * then we know that its counts, and the counts on the filesystems below it,
 * are already correct, so we don't have to update this filesystem.
 */
static void
dsl_dir_init_fs_ss_count(dsl_dir_t *dd, dmu_tx_t *tx)
{
	uint64_t my_fs_cnt = 0;
	uint64_t my_ss_cnt = 0;
	dsl_pool_t *dp = dd->dd_pool;
	objset_t *os = dp->dp_meta_objset;
	zap_cursor_t *zc;
	zap_attribute_t *za;
	dsl_dataset_t *ds;

	ASSERT(spa_feature_is_active(dp->dp_spa, SPA_FEATURE_FS_SS_LIMIT));
	ASSERT(dsl_pool_config_held(dp));
	ASSERT(dmu_tx_is_syncing(tx));

	dsl_dir_zapify(dd, tx);

	/*
	 * If the filesystem count has already been initialized then we
	 * don't need to recurse down any further.
	 */
	if (zap_contains(os, dd->dd_object, DD_FIELD_FILESYSTEM_COUNT) == 0)
		return;

	zc = kmem_alloc(sizeof (zap_cursor_t), KM_SLEEP);
	za = kmem_alloc(sizeof (zap_attribute_t), KM_SLEEP);

	/* Iterate my child dirs */
	for (zap_cursor_init(zc, os, dsl_dir_phys(dd)->dd_child_dir_zapobj);
	    zap_cursor_retrieve(zc, za) == 0; zap_cursor_advance(zc)) {
		dsl_dir_t *chld_dd;
		uint64_t count;

		VERIFY0(dsl_dir_hold_obj(dp, za->za_first_integer, NULL, FTAG,
		    &chld_dd));

		/*
		 * Ignore hidden ($FREE, $MOS & $ORIGIN) objsets.
		 */
		if (chld_dd->dd_myname[0] == '$') {
			dsl_dir_rele(chld_dd, FTAG);
			continue;
		}

		my_fs_cnt++;	/* count this child */

		dsl_dir_init_fs_ss_count(chld_dd, tx);

		VERIFY0(zap_lookup(os, chld_dd->dd_object,
		    DD_FIELD_FILESYSTEM_COUNT, sizeof (count), 1, &count));
		my_fs_cnt += count;
		VERIFY0(zap_lookup(os, chld_dd->dd_object,
		    DD_FIELD_SNAPSHOT_COUNT, sizeof (count), 1, &count));
		my_ss_cnt += count;

		dsl_dir_rele(chld_dd, FTAG);
	}
	zap_cursor_fini(zc);
	/* Count my snapshots (we counted children's snapshots above) */
	VERIFY0(dsl_dataset_hold_obj(dd->dd_pool,
	    dsl_dir_phys(dd)->dd_head_dataset_obj, FTAG, &ds));

	for (zap_cursor_init(zc, os, dsl_dataset_phys(ds)->ds_snapnames_zapobj);
	    zap_cursor_retrieve(zc, za) == 0;
	    zap_cursor_advance(zc)) {
		/* Don't count temporary snapshots */
		if (za->za_name[0] != '%')
			my_ss_cnt++;
	}
	zap_cursor_fini(zc);

	dsl_dataset_rele(ds, FTAG);

	kmem_free(zc, sizeof (zap_cursor_t));
	kmem_free(za, sizeof (zap_attribute_t));

	/* we're in a sync task, update counts */
	dmu_buf_will_dirty(dd->dd_dbuf, tx);
	VERIFY0(zap_add(os, dd->dd_object, DD_FIELD_FILESYSTEM_COUNT,
	    sizeof (my_fs_cnt), 1, &my_fs_cnt, tx));
	VERIFY0(zap_add(os, dd->dd_object, DD_FIELD_SNAPSHOT_COUNT,
	    sizeof (my_ss_cnt), 1, &my_ss_cnt, tx));
}

static int
dsl_dir_actv_fs_ss_limit_check(void *arg, dmu_tx_t *tx)
{
	char *ddname = (char *)arg;
	dsl_pool_t *dp = dmu_tx_pool(tx);
	dsl_dataset_t *ds;
	dsl_dir_t *dd;
	int error;

	error = dsl_dataset_hold(dp, ddname, FTAG, &ds);
	if (error != 0)
		return (error);

	if (!spa_feature_is_enabled(dp->dp_spa, SPA_FEATURE_FS_SS_LIMIT)) {
		dsl_dataset_rele(ds, FTAG);
		return (SET_ERROR(ENOTSUP));
	}

	dd = ds->ds_dir;
	if (spa_feature_is_active(dp->dp_spa, SPA_FEATURE_FS_SS_LIMIT) &&
	    dsl_dir_is_zapified(dd) &&
	    zap_contains(dp->dp_meta_objset, dd->dd_object,
	    DD_FIELD_FILESYSTEM_COUNT) == 0) {
		dsl_dataset_rele(ds, FTAG);
		return (SET_ERROR(EALREADY));
	}

	dsl_dataset_rele(ds, FTAG);
	return (0);
}

static void
dsl_dir_actv_fs_ss_limit_sync(void *arg, dmu_tx_t *tx)
{
	char *ddname = (char *)arg;
	dsl_pool_t *dp = dmu_tx_pool(tx);
	dsl_dataset_t *ds;
	spa_t *spa;

	VERIFY0(dsl_dataset_hold(dp, ddname, FTAG, &ds));

	spa = dsl_dataset_get_spa(ds);

	if (!spa_feature_is_active(spa, SPA_FEATURE_FS_SS_LIMIT)) {
		/*
		 * Since the feature was not active and we're now setting a
		 * limit, increment the feature-active counter so that the
		 * feature becomes active for the first time.
		 *
		 * We are already in a sync task so we can update the MOS.
		 */
		spa_feature_incr(spa, SPA_FEATURE_FS_SS_LIMIT, tx);
	}

	/*
	 * Since we are now setting a non-UINT64_MAX limit on the filesystem,
	 * we need to ensure the counts are correct. Descend down the tree from
	 * this point and update all of the counts to be accurate.
	 */
	dsl_dir_init_fs_ss_count(ds->ds_dir, tx);

	dsl_dataset_rele(ds, FTAG);
}

/*
 * Make sure the feature is enabled and activate it if necessary.
 * Since we're setting a limit, ensure the on-disk counts are valid.
 * This is only called by the ioctl path when setting a limit value.
 *
 * We do not need to validate the new limit, since users who can change the
 * limit are also allowed to exceed the limit.
 */
int
dsl_dir_activate_fs_ss_limit(const char *ddname)
{
	int error;

	error = dsl_sync_task(ddname, dsl_dir_actv_fs_ss_limit_check,
	    dsl_dir_actv_fs_ss_limit_sync, (void *)ddname, 0,
	    ZFS_SPACE_CHECK_RESERVED);

	if (error == EALREADY)
		error = 0;

	return (error);
}

/*
 * Used to determine if the filesystem_limit or snapshot_limit should be
 * enforced. We allow the limit to be exceeded if the user has permission to
 * write the property value. We pass in the creds that we got in the open
 * context since we will always be the GZ root in syncing context. We also have
 * to handle the case where we are allowed to change the limit on the current
 * dataset, but there may be another limit in the tree above.
 *
 * We can never modify these two properties within a non-global zone. In
 * addition, the other checks are modeled on zfs_secpolicy_write_perms. We
 * can't use that function since we are already holding the dp_config_rwlock.
 * In addition, we already have the dd and dealing with snapshots is simplified
 * in this code.
 */

typedef enum {
	ENFORCE_ALWAYS,
	ENFORCE_NEVER,
	ENFORCE_ABOVE
} enforce_res_t;

static enforce_res_t
dsl_enforce_ds_ss_limits(dsl_dir_t *dd, zfs_prop_t prop,
    cred_t *cr, proc_t *proc)
{
	enforce_res_t enforce = ENFORCE_ALWAYS;
	uint64_t obj;
	dsl_dataset_t *ds;
	uint64_t zoned;
	const char *zonedstr;

	ASSERT(prop == ZFS_PROP_FILESYSTEM_LIMIT ||
	    prop == ZFS_PROP_SNAPSHOT_LIMIT);

#ifdef _KERNEL
	if (crgetzoneid(cr) != GLOBAL_ZONEID)
		return (ENFORCE_ALWAYS);

	/*
	 * We are checking the saved credentials of the user process, which is
	 * not the current process.  Note that we can't use secpolicy_zfs(),
	 * because it only works if the cred is that of the current process (on
	 * Linux).
	 */
	if (secpolicy_zfs_proc(cr, proc) == 0)
		return (ENFORCE_NEVER);
#else
	(void) proc;
#endif

	if ((obj = dsl_dir_phys(dd)->dd_head_dataset_obj) == 0)
		return (ENFORCE_ALWAYS);

