aboutsummaryrefslogtreecommitdiff
path: root/sys/kern/vfs_vnops.c
blob: 832c717a33b71c553fae1481ea2bae45e030b09f (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
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
2959
2960
2961
2962
2963
2964
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
3015
3016
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
3329
3330
3331
3332
3333
3334
3335
3336
3337
3338
3339
3340
3341
3342
3343
3344
3345
3346
3347
3348
3349
3350
3351
3352
3353
3354
3355
3356
3357
3358
3359
3360
3361
3362
3363
3364
3365
3366
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
3378
3379
3380
3381
3382
3383
3384
3385
3386
3387
3388
3389
3390
3391
3392
3393
3394
3395
3396
3397
3398
3399
3400
3401
3402
3403
3404
3405
3406
3407
3408
3409
3410
3411
3412
3413
3414
3415
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437
3438
3439
3440
3441
3442
3443
3444
3445
3446
3447
3448
3449
3450
3451
3452
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
3489
3490
3491
3492
3493
/*-
 * SPDX-License-Identifier: BSD-3-Clause
 *
 * Copyright (c) 1982, 1986, 1989, 1993
 *	The Regents of the University of California.  All rights reserved.
 * (c) UNIX System Laboratories, Inc.
 * All or some portions of this file are derived from material licensed
 * to the University of California by American Telephone and Telegraph
 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
 * the permission of UNIX System Laboratories, Inc.
 *
 * Copyright (c) 2012 Konstantin Belousov <kib@FreeBSD.org>
 * Copyright (c) 2013, 2014 The FreeBSD Foundation
 *
 * Portions of this software were developed by Konstantin Belousov
 * under sponsorship from the FreeBSD Foundation.
 *
 * 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.
 * 3. Neither the name of the University nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``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 THE REGENTS OR CONTRIBUTORS 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.
 *
 *	@(#)vfs_vnops.c	8.2 (Berkeley) 1/21/94
 */

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

#include "opt_hwpmc_hooks.h"

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/disk.h>
#include <sys/fail.h>
#include <sys/fcntl.h>
#include <sys/file.h>
#include <sys/kdb.h>
#include <sys/ktr.h>
#include <sys/stat.h>
#include <sys/priv.h>
#include <sys/proc.h>
#include <sys/limits.h>
#include <sys/lock.h>
#include <sys/mman.h>
#include <sys/mount.h>
#include <sys/mutex.h>
#include <sys/namei.h>
#include <sys/vnode.h>
#include <sys/bio.h>
#include <sys/buf.h>
#include <sys/filio.h>
#include <sys/resourcevar.h>
#include <sys/rwlock.h>
#include <sys/prng.h>
#include <sys/sx.h>
#include <sys/sleepqueue.h>
#include <sys/sysctl.h>
#include <sys/ttycom.h>
#include <sys/conf.h>
#include <sys/syslog.h>
#include <sys/unistd.h>
#include <sys/user.h>

#include <security/audit/audit.h>
#include <security/mac/mac_framework.h>

#include <vm/vm.h>
#include <vm/vm_extern.h>
#include <vm/pmap.h>
#include <vm/vm_map.h>
#include <vm/vm_object.h>
#include <vm/vm_page.h>
#include <vm/vm_pager.h>

#ifdef HWPMC_HOOKS
#include <sys/pmckern.h>
#endif

static fo_rdwr_t	vn_read;
static fo_rdwr_t	vn_write;
static fo_rdwr_t	vn_io_fault;
static fo_truncate_t	vn_truncate;
static fo_ioctl_t	vn_ioctl;
static fo_poll_t	vn_poll;
static fo_kqfilter_t	vn_kqfilter;
static fo_close_t	vn_closefile;
static fo_mmap_t	vn_mmap;
static fo_fallocate_t	vn_fallocate;

struct 	fileops vnops = {
	.fo_read = vn_io_fault,
	.fo_write = vn_io_fault,
	.fo_truncate = vn_truncate,
	.fo_ioctl = vn_ioctl,
	.fo_poll = vn_poll,
	.fo_kqfilter = vn_kqfilter,
	.fo_stat = vn_statfile,
	.fo_close = vn_closefile,
	.fo_chmod = vn_chmod,
	.fo_chown = vn_chown,
	.fo_sendfile = vn_sendfile,
	.fo_seek = vn_seek,
	.fo_fill_kinfo = vn_fill_kinfo,
	.fo_mmap = vn_mmap,
	.fo_fallocate = vn_fallocate,
	.fo_flags = DFLAG_PASSABLE | DFLAG_SEEKABLE
};

const u_int io_hold_cnt = 16;
static int vn_io_fault_enable = 1;
SYSCTL_INT(_debug, OID_AUTO, vn_io_fault_enable, CTLFLAG_RWTUN,
    &vn_io_fault_enable, 0, "Enable vn_io_fault lock avoidance");
static int vn_io_fault_prefault = 0;
SYSCTL_INT(_debug, OID_AUTO, vn_io_fault_prefault, CTLFLAG_RWTUN,
    &vn_io_fault_prefault, 0, "Enable vn_io_fault prefaulting");
static int vn_io_pgcache_read_enable = 1;
SYSCTL_INT(_debug, OID_AUTO, vn_io_pgcache_read_enable, CTLFLAG_RWTUN,
    &vn_io_pgcache_read_enable, 0,
    "Enable copying from page cache for reads, avoiding fs");
static u_long vn_io_faults_cnt;
SYSCTL_ULONG(_debug, OID_AUTO, vn_io_faults, CTLFLAG_RD,
    &vn_io_faults_cnt, 0, "Count of vn_io_fault lock avoidance triggers");

static int vfs_allow_read_dir = 0;
SYSCTL_INT(_security_bsd, OID_AUTO, allow_read_dir, CTLFLAG_RW,
    &vfs_allow_read_dir, 0,
    "Enable read(2) of directory by root for filesystems that support it");

/*
 * Returns true if vn_io_fault mode of handling the i/o request should
 * be used.
 */
static bool
do_vn_io_fault(struct vnode *vp, struct uio *uio)
{
	struct mount *mp;

	return (uio->uio_segflg == UIO_USERSPACE && vp->v_type == VREG &&
	    (mp = vp->v_mount) != NULL &&
	    (mp->mnt_kern_flag & MNTK_NO_IOPF) != 0 && vn_io_fault_enable);
}

/*
 * Structure used to pass arguments to vn_io_fault1(), to do either
 * file- or vnode-based I/O calls.
 */
struct vn_io_fault_args {
	enum {
		VN_IO_FAULT_FOP,
		VN_IO_FAULT_VOP
	} kind;
	struct ucred *cred;
	int flags;
	union {
		struct fop_args_tag {
			struct file *fp;
			fo_rdwr_t *doio;
		} fop_args;
		struct vop_args_tag {
			struct vnode *vp;
		} vop_args;
	} args;
};

static int vn_io_fault1(struct vnode *vp, struct uio *uio,
    struct vn_io_fault_args *args, struct thread *td);

int
vn_open(struct nameidata *ndp, int *flagp, int cmode, struct file *fp)
{
	struct thread *td = ndp->ni_cnd.cn_thread;

	return (vn_open_cred(ndp, flagp, cmode, 0, td->td_ucred, fp));
}

static uint64_t
open2nameif(int fmode, u_int vn_open_flags)
{
	uint64_t res;

	res = ISOPEN | LOCKLEAF;
	if ((fmode & O_RESOLVE_BENEATH) != 0)
		res |= RBENEATH;
	if ((vn_open_flags & VN_OPEN_NOAUDIT) == 0)
		res |= AUDITVNODE1;
	if ((vn_open_flags & VN_OPEN_NOCAPCHECK) != 0)
		res |= NOCAPCHECK;
	return (res);
}

/*
 * Common code for vnode open operations via a name lookup.
 * Lookup the vnode and invoke VOP_CREATE if needed.
 * Check permissions, and call the VOP_OPEN or VOP_CREATE routine.
 *
 * Note that this does NOT free nameidata for the successful case,
 * due to the NDINIT being done elsewhere.
 */
int
vn_open_cred(struct nameidata *ndp, int *flagp, int cmode, u_int vn_open_flags,
    struct ucred *cred, struct file *fp)
{
	struct vnode *vp;
	struct mount *mp;
	struct thread *td = ndp->ni_cnd.cn_thread;
	struct vattr vat;
	struct vattr *vap = &vat;
	int fmode, error;
	bool first_open;

restart:
	first_open = false;
	fmode = *flagp;
	if ((fmode & (O_CREAT | O_EXCL | O_DIRECTORY)) == (O_CREAT |
	    O_EXCL | O_DIRECTORY))
		return (EINVAL);
	else if ((fmode & (O_CREAT | O_DIRECTORY)) == O_CREAT) {
		ndp->ni_cnd.cn_nameiop = CREATE;
		ndp->ni_cnd.cn_flags = open2nameif(fmode, vn_open_flags);
		/*
		 * Set NOCACHE to avoid flushing the cache when
		 * rolling in many files at once.
		 *
		 * Set NC_KEEPPOSENTRY to keep positive entries if they already
		 * exist despite NOCACHE.
		 */
		ndp->ni_cnd.cn_flags |= LOCKPARENT | NOCACHE | NC_KEEPPOSENTRY;
		if ((fmode & O_EXCL) == 0 && (fmode & O_NOFOLLOW) == 0)
			ndp->ni_cnd.cn_flags |= FOLLOW;
		if ((vn_open_flags & VN_OPEN_INVFS) == 0)
			bwillwrite();
		if ((error = namei(ndp)) != 0)
			return (error);
		if (ndp->ni_vp == NULL) {
			VATTR_NULL(vap);
			vap->va_type = VREG;
			vap->va_mode = cmode;
			if (fmode & O_EXCL)
				vap->va_vaflags |= VA_EXCLUSIVE;
			if (vn_start_write(ndp->ni_dvp, &mp, V_NOWAIT) != 0) {
				NDFREE(ndp, NDF_ONLY_PNBUF);
				vput(ndp->ni_dvp);
				if ((error = vn_start_write(NULL, &mp,
				    V_XSLEEP | PCATCH)) != 0)
					return (error);
				NDREINIT(ndp);
				goto restart;
			}
			if ((vn_open_flags & VN_OPEN_NAMECACHE) != 0)
				ndp->ni_cnd.cn_flags |= MAKEENTRY;
#ifdef MAC
			error = mac_vnode_check_create(cred, ndp->ni_dvp,
			    &ndp->ni_cnd, vap);
			if (error == 0)
#endif
				error = VOP_CREATE(ndp->ni_dvp, &ndp->ni_vp,
				    &ndp->ni_cnd, vap);
			vp = ndp->ni_vp;
			if (error == 0 && (fmode & O_EXCL) != 0 &&
			    (fmode & (O_EXLOCK | O_SHLOCK)) != 0) {
				VI_LOCK(vp);
				vp->v_iflag |= VI_FOPENING;
				VI_UNLOCK(vp);
				first_open = true;
			}
			VOP_VPUT_PAIR(ndp->ni_dvp, error == 0 ? &vp : NULL,
			    false);
			vn_finished_write(mp);
			if (error) {
				NDFREE(ndp, NDF_ONLY_PNBUF);
				if (error == ERELOOKUP) {
					NDREINIT(ndp);
					goto restart;
				}
				return (error);
			}
			fmode &= ~O_TRUNC;
		} else {
			if (ndp->ni_dvp == ndp->ni_vp)
				vrele(ndp->ni_dvp);
			else
				vput(ndp->ni_dvp);
			ndp->ni_dvp = NULL;
			vp = ndp->ni_vp;
			if (fmode & O_EXCL) {
				error = EEXIST;
				goto bad;
			}
			if (vp->v_type == VDIR) {
				error = EISDIR;
				goto bad;
			}
			fmode &= ~O_CREAT;
		}
	} else {
		ndp->ni_cnd.cn_nameiop = LOOKUP;
		ndp->ni_cnd.cn_flags = open2nameif(fmode, vn_open_flags);
		ndp->ni_cnd.cn_flags |= (fmode & O_NOFOLLOW) != 0 ? NOFOLLOW :
		    FOLLOW;
		if ((fmode & FWRITE) == 0)
			ndp->ni_cnd.cn_flags |= LOCKSHARED;
		if ((error = namei(ndp)) != 0)
			return (error);
		vp = ndp->ni_vp;
	}
	error = vn_open_vnode(vp, fmode, cred, td, fp);
	if (first_open) {
		VI_LOCK(vp);
		vp->v_iflag &= ~VI_FOPENING;
		wakeup(vp);
		VI_UNLOCK(vp);
	}
	if (error)
		goto bad;
	*flagp = fmode;
	return (0);
bad:
	NDFREE(ndp, NDF_ONLY_PNBUF);
	vput(vp);
	*flagp = fmode;
	ndp->ni_vp = NULL;
	return (error);
}

static int
vn_open_vnode_advlock(struct vnode *vp, int fmode, struct file *fp)
{
	struct flock lf;
	int error, lock_flags, type;

	ASSERT_VOP_LOCKED(vp, "vn_open_vnode_advlock");
	if ((fmode & (O_EXLOCK | O_SHLOCK)) == 0)
		return (0);
	KASSERT(fp != NULL, ("open with flock requires fp"));
	if (fp->f_type != DTYPE_NONE && fp->f_type != DTYPE_VNODE)
		return (EOPNOTSUPP);

	lock_flags = VOP_ISLOCKED(vp);
	VOP_UNLOCK(vp);

	lf.l_whence = SEEK_SET;
	lf.l_start = 0;
	lf.l_len = 0;
	lf.l_type = (fmode & O_EXLOCK) != 0 ? F_WRLCK : F_RDLCK;
	type = F_FLOCK;
	if ((fmode & FNONBLOCK) == 0)
		type |= F_WAIT;
	if ((fmode & (O_CREAT | O_EXCL)) == (O_CREAT | O_EXCL))
		type |= F_FIRSTOPEN;
	error = VOP_ADVLOCK(vp, (caddr_t)fp, F_SETLK, &lf, type);
	if (error == 0)
		fp->f_flag |= FHASLOCK;

	vn_lock(vp, lock_flags | LK_RETRY);
	return (error);
}

/*
 * Common code for vnode open operations once a vnode is located.
 * Check permissions, and call the VOP_OPEN routine.
 */
int
vn_open_vnode(struct vnode *vp, int fmode, struct ucred *cred,
    struct thread *td, struct file *fp)
{
	accmode_t accmode;
	int error;

	if (vp->v_type == VLNK) {
		if ((fmode & O_PATH) == 0 || (fmode & FEXEC) != 0)
			return (EMLINK);
	}
	if (vp->v_type == VSOCK)
		return (EOPNOTSUPP);
	if (vp->v_type != VDIR && fmode & O_DIRECTORY)
		return (ENOTDIR);

	accmode = 0;
	if ((fmode & O_PATH) == 0) {
		if ((fmode & (FWRITE | O_TRUNC)) != 0) {
			if (vp->v_type == VDIR)
				return (EISDIR);
			accmode |= VWRITE;
		}
		if ((fmode & FREAD) != 0)
			accmode |= VREAD;
		if ((fmode & O_APPEND) && (fmode & FWRITE))
			accmode |= VAPPEND;
#ifdef MAC
		if ((fmode & O_CREAT) != 0)
			accmode |= VCREAT;
#endif
	}
	if ((fmode & FEXEC) != 0)
		accmode |= VEXEC;
#ifdef MAC
	if ((fmode & O_VERIFY) != 0)
		accmode |= VVERIFY;
	error = mac_vnode_check_open(cred, vp, accmode);
	if (error != 0)
		return (error);

	accmode &= ~(VCREAT | VVERIFY);
#endif
	if ((fmode & O_CREAT) == 0 && accmode != 0) {
		error = VOP_ACCESS(vp, accmode, cred, td);
		if (error != 0)
			return (error);
	}
	if ((fmode & O_PATH) != 0) {
		if (vp->v_type == VFIFO)
			error = EPIPE;
		else
			error = VOP_ACCESS(vp, VREAD, cred, td);
		if (error == 0)
			fp->f_flag |= FKQALLOWED;
		return (0);
	}

	if (vp->v_type == VFIFO && VOP_ISLOCKED(vp) != LK_EXCLUSIVE)
		vn_lock(vp, LK_UPGRADE | LK_RETRY);
	error = VOP_OPEN(vp, fmode, cred, td, fp);
	if (error != 0)
		return (error);

	error = vn_open_vnode_advlock(vp, fmode, fp);
	if (error == 0 && (fmode & FWRITE) != 0) {
		error = VOP_ADD_WRITECOUNT(vp, 1);
		if (error == 0) {
			CTR3(KTR_VFS, "%s: vp %p v_writecount increased to %d",
			     __func__, vp, vp->v_writecount);
		}
	}

