aboutsummaryrefslogtreecommitdiff
path: root/sys/ufs/ffs/ffs_snapshot.c
blob: 32dc47653d1897f9159cc5b8679ec0727eb7ecc2 (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
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
/*-
 * SPDX-License-Identifier: BSD-2-Clause-FreeBSD
 *
 * Copyright 2000 Marshall Kirk McKusick. All Rights Reserved.
 *
 * Further information about snapshots can be obtained from:
 *
 *	Marshall Kirk McKusick		http://www.mckusick.com/softdep/
 *	1614 Oxford Street		mckusick@mckusick.com
 *	Berkeley, CA 94709-1608		+1-510-843-9542
 *	USA
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 *
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 *
 * THIS SOFTWARE IS PROVIDED BY MARSHALL KIRK MCKUSICK ``AS IS'' AND ANY
 * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
 * DISCLAIMED.  IN NO EVENT SHALL MARSHALL KIRK MCKUSICK BE LIABLE FOR
 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 *
 *	@(#)ffs_snapshot.c	8.11 (McKusick) 7/23/00
 */

#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");

#include "opt_quota.h"

#include <sys/param.h>
#include <sys/kernel.h>
#include <sys/systm.h>
#include <sys/conf.h>
#include <sys/gsb_crc32.h>
#include <sys/bio.h>
#include <sys/buf.h>
#include <sys/fcntl.h>
#include <sys/proc.h>
#include <sys/namei.h>
#include <sys/sched.h>
#include <sys/stat.h>
#include <sys/malloc.h>
#include <sys/mount.h>
#include <sys/resource.h>
#include <sys/resourcevar.h>
#include <sys/rwlock.h>
#include <sys/vnode.h>

#include <geom/geom.h>

#include <ufs/ufs/extattr.h>
#include <ufs/ufs/quota.h>
#include <ufs/ufs/ufsmount.h>
#include <ufs/ufs/inode.h>
#include <ufs/ufs/ufs_extern.h>

#include <ufs/ffs/fs.h>
#include <ufs/ffs/ffs_extern.h>

#define KERNCRED thread0.td_ucred

#include "opt_ffs.h"

#ifdef NO_FFS_SNAPSHOT
int
ffs_snapshot(mp, snapfile)
	struct mount *mp;
	char *snapfile;
{
	return (EINVAL);
}

int
ffs_snapblkfree(fs, devvp, bno, size, inum, vtype, wkhd)
	struct fs *fs;
	struct vnode *devvp;
	ufs2_daddr_t bno;
	long size;
	ino_t inum;
	enum vtype vtype;
	struct workhead *wkhd;
{
	return (EINVAL);
}

void
ffs_snapremove(vp)
	struct vnode *vp;
{
}

void
ffs_snapshot_mount(mp)
	struct mount *mp;
{
}

void
ffs_snapshot_unmount(mp)
	struct mount *mp;
{
}

void
ffs_snapgone(ip)
	struct inode *ip;
{
}

int
ffs_copyonwrite(devvp, bp)
	struct vnode *devvp;
	struct buf *bp;
{
	return (EINVAL);
}

void
ffs_sync_snap(mp, waitfor)
	struct mount *mp;
	int waitfor;
{
}

#else
FEATURE(ffs_snapshot, "FFS snapshot support");

LIST_HEAD(, snapdata) snapfree;
static struct mtx snapfree_lock;
MTX_SYSINIT(ffs_snapfree, &snapfree_lock, "snapdata free list", MTX_DEF);

static int cgaccount(int, struct vnode *, struct buf *, int);
static int expunge_ufs1(struct vnode *, struct inode *, struct fs *,
    int (*)(struct vnode *, ufs1_daddr_t *, ufs1_daddr_t *, struct fs *,
    ufs_lbn_t, int), int, int);
static int indiracct_ufs1(struct vnode *, struct vnode *, int,
    ufs1_daddr_t, ufs_lbn_t, ufs_lbn_t, ufs_lbn_t, ufs_lbn_t, struct fs *,
    int (*)(struct vnode *, ufs1_daddr_t *, ufs1_daddr_t *, struct fs *,
    ufs_lbn_t, int), int);
static int fullacct_ufs1(struct vnode *, ufs1_daddr_t *, ufs1_daddr_t *,
    struct fs *, ufs_lbn_t, int);
static int snapacct_ufs1(struct vnode *, ufs1_daddr_t *, ufs1_daddr_t *,
    struct fs *, ufs_lbn_t, int);
static int mapacct_ufs1(struct vnode *, ufs1_daddr_t *, ufs1_daddr_t *,
    struct fs *, ufs_lbn_t, int);
static int expunge_ufs2(struct vnode *, struct inode *, struct fs *,
    int (*)(struct vnode *, ufs2_daddr_t *, ufs2_daddr_t *, struct fs *,
    ufs_lbn_t, int), int, int);
static int indiracct_ufs2(struct vnode *, struct vnode *, int,
    ufs2_daddr_t, ufs_lbn_t, ufs_lbn_t, ufs_lbn_t, ufs_lbn_t, struct fs *,
    int (*)(struct vnode *, ufs2_daddr_t *, ufs2_daddr_t *, struct fs *,
    ufs_lbn_t, int), int);
static int fullacct_ufs2(struct vnode *, ufs2_daddr_t *, ufs2_daddr_t *,
    struct fs *, ufs_lbn_t, int);
static int snapacct_ufs2(struct vnode *, ufs2_daddr_t *, ufs2_daddr_t *,
    struct fs *, ufs_lbn_t, int);
static int mapacct_ufs2(struct vnode *, ufs2_daddr_t *, ufs2_daddr_t *,
    struct fs *, ufs_lbn_t, int);
static int readblock(struct vnode *vp, struct buf *, ufs2_daddr_t);
static void try_free_snapdata(struct vnode *devvp);
static struct snapdata *ffs_snapdata_acquire(struct vnode *devvp);
static int ffs_bp_snapblk(struct vnode *, struct buf *);

/*
 * To ensure the consistency of snapshots across crashes, we must
 * synchronously write out copied blocks before allowing the
 * originals to be modified. Because of the rather severe speed
 * penalty that this imposes, the code normally only ensures
 * persistence for the filesystem metadata contained within a
 * snapshot. Setting the following flag allows this crash
 * persistence to be enabled for file contents.
 */
int dopersistence = 0;

#ifdef DIAGNOSTIC
#include <sys/sysctl.h>
SYSCTL_INT(_debug, OID_AUTO, dopersistence, CTLFLAG_RW, &dopersistence, 0, "");
static int snapdebug = 0;
SYSCTL_INT(_debug, OID_AUTO, snapdebug, CTLFLAG_RW, &snapdebug, 0, "");
int collectsnapstats = 0;
SYSCTL_INT(_debug, OID_AUTO, collectsnapstats, CTLFLAG_RW, &collectsnapstats,
	0, "");
#endif /* DIAGNOSTIC */

/*
 * Create a snapshot file and initialize it for the filesystem.
 */
int
ffs_snapshot(mp, snapfile)
	struct mount *mp;
	char *snapfile;
{
	ufs2_daddr_t numblks, blkno, *blkp, *snapblklist;
	int error, cg, snaploc;
	int i, size, len, loc;
	ufs2_daddr_t blockno;
	uint64_t flag;
	char saved_nice = 0;
	long redo = 0, snaplistsize = 0;
	int32_t *lp;
	void *space;
	struct fs *copy_fs = NULL, *fs;
	struct thread *td = curthread;
	struct inode *ip, *xp;
	struct buf *bp, *nbp, *ibp;
	struct nameidata nd;
	struct mount *wrtmp;
	struct vattr vat;
	struct vnode *vp, *xvp, *mvp, *devvp;
	struct uio auio;
	struct iovec aiov;
	struct snapdata *sn;
	struct ufsmount *ump;
#ifdef DIAGNOSTIC
	struct timespec starttime = {0, 0}, endtime;
#endif

	ump = VFSTOUFS(mp);
	fs = ump->um_fs;
	sn = NULL;
	/*
	 * At the moment, journaled soft updates cannot support
	 * taking snapshots.
	 */
	if (MOUNTEDSUJ(mp)) {
		vfs_mount_error(mp, "%s: Snapshots are not yet supported when "
		    "running with journaled soft updates", fs->fs_fsmnt);
		return (EOPNOTSUPP);
	}
	MNT_ILOCK(mp);
	flag = mp->mnt_flag;
	MNT_IUNLOCK(mp);
	/*
	 * Need to serialize access to snapshot code per filesystem.
	 */
	/*
	 * Assign a snapshot slot in the superblock.
	 */
	UFS_LOCK(ump);
	for (snaploc = 0; snaploc < FSMAXSNAP; snaploc++)
		if (fs->fs_snapinum[snaploc] == 0)
			break;
	UFS_UNLOCK(ump);
	if (snaploc == FSMAXSNAP)
		return (ENOSPC);
	/*
	 * Create the snapshot file.
	 */
restart:
	NDINIT(&nd, CREATE, LOCKPARENT | LOCKLEAF | NOCACHE, UIO_SYSSPACE,
	    snapfile, td);
	if ((error = namei(&nd)) != 0)
		return (error);
	if (nd.ni_vp != NULL) {
		vput(nd.ni_vp);
		error = EEXIST;
	}
	if (nd.ni_dvp->v_mount != mp)
		error = EXDEV;
	if (error) {
		NDFREE(&nd, NDF_ONLY_PNBUF);
		if (nd.ni_dvp == nd.ni_vp)
			vrele(nd.ni_dvp);
		else
			vput(nd.ni_dvp);
		return (error);
	}
	VATTR_NULL(&vat);
	vat.va_type = VREG;
	vat.va_mode = S_IRUSR;
	vat.va_vaflags |= VA_EXCLUSIVE;
	if (VOP_GETWRITEMOUNT(nd.ni_dvp, &wrtmp))
		wrtmp = NULL;
	if (wrtmp != mp)
		panic("ffs_snapshot: mount mismatch");
	vfs_rel(wrtmp);
	if (vn_start_write(NULL, &wrtmp, V_NOWAIT) != 0) {
		NDFREE(&nd, NDF_ONLY_PNBUF);
		vput(nd.ni_dvp);
		if ((error = vn_start_write(NULL, &wrtmp,
		    V_XSLEEP | PCATCH)) != 0)
			return (error);
		goto restart;
	}
	error = VOP_CREATE(nd.ni_dvp, &nd.ni_vp, &nd.ni_cnd, &vat);
	VOP_UNLOCK(nd.ni_dvp);
	if (error) {
		NDFREE(&nd, NDF_ONLY_PNBUF);
		vn_finished_write(wrtmp);
		vrele(nd.ni_dvp);
		if (error == ERELOOKUP)
			goto restart;
		return (error);
	}
	vp = nd.ni_vp;
	vnode_create_vobject(nd.ni_vp, fs->fs_size, td);
	vp->v_vflag |= VV_SYSTEM;
	ip = VTOI(vp);
	devvp = ITODEVVP(ip);
	/*
	 * Allocate and copy the last block contents so as to be able
	 * to set size to that of the filesystem.
	 */
	numblks = howmany(fs->fs_size, fs->fs_frag);
	error = UFS_BALLOC(vp, lblktosize(fs, (off_t)(numblks - 1)),
	    fs->fs_bsize, KERNCRED, BA_CLRBUF, &bp);
	if (error)
		goto out;
	ip->i_size = lblktosize(fs, (off_t)numblks);
	DIP_SET(ip, i_size, ip->i_size);
	UFS_INODE_SET_FLAG(ip, IN_SIZEMOD | IN_CHANGE | IN_UPDATE);
	error = readblock(vp, bp, numblks - 1);
	bawrite(bp);
	if (error != 0)
		goto out;
	/*
	 * Preallocate critical data structures so that we can copy
	 * them in without further allocation after we suspend all
	 * operations on the filesystem. We would like to just release
	 * the allocated buffers without writing them since they will
	 * be filled in below once we are ready to go, but this upsets
	 * the soft update code, so we go ahead and write the new buffers.
	 *
	 * Allocate all indirect blocks and mark all of them as not
	 * needing to be copied.
	 */
	for (blkno = UFS_NDADDR; blkno < numblks; blkno += NINDIR(fs)) {
		error = UFS_BALLOC(vp, lblktosize(fs, (off_t)blkno),
		    fs->fs_bsize, td->td_ucred, BA_METAONLY, &ibp);
		if (error)
			goto out;
		bawrite(ibp);
	}
	/*
	 * Allocate copies for the superblock and its summary information.
	 */
	error = UFS_BALLOC(vp, fs->fs_sblockloc, fs->fs_sbsize, KERNCRED,
	    0, &nbp);
	if (error)
		goto out;
	bawrite(nbp);
	blkno = fragstoblks(fs, fs->fs_csaddr);
	len = howmany(fs->fs_cssize, fs->fs_bsize);
	for (loc = 0; loc < len; loc++) {
		error = UFS_BALLOC(vp, lblktosize(fs, (off_t)(blkno + loc)),
		    fs->fs_bsize, KERNCRED, 0, &nbp);
		if (error)
			goto out;
		bawrite(nbp);
	}
	/*
	 * Allocate all cylinder group blocks.
	 */
	for (cg = 0; cg < fs->fs_ncg; cg++) {
		error = UFS_BALLOC(vp, lfragtosize(fs, cgtod(fs, cg)),
		    fs->fs_bsize, KERNCRED, 0, &nbp);
		if (error)
			goto out;
		bawrite(nbp);
		if (cg % 10 == 0) {
			error = ffs_syncvnode(vp, MNT_WAIT, 0);
			/* vp possibly reclaimed if unlocked */
			if (error != 0)
				goto out;
		}
	}
	/*
	 * Copy all the cylinder group maps. Although the
	 * filesystem is still active, we hope that only a few
	 * cylinder groups will change between now and when we
	 * suspend operations. Thus, we will be able to quickly
	 * touch up the few cylinder groups that changed during
	 * the suspension period.
	 */
	len = roundup2(howmany(fs->fs_ncg, NBBY), sizeof(int));
	space = malloc(len, M_DEVBUF, M_WAITOK | M_ZERO);
	UFS_LOCK(ump);
	fs->fs_active = space;
	UFS_UNLOCK(ump);
	for (cg = 0; cg < fs->fs_ncg; cg++) {
		error = UFS_BALLOC(vp, lfragtosize(fs, cgtod(fs, cg)),
		    fs->fs_bsize, KERNCRED, 0, &nbp);
		if (error)
			goto out;
		error = cgaccount(cg, vp, nbp, 1);
		bawrite(nbp);
		if (cg % 10 == 0 && error == 0)
			error = ffs_syncvnode(vp, MNT_WAIT, 0);
		if (error)
			goto out;
	}
	/*
	 * Change inode to snapshot type file.
	 */
	ip->i_flags |= SF_SNAPSHOT;
	DIP_SET(ip, i_flags, ip->i_flags);
	UFS_INODE_SET_FLAG(ip, IN_CHANGE | IN_UPDATE);
	/*
	 * Ensure that the snapshot is completely on disk.
	 * Since we have marked it as a snapshot it is safe to
	 * unlock it as no process will be allowed to write to it.
	 */
	if ((error = ffs_syncvnode(vp, MNT_WAIT, 0)) != 0)
		goto out;
	VOP_UNLOCK(vp);
	/*
	 * All allocations are done, so we can now snapshot the system.
	 *
	 * Recind nice scheduling while running with the filesystem suspended.
	 */
	if (td->td_proc->p_nice > 0) {
		struct proc *p;

