summaryrefslogtreecommitdiffstats
path: root/sys/nlm/nlm_prot_impl.c
blob: 74fae870efa36fa9a629b0b450232a98187dd83c (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
/*-
 * Copyright (c) 2008 Isilon Inc http://www.isilon.com/
 * Authors: Doug Rabson <dfr@rabson.org>
 * Developed with Red Inc: Alfred Perlstein <alfred@freebsd.org>
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR 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 AUTHOR 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.
 */

#include "opt_inet6.h"

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

#include <sys/param.h>
#include <sys/fail.h>
#include <sys/fcntl.h>
#include <sys/kernel.h>
#include <sys/kthread.h>
#include <sys/lockf.h>
#include <sys/malloc.h>
#include <sys/mount.h>
#if __FreeBSD_version >= 700000
#include <sys/priv.h>
#endif
#include <sys/proc.h>
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/syscall.h>
#include <sys/sysctl.h>
#include <sys/sysent.h>
#include <sys/syslog.h>
#include <sys/sysproto.h>
#include <sys/systm.h>
#include <sys/taskqueue.h>
#include <sys/unistd.h>
#include <sys/vnode.h>

#include <nfs/nfsproto.h>
#include <nfs/nfs_lock.h>

#include <nlm/nlm_prot.h>
#include <nlm/sm_inter.h>
#include <nlm/nlm.h>
#include <rpc/rpc_com.h>
#include <rpc/rpcb_prot.h>

MALLOC_DEFINE(M_NLM, "NLM", "Network Lock Manager");

/*
 * If a host is inactive (and holds no locks) for this amount of
 * seconds, we consider it idle and stop tracking it.
 */
#define NLM_IDLE_TIMEOUT	30

/*
 * We check the host list for idle every few seconds.
 */
#define NLM_IDLE_PERIOD		5

/*
 * We only look for GRANTED_RES messages for a little while.
 */
#define NLM_EXPIRE_TIMEOUT	10

/*
 * Support for sysctl vfs.nlm.sysid
 */
static SYSCTL_NODE(_vfs, OID_AUTO, nlm, CTLFLAG_RW, NULL,
    "Network Lock Manager");
static SYSCTL_NODE(_vfs_nlm, OID_AUTO, sysid, CTLFLAG_RW, NULL, "");

/*
 * Syscall hooks
 */
static int nlm_syscall_offset = SYS_nlm_syscall;
static struct sysent nlm_syscall_prev_sysent;
#if __FreeBSD_version < 700000
static struct sysent nlm_syscall_sysent = {
	(sizeof(struct nlm_syscall_args) / sizeof(register_t)) | SYF_MPSAFE,
	(sy_call_t *) nlm_syscall
};
#else
MAKE_SYSENT(nlm_syscall);
#endif
static bool_t nlm_syscall_registered = FALSE;

/*
 * Debug level passed in from userland. We also support a sysctl hook
 * so that it can be changed on a live system.
 */
static int nlm_debug_level;
SYSCTL_INT(_debug, OID_AUTO, nlm_debug, CTLFLAG_RW, &nlm_debug_level, 0, "");

#define NLM_DEBUG(_level, args...)			\
	do {						\
		if (nlm_debug_level >= (_level))	\
			log(LOG_DEBUG, args);		\
	} while(0)
#define NLM_ERR(args...)			\
	do {					\
		log(LOG_ERR, args);		\
	} while(0)

/*
 * Grace period handling. The value of nlm_grace_threshold is the
 * value of time_uptime after which we are serving requests normally.
 */
static time_t nlm_grace_threshold;

/*
 * We check for idle hosts if time_uptime is greater than
 * nlm_next_idle_check,
 */
static time_t nlm_next_idle_check;

/*
 * A flag to indicate the server is already running.
 */
static int nlm_is_running;

/*
 * A socket to use for RPC - shared by all IPv4 RPC clients.
 */
static struct socket *nlm_socket;

#ifdef INET6

/*
 * A socket to use for RPC - shared by all IPv6 RPC clients.
 */
static struct socket *nlm_socket6;

#endif

/*
 * An RPC client handle that can be used to communicate with the local
 * NSM.
 */
static CLIENT *nlm_nsm;

/*
 * An AUTH handle for the server's creds.
 */
static AUTH *nlm_auth;

/*
 * A zero timeval for sending async RPC messages.
 */
struct timeval nlm_zero_tv = { 0, 0 };

/*
 * The local NSM state number
 */
int nlm_nsm_state;


/*
 * A lock to protect the host list and waiting lock list.
 */
static struct mtx nlm_global_lock;

/*
 * Locks:
 * (l)		locked by nh_lock
 * (s)		only accessed via server RPC which is single threaded
 * (g)		locked by nlm_global_lock
 * (c)		const until freeing
 * (a)		modified using atomic ops
 */

/*
 * A pending client-side lock request, stored on the nlm_waiting_locks
 * list.
 */
struct nlm_waiting_lock {
	TAILQ_ENTRY(nlm_waiting_lock) nw_link; /* (g) */
	bool_t		nw_waiting;	       /* (g) */
	nlm4_lock	nw_lock;	       /* (c) */
	union nfsfh	nw_fh;		       /* (c) */
	struct vnode	*nw_vp;		       /* (c) */
};
TAILQ_HEAD(nlm_waiting_lock_list, nlm_waiting_lock);

struct nlm_waiting_lock_list nlm_waiting_locks; /* (g) */

/*
 * A pending server-side asynchronous lock request, stored on the
 * nh_pending list of the NLM host.
 */
struct nlm_async_lock {
	TAILQ_ENTRY(nlm_async_lock) af_link; /* (l) host's list of locks */
	struct task	af_task;	/* (c) async callback details */
	void		*af_cookie;	/* (l) lock manager cancel token */
	struct vnode	*af_vp;		/* (l) vnode to lock */
	struct flock	af_fl;		/* (c) lock details */
	struct nlm_host *af_host;	/* (c) host which is locking */
	CLIENT		*af_rpc;	/* (c) rpc client to send message */
	nlm4_testargs	af_granted;	/* (c) notification details */
	time_t		af_expiretime;	/* (c) notification time */
};
TAILQ_HEAD(nlm_async_lock_list, nlm_async_lock);

/*
 * NLM host.
 */
enum nlm_host_state {
	NLM_UNMONITORED,
	NLM_MONITORED,
	NLM_MONITOR_FAILED,
	NLM_RECOVERING
};

struct nlm_rpc {
	CLIENT		*nr_client;    /* (l) RPC client handle */
	time_t		nr_create_time; /* (l) when client was created */
};

struct nlm_host {
	struct mtx	nh_lock;
	volatile u_int	nh_refs;       /* (a) reference count */
	TAILQ_ENTRY(nlm_host) nh_link; /* (g) global list of hosts */
	char		nh_caller_name[MAXNAMELEN]; /* (c) printable name of host */
	uint32_t	nh_sysid;	 /* (c) our allocaed system ID */
	char		nh_sysid_string[10]; /* (c) string rep. of sysid */
	struct sockaddr_storage	nh_addr; /* (s) remote address of host */
	struct nlm_rpc	nh_srvrpc;	 /* (l) RPC for server replies */
	struct nlm_rpc	nh_clntrpc;	 /* (l) RPC for client requests */
	rpcvers_t	nh_vers;	 /* (s) NLM version of host */
	int		nh_state;	 /* (s) last seen NSM state of host */
	enum nlm_host_state nh_monstate; /* (l) local NSM monitoring state */
	time_t		nh_idle_timeout; /* (s) Time at which host is idle */
	struct sysctl_ctx_list nh_sysctl; /* (c) vfs.nlm.sysid nodes */
	uint32_t	nh_grantcookie;  /* (l) grant cookie counter */
	struct nlm_async_lock_list nh_pending; /* (l) pending async locks */
	struct nlm_async_lock_list nh_granted; /* (l) granted locks */
	struct nlm_async_lock_list nh_finished; /* (l) finished async locks */
};
TAILQ_HEAD(nlm_host_list, nlm_host);

static struct nlm_host_list nlm_hosts; /* (g) */
static uint32_t nlm_next_sysid = 1;    /* (g) */

static void	nlm_host_unmonitor(struct nlm_host *);

struct nlm_grantcookie {
	uint32_t	ng_sysid;
	uint32_t	ng_cookie;
};

static inline uint32_t
ng_sysid(struct netobj *src)
{

	return ((struct nlm_grantcookie *)src->n_bytes)->ng_sysid;
}

static inline uint32_t
ng_cookie(struct netobj *src)
{

	return ((struct nlm_grantcookie *)src->n_bytes)->ng_cookie;
}

/**********************************************************************/

/*
 * Initialise NLM globals.
 */
static void
nlm_init(void *dummy)
{
	int error;

	mtx_init(&nlm_global_lock, "nlm_global_lock", NULL, MTX_DEF);
	TAILQ_INIT(&nlm_waiting_locks);
	TAILQ_INIT(&nlm_hosts);

	error = syscall_register(&nlm_syscall_offset, &nlm_syscall_sysent,
	    &nlm_syscall_prev_sysent);
	if (error)
		NLM_ERR("Can't register NLM syscall\n");
	else
		nlm_syscall_registered = TRUE;
}
SYSINIT(nlm_init, SI_SUB_LOCK, SI_ORDER_FIRST, nlm_init, NULL);

static void
nlm_uninit(void *dummy)
{

	if (nlm_syscall_registered)
		syscall_deregister(&nlm_syscall_offset,
		    &nlm_syscall_prev_sysent);
}
SYSUNINIT(nlm_uninit, SI_SUB_LOCK, SI_ORDER_FIRST, nlm_uninit, NULL);