	/*
	 * Error from advlock or VOP_ADD_WRITECOUNT() still requires
	 * calling VOP_CLOSE() to pair with earlier VOP_OPEN().
	 */
	if (error != 0) {
		if (fp != NULL) {
			/*
			 * Arrange the call by having fdrop() to use
			 * vn_closefile().  This is to satisfy
			 * filesystems like devfs or tmpfs, which
			 * override fo_close().
			 */
			fp->f_flag |= FOPENFAILED;
			fp->f_vnode = vp;
			if (fp->f_ops == &badfileops) {
				fp->f_type = DTYPE_VNODE;
				fp->f_ops = &vnops;
			}
			vref(vp);
		} else {
			/*
			 * If there is no fp, due to kernel-mode open,
			 * we can call VOP_CLOSE() now.
			 */
			if (vp->v_type != VFIFO && (fmode & FWRITE) != 0 &&
			    !MNT_EXTENDED_SHARED(vp->v_mount) &&
			    VOP_ISLOCKED(vp) != LK_EXCLUSIVE)
				vn_lock(vp, LK_UPGRADE | LK_RETRY);
			(void)VOP_CLOSE(vp, fmode & (FREAD | FWRITE | FEXEC),
			    cred, td);
		}
	}

	ASSERT_VOP_LOCKED(vp, "vn_open_vnode");
	return (error);

}

/*
 * Check for write permissions on the specified vnode.
 * Prototype text segments cannot be written.
 * It is racy.
 */
int
vn_writechk(struct vnode *vp)
{

	ASSERT_VOP_LOCKED(vp, "vn_writechk");
	/*
	 * If there's shared text associated with
	 * the vnode, try to free it up once.  If
	 * we fail, we can't allow writing.
	 */
	if (VOP_IS_TEXT(vp))
		return (ETXTBSY);

	return (0);
}

/*
 * Vnode close call
 */
static int
vn_close1(struct vnode *vp, int flags, struct ucred *file_cred,
    struct thread *td, bool keep_ref)
{
	struct mount *mp;
	int error, lock_flags;

	if (vp->v_type != VFIFO && (flags & FWRITE) == 0 &&
	    MNT_EXTENDED_SHARED(vp->v_mount))
		lock_flags = LK_SHARED;
	else
		lock_flags = LK_EXCLUSIVE;

	vn_start_write(vp, &mp, V_WAIT);
	vn_lock(vp, lock_flags | LK_RETRY);
	AUDIT_ARG_VNODE1(vp);
	if ((flags & (FWRITE | FOPENFAILED)) == FWRITE) {
		VOP_ADD_WRITECOUNT_CHECKED(vp, -1);
		CTR3(KTR_VFS, "%s: vp %p v_writecount decreased to %d",
		    __func__, vp, vp->v_writecount);
	}
	error = VOP_CLOSE(vp, flags, file_cred, td);
	if (keep_ref)
		VOP_UNLOCK(vp);
	else
		vput(vp);
	vn_finished_write(mp);
	return (error);
}

int
vn_close(struct vnode *vp, int flags, struct ucred *file_cred,
    struct thread *td)
{

	return (vn_close1(vp, flags, file_cred, td, false));
}

/*
 * Heuristic to detect sequential operation.
 */
static int
sequential_heuristic(struct uio *uio, struct file *fp)
{
	enum uio_rw rw;

	ASSERT_VOP_LOCKED(fp->f_vnode, __func__);

	rw = uio->uio_rw;
	if (fp->f_flag & FRDAHEAD)
		return (fp->f_seqcount[rw] << IO_SEQSHIFT);

	/*
	 * Offset 0 is handled specially.  open() sets f_seqcount to 1 so
	 * that the first I/O is normally considered to be slightly
	 * sequential.  Seeking to offset 0 doesn't change sequentiality
	 * unless previous seeks have reduced f_seqcount to 0, in which
	 * case offset 0 is not special.
	 */
	if ((uio->uio_offset == 0 && fp->f_seqcount[rw] > 0) ||
	    uio->uio_offset == fp->f_nextoff[rw]) {
		/*
		 * f_seqcount is in units of fixed-size blocks so that it
		 * depends mainly on the amount of sequential I/O and not
		 * much on the number of sequential I/O's.  The fixed size
		 * of 16384 is hard-coded here since it is (not quite) just
		 * a magic size that works well here.  This size is more
		 * closely related to the best I/O size for real disks than
		 * to any block size used by software.
		 */
		if (uio->uio_resid >= IO_SEQMAX * 16384)
			fp->f_seqcount[rw] = IO_SEQMAX;
		else {
			fp->f_seqcount[rw] += howmany(uio->uio_resid, 16384);
			if (fp->f_seqcount[rw] > IO_SEQMAX)
				fp->f_seqcount[rw] = IO_SEQMAX;
		}
		return (fp->f_seqcount[rw] << IO_SEQSHIFT);
	}

	/* Not sequential.  Quickly draw-down sequentiality. */
	if (fp->f_seqcount[rw] > 1)
		fp->f_seqcount[rw] = 1;
	else
		fp->f_seqcount[rw] = 0;
	return (0);
}

/*
 * Package up an I/O request on a vnode into a uio and do it.
 */
int
vn_rdwr(enum uio_rw rw, struct vnode *vp, void *base, int len, off_t offset,
    enum uio_seg segflg, int ioflg, struct ucred *active_cred,
    struct ucred *file_cred, ssize_t *aresid, struct thread *td)
{
	struct uio auio;
	struct iovec aiov;
	struct mount *mp;
	struct ucred *cred;
	void *rl_cookie;
	struct vn_io_fault_args args;
	int error, lock_flags;

	if (offset < 0 && vp->v_type != VCHR)
		return (EINVAL);
	auio.uio_iov = &aiov;
	auio.uio_iovcnt = 1;
	aiov.iov_base = base;
	aiov.iov_len = len;
	auio.uio_resid = len;
	auio.uio_offset = offset;
	auio.uio_segflg = segflg;
	auio.uio_rw = rw;
	auio.uio_td = td;
	error = 0;

	if ((ioflg & IO_NODELOCKED) == 0) {
		if ((ioflg & IO_RANGELOCKED) == 0) {
			if (rw == UIO_READ) {
				rl_cookie = vn_rangelock_rlock(vp, offset,
				    offset + len);
			} else if ((ioflg & IO_APPEND) != 0) {
				rl_cookie = vn_rangelock_wlock(vp, 0, OFF_MAX);
			} else {
				rl_cookie = vn_rangelock_wlock(vp, offset,
				    offset + len);
			}
		} else
			rl_cookie = NULL;
		mp = NULL;
		if (rw == UIO_WRITE) { 
			if (vp->v_type != VCHR &&
			    (error = vn_start_write(vp, &mp, V_WAIT | PCATCH))
			    != 0)
				goto out;
			if (MNT_SHARED_WRITES(mp) ||
			    ((mp == NULL) && MNT_SHARED_WRITES(vp->v_mount)))
				lock_flags = LK_SHARED;
			else
				lock_flags = LK_EXCLUSIVE;
		} else
			lock_flags = LK_SHARED;
		vn_lock(vp, lock_flags | LK_RETRY);
	} else
		rl_cookie = NULL;

	ASSERT_VOP_LOCKED(vp, "IO_NODELOCKED with no vp lock held");
#ifdef MAC
	if ((ioflg & IO_NOMACCHECK) == 0) {
		if (rw == UIO_READ)
			error = mac_vnode_check_read(active_cred, file_cred,
			    vp);
		else
			error = mac_vnode_check_write(active_cred, file_cred,
			    vp);
	}
#endif
	if (error == 0) {
		if (file_cred != NULL)
			cred = file_cred;
		else
			cred = active_cred;
		if (do_vn_io_fault(vp, &auio)) {
			args.kind = VN_IO_FAULT_VOP;
			args.cred = cred;
			args.flags = ioflg;
			args.args.vop_args.vp = vp;
			error = vn_io_fault1(vp, &auio, &args, td);
		} else if (rw == UIO_READ) {
			error = VOP_READ(vp, &auio, ioflg, cred);
		} else /* if (rw == UIO_WRITE) */ {
			error = VOP_WRITE(vp, &auio, ioflg, cred);
		}
	}
	if (aresid)
		*aresid = auio.uio_resid;
	else
		if (auio.uio_resid && error == 0)
			error = EIO;
	if ((ioflg & IO_NODELOCKED) == 0) {
		VOP_UNLOCK(vp);
		if (mp != NULL)
			vn_finished_write(mp);
	}
 out:
	if (rl_cookie != NULL)
		vn_rangelock_unlock(vp, rl_cookie);
	return (error);
}

/*
 * Package up an I/O request on a vnode into a uio and do it.  The I/O
 * request is split up into smaller chunks and we try to avoid saturating
 * the buffer cache while potentially holding a vnode locked, so we 
 * check bwillwrite() before calling vn_rdwr().  We also call kern_yield()
 * to give other processes a chance to lock the vnode (either other processes
 * core'ing the same binary, or unrelated processes scanning the directory).
 */
int
vn_rdwr_inchunks(enum uio_rw rw, struct vnode *vp, void *base, size_t len,
    off_t offset, enum uio_seg segflg, int ioflg, struct ucred *active_cred,
    struct ucred *file_cred, size_t *aresid, struct thread *td)
{
	int error = 0;
	ssize_t iaresid;

	do {
		int chunk;

		/*
		 * Force `offset' to a multiple of MAXBSIZE except possibly
		 * for the first chunk, so that filesystems only need to
		 * write full blocks except possibly for the first and last
		 * chunks.
		 */
		chunk = MAXBSIZE - (uoff_t)offset % MAXBSIZE;

		if (chunk > len)
			chunk = len;
		if (rw != UIO_READ && vp->v_type == VREG)
			bwillwrite();
		iaresid = 0;
		error = vn_rdwr(rw, vp, base, chunk, offset, segflg,
		    ioflg, active_cred, file_cred, &iaresid, td);
		len -= chunk;	/* aresid calc already includes length */
		if (error)
			break;
		offset += chunk;
		base = (char *)base + chunk;
		kern_yield(PRI_USER);
	} while (len);
	if (aresid)
		*aresid = len + iaresid;
	return (error);
}

#if OFF_MAX <= LONG_MAX
off_t
foffset_lock(struct file *fp, int flags)
{
	volatile short *flagsp;
	off_t res;
	short state;

	KASSERT((flags & FOF_OFFSET) == 0, ("FOF_OFFSET passed"));

	if ((flags & FOF_NOLOCK) != 0)
		return (atomic_load_long(&fp->f_offset));

	/*
	 * According to McKusick the vn lock was protecting f_offset here.
	 * It is now protected by the FOFFSET_LOCKED flag.
	 */
	flagsp = &fp->f_vnread_flags;
	if (atomic_cmpset_acq_16(flagsp, 0, FOFFSET_LOCKED))
		return (atomic_load_long(&fp->f_offset));

	sleepq_lock(&fp->f_vnread_flags);
	state = atomic_load_16(flagsp);
	for (;;) {
		if ((state & FOFFSET_LOCKED) == 0) {
			if (!atomic_fcmpset_acq_16(flagsp, &state,
			    FOFFSET_LOCKED))
				continue;
			break;
		}
		if ((state & FOFFSET_LOCK_WAITING) == 0) {
			if (!atomic_fcmpset_acq_16(flagsp, &state,
			    state | FOFFSET_LOCK_WAITING))
				continue;
		}
		DROP_GIANT();
		sleepq_add(&fp->f_vnread_flags, NULL, "vofflock", 0, 0);
		sleepq_wait(&fp->f_vnread_flags, PUSER -1);
		PICKUP_GIANT();
		sleepq_lock(&fp->f_vnread_flags);
		state = atomic_load_16(flagsp);
	}
	res = atomic_load_long(&fp->f_offset);
	sleepq_release(&fp->f_vnread_flags);
	return (res);
}

void
foffset_unlock(struct file *fp, off_t val, int flags)
{
	volatile short *flagsp;
	short state;

	KASSERT((flags & FOF_OFFSET) == 0, ("FOF_OFFSET passed"));

	if ((flags & FOF_NOUPDATE) == 0)
		atomic_store_long(&fp->f_offset, val);
	if ((flags & FOF_NEXTOFF_R) != 0)
		fp->f_nextoff[UIO_READ] = val;
	if ((flags & FOF_NEXTOFF_W) != 0)
		fp->f_nextoff[UIO_WRITE] = val;

	if ((flags & FOF_NOLOCK) != 0)
		return;

	flagsp = &fp->f_vnread_flags;
	state = atomic_load_16(flagsp);
	if ((state & FOFFSET_LOCK_WAITING) == 0 &&
	    atomic_cmpset_rel_16(flagsp, state, 0))
		return;

	sleepq_lock(&fp->f_vnread_flags);
	MPASS((fp->f_vnread_flags & FOFFSET_LOCKED) != 0);
	MPASS((fp->f_vnread_flags & FOFFSET_LOCK_WAITING) != 0);
	fp->f_vnread_flags = 0;
	sleepq_broadcast(&fp->f_vnread_flags, SLEEPQ_SLEEP, 0, 0);
	sleepq_release(&fp->f_vnread_flags);
}
#else
off_t
foffset_lock(struct file *fp, int flags)
{
	struct mtx *mtxp;
	off_t res;