		p = td->td_proc;
		PROC_LOCK(p);
		saved_nice = p->p_nice;
		sched_nice(p, 0);
		PROC_UNLOCK(p);
	}
	/*
	 * Suspend operation on filesystem.
	 */
	for (;;) {
		vn_finished_write(wrtmp);
		if ((error = vfs_write_suspend(vp->v_mount, 0)) != 0) {
			vn_start_write(NULL, &wrtmp, V_WAIT);
			vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
			goto out;
		}
		if (mp->mnt_kern_flag & MNTK_SUSPENDED)
			break;
		vn_start_write(NULL, &wrtmp, V_WAIT);
	}
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
	if (ip->i_effnlink == 0) {
		error = ENOENT;		/* Snapshot file unlinked */
		goto out1;
	}
#ifdef DIAGNOSTIC
	if (collectsnapstats)
		nanotime(&starttime);
#endif

	/* The last block might have changed.  Copy it again to be sure. */
	error = UFS_BALLOC(vp, lblktosize(fs, (off_t)(numblks - 1)),
	    fs->fs_bsize, KERNCRED, BA_CLRBUF, &bp);
	if (error != 0)
		goto out1;
	error = readblock(vp, bp, numblks - 1);
	bp->b_flags |= B_VALIDSUSPWRT;
	bawrite(bp);
	if (error != 0)
		goto out1;
	/*
	 * First, copy all the cylinder group maps that have changed.
	 */
	for (cg = 0; cg < fs->fs_ncg; cg++) {
		if ((ACTIVECGNUM(fs, cg) & ACTIVECGOFF(cg)) != 0)
			continue;
		redo++;
		error = UFS_BALLOC(vp, lfragtosize(fs, cgtod(fs, cg)),
		    fs->fs_bsize, KERNCRED, 0, &nbp);
		if (error)
			goto out1;
		error = cgaccount(cg, vp, nbp, 2);
		bawrite(nbp);
		if (error)
			goto out1;
	}
	/*
	 * Grab a copy of the superblock and its summary information.
	 * We delay writing it until the suspension is released below.
	 */
	copy_fs = malloc((u_long)fs->fs_bsize, M_UFSMNT, M_WAITOK);
	bcopy(fs, copy_fs, fs->fs_sbsize);
	copy_fs->fs_si = malloc(sizeof(struct fs_summary_info), M_UFSMNT,
	    M_ZERO | M_WAITOK);
	if ((fs->fs_flags & (FS_UNCLEAN | FS_NEEDSFSCK)) == 0)
		copy_fs->fs_clean = 1;
	size = fs->fs_bsize < SBLOCKSIZE ? fs->fs_bsize : SBLOCKSIZE;
	if (fs->fs_sbsize < size)
		bzero(&((char *)copy_fs)[fs->fs_sbsize],
		    size - fs->fs_sbsize);
	size = blkroundup(fs, fs->fs_cssize);
	if (fs->fs_contigsumsize > 0)
		size += fs->fs_ncg * sizeof(int32_t);
	space = malloc((u_long)size, M_UFSMNT, M_WAITOK);
	copy_fs->fs_csp = space;
	bcopy(fs->fs_csp, copy_fs->fs_csp, fs->fs_cssize);
	space = (char *)space + fs->fs_cssize;
	loc = howmany(fs->fs_cssize, fs->fs_fsize);
	i = fs->fs_frag - loc % fs->fs_frag;
	len = (i == fs->fs_frag) ? 0 : i * fs->fs_fsize;
	if (len > 0) {
		if ((error = bread(devvp, fsbtodb(fs, fs->fs_csaddr + loc),
		    len, KERNCRED, &bp)) != 0) {
			brelse(bp);
			free(copy_fs->fs_csp, M_UFSMNT);
			free(copy_fs->fs_si, M_UFSMNT);
			free(copy_fs, M_UFSMNT);
			copy_fs = NULL;
			goto out1;
		}
		bcopy(bp->b_data, space, (u_int)len);
		space = (char *)space + len;
		bp->b_flags |= B_INVAL | B_NOCACHE;
		brelse(bp);
	}
	if (fs->fs_contigsumsize > 0) {
		copy_fs->fs_maxcluster = lp = space;
		for (i = 0; i < fs->fs_ncg; i++)
			*lp++ = fs->fs_contigsumsize;
	}
	/*
	 * We must check for active files that have been unlinked
	 * (e.g., with a zero link count). We have to expunge all
	 * trace of these files from the snapshot so that they are
	 * not reclaimed prematurely by fsck or unnecessarily dumped.
	 * We turn off the MNTK_SUSPENDED flag to avoid a panic from
	 * spec_strategy about writing on a suspended filesystem.
	 * Note that we skip unlinked snapshot files as they will
	 * be handled separately below.
	 *
	 * We also calculate the needed size for the snapshot list.
	 */
	snaplistsize = fs->fs_ncg + howmany(fs->fs_cssize, fs->fs_bsize) +
	    FSMAXSNAP + 1 /* superblock */ + 1 /* last block */ + 1 /* size */;
	MNT_ILOCK(mp);
	mp->mnt_kern_flag &= ~MNTK_SUSPENDED;
	MNT_IUNLOCK(mp);
loop:
	MNT_VNODE_FOREACH_ALL(xvp, mp, mvp) {
		if ((xvp->v_usecount == 0 &&
		     (xvp->v_iflag & (VI_OWEINACT | VI_DOINGINACT)) == 0) ||
		    xvp->v_type == VNON ||
		    IS_SNAPSHOT(VTOI(xvp))) {
			VI_UNLOCK(xvp);
			continue;
		}
		/*
		 * We can skip parent directory vnode because it must have
		 * this snapshot file in it.
		 */
		if (xvp == nd.ni_dvp) {
			VI_UNLOCK(xvp);
			continue;
		}
		vholdl(xvp);
		if (vn_lock(xvp, LK_EXCLUSIVE | LK_INTERLOCK) != 0) {
			MNT_VNODE_FOREACH_ALL_ABORT(mp, mvp);
			vdrop(xvp);
			goto loop;
		}
		VI_LOCK(xvp);
		if (xvp->v_usecount == 0 &&
		    (xvp->v_iflag & (VI_OWEINACT | VI_DOINGINACT)) == 0) {
			VI_UNLOCK(xvp);
			VOP_UNLOCK(xvp);
			vdrop(xvp);
			continue;
		}
		VI_UNLOCK(xvp);
#ifdef DIAGNOSTIC
		if (snapdebug)
			vn_printf(xvp, "ffs_snapshot: busy vnode ");
#endif
		if (VOP_GETATTR(xvp, &vat, td->td_ucred) == 0 &&
		    vat.va_nlink > 0) {
			VOP_UNLOCK(xvp);
			vdrop(xvp);
			continue;
		}
		xp = VTOI(xvp);
		if (ffs_checkfreefile(copy_fs, vp, xp->i_number)) {
			VOP_UNLOCK(xvp);
			vdrop(xvp);
			continue;
		}
		/*
		 * If there is a fragment, clear it here.
		 */
		blkno = 0;
		loc = howmany(xp->i_size, fs->fs_bsize) - 1;
		if (loc < UFS_NDADDR) {
			len = fragroundup(fs, blkoff(fs, xp->i_size));
			if (len != 0 && len < fs->fs_bsize) {
				ffs_blkfree(ump, copy_fs, vp,
				    DIP(xp, i_db[loc]), len, xp->i_number,
				    xvp->v_type, NULL, SINGLETON_KEY);
				blkno = DIP(xp, i_db[loc]);
				DIP_SET(xp, i_db[loc], 0);
			}
		}
		snaplistsize += 1;
		if (I_IS_UFS1(xp))
			error = expunge_ufs1(vp, xp, copy_fs, fullacct_ufs1,
			    BLK_NOCOPY, 1);
		else
			error = expunge_ufs2(vp, xp, copy_fs, fullacct_ufs2,
			    BLK_NOCOPY, 1);
		if (blkno)
			DIP_SET(xp, i_db[loc], blkno);
		if (!error)
			error = ffs_freefile(ump, copy_fs, vp, xp->i_number,
			    xp->i_mode, NULL);
		VOP_UNLOCK(xvp);
		vdrop(xvp);
		if (error) {
			free(copy_fs->fs_csp, M_UFSMNT);
			free(copy_fs->fs_si, M_UFSMNT);
			free(copy_fs, M_UFSMNT);
			copy_fs = NULL;
			MNT_VNODE_FOREACH_ALL_ABORT(mp, mvp);
			goto out1;
		}
	}
	/*
	 * Erase the journal file from the snapshot.
	 */
	if (fs->fs_flags & FS_SUJ) {
		error = softdep_journal_lookup(mp, &xvp);
		if (error) {
			free(copy_fs->fs_csp, M_UFSMNT);
			free(copy_fs->fs_si, M_UFSMNT);
			free(copy_fs, M_UFSMNT);
			copy_fs = NULL;
			goto out1;
		}
		xp = VTOI(xvp);
		if (I_IS_UFS1(xp))
			error = expunge_ufs1(vp, xp, copy_fs, fullacct_ufs1,
			    BLK_NOCOPY, 0);
		else
			error = expunge_ufs2(vp, xp, copy_fs, fullacct_ufs2,
			    BLK_NOCOPY, 0);
		vput(xvp);
	}
	/*
	 * Acquire a lock on the snapdata structure, creating it if necessary.
	 */
	sn = ffs_snapdata_acquire(devvp);
	/* 
	 * Change vnode to use shared snapshot lock instead of the original
	 * private lock.
	 */
	vp->v_vnlock = &sn->sn_lock;
	lockmgr(&vp->v_lock, LK_RELEASE, NULL);
	xp = TAILQ_FIRST(&sn->sn_head);
	/*
	 * If this is the first snapshot on this filesystem, then we need
	 * to allocate the space for the list of preallocated snapshot blocks.
	 * This list will be refined below, but this preliminary one will
	 * keep us out of deadlock until the full one is ready.
	 */
	if (xp == NULL) {
		snapblklist = malloc(snaplistsize * sizeof(daddr_t),
		    M_UFSMNT, M_WAITOK);
		blkp = &snapblklist[1];
		*blkp++ = lblkno(fs, fs->fs_sblockloc);
		blkno = fragstoblks(fs, fs->fs_csaddr);
		for (cg = 0; cg < fs->fs_ncg; cg++) {
			if (fragstoblks(fs, cgtod(fs, cg) > blkno))
				break;
			*blkp++ = fragstoblks(fs, cgtod(fs, cg));
		}
		len = howmany(fs->fs_cssize, fs->fs_bsize);
		for (loc = 0; loc < len; loc++)
			*blkp++ = blkno + loc;
		for (; cg < fs->fs_ncg; cg++)
			*blkp++ = fragstoblks(fs, cgtod(fs, cg));
		snapblklist[0] = blkp - snapblklist;
		VI_LOCK(devvp);
		if (sn->sn_blklist != NULL)
			panic("ffs_snapshot: non-empty list");
		sn->sn_blklist = snapblklist;
		sn->sn_listsize = blkp - snapblklist;
		VI_UNLOCK(devvp);
	}
	/*
	 * Record snapshot inode. Since this is the newest snapshot,
	 * it must be placed at the end of the list.
	 */
	VI_LOCK(devvp);
	fs->fs_snapinum[snaploc] = ip->i_number;
	if (ip->i_nextsnap.tqe_prev != 0)
		panic("ffs_snapshot: %ju already on list",
		    (uintmax_t)ip->i_number);
	TAILQ_INSERT_TAIL(&sn->sn_head, ip, i_nextsnap);
	devvp->v_vflag |= VV_COPYONWRITE;
	VI_UNLOCK(devvp);
	ASSERT_VOP_LOCKED(vp, "ffs_snapshot vp");
out1:
	KASSERT((sn != NULL && copy_fs != NULL && error == 0) ||
		(sn == NULL && copy_fs == NULL && error != 0),
		("email phk@ and mckusick@"));
	/*
	 * Resume operation on filesystem.
	 */
	vfs_write_resume(vp->v_mount, VR_START_WRITE | VR_NO_SUSPCLR);
#ifdef DIAGNOSTIC
	if (collectsnapstats && starttime.tv_sec > 0) {
		nanotime(&endtime);
		timespecsub(&endtime, &starttime, &endtime);
		printf("%s: suspended %ld.%03ld sec, redo %ld of %d\n",
		    vp->v_mount->mnt_stat.f_mntonname, (long)endtime.tv_sec,
		    endtime.tv_nsec / 1000000, redo, fs->fs_ncg);
	}
#endif
	if (copy_fs == NULL)
		goto out;
	/*
	 * Copy allocation information from all the snapshots in
	 * this snapshot and then expunge them from its view.
	 */
	TAILQ_FOREACH(xp, &sn->sn_head, i_nextsnap) {
		if (xp == ip)
			break;
		if (I_IS_UFS1(xp))
			error = expunge_ufs1(vp, xp, fs, snapacct_ufs1,
			    BLK_SNAP, 0);
		else
			error = expunge_ufs2(vp, xp, fs, snapacct_ufs2,
			    BLK_SNAP, 0);
		if (error == 0 && xp->i_effnlink == 0) {
			error = ffs_freefile(ump,
					     copy_fs,
					     vp,
					     xp->i_number,
					     xp->i_mode, NULL);
		}
		if (error) {
			fs->fs_snapinum[snaploc] = 0;
			goto done;
		}
	}
	/*
	 * Allocate space for the full list of preallocated snapshot blocks.
	 */
	snapblklist = malloc(snaplistsize * sizeof(daddr_t),
	    M_UFSMNT, M_WAITOK);
	ip->i_snapblklist = &snapblklist[1];
	/*
	 * Expunge the blocks used by the snapshots from the set of
	 * blocks marked as used in the snapshot bitmaps. Also, collect
	 * the list of allocated blocks in i_snapblklist.
	 */
	if (I_IS_UFS1(ip))
		error = expunge_ufs1(vp, ip, copy_fs, mapacct_ufs1,
		    BLK_SNAP, 0);
	else
		error = expunge_ufs2(vp, ip, copy_fs, mapacct_ufs2,
		    BLK_SNAP, 0);
	if (error) {
		fs->fs_snapinum[snaploc] = 0;
		free(snapblklist, M_UFSMNT);
		goto done;
	}
	if (snaplistsize < ip->i_snapblklist - snapblklist)
		panic("ffs_snapshot: list too small");
	snaplistsize = ip->i_snapblklist - snapblklist;
	snapblklist[0] = snaplistsize;
	ip->i_snapblklist = 0;
	/*
	 * Write out the list of allocated blocks to the end of the snapshot.
	 */
	auio.uio_iov = &aiov;
	auio.uio_iovcnt = 1;
	aiov.iov_base = (void *)snapblklist;
	aiov.iov_len = snaplistsize * sizeof(daddr_t);
	auio.uio_resid = aiov.iov_len;
	auio.uio_offset = ip->i_size;
	auio.uio_segflg = UIO_SYSSPACE;
	auio.uio_rw = UIO_WRITE;
	auio.uio_td = td;
	if ((error = VOP_WRITE(vp, &auio, IO_UNIT, td->td_ucred)) != 0) {
		fs->fs_snapinum[snaploc] = 0;
		free(snapblklist, M_UFSMNT);
		goto done;
	}
	/*
	 * Write the superblock and its summary information
	 * to the snapshot.
	 */
	blkno = fragstoblks(fs, fs->fs_csaddr);
	len = howmany(fs->fs_cssize, fs->fs_bsize);
	space = copy_fs->fs_csp;
	for (loc = 0; loc < len; loc++) {
		error = bread(vp, blkno + loc, fs->fs_bsize, KERNCRED, &nbp);
		if (error) {
			fs->fs_snapinum[snaploc] = 0;
			free(snapblklist, M_UFSMNT);
			goto done;
		}
		bcopy(space, nbp->b_data, fs->fs_bsize);
		space = (char *)space + fs->fs_bsize;
		bawrite(nbp);
	}
	error = bread(vp, lblkno(fs, fs->fs_sblockloc), fs->fs_bsize,
	    KERNCRED, &nbp);
	if (error) {
		brelse(nbp);
	} else {
		loc = blkoff(fs, fs->fs_sblockloc);
		copy_fs->fs_fmod = 0;
		copy_fs->fs_ckhash = ffs_calc_sbhash(copy_fs);
		bcopy((char *)copy_fs, &nbp->b_data[loc], (u_int)fs->fs_sbsize);
		bawrite(nbp);
	}
	/*
	 * As this is the newest list, it is the most inclusive, so
	 * should replace the previous list.
	 */
	VI_LOCK(devvp);
	space = sn->sn_blklist;
	sn->sn_blklist = snapblklist;
	sn->sn_listsize = snaplistsize;
	VI_UNLOCK(devvp);
	if (space != NULL)
		free(space, M_UFSMNT);
	/*
	 * Preallocate all the direct blocks in the snapshot inode so
	 * that we never have to write the inode itself to commit an
	 * update to the contents of the snapshot. Note that once
	 * created, the size of the snapshot will never change, so
	 * there will never be a need to write the inode except to
	 * update the non-integrity-critical time fields and
	 * allocated-block count.
	 */
	for (blockno = 0; blockno < UFS_NDADDR; blockno++) {
		if (DIP(ip, i_db[blockno]) != 0)
			continue;
		error = UFS_BALLOC(vp, lblktosize(fs, blockno),
		    fs->fs_bsize, KERNCRED, BA_CLRBUF, &bp);
		if (error)
			break;
		error = readblock(vp, bp, blockno);
		bawrite(bp);
		if (error != 0)
			break;
	}
done:
	free(copy_fs->fs_csp, M_UFSMNT);
	free(copy_fs->fs_si, M_UFSMNT);
	free(copy_fs, M_UFSMNT);
	copy_fs = NULL;
out:
	NDFREE(&nd, NDF_ONLY_PNBUF);
	if (saved_nice > 0) {
		struct proc *p;