/*
 * Create a netobj from an arbitrary source.
 */
void
nlm_make_netobj(struct netobj *dst, caddr_t src, size_t srcsize,
    struct malloc_type *type)
{

	dst->n_len = srcsize;
	dst->n_bytes = malloc(srcsize, type, M_WAITOK);
	memcpy(dst->n_bytes, src, srcsize);
}

/*
 * Copy a struct netobj.
 */ 
void
nlm_copy_netobj(struct netobj *dst, struct netobj *src,
    struct malloc_type *type)
{

	nlm_make_netobj(dst, src->n_bytes, src->n_len, type);
}


/*
 * Create an RPC client handle for the given (address,prog,vers)
 * triple using UDP.
 */
static CLIENT *
nlm_get_rpc(struct sockaddr *sa, rpcprog_t prog, rpcvers_t vers)
{
	char *wchan = "nlmrcv";
	const char* protofmly;
	struct sockaddr_storage ss;
	struct socket *so;
	CLIENT *rpcb;
	struct timeval timo;
	RPCB parms;
	char *uaddr;
	enum clnt_stat stat = RPC_SUCCESS;
	int rpcvers = RPCBVERS4;
	bool_t do_tcp = FALSE;
	bool_t tryagain = FALSE;
	struct portmap mapping;
	u_short port = 0;

	/*
	 * First we need to contact the remote RPCBIND service to find
	 * the right port.
	 */
	memcpy(&ss, sa, sa->sa_len);
	switch (ss.ss_family) {
	case AF_INET:
		((struct sockaddr_in *)&ss)->sin_port = htons(111);
		protofmly = "inet";
		so = nlm_socket;
		break;
		
#ifdef INET6
	case AF_INET6:
		((struct sockaddr_in6 *)&ss)->sin6_port = htons(111);
		protofmly = "inet6";
		so = nlm_socket6;
		break;
#endif

	default:
		/*
		 * Unsupported address family - fail.
		 */
		return (NULL);
	}

	rpcb = clnt_dg_create(so, (struct sockaddr *)&ss,
	    RPCBPROG, rpcvers, 0, 0);
	if (!rpcb)
		return (NULL);

try_tcp:
	parms.r_prog = prog;
	parms.r_vers = vers;
	if (do_tcp)
		parms.r_netid = "tcp";
	else
		parms.r_netid = "udp";
	parms.r_addr = "";
	parms.r_owner = "";

	/*
	 * Use the default timeout.
	 */
	timo.tv_sec = 25;
	timo.tv_usec = 0;
again:
	switch (rpcvers) {
	case RPCBVERS4:
	case RPCBVERS:
		/*
		 * Try RPCBIND 4 then 3.
		 */
		uaddr = NULL;
		stat = CLNT_CALL(rpcb, (rpcprog_t) RPCBPROC_GETADDR,
		    (xdrproc_t) xdr_rpcb, &parms,
		    (xdrproc_t) xdr_wrapstring, &uaddr, timo);
		if (stat == RPC_SUCCESS) {
			/*
			 * We have a reply from the remote RPCBIND - turn it
			 * into an appropriate address and make a new client
			 * that can talk to the remote NLM.
			 *
			 * XXX fixup IPv6 scope ID.
			 */
			struct netbuf *a;
			a = __rpc_uaddr2taddr_af(ss.ss_family, uaddr);
			if (!a) {
				tryagain = TRUE;
			} else {
				tryagain = FALSE;
				memcpy(&ss, a->buf, a->len);
				free(a->buf, M_RPC);
				free(a, M_RPC);
				xdr_free((xdrproc_t) xdr_wrapstring, &uaddr);
			}
		}
		if (tryagain || stat == RPC_PROGVERSMISMATCH) {
			if (rpcvers == RPCBVERS4)
				rpcvers = RPCBVERS;
			else if (rpcvers == RPCBVERS)
				rpcvers = PMAPVERS;
			CLNT_CONTROL(rpcb, CLSET_VERS, &rpcvers);
			goto again;
		}
		break;
	case PMAPVERS:
		/*
		 * Try portmap.
		 */
		mapping.pm_prog = parms.r_prog;
		mapping.pm_vers = parms.r_vers;
		mapping.pm_prot = do_tcp ? IPPROTO_TCP : IPPROTO_UDP;
		mapping.pm_port = 0;

		stat = CLNT_CALL(rpcb, (rpcprog_t) PMAPPROC_GETPORT,
		    (xdrproc_t) xdr_portmap, &mapping,
		    (xdrproc_t) xdr_u_short, &port, timo);

		if (stat == RPC_SUCCESS) {
			switch (ss.ss_family) {
			case AF_INET:
				((struct sockaddr_in *)&ss)->sin_port =
					htons(port);
				break;
		
#ifdef INET6
			case AF_INET6:
				((struct sockaddr_in6 *)&ss)->sin6_port =
					htons(port);
				break;
#endif
			}
		}
		break;
	default:
		panic("invalid rpcvers %d", rpcvers);
	}
	/*
	 * We may have a positive response from the portmapper, but the NLM
	 * service was not found. Make sure we received a valid port.
	 */
	switch (ss.ss_family) {
	case AF_INET:
		port = ((struct sockaddr_in *)&ss)->sin_port;
		break;
#ifdef INET6
	case AF_INET6:
		port = ((struct sockaddr_in6 *)&ss)->sin6_port;
		break;
#endif
	}
	if (stat != RPC_SUCCESS || !port) {
		/*
		 * If we were able to talk to rpcbind or portmap, but the udp
		 * variant wasn't available, ask about tcp.
		 *
		 * XXX - We could also check for a TCP portmapper, but
		 * if the host is running a portmapper at all, we should be able
		 * to hail it over UDP.
		 */
		if (stat == RPC_SUCCESS && !do_tcp) {
			do_tcp = TRUE;
			goto try_tcp;
		}

		/* Otherwise, bad news. */
		NLM_ERR("NLM: failed to contact remote rpcbind, "
		    "stat = %d, port = %d\n", (int) stat, port);
		CLNT_DESTROY(rpcb);
		return (NULL);
	}

	if (do_tcp) {
		/*
		 * Destroy the UDP client we used to speak to rpcbind and
		 * recreate as a TCP client.
		 */
		struct netconfig *nconf = NULL;

		CLNT_DESTROY(rpcb);

		switch (ss.ss_family) {
		case AF_INET:
			nconf = getnetconfigent("tcp");
			break;
#ifdef INET6
		case AF_INET6:
			nconf = getnetconfigent("tcp6");
			break;
#endif
		}

		rpcb = clnt_reconnect_create(nconf, (struct sockaddr *)&ss,
		    prog, vers, 0, 0);
		CLNT_CONTROL(rpcb, CLSET_WAITCHAN, wchan);
		rpcb->cl_auth = nlm_auth;
		
	} else {
		/*
		 * Re-use the client we used to speak to rpcbind.
		 */
		CLNT_CONTROL(rpcb, CLSET_SVC_ADDR, &ss);
		CLNT_CONTROL(rpcb, CLSET_PROG, &prog);
		CLNT_CONTROL(rpcb, CLSET_VERS, &vers);
		CLNT_CONTROL(rpcb, CLSET_WAITCHAN, wchan);
		rpcb->cl_auth = nlm_auth;
	}

	return (rpcb);
}

/*
 * This async callback after when an async lock request has been
 * granted. We notify the host which initiated the request.
 */
static void
nlm_lock_callback(void *arg, int pending)
{
	struct nlm_async_lock *af = (struct nlm_async_lock *) arg;
	struct rpc_callextra ext;

	NLM_DEBUG(2, "NLM: async lock %p for %s (sysid %d) granted,"
	    " cookie %d:%d\n", af, af->af_host->nh_caller_name,
	    af->af_host->nh_sysid, ng_sysid(&af->af_granted.cookie),
	    ng_cookie(&af->af_granted.cookie));