	KASSERT((flags & FOF_OFFSET) == 0, ("FOF_OFFSET passed"));

	mtxp = mtx_pool_find(mtxpool_sleep, fp);
	mtx_lock(mtxp);
	if ((flags & FOF_NOLOCK) == 0) {
		while (fp->f_vnread_flags & FOFFSET_LOCKED) {
			fp->f_vnread_flags |= FOFFSET_LOCK_WAITING;
			msleep(&fp->f_vnread_flags, mtxp, PUSER -1,
			    "vofflock", 0);
		}
		fp->f_vnread_flags |= FOFFSET_LOCKED;
	}
	res = fp->f_offset;
	mtx_unlock(mtxp);
	return (res);
}

void
foffset_unlock(struct file *fp, off_t val, int flags)
{
	struct mtx *mtxp;

	KASSERT((flags & FOF_OFFSET) == 0, ("FOF_OFFSET passed"));

	mtxp = mtx_pool_find(mtxpool_sleep, fp);
	mtx_lock(mtxp);
	if ((flags & FOF_NOUPDATE) == 0)
		fp->f_offset = val;
	if ((flags & FOF_NEXTOFF_R) != 0)
		fp->f_nextoff[UIO_READ] = val;
	if ((flags & FOF_NEXTOFF_W) != 0)
		fp->f_nextoff[UIO_WRITE] = val;
	if ((flags & FOF_NOLOCK) == 0) {
		KASSERT((fp->f_vnread_flags & FOFFSET_LOCKED) != 0,
		    ("Lost FOFFSET_LOCKED"));
		if (fp->f_vnread_flags & FOFFSET_LOCK_WAITING)
			wakeup(&fp->f_vnread_flags);
		fp->f_vnread_flags = 0;
	}
	mtx_unlock(mtxp);
}
#endif

void
foffset_lock_uio(struct file *fp, struct uio *uio, int flags)
{

	if ((flags & FOF_OFFSET) == 0)
		uio->uio_offset = foffset_lock(fp, flags);
}

void
foffset_unlock_uio(struct file *fp, struct uio *uio, int flags)
{

	if ((flags & FOF_OFFSET) == 0)
		foffset_unlock(fp, uio->uio_offset, flags);
}

static int
get_advice(struct file *fp, struct uio *uio)
{
	struct mtx *mtxp;
	int ret;

	ret = POSIX_FADV_NORMAL;
	if (fp->f_advice == NULL || fp->f_vnode->v_type != VREG)
		return (ret);

	mtxp = mtx_pool_find(mtxpool_sleep, fp);
	mtx_lock(mtxp);
	if (fp->f_advice != NULL &&
	    uio->uio_offset >= fp->f_advice->fa_start &&
	    uio->uio_offset + uio->uio_resid <= fp->f_advice->fa_end)
		ret = fp->f_advice->fa_advice;
	mtx_unlock(mtxp);
	return (ret);
}

int
vn_read_from_obj(struct vnode *vp, struct uio *uio)
{
	vm_object_t obj;
	vm_page_t ma[io_hold_cnt + 2];
	off_t off, vsz;
	ssize_t resid;
	int error, i, j;

	MPASS(uio->uio_resid <= ptoa(io_hold_cnt + 2));
	obj = atomic_load_ptr(&vp->v_object);
	if (obj == NULL)
		return (EJUSTRETURN);

	/*
	 * Depends on type stability of vm_objects.
	 */
	vm_object_pip_add(obj, 1);
	if ((obj->flags & OBJ_DEAD) != 0) {
		/*
		 * Note that object might be already reused from the
		 * vnode, and the OBJ_DEAD flag cleared.  This is fine,
		 * we recheck for DOOMED vnode state after all pages
		 * are busied, and retract then.
		 *
		 * But we check for OBJ_DEAD to ensure that we do not
		 * busy pages while vm_object_terminate_pages()
		 * processes the queue.
		 */
		error = EJUSTRETURN;
		goto out_pip;
	}

	resid = uio->uio_resid;
	off = uio->uio_offset;
	for (i = 0; resid > 0; i++) {
		MPASS(i < io_hold_cnt + 2);
		ma[i] = vm_page_grab_unlocked(obj, atop(off),
		    VM_ALLOC_NOCREAT | VM_ALLOC_SBUSY | VM_ALLOC_IGN_SBUSY |
		    VM_ALLOC_NOWAIT);
		if (ma[i] == NULL)
			break;

		/*
		 * Skip invalid pages.  Valid mask can be partial only
		 * at EOF, and we clip later.
		 */
		if (vm_page_none_valid(ma[i])) {
			vm_page_sunbusy(ma[i]);
			break;
		}

		resid -= PAGE_SIZE;
		off += PAGE_SIZE;
	}
	if (i == 0) {
		error = EJUSTRETURN;
		goto out_pip;
	}

	/*
	 * Check VIRF_DOOMED after we busied our pages.  Since
	 * vgonel() terminates the vnode' vm_object, it cannot
	 * process past pages busied by us.
	 */
	if (VN_IS_DOOMED(vp)) {
		error = EJUSTRETURN;
		goto out;
	}

	resid = PAGE_SIZE - (uio->uio_offset & PAGE_MASK) + ptoa(i - 1);
	if (resid > uio->uio_resid)
		resid = uio->uio_resid;

	/*
	 * Unlocked read of vnp_size is safe because truncation cannot
	 * pass busied page.  But we load vnp_size into a local
	 * variable so that possible concurrent extension does not
	 * break calculation.
	 */
#if defined(__powerpc__) && !defined(__powerpc64__)
	vsz = obj->un_pager.vnp.vnp_size;
#else
	vsz = atomic_load_64(&obj->un_pager.vnp.vnp_size);
#endif
	if (uio->uio_offset >= vsz) {
		error = EJUSTRETURN;
		goto out;
	}
	if (uio->uio_offset + resid > vsz)
		resid = vsz - uio->uio_offset;

	error = vn_io_fault_pgmove(ma, uio->uio_offset & PAGE_MASK, resid, uio);

out:
	for (j = 0; j < i; j++) {
		if (error == 0)
			vm_page_reference(ma[j]);
		vm_page_sunbusy(ma[j]);
	}
out_pip:
	vm_object_pip_wakeup(obj);
	if (error != 0)
		return (error);
	return (uio->uio_resid == 0 ? 0 : EJUSTRETURN);
}

/*
 * File table vnode read routine.
 */
static int
vn_read(struct file *fp, struct uio *uio, struct ucred *active_cred, int flags,
    struct thread *td)
{
	struct vnode *vp;
	off_t orig_offset;
	int error, ioflag;
	int advice;

	KASSERT(uio->uio_td == td, ("uio_td %p is not td %p",
	    uio->uio_td, td));
	KASSERT(flags & FOF_OFFSET, ("No FOF_OFFSET"));
	vp = fp->f_vnode;
	ioflag = 0;
	if (fp->f_flag & FNONBLOCK)
		ioflag |= IO_NDELAY;
	if (fp->f_flag & O_DIRECT)
		ioflag |= IO_DIRECT;

	/*
	 * Try to read from page cache.  VIRF_DOOMED check is racy but
	 * allows us to avoid unneeded work outright.
	 */
	if (vn_io_pgcache_read_enable && !mac_vnode_check_read_enabled() &&
	    (vn_irflag_read(vp) & (VIRF_DOOMED | VIRF_PGREAD)) == VIRF_PGREAD) {
		error = VOP_READ_PGCACHE(vp, uio, ioflag, fp->f_cred);
		if (error == 0) {
			fp->f_nextoff[UIO_READ] = uio->uio_offset;
			return (0);
		}
		if (error != EJUSTRETURN)
			return (error);
	}

	advice = get_advice(fp, uio);
	vn_lock(vp, LK_SHARED | LK_RETRY);

	switch (advice) {
	case POSIX_FADV_NORMAL:
	case POSIX_FADV_SEQUENTIAL:
	case POSIX_FADV_NOREUSE:
		ioflag |= sequential_heuristic(uio, fp);
		break;
	case POSIX_FADV_RANDOM:
		/* Disable read-ahead for random I/O. */
		break;
	}
	orig_offset = uio->uio_offset;

#ifdef MAC
	error = mac_vnode_check_read(active_cred, fp->f_cred, vp);
	if (error == 0)
#endif
		error = VOP_READ(vp, uio, ioflag, fp->f_cred);
	fp->f_nextoff[UIO_READ] = uio->uio_offset;
	VOP_UNLOCK(vp);
	if (error == 0 && advice == POSIX_FADV_NOREUSE &&
	    orig_offset != uio->uio_offset)
		/*
		 * Use POSIX_FADV_DONTNEED to flush pages and buffers
		 * for the backing file after a POSIX_FADV_NOREUSE
		 * read(2).
		 */
		error = VOP_ADVISE(vp, orig_offset, uio->uio_offset - 1,
		    POSIX_FADV_DONTNEED);
	return (error);
}

/*
 * File table vnode write routine.
 */
static int
vn_write(struct file *fp, struct uio *uio, struct ucred *active_cred, int flags,
    struct thread *td)
{
	struct vnode *vp;
	struct mount *mp;
	off_t orig_offset;
	int error, ioflag, lock_flags;
	int advice;

	KASSERT(uio->uio_td == td, ("uio_td %p is not td %p",
	    uio->uio_td, td));
	KASSERT(flags & FOF_OFFSET, ("No FOF_OFFSET"));
	vp = fp->f_vnode;
	if (vp->v_type == VREG)
		bwillwrite();
	ioflag = IO_UNIT;
	if (vp->v_type == VREG && (fp->f_flag & O_APPEND))
		ioflag |= IO_APPEND;
	if (fp->f_flag & FNONBLOCK)
		ioflag |= IO_NDELAY;
	if (fp->f_flag & O_DIRECT)
		ioflag |= IO_DIRECT;
	if ((fp->f_flag & O_FSYNC) ||
	    (vp->v_mount && (vp->v_mount->mnt_flag & MNT_SYNCHRONOUS)))
		ioflag |= IO_SYNC;
	/*
	 * For O_DSYNC we set both IO_SYNC and IO_DATASYNC, so that VOP_WRITE()
	 * implementations that don't understand IO_DATASYNC fall back to full
	 * O_SYNC behavior.
	 */
	if (fp->f_flag & O_DSYNC)
		ioflag |= IO_SYNC | IO_DATASYNC;
	mp = NULL;
	if (vp->v_type != VCHR &&
	    (error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0)
		goto unlock;

	advice = get_advice(fp, uio);

	if (MNT_SHARED_WRITES(mp) ||
	    (mp == NULL && MNT_SHARED_WRITES(vp->v_mount))) {
		lock_flags = LK_SHARED;
	} else {
		lock_flags = LK_EXCLUSIVE;
	}

	vn_lock(vp, lock_flags | LK_RETRY);
	switch (advice) {
	case POSIX_FADV_NORMAL:
	case POSIX_FADV_SEQUENTIAL:
	case POSIX_FADV_NOREUSE:
		ioflag |= sequential_heuristic(uio, fp);
		break;
	case POSIX_FADV_RANDOM:
		/* XXX: Is this correct? */
		break;
	}
	orig_offset = uio->uio_offset;

#ifdef MAC
	error = mac_vnode_check_write(active_cred, fp->f_cred, vp);
	if (error == 0)
#endif
		error = VOP_WRITE(vp, uio, ioflag, fp->f_cred);
	fp->f_nextoff[UIO_WRITE] = uio->uio_offset;
	VOP_UNLOCK(vp);
	if (vp->v_type != VCHR)
		vn_finished_write(mp);
	if (error == 0 && advice == POSIX_FADV_NOREUSE &&
	    orig_offset != uio->uio_offset)
		/*
		 * Use POSIX_FADV_DONTNEED to flush pages and buffers
		 * for the backing file after a POSIX_FADV_NOREUSE
		 * write(2).
		 */
		error = VOP_ADVISE(vp, orig_offset, uio->uio_offset - 1,
		    POSIX_FADV_DONTNEED);
unlock:
	return (error);
}

/*
 * The vn_io_fault() is a wrapper around vn_read() and vn_write() to
 * prevent the following deadlock:
 *
 * Assume that the thread A reads from the vnode vp1 into userspace
 * buffer buf1 backed by the pages of vnode vp2.  If a page in buf1 is
 * currently not resident, then system ends up with the call chain
 *   vn_read() -> VOP_READ(vp1) -> uiomove() -> [Page Fault] ->
 *     vm_fault(buf1) -> vnode_pager_getpages(vp2) -> VOP_GETPAGES(vp2)
 * which establishes lock order vp1->vn_lock, then vp2->vn_lock.
 * If, at the same time, thread B reads from vnode vp2 into buffer buf2
 * backed by the pages of vnode vp1, and some page in buf2 is not
 * resident, we get a reversed order vp2->vn_lock, then vp1->vn_lock.
 *
 * To prevent the lock order reversal and deadlock, vn_io_fault() does
 * not allow page faults to happen during VOP_READ() or VOP_WRITE().
 * Instead, it first tries to do the whole range i/o with pagefaults
 * disabled. If all pages in the i/o buffer are resident and mapped,
 * VOP will succeed (ignoring the genuine filesystem errors).
 * Otherwise, we get back EFAULT, and vn_io_fault() falls back to do
 * i/o in chunks, with all pages in the chunk prefaulted and held
 * using vm_fault_quick_hold_pages().
 *
 * Filesystems using this deadlock avoidance scheme should use the
 * array of the held pages from uio, saved in the curthread->td_ma,
 * instead of doing uiomove().  A helper function
 * vn_io_fault_uiomove() converts uiomove request into
 * uiomove_fromphys() over td_ma array.
 *
 * Since vnode locks do not cover the whole i/o anymore, rangelocks
 * make the current i/o request atomic with respect to other i/os and
 * truncations.
 */

/*
 * Decode vn_io_fault_args and perform the corresponding i/o.
 */
static int
vn_io_fault_doio(struct vn_io_fault_args *args, struct uio *uio,
    struct thread *td)
{
	int error, save;

	error = 0;
	save = vm_fault_disable_pagefaults();
	switch (args->kind) {
	case VN_IO_FAULT_FOP:
		error = (args->args.fop_args.doio)(args->args.fop_args.fp,
		    uio, args->cred, args->flags, td);
		break;
	case VN_IO_FAULT_VOP:
		if (uio->uio_rw == UIO_READ) {
			error = VOP_READ(args->args.vop_args.vp, uio,
			    args->flags, args->cred);
		} else if (uio->uio_rw == UIO_WRITE) {
			error = VOP_WRITE(args->args.vop_args.vp, uio,
			    args->flags, args->cred);
		}
		break;
	default:
		panic("vn_io_fault_doio: unknown kind of io %d %d",
		    args->kind, uio->uio_rw);
	}
	vm_fault_enable_pagefaults(save);
	return (error);
}

static int
vn_io_fault_touch(char *base, const struct uio *uio)
{
	int r;

	r = fubyte(base);
	if (r == -1 || (uio->uio_rw == UIO_READ && subyte(base, r) == -1))
		return (EFAULT);
	return (0);
}

static int
vn_io_fault_prefault_user(const struct uio *uio)
{
	char *base;
	const struct iovec *iov;
	size_t len;
	ssize_t resid;
	int error, i;