	ASSERT(dsl_pool_config_held(dd->dd_pool));

	if (dsl_dataset_hold_obj(dd->dd_pool, obj, FTAG, &ds) != 0)
		return (ENFORCE_ALWAYS);

	zonedstr = zfs_prop_to_name(ZFS_PROP_ZONED);
	if (dsl_prop_get_ds(ds, zonedstr, 8, 1, &zoned, NULL) || zoned) {
		/* Only root can access zoned fs's from the GZ */
		enforce = ENFORCE_ALWAYS;
	} else {
		if (dsl_deleg_access_impl(ds, zfs_prop_to_name(prop), cr) == 0)
			enforce = ENFORCE_ABOVE;
	}

	dsl_dataset_rele(ds, FTAG);
	return (enforce);
}

/*
 * Check if adding additional child filesystem(s) would exceed any filesystem
 * limits or adding additional snapshot(s) would exceed any snapshot limits.
 * The prop argument indicates which limit to check.
 *
 * Note that all filesystem limits up to the root (or the highest
 * initialized) filesystem or the given ancestor must be satisfied.
 */
int
dsl_fs_ss_limit_check(dsl_dir_t *dd, uint64_t delta, zfs_prop_t prop,
    dsl_dir_t *ancestor, cred_t *cr, proc_t *proc)
{
	objset_t *os = dd->dd_pool->dp_meta_objset;
	uint64_t limit, count;
	const char *count_prop;
	enforce_res_t enforce;
	int err = 0;

	ASSERT(dsl_pool_config_held(dd->dd_pool));
	ASSERT(prop == ZFS_PROP_FILESYSTEM_LIMIT ||
	    prop == ZFS_PROP_SNAPSHOT_LIMIT);

	/*
	 * If we're allowed to change the limit, don't enforce the limit
	 * e.g. this can happen if a snapshot is taken by an administrative
	 * user in the global zone (i.e. a recursive snapshot by root).
	 * However, we must handle the case of delegated permissions where we
	 * are allowed to change the limit on the current dataset, but there
	 * is another limit in the tree above.
	 */
	enforce = dsl_enforce_ds_ss_limits(dd, prop, cr, proc);
	if (enforce == ENFORCE_NEVER)
		return (0);

	/*
	 * e.g. if renaming a dataset with no snapshots, count adjustment
	 * is 0.
	 */
	if (delta == 0)
		return (0);

	if (prop == ZFS_PROP_SNAPSHOT_LIMIT) {
		/*
		 * We don't enforce the limit for temporary snapshots. This is
		 * indicated by a NULL cred_t argument.
		 */
		if (cr == NULL)
			return (0);

		count_prop = DD_FIELD_SNAPSHOT_COUNT;
	} else {
		count_prop = DD_FIELD_FILESYSTEM_COUNT;
	}

	/*
	 * If an ancestor has been provided, stop checking the limit once we
	 * hit that dir. We need this during rename so that we don't overcount
	 * the check once we recurse up to the common ancestor.
	 */
	if (ancestor == dd)
		return (0);

	/*
	 * If we hit an uninitialized node while recursing up the tree, we can
	 * stop since we know there is no limit here (or above). The counts are
	 * not valid on this node and we know we won't touch this node's counts.
	 */
	if (!dsl_dir_is_zapified(dd))
		return (0);
	err = zap_lookup(os, dd->dd_object,
	    count_prop, sizeof (count), 1, &count);
	if (err == ENOENT)
		return (0);
	if (err != 0)
		return (err);

	err = dsl_prop_get_dd(dd, zfs_prop_to_name(prop), 8, 1, &limit, NULL,
	    B_FALSE);
	if (err != 0)
		return (err);

	/* Is there a limit which we've hit? */
	if (enforce == ENFORCE_ALWAYS && (count + delta) > limit)
		return (SET_ERROR(EDQUOT));

	if (dd->dd_parent != NULL)
		err = dsl_fs_ss_limit_check(dd->dd_parent, delta, prop,
		    ancestor, cr, proc);

	return (err);
}

/*
 * Adjust the filesystem or snapshot count for the specified dsl_dir_t and all
 * parents. When a new filesystem/snapshot is created, increment the count on
 * all parents, and when a filesystem/snapshot is destroyed, decrement the
 * count.
 */
void
dsl_fs_ss_count_adjust(dsl_dir_t *dd, int64_t delta, const char *prop,
    dmu_tx_t *tx)
{
	int err;
	objset_t *os = dd->dd_pool->dp_meta_objset;
	uint64_t count;

	ASSERT(dsl_pool_config_held(dd->dd_pool));
	ASSERT(dmu_tx_is_syncing(tx));
	ASSERT(strcmp(prop, DD_FIELD_FILESYSTEM_COUNT) == 0 ||
	    strcmp(prop, DD_FIELD_SNAPSHOT_COUNT) == 0);

	/*
	 * We don't do accounting for hidden ($FREE, $MOS & $ORIGIN) objsets.
	 */
	if (dd->dd_myname[0] == '$' && strcmp(prop,
	    DD_FIELD_FILESYSTEM_COUNT) == 0) {
		return;
	}

	/*
	 * e.g. if renaming a dataset with no snapshots, count adjustment is 0
	 */
	if (delta == 0)
		return;

	/*
	 * If we hit an uninitialized node while recursing up the tree, we can
	 * stop since we know the counts are not valid on this node and we
	 * know we shouldn't touch this node's counts. An uninitialized count
	 * on the node indicates that either the feature has not yet been
	 * activated or there are no limits on this part of the tree.
	 */
	if (!dsl_dir_is_zapified(dd) || (err = zap_lookup(os, dd->dd_object,
	    prop, sizeof (count), 1, &count)) == ENOENT)
		return;
	VERIFY0(err);

	count += delta;
	/* Use a signed verify to make sure we're not neg. */
	VERIFY3S(count, >=, 0);

	VERIFY0(zap_update(os, dd->dd_object, prop, sizeof (count), 1, &count,
	    tx));

	/* Roll up this additional count into our ancestors */
	if (dd->dd_parent != NULL)
		dsl_fs_ss_count_adjust(dd->dd_parent, delta, prop, tx);
}

uint64_t
dsl_dir_create_sync(dsl_pool_t *dp, dsl_dir_t *pds, const char *name,
    dmu_tx_t *tx)
{
	objset_t *mos = dp->dp_meta_objset;
	uint64_t ddobj;
	dsl_dir_phys_t *ddphys;
	dmu_buf_t *dbuf;

	ddobj = dmu_object_alloc(mos, DMU_OT_DSL_DIR, 0,
	    DMU_OT_DSL_DIR, sizeof (dsl_dir_phys_t), tx);
	if (pds) {
		VERIFY0(zap_add(mos, dsl_dir_phys(pds)->dd_child_dir_zapobj,
		    name, sizeof (uint64_t), 1, &ddobj, tx));
	} else {
		/* it's the root dir */
		VERIFY0(zap_add(mos, DMU_POOL_DIRECTORY_OBJECT,
		    DMU_POOL_ROOT_DATASET, sizeof (uint64_t), 1, &ddobj, tx));
	}
	VERIFY0(dmu_bonus_hold(mos, ddobj, FTAG, &dbuf));
	dmu_buf_will_dirty(dbuf, tx);
	ddphys = dbuf->db_data;

	ddphys->dd_creation_time = gethrestime_sec();
	if (pds) {
		ddphys->dd_parent_obj = pds->dd_object;

		/* update the filesystem counts */
		dsl_fs_ss_count_adjust(pds, 1, DD_FIELD_FILESYSTEM_COUNT, tx);
	}
	ddphys->dd_props_zapobj = zap_create(mos,
	    DMU_OT_DSL_PROPS, DMU_OT_NONE, 0, tx);
	ddphys->dd_child_dir_zapobj = zap_create(mos,
	    DMU_OT_DSL_DIR_CHILD_MAP, DMU_OT_NONE, 0, tx);
	if (spa_version(dp->dp_spa) >= SPA_VERSION_USED_BREAKDOWN)
		ddphys->dd_flags |= DD_FLAG_USED_BREAKDOWN;

	dmu_buf_rele(dbuf, FTAG);

	return (ddobj);
}

boolean_t
dsl_dir_is_clone(dsl_dir_t *dd)
{
	return (dsl_dir_phys(dd)->dd_origin_obj &&
	    (dd->dd_pool->dp_origin_snap == NULL ||
	    dsl_dir_phys(dd)->dd_origin_obj !=
	    dd->dd_pool->dp_origin_snap->ds_object));
}

uint64_t
dsl_dir_get_used(dsl_dir_t *dd)
{
	return (dsl_dir_phys(dd)->dd_used_bytes);
}

uint64_t
dsl_dir_get_compressed(dsl_dir_t *dd)
{
	return (dsl_dir_phys(dd)->dd_compressed_bytes);
}

uint64_t
dsl_dir_get_quota(dsl_dir_t *dd)
{
	return (dsl_dir_phys(dd)->dd_quota);
}

uint64_t
dsl_dir_get_reservation(dsl_dir_t *dd)
{
	return (dsl_dir_phys(dd)->dd_reserved);
}

uint64_t
dsl_dir_get_compressratio(dsl_dir_t *dd)
{
	/* a fixed point number, 100x the ratio */
	return (dsl_dir_phys(dd)->dd_compressed_bytes == 0 ? 100 :
	    (dsl_dir_phys(dd)->dd_uncompressed_bytes * 100 /
	    dsl_dir_phys(dd)->dd_compressed_bytes));
}