	/*
	 * Send the results back to the host.
	 *
	 * Note: there is a possible race here with nlm_host_notify
	 * destroying the RPC client. To avoid problems, the first
	 * thing nlm_host_notify does is to cancel pending async lock
	 * requests.
	 */
	memset(&ext, 0, sizeof(ext));
	ext.rc_auth = nlm_auth;
	if (af->af_host->nh_vers == NLM_VERS4) {
		nlm4_granted_msg_4(&af->af_granted,
		    NULL, af->af_rpc, &ext, nlm_zero_tv);
	} else {
		/*
		 * Back-convert to legacy protocol
		 */
		nlm_testargs granted;
		granted.cookie = af->af_granted.cookie;
		granted.exclusive = af->af_granted.exclusive;
		granted.alock.caller_name =
			af->af_granted.alock.caller_name;
		granted.alock.fh = af->af_granted.alock.fh;
		granted.alock.oh = af->af_granted.alock.oh;
		granted.alock.svid = af->af_granted.alock.svid;
		granted.alock.l_offset =
			af->af_granted.alock.l_offset;
		granted.alock.l_len =
			af->af_granted.alock.l_len;

		nlm_granted_msg_1(&granted,
		    NULL, af->af_rpc, &ext, nlm_zero_tv);
	}

	/*
	 * Move this entry to the nh_granted list.
	 */
	af->af_expiretime = time_uptime + NLM_EXPIRE_TIMEOUT;
	mtx_lock(&af->af_host->nh_lock);
	TAILQ_REMOVE(&af->af_host->nh_pending, af, af_link);
	TAILQ_INSERT_TAIL(&af->af_host->nh_granted, af, af_link);
	mtx_unlock(&af->af_host->nh_lock);
}

/*
 * Free an async lock request. The request must have been removed from
 * any list.
 */
static void
nlm_free_async_lock(struct nlm_async_lock *af)
{
	/*
	 * Free an async lock.
	 */
	if (af->af_rpc)
		CLNT_RELEASE(af->af_rpc);
	xdr_free((xdrproc_t) xdr_nlm4_testargs, &af->af_granted);
	if (af->af_vp)
		vrele(af->af_vp);
	free(af, M_NLM);
}

/*
 * Cancel our async request - this must be called with
 * af->nh_host->nh_lock held. This is slightly complicated by a
 * potential race with our own callback. If we fail to cancel the
 * lock, it must already have been granted - we make sure our async
 * task has completed by calling taskqueue_drain in this case.
 */
static int
nlm_cancel_async_lock(struct nlm_async_lock *af)
{
	struct nlm_host *host = af->af_host;
	int error;

	mtx_assert(&host->nh_lock, MA_OWNED);

	mtx_unlock(&host->nh_lock);

	error = VOP_ADVLOCKASYNC(af->af_vp, NULL, F_CANCEL, &af->af_fl,
	    F_REMOTE, NULL, &af->af_cookie);

	if (error) {
		/*
		 * We failed to cancel - make sure our callback has
		 * completed before we continue.
		 */
		taskqueue_drain(taskqueue_thread, &af->af_task);
	}

	mtx_lock(&host->nh_lock);
	
	if (!error) {
		NLM_DEBUG(2, "NLM: async lock %p for %s (sysid %d) "
		    "cancelled\n", af, host->nh_caller_name, host->nh_sysid);

		/*
		 * Remove from the nh_pending list and free now that
		 * we are safe from the callback.
		 */
		TAILQ_REMOVE(&host->nh_pending, af, af_link);
		mtx_unlock(&host->nh_lock);
		nlm_free_async_lock(af);
		mtx_lock(&host->nh_lock);
	}

	return (error);
}

static void
nlm_check_expired_locks(struct nlm_host *host)
{
	struct nlm_async_lock *af;
	time_t uptime = time_uptime;

	mtx_lock(&host->nh_lock);
	while ((af = TAILQ_FIRST(&host->nh_granted)) != NULL
	    && uptime >= af->af_expiretime) {
		NLM_DEBUG(2, "NLM: async lock %p for %s (sysid %d) expired,"
		    " cookie %d:%d\n", af, af->af_host->nh_caller_name,
		    af->af_host->nh_sysid, ng_sysid(&af->af_granted.cookie),
		    ng_cookie(&af->af_granted.cookie));
		TAILQ_REMOVE(&host->nh_granted, af, af_link);
		mtx_unlock(&host->nh_lock);
		nlm_free_async_lock(af);
		mtx_lock(&host->nh_lock);
	}
	while ((af = TAILQ_FIRST(&host->nh_finished)) != NULL) {
		TAILQ_REMOVE(&host->nh_finished, af, af_link);
		mtx_unlock(&host->nh_lock);
		nlm_free_async_lock(af);
		mtx_lock(&host->nh_lock);
	}
	mtx_unlock(&host->nh_lock);
}

/*
 * Free resources used by a host. This is called after the reference
 * count has reached zero so it doesn't need to worry about locks.
 */
static void
nlm_host_destroy(struct nlm_host *host)
{

	mtx_lock(&nlm_global_lock);
	TAILQ_REMOVE(&nlm_hosts, host, nh_link);
	mtx_unlock(&nlm_global_lock);

	if (host->nh_srvrpc.nr_client)
		CLNT_RELEASE(host->nh_srvrpc.nr_client);
	if (host->nh_clntrpc.nr_client)
		CLNT_RELEASE(host->nh_clntrpc.nr_client);
	mtx_destroy(&host->nh_lock);
	sysctl_ctx_free(&host->nh_sysctl);
	free(host, M_NLM);
}

/*
 * Thread start callback for client lock recovery
 */
static void
nlm_client_recovery_start(void *arg)
{
	struct nlm_host *host = (struct nlm_host *) arg;

	NLM_DEBUG(1, "NLM: client lock recovery for %s started\n",
	    host->nh_caller_name);

	nlm_client_recovery(host);

	NLM_DEBUG(1, "NLM: client lock recovery for %s completed\n",
	    host->nh_caller_name);

	host->nh_monstate = NLM_MONITORED;
	nlm_host_release(host);

	kthread_exit();
}

/*
 * This is called when we receive a host state change notification. We
 * unlock any active locks owned by the host. When rpc.lockd is
 * shutting down, this function is called with newstate set to zero
 * which allows us to cancel any pending async locks and clear the
 * locking state.
 */
static void
nlm_host_notify(struct nlm_host *host, int newstate)
{
	struct nlm_async_lock *af;

	if (newstate) {
		NLM_DEBUG(1, "NLM: host %s (sysid %d) rebooted, new "
		    "state is %d\n", host->nh_caller_name,
		    host->nh_sysid, newstate);
	}

	/*
	 * Cancel any pending async locks for this host.
	 */
	mtx_lock(&host->nh_lock);
	while ((af = TAILQ_FIRST(&host->nh_pending)) != NULL) {
		/*
		 * nlm_cancel_async_lock will remove the entry from
		 * nh_pending and free it.
		 */
		nlm_cancel_async_lock(af);
	}
	mtx_unlock(&host->nh_lock);
	nlm_check_expired_locks(host);

	/*
	 * The host just rebooted - trash its locks.
	 */
	lf_clearremotesys(host->nh_sysid);
	host->nh_state = newstate;

	/*
	 * If we have any remote locks for this host (i.e. it
	 * represents a remote NFS server that our local NFS client
	 * has locks for), start a recovery thread.
	 */
	if (newstate != 0
	    && host->nh_monstate != NLM_RECOVERING
	    && lf_countlocks(NLM_SYSID_CLIENT | host->nh_sysid) > 0) {
		struct thread *td;
		host->nh_monstate = NLM_RECOVERING;
		refcount_acquire(&host->nh_refs);
		kthread_add(nlm_client_recovery_start, host, curproc, &td, 0, 0,
		    "NFS lock recovery for %s", host->nh_caller_name);
	}
}

/*
 * Sysctl handler to count the number of locks for a sysid.
 */
static int
nlm_host_lock_count_sysctl(SYSCTL_HANDLER_ARGS)
{
	struct nlm_host *host;
	int count;

	host = oidp->oid_arg1;
	count = lf_countlocks(host->nh_sysid);
	return sysctl_handle_int(oidp, &count, 0, req);
}

/*
 * Sysctl handler to count the number of client locks for a sysid.
 */
static int
nlm_host_client_lock_count_sysctl(SYSCTL_HANDLER_ARGS)
{
	struct nlm_host *host;
	int count;

	host = oidp->oid_arg1;
	count = lf_countlocks(NLM_SYSID_CLIENT | host->nh_sysid);
	return sysctl_handle_int(oidp, &count, 0, req);
}

/*
 * Create a new NLM host.
 */
static struct nlm_host *
nlm_create_host(const char* caller_name)
{
	struct nlm_host *host;
	struct sysctl_oid *oid;

	mtx_assert(&nlm_global_lock, MA_OWNED);