	KASSERT(uio->uio_segflg == UIO_USERSPACE,
	    ("vn_io_fault_prefault userspace"));

	error = i = 0;
	iov = uio->uio_iov;
	resid = uio->uio_resid;
	base = iov->iov_base;
	len = iov->iov_len;
	while (resid > 0) {
		error = vn_io_fault_touch(base, uio);
		if (error != 0)
			break;
		if (len < PAGE_SIZE) {
			if (len != 0) {
				error = vn_io_fault_touch(base + len - 1, uio);
				if (error != 0)
					break;
				resid -= len;
			}
			if (++i >= uio->uio_iovcnt)
				break;
			iov = uio->uio_iov + i;
			base = iov->iov_base;
			len = iov->iov_len;
		} else {
			len -= PAGE_SIZE;
			base += PAGE_SIZE;
			resid -= PAGE_SIZE;
		}
	}
	return (error);
}

/*
 * Common code for vn_io_fault(), agnostic to the kind of i/o request.
 * Uses vn_io_fault_doio() to make the call to an actual i/o function.
 * Used from vn_rdwr() and vn_io_fault(), which encode the i/o request
 * into args and call vn_io_fault1() to handle faults during the user
 * mode buffer accesses.
 */
static int
vn_io_fault1(struct vnode *vp, struct uio *uio, struct vn_io_fault_args *args,
    struct thread *td)
{
	vm_page_t ma[io_hold_cnt + 2];
	struct uio *uio_clone, short_uio;
	struct iovec short_iovec[1];
	vm_page_t *prev_td_ma;
	vm_prot_t prot;
	vm_offset_t addr, end;
	size_t len, resid;
	ssize_t adv;
	int error, cnt, saveheld, prev_td_ma_cnt;

	if (vn_io_fault_prefault) {
		error = vn_io_fault_prefault_user(uio);
		if (error != 0)
			return (error); /* Or ignore ? */
	}

	prot = uio->uio_rw == UIO_READ ? VM_PROT_WRITE : VM_PROT_READ;

	/*
	 * The UFS follows IO_UNIT directive and replays back both
	 * uio_offset and uio_resid if an error is encountered during the
	 * operation.  But, since the iovec may be already advanced,
	 * uio is still in an inconsistent state.
	 *
	 * Cache a copy of the original uio, which is advanced to the redo
	 * point using UIO_NOCOPY below.
	 */
	uio_clone = cloneuio(uio);
	resid = uio->uio_resid;

	short_uio.uio_segflg = UIO_USERSPACE;
	short_uio.uio_rw = uio->uio_rw;
	short_uio.uio_td = uio->uio_td;

	error = vn_io_fault_doio(args, uio, td);
	if (error != EFAULT)
		goto out;

	atomic_add_long(&vn_io_faults_cnt, 1);
	uio_clone->uio_segflg = UIO_NOCOPY;
	uiomove(NULL, resid - uio->uio_resid, uio_clone);
	uio_clone->uio_segflg = uio->uio_segflg;

	saveheld = curthread_pflags_set(TDP_UIOHELD);
	prev_td_ma = td->td_ma;
	prev_td_ma_cnt = td->td_ma_cnt;

	while (uio_clone->uio_resid != 0) {
		len = uio_clone->uio_iov->iov_len;
		if (len == 0) {
			KASSERT(uio_clone->uio_iovcnt >= 1,
			    ("iovcnt underflow"));
			uio_clone->uio_iov++;
			uio_clone->uio_iovcnt--;
			continue;
		}
		if (len > ptoa(io_hold_cnt))
			len = ptoa(io_hold_cnt);
		addr = (uintptr_t)uio_clone->uio_iov->iov_base;
		end = round_page(addr + len);
		if (end < addr) {
			error = EFAULT;
			break;
		}
		cnt = atop(end - trunc_page(addr));
		/*
		 * A perfectly misaligned address and length could cause
		 * both the start and the end of the chunk to use partial
		 * page.  +2 accounts for such a situation.
		 */
		cnt = vm_fault_quick_hold_pages(&td->td_proc->p_vmspace->vm_map,
		    addr, len, prot, ma, io_hold_cnt + 2);
		if (cnt == -1) {
			error = EFAULT;
			break;
		}
		short_uio.uio_iov = &short_iovec[0];
		short_iovec[0].iov_base = (void *)addr;
		short_uio.uio_iovcnt = 1;
		short_uio.uio_resid = short_iovec[0].iov_len = len;
		short_uio.uio_offset = uio_clone->uio_offset;
		td->td_ma = ma;
		td->td_ma_cnt = cnt;

		error = vn_io_fault_doio(args, &short_uio, td);
		vm_page_unhold_pages(ma, cnt);
		adv = len - short_uio.uio_resid;

		uio_clone->uio_iov->iov_base =
		    (char *)uio_clone->uio_iov->iov_base + adv;
		uio_clone->uio_iov->iov_len -= adv;
		uio_clone->uio_resid -= adv;
		uio_clone->uio_offset += adv;

		uio->uio_resid -= adv;
		uio->uio_offset += adv;

		if (error != 0 || adv == 0)
			break;
	}
	td->td_ma = prev_td_ma;
	td->td_ma_cnt = prev_td_ma_cnt;
	curthread_pflags_restore(saveheld);
out:
	free(uio_clone, M_IOV);
	return (error);
}

static int
vn_io_fault(struct file *fp, struct uio *uio, struct ucred *active_cred,
    int flags, struct thread *td)
{
	fo_rdwr_t *doio;
	struct vnode *vp;
	void *rl_cookie;
	struct vn_io_fault_args args;
	int error;

	doio = uio->uio_rw == UIO_READ ? vn_read : vn_write;
	vp = fp->f_vnode;

	/*
	 * The ability to read(2) on a directory has historically been
	 * allowed for all users, but this can and has been the source of
	 * at least one security issue in the past.  As such, it is now hidden
	 * away behind a sysctl for those that actually need it to use it, and
	 * restricted to root when it's turned on to make it relatively safe to
	 * leave on for longer sessions of need.
	 */
	if (vp->v_type == VDIR) {
		KASSERT(uio->uio_rw == UIO_READ,
		    ("illegal write attempted on a directory"));
		if (!vfs_allow_read_dir)
			return (EISDIR);
		if ((error = priv_check(td, PRIV_VFS_READ_DIR)) != 0)
			return (EISDIR);
	}

	foffset_lock_uio(fp, uio, flags);
	if (do_vn_io_fault(vp, uio)) {
		args.kind = VN_IO_FAULT_FOP;
		args.args.fop_args.fp = fp;
		args.args.fop_args.doio = doio;
		args.cred = active_cred;
		args.flags = flags | FOF_OFFSET;
		if (uio->uio_rw == UIO_READ) {
			rl_cookie = vn_rangelock_rlock(vp, uio->uio_offset,
			    uio->uio_offset + uio->uio_resid);
		} else if ((fp->f_flag & O_APPEND) != 0 ||
		    (flags & FOF_OFFSET) == 0) {
			/* For appenders, punt and lock the whole range. */
			rl_cookie = vn_rangelock_wlock(vp, 0, OFF_MAX);
		} else {
			rl_cookie = vn_rangelock_wlock(vp, uio->uio_offset,
			    uio->uio_offset + uio->uio_resid);
		}
		error = vn_io_fault1(vp, uio, &args, td);
		vn_rangelock_unlock(vp, rl_cookie);
	} else {
		error = doio(fp, uio, active_cred, flags | FOF_OFFSET, td);
	}
	foffset_unlock_uio(fp, uio, flags);
	return (error);
}

/*
 * Helper function to perform the requested uiomove operation using
 * the held pages for io->uio_iov[0].iov_base buffer instead of
 * copyin/copyout.  Access to the pages with uiomove_fromphys()
 * instead of iov_base prevents page faults that could occur due to
 * pmap_collect() invalidating the mapping created by
 * vm_fault_quick_hold_pages(), or pageout daemon, page laundry or
 * object cleanup revoking the write access from page mappings.
 *
 * Filesystems specified MNTK_NO_IOPF shall use vn_io_fault_uiomove()
 * instead of plain uiomove().
 */
int
vn_io_fault_uiomove(char *data, int xfersize, struct uio *uio)
{
	struct uio transp_uio;
	struct iovec transp_iov[1];
	struct thread *td;
	size_t adv;
	int error, pgadv;

	td = curthread;
	if ((td->td_pflags & TDP_UIOHELD) == 0 ||
	    uio->uio_segflg != UIO_USERSPACE)
		return (uiomove(data, xfersize, uio));

	KASSERT(uio->uio_iovcnt == 1, ("uio_iovcnt %d", uio->uio_iovcnt));
	transp_iov[0].iov_base = data;
	transp_uio.uio_iov = &transp_iov[0];
	transp_uio.uio_iovcnt = 1;
	if (xfersize > uio->uio_resid)
		xfersize = uio->uio_resid;
	transp_uio.uio_resid = transp_iov[0].iov_len = xfersize;
	transp_uio.uio_offset = 0;
	transp_uio.uio_segflg = UIO_SYSSPACE;
	/*
	 * Since transp_iov points to data, and td_ma page array
	 * corresponds to original uio->uio_iov, we need to invert the
	 * direction of the i/o operation as passed to
	 * uiomove_fromphys().
	 */
	switch (uio->uio_rw) {
	case UIO_WRITE:
		transp_uio.uio_rw = UIO_READ;
		break;
	case UIO_READ:
		transp_uio.uio_rw = UIO_WRITE;
		break;
	}
	transp_uio.uio_td = uio->uio_td;
	error = uiomove_fromphys(td->td_ma,
	    ((vm_offset_t)uio->uio_iov->iov_base) & PAGE_MASK,
	    xfersize, &transp_uio);
	adv = xfersize - transp_uio.uio_resid;
	pgadv =
	    (((vm_offset_t)uio->uio_iov->iov_base + adv) >> PAGE_SHIFT) -
	    (((vm_offset_t)uio->uio_iov->iov_base) >> PAGE_SHIFT);
	td->td_ma += pgadv;
	KASSERT(td->td_ma_cnt >= pgadv, ("consumed pages %d %d", td->td_ma_cnt,
	    pgadv));
	td->td_ma_cnt -= pgadv;
	uio->uio_iov->iov_base = (char *)uio->uio_iov->iov_base + adv;
	uio->uio_iov->iov_len -= adv;
	uio->uio_resid -= adv;
	uio->uio_offset += adv;
	return (error);
}

int
vn_io_fault_pgmove(vm_page_t ma[], vm_offset_t offset, int xfersize,
    struct uio *uio)
{
	struct thread *td;
	vm_offset_t iov_base;
	int cnt, pgadv;

	td = curthread;
	if ((td->td_pflags & TDP_UIOHELD) == 0 ||
	    uio->uio_segflg != UIO_USERSPACE)
		return (uiomove_fromphys(ma, offset, xfersize, uio));

	KASSERT(uio->uio_iovcnt == 1, ("uio_iovcnt %d", uio->uio_iovcnt));
	cnt = xfersize > uio->uio_resid ? uio->uio_resid : xfersize;
	iov_base = (vm_offset_t)uio->uio_iov->iov_base;
	switch (uio->uio_rw) {
	case UIO_WRITE:
		pmap_copy_pages(td->td_ma, iov_base & PAGE_MASK, ma,
		    offset, cnt);
		break;
	case UIO_READ:
		pmap_copy_pages(ma, offset, td->td_ma, iov_base & PAGE_MASK,
		    cnt);
		break;
	}
	pgadv = ((iov_base + cnt) >> PAGE_SHIFT) - (iov_base >> PAGE_SHIFT);
	td->td_ma += pgadv;
	KASSERT(td->td_ma_cnt >= pgadv, ("consumed pages %d %d", td->td_ma_cnt,
	    pgadv));
	td->td_ma_cnt -= pgadv;
	uio->uio_iov->iov_base = (char *)(iov_base + cnt);
	uio->uio_iov->iov_len -= cnt;
	uio->uio_resid -= cnt;
	uio->uio_offset += cnt;
	return (0);
}

/*
 * File table truncate routine.
 */
static int
vn_truncate(struct file *fp, off_t length, struct ucred *active_cred,
    struct thread *td)
{
	struct mount *mp;
	struct vnode *vp;
	void *rl_cookie;
	int error;

	vp = fp->f_vnode;

retry:
	/*
	 * Lock the whole range for truncation.  Otherwise split i/o
	 * might happen partly before and partly after the truncation.
	 */
	rl_cookie = vn_rangelock_wlock(vp, 0, OFF_MAX);
	error = vn_start_write(vp, &mp, V_WAIT | PCATCH);
	if (error)
		goto out1;
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
	AUDIT_ARG_VNODE1(vp);
	if (vp->v_type == VDIR) {
		error = EISDIR;
		goto out;
	}
#ifdef MAC
	error = mac_vnode_check_write(active_cred, fp->f_cred, vp);
	if (error)
		goto out;
#endif
	error = vn_truncate_locked(vp, length, (fp->f_flag & O_FSYNC) != 0,
	    fp->f_cred);
out:
	VOP_UNLOCK(vp);
	vn_finished_write(mp);
out1:
	vn_rangelock_unlock(vp, rl_cookie);
	if (error == ERELOOKUP)
		goto retry;
	return (error);
}

/*
 * Truncate a file that is already locked.
 */
int
vn_truncate_locked(struct vnode *vp, off_t length, bool sync,
    struct ucred *cred)
{
	struct vattr vattr;
	int error;

	error = VOP_ADD_WRITECOUNT(vp, 1);
	if (error == 0) {
		VATTR_NULL(&vattr);
		vattr.va_size = length;
		if (sync)
			vattr.va_vaflags |= VA_SYNC;
		error = VOP_SETATTR(vp, &vattr, cred);
		VOP_ADD_WRITECOUNT_CHECKED(vp, -1);
	}
	return (error);
}