uint64_t
dsl_dir_get_logicalused(dsl_dir_t *dd)
{
	return (dsl_dir_phys(dd)->dd_uncompressed_bytes);
}

uint64_t
dsl_dir_get_usedsnap(dsl_dir_t *dd)
{
	return (dsl_dir_phys(dd)->dd_used_breakdown[DD_USED_SNAP]);
}

uint64_t
dsl_dir_get_usedds(dsl_dir_t *dd)
{
	return (dsl_dir_phys(dd)->dd_used_breakdown[DD_USED_HEAD]);
}

uint64_t
dsl_dir_get_usedrefreserv(dsl_dir_t *dd)
{
	return (dsl_dir_phys(dd)->dd_used_breakdown[DD_USED_REFRSRV]);
}

uint64_t
dsl_dir_get_usedchild(dsl_dir_t *dd)
{
	return (dsl_dir_phys(dd)->dd_used_breakdown[DD_USED_CHILD] +
	    dsl_dir_phys(dd)->dd_used_breakdown[DD_USED_CHILD_RSRV]);
}

void
dsl_dir_get_origin(dsl_dir_t *dd, char *buf)
{
	dsl_dataset_t *ds;
	VERIFY0(dsl_dataset_hold_obj(dd->dd_pool,
	    dsl_dir_phys(dd)->dd_origin_obj, FTAG, &ds));

	dsl_dataset_name(ds, buf);

	dsl_dataset_rele(ds, FTAG);
}

int
dsl_dir_get_filesystem_count(dsl_dir_t *dd, uint64_t *count)
{
	if (dsl_dir_is_zapified(dd)) {
		objset_t *os = dd->dd_pool->dp_meta_objset;
		return (zap_lookup(os, dd->dd_object, DD_FIELD_FILESYSTEM_COUNT,
		    sizeof (*count), 1, count));
	} else {
		return (SET_ERROR(ENOENT));
	}
}

int
dsl_dir_get_snapshot_count(dsl_dir_t *dd, uint64_t *count)
{
	if (dsl_dir_is_zapified(dd)) {
		objset_t *os = dd->dd_pool->dp_meta_objset;
		return (zap_lookup(os, dd->dd_object, DD_FIELD_SNAPSHOT_COUNT,
		    sizeof (*count), 1, count));
	} else {
		return (SET_ERROR(ENOENT));
	}
}

void
dsl_dir_stats(dsl_dir_t *dd, nvlist_t *nv)
{
	mutex_enter(&dd->dd_lock);
	dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_QUOTA,
	    dsl_dir_get_quota(dd));
	dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_RESERVATION,
	    dsl_dir_get_reservation(dd));
	dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_LOGICALUSED,
	    dsl_dir_get_logicalused(dd));
	if (dsl_dir_phys(dd)->dd_flags & DD_FLAG_USED_BREAKDOWN) {
		dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_USEDSNAP,
		    dsl_dir_get_usedsnap(dd));
		dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_USEDDS,
		    dsl_dir_get_usedds(dd));
		dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_USEDREFRESERV,
		    dsl_dir_get_usedrefreserv(dd));
		dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_USEDCHILD,
		    dsl_dir_get_usedchild(dd));
	}
	mutex_exit(&dd->dd_lock);

	uint64_t count;
	if (dsl_dir_get_filesystem_count(dd, &count) == 0) {
		dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_FILESYSTEM_COUNT,
		    count);
	}
	if (dsl_dir_get_snapshot_count(dd, &count) == 0) {
		dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_SNAPSHOT_COUNT,
		    count);
	}

	if (dsl_dir_is_clone(dd)) {
		char buf[ZFS_MAX_DATASET_NAME_LEN];
		dsl_dir_get_origin(dd, buf);
		dsl_prop_nvlist_add_string(nv, ZFS_PROP_ORIGIN, buf);
	}

}

void
dsl_dir_dirty(dsl_dir_t *dd, dmu_tx_t *tx)
{
	dsl_pool_t *dp = dd->dd_pool;

	ASSERT(dsl_dir_phys(dd));

	if (txg_list_add(&dp->dp_dirty_dirs, dd, tx->tx_txg)) {
		/* up the hold count until we can be written out */
		dmu_buf_add_ref(dd->dd_dbuf, dd);
	}
}

static int64_t
parent_delta(dsl_dir_t *dd, uint64_t used, int64_t delta)
{
	uint64_t old_accounted = MAX(used, dsl_dir_phys(dd)->dd_reserved);
	uint64_t new_accounted =
	    MAX(used + delta, dsl_dir_phys(dd)->dd_reserved);
	return (new_accounted - old_accounted);
}

void
dsl_dir_sync(dsl_dir_t *dd, dmu_tx_t *tx)
{
	ASSERT(dmu_tx_is_syncing(tx));

	mutex_enter(&dd->dd_lock);
	ASSERT0(dd->dd_tempreserved[tx->tx_txg & TXG_MASK]);
	dprintf_dd(dd, "txg=%llu towrite=%lluK\n", (u_longlong_t)tx->tx_txg,
	    (u_longlong_t)dd->dd_space_towrite[tx->tx_txg & TXG_MASK] / 1024);
	dd->dd_space_towrite[tx->tx_txg & TXG_MASK] = 0;
	mutex_exit(&dd->dd_lock);

	/* release the hold from dsl_dir_dirty */
	dmu_buf_rele(dd->dd_dbuf, dd);
}

static uint64_t
dsl_dir_space_towrite(dsl_dir_t *dd)
{
	uint64_t space = 0;

	ASSERT(MUTEX_HELD(&dd->dd_lock));

	for (int i = 0; i < TXG_SIZE; i++) {
		space += dd->dd_space_towrite[i & TXG_MASK];
		ASSERT3U(dd->dd_space_towrite[i & TXG_MASK], >=, 0);
	}
	return (space);
}

/*
 * How much space would dd have available if ancestor had delta applied
 * to it?  If ondiskonly is set, we're only interested in what's
 * on-disk, not estimated pending changes.
 */
uint64_t
dsl_dir_space_available(dsl_dir_t *dd,
    dsl_dir_t *ancestor, int64_t delta, int ondiskonly)
{
	uint64_t parentspace, myspace, quota, used;

	/*
	 * If there are no restrictions otherwise, assume we have
	 * unlimited space available.
	 */
	quota = UINT64_MAX;
	parentspace = UINT64_MAX;

	if (dd->dd_parent != NULL) {
		parentspace = dsl_dir_space_available(dd->dd_parent,
		    ancestor, delta, ondiskonly);
	}

	mutex_enter(&dd->dd_lock);
	if (dsl_dir_phys(dd)->dd_quota != 0)
		quota = dsl_dir_phys(dd)->dd_quota;
	used = dsl_dir_phys(dd)->dd_used_bytes;
	if (!ondiskonly)
		used += dsl_dir_space_towrite(dd);

	if (dd->dd_parent == NULL) {
		uint64_t poolsize = dsl_pool_adjustedsize(dd->dd_pool,
		    ZFS_SPACE_CHECK_NORMAL);
		quota = MIN(quota, poolsize);
	}

	if (dsl_dir_phys(dd)->dd_reserved > used && parentspace != UINT64_MAX) {
		/*
		 * We have some space reserved, in addition to what our
		 * parent gave us.
		 */
		parentspace += dsl_dir_phys(dd)->dd_reserved - used;
	}

	if (dd == ancestor) {
		ASSERT(delta <= 0);
		ASSERT(used >= -delta);
		used += delta;
		if (parentspace != UINT64_MAX)
			parentspace -= delta;
	}

	if (used > quota) {
		/* over quota */
		myspace = 0;
	} else {
		/*
		 * the lesser of the space provided by our parent and
		 * the space left in our quota
		 */
		myspace = MIN(parentspace, quota - used);
	}

	mutex_exit(&dd->dd_lock);

	return (myspace);
}

struct tempreserve {
	list_node_t tr_node;
	dsl_dir_t *tr_ds;
	uint64_t tr_size;
};

static int
dsl_dir_tempreserve_impl(dsl_dir_t *dd, uint64_t asize, boolean_t netfree,
    boolean_t ignorequota, list_t *tr_list,
    dmu_tx_t *tx, boolean_t first)
{
	uint64_t txg;
	uint64_t quota;
	struct tempreserve *tr;
	int retval;
	uint64_t ref_rsrv;

top_of_function:
	txg = tx->tx_txg;
	retval = EDQUOT;
	ref_rsrv = 0;

	ASSERT3U(txg, !=, 0);
	ASSERT3S(asize, >, 0);

	mutex_enter(&dd->dd_lock);

	/*
	 * Check against the dsl_dir's quota.  We don't add in the delta
	 * when checking for over-quota because they get one free hit.
	 */
	uint64_t est_inflight = dsl_dir_space_towrite(dd);
	for (int i = 0; i < TXG_SIZE; i++)
		est_inflight += dd->dd_tempreserved[i];
	uint64_t used_on_disk = dsl_dir_phys(dd)->dd_used_bytes;