		p = td->td_proc;
		PROC_LOCK(p);
		sched_nice(td->td_proc, saved_nice);
		PROC_UNLOCK(td->td_proc);
	}
	UFS_LOCK(ump);
	if (fs->fs_active != 0) {
		free(fs->fs_active, M_DEVBUF);
		fs->fs_active = 0;
	}
	UFS_UNLOCK(ump);
	MNT_ILOCK(mp);
	mp->mnt_flag = (mp->mnt_flag & MNT_QUOTA) | (flag & ~MNT_QUOTA);
	MNT_IUNLOCK(mp);
	if (error)
		(void) ffs_truncate(vp, (off_t)0, 0, NOCRED);
	(void) ffs_syncvnode(vp, MNT_WAIT, 0);
	if (error)
		vput(vp);
	else
		VOP_UNLOCK(vp);
	vrele(nd.ni_dvp);
	vn_finished_write(wrtmp);
	process_deferred_inactive(mp);
	return (error);
}

/*
 * Copy a cylinder group map. All the unallocated blocks are marked
 * BLK_NOCOPY so that the snapshot knows that it need not copy them
 * if they are later written. If passno is one, then this is a first
 * pass, so only setting needs to be done. If passno is 2, then this
 * is a revision to a previous pass which must be undone as the
 * replacement pass is done.
 */
static int
cgaccount(cg, vp, nbp, passno)
	int cg;
	struct vnode *vp;
	struct buf *nbp;
	int passno;
{
	struct buf *bp, *ibp;
	struct inode *ip;
	struct cg *cgp;
	struct fs *fs;
	ufs2_daddr_t base, numblks;
	int error, len, loc, indiroff;

	ip = VTOI(vp);
	fs = ITOFS(ip);
	if ((error = ffs_getcg(fs, ITODEVVP(ip), cg, 0, &bp, &cgp)) != 0)
		return (error);
	UFS_LOCK(ITOUMP(ip));
	ACTIVESET(fs, cg);
	/*
	 * Recomputation of summary information might not have been performed
	 * at mount time.  Sync up summary information for current cylinder
	 * group while data is in memory to ensure that result of background
	 * fsck is slightly more consistent.
	 */
	fs->fs_cs(fs, cg) = cgp->cg_cs;
	UFS_UNLOCK(ITOUMP(ip));
	bcopy(bp->b_data, nbp->b_data, fs->fs_cgsize);
	if (fs->fs_cgsize < fs->fs_bsize)
		bzero(&nbp->b_data[fs->fs_cgsize],
		    fs->fs_bsize - fs->fs_cgsize);
	cgp = (struct cg *)nbp->b_data;
	bqrelse(bp);
	if (passno == 2)
		nbp->b_flags |= B_VALIDSUSPWRT;
	numblks = howmany(fs->fs_size, fs->fs_frag);
	len = howmany(fs->fs_fpg, fs->fs_frag);
	base = cgbase(fs, cg) / fs->fs_frag;
	if (base + len >= numblks)
		len = numblks - base - 1;
	loc = 0;
	if (base < UFS_NDADDR) {
		for ( ; loc < UFS_NDADDR; loc++) {
			if (ffs_isblock(fs, cg_blksfree(cgp), loc))
				DIP_SET(ip, i_db[loc], BLK_NOCOPY);
			else if (passno == 2 && DIP(ip, i_db[loc])== BLK_NOCOPY)
				DIP_SET(ip, i_db[loc], 0);
			else if (passno == 1 && DIP(ip, i_db[loc])== BLK_NOCOPY)
				panic("ffs_snapshot: lost direct block");
		}
	}
	error = UFS_BALLOC(vp, lblktosize(fs, (off_t)(base + loc)),
	    fs->fs_bsize, KERNCRED, BA_METAONLY, &ibp);
	if (error) {
		goto out;
	}
	indiroff = (base + loc - UFS_NDADDR) % NINDIR(fs);
	for ( ; loc < len; loc++, indiroff++) {
		if (indiroff >= NINDIR(fs)) {
			if (passno == 2)
				ibp->b_flags |= B_VALIDSUSPWRT;
			bawrite(ibp);
			error = UFS_BALLOC(vp,
			    lblktosize(fs, (off_t)(base + loc)),
			    fs->fs_bsize, KERNCRED, BA_METAONLY, &ibp);
			if (error) {
				goto out;
			}
			indiroff = 0;
		}
		if (I_IS_UFS1(ip)) {
			if (ffs_isblock(fs, cg_blksfree(cgp), loc))
				((ufs1_daddr_t *)(ibp->b_data))[indiroff] =
				    BLK_NOCOPY;
			else if (passno == 2 && ((ufs1_daddr_t *)(ibp->b_data))
			    [indiroff] == BLK_NOCOPY)
				((ufs1_daddr_t *)(ibp->b_data))[indiroff] = 0;
			else if (passno == 1 && ((ufs1_daddr_t *)(ibp->b_data))
			    [indiroff] == BLK_NOCOPY)
				panic("ffs_snapshot: lost indirect block");
			continue;
		}
		if (ffs_isblock(fs, cg_blksfree(cgp), loc))
			((ufs2_daddr_t *)(ibp->b_data))[indiroff] = BLK_NOCOPY;
		else if (passno == 2 &&
		    ((ufs2_daddr_t *)(ibp->b_data)) [indiroff] == BLK_NOCOPY)
			((ufs2_daddr_t *)(ibp->b_data))[indiroff] = 0;
		else if (passno == 1 &&
		    ((ufs2_daddr_t *)(ibp->b_data)) [indiroff] == BLK_NOCOPY)
			panic("ffs_snapshot: lost indirect block");
	}
	if (passno == 2)
		ibp->b_flags |= B_VALIDSUSPWRT;
	bdwrite(ibp);
out:
	/*
	 * We have to calculate the crc32c here rather than just setting the
	 * BX_CYLGRP b_xflags because the allocation of the block for the
	 * the cylinder group map will always be a full size block (fs_bsize)
	 * even though the cylinder group may be smaller (fs_cgsize). The
	 * crc32c must be computed only over fs_cgsize whereas the BX_CYLGRP
	 * flag causes it to be computed over the size of the buffer.
	 */
	if ((fs->fs_metackhash & CK_CYLGRP) != 0) {
		((struct cg *)nbp->b_data)->cg_ckhash = 0;
		((struct cg *)nbp->b_data)->cg_ckhash =
		    calculate_crc32c(~0L, nbp->b_data, fs->fs_cgsize);
	}
	return (error);
}