/*
 * File table vnode stat routine.
 */
int
vn_statfile(struct file *fp, struct stat *sb, struct ucred *active_cred,
    struct thread *td)
{
	struct vnode *vp = fp->f_vnode;
	int error;

	vn_lock(vp, LK_SHARED | LK_RETRY);
	error = VOP_STAT(vp, sb, active_cred, fp->f_cred, td);
	VOP_UNLOCK(vp);

	return (error);
}

/*
 * File table vnode ioctl routine.
 */
static int
vn_ioctl(struct file *fp, u_long com, void *data, struct ucred *active_cred,
    struct thread *td)
{
	struct vattr vattr;
	struct vnode *vp;
	struct fiobmap2_arg *bmarg;
	int error;

	vp = fp->f_vnode;
	switch (vp->v_type) {
	case VDIR:
	case VREG:
		switch (com) {
		case FIONREAD:
			vn_lock(vp, LK_SHARED | LK_RETRY);
			error = VOP_GETATTR(vp, &vattr, active_cred);
			VOP_UNLOCK(vp);
			if (error == 0)
				*(int *)data = vattr.va_size - fp->f_offset;
			return (error);
		case FIOBMAP2:
			bmarg = (struct fiobmap2_arg *)data;
			vn_lock(vp, LK_SHARED | LK_RETRY);
#ifdef MAC
			error = mac_vnode_check_read(active_cred, fp->f_cred,
			    vp);
			if (error == 0)
#endif
				error = VOP_BMAP(vp, bmarg->bn, NULL,
				    &bmarg->bn, &bmarg->runp, &bmarg->runb);
			VOP_UNLOCK(vp);
			return (error);
		case FIONBIO:
		case FIOASYNC:
			return (0);
		default:
			return (VOP_IOCTL(vp, com, data, fp->f_flag,
			    active_cred, td));
		}
		break;
	case VCHR:
		return (VOP_IOCTL(vp, com, data, fp->f_flag,
		    active_cred, td));
	default:
		return (ENOTTY);
	}
}

/*
 * File table vnode poll routine.
 */
static int
vn_poll(struct file *fp, int events, struct ucred *active_cred,
    struct thread *td)
{
	struct vnode *vp;
	int error;

	vp = fp->f_vnode;
#if defined(MAC) || defined(AUDIT)
	if (AUDITING_TD(td) || mac_vnode_check_poll_enabled()) {
		vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
		AUDIT_ARG_VNODE1(vp);
		error = mac_vnode_check_poll(active_cred, fp->f_cred, vp);
		VOP_UNLOCK(vp);
		if (error != 0)
			return (error);
	}
#endif
	error = VOP_POLL(vp, events, fp->f_cred, td);
	return (error);
}

/*
 * Acquire the requested lock and then check for validity.  LK_RETRY
 * permits vn_lock to return doomed vnodes.
 */
static int __noinline
_vn_lock_fallback(struct vnode *vp, int flags, const char *file, int line,
    int error)
{

	KASSERT((flags & LK_RETRY) == 0 || error == 0,
	    ("vn_lock: error %d incompatible with flags %#x", error, flags));

	if (error == 0)
		VNASSERT(VN_IS_DOOMED(vp), vp, ("vnode not doomed"));

	if ((flags & LK_RETRY) == 0) {
		if (error == 0) {
			VOP_UNLOCK(vp);
			error = ENOENT;
		}
		return (error);
	}

	/*
	 * LK_RETRY case.
	 *
	 * Nothing to do if we got the lock.
	 */
	if (error == 0)
		return (0);

	/*
	 * Interlock was dropped by the call in _vn_lock.
	 */
	flags &= ~LK_INTERLOCK;
	do {
		error = VOP_LOCK1(vp, flags, file, line);
	} while (error != 0);
	return (0);
}

int
_vn_lock(struct vnode *vp, int flags, const char *file, int line)
{
	int error;

	VNASSERT((flags & LK_TYPE_MASK) != 0, vp,
	    ("vn_lock: no locktype (%d passed)", flags));
	VNPASS(vp->v_holdcnt > 0, vp);
	error = VOP_LOCK1(vp, flags, file, line);
	if (__predict_false(error != 0 || VN_IS_DOOMED(vp)))
		return (_vn_lock_fallback(vp, flags, file, line, error));
	return (0);
}

/*
 * File table vnode close routine.
 */
static int
vn_closefile(struct file *fp, struct thread *td)
{
	struct vnode *vp;
	struct flock lf;
	int error;
	bool ref;

	vp = fp->f_vnode;
	fp->f_ops = &badfileops;
	ref = (fp->f_flag & FHASLOCK) != 0 && fp->f_type == DTYPE_VNODE;

	error = vn_close1(vp, fp->f_flag, fp->f_cred, td, ref);

	if (__predict_false(ref)) {
		lf.l_whence = SEEK_SET;
		lf.l_start = 0;
		lf.l_len = 0;
		lf.l_type = F_UNLCK;
		(void) VOP_ADVLOCK(vp, fp, F_UNLCK, &lf, F_FLOCK);
		vrele(vp);
	}
	return (error);
}

/*
 * Preparing to start a filesystem write operation. If the operation is
 * permitted, then we bump the count of operations in progress and
 * proceed. If a suspend request is in progress, we wait until the
 * suspension is over, and then proceed.
 */
static int
vn_start_write_refed(struct mount *mp, int flags, bool mplocked)
{
	struct mount_pcpu *mpcpu;
	int error, mflags;

	if (__predict_true(!mplocked) && (flags & V_XSLEEP) == 0 &&
	    vfs_op_thread_enter(mp, mpcpu)) {
		MPASS((mp->mnt_kern_flag & MNTK_SUSPEND) == 0);
		vfs_mp_count_add_pcpu(mpcpu, writeopcount, 1);
		vfs_op_thread_exit(mp, mpcpu);
		return (0);
	}

	if (mplocked)
		mtx_assert(MNT_MTX(mp), MA_OWNED);
	else
		MNT_ILOCK(mp);

	error = 0;

	/*
	 * Check on status of suspension.
	 */
	if ((curthread->td_pflags & TDP_IGNSUSP) == 0 ||
	    mp->mnt_susp_owner != curthread) {
		mflags = ((mp->mnt_vfc->vfc_flags & VFCF_SBDRY) != 0 ?
		    (flags & PCATCH) : 0) | (PUSER - 1);
		while ((mp->mnt_kern_flag & MNTK_SUSPEND) != 0) {
			if (flags & V_NOWAIT) {
				error = EWOULDBLOCK;
				goto unlock;
			}
			error = msleep(&mp->mnt_flag, MNT_MTX(mp), mflags,
			    "suspfs", 0);
			if (error)
				goto unlock;
		}
	}
	if (flags & V_XSLEEP)
		goto unlock;
	mp->mnt_writeopcount++;
unlock:
	if (error != 0 || (flags & V_XSLEEP) != 0)
		MNT_REL(mp);
	MNT_IUNLOCK(mp);
	return (error);
}

int
vn_start_write(struct vnode *vp, struct mount **mpp, int flags)
{
	struct mount *mp;
	int error;

	KASSERT((flags & V_MNTREF) == 0 || (*mpp != NULL && vp == NULL),
	    ("V_MNTREF requires mp"));

	error = 0;
	/*
	 * If a vnode is provided, get and return the mount point that
	 * to which it will write.
	 */
	if (vp != NULL) {
		if ((error = VOP_GETWRITEMOUNT(vp, mpp)) != 0) {
			*mpp = NULL;
			if (error != EOPNOTSUPP)
				return (error);
			return (0);
		}
	}
	if ((mp = *mpp) == NULL)
		return (0);

	/*
	 * VOP_GETWRITEMOUNT() returns with the mp refcount held through
	 * a vfs_ref().
	 * As long as a vnode is not provided we need to acquire a
	 * refcount for the provided mountpoint too, in order to
	 * emulate a vfs_ref().
	 */
	if (vp == NULL && (flags & V_MNTREF) == 0)
		vfs_ref(mp);

	return (vn_start_write_refed(mp, flags, false));
}

/*
 * Secondary suspension. Used by operations such as vop_inactive
 * routines that are needed by the higher level functions. These
 * are allowed to proceed until all the higher level functions have
 * completed (indicated by mnt_writeopcount dropping to zero). At that
 * time, these operations are halted until the suspension is over.
 */
int
vn_start_secondary_write(struct vnode *vp, struct mount **mpp, int flags)
{
	struct mount *mp;
	int error;

	KASSERT((flags & V_MNTREF) == 0 || (*mpp != NULL && vp == NULL),
	    ("V_MNTREF requires mp"));

 retry:
	if (vp != NULL) {
		if ((error = VOP_GETWRITEMOUNT(vp, mpp)) != 0) {
			*mpp = NULL;
			if (error != EOPNOTSUPP)
				return (error);
			return (0);
		}
	}
	/*
	 * If we are not suspended or have not yet reached suspended
	 * mode, then let the operation proceed.
	 */
	if ((mp = *mpp) == NULL)
		return (0);

	/*
	 * VOP_GETWRITEMOUNT() returns with the mp refcount held through
	 * a vfs_ref().
	 * As long as a vnode is not provided we need to acquire a
	 * refcount for the provided mountpoint too, in order to
	 * emulate a vfs_ref().
	 */
	MNT_ILOCK(mp);
	if (vp == NULL && (flags & V_MNTREF) == 0)
		MNT_REF(mp);
	if ((mp->mnt_kern_flag & (MNTK_SUSPENDED | MNTK_SUSPEND2)) == 0) {
		mp->mnt_secondary_writes++;
		mp->mnt_secondary_accwrites++;
		MNT_IUNLOCK(mp);
		return (0);
	}
	if (flags & V_NOWAIT) {
		MNT_REL(mp);
		MNT_IUNLOCK(mp);
		return (EWOULDBLOCK);
	}
	/*
	 * Wait for the suspension to finish.
	 */
	error = msleep(&mp->mnt_flag, MNT_MTX(mp), (PUSER - 1) | PDROP |
	    ((mp->mnt_vfc->vfc_flags & VFCF_SBDRY) != 0 ? (flags & PCATCH) : 0),
	    "suspfs", 0);
	vfs_rel(mp);
	if (error == 0)
		goto retry;
	return (error);
}

/*
 * Filesystem write operation has completed. If we are suspending and this
 * operation is the last one, notify the suspender that the suspension is
 * now in effect.
 */
void
vn_finished_write(struct mount *mp)
{
	struct mount_pcpu *mpcpu;
	int c;

	if (mp == NULL)
		return;

	if (vfs_op_thread_enter(mp, mpcpu)) {
		vfs_mp_count_sub_pcpu(mpcpu, writeopcount, 1);
		vfs_mp_count_sub_pcpu(mpcpu, ref, 1);
		vfs_op_thread_exit(mp, mpcpu);
		return;
	}

	MNT_ILOCK(mp);
	vfs_assert_mount_counters(mp);
	MNT_REL(mp);
	c = --mp->mnt_writeopcount;
	if (mp->mnt_vfs_ops == 0) {
		MPASS((mp->mnt_kern_flag & MNTK_SUSPEND) == 0);
		MNT_IUNLOCK(mp);
		return;
	}
	if (c < 0)
		vfs_dump_mount_counters(mp);
	if ((mp->mnt_kern_flag & MNTK_SUSPEND) != 0 && c == 0)
		wakeup(&mp->mnt_writeopcount);
	MNT_IUNLOCK(mp);
}

/*
 * Filesystem secondary write operation has completed. If we are
 * suspending and this operation is the last one, notify the suspender
 * that the suspension is now in effect.
 */
void
vn_finished_secondary_write(struct mount *mp)
{
	if (mp == NULL)
		return;
	MNT_ILOCK(mp);
	MNT_REL(mp);
	mp->mnt_secondary_writes--;
	if (mp->mnt_secondary_writes < 0)
		panic("vn_finished_secondary_write: neg cnt");
	if ((mp->mnt_kern_flag & MNTK_SUSPEND) != 0 &&
	    mp->mnt_secondary_writes <= 0)
		wakeup(&mp->mnt_secondary_writes);
	MNT_IUNLOCK(mp);
}

/*
 * Request a filesystem to suspend write operations.
 */
int
vfs_write_suspend(struct mount *mp, int flags)
{
	int error;

	vfs_op_enter(mp);

	MNT_ILOCK(mp);
	vfs_assert_mount_counters(mp);
	if (mp->mnt_susp_owner == curthread) {
		vfs_op_exit_locked(mp);
		MNT_IUNLOCK(mp);
		return (EALREADY);
	}
	while (mp->mnt_kern_flag & MNTK_SUSPEND)
		msleep(&mp->mnt_flag, MNT_MTX(mp), PUSER - 1, "wsuspfs", 0);

	/*
	 * Unmount holds a write reference on the mount point.  If we
	 * own busy reference and drain for writers, we deadlock with
	 * the reference draining in the unmount path.  Callers of
	 * vfs_write_suspend() must specify VS_SKIP_UNMOUNT if
	 * vfs_busy() reference is owned and caller is not in the
	 * unmount context.
	 */
	if ((flags & VS_SKIP_UNMOUNT) != 0 &&
	    (mp->mnt_kern_flag & MNTK_UNMOUNT) != 0) {
		vfs_op_exit_locked(mp);
		MNT_IUNLOCK(mp);
		return (EBUSY);
	}

	mp->mnt_kern_flag |= MNTK_SUSPEND;
	mp->mnt_susp_owner = curthread;
	if (mp->mnt_writeopcount > 0)
		(void) msleep(&mp->mnt_writeopcount, 
		    MNT_MTX(mp), (PUSER - 1)|PDROP, "suspwt", 0);
	else
		MNT_IUNLOCK(mp);
	if ((error = VFS_SYNC(mp, MNT_SUSPEND)) != 0) {
		vfs_write_resume(mp, 0);
		/* vfs_write_resume does vfs_op_exit() for us */
	}
	return (error);
}

/*
 * Request a filesystem to resume write operations.
 */
void
vfs_write_resume(struct mount *mp, int flags)
{

	MNT_ILOCK(mp);
	if ((mp->mnt_kern_flag & MNTK_SUSPEND) != 0) {
		KASSERT(mp->mnt_susp_owner == curthread, ("mnt_susp_owner"));
		mp->mnt_kern_flag &= ~(MNTK_SUSPEND | MNTK_SUSPEND2 |
				       MNTK_SUSPENDED);
		mp->mnt_susp_owner = NULL;
		wakeup(&mp->mnt_writeopcount);
		wakeup(&mp->mnt_flag);
		curthread->td_pflags &= ~TDP_IGNSUSP;
		if ((flags & VR_START_WRITE) != 0) {
			MNT_REF(mp);
			mp->mnt_writeopcount++;
		}
		MNT_IUNLOCK(mp);
		if ((flags & VR_NO_SUSPCLR) == 0)
			VFS_SUSP_CLEAN(mp);
		vfs_op_exit(mp);
	} else if ((flags & VR_START_WRITE) != 0) {
		MNT_REF(mp);
		vn_start_write_refed(mp, 0, true);
	} else {
		MNT_IUNLOCK(mp);
	}
}

/*
 * Helper loop around vfs_write_suspend() for filesystem unmount VFS
 * methods.
 */
int
vfs_write_suspend_umnt(struct mount *mp)
{
	int error;

	KASSERT((curthread->td_pflags & TDP_IGNSUSP) == 0,
	    ("vfs_write_suspend_umnt: recursed"));

	/* dounmount() already called vn_start_write(). */
	for (;;) {
		vn_finished_write(mp);
		error = vfs_write_suspend(mp, 0);
		if (error != 0) {
			vn_start_write(NULL, &mp, V_WAIT);
			return (error);
		}
		MNT_ILOCK(mp);
		if ((mp->mnt_kern_flag & MNTK_SUSPENDED) != 0)
			break;
		MNT_IUNLOCK(mp);
		vn_start_write(NULL, &mp, V_WAIT);
	}
	mp->mnt_kern_flag &= ~(MNTK_SUSPENDED | MNTK_SUSPEND2);
	wakeup(&mp->mnt_flag);
	MNT_IUNLOCK(mp);
	curthread->td_pflags |= TDP_IGNSUSP;
	return (0);
}