	NLM_DEBUG(1, "NLM: new host %s (sysid %d)\n",
	    caller_name, nlm_next_sysid);
	host = malloc(sizeof(struct nlm_host), M_NLM, M_NOWAIT|M_ZERO);
	if (!host)
		return (NULL);
	mtx_init(&host->nh_lock, "nh_lock", NULL, MTX_DEF);
	host->nh_refs = 1;
	strlcpy(host->nh_caller_name, caller_name, MAXNAMELEN);
	host->nh_sysid = nlm_next_sysid++;
	snprintf(host->nh_sysid_string, sizeof(host->nh_sysid_string),
		"%d", host->nh_sysid);
	host->nh_vers = 0;
	host->nh_state = 0;
	host->nh_monstate = NLM_UNMONITORED;
	host->nh_grantcookie = 1;
	TAILQ_INIT(&host->nh_pending);
	TAILQ_INIT(&host->nh_granted);
	TAILQ_INIT(&host->nh_finished);
	TAILQ_INSERT_TAIL(&nlm_hosts, host, nh_link);

	mtx_unlock(&nlm_global_lock);

	sysctl_ctx_init(&host->nh_sysctl);
	oid = SYSCTL_ADD_NODE(&host->nh_sysctl,
	    SYSCTL_STATIC_CHILDREN(_vfs_nlm_sysid),
	    OID_AUTO, host->nh_sysid_string, CTLFLAG_RD, NULL, "");
	SYSCTL_ADD_STRING(&host->nh_sysctl, SYSCTL_CHILDREN(oid), OID_AUTO,
	    "hostname", CTLFLAG_RD, host->nh_caller_name, 0, "");
	SYSCTL_ADD_UINT(&host->nh_sysctl, SYSCTL_CHILDREN(oid), OID_AUTO,
	    "version", CTLFLAG_RD, &host->nh_vers, 0, "");
	SYSCTL_ADD_UINT(&host->nh_sysctl, SYSCTL_CHILDREN(oid), OID_AUTO,
	    "monitored", CTLFLAG_RD, &host->nh_monstate, 0, "");
	SYSCTL_ADD_PROC(&host->nh_sysctl, SYSCTL_CHILDREN(oid), OID_AUTO,
	    "lock_count", CTLTYPE_INT | CTLFLAG_RD, host, 0,
	    nlm_host_lock_count_sysctl, "I", "");
	SYSCTL_ADD_PROC(&host->nh_sysctl, SYSCTL_CHILDREN(oid), OID_AUTO,
	    "client_lock_count", CTLTYPE_INT | CTLFLAG_RD, host, 0,
	    nlm_host_client_lock_count_sysctl, "I", "");

	mtx_lock(&nlm_global_lock);

	return (host);
}

/*
 * Acquire the next sysid for remote locks not handled by the NLM.
 */
uint32_t
nlm_acquire_next_sysid(void)
{
	uint32_t next_sysid;

	mtx_lock(&nlm_global_lock);
	next_sysid = nlm_next_sysid++;
	mtx_unlock(&nlm_global_lock);
	return (next_sysid);
}

/*
 * Return non-zero if the address parts of the two sockaddrs are the
 * same.
 */
static int
nlm_compare_addr(const struct sockaddr *a, const struct sockaddr *b)
{
	const struct sockaddr_in *a4, *b4;
#ifdef INET6
	const struct sockaddr_in6 *a6, *b6;
#endif

	if (a->sa_family != b->sa_family)
		return (FALSE);

	switch (a->sa_family) {
	case AF_INET:
		a4 = (const struct sockaddr_in *) a;
		b4 = (const struct sockaddr_in *) b;
		return !memcmp(&a4->sin_addr, &b4->sin_addr,
		    sizeof(a4->sin_addr));
#ifdef INET6
	case AF_INET6:
		a6 = (const struct sockaddr_in6 *) a;
		b6 = (const struct sockaddr_in6 *) b;
		return !memcmp(&a6->sin6_addr, &b6->sin6_addr,
		    sizeof(a6->sin6_addr));
#endif
	}

	return (0);
}

/*
 * Check for idle hosts and stop monitoring them. We could also free
 * the host structure here, possibly after a larger timeout but that
 * would require some care to avoid races with
 * e.g. nlm_host_lock_count_sysctl.
 */
static void
nlm_check_idle(void)
{
	struct nlm_host *host;

	mtx_assert(&nlm_global_lock, MA_OWNED);

	if (time_uptime <= nlm_next_idle_check)
		return;

	nlm_next_idle_check = time_uptime + NLM_IDLE_PERIOD;

	TAILQ_FOREACH(host, &nlm_hosts, nh_link) {
		if (host->nh_monstate == NLM_MONITORED
		    && time_uptime > host->nh_idle_timeout) {
			mtx_unlock(&nlm_global_lock);
			if (lf_countlocks(host->nh_sysid) > 0
			    || lf_countlocks(NLM_SYSID_CLIENT
				+ host->nh_sysid)) {
				host->nh_idle_timeout =
					time_uptime + NLM_IDLE_TIMEOUT;
				mtx_lock(&nlm_global_lock);
				continue;
			}
			nlm_host_unmonitor(host);
			mtx_lock(&nlm_global_lock);
		} 
	}
}

/*
 * Search for an existing NLM host that matches the given name
 * (typically the caller_name element of an nlm4_lock).  If none is
 * found, create a new host. If 'addr' is non-NULL, record the remote
 * address of the host so that we can call it back for async
 * responses. If 'vers' is greater than zero then record the NLM
 * program version to use to communicate with this client.
 */
struct nlm_host *
nlm_find_host_by_name(const char *name, const struct sockaddr *addr,
    rpcvers_t vers)
{
	struct nlm_host *host;

	mtx_lock(&nlm_global_lock);

	/*
	 * The remote host is determined by caller_name.
	 */
	TAILQ_FOREACH(host, &nlm_hosts, nh_link) {
		if (!strcmp(host->nh_caller_name, name))
			break;
	}

	if (!host) {
		host = nlm_create_host(name);
		if (!host) {
			mtx_unlock(&nlm_global_lock);
			return (NULL);
		}
	}
	refcount_acquire(&host->nh_refs);

	host->nh_idle_timeout = time_uptime + NLM_IDLE_TIMEOUT;

	/*
	 * If we have an address for the host, record it so that we
	 * can send async replies etc.
	 */
	if (addr) {
		
		KASSERT(addr->sa_len < sizeof(struct sockaddr_storage),
		    ("Strange remote transport address length"));

		/*
		 * If we have seen an address before and we currently
		 * have an RPC client handle, make sure the address is
		 * the same, otherwise discard the client handle.
		 */
		if (host->nh_addr.ss_len && host->nh_srvrpc.nr_client) {
			if (!nlm_compare_addr(
				    (struct sockaddr *) &host->nh_addr,
				    addr)
			    || host->nh_vers != vers) {
				CLIENT *client;
				mtx_lock(&host->nh_lock);
				client = host->nh_srvrpc.nr_client;
				host->nh_srvrpc.nr_client = NULL;
				mtx_unlock(&host->nh_lock);
				if (client) {
					CLNT_RELEASE(client);
				}
			}
		}
		memcpy(&host->nh_addr, addr, addr->sa_len);
		host->nh_vers = vers;
	}

	nlm_check_idle();

	mtx_unlock(&nlm_global_lock);

	return (host);
}

/*
 * Search for an existing NLM host that matches the given remote
 * address. If none is found, create a new host with the requested
 * address and remember 'vers' as the NLM protocol version to use for
 * that host.
 */
struct nlm_host *
nlm_find_host_by_addr(const struct sockaddr *addr, int vers)
{
	/*
	 * Fake up a name using inet_ntop. This buffer is
	 * large enough for an IPv6 address.
	 */
	char tmp[sizeof "ffff:ffff:ffff:ffff:ffff:ffff:255.255.255.255"];
	struct nlm_host *host;

	switch (addr->sa_family) {
	case AF_INET:
		inet_ntop(AF_INET,
		    &((const struct sockaddr_in *) addr)->sin_addr,
		    tmp, sizeof tmp);
		break;
#ifdef INET6
	case AF_INET6:
		inet_ntop(AF_INET6,
		    &((const struct sockaddr_in6 *) addr)->sin6_addr,
		    tmp, sizeof tmp);
		break;
#endif
	default:
		strlcpy(tmp, "<unknown>", sizeof(tmp));
	}


	mtx_lock(&nlm_global_lock);

	/*
	 * The remote host is determined by caller_name.
	 */
	TAILQ_FOREACH(host, &nlm_hosts, nh_link) {
		if (nlm_compare_addr(addr,
			(const struct sockaddr *) &host->nh_addr))
			break;
	}

	if (!host) {
		host = nlm_create_host(tmp);
		if (!host) {
			mtx_unlock(&nlm_global_lock);
			return (NULL);
		}
		memcpy(&host->nh_addr, addr, addr->sa_len);
		host->nh_vers = vers;
	}
	refcount_acquire(&host->nh_refs);

	host->nh_idle_timeout = time_uptime + NLM_IDLE_TIMEOUT;

	nlm_check_idle();

	mtx_unlock(&nlm_global_lock);

	return (host);
}

/*
 * Find the NLM host that matches the value of 'sysid'. If none
 * exists, return NULL.
 */
static struct nlm_host *
nlm_find_host_by_sysid(int sysid)
{
	struct nlm_host *host;

	TAILQ_FOREACH(host, &nlm_hosts, nh_link) {
		if (host->nh_sysid == sysid) {
			refcount_acquire(&host->nh_refs);
			return (host);
		}
	}

	return (NULL);
}

void nlm_host_release(struct nlm_host *host)
{
	if (refcount_release(&host->nh_refs)) {
		/*
		 * Free the host
		 */
		nlm_host_destroy(host);
	}
}

/*
 * Unregister this NLM host with the local NSM due to idleness.
 */
static void
nlm_host_unmonitor(struct nlm_host *host)
{
	mon_id smmonid;
	sm_stat_res smstat;
	struct timeval timo;
	enum clnt_stat stat;