/*
 * Before expunging a snapshot inode, note all the
 * blocks that it claims with BLK_SNAP so that fsck will
 * be able to account for those blocks properly and so
 * that this snapshot knows that it need not copy them
 * if the other snapshot holding them is freed. This code
 * is reproduced once each for UFS1 and UFS2.
 */
static int
expunge_ufs1(snapvp, cancelip, fs, acctfunc, expungetype, clearmode)
	struct vnode *snapvp;
	struct inode *cancelip;
	struct fs *fs;
	int (*acctfunc)(struct vnode *, ufs1_daddr_t *, ufs1_daddr_t *,
	    struct fs *, ufs_lbn_t, int);
	int expungetype;
	int clearmode;
{
	int i, error, indiroff;
	ufs_lbn_t lbn, rlbn;
	ufs2_daddr_t len, blkno, numblks, blksperindir;
	struct ufs1_dinode *dip;
	struct thread *td = curthread;
	struct buf *bp;

	/*
	 * Prepare to expunge the inode. If its inode block has not
	 * yet been copied, then allocate and fill the copy.
	 */
	lbn = fragstoblks(fs, ino_to_fsba(fs, cancelip->i_number));
	blkno = 0;
	if (lbn < UFS_NDADDR) {
		blkno = VTOI(snapvp)->i_din1->di_db[lbn];
	} else {
		if (DOINGSOFTDEP(snapvp))
			softdep_prealloc(snapvp, MNT_WAIT);
		td->td_pflags |= TDP_COWINPROGRESS;
		error = ffs_balloc_ufs1(snapvp, lblktosize(fs, (off_t)lbn),
		   fs->fs_bsize, KERNCRED, BA_METAONLY, &bp);
		td->td_pflags &= ~TDP_COWINPROGRESS;
		if (error)
			return (error);
		indiroff = (lbn - UFS_NDADDR) % NINDIR(fs);
		blkno = ((ufs1_daddr_t *)(bp->b_data))[indiroff];
		bqrelse(bp);
	}
	if (blkno != 0) {
		if ((error = bread(snapvp, lbn, fs->fs_bsize, KERNCRED, &bp)))
			return (error);
	} else {
		error = ffs_balloc_ufs1(snapvp, lblktosize(fs, (off_t)lbn),
		    fs->fs_bsize, KERNCRED, 0, &bp);
		if (error)
			return (error);
		if ((error = readblock(snapvp, bp, lbn)) != 0)
			return (error);
	}
	/*
	 * Set a snapshot inode to be a zero length file, regular files
	 * or unlinked snapshots to be completely unallocated.
	 */
	dip = (struct ufs1_dinode *)bp->b_data +
	    ino_to_fsbo(fs, cancelip->i_number);
	if (clearmode || cancelip->i_effnlink == 0)
		dip->di_mode = 0;
	dip->di_size = 0;
	dip->di_blocks = 0;
	dip->di_flags &= ~SF_SNAPSHOT;
	bzero(&dip->di_db[0], (UFS_NDADDR + UFS_NIADDR) * sizeof(ufs1_daddr_t));
	bdwrite(bp);
	/*
	 * Now go through and expunge all the blocks in the file
	 * using the function requested.
	 */
	numblks = howmany(cancelip->i_size, fs->fs_bsize);
	if ((error = (*acctfunc)(snapvp, &cancelip->i_din1->di_db[0],
	    &cancelip->i_din1->di_db[UFS_NDADDR], fs, 0, expungetype)))
		return (error);
	if ((error = (*acctfunc)(snapvp, &cancelip->i_din1->di_ib[0],
	    &cancelip->i_din1->di_ib[UFS_NIADDR], fs, -1, expungetype)))
		return (error);
	blksperindir = 1;
	lbn = -UFS_NDADDR;
	len = numblks - UFS_NDADDR;
	rlbn = UFS_NDADDR;
	for (i = 0; len > 0 && i < UFS_NIADDR; i++) {
		error = indiracct_ufs1(snapvp, ITOV(cancelip), i,
		    cancelip->i_din1->di_ib[i], lbn, rlbn, len,
		    blksperindir, fs, acctfunc, expungetype);
		if (error)
			return (error);
		blksperindir *= NINDIR(fs);
		lbn -= blksperindir + 1;
		len -= blksperindir;
		rlbn += blksperindir;
	}
	return (0);
}

/*
 * Descend an indirect block chain for vnode cancelvp accounting for all
 * its indirect blocks in snapvp.
 */ 
static int
indiracct_ufs1(snapvp, cancelvp, level, blkno, lbn, rlbn, remblks,
	    blksperindir, fs, acctfunc, expungetype)
	struct vnode *snapvp;
	struct vnode *cancelvp;
	int level;
	ufs1_daddr_t blkno;
	ufs_lbn_t lbn;
	ufs_lbn_t rlbn;
	ufs_lbn_t remblks;
	ufs_lbn_t blksperindir;
	struct fs *fs;
	int (*acctfunc)(struct vnode *, ufs1_daddr_t *, ufs1_daddr_t *,
	    struct fs *, ufs_lbn_t, int);
	int expungetype;
{
	int error, num, i;
	ufs_lbn_t subblksperindir;
	struct indir indirs[UFS_NIADDR + 2];
	ufs1_daddr_t last, *bap;
	struct buf *bp;

	if (blkno == 0) {
		if (expungetype == BLK_NOCOPY)
			return (0);
		panic("indiracct_ufs1: missing indir");
	}
	if ((error = ufs_getlbns(cancelvp, rlbn, indirs, &num)) != 0)
		return (error);
	if (lbn != indirs[num - 1 - level].in_lbn || num < 2)
		panic("indiracct_ufs1: botched params");
	/*
	 * We have to expand bread here since it will deadlock looking
	 * up the block number for any blocks that are not in the cache.
	 */
	bp = getblk(cancelvp, lbn, fs->fs_bsize, 0, 0, 0);
	bp->b_blkno = fsbtodb(fs, blkno);
	if ((bp->b_flags & (B_DONE | B_DELWRI)) == 0 &&
	    (error = readblock(cancelvp, bp, fragstoblks(fs, blkno)))) {
		brelse(bp);
		return (error);
	}
	/*
	 * Account for the block pointers in this indirect block.
	 */
	last = howmany(remblks, blksperindir);
	if (last > NINDIR(fs))
		last = NINDIR(fs);
	bap = malloc(fs->fs_bsize, M_DEVBUF, M_WAITOK);
	bcopy(bp->b_data, (caddr_t)bap, fs->fs_bsize);
	bqrelse(bp);
	error = (*acctfunc)(snapvp, &bap[0], &bap[last], fs,
	    level == 0 ? rlbn : -1, expungetype);
	if (error || level == 0)
		goto out;
	/*
	 * Account for the block pointers in each of the indirect blocks
	 * in the levels below us.
	 */
	subblksperindir = blksperindir / NINDIR(fs);
	for (lbn++, level--, i = 0; i < last; i++) {
		error = indiracct_ufs1(snapvp, cancelvp, level, bap[i], lbn,
		    rlbn, remblks, subblksperindir, fs, acctfunc, expungetype);
		if (error)
			goto out;
		rlbn += blksperindir;
		lbn -= blksperindir;
		remblks -= blksperindir;
	}
out:
	free(bap, M_DEVBUF);
	return (error);
}

/*
 * Do both snap accounting and map accounting.
 */
static int
fullacct_ufs1(vp, oldblkp, lastblkp, fs, lblkno, exptype)
	struct vnode *vp;
	ufs1_daddr_t *oldblkp, *lastblkp;
	struct fs *fs;
	ufs_lbn_t lblkno;
	int exptype;	/* BLK_SNAP or BLK_NOCOPY */
{
	int error;

	if ((error = snapacct_ufs1(vp, oldblkp, lastblkp, fs, lblkno, exptype)))
		return (error);
	return (mapacct_ufs1(vp, oldblkp, lastblkp, fs, lblkno, exptype));
}

/*
 * Identify a set of blocks allocated in a snapshot inode.
 */
static int
snapacct_ufs1(vp, oldblkp, lastblkp, fs, lblkno, expungetype)
	struct vnode *vp;
	ufs1_daddr_t *oldblkp, *lastblkp;
	struct fs *fs;
	ufs_lbn_t lblkno;
	int expungetype;	/* BLK_SNAP or BLK_NOCOPY */
{
	struct inode *ip = VTOI(vp);
	ufs1_daddr_t blkno, *blkp;
	ufs_lbn_t lbn;
	struct buf *ibp;
	int error;

	for ( ; oldblkp < lastblkp; oldblkp++) {
		blkno = *oldblkp;
		if (blkno == 0 || blkno == BLK_NOCOPY || blkno == BLK_SNAP)
			continue;
		lbn = fragstoblks(fs, blkno);
		if (lbn < UFS_NDADDR) {
			blkp = &ip->i_din1->di_db[lbn];
			UFS_INODE_SET_FLAG(ip, IN_CHANGE | IN_UPDATE);
		} else {
			error = ffs_balloc_ufs1(vp, lblktosize(fs, (off_t)lbn),
			    fs->fs_bsize, KERNCRED, BA_METAONLY, &ibp);
			if (error)
				return (error);
			blkp = &((ufs1_daddr_t *)(ibp->b_data))
			    [(lbn - UFS_NDADDR) % NINDIR(fs)];
		}
		/*
		 * If we are expunging a snapshot vnode and we
		 * find a block marked BLK_NOCOPY, then it is
		 * one that has been allocated to this snapshot after
		 * we took our current snapshot and can be ignored.
		 */
		if (expungetype == BLK_SNAP && *blkp == BLK_NOCOPY) {
			if (lbn >= UFS_NDADDR)
				brelse(ibp);
		} else {
			if (*blkp != 0)
				panic("snapacct_ufs1: bad block");
			*blkp = expungetype;
			if (lbn >= UFS_NDADDR)
				bdwrite(ibp);
		}
	}
	return (0);
}

/*
 * Account for a set of blocks allocated in a snapshot inode.
 */
static int
mapacct_ufs1(vp, oldblkp, lastblkp, fs, lblkno, expungetype)
	struct vnode *vp;
	ufs1_daddr_t *oldblkp, *lastblkp;
	struct fs *fs;
	ufs_lbn_t lblkno;
	int expungetype;
{
	ufs1_daddr_t blkno;
	struct inode *ip;
	ino_t inum;
	int acctit;

	ip = VTOI(vp);
	inum = ip->i_number;
	if (lblkno == -1)
		acctit = 0;
	else
		acctit = 1;
	for ( ; oldblkp < lastblkp; oldblkp++, lblkno++) {
		blkno = *oldblkp;
		if (blkno == 0 || blkno == BLK_NOCOPY)
			continue;
		if (acctit && expungetype == BLK_SNAP && blkno != BLK_SNAP)
			*ip->i_snapblklist++ = lblkno;
		if (blkno == BLK_SNAP)
			blkno = blkstofrags(fs, lblkno);
		ffs_blkfree(ITOUMP(ip), fs, vp, blkno, fs->fs_bsize, inum,
		    vp->v_type, NULL, SINGLETON_KEY);
	}
	return (0);
}

/*
 * Before expunging a snapshot inode, note all the
 * blocks that it claims with BLK_SNAP so that fsck will
 * be able to account for those blocks properly and so
 * that this snapshot knows that it need not copy them
 * if the other snapshot holding them is freed. This code
 * is reproduced once each for UFS1 and UFS2.
 */
static int
expunge_ufs2(snapvp, cancelip, fs, acctfunc, expungetype, clearmode)
	struct vnode *snapvp;
	struct inode *cancelip;
	struct fs *fs;
	int (*acctfunc)(struct vnode *, ufs2_daddr_t *, ufs2_daddr_t *,
	    struct fs *, ufs_lbn_t, int);
	int expungetype;
	int clearmode;
{
	int i, error, indiroff;
	ufs_lbn_t lbn, rlbn;
	ufs2_daddr_t len, blkno, numblks, blksperindir;
	struct ufs2_dinode *dip;
	struct thread *td = curthread;
	struct buf *bp;