/*
 * Implement kqueues for files by translating it to vnode operation.
 */
static int
vn_kqfilter(struct file *fp, struct knote *kn)
{

	return (VOP_KQFILTER(fp->f_vnode, kn));
}

int
vn_kqfilter_opath(struct file *fp, struct knote *kn)
{
	if ((fp->f_flag & FKQALLOWED) == 0)
		return (EBADF);
	return (vn_kqfilter(fp, kn));
}

/*
 * Simplified in-kernel wrapper calls for extended attribute access.
 * Both calls pass in a NULL credential, authorizing as "kernel" access.
 * Set IO_NODELOCKED in ioflg if the vnode is already locked.
 */
int
vn_extattr_get(struct vnode *vp, int ioflg, int attrnamespace,
    const char *attrname, int *buflen, char *buf, struct thread *td)
{
	struct uio	auio;
	struct iovec	iov;
	int	error;

	iov.iov_len = *buflen;
	iov.iov_base = buf;

	auio.uio_iov = &iov;
	auio.uio_iovcnt = 1;
	auio.uio_rw = UIO_READ;
	auio.uio_segflg = UIO_SYSSPACE;
	auio.uio_td = td;
	auio.uio_offset = 0;
	auio.uio_resid = *buflen;

	if ((ioflg & IO_NODELOCKED) == 0)
		vn_lock(vp, LK_SHARED | LK_RETRY);

	ASSERT_VOP_LOCKED(vp, "IO_NODELOCKED with no vp lock held");

	/* authorize attribute retrieval as kernel */
	error = VOP_GETEXTATTR(vp, attrnamespace, attrname, &auio, NULL, NULL,
	    td);

	if ((ioflg & IO_NODELOCKED) == 0)
		VOP_UNLOCK(vp);

	if (error == 0) {
		*buflen = *buflen - auio.uio_resid;
	}

	return (error);
}

/*
 * XXX failure mode if partially written?
 */
int
vn_extattr_set(struct vnode *vp, int ioflg, int attrnamespace,
    const char *attrname, int buflen, char *buf, struct thread *td)
{
	struct uio	auio;
	struct iovec	iov;
	struct mount	*mp;
	int	error;

	iov.iov_len = buflen;
	iov.iov_base = buf;

	auio.uio_iov = &iov;
	auio.uio_iovcnt = 1;
	auio.uio_rw = UIO_WRITE;
	auio.uio_segflg = UIO_SYSSPACE;
	auio.uio_td = td;
	auio.uio_offset = 0;
	auio.uio_resid = buflen;

	if ((ioflg & IO_NODELOCKED) == 0) {
		if ((error = vn_start_write(vp, &mp, V_WAIT)) != 0)
			return (error);
		vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
	}

	ASSERT_VOP_LOCKED(vp, "IO_NODELOCKED with no vp lock held");

	/* authorize attribute setting as kernel */
	error = VOP_SETEXTATTR(vp, attrnamespace, attrname, &auio, NULL, td);

	if ((ioflg & IO_NODELOCKED) == 0) {
		vn_finished_write(mp);
		VOP_UNLOCK(vp);
	}

	return (error);
}

int
vn_extattr_rm(struct vnode *vp, int ioflg, int attrnamespace,
    const char *attrname, struct thread *td)
{
	struct mount	*mp;
	int	error;

	if ((ioflg & IO_NODELOCKED) == 0) {
		if ((error = vn_start_write(vp, &mp, V_WAIT)) != 0)
			return (error);
		vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
	}

	ASSERT_VOP_LOCKED(vp, "IO_NODELOCKED with no vp lock held");

	/* authorize attribute removal as kernel */
	error = VOP_DELETEEXTATTR(vp, attrnamespace, attrname, NULL, td);
	if (error == EOPNOTSUPP)
		error = VOP_SETEXTATTR(vp, attrnamespace, attrname, NULL,
		    NULL, td);

	if ((ioflg & IO_NODELOCKED) == 0) {
		vn_finished_write(mp);
		VOP_UNLOCK(vp);
	}

	return (error);
}

static int
vn_get_ino_alloc_vget(struct mount *mp, void *arg, int lkflags,
    struct vnode **rvp)
{

	return (VFS_VGET(mp, *(ino_t *)arg, lkflags, rvp));
}

int
vn_vget_ino(struct vnode *vp, ino_t ino, int lkflags, struct vnode **rvp)
{

	return (vn_vget_ino_gen(vp, vn_get_ino_alloc_vget, &ino,
	    lkflags, rvp));
}

int
vn_vget_ino_gen(struct vnode *vp, vn_get_ino_t alloc, void *alloc_arg,
    int lkflags, struct vnode **rvp)
{
	struct mount *mp;
	int ltype, error;

	ASSERT_VOP_LOCKED(vp, "vn_vget_ino_get");
	mp = vp->v_mount;
	ltype = VOP_ISLOCKED(vp);
	KASSERT(ltype == LK_EXCLUSIVE || ltype == LK_SHARED,
	    ("vn_vget_ino: vp not locked"));
	error = vfs_busy(mp, MBF_NOWAIT);
	if (error != 0) {
		vfs_ref(mp);
		VOP_UNLOCK(vp);
		error = vfs_busy(mp, 0);
		vn_lock(vp, ltype | LK_RETRY);
		vfs_rel(mp);
		if (error != 0)
			return (ENOENT);
		if (VN_IS_DOOMED(vp)) {
			vfs_unbusy(mp);
			return (ENOENT);
		}
	}
	VOP_UNLOCK(vp);
	error = alloc(mp, alloc_arg, lkflags, rvp);
	vfs_unbusy(mp);
	if (error != 0 || *rvp != vp)
		vn_lock(vp, ltype | LK_RETRY);
	if (VN_IS_DOOMED(vp)) {
		if (error == 0) {
			if (*rvp == vp)
				vunref(vp);
			else
				vput(*rvp);
		}
		error = ENOENT;
	}
	return (error);
}

int
vn_rlimit_fsize(const struct vnode *vp, const struct uio *uio,
    struct thread *td)
{

	if (vp->v_type != VREG || td == NULL)
		return (0);
	if ((uoff_t)uio->uio_offset + uio->uio_resid >
	    lim_cur(td, RLIMIT_FSIZE)) {
		PROC_LOCK(td->td_proc);
		kern_psignal(td->td_proc, SIGXFSZ);
		PROC_UNLOCK(td->td_proc);
		return (EFBIG);
	}
	return (0);
}

int
vn_chmod(struct file *fp, mode_t mode, struct ucred *active_cred,
    struct thread *td)
{
	struct vnode *vp;

	vp = fp->f_vnode;
#ifdef AUDIT
	vn_lock(vp, LK_SHARED | LK_RETRY);
	AUDIT_ARG_VNODE1(vp);
	VOP_UNLOCK(vp);
#endif
	return (setfmode(td, active_cred, vp, mode));
}

int
vn_chown(struct file *fp, uid_t uid, gid_t gid, struct ucred *active_cred,
    struct thread *td)
{
	struct vnode *vp;

	vp = fp->f_vnode;
#ifdef AUDIT
	vn_lock(vp, LK_SHARED | LK_RETRY);
	AUDIT_ARG_VNODE1(vp);
	VOP_UNLOCK(vp);
#endif
	return (setfown(td, active_cred, vp, uid, gid));
}

void
vn_pages_remove(struct vnode *vp, vm_pindex_t start, vm_pindex_t end)
{
	vm_object_t object;

	if ((object = vp->v_object) == NULL)
		return;
	VM_OBJECT_WLOCK(object);
	vm_object_page_remove(object, start, end, 0);
	VM_OBJECT_WUNLOCK(object);
}

int
vn_bmap_seekhole(struct vnode *vp, u_long cmd, off_t *off, struct ucred *cred)
{
	struct vattr va;
	daddr_t bn, bnp;
	uint64_t bsize;
	off_t noff;
	int error;

	KASSERT(cmd == FIOSEEKHOLE || cmd == FIOSEEKDATA,
	    ("Wrong command %lu", cmd));

	if (vn_lock(vp, LK_SHARED) != 0)
		return (EBADF);
	if (vp->v_type != VREG) {
		error = ENOTTY;
		goto unlock;
	}
	error = VOP_GETATTR(vp, &va, cred);
	if (error != 0)
		goto unlock;
	noff = *off;
	if (noff >= va.va_size) {
		error = ENXIO;
		goto unlock;
	}
	bsize = vp->v_mount->mnt_stat.f_iosize;
	for (bn = noff / bsize; noff < va.va_size; bn++, noff += bsize -
	    noff % bsize) {
		error = VOP_BMAP(vp, bn, NULL, &bnp, NULL, NULL);
		if (error == EOPNOTSUPP) {
			error = ENOTTY;
			goto unlock;
		}
		if ((bnp == -1 && cmd == FIOSEEKHOLE) ||
		    (bnp != -1 && cmd == FIOSEEKDATA)) {
			noff = bn * bsize;
			if (noff < *off)
				noff = *off;
			goto unlock;
		}
	}
	if (noff > va.va_size)
		noff = va.va_size;
	/* noff == va.va_size. There is an implicit hole at the end of file. */
	if (cmd == FIOSEEKDATA)
		error = ENXIO;
unlock:
	VOP_UNLOCK(vp);
	if (error == 0)
		*off = noff;
	return (error);
}

int
vn_seek(struct file *fp, off_t offset, int whence, struct thread *td)
{
	struct ucred *cred;
	struct vnode *vp;
	struct vattr vattr;
	off_t foffset, size;
	int error, noneg;

	cred = td->td_ucred;
	vp = fp->f_vnode;
	foffset = foffset_lock(fp, 0);
	noneg = (vp->v_type != VCHR);
	error = 0;
	switch (whence) {
	case L_INCR:
		if (noneg &&
		    (foffset < 0 ||
		    (offset > 0 && foffset > OFF_MAX - offset))) {
			error = EOVERFLOW;
			break;
		}
		offset += foffset;
		break;
	case L_XTND:
		vn_lock(vp, LK_SHARED | LK_RETRY);
		error = VOP_GETATTR(vp, &vattr, cred);
		VOP_UNLOCK(vp);
		if (error)
			break;

		/*
		 * If the file references a disk device, then fetch
		 * the media size and use that to determine the ending
		 * offset.
		 */
		if (vattr.va_size == 0 && vp->v_type == VCHR &&
		    fo_ioctl(fp, DIOCGMEDIASIZE, &size, cred, td) == 0)
			vattr.va_size = size;
		if (noneg &&
		    (vattr.va_size > OFF_MAX ||
		    (offset > 0 && vattr.va_size > OFF_MAX - offset))) {
			error = EOVERFLOW;
			break;
		}
		offset += vattr.va_size;
		break;
	case L_SET:
		break;
	case SEEK_DATA:
		error = fo_ioctl(fp, FIOSEEKDATA, &offset, cred, td);
		if (error == ENOTTY)
			error = EINVAL;
		break;
	case SEEK_HOLE:
		error = fo_ioctl(fp, FIOSEEKHOLE, &offset, cred, td);
		if (error == ENOTTY)
			error = EINVAL;
		break;
	default:
		error = EINVAL;
	}
	if (error == 0 && noneg && offset < 0)
		error = EINVAL;
	if (error != 0)
		goto drop;
	VFS_KNOTE_UNLOCKED(vp, 0);
	td->td_uretoff.tdu_off = offset;
drop:
	foffset_unlock(fp, offset, error != 0 ? FOF_NOUPDATE : 0);
	return (error);
}

int
vn_utimes_perm(struct vnode *vp, struct vattr *vap, struct ucred *cred,
    struct thread *td)
{
	int error;

	/*
	 * Grant permission if the caller is the owner of the file, or
	 * the super-user, or has ACL_WRITE_ATTRIBUTES permission on
	 * on the file.  If the time pointer is null, then write
	 * permission on the file is also sufficient.
	 *
	 * From NFSv4.1, draft 21, 6.2.1.3.1, Discussion of Mask Attributes:
	 * A user having ACL_WRITE_DATA or ACL_WRITE_ATTRIBUTES
	 * will be allowed to set the times [..] to the current
	 * server time.
	 */
	error = VOP_ACCESSX(vp, VWRITE_ATTRIBUTES, cred, td);
	if (error != 0 && (vap->va_vaflags & VA_UTIMES_NULL) != 0)
		error = VOP_ACCESS(vp, VWRITE, cred, td);
	return (error);
}

int
vn_fill_kinfo(struct file *fp, struct kinfo_file *kif, struct filedesc *fdp)
{
	struct vnode *vp;
	int error;

	if (fp->f_type == DTYPE_FIFO)
		kif->kf_type = KF_TYPE_FIFO;
	else
		kif->kf_type = KF_TYPE_VNODE;
	vp = fp->f_vnode;
	vref(vp);
	FILEDESC_SUNLOCK(fdp);
	error = vn_fill_kinfo_vnode(vp, kif);
	vrele(vp);
	FILEDESC_SLOCK(fdp);
	return (error);
}

static inline void
vn_fill_junk(struct kinfo_file *kif)
{
	size_t len, olen;

	/*
	 * Simulate vn_fullpath returning changing values for a given
	 * vp during e.g. coredump.
	 */
	len = (arc4random() % (sizeof(kif->kf_path) - 2)) + 1;
	olen = strlen(kif->kf_path);
	if (len < olen)
		strcpy(&kif->kf_path[len - 1], "$");
	else
		for (; olen < len; olen++)
			strcpy(&kif->kf_path[olen], "A");
}

int
vn_fill_kinfo_vnode(struct vnode *vp, struct kinfo_file *kif)
{
	struct vattr va;
	char *fullpath, *freepath;
	int error;

	kif->kf_un.kf_file.kf_file_type = vntype_to_kinfo(vp->v_type);
	freepath = NULL;
	fullpath = "-";
	error = vn_fullpath(vp, &fullpath, &freepath);
	if (error == 0) {
		strlcpy(kif->kf_path, fullpath, sizeof(kif->kf_path));
	}
	if (freepath != NULL)
		free(freepath, M_TEMP);

	KFAIL_POINT_CODE(DEBUG_FP, fill_kinfo_vnode__random_path,
		vn_fill_junk(kif);
	);