	NLM_DEBUG(1, "NLM: unmonitoring %s (sysid %d)\n",
	    host->nh_caller_name, host->nh_sysid);

	/*
	 * We put our assigned system ID value in the priv field to
	 * make it simpler to find the host if we are notified of a
	 * host restart.
	 */
	smmonid.mon_name = host->nh_caller_name;
	smmonid.my_id.my_name = "localhost";
	smmonid.my_id.my_prog = NLM_PROG;
	smmonid.my_id.my_vers = NLM_SM;
	smmonid.my_id.my_proc = NLM_SM_NOTIFY;

	timo.tv_sec = 25;
	timo.tv_usec = 0;
	stat = CLNT_CALL(nlm_nsm, SM_UNMON,
	    (xdrproc_t) xdr_mon, &smmonid,
	    (xdrproc_t) xdr_sm_stat, &smstat, timo);

	if (stat != RPC_SUCCESS) {
		NLM_ERR("Failed to contact local NSM - rpc error %d\n", stat);
		return;
	}
	if (smstat.res_stat == stat_fail) {
		NLM_ERR("Local NSM refuses to unmonitor %s\n",
		    host->nh_caller_name);
		return;
	}

	host->nh_monstate = NLM_UNMONITORED;
}

/*
 * Register this NLM host with the local NSM so that we can be
 * notified if it reboots.
 */
void
nlm_host_monitor(struct nlm_host *host, int state)
{
	mon smmon;
	sm_stat_res smstat;
	struct timeval timo;
	enum clnt_stat stat;

	if (state && !host->nh_state) {
		/*
		 * This is the first time we have seen an NSM state
		 * value for this host. We record it here to help
		 * detect host reboots.
		 */
		host->nh_state = state;
		NLM_DEBUG(1, "NLM: host %s (sysid %d) has NSM state %d\n",
		    host->nh_caller_name, host->nh_sysid, state);
	}

	mtx_lock(&host->nh_lock);
	if (host->nh_monstate != NLM_UNMONITORED) {
		mtx_unlock(&host->nh_lock);
		return;
	}
	host->nh_monstate = NLM_MONITORED;
	mtx_unlock(&host->nh_lock);

	NLM_DEBUG(1, "NLM: monitoring %s (sysid %d)\n",
	    host->nh_caller_name, host->nh_sysid);

	/*
	 * We put our assigned system ID value in the priv field to
	 * make it simpler to find the host if we are notified of a
	 * host restart.
	 */
	smmon.mon_id.mon_name = host->nh_caller_name;
	smmon.mon_id.my_id.my_name = "localhost";
	smmon.mon_id.my_id.my_prog = NLM_PROG;
	smmon.mon_id.my_id.my_vers = NLM_SM;
	smmon.mon_id.my_id.my_proc = NLM_SM_NOTIFY;
	memcpy(smmon.priv, &host->nh_sysid, sizeof(host->nh_sysid));

	timo.tv_sec = 25;
	timo.tv_usec = 0;
	stat = CLNT_CALL(nlm_nsm, SM_MON,
	    (xdrproc_t) xdr_mon, &smmon,
	    (xdrproc_t) xdr_sm_stat, &smstat, timo);

	if (stat != RPC_SUCCESS) {
		NLM_ERR("Failed to contact local NSM - rpc error %d\n", stat);
		return;
	}
	if (smstat.res_stat == stat_fail) {
		NLM_ERR("Local NSM refuses to monitor %s\n",
		    host->nh_caller_name);
		mtx_lock(&host->nh_lock);
		host->nh_monstate = NLM_MONITOR_FAILED;
		mtx_unlock(&host->nh_lock);
		return;
	}

	host->nh_monstate = NLM_MONITORED;
}

/*
 * Return an RPC client handle that can be used to talk to the NLM
 * running on the given host.
 */
CLIENT *
nlm_host_get_rpc(struct nlm_host *host, bool_t isserver)
{
	struct nlm_rpc *rpc;
	CLIENT *client;

	mtx_lock(&host->nh_lock);

	if (isserver)
		rpc = &host->nh_srvrpc;
	else
		rpc = &host->nh_clntrpc;

	/*
	 * We can't hold onto RPC handles for too long - the async
	 * call/reply protocol used by some NLM clients makes it hard
	 * to tell when they change port numbers (e.g. after a
	 * reboot). Note that if a client reboots while it isn't
	 * holding any locks, it won't bother to notify us. We
	 * expire the RPC handles after two minutes.
	 */
	if (rpc->nr_client && time_uptime > rpc->nr_create_time + 2*60) {
		client = rpc->nr_client;
		rpc->nr_client = NULL;
		mtx_unlock(&host->nh_lock);
		CLNT_RELEASE(client);
		mtx_lock(&host->nh_lock);
	}

	if (!rpc->nr_client) {
		mtx_unlock(&host->nh_lock);
		client = nlm_get_rpc((struct sockaddr *)&host->nh_addr,
		    NLM_PROG, host->nh_vers);
		mtx_lock(&host->nh_lock);

		if (client) {
			if (rpc->nr_client) {
				mtx_unlock(&host->nh_lock);
				CLNT_DESTROY(client);
				mtx_lock(&host->nh_lock);
			} else {
				rpc->nr_client = client;
				rpc->nr_create_time = time_uptime;
			}
		}
	}

	client = rpc->nr_client;
	if (client)
		CLNT_ACQUIRE(client);
	mtx_unlock(&host->nh_lock);

	return (client);

}

int nlm_host_get_sysid(struct nlm_host *host)
{

	return (host->nh_sysid);
}

int
nlm_host_get_state(struct nlm_host *host)
{

	return (host->nh_state);
}

void *
nlm_register_wait_lock(struct nlm4_lock *lock, struct vnode *vp)
{
	struct nlm_waiting_lock *nw;

	nw = malloc(sizeof(struct nlm_waiting_lock), M_NLM, M_WAITOK);
	nw->nw_lock = *lock;
	memcpy(&nw->nw_fh.fh_bytes, nw->nw_lock.fh.n_bytes,
	    nw->nw_lock.fh.n_len);
	nw->nw_lock.fh.n_bytes = nw->nw_fh.fh_bytes;
	nw->nw_waiting = TRUE;
	nw->nw_vp = vp;
	mtx_lock(&nlm_global_lock);
	TAILQ_INSERT_TAIL(&nlm_waiting_locks, nw, nw_link);
	mtx_unlock(&nlm_global_lock);

	return nw;
}

void
nlm_deregister_wait_lock(void *handle)
{
	struct nlm_waiting_lock *nw = handle;

	mtx_lock(&nlm_global_lock);
	TAILQ_REMOVE(&nlm_waiting_locks, nw, nw_link);
	mtx_unlock(&nlm_global_lock);
	
	free(nw, M_NLM);
}

int
nlm_wait_lock(void *handle, int timo)
{
	struct nlm_waiting_lock *nw = handle;
	int error;

	/*
	 * If the granted message arrived before we got here,
	 * nw->nw_waiting will be FALSE - in that case, don't sleep.
	 */
	mtx_lock(&nlm_global_lock);
	error = 0;
	if (nw->nw_waiting)
		error = msleep(nw, &nlm_global_lock, PCATCH, "nlmlock", timo);
	TAILQ_REMOVE(&nlm_waiting_locks, nw, nw_link);
	if (error) {
		/*
		 * The granted message may arrive after the
		 * interrupt/timeout but before we manage to lock the
		 * mutex. Detect this by examining nw_lock.
		 */
		if (!nw->nw_waiting)
			error = 0;
	} else {
		/*
		 * If nlm_cancel_wait is called, then error will be
		 * zero but nw_waiting will still be TRUE. We
		 * translate this into EINTR.
		 */
		if (nw->nw_waiting)
			error = EINTR;
	}
	mtx_unlock(&nlm_global_lock);

	free(nw, M_NLM);

	return (error);
}

void
nlm_cancel_wait(struct vnode *vp)
{
	struct nlm_waiting_lock *nw;

	mtx_lock(&nlm_global_lock);
	TAILQ_FOREACH(nw, &nlm_waiting_locks, nw_link) {
		if (nw->nw_vp == vp) {
			wakeup(nw);
		}
	}
	mtx_unlock(&nlm_global_lock);
}


/**********************************************************************/

/*
 * Syscall interface with userland.
 */

extern void nlm_prog_0(struct svc_req *rqstp, SVCXPRT *transp);
extern void nlm_prog_1(struct svc_req *rqstp, SVCXPRT *transp);
extern void nlm_prog_3(struct svc_req *rqstp, SVCXPRT *transp);
extern void nlm_prog_4(struct svc_req *rqstp, SVCXPRT *transp);

static int
nlm_register_services(SVCPOOL *pool, int addr_count, char **addrs)
{
	static rpcvers_t versions[] = {
		NLM_SM, NLM_VERS, NLM_VERSX, NLM_VERS4
	};
	static void (*dispatchers[])(struct svc_req *, SVCXPRT *) = {
		nlm_prog_0, nlm_prog_1, nlm_prog_3, nlm_prog_4
	};
	static const int version_count = sizeof(versions) / sizeof(versions[0]);