	/*
	 * On the first iteration, fetch the dataset's used-on-disk and
	 * refreservation values. Also, if checkrefquota is set, test if
	 * allocating this space would exceed the dataset's refquota.
	 */
	if (first && tx->tx_objset) {
		int error;
		dsl_dataset_t *ds = tx->tx_objset->os_dsl_dataset;

		error = dsl_dataset_check_quota(ds, !netfree,
		    asize, est_inflight, &used_on_disk, &ref_rsrv);
		if (error != 0) {
			mutex_exit(&dd->dd_lock);
			DMU_TX_STAT_BUMP(dmu_tx_quota);
			return (error);
		}
	}

	/*
	 * If this transaction will result in a net free of space,
	 * we want to let it through.
	 */
	if (ignorequota || netfree || dsl_dir_phys(dd)->dd_quota == 0)
		quota = UINT64_MAX;
	else
		quota = dsl_dir_phys(dd)->dd_quota;

	/*
	 * Adjust the quota against the actual pool size at the root
	 * minus any outstanding deferred frees.
	 * To ensure that it's possible to remove files from a full
	 * pool without inducing transient overcommits, we throttle
	 * netfree transactions against a quota that is slightly larger,
	 * but still within the pool's allocation slop.  In cases where
	 * we're very close to full, this will allow a steady trickle of
	 * removes to get through.
	 */
	if (dd->dd_parent == NULL) {
		uint64_t avail = dsl_pool_unreserved_space(dd->dd_pool,
		    (netfree) ?
		    ZFS_SPACE_CHECK_RESERVED : ZFS_SPACE_CHECK_NORMAL);

		if (avail < quota) {
			quota = avail;
			retval = SET_ERROR(ENOSPC);
		}
	}

	/*
	 * If they are requesting more space, and our current estimate
	 * is over quota, they get to try again unless the actual
	 * on-disk is over quota and there are no pending changes
	 * or deferred frees (which may free up space for us).
	 */
	if (used_on_disk + est_inflight >= quota) {
		if (est_inflight > 0 || used_on_disk < quota) {
			retval = SET_ERROR(ERESTART);
		} else {
			ASSERT3U(used_on_disk, >=, quota);

			if (retval == ENOSPC && (used_on_disk - quota) <
			    dsl_pool_deferred_space(dd->dd_pool)) {
				retval = SET_ERROR(ERESTART);
			}
		}

		dprintf_dd(dd, "failing: used=%lluK inflight = %lluK "
		    "quota=%lluK tr=%lluK err=%d\n",
		    (u_longlong_t)used_on_disk>>10,
		    (u_longlong_t)est_inflight>>10,
		    (u_longlong_t)quota>>10, (u_longlong_t)asize>>10, retval);
		mutex_exit(&dd->dd_lock);
		DMU_TX_STAT_BUMP(dmu_tx_quota);
		return (retval);
	}

	/* We need to up our estimated delta before dropping dd_lock */
	dd->dd_tempreserved[txg & TXG_MASK] += asize;

	uint64_t parent_rsrv = parent_delta(dd, used_on_disk + est_inflight,
	    asize - ref_rsrv);
	mutex_exit(&dd->dd_lock);

	tr = kmem_zalloc(sizeof (struct tempreserve), KM_SLEEP);
	tr->tr_ds = dd;
	tr->tr_size = asize;
	list_insert_tail(tr_list, tr);

	/* see if it's OK with our parent */
	if (dd->dd_parent != NULL && parent_rsrv != 0) {
		/*
		 * Recurse on our parent without recursion. This has been
		 * observed to be potentially large stack usage even within
		 * the test suite. Largest seen stack was 7632 bytes on linux.
		 */

		dd = dd->dd_parent;
		asize = parent_rsrv;
		ignorequota = (dsl_dir_phys(dd)->dd_head_dataset_obj == 0);
		first = B_FALSE;
		goto top_of_function;

	} else {
		return (0);
	}
}

/*
 * Reserve space in this dsl_dir, to be used in this tx's txg.
 * After the space has been dirtied (and dsl_dir_willuse_space()
 * has been called), the reservation should be canceled, using
 * dsl_dir_tempreserve_clear().
 */
int
dsl_dir_tempreserve_space(dsl_dir_t *dd, uint64_t lsize, uint64_t asize,
    boolean_t netfree, void **tr_cookiep, dmu_tx_t *tx)
{
	int err;
	list_t *tr_list;

	if (asize == 0) {
		*tr_cookiep = NULL;
		return (0);
	}

	tr_list = kmem_alloc(sizeof (list_t), KM_SLEEP);
	list_create(tr_list, sizeof (struct tempreserve),
	    offsetof(struct tempreserve, tr_node));
	ASSERT3S(asize, >, 0);

	err = arc_tempreserve_space(dd->dd_pool->dp_spa, lsize, tx->tx_txg);
	if (err == 0) {
		struct tempreserve *tr;

		tr = kmem_zalloc(sizeof (struct tempreserve), KM_SLEEP);
		tr->tr_size = lsize;
		list_insert_tail(tr_list, tr);
	} else {
		if (err == EAGAIN) {
			/*
			 * If arc_memory_throttle() detected that pageout
			 * is running and we are low on memory, we delay new
			 * non-pageout transactions to give pageout an
			 * advantage.
			 *
			 * It is unfortunate to be delaying while the caller's
			 * locks are held.
			 */
			txg_delay(dd->dd_pool, tx->tx_txg,
			    MSEC2NSEC(10), MSEC2NSEC(10));
			err = SET_ERROR(ERESTART);
		}
	}

	if (err == 0) {
		err = dsl_dir_tempreserve_impl(dd, asize, netfree,
		    B_FALSE, tr_list, tx, B_TRUE);
	}

	if (err != 0)
		dsl_dir_tempreserve_clear(tr_list, tx);
	else
		*tr_cookiep = tr_list;

	return (err);
}

/*
 * Clear a temporary reservation that we previously made with
 * dsl_dir_tempreserve_space().
 */
void
dsl_dir_tempreserve_clear(void *tr_cookie, dmu_tx_t *tx)
{
	int txgidx = tx->tx_txg & TXG_MASK;
	list_t *tr_list = tr_cookie;
	struct tempreserve *tr;

	ASSERT3U(tx->tx_txg, !=, 0);

	if (tr_cookie == NULL)
		return;

	while ((tr = list_head(tr_list)) != NULL) {
		if (tr->tr_ds) {
			mutex_enter(&tr->tr_ds->dd_lock);
			ASSERT3U(tr->tr_ds->dd_tempreserved[txgidx], >=,
			    tr->tr_size);
			tr->tr_ds->dd_tempreserved[txgidx] -= tr->tr_size;
			mutex_exit(&tr->tr_ds->dd_lock);
		} else {
			arc_tempreserve_clear(tr->tr_size);
		}
		list_remove(tr_list, tr);
		kmem_free(tr, sizeof (struct tempreserve));
	}

	kmem_free(tr_list, sizeof (list_t));
}

/*
 * This should be called from open context when we think we're going to write
 * or free space, for example when dirtying data. Be conservative; it's okay
 * to write less space or free more, but we don't want to write more or free
 * less than the amount specified.
 *
 * NOTE: The behavior of this function is identical to the Illumos / FreeBSD
 * version however it has been adjusted to use an iterative rather than
 * recursive algorithm to minimize stack usage.
 */
void
dsl_dir_willuse_space(dsl_dir_t *dd, int64_t space, dmu_tx_t *tx)
{
	int64_t parent_space;
	uint64_t est_used;

	do {
		mutex_enter(&dd->dd_lock);
		if (space > 0)
			dd->dd_space_towrite[tx->tx_txg & TXG_MASK] += space;

		est_used = dsl_dir_space_towrite(dd) +
		    dsl_dir_phys(dd)->dd_used_bytes;
		parent_space = parent_delta(dd, est_used, space);
		mutex_exit(&dd->dd_lock);

		/* Make sure that we clean up dd_space_to* */
		dsl_dir_dirty(dd, tx);

		dd = dd->dd_parent;
		space = parent_space;
	} while (space && dd);
}

/* call from syncing context when we actually write/free space for this dd */
void
dsl_dir_diduse_space(dsl_dir_t *dd, dd_used_t type,
    int64_t used, int64_t compressed, int64_t uncompressed, dmu_tx_t *tx)
{
	int64_t accounted_delta;

	ASSERT(dmu_tx_is_syncing(tx));
	ASSERT(type < DD_USED_NUM);

	dmu_buf_will_dirty(dd->dd_dbuf, tx);