	/*
	 * Retrieve vnode attributes.
	 */
	va.va_fsid = VNOVAL;
	va.va_rdev = NODEV;
	vn_lock(vp, LK_SHARED | LK_RETRY);
	error = VOP_GETATTR(vp, &va, curthread->td_ucred);
	VOP_UNLOCK(vp);
	if (error != 0)
		return (error);
	if (va.va_fsid != VNOVAL)
		kif->kf_un.kf_file.kf_file_fsid = va.va_fsid;
	else
		kif->kf_un.kf_file.kf_file_fsid =
		    vp->v_mount->mnt_stat.f_fsid.val[0];
	kif->kf_un.kf_file.kf_file_fsid_freebsd11 =
	    kif->kf_un.kf_file.kf_file_fsid; /* truncate */
	kif->kf_un.kf_file.kf_file_fileid = va.va_fileid;
	kif->kf_un.kf_file.kf_file_mode = MAKEIMODE(va.va_type, va.va_mode);
	kif->kf_un.kf_file.kf_file_size = va.va_size;
	kif->kf_un.kf_file.kf_file_rdev = va.va_rdev;
	kif->kf_un.kf_file.kf_file_rdev_freebsd11 =
	    kif->kf_un.kf_file.kf_file_rdev; /* truncate */
	return (0);
}

int
vn_mmap(struct file *fp, vm_map_t map, vm_offset_t *addr, vm_size_t size,
    vm_prot_t prot, vm_prot_t cap_maxprot, int flags, vm_ooffset_t foff,
    struct thread *td)
{
#ifdef HWPMC_HOOKS
	struct pmckern_map_in pkm;
#endif
	struct mount *mp;
	struct vnode *vp;
	vm_object_t object;
	vm_prot_t maxprot;
	boolean_t writecounted;
	int error;

#if defined(COMPAT_FREEBSD7) || defined(COMPAT_FREEBSD6) || \
    defined(COMPAT_FREEBSD5) || defined(COMPAT_FREEBSD4)
	/*
	 * POSIX shared-memory objects are defined to have
	 * kernel persistence, and are not defined to support
	 * read(2)/write(2) -- or even open(2).  Thus, we can
	 * use MAP_ASYNC to trade on-disk coherence for speed.
	 * The shm_open(3) library routine turns on the FPOSIXSHM
	 * flag to request this behavior.
	 */
	if ((fp->f_flag & FPOSIXSHM) != 0)
		flags |= MAP_NOSYNC;
#endif
	vp = fp->f_vnode;

	/*
	 * Ensure that file and memory protections are
	 * compatible.  Note that we only worry about
	 * writability if mapping is shared; in this case,
	 * current and max prot are dictated by the open file.
	 * XXX use the vnode instead?  Problem is: what
	 * credentials do we use for determination? What if
	 * proc does a setuid?
	 */
	mp = vp->v_mount;
	if (mp != NULL && (mp->mnt_flag & MNT_NOEXEC) != 0) {
		maxprot = VM_PROT_NONE;
		if ((prot & VM_PROT_EXECUTE) != 0)
			return (EACCES);
	} else
		maxprot = VM_PROT_EXECUTE;
	if ((fp->f_flag & FREAD) != 0)
		maxprot |= VM_PROT_READ;
	else if ((prot & VM_PROT_READ) != 0)
		return (EACCES);

	/*
	 * If we are sharing potential changes via MAP_SHARED and we
	 * are trying to get write permission although we opened it
	 * without asking for it, bail out.
	 */
	if ((flags & MAP_SHARED) != 0) {
		if ((fp->f_flag & FWRITE) != 0)
			maxprot |= VM_PROT_WRITE;
		else if ((prot & VM_PROT_WRITE) != 0)
			return (EACCES);
	} else {
		maxprot |= VM_PROT_WRITE;
		cap_maxprot |= VM_PROT_WRITE;
	}
	maxprot &= cap_maxprot;

	/*
	 * For regular files and shared memory, POSIX requires that
	 * the value of foff be a legitimate offset within the data
	 * object.  In particular, negative offsets are invalid.
	 * Blocking negative offsets and overflows here avoids
	 * possible wraparound or user-level access into reserved
	 * ranges of the data object later.  In contrast, POSIX does
	 * not dictate how offsets are used by device drivers, so in
	 * the case of a device mapping a negative offset is passed
	 * on.
	 */
	if (
#ifdef _LP64
	    size > OFF_MAX ||
#endif
	    foff > OFF_MAX - size)
		return (EINVAL);

	writecounted = FALSE;
	error = vm_mmap_vnode(td, size, prot, &maxprot, &flags, vp,
	    &foff, &object, &writecounted);
	if (error != 0)
		return (error);
	error = vm_mmap_object(map, addr, size, prot, maxprot, flags, object,
	    foff, writecounted, td);
	if (error != 0) {
		/*
		 * If this mapping was accounted for in the vnode's
		 * writecount, then undo that now.
		 */
		if (writecounted)
			vm_pager_release_writecount(object, 0, size);
		vm_object_deallocate(object);
	}
#ifdef HWPMC_HOOKS
	/* Inform hwpmc(4) if an executable is being mapped. */
	if (PMC_HOOK_INSTALLED(PMC_FN_MMAP)) {
		if ((prot & VM_PROT_EXECUTE) != 0 && error == 0) {
			pkm.pm_file = vp;
			pkm.pm_address = (uintptr_t) *addr;
			PMC_CALL_HOOK_UNLOCKED(td, PMC_FN_MMAP, (void *) &pkm);
		}
	}
#endif
	return (error);
}

void
vn_fsid(struct vnode *vp, struct vattr *va)
{
	fsid_t *f;

	f = &vp->v_mount->mnt_stat.f_fsid;
	va->va_fsid = (uint32_t)f->val[1];
	va->va_fsid <<= sizeof(f->val[1]) * NBBY;
	va->va_fsid += (uint32_t)f->val[0];
}

int
vn_fsync_buf(struct vnode *vp, int waitfor)
{
	struct buf *bp, *nbp;
	struct bufobj *bo;
	struct mount *mp;
	int error, maxretry;

	error = 0;
	maxretry = 10000;     /* large, arbitrarily chosen */
	mp = NULL;
	if (vp->v_type == VCHR) {
		VI_LOCK(vp);
		mp = vp->v_rdev->si_mountpt;
		VI_UNLOCK(vp);
	}
	bo = &vp->v_bufobj;
	BO_LOCK(bo);
loop1:
	/*
	 * MARK/SCAN initialization to avoid infinite loops.
	 */
        TAILQ_FOREACH(bp, &bo->bo_dirty.bv_hd, b_bobufs) {
		bp->b_vflags &= ~BV_SCANNED;
		bp->b_error = 0;
	}

	/*
	 * Flush all dirty buffers associated with a vnode.
	 */
loop2:
	TAILQ_FOREACH_SAFE(bp, &bo->bo_dirty.bv_hd, b_bobufs, nbp) {
		if ((bp->b_vflags & BV_SCANNED) != 0)
			continue;
		bp->b_vflags |= BV_SCANNED;
		if (BUF_LOCK(bp, LK_EXCLUSIVE | LK_NOWAIT, NULL)) {
			if (waitfor != MNT_WAIT)
				continue;
			if (BUF_LOCK(bp,
			    LK_EXCLUSIVE | LK_INTERLOCK | LK_SLEEPFAIL,
			    BO_LOCKPTR(bo)) != 0) {
				BO_LOCK(bo);
				goto loop1;
			}
			BO_LOCK(bo);
		}
		BO_UNLOCK(bo);
		KASSERT(bp->b_bufobj == bo,
		    ("bp %p wrong b_bufobj %p should be %p",
		    bp, bp->b_bufobj, bo));
		if ((bp->b_flags & B_DELWRI) == 0)
			panic("fsync: not dirty");
		if ((vp->v_object != NULL) && (bp->b_flags & B_CLUSTEROK)) {
			vfs_bio_awrite(bp);
		} else {
			bremfree(bp);
			bawrite(bp);
		}
		if (maxretry < 1000)
			pause("dirty", hz < 1000 ? 1 : hz / 1000);
		BO_LOCK(bo);
		goto loop2;
	}

	/*
	 * If synchronous the caller expects us to completely resolve all
	 * dirty buffers in the system.  Wait for in-progress I/O to
	 * complete (which could include background bitmap writes), then
	 * retry if dirty blocks still exist.
	 */
	if (waitfor == MNT_WAIT) {
		bufobj_wwait(bo, 0, 0);
		if (bo->bo_dirty.bv_cnt > 0) {
			/*
			 * If we are unable to write any of these buffers
			 * then we fail now rather than trying endlessly
			 * to write them out.
			 */
			TAILQ_FOREACH(bp, &bo->bo_dirty.bv_hd, b_bobufs)
				if ((error = bp->b_error) != 0)
					break;
			if ((mp != NULL && mp->mnt_secondary_writes > 0) ||
			    (error == 0 && --maxretry >= 0))
				goto loop1;
			if (error == 0)
				error = EAGAIN;
		}
	}
	BO_UNLOCK(bo);
	if (error != 0)
		vn_printf(vp, "fsync: giving up on dirty (error = %d) ", error);

	return (error);
}

/*
 * Copies a byte range from invp to outvp.  Calls VOP_COPY_FILE_RANGE()
 * or vn_generic_copy_file_range() after rangelocking the byte ranges,
 * to do the actual copy.
 * vn_generic_copy_file_range() is factored out, so it can be called
 * from a VOP_COPY_FILE_RANGE() call as well, but handles vnodes from
 * different file systems.
 */
int
vn_copy_file_range(struct vnode *invp, off_t *inoffp, struct vnode *outvp,
    off_t *outoffp, size_t *lenp, unsigned int flags, struct ucred *incred,
    struct ucred *outcred, struct thread *fsize_td)
{
	int error;
	size_t len;
	uint64_t uval;

	len = *lenp;
	*lenp = 0;		/* For error returns. */
	error = 0;

	/* Do some sanity checks on the arguments. */
	if (invp->v_type == VDIR || outvp->v_type == VDIR)
		error = EISDIR;
	else if (*inoffp < 0 || *outoffp < 0 ||
	    invp->v_type != VREG || outvp->v_type != VREG)
		error = EINVAL;
	if (error != 0)
		goto out;

	/* Ensure offset + len does not wrap around. */
	uval = *inoffp;
	uval += len;
	if (uval > INT64_MAX)
		len = INT64_MAX - *inoffp;
	uval = *outoffp;
	uval += len;
	if (uval > INT64_MAX)
		len = INT64_MAX - *outoffp;
	if (len == 0)
		goto out;

	/*
	 * If the two vnode are for the same file system, call
	 * VOP_COPY_FILE_RANGE(), otherwise call vn_generic_copy_file_range()
	 * which can handle copies across multiple file systems.
	 */
	*lenp = len;
	if (invp->v_mount == outvp->v_mount)
		error = VOP_COPY_FILE_RANGE(invp, inoffp, outvp, outoffp,
		    lenp, flags, incred, outcred, fsize_td);
	else
		error = vn_generic_copy_file_range(invp, inoffp, outvp,
		    outoffp, lenp, flags, incred, outcred, fsize_td);
out:
	return (error);
}

/*
 * Test len bytes of data starting at dat for all bytes == 0.
 * Return true if all bytes are zero, false otherwise.
 * Expects dat to be well aligned.
 */
static bool
mem_iszero(void *dat, int len)
{
	int i;
	const u_int *p;
	const char *cp;

	for (p = dat; len > 0; len -= sizeof(*p), p++) {
		if (len >= sizeof(*p)) {
			if (*p != 0)
				return (false);
		} else {
			cp = (const char *)p;
			for (i = 0; i < len; i++, cp++)
				if (*cp != '\0')
					return (false);
		}
	}
	return (true);
}

/*
 * Look for a hole in the output file and, if found, adjust *outoffp
 * and *xferp to skip past the hole.
 * *xferp is the entire hole length to be written and xfer2 is how many bytes
 * to be written as 0's upon return.
 */
static off_t
vn_skip_hole(struct vnode *outvp, off_t xfer2, off_t *outoffp, off_t *xferp,
    off_t *dataoffp, off_t *holeoffp, struct ucred *cred)
{
	int error;
	off_t delta;

	if (*holeoffp == 0 || *holeoffp <= *outoffp) {
		*dataoffp = *outoffp;
		error = VOP_IOCTL(outvp, FIOSEEKDATA, dataoffp, 0, cred,
		    curthread);
		if (error == 0) {
			*holeoffp = *dataoffp;
			error = VOP_IOCTL(outvp, FIOSEEKHOLE, holeoffp, 0, cred,
			    curthread);
		}
		if (error != 0 || *holeoffp == *dataoffp) {
			/*
			 * Since outvp is unlocked, it may be possible for
			 * another thread to do a truncate(), lseek(), write()
			 * creating a hole at startoff between the above
			 * VOP_IOCTL() calls, if the other thread does not do
			 * rangelocking.
			 * If that happens, *holeoffp == *dataoffp and finding
			 * the hole has failed, so disable vn_skip_hole().
			 */
			*holeoffp = -1;	/* Disable use of vn_skip_hole(). */
			return (xfer2);
		}
		KASSERT(*dataoffp >= *outoffp,
		    ("vn_skip_hole: dataoff=%jd < outoff=%jd",
		    (intmax_t)*dataoffp, (intmax_t)*outoffp));
		KASSERT(*holeoffp > *dataoffp,
		    ("vn_skip_hole: holeoff=%jd <= dataoff=%jd",
		    (intmax_t)*holeoffp, (intmax_t)*dataoffp));
	}

	/*
	 * If there is a hole before the data starts, advance *outoffp and
	 * *xferp past the hole.
	 */
	if (*dataoffp > *outoffp) {
		delta = *dataoffp - *outoffp;
		if (delta >= *xferp) {
			/* Entire *xferp is a hole. */
			*outoffp += *xferp;
			*xferp = 0;
			return (0);
		}
		*xferp -= delta;
		*outoffp += delta;
		xfer2 = MIN(xfer2, *xferp);
	}

	/*
	 * If a hole starts before the end of this xfer2, reduce this xfer2 so
	 * that the write ends at the start of the hole.
	 * *holeoffp should always be greater than *outoffp, but for the
	 * non-INVARIANTS case, check this to make sure xfer2 remains a sane
	 * value.
	 */
	if (*holeoffp > *outoffp && *holeoffp < *outoffp + xfer2)
		xfer2 = *holeoffp - *outoffp;
	return (xfer2);
}

/*
 * Write an xfer sized chunk to outvp in blksize blocks from dat.
 * dat is a maximum of blksize in length and can be written repeatedly in
 * the chunk.
 * If growfile == true, just grow the file via vn_truncate_locked() instead
 * of doing actual writes.
 * If checkhole == true, a hole is being punched, so skip over any hole
 * already in the output file.
 */
static int
vn_write_outvp(struct vnode *outvp, char *dat, off_t outoff, off_t xfer,
    u_long blksize, bool growfile, bool checkhole, struct ucred *cred)
{
	struct mount *mp;
	off_t dataoff, holeoff, xfer2;
	int error, lckf;