	SVCXPRT **xprts;
	char netid[16];
	char uaddr[128];
	struct netconfig *nconf;
	int i, j, error;

	if (!addr_count) {
		NLM_ERR("NLM: no service addresses given - can't start server");
		return (EINVAL);
	}

	xprts = malloc(addr_count * sizeof(SVCXPRT *), M_NLM, M_WAITOK|M_ZERO);
	for (i = 0; i < version_count; i++) {
		for (j = 0; j < addr_count; j++) {
			/*
			 * Create transports for the first version and
			 * then just register everything else to the
			 * same transports.
			 */
			if (i == 0) {
				char *up;

				error = copyin(&addrs[2*j], &up,
				    sizeof(char*));
				if (error)
					goto out;
				error = copyinstr(up, netid, sizeof(netid),
				    NULL);
				if (error)
					goto out;
				error = copyin(&addrs[2*j+1], &up,
				    sizeof(char*));
				if (error)
					goto out;
				error = copyinstr(up, uaddr, sizeof(uaddr),
				    NULL);
				if (error)
					goto out;
				nconf = getnetconfigent(netid);
				if (!nconf) {
					NLM_ERR("Can't lookup netid %s\n",
					    netid);
					error = EINVAL;
					goto out;
				}
				xprts[j] = svc_tp_create(pool, dispatchers[i],
				    NLM_PROG, versions[i], uaddr, nconf);
				if (!xprts[j]) {
					NLM_ERR("NLM: unable to create "
					    "(NLM_PROG, %d).\n", versions[i]);
					error = EINVAL;
					goto out;
				}
				freenetconfigent(nconf);
			} else {
				nconf = getnetconfigent(xprts[j]->xp_netid);
				rpcb_unset(NLM_PROG, versions[i], nconf);
				if (!svc_reg(xprts[j], NLM_PROG, versions[i],
					dispatchers[i], nconf)) {
					NLM_ERR("NLM: can't register "
					    "(NLM_PROG, %d)\n", versions[i]);
					error = EINVAL;
					goto out;
				}
			}
		}
	}
	error = 0;
out:
	for (j = 0; j < addr_count; j++) {
		if (xprts[j])
			SVC_RELEASE(xprts[j]);
	}
	free(xprts, M_NLM);
	return (error);
}

/*
 * Main server entry point. Contacts the local NSM to get its current
 * state and send SM_UNMON_ALL. Registers the NLM services and then
 * services requests. Does not return until the server is interrupted
 * by a signal.
 */
static int
nlm_server_main(int addr_count, char **addrs)
{
	struct thread *td = curthread;
	int error;
	SVCPOOL *pool = NULL;
	struct sockopt opt;
	int portlow;
#ifdef INET6
	struct sockaddr_in6 sin6;
#endif
	struct sockaddr_in sin;
	my_id id;
	sm_stat smstat;
	struct timeval timo;
	enum clnt_stat stat;
	struct nlm_host *host, *nhost;
	struct nlm_waiting_lock *nw;
	vop_advlock_t *old_nfs_advlock;
	vop_reclaim_t *old_nfs_reclaim;

	if (nlm_is_running != 0) {
		NLM_ERR("NLM: can't start server - "
		    "it appears to be running already\n");
		return (EPERM);
	}

	if (nlm_socket == NULL) {
		memset(&opt, 0, sizeof(opt));

		error = socreate(AF_INET, &nlm_socket, SOCK_DGRAM, 0,
		    td->td_ucred, td);
		if (error) {
			NLM_ERR("NLM: can't create IPv4 socket - error %d\n",
			    error);
			return (error);
		}
		opt.sopt_dir = SOPT_SET;
		opt.sopt_level = IPPROTO_IP;
		opt.sopt_name = IP_PORTRANGE;
		portlow = IP_PORTRANGE_LOW;
		opt.sopt_val = &portlow;
		opt.sopt_valsize = sizeof(portlow);
		sosetopt(nlm_socket, &opt);

#ifdef INET6
		nlm_socket6 = NULL;
		error = socreate(AF_INET6, &nlm_socket6, SOCK_DGRAM, 0,
		    td->td_ucred, td);
		if (error) {
			NLM_ERR("NLM: can't create IPv6 socket - error %d\n",
			    error);
			soclose(nlm_socket);
			nlm_socket = NULL;
			return (error);
		}
		opt.sopt_dir = SOPT_SET;
		opt.sopt_level = IPPROTO_IPV6;
		opt.sopt_name = IPV6_PORTRANGE;
		portlow = IPV6_PORTRANGE_LOW;
		opt.sopt_val = &portlow;
		opt.sopt_valsize = sizeof(portlow);
		sosetopt(nlm_socket6, &opt);
#endif
	}

	nlm_auth = authunix_create(curthread->td_ucred);

#ifdef INET6
	memset(&sin6, 0, sizeof(sin6));
	sin6.sin6_len = sizeof(sin6);
	sin6.sin6_family = AF_INET6;
	sin6.sin6_addr = in6addr_loopback;
	nlm_nsm = nlm_get_rpc((struct sockaddr *) &sin6, SM_PROG, SM_VERS);
	if (!nlm_nsm) {
#endif
		memset(&sin, 0, sizeof(sin));
		sin.sin_len = sizeof(sin);
		sin.sin_family = AF_INET;
		sin.sin_addr.s_addr = htonl(INADDR_LOOPBACK);
		nlm_nsm = nlm_get_rpc((struct sockaddr *) &sin, SM_PROG,
		    SM_VERS);
#ifdef INET6
	}
#endif

	if (!nlm_nsm) {
		NLM_ERR("Can't start NLM - unable to contact NSM\n");
		error = EINVAL;
		goto out;
	}

	pool = svcpool_create("NLM", NULL);

	error = nlm_register_services(pool, addr_count, addrs);
	if (error)
		goto out;

	memset(&id, 0, sizeof(id));
	id.my_name = "NFS NLM";

	timo.tv_sec = 25;
	timo.tv_usec = 0;
	stat = CLNT_CALL(nlm_nsm, SM_UNMON_ALL,
	    (xdrproc_t) xdr_my_id, &id,
	    (xdrproc_t) xdr_sm_stat, &smstat, timo);

	if (stat != RPC_SUCCESS) {
		struct rpc_err err;

		CLNT_GETERR(nlm_nsm, &err);
		NLM_ERR("NLM: unexpected error contacting NSM, "
		    "stat=%d, errno=%d\n", stat, err.re_errno);
		error = EINVAL;
		goto out;
	}
	nlm_is_running = 1;

	NLM_DEBUG(1, "NLM: local NSM state is %d\n", smstat.state);
	nlm_nsm_state = smstat.state;

	old_nfs_advlock = nfs_advlock_p;
	nfs_advlock_p = nlm_advlock;
	old_nfs_reclaim = nfs_reclaim_p;
	nfs_reclaim_p = nlm_reclaim;

	svc_run(pool);
	error = 0;

	nfs_advlock_p = old_nfs_advlock;
	nfs_reclaim_p = old_nfs_reclaim;

out:
	nlm_is_running = 0;
	if (pool)
		svcpool_destroy(pool);

	/*
	 * We are finished communicating with the NSM.
	 */
	if (nlm_nsm) {
		CLNT_RELEASE(nlm_nsm);
		nlm_nsm = NULL;
	}

	/*
	 * Trash all the existing state so that if the server
	 * restarts, it gets a clean slate. This is complicated by the
	 * possibility that there may be other threads trying to make
	 * client locking requests.
	 *
	 * First we fake a client reboot notification which will
	 * cancel any pending async locks and purge remote lock state
	 * from the local lock manager. We release the reference from
	 * nlm_hosts to the host (which may remove it from the list
	 * and free it). After this phase, the only entries in the
	 * nlm_host list should be from other threads performing
	 * client lock requests.
	 */
	mtx_lock(&nlm_global_lock);
	TAILQ_FOREACH(nw, &nlm_waiting_locks, nw_link) {
		wakeup(nw);
	}
	TAILQ_FOREACH_SAFE(host, &nlm_hosts, nh_link, nhost) {
		mtx_unlock(&nlm_global_lock);
		nlm_host_notify(host, 0);
		nlm_host_release(host);
		mtx_lock(&nlm_global_lock);
	}
	mtx_unlock(&nlm_global_lock);

	AUTH_DESTROY(nlm_auth);

	return (error);
}

int
sys_nlm_syscall(struct thread *td, struct nlm_syscall_args *uap)
{
	int error;

#if __FreeBSD_version >= 700000
	error = priv_check(td, PRIV_NFS_LOCKD);
#else
	error = suser(td);
#endif
	if (error)
		return (error);

	nlm_debug_level = uap->debug_level;
	nlm_grace_threshold = time_uptime + uap->grace_period;
	nlm_next_idle_check = time_uptime + NLM_IDLE_PERIOD;

	return nlm_server_main(uap->addr_count, uap->addrs);
}

/**********************************************************************/

/*
 * NLM implementation details, called from the RPC stubs.
 */


void
nlm_sm_notify(struct nlm_sm_status *argp)
{
	uint32_t sysid;
	struct nlm_host *host;