	/*
	 * dsl_dataset_set_refreservation_sync_impl() calls this with
	 * dd_lock held, so that it can atomically update
	 * ds->ds_reserved and the dsl_dir accounting, so that
	 * dsl_dataset_check_quota() can see dataset and dir accounting
	 * consistently.
	 */
	boolean_t needlock = !MUTEX_HELD(&dd->dd_lock);
	if (needlock)
		mutex_enter(&dd->dd_lock);
	dsl_dir_phys_t *ddp = dsl_dir_phys(dd);
	accounted_delta = parent_delta(dd, ddp->dd_used_bytes, used);
	ASSERT(used >= 0 || ddp->dd_used_bytes >= -used);
	ASSERT(compressed >= 0 || ddp->dd_compressed_bytes >= -compressed);
	ASSERT(uncompressed >= 0 ||
	    ddp->dd_uncompressed_bytes >= -uncompressed);
	ddp->dd_used_bytes += used;
	ddp->dd_uncompressed_bytes += uncompressed;
	ddp->dd_compressed_bytes += compressed;

	if (ddp->dd_flags & DD_FLAG_USED_BREAKDOWN) {
		ASSERT(used >= 0 || ddp->dd_used_breakdown[type] >= -used);
		ddp->dd_used_breakdown[type] += used;
#ifdef ZFS_DEBUG
		{
			dd_used_t t;
			uint64_t u = 0;
			for (t = 0; t < DD_USED_NUM; t++)
				u += ddp->dd_used_breakdown[t];
			ASSERT3U(u, ==, ddp->dd_used_bytes);
		}
#endif
	}
	if (needlock)
		mutex_exit(&dd->dd_lock);

	if (dd->dd_parent != NULL) {
		dsl_dir_diduse_transfer_space(dd->dd_parent,
		    accounted_delta, compressed, uncompressed,
		    used, DD_USED_CHILD_RSRV, DD_USED_CHILD, tx);
	}
}

void
dsl_dir_transfer_space(dsl_dir_t *dd, int64_t delta,
    dd_used_t oldtype, dd_used_t newtype, dmu_tx_t *tx)
{
	ASSERT(dmu_tx_is_syncing(tx));
	ASSERT(oldtype < DD_USED_NUM);
	ASSERT(newtype < DD_USED_NUM);

	dsl_dir_phys_t *ddp = dsl_dir_phys(dd);
	if (delta == 0 ||
	    !(ddp->dd_flags & DD_FLAG_USED_BREAKDOWN))
		return;

	dmu_buf_will_dirty(dd->dd_dbuf, tx);
	mutex_enter(&dd->dd_lock);
	ASSERT(delta > 0 ?
	    ddp->dd_used_breakdown[oldtype] >= delta :
	    ddp->dd_used_breakdown[newtype] >= -delta);
	ASSERT(ddp->dd_used_bytes >= ABS(delta));
	ddp->dd_used_breakdown[oldtype] -= delta;
	ddp->dd_used_breakdown[newtype] += delta;
	mutex_exit(&dd->dd_lock);
}

void
dsl_dir_diduse_transfer_space(dsl_dir_t *dd, int64_t used,
    int64_t compressed, int64_t uncompressed, int64_t tonew,
    dd_used_t oldtype, dd_used_t newtype, dmu_tx_t *tx)
{
	int64_t accounted_delta;

	ASSERT(dmu_tx_is_syncing(tx));
	ASSERT(oldtype < DD_USED_NUM);
	ASSERT(newtype < DD_USED_NUM);

	dmu_buf_will_dirty(dd->dd_dbuf, tx);

	mutex_enter(&dd->dd_lock);
	dsl_dir_phys_t *ddp = dsl_dir_phys(dd);
	accounted_delta = parent_delta(dd, ddp->dd_used_bytes, used);
	ASSERT(used >= 0 || ddp->dd_used_bytes >= -used);
	ASSERT(compressed >= 0 || ddp->dd_compressed_bytes >= -compressed);
	ASSERT(uncompressed >= 0 ||
	    ddp->dd_uncompressed_bytes >= -uncompressed);
	ddp->dd_used_bytes += used;
	ddp->dd_uncompressed_bytes += uncompressed;
	ddp->dd_compressed_bytes += compressed;

	if (ddp->dd_flags & DD_FLAG_USED_BREAKDOWN) {
		ASSERT(tonew - used <= 0 ||
		    ddp->dd_used_breakdown[oldtype] >= tonew - used);
		ASSERT(tonew >= 0 ||
		    ddp->dd_used_breakdown[newtype] >= -tonew);
		ddp->dd_used_breakdown[oldtype] -= tonew - used;
		ddp->dd_used_breakdown[newtype] += tonew;
#ifdef ZFS_DEBUG
		{
			dd_used_t t;
			uint64_t u = 0;
			for (t = 0; t < DD_USED_NUM; t++)
				u += ddp->dd_used_breakdown[t];
			ASSERT3U(u, ==, ddp->dd_used_bytes);
		}
#endif
	}
	mutex_exit(&dd->dd_lock);

	if (dd->dd_parent != NULL) {
		dsl_dir_diduse_transfer_space(dd->dd_parent,
		    accounted_delta, compressed, uncompressed,
		    used, DD_USED_CHILD_RSRV, DD_USED_CHILD, tx);
	}
}

typedef struct dsl_dir_set_qr_arg {
	const char *ddsqra_name;
	zprop_source_t ddsqra_source;
	uint64_t ddsqra_value;
} dsl_dir_set_qr_arg_t;

static int
dsl_dir_set_quota_check(void *arg, dmu_tx_t *tx)
{
	dsl_dir_set_qr_arg_t *ddsqra = arg;
	dsl_pool_t *dp = dmu_tx_pool(tx);
	dsl_dataset_t *ds;
	int error;
	uint64_t towrite, newval;

	error = dsl_dataset_hold(dp, ddsqra->ddsqra_name, FTAG, &ds);
	if (error != 0)
		return (error);

	error = dsl_prop_predict(ds->ds_dir, "quota",
	    ddsqra->ddsqra_source, ddsqra->ddsqra_value, &newval);
	if (error != 0) {
		dsl_dataset_rele(ds, FTAG);
		return (error);
	}

	if (newval == 0) {
		dsl_dataset_rele(ds, FTAG);
		return (0);
	}

	mutex_enter(&ds->ds_dir->dd_lock);
	/*
	 * If we are doing the preliminary check in open context, and
	 * there are pending changes, then don't fail it, since the
	 * pending changes could under-estimate the amount of space to be
	 * freed up.
	 */
	towrite = dsl_dir_space_towrite(ds->ds_dir);
	if ((dmu_tx_is_syncing(tx) || towrite == 0) &&
	    (newval < dsl_dir_phys(ds->ds_dir)->dd_reserved ||
	    newval < dsl_dir_phys(ds->ds_dir)->dd_used_bytes + towrite)) {
		error = SET_ERROR(ENOSPC);
	}
	mutex_exit(&ds->ds_dir->dd_lock);
	dsl_dataset_rele(ds, FTAG);
	return (error);
}

static void
dsl_dir_set_quota_sync(void *arg, dmu_tx_t *tx)
{
	dsl_dir_set_qr_arg_t *ddsqra = arg;
	dsl_pool_t *dp = dmu_tx_pool(tx);
	dsl_dataset_t *ds;
	uint64_t newval;

	VERIFY0(dsl_dataset_hold(dp, ddsqra->ddsqra_name, FTAG, &ds));

	if (spa_version(dp->dp_spa) >= SPA_VERSION_RECVD_PROPS) {
		dsl_prop_set_sync_impl(ds, zfs_prop_to_name(ZFS_PROP_QUOTA),
		    ddsqra->ddsqra_source, sizeof (ddsqra->ddsqra_value), 1,
		    &ddsqra->ddsqra_value, tx);

		VERIFY0(dsl_prop_get_int_ds(ds,
		    zfs_prop_to_name(ZFS_PROP_QUOTA), &newval));
	} else {
		newval = ddsqra->ddsqra_value;
		spa_history_log_internal_ds(ds, "set", tx, "%s=%lld",
		    zfs_prop_to_name(ZFS_PROP_QUOTA), (longlong_t)newval);
	}

	dmu_buf_will_dirty(ds->ds_dir->dd_dbuf, tx);
	mutex_enter(&ds->ds_dir->dd_lock);
	dsl_dir_phys(ds->ds_dir)->dd_quota = newval;
	mutex_exit(&ds->ds_dir->dd_lock);
	dsl_dataset_rele(ds, FTAG);
}

int
dsl_dir_set_quota(const char *ddname, zprop_source_t source, uint64_t quota)
{
	dsl_dir_set_qr_arg_t ddsqra;

	ddsqra.ddsqra_name = ddname;
	ddsqra.ddsqra_source = source;
	ddsqra.ddsqra_value = quota;

	return (dsl_sync_task(ddname, dsl_dir_set_quota_check,
	    dsl_dir_set_quota_sync, &ddsqra, 0,
	    ZFS_SPACE_CHECK_EXTRA_RESERVED));
}

static int
dsl_dir_set_reservation_check(void *arg, dmu_tx_t *tx)
{
	dsl_dir_set_qr_arg_t *ddsqra = arg;
	dsl_pool_t *dp = dmu_tx_pool(tx);
	dsl_dataset_t *ds;
	dsl_dir_t *dd;
	uint64_t newval, used, avail;
	int error;

	error = dsl_dataset_hold(dp, ddsqra->ddsqra_name, FTAG, &ds);
	if (error != 0)
		return (error);
	dd = ds->ds_dir;