	/*
	 * Prepare to expunge the inode. If its inode block has not
	 * yet been copied, then allocate and fill the copy.
	 */
	lbn = fragstoblks(fs, ino_to_fsba(fs, cancelip->i_number));
	blkno = 0;
	if (lbn < UFS_NDADDR) {
		blkno = VTOI(snapvp)->i_din2->di_db[lbn];
	} else {
		if (DOINGSOFTDEP(snapvp))
			softdep_prealloc(snapvp, MNT_WAIT);
		td->td_pflags |= TDP_COWINPROGRESS;
		error = ffs_balloc_ufs2(snapvp, lblktosize(fs, (off_t)lbn),
		   fs->fs_bsize, KERNCRED, BA_METAONLY, &bp);
		td->td_pflags &= ~TDP_COWINPROGRESS;
		if (error)
			return (error);
		indiroff = (lbn - UFS_NDADDR) % NINDIR(fs);
		blkno = ((ufs2_daddr_t *)(bp->b_data))[indiroff];
		bqrelse(bp);
	}
	if (blkno != 0) {
		if ((error = bread(snapvp, lbn, fs->fs_bsize, KERNCRED, &bp)))
			return (error);
	} else {
		error = ffs_balloc_ufs2(snapvp, lblktosize(fs, (off_t)lbn),
		    fs->fs_bsize, KERNCRED, 0, &bp);
		if (error)
			return (error);
		if ((error = readblock(snapvp, bp, lbn)) != 0)
			return (error);
	}
	/*
	 * Set a snapshot inode to be a zero length file, regular files
	 * to be completely unallocated.
	 */
	dip = (struct ufs2_dinode *)bp->b_data +
	    ino_to_fsbo(fs, cancelip->i_number);
	dip->di_size = 0;
	dip->di_blocks = 0;
	dip->di_flags &= ~SF_SNAPSHOT;
	bzero(&dip->di_db[0], (UFS_NDADDR + UFS_NIADDR) * sizeof(ufs2_daddr_t));
	if (clearmode || cancelip->i_effnlink == 0)
		dip->di_mode = 0;
	else
		ffs_update_dinode_ckhash(fs, dip);
	bdwrite(bp);
	/*
	 * Now go through and expunge all the blocks in the file
	 * using the function requested.
	 */
	numblks = howmany(cancelip->i_size, fs->fs_bsize);
	if ((error = (*acctfunc)(snapvp, &cancelip->i_din2->di_db[0],
	    &cancelip->i_din2->di_db[UFS_NDADDR], fs, 0, expungetype)))
		return (error);
	if ((error = (*acctfunc)(snapvp, &cancelip->i_din2->di_ib[0],
	    &cancelip->i_din2->di_ib[UFS_NIADDR], fs, -1, expungetype)))
		return (error);
	blksperindir = 1;
	lbn = -UFS_NDADDR;
	len = numblks - UFS_NDADDR;
	rlbn = UFS_NDADDR;
	for (i = 0; len > 0 && i < UFS_NIADDR; i++) {
		error = indiracct_ufs2(snapvp, ITOV(cancelip), i,
		    cancelip->i_din2->di_ib[i], lbn, rlbn, len,
		    blksperindir, fs, acctfunc, expungetype);
		if (error)
			return (error);
		blksperindir *= NINDIR(fs);
		lbn -= blksperindir + 1;
		len -= blksperindir;
		rlbn += blksperindir;
	}
	return (0);
}

/*
 * Descend an indirect block chain for vnode cancelvp accounting for all
 * its indirect blocks in snapvp.
 */ 
static int
indiracct_ufs2(snapvp, cancelvp, level, blkno, lbn, rlbn, remblks,
	    blksperindir, fs, acctfunc, expungetype)
	struct vnode *snapvp;
	struct vnode *cancelvp;
	int level;
	ufs2_daddr_t blkno;
	ufs_lbn_t lbn;
	ufs_lbn_t rlbn;
	ufs_lbn_t remblks;
	ufs_lbn_t blksperindir;
	struct fs *fs;
	int (*acctfunc)(struct vnode *, ufs2_daddr_t *, ufs2_daddr_t *,
	    struct fs *, ufs_lbn_t, int);
	int expungetype;
{
	int error, num, i;
	ufs_lbn_t subblksperindir;
	struct indir indirs[UFS_NIADDR + 2];
	ufs2_daddr_t last, *bap;
	struct buf *bp;

	if (blkno == 0) {
		if (expungetype == BLK_NOCOPY)
			return (0);
		panic("indiracct_ufs2: missing indir");
	}
	if ((error = ufs_getlbns(cancelvp, rlbn, indirs, &num)) != 0)
		return (error);
	if (lbn != indirs[num - 1 - level].in_lbn || num < 2)
		panic("indiracct_ufs2: botched params");
	/*
	 * We have to expand bread here since it will deadlock looking
	 * up the block number for any blocks that are not in the cache.
	 */
	bp = getblk(cancelvp, lbn, fs->fs_bsize, 0, 0, 0);
	bp->b_blkno = fsbtodb(fs, blkno);
	if ((bp->b_flags & B_CACHE) == 0 &&
	    (error = readblock(cancelvp, bp, fragstoblks(fs, blkno)))) {
		brelse(bp);
		return (error);
	}
	/*
	 * Account for the block pointers in this indirect block.
	 */
	last = howmany(remblks, blksperindir);
	if (last > NINDIR(fs))
		last = NINDIR(fs);
	bap = malloc(fs->fs_bsize, M_DEVBUF, M_WAITOK);
	bcopy(bp->b_data, (caddr_t)bap, fs->fs_bsize);
	bqrelse(bp);
	error = (*acctfunc)(snapvp, &bap[0], &bap[last], fs,
	    level == 0 ? rlbn : -1, expungetype);
	if (error || level == 0)
		goto out;
	/*
	 * Account for the block pointers in each of the indirect blocks
	 * in the levels below us.
	 */
	subblksperindir = blksperindir / NINDIR(fs);
	for (lbn++, level--, i = 0; i < last; i++) {
		error = indiracct_ufs2(snapvp, cancelvp, level, bap[i], lbn,
		    rlbn, remblks, subblksperindir, fs, acctfunc, expungetype);
		if (error)
			goto out;
		rlbn += blksperindir;
		lbn -= blksperindir;
		remblks -= blksperindir;
	}
out:
	free(bap, M_DEVBUF);
	return (error);
}

/*
 * Do both snap accounting and map accounting.
 */
static int
fullacct_ufs2(vp, oldblkp, lastblkp, fs, lblkno, exptype)
	struct vnode *vp;
	ufs2_daddr_t *oldblkp, *lastblkp;
	struct fs *fs;
	ufs_lbn_t lblkno;
	int exptype;	/* BLK_SNAP or BLK_NOCOPY */
{
	int error;

	if ((error = snapacct_ufs2(vp, oldblkp, lastblkp, fs, lblkno, exptype)))
		return (error);
	return (mapacct_ufs2(vp, oldblkp, lastblkp, fs, lblkno, exptype));
}

/*
 * Identify a set of blocks allocated in a snapshot inode.
 */
static int
snapacct_ufs2(vp, oldblkp, lastblkp, fs, lblkno, expungetype)
	struct vnode *vp;
	ufs2_daddr_t *oldblkp, *lastblkp;
	struct fs *fs;
	ufs_lbn_t lblkno;
	int expungetype;	/* BLK_SNAP or BLK_NOCOPY */
{
	struct inode *ip = VTOI(vp);
	ufs2_daddr_t blkno, *blkp;
	ufs_lbn_t lbn;
	struct buf *ibp;
	int error;

	for ( ; oldblkp < lastblkp; oldblkp++) {
		blkno = *oldblkp;
		if (blkno == 0 || blkno == BLK_NOCOPY || blkno == BLK_SNAP)
			continue;
		lbn = fragstoblks(fs, blkno);
		if (lbn < UFS_NDADDR) {
			blkp = &ip->i_din2->di_db[lbn];
			UFS_INODE_SET_FLAG(ip, IN_CHANGE | IN_UPDATE);
		} else {
			error = ffs_balloc_ufs2(vp, lblktosize(fs, (off_t)lbn),
			    fs->fs_bsize, KERNCRED, BA_METAONLY, &ibp);
			if (error)
				return (error);
			blkp = &((ufs2_daddr_t *)(ibp->b_data))
			    [(lbn - UFS_NDADDR) % NINDIR(fs)];
		}
		/*
		 * If we are expunging a snapshot vnode and we
		 * find a block marked BLK_NOCOPY, then it is
		 * one that has been allocated to this snapshot after
		 * we took our current snapshot and can be ignored.
		 */
		if (expungetype == BLK_SNAP && *blkp == BLK_NOCOPY) {
			if (lbn >= UFS_NDADDR)
				brelse(ibp);
		} else {
			if (*blkp != 0)
				panic("snapacct_ufs2: bad block");
			*blkp = expungetype;
			if (lbn >= UFS_NDADDR)
				bdwrite(ibp);
		}
	}
	return (0);
}

/*
 * Account for a set of blocks allocated in a snapshot inode.
 */
static int
mapacct_ufs2(vp, oldblkp, lastblkp, fs, lblkno, expungetype)
	struct vnode *vp;
	ufs2_daddr_t *oldblkp, *lastblkp;
	struct fs *fs;
	ufs_lbn_t lblkno;
	int expungetype;
{
	ufs2_daddr_t blkno;
	struct inode *ip;
	ino_t inum;
	int acctit;

	ip = VTOI(vp);
	inum = ip->i_number;
	if (lblkno == -1)
		acctit = 0;
	else
		acctit = 1;
	for ( ; oldblkp < lastblkp; oldblkp++, lblkno++) {
		blkno = *oldblkp;
		if (blkno == 0 || blkno == BLK_NOCOPY)
			continue;
		if (acctit && expungetype == BLK_SNAP && blkno != BLK_SNAP)
			*ip->i_snapblklist++ = lblkno;
		if (blkno == BLK_SNAP)
			blkno = blkstofrags(fs, lblkno);
		ffs_blkfree(ITOUMP(ip), fs, vp, blkno, fs->fs_bsize, inum,
		    vp->v_type, NULL, SINGLETON_KEY);
	}
	return (0);
}

/*
 * Decrement extra reference on snapshot when last name is removed.
 * It will not be freed until the last open reference goes away.
 */
void
ffs_snapgone(ip)
	struct inode *ip;
{
	struct inode *xp;
	struct fs *fs;
	int snaploc;
	struct snapdata *sn;
	struct ufsmount *ump;

	/*
	 * Find snapshot in incore list.
	 */
	xp = NULL;
	sn = ITODEVVP(ip)->v_rdev->si_snapdata;
	if (sn != NULL)
		TAILQ_FOREACH(xp, &sn->sn_head, i_nextsnap)
			if (xp == ip)
				break;
	if (xp != NULL)
		vrele(ITOV(ip));
#ifdef DIAGNOSTIC
	else if (snapdebug)
		printf("ffs_snapgone: lost snapshot vnode %ju\n",
		    (uintmax_t)ip->i_number);
#endif
	/*
	 * Delete snapshot inode from superblock. Keep list dense.
	 */
	ump = ITOUMP(ip);
	fs = ump->um_fs;
	UFS_LOCK(ump);
	for (snaploc = 0; snaploc < FSMAXSNAP; snaploc++)
		if (fs->fs_snapinum[snaploc] == ip->i_number)
			break;
	if (snaploc < FSMAXSNAP) {
		for (snaploc++; snaploc < FSMAXSNAP; snaploc++) {
			if (fs->fs_snapinum[snaploc] == 0)
				break;
			fs->fs_snapinum[snaploc - 1] = fs->fs_snapinum[snaploc];
		}
		fs->fs_snapinum[snaploc - 1] = 0;
	}
	UFS_UNLOCK(ump);
}

/*
 * Prepare a snapshot file for being removed.
 */
void
ffs_snapremove(vp)
	struct vnode *vp;
{
	struct inode *ip;
	struct vnode *devvp;
	struct buf *ibp;
	struct fs *fs;
	ufs2_daddr_t numblks, blkno, dblk;
	int error, i, last, loc;
	struct snapdata *sn;