	NLM_DEBUG(3, "nlm_sm_notify(): mon_name = %s\n", argp->mon_name);
	memcpy(&sysid, &argp->priv, sizeof(sysid));
	host = nlm_find_host_by_sysid(sysid);
	if (host) {
		nlm_host_notify(host, argp->state);
		nlm_host_release(host);
	}
}

static void
nlm_convert_to_fhandle_t(fhandle_t *fhp, struct netobj *p)
{
	memcpy(fhp, p->n_bytes, sizeof(fhandle_t));
}

struct vfs_state {
	struct mount	*vs_mp;
	struct vnode	*vs_vp;
	int		vs_vnlocked;
};

static int
nlm_get_vfs_state(struct nlm_host *host, struct svc_req *rqstp,
    fhandle_t *fhp, struct vfs_state *vs, accmode_t accmode)
{
	int error, exflags;
	struct ucred *cred = NULL, *credanon = NULL;
	
	memset(vs, 0, sizeof(*vs));

	vs->vs_mp = vfs_getvfs(&fhp->fh_fsid);
	if (!vs->vs_mp) {
		return (ESTALE);
	}

	/* accmode == 0 means don't check, since it is an unlock. */
	if (accmode != 0) {
		error = VFS_CHECKEXP(vs->vs_mp,
		    (struct sockaddr *)&host->nh_addr, &exflags, &credanon,
		    NULL, NULL);
		if (error)
			goto out;

		if (exflags & MNT_EXRDONLY ||
		    (vs->vs_mp->mnt_flag & MNT_RDONLY)) {
			error = EROFS;
			goto out;
		}
	}

	error = VFS_FHTOVP(vs->vs_mp, &fhp->fh_fid, LK_EXCLUSIVE, &vs->vs_vp);
	if (error)
		goto out;
	vs->vs_vnlocked = TRUE;

	if (accmode != 0) {
		if (!svc_getcred(rqstp, &cred, NULL)) {
			error = EINVAL;
			goto out;
		}
		if (cred->cr_uid == 0 || (exflags & MNT_EXPORTANON)) {
			crfree(cred);
			cred = credanon;
			credanon = NULL;
		}

		/*
		 * Check cred.
		 */
		error = VOP_ACCESS(vs->vs_vp, accmode, cred, curthread);
		/*
		 * If this failed and accmode != VWRITE, try again with
		 * VWRITE to maintain backwards compatibility with the
		 * old code that always used VWRITE.
		 */
		if (error != 0 && accmode != VWRITE)
			error = VOP_ACCESS(vs->vs_vp, VWRITE, cred, curthread);
		if (error)
			goto out;
	}

#if __FreeBSD_version < 800011
	VOP_UNLOCK(vs->vs_vp, 0, curthread);
#else
	VOP_UNLOCK(vs->vs_vp, 0);
#endif
	vs->vs_vnlocked = FALSE;

out:
	if (cred)
		crfree(cred);
	if (credanon)
		crfree(credanon);

	return (error);
}

static void
nlm_release_vfs_state(struct vfs_state *vs)
{

	if (vs->vs_vp) {
		if (vs->vs_vnlocked)
			vput(vs->vs_vp);
		else
			vrele(vs->vs_vp);
	}
	if (vs->vs_mp)
		vfs_rel(vs->vs_mp);
}

static nlm4_stats
nlm_convert_error(int error)
{

	if (error == ESTALE)
		return nlm4_stale_fh;
	else if (error == EROFS)
		return nlm4_rofs;
	else
		return nlm4_failed;
}

int
nlm_do_test(nlm4_testargs *argp, nlm4_testres *result, struct svc_req *rqstp,
	CLIENT **rpcp)
{
	fhandle_t fh;
	struct vfs_state vs;
	struct nlm_host *host, *bhost;
	int error, sysid;
	struct flock fl;
	accmode_t accmode;
	
	memset(result, 0, sizeof(*result));
	memset(&vs, 0, sizeof(vs));

	host = nlm_find_host_by_name(argp->alock.caller_name,
	    svc_getrpccaller(rqstp), rqstp->rq_vers);
	if (!host) {
		result->stat.stat = nlm4_denied_nolocks;
		return (ENOMEM);
	}

	NLM_DEBUG(3, "nlm_do_test(): caller_name = %s (sysid = %d)\n",
	    host->nh_caller_name, host->nh_sysid);

	nlm_check_expired_locks(host);
	sysid = host->nh_sysid;

	nlm_convert_to_fhandle_t(&fh, &argp->alock.fh);
	nlm_copy_netobj(&result->cookie, &argp->cookie, M_RPC);

	if (time_uptime < nlm_grace_threshold) {
		result->stat.stat = nlm4_denied_grace_period;
		goto out;
	}

	accmode = argp->exclusive ? VWRITE : VREAD;
	error = nlm_get_vfs_state(host, rqstp, &fh, &vs, accmode);
	if (error) {
		result->stat.stat = nlm_convert_error(error);
		goto out;
	}

	fl.l_start = argp->alock.l_offset;
	fl.l_len = argp->alock.l_len;
	fl.l_pid = argp->alock.svid;
	fl.l_sysid = sysid;
	fl.l_whence = SEEK_SET;
	if (argp->exclusive)
		fl.l_type = F_WRLCK;
	else
		fl.l_type = F_RDLCK;
	error = VOP_ADVLOCK(vs.vs_vp, NULL, F_GETLK, &fl, F_REMOTE);
	if (error) {
		result->stat.stat = nlm4_failed;
		goto out;
	}

	if (fl.l_type == F_UNLCK) {
		result->stat.stat = nlm4_granted;
	} else {
		result->stat.stat = nlm4_denied;
		result->stat.nlm4_testrply_u.holder.exclusive =
			(fl.l_type == F_WRLCK);
		result->stat.nlm4_testrply_u.holder.svid = fl.l_pid;
		bhost = nlm_find_host_by_sysid(fl.l_sysid);
		if (bhost) {
			/*
			 * We don't have any useful way of recording
			 * the value of oh used in the original lock
			 * request. Ideally, the test reply would have
			 * a space for the owning host's name allowing
			 * our caller's NLM to keep track.
			 *
			 * As far as I can see, Solaris uses an eight
			 * byte structure for oh which contains a four
			 * byte pid encoded in local byte order and
			 * the first four bytes of the host
			 * name. Linux uses a variable length string
			 * 'pid@hostname' in ascii but doesn't even
			 * return that in test replies.
			 *
			 * For the moment, return nothing in oh
			 * (already zero'ed above).
			 */
			nlm_host_release(bhost);
		}
		result->stat.nlm4_testrply_u.holder.l_offset = fl.l_start;
		result->stat.nlm4_testrply_u.holder.l_len = fl.l_len;
	}

out:
	nlm_release_vfs_state(&vs);
	if (rpcp)
		*rpcp = nlm_host_get_rpc(host, TRUE);
	nlm_host_release(host);
	return (0);
}

int
nlm_do_lock(nlm4_lockargs *argp, nlm4_res *result, struct svc_req *rqstp,
    bool_t monitor, CLIENT **rpcp)
{
	fhandle_t fh;
	struct vfs_state vs;
	struct nlm_host *host;
	int error, sysid;
	struct flock fl;
	accmode_t accmode;
	
	memset(result, 0, sizeof(*result));
	memset(&vs, 0, sizeof(vs));

	host = nlm_find_host_by_name(argp->alock.caller_name,
	    svc_getrpccaller(rqstp), rqstp->rq_vers);
	if (!host) {
		result->stat.stat = nlm4_denied_nolocks;
		return (ENOMEM);
	}

	NLM_DEBUG(3, "nlm_do_lock(): caller_name = %s (sysid = %d)\n",
	    host->nh_caller_name, host->nh_sysid);

	if (monitor && host->nh_state && argp->state
	    && host->nh_state != argp->state) {
		/*
		 * The host rebooted without telling us. Trash its
		 * locks.
		 */
		nlm_host_notify(host, argp->state);
	}

	nlm_check_expired_locks(host);
	sysid = host->nh_sysid;

	nlm_convert_to_fhandle_t(&fh, &argp->alock.fh);
	nlm_copy_netobj(&result->cookie, &argp->cookie, M_RPC);

	if (time_uptime < nlm_grace_threshold && !argp->reclaim) {
		result->stat.stat = nlm4_denied_grace_period;
		goto out;
	}

	accmode = argp->exclusive ? VWRITE : VREAD;
	error = nlm_get_vfs_state(host, rqstp, &fh, &vs, accmode);
	if (error) {
		result->stat.stat = nlm_convert_error(error);
		goto out;
	}

	fl.l_start = argp->alock.l_offset;
	fl.l_len = argp->alock.l_len;
	fl.l_pid = argp->alock.svid;
	fl.l_sysid = sysid;
	fl.l_whence = SEEK_SET;
	if (argp->exclusive)
		fl.l_type = F_WRLCK;
	else
		fl.l_type = F_RDLCK;
	if (argp->block) {
		struct nlm_async_lock *af;
		CLIENT *client;
		struct nlm_grantcookie cookie;