	/*
	 * Loop around doing writes of blksize until write has been completed.
	 * Lock/unlock on each loop iteration so that a bwillwrite() can be
	 * done for each iteration, since the xfer argument can be very
	 * large if there is a large hole to punch in the output file.
	 */
	error = 0;
	holeoff = 0;
	do {
		xfer2 = MIN(xfer, blksize);
		if (checkhole) {
			/*
			 * Punching a hole.  Skip writing if there is
			 * already a hole in the output file.
			 */
			xfer2 = vn_skip_hole(outvp, xfer2, &outoff, &xfer,
			    &dataoff, &holeoff, cred);
			if (xfer == 0)
				break;
			if (holeoff < 0)
				checkhole = false;
			KASSERT(xfer2 > 0, ("vn_write_outvp: xfer2=%jd",
			    (intmax_t)xfer2));
		}
		bwillwrite();
		mp = NULL;
		error = vn_start_write(outvp, &mp, V_WAIT);
		if (error != 0)
			break;
		if (growfile) {
			error = vn_lock(outvp, LK_EXCLUSIVE);
			if (error == 0) {
				error = vn_truncate_locked(outvp, outoff + xfer,
				    false, cred);
				VOP_UNLOCK(outvp);
			}
		} else {
			if (MNT_SHARED_WRITES(mp))
				lckf = LK_SHARED;
			else
				lckf = LK_EXCLUSIVE;
			error = vn_lock(outvp, lckf);
			if (error == 0) {
				error = vn_rdwr(UIO_WRITE, outvp, dat, xfer2,
				    outoff, UIO_SYSSPACE, IO_NODELOCKED,
				    curthread->td_ucred, cred, NULL, curthread);
				outoff += xfer2;
				xfer -= xfer2;
				VOP_UNLOCK(outvp);
			}
		}
		if (mp != NULL)
			vn_finished_write(mp);
	} while (!growfile && xfer > 0 && error == 0);
	return (error);
}

/*
 * Copy a byte range of one file to another.  This function can handle the
 * case where invp and outvp are on different file systems.
 * It can also be called by a VOP_COPY_FILE_RANGE() to do the work, if there
 * is no better file system specific way to do it.
 */
int
vn_generic_copy_file_range(struct vnode *invp, off_t *inoffp,
    struct vnode *outvp, off_t *outoffp, size_t *lenp, unsigned int flags,
    struct ucred *incred, struct ucred *outcred, struct thread *fsize_td)
{
	struct vattr va;
	struct mount *mp;
	struct uio io;
	off_t startoff, endoff, xfer, xfer2;
	u_long blksize;
	int error, interrupted;
	bool cantseek, readzeros, eof, lastblock;
	ssize_t aresid;
	size_t copylen, len, rem, savlen;
	char *dat;
	long holein, holeout;

	holein = holeout = 0;
	savlen = len = *lenp;
	error = 0;
	interrupted = 0;
	dat = NULL;

	error = vn_lock(invp, LK_SHARED);
	if (error != 0)
		goto out;
	if (VOP_PATHCONF(invp, _PC_MIN_HOLE_SIZE, &holein) != 0)
		holein = 0;
	VOP_UNLOCK(invp);

	mp = NULL;
	error = vn_start_write(outvp, &mp, V_WAIT);
	if (error == 0)
		error = vn_lock(outvp, LK_EXCLUSIVE);
	if (error == 0) {
		/*
		 * If fsize_td != NULL, do a vn_rlimit_fsize() call,
		 * now that outvp is locked.
		 */
		if (fsize_td != NULL) {
			io.uio_offset = *outoffp;
			io.uio_resid = len;
			error = vn_rlimit_fsize(outvp, &io, fsize_td);
			if (error != 0)
				error = EFBIG;
		}
		if (VOP_PATHCONF(outvp, _PC_MIN_HOLE_SIZE, &holeout) != 0)
			holeout = 0;
		/*
		 * Holes that are past EOF do not need to be written as a block
		 * of zero bytes.  So, truncate the output file as far as
		 * possible and then use va.va_size to decide if writing 0
		 * bytes is necessary in the loop below.
		 */
		if (error == 0)
			error = VOP_GETATTR(outvp, &va, outcred);
		if (error == 0 && va.va_size > *outoffp && va.va_size <=
		    *outoffp + len) {
#ifdef MAC
			error = mac_vnode_check_write(curthread->td_ucred,
			    outcred, outvp);
			if (error == 0)
#endif
				error = vn_truncate_locked(outvp, *outoffp,
				    false, outcred);
			if (error == 0)
				va.va_size = *outoffp;
		}
		VOP_UNLOCK(outvp);
	}
	if (mp != NULL)
		vn_finished_write(mp);
	if (error != 0)
		goto out;

	/*
	 * Set the blksize to the larger of the hole sizes for invp and outvp.
	 * If hole sizes aren't available, set the blksize to the larger 
	 * f_iosize of invp and outvp.
	 * This code expects the hole sizes and f_iosizes to be powers of 2.
	 * This value is clipped at 4Kbytes and 1Mbyte.
	 */
	blksize = MAX(holein, holeout);

	/* Clip len to end at an exact multiple of hole size. */
	if (blksize > 1) {
		rem = *inoffp % blksize;
		if (rem > 0)
			rem = blksize - rem;
		if (len > rem && len - rem > blksize)
			len = savlen = rounddown(len - rem, blksize) + rem;
	}

	if (blksize <= 1)
		blksize = MAX(invp->v_mount->mnt_stat.f_iosize,
		    outvp->v_mount->mnt_stat.f_iosize);
	if (blksize < 4096)
		blksize = 4096;
	else if (blksize > 1024 * 1024)
		blksize = 1024 * 1024;
	dat = malloc(blksize, M_TEMP, M_WAITOK);

	/*
	 * If VOP_IOCTL(FIOSEEKHOLE) works for invp, use it and FIOSEEKDATA
	 * to find holes.  Otherwise, just scan the read block for all 0s
	 * in the inner loop where the data copying is done.
	 * Note that some file systems such as NFSv3, NFSv4.0 and NFSv4.1 may
	 * support holes on the server, but do not support FIOSEEKHOLE.
	 */
	eof = false;
	while (len > 0 && error == 0 && !eof && interrupted == 0) {
		endoff = 0;			/* To shut up compilers. */
		cantseek = true;
		startoff = *inoffp;
		copylen = len;

		/*
		 * Find the next data area.  If there is just a hole to EOF,
		 * FIOSEEKDATA should fail and then we drop down into the
		 * inner loop and create the hole on the outvp file.
		 * (I do not know if any file system will report a hole to
		 *  EOF via FIOSEEKHOLE, but I am pretty sure FIOSEEKDATA
		 *  will fail for those file systems.)
		 *
		 * For input files that don't support FIOSEEKDATA/FIOSEEKHOLE,
		 * the code just falls through to the inner copy loop.
		 */
		error = EINVAL;
		if (holein > 0)
			error = VOP_IOCTL(invp, FIOSEEKDATA, &startoff, 0,
			    incred, curthread);
		if (error == 0) {
			endoff = startoff;
			error = VOP_IOCTL(invp, FIOSEEKHOLE, &endoff, 0,
			    incred, curthread);
			/*
			 * Since invp is unlocked, it may be possible for
			 * another thread to do a truncate(), lseek(), write()
			 * creating a hole at startoff between the above
			 * VOP_IOCTL() calls, if the other thread does not do
			 * rangelocking.
			 * If that happens, startoff == endoff and finding
			 * the hole has failed, so set an error.
			 */
			if (error == 0 && startoff == endoff)
				error = EINVAL; /* Any error. Reset to 0. */
		}
		if (error == 0) {
			if (startoff > *inoffp) {
				/* Found hole before data block. */
				xfer = MIN(startoff - *inoffp, len);
				if (*outoffp < va.va_size) {
					/* Must write 0s to punch hole. */
					xfer2 = MIN(va.va_size - *outoffp,
					    xfer);
					memset(dat, 0, MIN(xfer2, blksize));
					error = vn_write_outvp(outvp, dat,
					    *outoffp, xfer2, blksize, false,
					    holeout > 0, outcred);
				}

				if (error == 0 && *outoffp + xfer >
				    va.va_size && xfer == len)
					/* Grow last block. */
					error = vn_write_outvp(outvp, dat,
					    *outoffp, xfer, blksize, true,
					    false, outcred);
				if (error == 0) {
					*inoffp += xfer;
					*outoffp += xfer;
					len -= xfer;
					if (len < savlen)
						interrupted = sig_intr();
				}
			}
			copylen = MIN(len, endoff - startoff);
			cantseek = false;
		} else {
			cantseek = true;
			startoff = *inoffp;
			copylen = len;
			error = 0;
		}

		xfer = blksize;
		if (cantseek) {
			/*
			 * Set first xfer to end at a block boundary, so that
			 * holes are more likely detected in the loop below via
			 * the for all bytes 0 method.
			 */
			xfer -= (*inoffp % blksize);
		}
		/* Loop copying the data block. */
		while (copylen > 0 && error == 0 && !eof && interrupted == 0) {
			if (copylen < xfer)
				xfer = copylen;
			error = vn_lock(invp, LK_SHARED);
			if (error != 0)
				goto out;
			error = vn_rdwr(UIO_READ, invp, dat, xfer,
			    startoff, UIO_SYSSPACE, IO_NODELOCKED,
			    curthread->td_ucred, incred, &aresid,
			    curthread);
			VOP_UNLOCK(invp);
			lastblock = false;
			if (error == 0 && aresid > 0) {
				/* Stop the copy at EOF on the input file. */
				xfer -= aresid;
				eof = true;
				lastblock = true;
			}
			if (error == 0) {
				/*
				 * Skip the write for holes past the initial EOF
				 * of the output file, unless this is the last
				 * write of the output file at EOF.
				 */
				readzeros = cantseek ? mem_iszero(dat, xfer) :
				    false;
				if (xfer == len)
					lastblock = true;
				if (!cantseek || *outoffp < va.va_size ||
				    lastblock || !readzeros)
					error = vn_write_outvp(outvp, dat,
					    *outoffp, xfer, blksize,
					    readzeros && lastblock &&
					    *outoffp >= va.va_size, false,
					    outcred);
				if (error == 0) {
					*inoffp += xfer;
					startoff += xfer;
					*outoffp += xfer;
					copylen -= xfer;
					len -= xfer;
					if (len < savlen)
						interrupted = sig_intr();
				}
			}
			xfer = blksize;
		}
	}
out:
	*lenp = savlen - len;
	free(dat, M_TEMP);
	return (error);
}

static int
vn_fallocate(struct file *fp, off_t offset, off_t len, struct thread *td)
{
	struct mount *mp;
	struct vnode *vp;
	off_t olen, ooffset;
	int error;
#ifdef AUDIT
	int audited_vnode1 = 0;
#endif

	vp = fp->f_vnode;
	if (vp->v_type != VREG)
		return (ENODEV);

	/* Allocating blocks may take a long time, so iterate. */
	for (;;) {
		olen = len;
		ooffset = offset;

		bwillwrite();
		mp = NULL;
		error = vn_start_write(vp, &mp, V_WAIT | PCATCH);
		if (error != 0)
			break;
		error = vn_lock(vp, LK_EXCLUSIVE);
		if (error != 0) {
			vn_finished_write(mp);
			break;
		}
#ifdef AUDIT
		if (!audited_vnode1) {
			AUDIT_ARG_VNODE1(vp);
			audited_vnode1 = 1;
		}
#endif
#ifdef MAC
		error = mac_vnode_check_write(td->td_ucred, fp->f_cred, vp);
		if (error == 0)
#endif
			error = VOP_ALLOCATE(vp, &offset, &len);
		VOP_UNLOCK(vp);
		vn_finished_write(mp);

		if (olen + ooffset != offset + len) {
			panic("offset + len changed from %jx/%jx to %jx/%jx",
			    ooffset, olen, offset, len);
		}
		if (error != 0 || len == 0)
			break;
		KASSERT(olen > len, ("Iteration did not make progress?"));
		maybe_yield();
	}

	return (error);
}

static u_long vn_lock_pair_pause_cnt;
SYSCTL_ULONG(_debug, OID_AUTO, vn_lock_pair_pause, CTLFLAG_RD,
    &vn_lock_pair_pause_cnt, 0,
    "Count of vn_lock_pair deadlocks");

u_int vn_lock_pair_pause_max;
SYSCTL_UINT(_debug, OID_AUTO, vn_lock_pair_pause_max, CTLFLAG_RW,
    &vn_lock_pair_pause_max, 0,
    "Max ticks for vn_lock_pair deadlock avoidance sleep");

static void
vn_lock_pair_pause(const char *wmesg)
{
	atomic_add_long(&vn_lock_pair_pause_cnt, 1);
	pause(wmesg, prng32_bounded(vn_lock_pair_pause_max));
}

/*
 * Lock pair of vnodes vp1, vp2, avoiding lock order reversal.
 * vp1_locked indicates whether vp1 is exclusively locked; if not, vp1
 * must be unlocked.  Same for vp2 and vp2_locked.  One of the vnodes
 * can be NULL.
 *
 * The function returns with both vnodes exclusively locked, and
 * guarantees that it does not create lock order reversal with other
 * threads during its execution.  Both vnodes could be unlocked
 * temporary (and reclaimed).
 */
void
vn_lock_pair(struct vnode *vp1, bool vp1_locked, struct vnode *vp2,
    bool vp2_locked)
{
	int error;

	if (vp1 == NULL && vp2 == NULL)
		return;
	if (vp1 != NULL) {
		if (vp1_locked)
			ASSERT_VOP_ELOCKED(vp1, "vp1");
		else
			ASSERT_VOP_UNLOCKED(vp1, "vp1");
	} else {
		vp1_locked = true;
	}
	if (vp2 != NULL) {
		if (vp2_locked)
			ASSERT_VOP_ELOCKED(vp2, "vp2");
		else
			ASSERT_VOP_UNLOCKED(vp2, "vp2");
	} else {
		vp2_locked = true;
	}
	if (!vp1_locked && !vp2_locked) {
		vn_lock(vp1, LK_EXCLUSIVE | LK_RETRY);
		vp1_locked = true;
	}

	for (;;) {
		if (vp1_locked && vp2_locked)
			break;
		if (vp1_locked && vp2 != NULL) {
			if (vp1 != NULL) {
				error = VOP_LOCK1(vp2, LK_EXCLUSIVE | LK_NOWAIT,
				    __FILE__, __LINE__);
				if (error == 0)
					break;
				VOP_UNLOCK(vp1);
				vp1_locked = false;
				vn_lock_pair_pause("vlp1");
			}
			vn_lock(vp2, LK_EXCLUSIVE | LK_RETRY);
			vp2_locked = true;
		}
		if (vp2_locked && vp1 != NULL) {
			if (vp2 != NULL) {
				error = VOP_LOCK1(vp1, LK_EXCLUSIVE | LK_NOWAIT,
				    __FILE__, __LINE__);
				if (error == 0)
					break;
				VOP_UNLOCK(vp2);
				vp2_locked = false;
				vn_lock_pair_pause("vlp2");
			}
			vn_lock(vp1, LK_EXCLUSIVE | LK_RETRY);
			vp1_locked = true;
		}
	}
	if (vp1 != NULL)
		ASSERT_VOP_ELOCKED(vp1, "vp1 ret");
	if (vp2 != NULL)
		ASSERT_VOP_ELOCKED(vp2, "vp2 ret");
}