	/*
	 * If we are doing the preliminary check in open context, the
	 * space estimates may be inaccurate.
	 */
	if (!dmu_tx_is_syncing(tx)) {
		dsl_dataset_rele(ds, FTAG);
		return (0);
	}

	error = dsl_prop_predict(ds->ds_dir,
	    zfs_prop_to_name(ZFS_PROP_RESERVATION),
	    ddsqra->ddsqra_source, ddsqra->ddsqra_value, &newval);
	if (error != 0) {
		dsl_dataset_rele(ds, FTAG);
		return (error);
	}

	mutex_enter(&dd->dd_lock);
	used = dsl_dir_phys(dd)->dd_used_bytes;
	mutex_exit(&dd->dd_lock);

	if (dd->dd_parent) {
		avail = dsl_dir_space_available(dd->dd_parent,
		    NULL, 0, FALSE);
	} else {
		avail = dsl_pool_adjustedsize(dd->dd_pool,
		    ZFS_SPACE_CHECK_NORMAL) - used;
	}

	if (MAX(used, newval) > MAX(used, dsl_dir_phys(dd)->dd_reserved)) {
		uint64_t delta = MAX(used, newval) -
		    MAX(used, dsl_dir_phys(dd)->dd_reserved);

		if (delta > avail ||
		    (dsl_dir_phys(dd)->dd_quota > 0 &&
		    newval > dsl_dir_phys(dd)->dd_quota))
			error = SET_ERROR(ENOSPC);
	}

	dsl_dataset_rele(ds, FTAG);
	return (error);
}

void
dsl_dir_set_reservation_sync_impl(dsl_dir_t *dd, uint64_t value, dmu_tx_t *tx)
{
	uint64_t used;
	int64_t delta;

	dmu_buf_will_dirty(dd->dd_dbuf, tx);

	mutex_enter(&dd->dd_lock);
	used = dsl_dir_phys(dd)->dd_used_bytes;
	delta = MAX(used, value) - MAX(used, dsl_dir_phys(dd)->dd_reserved);
	dsl_dir_phys(dd)->dd_reserved = value;

	if (dd->dd_parent != NULL) {
		/* Roll up this additional usage into our ancestors */
		dsl_dir_diduse_space(dd->dd_parent, DD_USED_CHILD_RSRV,
		    delta, 0, 0, tx);
	}
	mutex_exit(&dd->dd_lock);
}

static void
dsl_dir_set_reservation_sync(void *arg, dmu_tx_t *tx)
{
	dsl_dir_set_qr_arg_t *ddsqra = arg;
	dsl_pool_t *dp = dmu_tx_pool(tx);
	dsl_dataset_t *ds;
	uint64_t newval;

	VERIFY0(dsl_dataset_hold(dp, ddsqra->ddsqra_name, FTAG, &ds));

	if (spa_version(dp->dp_spa) >= SPA_VERSION_RECVD_PROPS) {
		dsl_prop_set_sync_impl(ds,
		    zfs_prop_to_name(ZFS_PROP_RESERVATION),
		    ddsqra->ddsqra_source, sizeof (ddsqra->ddsqra_value), 1,
		    &ddsqra->ddsqra_value, tx);

		VERIFY0(dsl_prop_get_int_ds(ds,
		    zfs_prop_to_name(ZFS_PROP_RESERVATION), &newval));
	} else {
		newval = ddsqra->ddsqra_value;
		spa_history_log_internal_ds(ds, "set", tx, "%s=%lld",
		    zfs_prop_to_name(ZFS_PROP_RESERVATION),
		    (longlong_t)newval);
	}

	dsl_dir_set_reservation_sync_impl(ds->ds_dir, newval, tx);
	dsl_dataset_rele(ds, FTAG);
}

int
dsl_dir_set_reservation(const char *ddname, zprop_source_t source,
    uint64_t reservation)
{
	dsl_dir_set_qr_arg_t ddsqra;

	ddsqra.ddsqra_name = ddname;
	ddsqra.ddsqra_source = source;
	ddsqra.ddsqra_value = reservation;

	return (dsl_sync_task(ddname, dsl_dir_set_reservation_check,
	    dsl_dir_set_reservation_sync, &ddsqra, 0,
	    ZFS_SPACE_CHECK_EXTRA_RESERVED));
}

static dsl_dir_t *
closest_common_ancestor(dsl_dir_t *ds1, dsl_dir_t *ds2)
{
	for (; ds1; ds1 = ds1->dd_parent) {
		dsl_dir_t *dd;
		for (dd = ds2; dd; dd = dd->dd_parent) {
			if (ds1 == dd)
				return (dd);
		}
	}
	return (NULL);
}

/*
 * If delta is applied to dd, how much of that delta would be applied to
 * ancestor?  Syncing context only.
 */
static int64_t
would_change(dsl_dir_t *dd, int64_t delta, dsl_dir_t *ancestor)
{
	if (dd == ancestor)
		return (delta);

	mutex_enter(&dd->dd_lock);
	delta = parent_delta(dd, dsl_dir_phys(dd)->dd_used_bytes, delta);
	mutex_exit(&dd->dd_lock);
	return (would_change(dd->dd_parent, delta, ancestor));
}

typedef struct dsl_dir_rename_arg {
	const char *ddra_oldname;
	const char *ddra_newname;
	cred_t *ddra_cred;
	proc_t *ddra_proc;
} dsl_dir_rename_arg_t;

typedef struct dsl_valid_rename_arg {
	int char_delta;
	int nest_delta;
} dsl_valid_rename_arg_t;

static int
dsl_valid_rename(dsl_pool_t *dp, dsl_dataset_t *ds, void *arg)
{
	(void) dp;
	dsl_valid_rename_arg_t *dvra = arg;
	char namebuf[ZFS_MAX_DATASET_NAME_LEN];

	dsl_dataset_name(ds, namebuf);

	ASSERT3U(strnlen(namebuf, ZFS_MAX_DATASET_NAME_LEN),
	    <, ZFS_MAX_DATASET_NAME_LEN);
	int namelen = strlen(namebuf) + dvra->char_delta;
	int depth = get_dataset_depth(namebuf) + dvra->nest_delta;

	if (namelen >= ZFS_MAX_DATASET_NAME_LEN)
		return (SET_ERROR(ENAMETOOLONG));
	if (dvra->nest_delta > 0 && depth >= zfs_max_dataset_nesting)
		return (SET_ERROR(ENAMETOOLONG));
	return (0);
}

static int
dsl_dir_rename_check(void *arg, dmu_tx_t *tx)
{
	dsl_dir_rename_arg_t *ddra = arg;
	dsl_pool_t *dp = dmu_tx_pool(tx);
	dsl_dir_t *dd, *newparent;
	dsl_valid_rename_arg_t dvra;
	dsl_dataset_t *parentds;
	objset_t *parentos;
	const char *mynewname;
	int error;

	/* target dir should exist */
	error = dsl_dir_hold(dp, ddra->ddra_oldname, FTAG, &dd, NULL);
	if (error != 0)
		return (error);

	/* new parent should exist */
	error = dsl_dir_hold(dp, ddra->ddra_newname, FTAG,
	    &newparent, &mynewname);
	if (error != 0) {
		dsl_dir_rele(dd, FTAG);
		return (error);
	}

	/* can't rename to different pool */
	if (dd->dd_pool != newparent->dd_pool) {
		dsl_dir_rele(newparent, FTAG);
		dsl_dir_rele(dd, FTAG);
		return (SET_ERROR(EXDEV));
	}

	/* new name should not already exist */
	if (mynewname == NULL) {
		dsl_dir_rele(newparent, FTAG);
		dsl_dir_rele(dd, FTAG);
		return (SET_ERROR(EEXIST));
	}

	/* can't rename below anything but filesystems (eg. no ZVOLs) */
	error = dsl_dataset_hold_obj(newparent->dd_pool,
	    dsl_dir_phys(newparent)->dd_head_dataset_obj, FTAG, &parentds);
	if (error != 0) {
		dsl_dir_rele(newparent, FTAG);
		dsl_dir_rele(dd, FTAG);
		return (error);
	}
	error = dmu_objset_from_ds(parentds, &parentos);
	if (error != 0) {
		dsl_dataset_rele(parentds, FTAG);
		dsl_dir_rele(newparent, FTAG);
		dsl_dir_rele(dd, FTAG);
		return (error);
	}
	if (dmu_objset_type(parentos) != DMU_OST_ZFS) {
		dsl_dataset_rele(parentds, FTAG);
		dsl_dir_rele(newparent, FTAG);
		dsl_dir_rele(dd, FTAG);
		return (SET_ERROR(ZFS_ERR_WRONG_PARENT));
	}
	dsl_dataset_rele(parentds, FTAG);