	ip = VTOI(vp);
	fs = ITOFS(ip);
	devvp = ITODEVVP(ip);
	/*
	 * If active, delete from incore list (this snapshot may
	 * already have been in the process of being deleted, so
	 * would not have been active).
	 *
	 * Clear copy-on-write flag if last snapshot.
	 */
	VI_LOCK(devvp);
	if (ip->i_nextsnap.tqe_prev != 0) {
		sn = devvp->v_rdev->si_snapdata;
		TAILQ_REMOVE(&sn->sn_head, ip, i_nextsnap);
		ip->i_nextsnap.tqe_prev = 0;
		VI_UNLOCK(devvp);
		lockmgr(&vp->v_lock, LK_EXCLUSIVE, NULL);
		for (i = 0; i < sn->sn_lock.lk_recurse; i++)
			lockmgr(&vp->v_lock, LK_EXCLUSIVE, NULL);
		KASSERT(vp->v_vnlock == &sn->sn_lock,
			("ffs_snapremove: lost lock mutation")); 
		vp->v_vnlock = &vp->v_lock;
		VI_LOCK(devvp);
		while (sn->sn_lock.lk_recurse > 0)
			lockmgr(&sn->sn_lock, LK_RELEASE, NULL);
		lockmgr(&sn->sn_lock, LK_RELEASE, NULL);
		try_free_snapdata(devvp);
	} else
		VI_UNLOCK(devvp);
	/*
	 * Clear all BLK_NOCOPY fields. Pass any block claims to other
	 * snapshots that want them (see ffs_snapblkfree below).
	 */
	for (blkno = 1; blkno < UFS_NDADDR; blkno++) {
		dblk = DIP(ip, i_db[blkno]);
		if (dblk == 0)
			continue;
		if (dblk == BLK_NOCOPY || dblk == BLK_SNAP)
			DIP_SET(ip, i_db[blkno], 0);
		else if ((dblk == blkstofrags(fs, blkno) &&
		     ffs_snapblkfree(fs, ITODEVVP(ip), dblk, fs->fs_bsize,
		     ip->i_number, vp->v_type, NULL))) {
			DIP_SET(ip, i_blocks, DIP(ip, i_blocks) -
			    btodb(fs->fs_bsize));
			DIP_SET(ip, i_db[blkno], 0);
		}
	}
	numblks = howmany(ip->i_size, fs->fs_bsize);
	for (blkno = UFS_NDADDR; blkno < numblks; blkno += NINDIR(fs)) {
		error = UFS_BALLOC(vp, lblktosize(fs, (off_t)blkno),
		    fs->fs_bsize, KERNCRED, BA_METAONLY, &ibp);
		if (error)
			continue;
		if (fs->fs_size - blkno > NINDIR(fs))
			last = NINDIR(fs);
		else
			last = fs->fs_size - blkno;
		for (loc = 0; loc < last; loc++) {
			if (I_IS_UFS1(ip)) {
				dblk = ((ufs1_daddr_t *)(ibp->b_data))[loc];
				if (dblk == 0)
					continue;
				if (dblk == BLK_NOCOPY || dblk == BLK_SNAP)
					((ufs1_daddr_t *)(ibp->b_data))[loc]= 0;
				else if ((dblk == blkstofrags(fs, blkno) &&
				     ffs_snapblkfree(fs, ITODEVVP(ip), dblk,
				     fs->fs_bsize, ip->i_number, vp->v_type,
				     NULL))) {
					ip->i_din1->di_blocks -=
					    btodb(fs->fs_bsize);
					((ufs1_daddr_t *)(ibp->b_data))[loc]= 0;
				}
				continue;
			}
			dblk = ((ufs2_daddr_t *)(ibp->b_data))[loc];
			if (dblk == 0)
				continue;
			if (dblk == BLK_NOCOPY || dblk == BLK_SNAP)
				((ufs2_daddr_t *)(ibp->b_data))[loc] = 0;
			else if ((dblk == blkstofrags(fs, blkno) &&
			     ffs_snapblkfree(fs, ITODEVVP(ip), dblk,
			     fs->fs_bsize, ip->i_number, vp->v_type, NULL))) {
				ip->i_din2->di_blocks -= btodb(fs->fs_bsize);
				((ufs2_daddr_t *)(ibp->b_data))[loc] = 0;
			}
		}
		bawrite(ibp);
	}
	/*
	 * Clear snapshot flag and drop reference.
	 */
	ip->i_flags &= ~SF_SNAPSHOT;
	DIP_SET(ip, i_flags, ip->i_flags);
	UFS_INODE_SET_FLAG(ip, IN_CHANGE | IN_UPDATE);
	/*
	 * The dirtied indirects must be written out before
	 * softdep_setup_freeblocks() is called.  Otherwise indir_trunc()
	 * may find indirect pointers using the magic BLK_* values.
	 */
	if (DOINGSOFTDEP(vp))
		ffs_syncvnode(vp, MNT_WAIT, 0);
#ifdef QUOTA
	/*
	 * Reenable disk quotas for ex-snapshot file.
	 */
	if (!getinoquota(ip))
		(void) chkdq(ip, DIP(ip, i_blocks), KERNCRED, FORCE);
#endif
}

/*
 * Notification that a block is being freed. Return zero if the free
 * should be allowed to proceed. Return non-zero if the snapshot file
 * wants to claim the block. The block will be claimed if it is an
 * uncopied part of one of the snapshots. It will be freed if it is
 * either a BLK_NOCOPY or has already been copied in all of the snapshots.
 * If a fragment is being freed, then all snapshots that care about
 * it must make a copy since a snapshot file can only claim full sized
 * blocks. Note that if more than one snapshot file maps the block,
 * we can pick one at random to claim it. Since none of the snapshots
 * can change, we are assurred that they will all see the same unmodified
 * image. When deleting a snapshot file (see ffs_snapremove above), we
 * must push any of these claimed blocks to one of the other snapshots
 * that maps it. These claimed blocks are easily identified as they will
 * have a block number equal to their logical block number within the
 * snapshot. A copied block can never have this property because they
 * must always have been allocated from a BLK_NOCOPY location.
 */
int
ffs_snapblkfree(fs, devvp, bno, size, inum, vtype, wkhd)
	struct fs *fs;
	struct vnode *devvp;
	ufs2_daddr_t bno;
	long size;
	ino_t inum;
	enum vtype vtype;
	struct workhead *wkhd;
{
	struct buf *ibp, *cbp, *savedcbp = NULL;
	struct thread *td = curthread;
	struct inode *ip;
	struct vnode *vp = NULL;
	ufs_lbn_t lbn;
	ufs2_daddr_t blkno;
	int indiroff = 0, error = 0, claimedblk = 0;
	struct snapdata *sn;

	lbn = fragstoblks(fs, bno);
retry:
	VI_LOCK(devvp);
	sn = devvp->v_rdev->si_snapdata;
	if (sn == NULL) {
		VI_UNLOCK(devvp);
		return (0);
	}
	if (lockmgr(&sn->sn_lock, LK_INTERLOCK | LK_EXCLUSIVE | LK_SLEEPFAIL,
	    VI_MTX(devvp)) != 0)
		goto retry;
	TAILQ_FOREACH(ip, &sn->sn_head, i_nextsnap) {
		vp = ITOV(ip);
		if (DOINGSOFTDEP(vp))
			softdep_prealloc(vp, MNT_WAIT);
		/*
		 * Lookup block being written.
		 */
		if (lbn < UFS_NDADDR) {
			blkno = DIP(ip, i_db[lbn]);
		} else {
			td->td_pflags |= TDP_COWINPROGRESS;
			error = UFS_BALLOC(vp, lblktosize(fs, (off_t)lbn),
			    fs->fs_bsize, KERNCRED, BA_METAONLY, &ibp);
			td->td_pflags &= ~TDP_COWINPROGRESS;
			if (error)
				break;
			indiroff = (lbn - UFS_NDADDR) % NINDIR(fs);
			if (I_IS_UFS1(ip))
				blkno=((ufs1_daddr_t *)(ibp->b_data))[indiroff];
			else
				blkno=((ufs2_daddr_t *)(ibp->b_data))[indiroff];
		}
		/*
		 * Check to see if block needs to be copied.
		 */
		if (blkno == 0) {
			/*
			 * A block that we map is being freed. If it has not
			 * been claimed yet, we will claim or copy it (below).
			 */
			claimedblk = 1;
		} else if (blkno == BLK_SNAP) {
			/*
			 * No previous snapshot claimed the block,
			 * so it will be freed and become a BLK_NOCOPY
			 * (don't care) for us.
			 */
			if (claimedblk)
				panic("snapblkfree: inconsistent block type");
			if (lbn < UFS_NDADDR) {
				DIP_SET(ip, i_db[lbn], BLK_NOCOPY);
				UFS_INODE_SET_FLAG(ip, IN_CHANGE | IN_UPDATE);
			} else if (I_IS_UFS1(ip)) {
				((ufs1_daddr_t *)(ibp->b_data))[indiroff] =
				    BLK_NOCOPY;
				bdwrite(ibp);
			} else {
				((ufs2_daddr_t *)(ibp->b_data))[indiroff] =
				    BLK_NOCOPY;
				bdwrite(ibp);
			}
			continue;
		} else /* BLK_NOCOPY or default */ {
			/*
			 * If the snapshot has already copied the block
			 * (default), or does not care about the block,
			 * it is not needed.
			 */
			if (lbn >= UFS_NDADDR)
				bqrelse(ibp);
			continue;
		}
		/*
		 * If this is a full size block, we will just grab it
		 * and assign it to the snapshot inode. Otherwise we
		 * will proceed to copy it. See explanation for this
		 * routine as to why only a single snapshot needs to
		 * claim this block.
		 */
		if (size == fs->fs_bsize) {
#ifdef DIAGNOSTIC
			if (snapdebug)
				printf("%s %ju lbn %jd from inum %ju\n",
				    "Grabonremove: snapino",
				    (uintmax_t)ip->i_number,
				    (intmax_t)lbn, (uintmax_t)inum);
#endif
			/*
			 * If journaling is tracking this write we must add
			 * the work to the inode or indirect being written.
			 */
			if (wkhd != NULL) {
				if (lbn < UFS_NDADDR)
					softdep_inode_append(ip,
					    curthread->td_ucred, wkhd);
				else
					softdep_buf_append(ibp, wkhd);
			}
			if (lbn < UFS_NDADDR) {
				DIP_SET(ip, i_db[lbn], bno);
			} else if (I_IS_UFS1(ip)) {
				((ufs1_daddr_t *)(ibp->b_data))[indiroff] = bno;
				bdwrite(ibp);
			} else {
				((ufs2_daddr_t *)(ibp->b_data))[indiroff] = bno;
				bdwrite(ibp);
			}
			DIP_SET(ip, i_blocks, DIP(ip, i_blocks) + btodb(size));
			UFS_INODE_SET_FLAG(ip, IN_CHANGE | IN_UPDATE);
			lockmgr(vp->v_vnlock, LK_RELEASE, NULL);
			return (1);
		}
		if (lbn >= UFS_NDADDR)
			bqrelse(ibp);
		/*
		 * Allocate the block into which to do the copy. Note that this
		 * allocation will never require any additional allocations for
		 * the snapshot inode.
		 */
		td->td_pflags |= TDP_COWINPROGRESS;
		error = UFS_BALLOC(vp, lblktosize(fs, (off_t)lbn),
		    fs->fs_bsize, KERNCRED, 0, &cbp);
		td->td_pflags &= ~TDP_COWINPROGRESS;
		if (error)
			break;
#ifdef DIAGNOSTIC
		if (snapdebug)
			printf("%s%ju lbn %jd %s %ju size %ld to blkno %jd\n",
			    "Copyonremove: snapino ", (uintmax_t)ip->i_number,
			    (intmax_t)lbn, "for inum", (uintmax_t)inum, size,
			    (intmax_t)cbp->b_blkno);
#endif
		/*
		 * If we have already read the old block contents, then
		 * simply copy them to the new block. Note that we need
		 * to synchronously write snapshots that have not been
		 * unlinked, and hence will be visible after a crash,
		 * to ensure their integrity. At a minimum we ensure the
		 * integrity of the filesystem metadata, but use the
		 * dopersistence sysctl-setable flag to decide on the
		 * persistence needed for file content data.
		 */
		if (savedcbp != NULL) {
			bcopy(savedcbp->b_data, cbp->b_data, fs->fs_bsize);
			bawrite(cbp);
			if ((vtype == VDIR || dopersistence) &&
			    ip->i_effnlink > 0)
				(void) ffs_syncvnode(vp, MNT_WAIT, NO_INO_UPDT);
			continue;
		}
		/*
		 * Otherwise, read the old block contents into the buffer.
		 */
		if ((error = readblock(vp, cbp, lbn)) != 0) {
			bzero(cbp->b_data, fs->fs_bsize);
			bawrite(cbp);
			if ((vtype == VDIR || dopersistence) &&
			    ip->i_effnlink > 0)
				(void) ffs_syncvnode(vp, MNT_WAIT, NO_INO_UPDT);
			break;
		}
		savedcbp = cbp;
	}
	/*
	 * Note that we need to synchronously write snapshots that
	 * have not been unlinked, and hence will be visible after
	 * a crash, to ensure their integrity. At a minimum we
	 * ensure the integrity of the filesystem metadata, but
	 * use the dopersistence sysctl-setable flag to decide on
	 * the persistence needed for file content data.
	 */
	if (savedcbp) {
		vp = savedcbp->b_vp;
		bawrite(savedcbp);
		if ((vtype == VDIR || dopersistence) &&
		    VTOI(vp)->i_effnlink > 0)
			(void) ffs_syncvnode(vp, MNT_WAIT, NO_INO_UPDT);
	}
	/*
	 * If we have been unable to allocate a block in which to do
	 * the copy, then return non-zero so that the fragment will
	 * not be freed. Although space will be lost, the snapshot
	 * will stay consistent.
	 */
	if (error != 0 && wkhd != NULL)
		softdep_freework(wkhd);
	lockmgr(&sn->sn_lock, LK_RELEASE, NULL);
	return (error);
}

/*
 * Associate snapshot files when mounting.
 */
void
ffs_snapshot_mount(mp)
	struct mount *mp;
{
	struct ufsmount *ump = VFSTOUFS(mp);
	struct vnode *devvp = ump->um_devvp;
	struct fs *fs = ump->um_fs;
	struct thread *td = curthread;
	struct snapdata *sn;
	struct vnode *vp;
	struct vnode *lastvp;
	struct inode *ip;
	struct uio auio;
	struct iovec aiov;
	void *snapblklist;
	char *reason;
	daddr_t snaplistsize;
	int error, snaploc, loc;

	/*
	 * XXX The following needs to be set before ffs_truncate or
	 * VOP_READ can be called.
	 */
	mp->mnt_stat.f_iosize = fs->fs_bsize;
	/*
	 * Process each snapshot listed in the superblock.
	 */
	vp = NULL;
	lastvp = NULL;
	sn = NULL;
	for (snaploc = 0; snaploc < FSMAXSNAP; snaploc++) {
		if (fs->fs_snapinum[snaploc] == 0)
			break;
		if ((error = ffs_vget(mp, fs->fs_snapinum[snaploc],
		    LK_EXCLUSIVE, &vp)) != 0){
			printf("ffs_snapshot_mount: vget failed %d\n", error);
			continue;
		}
		ip = VTOI(vp);
		if (vp->v_type != VREG) {
			reason = "non-file snapshot";
		} else if (!IS_SNAPSHOT(ip)) {
			reason = "non-snapshot";
		} else if (ip->i_size ==
		    lblktosize(fs, howmany(fs->fs_size, fs->fs_frag))) {
			reason = "old format snapshot";
			(void)ffs_truncate(vp, (off_t)0, 0, NOCRED);
			(void)ffs_syncvnode(vp, MNT_WAIT, 0);
		} else {
			reason = NULL;
		}
		if (reason != NULL) {
			printf("ffs_snapshot_mount: %s inode %d\n",
			    reason, fs->fs_snapinum[snaploc]);
			vput(vp);
			vp = NULL;
			for (loc = snaploc + 1; loc < FSMAXSNAP; loc++) {
				if (fs->fs_snapinum[loc] == 0)
					break;
				fs->fs_snapinum[loc - 1] = fs->fs_snapinum[loc];
			}
			fs->fs_snapinum[loc - 1] = 0;
			snaploc--;
			continue;
		}
		/*
		 * Acquire a lock on the snapdata structure, creating it if
		 * necessary.
		 */
		sn = ffs_snapdata_acquire(devvp);
		/* 
		 * Change vnode to use shared snapshot lock instead of the
		 * original private lock.
		 */
		vp->v_vnlock = &sn->sn_lock;
		lockmgr(&vp->v_lock, LK_RELEASE, NULL);
		/*
		 * Link it onto the active snapshot list.
		 */
		VI_LOCK(devvp);
		if (ip->i_nextsnap.tqe_prev != 0)
			panic("ffs_snapshot_mount: %ju already on list",
			    (uintmax_t)ip->i_number);
		else
			TAILQ_INSERT_TAIL(&sn->sn_head, ip, i_nextsnap);
		vp->v_vflag |= VV_SYSTEM;
		VI_UNLOCK(devvp);
		VOP_UNLOCK(vp);
		lastvp = vp;
	}
	vp = lastvp;
	/*
	 * No usable snapshots found.
	 */
	if (sn == NULL || vp == NULL)
		return;
	/*
	 * Allocate the space for the block hints list. We always want to
	 * use the list from the newest snapshot.
	 */
	auio.uio_iov = &aiov;
	auio.uio_iovcnt = 1;
	aiov.iov_base = (void *)&snaplistsize;
	aiov.iov_len = sizeof(snaplistsize);
	auio.uio_resid = aiov.iov_len;
	auio.uio_offset =
	    lblktosize(fs, howmany(fs->fs_size, fs->fs_frag));
	auio.uio_segflg = UIO_SYSSPACE;
	auio.uio_rw = UIO_READ;
	auio.uio_td = td;
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
	if ((error = VOP_READ(vp, &auio, IO_UNIT, td->td_ucred)) != 0) {
		printf("ffs_snapshot_mount: read_1 failed %d\n", error);
		VOP_UNLOCK(vp);
		return;
	}
	snapblklist = malloc(snaplistsize * sizeof(daddr_t),
	    M_UFSMNT, M_WAITOK);
	auio.uio_iovcnt = 1;
	aiov.iov_base = snapblklist;
	aiov.iov_len = snaplistsize * sizeof (daddr_t);
	auio.uio_resid = aiov.iov_len;
	auio.uio_offset -= sizeof(snaplistsize);
	if ((error = VOP_READ(vp, &auio, IO_UNIT, td->td_ucred)) != 0) {
		printf("ffs_snapshot_mount: read_2 failed %d\n", error);
		VOP_UNLOCK(vp);
		free(snapblklist, M_UFSMNT);
		return;
	}
	VOP_UNLOCK(vp);
	VI_LOCK(devvp);
	ASSERT_VOP_LOCKED(devvp, "ffs_snapshot_mount");
	sn->sn_listsize = snaplistsize;
	sn->sn_blklist = (daddr_t *)snapblklist;
	devvp->v_vflag |= VV_COPYONWRITE;
	VI_UNLOCK(devvp);
}

/*
 * Disassociate snapshot files when unmounting.
 */
void
ffs_snapshot_unmount(mp)
	struct mount *mp;
{
	struct vnode *devvp = VFSTOUFS(mp)->um_devvp;
	struct snapdata *sn;
	struct inode *xp;
	struct vnode *vp;

	VI_LOCK(devvp);
	sn = devvp->v_rdev->si_snapdata;
	while (sn != NULL && (xp = TAILQ_FIRST(&sn->sn_head)) != NULL) {
		vp = ITOV(xp);
		TAILQ_REMOVE(&sn->sn_head, xp, i_nextsnap);
		xp->i_nextsnap.tqe_prev = 0;
		lockmgr(&sn->sn_lock, LK_INTERLOCK | LK_EXCLUSIVE,
		    VI_MTX(devvp));
		/*
		 * Avoid LOR with above snapshot lock. The LK_NOWAIT should
		 * never fail as the lock is currently unused. Rather than
		 * panic, we recover by doing the blocking lock.
		 */
		if (lockmgr(&vp->v_lock, LK_EXCLUSIVE | LK_NOWAIT, NULL) != 0) {
			printf("ffs_snapshot_unmount: Unexpected LK_NOWAIT "
			    "failure\n");
			lockmgr(&vp->v_lock, LK_EXCLUSIVE, NULL);
		}
		KASSERT(vp->v_vnlock == &sn->sn_lock,
		("ffs_snapshot_unmount: lost lock mutation")); 
		vp->v_vnlock = &vp->v_lock;
		lockmgr(&vp->v_lock, LK_RELEASE, NULL);
		lockmgr(&sn->sn_lock, LK_RELEASE, NULL);
		if (xp->i_effnlink > 0)
			vrele(vp);
		VI_LOCK(devvp);
		sn = devvp->v_rdev->si_snapdata;
	}
	try_free_snapdata(devvp);
	ASSERT_VOP_LOCKED(devvp, "ffs_snapshot_unmount");
}

/*
 * Check the buffer block to be belong to device buffer that shall be
 * locked after snaplk. devvp shall be locked on entry, and will be
 * leaved locked upon exit.
 */
static int
ffs_bp_snapblk(devvp, bp)
	struct vnode *devvp;
	struct buf *bp;
{
	struct snapdata *sn;
	struct fs *fs;
	ufs2_daddr_t lbn, *snapblklist;
	int lower, upper, mid;

	ASSERT_VI_LOCKED(devvp, "ffs_bp_snapblk");
	KASSERT(devvp->v_type == VCHR, ("Not a device %p", devvp));
	sn = devvp->v_rdev->si_snapdata;
	if (sn == NULL || TAILQ_FIRST(&sn->sn_head) == NULL)
		return (0);
	fs = ITOFS(TAILQ_FIRST(&sn->sn_head));
	lbn = fragstoblks(fs, dbtofsb(fs, bp->b_blkno));
	snapblklist = sn->sn_blklist;
	upper = sn->sn_listsize - 1;
	lower = 1;
	while (lower <= upper) {
		mid = (lower + upper) / 2;
		if (snapblklist[mid] == lbn)
			break;
		if (snapblklist[mid] < lbn)
			lower = mid + 1;
		else
			upper = mid - 1;
	}
	if (lower <= upper)
		return (1);
	return (0);
}

void
ffs_bdflush(bo, bp)
	struct bufobj *bo;
	struct buf *bp;
{
	struct thread *td;
	struct vnode *vp, *devvp;
	struct buf *nbp;
	int bp_bdskip;

	if (bo->bo_dirty.bv_cnt <= dirtybufthresh)
		return;

	td = curthread;
	vp = bp->b_vp;
	devvp = bo2vnode(bo);
	KASSERT(vp == devvp, ("devvp != vp %p %p", bo, bp));

	VI_LOCK(devvp);
	bp_bdskip = ffs_bp_snapblk(devvp, bp);
	if (bp_bdskip)
		bdwriteskip++;
	VI_UNLOCK(devvp);
	if (bo->bo_dirty.bv_cnt > dirtybufthresh + 10 && !bp_bdskip) {
		(void) VOP_FSYNC(vp, MNT_NOWAIT, td);
		altbufferflushes++;
	} else {
		BO_LOCK(bo);
		/*
		 * Try to find a buffer to flush.
		 */
		TAILQ_FOREACH(nbp, &bo->bo_dirty.bv_hd, b_bobufs) {
			if ((nbp->b_vflags & BV_BKGRDINPROG) ||
			    BUF_LOCK(nbp,
				     LK_EXCLUSIVE | LK_NOWAIT, NULL))
				continue;
			if (bp == nbp)
				panic("bdwrite: found ourselves");
			BO_UNLOCK(bo);
			/*
			 * Don't countdeps with the bo lock
			 * held.
			 */
			if (buf_countdeps(nbp, 0)) {
				BO_LOCK(bo);
				BUF_UNLOCK(nbp);
				continue;
			}
			if (bp_bdskip) {
				VI_LOCK(devvp);
				if (!ffs_bp_snapblk(vp, nbp)) {
					VI_UNLOCK(devvp);
					BO_LOCK(bo);
					BUF_UNLOCK(nbp);
					continue;
				}
				VI_UNLOCK(devvp);
			}
			if (nbp->b_flags & B_CLUSTEROK) {
				vfs_bio_awrite(nbp);
			} else {
				bremfree(nbp);
				bawrite(nbp);
			}
			dirtybufferflushes++;
			break;
		}
		if (nbp == NULL)
			BO_UNLOCK(bo);
	}
}

/*
 * Check for need to copy block that is about to be written,
 * copying the block if necessary.
 */
int
ffs_copyonwrite(devvp, bp)
	struct vnode *devvp;
	struct buf *bp;
{
	struct snapdata *sn;
	struct buf *ibp, *cbp, *savedcbp = NULL;
	struct thread *td = curthread;
	struct fs *fs;
	struct inode *ip;
	struct vnode *vp = NULL;
	ufs2_daddr_t lbn, blkno, *snapblklist;
	int lower, upper, mid, indiroff, error = 0;
	int launched_async_io, prev_norunningbuf;
	long saved_runningbufspace;