	ASSERT3U(strnlen(ddra->ddra_newname, ZFS_MAX_DATASET_NAME_LEN),
	    <, ZFS_MAX_DATASET_NAME_LEN);
	ASSERT3U(strnlen(ddra->ddra_oldname, ZFS_MAX_DATASET_NAME_LEN),
	    <, ZFS_MAX_DATASET_NAME_LEN);
	dvra.char_delta = strlen(ddra->ddra_newname)
	    - strlen(ddra->ddra_oldname);
	dvra.nest_delta = get_dataset_depth(ddra->ddra_newname)
	    - get_dataset_depth(ddra->ddra_oldname);

	/* if the name length is growing, validate child name lengths */
	if (dvra.char_delta > 0 || dvra.nest_delta > 0) {
		error = dmu_objset_find_dp(dp, dd->dd_object, dsl_valid_rename,
		    &dvra, DS_FIND_CHILDREN | DS_FIND_SNAPSHOTS);
		if (error != 0) {
			dsl_dir_rele(newparent, FTAG);
			dsl_dir_rele(dd, FTAG);
			return (error);
		}
	}

	if (dmu_tx_is_syncing(tx)) {
		if (spa_feature_is_active(dp->dp_spa,
		    SPA_FEATURE_FS_SS_LIMIT)) {
			/*
			 * Although this is the check function and we don't
			 * normally make on-disk changes in check functions,
			 * we need to do that here.
			 *
			 * Ensure this portion of the tree's counts have been
			 * initialized in case the new parent has limits set.
			 */
			dsl_dir_init_fs_ss_count(dd, tx);
		}
	}

	if (newparent != dd->dd_parent) {
		/* is there enough space? */
		uint64_t myspace =
		    MAX(dsl_dir_phys(dd)->dd_used_bytes,
		    dsl_dir_phys(dd)->dd_reserved);
		objset_t *os = dd->dd_pool->dp_meta_objset;
		uint64_t fs_cnt = 0;
		uint64_t ss_cnt = 0;

		if (dsl_dir_is_zapified(dd)) {
			int err;

			err = zap_lookup(os, dd->dd_object,
			    DD_FIELD_FILESYSTEM_COUNT, sizeof (fs_cnt), 1,
			    &fs_cnt);
			if (err != ENOENT && err != 0) {
				dsl_dir_rele(newparent, FTAG);
				dsl_dir_rele(dd, FTAG);
				return (err);
			}

			/*
			 * have to add 1 for the filesystem itself that we're
			 * moving
			 */
			fs_cnt++;

			err = zap_lookup(os, dd->dd_object,
			    DD_FIELD_SNAPSHOT_COUNT, sizeof (ss_cnt), 1,
			    &ss_cnt);
			if (err != ENOENT && err != 0) {
				dsl_dir_rele(newparent, FTAG);
				dsl_dir_rele(dd, FTAG);
				return (err);
			}
		}

		/* check for encryption errors */
		error = dsl_dir_rename_crypt_check(dd, newparent);
		if (error != 0) {
			dsl_dir_rele(newparent, FTAG);
			dsl_dir_rele(dd, FTAG);
			return (SET_ERROR(EACCES));
		}

		/* no rename into our descendant */
		if (closest_common_ancestor(dd, newparent) == dd) {
			dsl_dir_rele(newparent, FTAG);
			dsl_dir_rele(dd, FTAG);
			return (SET_ERROR(EINVAL));
		}

		error = dsl_dir_transfer_possible(dd->dd_parent,
		    newparent, fs_cnt, ss_cnt, myspace,
		    ddra->ddra_cred, ddra->ddra_proc);
		if (error != 0) {
			dsl_dir_rele(newparent, FTAG);
			dsl_dir_rele(dd, FTAG);
			return (error);
		}
	}

	dsl_dir_rele(newparent, FTAG);
	dsl_dir_rele(dd, FTAG);
	return (0);
}

static void
dsl_dir_rename_sync(void *arg, dmu_tx_t *tx)
{
	dsl_dir_rename_arg_t *ddra = arg;
	dsl_pool_t *dp = dmu_tx_pool(tx);
	dsl_dir_t *dd, *newparent;
	const char *mynewname;
	objset_t *mos = dp->dp_meta_objset;

	VERIFY0(dsl_dir_hold(dp, ddra->ddra_oldname, FTAG, &dd, NULL));
	VERIFY0(dsl_dir_hold(dp, ddra->ddra_newname, FTAG, &newparent,
	    &mynewname));

	/* Log this before we change the name. */
	spa_history_log_internal_dd(dd, "rename", tx,
	    "-> %s", ddra->ddra_newname);

	if (newparent != dd->dd_parent) {
		objset_t *os = dd->dd_pool->dp_meta_objset;
		uint64_t fs_cnt = 0;
		uint64_t ss_cnt = 0;

		/*
		 * We already made sure the dd counts were initialized in the
		 * check function.
		 */
		if (spa_feature_is_active(dp->dp_spa,
		    SPA_FEATURE_FS_SS_LIMIT)) {
			VERIFY0(zap_lookup(os, dd->dd_object,
			    DD_FIELD_FILESYSTEM_COUNT, sizeof (fs_cnt), 1,
			    &fs_cnt));
			/* add 1 for the filesystem itself that we're moving */
			fs_cnt++;

			VERIFY0(zap_lookup(os, dd->dd_object,
			    DD_FIELD_SNAPSHOT_COUNT, sizeof (ss_cnt), 1,
			    &ss_cnt));
		}

		dsl_fs_ss_count_adjust(dd->dd_parent, -fs_cnt,
		    DD_FIELD_FILESYSTEM_COUNT, tx);
		dsl_fs_ss_count_adjust(newparent, fs_cnt,
		    DD_FIELD_FILESYSTEM_COUNT, tx);

		dsl_fs_ss_count_adjust(dd->dd_parent, -ss_cnt,
		    DD_FIELD_SNAPSHOT_COUNT, tx);
		dsl_fs_ss_count_adjust(newparent, ss_cnt,
		    DD_FIELD_SNAPSHOT_COUNT, tx);

		dsl_dir_diduse_space(dd->dd_parent, DD_USED_CHILD,
		    -dsl_dir_phys(dd)->dd_used_bytes,
		    -dsl_dir_phys(dd)->dd_compressed_bytes,
		    -dsl_dir_phys(dd)->dd_uncompressed_bytes, tx);
		dsl_dir_diduse_space(newparent, DD_USED_CHILD,
		    dsl_dir_phys(dd)->dd_used_bytes,
		    dsl_dir_phys(dd)->dd_compressed_bytes,
		    dsl_dir_phys(dd)->dd_uncompressed_bytes, tx);

		if (dsl_dir_phys(dd)->dd_reserved >
		    dsl_dir_phys(dd)->dd_used_bytes) {
			uint64_t unused_rsrv = dsl_dir_phys(dd)->dd_reserved -
			    dsl_dir_phys(dd)->dd_used_bytes;

			dsl_dir_diduse_space(dd->dd_parent, DD_USED_CHILD_RSRV,
			    -unused_rsrv, 0, 0, tx);
			dsl_dir_diduse_space(newparent, DD_USED_CHILD_RSRV,
			    unused_rsrv, 0, 0, tx);
		}
	}

	dmu_buf_will_dirty(dd->dd_dbuf, tx);

	/* remove from old parent zapobj */
	VERIFY0(zap_remove(mos,
	    dsl_dir_phys(dd->dd_parent)->dd_child_dir_zapobj,
	    dd->dd_myname, tx));

	(void) strlcpy(dd->dd_myname, mynewname,
	    sizeof (dd->dd_myname));
	dsl_dir_rele(dd->dd_parent, dd);
	dsl_dir_phys(dd)->dd_parent_obj = newparent->dd_object;
	VERIFY0(dsl_dir_hold_obj(dp,
	    newparent->dd_object, NULL, dd, &dd->dd_parent));

	/* add to new parent zapobj */
	VERIFY0(zap_add(mos, dsl_dir_phys(newparent)->dd_child_dir_zapobj,
	    dd->dd_myname, 8, 1, &dd->dd_object, tx));