	if (devvp != bp->b_vp && IS_SNAPSHOT(VTOI(bp->b_vp)))
		return (0);		/* Update on a snapshot file */
	if (td->td_pflags & TDP_COWINPROGRESS)
		panic("ffs_copyonwrite: recursive call");
	/*
	 * First check to see if it is in the preallocated list.
	 * By doing this check we avoid several potential deadlocks.
	 */
	VI_LOCK(devvp);
	sn = devvp->v_rdev->si_snapdata;
	if (sn == NULL ||
	    TAILQ_EMPTY(&sn->sn_head)) {
		VI_UNLOCK(devvp);
		return (0);		/* No snapshot */
	}
	ip = TAILQ_FIRST(&sn->sn_head);
	fs = ITOFS(ip);
	lbn = fragstoblks(fs, dbtofsb(fs, bp->b_blkno));
	snapblklist = sn->sn_blklist;
	upper = sn->sn_listsize - 1;
	lower = 1;
	while (lower <= upper) {
		mid = (lower + upper) / 2;
		if (snapblklist[mid] == lbn)
			break;
		if (snapblklist[mid] < lbn)
			lower = mid + 1;
		else
			upper = mid - 1;
	}
	if (lower <= upper) {
		VI_UNLOCK(devvp);
		return (0);
	}
	launched_async_io = 0;
	prev_norunningbuf = td->td_pflags & TDP_NORUNNINGBUF;
	/*
	 * Since I/O on bp isn't yet in progress and it may be blocked
	 * for a long time waiting on snaplk, back it out of
	 * runningbufspace, possibly waking other threads waiting for space.
	 */
	saved_runningbufspace = bp->b_runningbufspace;
	if (saved_runningbufspace != 0)
		runningbufwakeup(bp);
	/*
	 * Not in the precomputed list, so check the snapshots.
	 */
	while (lockmgr(&sn->sn_lock, LK_INTERLOCK | LK_EXCLUSIVE | LK_SLEEPFAIL,
	    VI_MTX(devvp)) != 0) {
		VI_LOCK(devvp);
		sn = devvp->v_rdev->si_snapdata;
		if (sn == NULL ||
		    TAILQ_EMPTY(&sn->sn_head)) {
			VI_UNLOCK(devvp);
			if (saved_runningbufspace != 0) {
				bp->b_runningbufspace = saved_runningbufspace;
				atomic_add_long(&runningbufspace,
					       bp->b_runningbufspace);
			}
			return (0);		/* Snapshot gone */
		}
	}
	TAILQ_FOREACH(ip, &sn->sn_head, i_nextsnap) {
		vp = ITOV(ip);
		if (DOINGSOFTDEP(vp))
			softdep_prealloc(vp, MNT_WAIT);
		/*
		 * We ensure that everything of our own that needs to be
		 * copied will be done at the time that ffs_snapshot is
		 * called. Thus we can skip the check here which can
		 * deadlock in doing the lookup in UFS_BALLOC.
		 */
		if (bp->b_vp == vp)
			continue;
		/*
		 * Check to see if block needs to be copied. We do not have
		 * to hold the snapshot lock while doing this lookup as it
		 * will never require any additional allocations for the
		 * snapshot inode.
		 */
		if (lbn < UFS_NDADDR) {
			blkno = DIP(ip, i_db[lbn]);
		} else {
			td->td_pflags |= TDP_COWINPROGRESS | TDP_NORUNNINGBUF;
			error = UFS_BALLOC(vp, lblktosize(fs, (off_t)lbn),
			   fs->fs_bsize, KERNCRED, BA_METAONLY, &ibp);
			td->td_pflags &= ~TDP_COWINPROGRESS;
			if (error)
				break;
			indiroff = (lbn - UFS_NDADDR) % NINDIR(fs);
			if (I_IS_UFS1(ip))
				blkno=((ufs1_daddr_t *)(ibp->b_data))[indiroff];
			else
				blkno=((ufs2_daddr_t *)(ibp->b_data))[indiroff];
			bqrelse(ibp);
		}
#ifdef INVARIANTS
		if (blkno == BLK_SNAP && bp->b_lblkno >= 0)
			panic("ffs_copyonwrite: bad copy block");
#endif
		if (blkno != 0)
			continue;
		/*
		 * Allocate the block into which to do the copy. Since
		 * multiple processes may all try to copy the same block,
		 * we have to recheck our need to do a copy if we sleep
		 * waiting for the lock.
		 *
		 * Because all snapshots on a filesystem share a single
		 * lock, we ensure that we will never be in competition
		 * with another process to allocate a block.
		 */
		td->td_pflags |= TDP_COWINPROGRESS | TDP_NORUNNINGBUF;
		error = UFS_BALLOC(vp, lblktosize(fs, (off_t)lbn),
		    fs->fs_bsize, KERNCRED, 0, &cbp);
		td->td_pflags &= ~TDP_COWINPROGRESS;
		if (error)
			break;
#ifdef DIAGNOSTIC
		if (snapdebug) {
			printf("Copyonwrite: snapino %ju lbn %jd for ",
			    (uintmax_t)ip->i_number, (intmax_t)lbn);
			if (bp->b_vp == devvp)
				printf("fs metadata");
			else
				printf("inum %ju",
				    (uintmax_t)VTOI(bp->b_vp)->i_number);
			printf(" lblkno %jd to blkno %jd\n",
			    (intmax_t)bp->b_lblkno, (intmax_t)cbp->b_blkno);
		}
#endif
		/*
		 * If we have already read the old block contents, then
		 * simply copy them to the new block. Note that we need
		 * to synchronously write snapshots that have not been
		 * unlinked, and hence will be visible after a crash,
		 * to ensure their integrity. At a minimum we ensure the
		 * integrity of the filesystem metadata, but use the
		 * dopersistence sysctl-setable flag to decide on the
		 * persistence needed for file content data.
		 */
		if (savedcbp != NULL) {
			bcopy(savedcbp->b_data, cbp->b_data, fs->fs_bsize);
			bawrite(cbp);
			if ((devvp == bp->b_vp || bp->b_vp->v_type == VDIR ||
			    dopersistence) && ip->i_effnlink > 0)
				(void) ffs_syncvnode(vp, MNT_WAIT, NO_INO_UPDT);
			else
				launched_async_io = 1;
			continue;
		}
		/*
		 * Otherwise, read the old block contents into the buffer.
		 */
		if ((error = readblock(vp, cbp, lbn)) != 0) {
			bzero(cbp->b_data, fs->fs_bsize);
			bawrite(cbp);
			if ((devvp == bp->b_vp || bp->b_vp->v_type == VDIR ||
			    dopersistence) && ip->i_effnlink > 0)
				(void) ffs_syncvnode(vp, MNT_WAIT, NO_INO_UPDT);
			else
				launched_async_io = 1;
			break;
		}
		savedcbp = cbp;
	}
	/*
	 * Note that we need to synchronously write snapshots that
	 * have not been unlinked, and hence will be visible after
	 * a crash, to ensure their integrity. At a minimum we
	 * ensure the integrity of the filesystem metadata, but
	 * use the dopersistence sysctl-setable flag to decide on
	 * the persistence needed for file content data.
	 */
	if (savedcbp) {
		vp = savedcbp->b_vp;
		bawrite(savedcbp);
		if ((devvp == bp->b_vp || bp->b_vp->v_type == VDIR ||
		    dopersistence) && VTOI(vp)->i_effnlink > 0)
			(void) ffs_syncvnode(vp, MNT_WAIT, NO_INO_UPDT);
		else
			launched_async_io = 1;
	}
	lockmgr(vp->v_vnlock, LK_RELEASE, NULL);
	td->td_pflags = (td->td_pflags & ~TDP_NORUNNINGBUF) |
		prev_norunningbuf;
	if (launched_async_io && (td->td_pflags & TDP_NORUNNINGBUF) == 0)
		waitrunningbufspace();
	/*
	 * I/O on bp will now be started, so count it in runningbufspace.
	 */
	if (saved_runningbufspace != 0) {
		bp->b_runningbufspace = saved_runningbufspace;
		atomic_add_long(&runningbufspace, bp->b_runningbufspace);
	}
	return (error);
}

/*
 * sync snapshots to force freework records waiting on snapshots to claim
 * blocks to free.
 */
void
ffs_sync_snap(mp, waitfor)
	struct mount *mp;
	int waitfor;
{
	struct snapdata *sn;
	struct vnode *devvp;
	struct vnode *vp;
	struct inode *ip;

	devvp = VFSTOUFS(mp)->um_devvp;
	if ((devvp->v_vflag & VV_COPYONWRITE) == 0)
		return;
	for (;;) {
		VI_LOCK(devvp);
		sn = devvp->v_rdev->si_snapdata;
		if (sn == NULL) {
			VI_UNLOCK(devvp);
			return;
		}
		if (lockmgr(&sn->sn_lock,
		    LK_INTERLOCK | LK_EXCLUSIVE | LK_SLEEPFAIL,
		    VI_MTX(devvp)) == 0)
			break;
	}
	TAILQ_FOREACH(ip, &sn->sn_head, i_nextsnap) {
		vp = ITOV(ip);
		ffs_syncvnode(vp, waitfor, NO_INO_UPDT);
	}
	lockmgr(&sn->sn_lock, LK_RELEASE, NULL);
}

/*
 * Read the specified block into the given buffer.
 * Much of this boiler-plate comes from bwrite().
 */
static int
readblock(vp, bp, lbn)
	struct vnode *vp;
	struct buf *bp;
	ufs2_daddr_t lbn;
{
	struct inode *ip;
	struct bio *bip;
	struct fs *fs;

	ip = VTOI(vp);
	fs = ITOFS(ip);

	bip = g_alloc_bio();
	bip->bio_cmd = BIO_READ;
	bip->bio_offset = dbtob(fsbtodb(fs, blkstofrags(fs, lbn)));
	bip->bio_data = bp->b_data;
	bip->bio_length = bp->b_bcount;
	bip->bio_done = NULL;

	g_io_request(bip, ITODEVVP(ip)->v_bufobj.bo_private);
	bp->b_error = biowait(bip, "snaprdb");
	g_destroy_bio(bip);
	return (bp->b_error);
}

#endif

/*
 * Process file deletes that were deferred by ufs_inactive() due to
 * the file system being suspended. Transfer IN_LAZYACCESS into
 * IN_MODIFIED for vnodes that were accessed during suspension.
 */
void
process_deferred_inactive(struct mount *mp)
{
	struct vnode *vp, *mvp;
	struct inode *ip;
	int error;

	(void) vn_start_secondary_write(NULL, &mp, V_WAIT);
 loop:
	MNT_VNODE_FOREACH_ALL(vp, mp, mvp) {
		/*
		 * IN_LAZYACCESS is checked here without holding any
		 * vnode lock, but this flag is set only while holding
		 * vnode interlock.
		 */
		if (vp->v_type == VNON ||
		    ((VTOI(vp)->i_flag & IN_LAZYACCESS) == 0 &&
		    ((vp->v_iflag & VI_OWEINACT) == 0 || vp->v_usecount > 0))) {
			VI_UNLOCK(vp);
			continue;
		}
		vholdl(vp);
		error = vn_lock(vp, LK_EXCLUSIVE | LK_INTERLOCK);
		if (error != 0) {
			vdrop(vp);
			if (error == ENOENT)
				continue;	/* vnode recycled */
			MNT_VNODE_FOREACH_ALL_ABORT(mp, mvp);
			goto loop;
		}
		ip = VTOI(vp);
		if ((ip->i_flag & IN_LAZYACCESS) != 0) {
			ip->i_flag &= ~IN_LAZYACCESS;
			UFS_INODE_SET_FLAG(ip, IN_MODIFIED);
		}
		VI_LOCK(vp);
		vinactive(vp);
		VI_UNLOCK(vp);
		VOP_UNLOCK(vp);
		vdrop(vp);
	}
	vn_finished_secondary_write(mp);
}

#ifndef NO_FFS_SNAPSHOT

static struct snapdata *
ffs_snapdata_alloc(void)
{
	struct snapdata *sn;

	/*
	 * Fetch a snapdata from the free list if there is one available.
	 */
	mtx_lock(&snapfree_lock);
	sn = LIST_FIRST(&snapfree);
	if (sn != NULL)
		LIST_REMOVE(sn, sn_link);
	mtx_unlock(&snapfree_lock);
	if (sn != NULL)
		return (sn);
	/*
 	 * If there were no free snapdatas allocate one.
	 */
	sn = malloc(sizeof *sn, M_UFSMNT, M_WAITOK | M_ZERO);
	TAILQ_INIT(&sn->sn_head);
	lockinit(&sn->sn_lock, PVFS, "snaplk", VLKTIMEOUT,
	    LK_CANRECURSE | LK_NOSHARE);
	return (sn);
}

/*
 * The snapdata is never freed because we can not be certain that
 * there are no threads sleeping on the snap lock.  Persisting
 * them permanently avoids costly synchronization in ffs_lock().
 */
static void
ffs_snapdata_free(struct snapdata *sn)
{
	mtx_lock(&snapfree_lock);
	LIST_INSERT_HEAD(&snapfree, sn, sn_link);
	mtx_unlock(&snapfree_lock);
}

/* Try to free snapdata associated with devvp */
static void
try_free_snapdata(struct vnode *devvp)
{
	struct snapdata *sn;
	ufs2_daddr_t *snapblklist;

	ASSERT_VI_LOCKED(devvp, "try_free_snapdata");
	sn = devvp->v_rdev->si_snapdata;

	if (sn == NULL || TAILQ_FIRST(&sn->sn_head) != NULL ||
	    (devvp->v_vflag & VV_COPYONWRITE) == 0) {
		VI_UNLOCK(devvp);
		return;
	}

	devvp->v_rdev->si_snapdata = NULL;
	devvp->v_vflag &= ~VV_COPYONWRITE;
	lockmgr(&sn->sn_lock, LK_DRAIN|LK_INTERLOCK, VI_MTX(devvp));
	snapblklist = sn->sn_blklist;
	sn->sn_blklist = NULL;
	sn->sn_listsize = 0;
	lockmgr(&sn->sn_lock, LK_RELEASE, NULL);
	if (snapblklist != NULL)
		free(snapblklist, M_UFSMNT);
	ffs_snapdata_free(sn);
}

static struct snapdata *
ffs_snapdata_acquire(struct vnode *devvp)
{
	struct snapdata *nsn, *sn;
	int error;

	/*
	 * Allocate a free snapdata.  This is done before acquiring the
	 * devvp lock to avoid allocation while the devvp interlock is
	 * held.
	 */
	nsn = ffs_snapdata_alloc();

	for (;;) {
		VI_LOCK(devvp);
		sn = devvp->v_rdev->si_snapdata;
		if (sn == NULL) {
			/*
			 * This is the first snapshot on this
			 * filesystem and we use our pre-allocated
			 * snapdata.  Publish sn with the sn_lock
			 * owned by us, to avoid the race.
			 */
			error = lockmgr(&nsn->sn_lock, LK_EXCLUSIVE |
			    LK_NOWAIT, NULL);
			if (error != 0)
				panic("leaked sn, lockmgr error %d", error);
			sn = devvp->v_rdev->si_snapdata = nsn;
			VI_UNLOCK(devvp);
			nsn = NULL;
			break;
		}

		/*
		 * There is a snapshots which already exists on this
		 * filesystem, grab a reference to the common lock.
		 */
		error = lockmgr(&sn->sn_lock, LK_INTERLOCK |
		    LK_EXCLUSIVE | LK_SLEEPFAIL, VI_MTX(devvp));
		if (error == 0)
			break;
	}

	/*
	 * Free any unused snapdata.
	 */
	if (nsn != NULL)
		ffs_snapdata_free(nsn);

	return (sn);
}

#endif