	/* TODO: A rename callback to avoid these layering violations. */
	zfsvfs_update_fromname(ddra->ddra_oldname, ddra->ddra_newname);
	zvol_rename_minors(dp->dp_spa, ddra->ddra_oldname,
	    ddra->ddra_newname, B_TRUE);

	dsl_prop_notify_all(dd);

	dsl_dir_rele(newparent, FTAG);
	dsl_dir_rele(dd, FTAG);
}

int
dsl_dir_rename(const char *oldname, const char *newname)
{
	dsl_dir_rename_arg_t ddra;

	ddra.ddra_oldname = oldname;
	ddra.ddra_newname = newname;
	ddra.ddra_cred = CRED();
	ddra.ddra_proc = curproc;

	return (dsl_sync_task(oldname,
	    dsl_dir_rename_check, dsl_dir_rename_sync, &ddra,
	    3, ZFS_SPACE_CHECK_RESERVED));
}

int
dsl_dir_transfer_possible(dsl_dir_t *sdd, dsl_dir_t *tdd,
    uint64_t fs_cnt, uint64_t ss_cnt, uint64_t space,
    cred_t *cr, proc_t *proc)
{
	dsl_dir_t *ancestor;
	int64_t adelta;
	uint64_t avail;
	int err;

	ancestor = closest_common_ancestor(sdd, tdd);
	adelta = would_change(sdd, -space, ancestor);
	avail = dsl_dir_space_available(tdd, ancestor, adelta, FALSE);
	if (avail < space)
		return (SET_ERROR(ENOSPC));

	err = dsl_fs_ss_limit_check(tdd, fs_cnt, ZFS_PROP_FILESYSTEM_LIMIT,
	    ancestor, cr, proc);
	if (err != 0)
		return (err);
	err = dsl_fs_ss_limit_check(tdd, ss_cnt, ZFS_PROP_SNAPSHOT_LIMIT,
	    ancestor, cr, proc);
	if (err != 0)
		return (err);

	return (0);
}

inode_timespec_t
dsl_dir_snap_cmtime(dsl_dir_t *dd)
{
	inode_timespec_t t;

	mutex_enter(&dd->dd_lock);
	t = dd->dd_snap_cmtime;
	mutex_exit(&dd->dd_lock);

	return (t);
}

void
dsl_dir_snap_cmtime_update(dsl_dir_t *dd, dmu_tx_t *tx)
{
	inode_timespec_t t;
	objset_t *mos = dd->dd_pool->dp_meta_objset;
	uint64_t zapobj = dsl_dir_phys(dd)->dd_props_zapobj;
	const char *prop_name = zfs_prop_to_name(ZFS_PROP_SNAPSHOTS_CHANGED);

	gethrestime(&t);
	mutex_enter(&dd->dd_lock);
	dd->dd_snap_cmtime = t;
	VERIFY0(zap_update(mos, zapobj, prop_name, sizeof (uint64_t),
	    sizeof (inode_timespec_t) / sizeof (uint64_t), &t, tx));
	mutex_exit(&dd->dd_lock);
}

void
dsl_dir_zapify(dsl_dir_t *dd, dmu_tx_t *tx)
{
	objset_t *mos = dd->dd_pool->dp_meta_objset;
	dmu_object_zapify(mos, dd->dd_object, DMU_OT_DSL_DIR, tx);
}

boolean_t
dsl_dir_is_zapified(dsl_dir_t *dd)
{
	dmu_object_info_t doi;

	dmu_object_info_from_db(dd->dd_dbuf, &doi);
	return (doi.doi_type == DMU_OTN_ZAP_METADATA);
}

void
dsl_dir_livelist_open(dsl_dir_t *dd, uint64_t obj)
{
	objset_t *mos = dd->dd_pool->dp_meta_objset;
	ASSERT(spa_feature_is_active(dd->dd_pool->dp_spa,
	    SPA_FEATURE_LIVELIST));
	dsl_deadlist_open(&dd->dd_livelist, mos, obj);
	bplist_create(&dd->dd_pending_allocs);
	bplist_create(&dd->dd_pending_frees);
}

void
dsl_dir_livelist_close(dsl_dir_t *dd)
{
	dsl_deadlist_close(&dd->dd_livelist);
	bplist_destroy(&dd->dd_pending_allocs);
	bplist_destroy(&dd->dd_pending_frees);
}

void
dsl_dir_remove_livelist(dsl_dir_t *dd, dmu_tx_t *tx, boolean_t total)
{
	uint64_t obj;
	dsl_pool_t *dp = dmu_tx_pool(tx);
	spa_t *spa = dp->dp_spa;
	livelist_condense_entry_t to_condense = spa->spa_to_condense;

	if (!dsl_deadlist_is_open(&dd->dd_livelist))
		return;

	/*
	 * If the livelist being removed is set to be condensed, stop the
	 * condense zthr and indicate the cancellation in the spa_to_condense
	 * struct in case the condense no-wait synctask has already started
	 */
	zthr_t *ll_condense_thread = spa->spa_livelist_condense_zthr;
	if (ll_condense_thread != NULL &&
	    (to_condense.ds != NULL) && (to_condense.ds->ds_dir == dd)) {
		/*
		 * We use zthr_wait_cycle_done instead of zthr_cancel
		 * because we don't want to destroy the zthr, just have
		 * it skip its current task.
		 */
		spa->spa_to_condense.cancelled = B_TRUE;
		zthr_wait_cycle_done(ll_condense_thread);
		/*
		 * If we've returned from zthr_wait_cycle_done without
		 * clearing the to_condense data structure it's either
		 * because the no-wait synctask has started (which is
		 * indicated by 'syncing' field of to_condense) and we
		 * can expect it to clear to_condense on its own.
		 * Otherwise, we returned before the zthr ran. The
		 * checkfunc will now fail as cancelled == B_TRUE so we
		 * can safely NULL out ds, allowing a different dir's
		 * livelist to be condensed.
		 *
		 * We can be sure that the to_condense struct will not
		 * be repopulated at this stage because both this
		 * function and dsl_livelist_try_condense execute in
		 * syncing context.
		 */
		if ((spa->spa_to_condense.ds != NULL) &&
		    !spa->spa_to_condense.syncing) {
			dmu_buf_rele(spa->spa_to_condense.ds->ds_dbuf,
			    spa);
			spa->spa_to_condense.ds = NULL;
		}
	}

	dsl_dir_livelist_close(dd);
	VERIFY0(zap_lookup(dp->dp_meta_objset, dd->dd_object,
	    DD_FIELD_LIVELIST, sizeof (uint64_t), 1, &obj));
	VERIFY0(zap_remove(dp->dp_meta_objset, dd->dd_object,
	    DD_FIELD_LIVELIST, tx));
	if (total) {
		dsl_deadlist_free(dp->dp_meta_objset, obj, tx);
		spa_feature_decr(spa, SPA_FEATURE_LIVELIST, tx);
	}
}

static int
dsl_dir_activity_in_progress(dsl_dir_t *dd, dsl_dataset_t *ds,
    zfs_wait_activity_t activity, boolean_t *in_progress)
{
	int error = 0;

	ASSERT(MUTEX_HELD(&dd->dd_activity_lock));

	switch (activity) {
	case ZFS_WAIT_DELETEQ: {
#ifdef _KERNEL
		objset_t *os;
		error = dmu_objset_from_ds(ds, &os);
		if (error != 0)
			break;

		mutex_enter(&os->os_user_ptr_lock);
		void *user = dmu_objset_get_user(os);
		mutex_exit(&os->os_user_ptr_lock);
		if (dmu_objset_type(os) != DMU_OST_ZFS ||
		    user == NULL || zfs_get_vfs_flag_unmounted(os)) {
			*in_progress = B_FALSE;
			return (0);
		}

		uint64_t readonly = B_FALSE;
		error = zfs_get_temporary_prop(ds, ZFS_PROP_READONLY, &readonly,
		    NULL);

		if (error != 0)
			break;

		if (readonly || !spa_writeable(dd->dd_pool->dp_spa)) {
			*in_progress = B_FALSE;
			return (0);
		}

		uint64_t count, unlinked_obj;
		error = zap_lookup(os, MASTER_NODE_OBJ, ZFS_UNLINKED_SET, 8, 1,
		    &unlinked_obj);
		if (error != 0) {
			dsl_dataset_rele(ds, FTAG);
			break;
		}
		error = zap_count(os, unlinked_obj, &count);

		if (error == 0)
			*in_progress = (count != 0);
		break;
#else
		/*
		 * The delete queue is ZPL specific, and libzpool doesn't have
		 * it. It doesn't make sense to wait for it.
		 */
		(void) ds;
		*in_progress = B_FALSE;
		break;
#endif
	}
	default:
		panic("unrecognized value for activity %d", activity);
	}

	return (error);
}

int
dsl_dir_wait(dsl_dir_t *dd, dsl_dataset_t *ds, zfs_wait_activity_t activity,
    boolean_t *waited)
{
	int error = 0;
	boolean_t in_progress;
	dsl_pool_t *dp = dd->dd_pool;
	for (;;) {
		dsl_pool_config_enter(dp, FTAG);
		error = dsl_dir_activity_in_progress(dd, ds, activity,
		    &in_progress);
		dsl_pool_config_exit(dp, FTAG);
		if (error != 0 || !in_progress)
			break;

		*waited = B_TRUE;

		if (cv_wait_sig(&dd->dd_activity_cv, &dd->dd_activity_lock) ==
		    0 || dd->dd_activity_cancelled) {
			error = SET_ERROR(EINTR);
			break;
		}
	}
	return (error);
}

void
dsl_dir_cancel_waiters(dsl_dir_t *dd)
{
	mutex_enter(&dd->dd_activity_lock);
	dd->dd_activity_cancelled = B_TRUE;
	cv_broadcast(&dd->dd_activity_cv);
	while (dd->dd_activity_waiters > 0)
		cv_wait(&dd->dd_activity_cv, &dd->dd_activity_lock);
	mutex_exit(&dd->dd_activity_lock);
}

#if defined(_KERNEL)
EXPORT_SYMBOL(dsl_dir_set_quota);
EXPORT_SYMBOL(dsl_dir_set_reservation);
#endif