		/*
		 * First, make sure we can contact the host's NLM.
		 */
		client = nlm_host_get_rpc(host, TRUE);
		if (!client) {
			result->stat.stat = nlm4_failed;
			goto out;
		}

		/*
		 * First we need to check and see if there is an
		 * existing blocked lock that matches. This could be a
		 * badly behaved client or an RPC re-send. If we find
		 * one, just return nlm4_blocked.
		 */
		mtx_lock(&host->nh_lock);
		TAILQ_FOREACH(af, &host->nh_pending, af_link) {
			if (af->af_fl.l_start == fl.l_start
			    && af->af_fl.l_len == fl.l_len
			    && af->af_fl.l_pid == fl.l_pid
			    && af->af_fl.l_type == fl.l_type) {
				break;
			}
		}
		if (!af) {
			cookie.ng_sysid = host->nh_sysid;
			cookie.ng_cookie = host->nh_grantcookie++;
		}
		mtx_unlock(&host->nh_lock);
		if (af) {
			CLNT_RELEASE(client);
			result->stat.stat = nlm4_blocked;
			goto out;
		}

		af = malloc(sizeof(struct nlm_async_lock), M_NLM,
		    M_WAITOK|M_ZERO);
		TASK_INIT(&af->af_task, 0, nlm_lock_callback, af);
		af->af_vp = vs.vs_vp;
		af->af_fl = fl;
		af->af_host = host;
		af->af_rpc = client;
		/*
		 * We use M_RPC here so that we can xdr_free the thing
		 * later.
		 */
		nlm_make_netobj(&af->af_granted.cookie,
		    (caddr_t)&cookie, sizeof(cookie), M_RPC);
		af->af_granted.exclusive = argp->exclusive;
		af->af_granted.alock.caller_name =
			strdup(argp->alock.caller_name, M_RPC);
		nlm_copy_netobj(&af->af_granted.alock.fh,
		    &argp->alock.fh, M_RPC);
		nlm_copy_netobj(&af->af_granted.alock.oh,
		    &argp->alock.oh, M_RPC);
		af->af_granted.alock.svid = argp->alock.svid;
		af->af_granted.alock.l_offset = argp->alock.l_offset;
		af->af_granted.alock.l_len = argp->alock.l_len;

		/*
		 * Put the entry on the pending list before calling
		 * VOP_ADVLOCKASYNC. We do this in case the lock
		 * request was blocked (returning EINPROGRESS) but
		 * then granted before we manage to run again. The
		 * client may receive the granted message before we
		 * send our blocked reply but thats their problem.
		 */
		mtx_lock(&host->nh_lock);
		TAILQ_INSERT_TAIL(&host->nh_pending, af, af_link);
		mtx_unlock(&host->nh_lock);

		error = VOP_ADVLOCKASYNC(vs.vs_vp, NULL, F_SETLK, &fl, F_REMOTE,
		    &af->af_task, &af->af_cookie);

		/*
		 * If the lock completed synchronously, just free the
		 * tracking structure now.
		 */
		if (error != EINPROGRESS) {
			CLNT_RELEASE(af->af_rpc);
			mtx_lock(&host->nh_lock);
			TAILQ_REMOVE(&host->nh_pending, af, af_link);
			mtx_unlock(&host->nh_lock);
			xdr_free((xdrproc_t) xdr_nlm4_testargs,
			    &af->af_granted);
			free(af, M_NLM);
		} else {
			NLM_DEBUG(2, "NLM: pending async lock %p for %s "
			    "(sysid %d)\n", af, host->nh_caller_name, sysid);
			/*
			 * Don't vrele the vnode just yet - this must
			 * wait until either the async callback
			 * happens or the lock is cancelled.
			 */
			vs.vs_vp = NULL;
		}
	} else {
		error = VOP_ADVLOCK(vs.vs_vp, NULL, F_SETLK, &fl, F_REMOTE);
	}

	if (error) {
		if (error == EINPROGRESS) {
			result->stat.stat = nlm4_blocked;
		} else if (error == EDEADLK) {
			result->stat.stat = nlm4_deadlck;
		} else if (error == EAGAIN) {
			result->stat.stat = nlm4_denied;
		} else {
			result->stat.stat = nlm4_failed;
		}
	} else {
		if (monitor)
			nlm_host_monitor(host, argp->state);
		result->stat.stat = nlm4_granted;
	}       

out:
	nlm_release_vfs_state(&vs);
	if (rpcp)
		*rpcp = nlm_host_get_rpc(host, TRUE);
	nlm_host_release(host);
	return (0);
}

int
nlm_do_cancel(nlm4_cancargs *argp, nlm4_res *result, struct svc_req *rqstp,
    CLIENT **rpcp)
{
	fhandle_t fh;
	struct vfs_state vs;
	struct nlm_host *host;
	int error, sysid;
	struct flock fl;
	struct nlm_async_lock *af;
	
	memset(result, 0, sizeof(*result));
	memset(&vs, 0, sizeof(vs));

	host = nlm_find_host_by_name(argp->alock.caller_name,
	    svc_getrpccaller(rqstp), rqstp->rq_vers);
	if (!host) {
		result->stat.stat = nlm4_denied_nolocks;
		return (ENOMEM);
	}

	NLM_DEBUG(3, "nlm_do_cancel(): caller_name = %s (sysid = %d)\n",
	    host->nh_caller_name, host->nh_sysid);

	nlm_check_expired_locks(host);
	sysid = host->nh_sysid;

	nlm_convert_to_fhandle_t(&fh, &argp->alock.fh);
	nlm_copy_netobj(&result->cookie, &argp->cookie, M_RPC);

	if (time_uptime < nlm_grace_threshold) {
		result->stat.stat = nlm4_denied_grace_period;
		goto out;
	}

	error = nlm_get_vfs_state(host, rqstp, &fh, &vs, (accmode_t)0);
	if (error) {
		result->stat.stat = nlm_convert_error(error);
		goto out;
	}

	fl.l_start = argp->alock.l_offset;
	fl.l_len = argp->alock.l_len;
	fl.l_pid = argp->alock.svid;
	fl.l_sysid = sysid;
	fl.l_whence = SEEK_SET;
	if (argp->exclusive)
		fl.l_type = F_WRLCK;
	else
		fl.l_type = F_RDLCK;

	/*
	 * First we need to try and find the async lock request - if
	 * there isn't one, we give up and return nlm4_denied.
	 */
	mtx_lock(&host->nh_lock);

	TAILQ_FOREACH(af, &host->nh_pending, af_link) {
		if (af->af_fl.l_start == fl.l_start
		    && af->af_fl.l_len == fl.l_len
		    && af->af_fl.l_pid == fl.l_pid
		    && af->af_fl.l_type == fl.l_type) {
			break;
		}
	}

	if (!af) {
		mtx_unlock(&host->nh_lock);
		result->stat.stat = nlm4_denied;
		goto out;
	}

	error = nlm_cancel_async_lock(af);

	if (error) {
		result->stat.stat = nlm4_denied;
	} else {
		result->stat.stat = nlm4_granted;
	}

	mtx_unlock(&host->nh_lock);

out:
	nlm_release_vfs_state(&vs);
	if (rpcp)
		*rpcp = nlm_host_get_rpc(host, TRUE);
	nlm_host_release(host);
	return (0);
}

int
nlm_do_unlock(nlm4_unlockargs *argp, nlm4_res *result, struct svc_req *rqstp,
    CLIENT **rpcp)
{
	fhandle_t fh;
	struct vfs_state vs;
	struct nlm_host *host;
	int error, sysid;
	struct flock fl;
	
	memset(result, 0, sizeof(*result));
	memset(&vs, 0, sizeof(vs));

	host = nlm_find_host_by_name(argp->alock.caller_name,
	    svc_getrpccaller(rqstp), rqstp->rq_vers);
	if (!host) {
		result->stat.stat = nlm4_denied_nolocks;
		return (ENOMEM);
	}

	NLM_DEBUG(3, "nlm_do_unlock(): caller_name = %s (sysid = %d)\n",
	    host->nh_caller_name, host->nh_sysid);

	nlm_check_expired_locks(host);
	sysid = host->nh_sysid;

	nlm_convert_to_fhandle_t(&fh, &argp->alock.fh);
	nlm_copy_netobj(&result->cookie, &argp->cookie, M_RPC);

	if (time_uptime < nlm_grace_threshold) {
		result->stat.stat = nlm4_denied_grace_period;
		goto out;
	}

	error = nlm_get_vfs_state(host, rqstp, &fh, &vs, (accmode_t)0);
	if (error) {
		result->stat.stat = nlm_convert_error(error);
		goto out;
	}

	fl.l_start = argp->alock.l_offset;
	fl.l_len = argp->alock.l_len;
	fl.l_pid = argp->alock.svid;
	fl.l_sysid = sysid;
	fl.l_whence = SEEK_SET;
	fl.l_type = F_UNLCK;
	error = VOP_ADVLOCK(vs.vs_vp, NULL, F_UNLCK, &fl, F_REMOTE);

	/*
	 * Ignore the error - there is no result code for failure,
	 * only for grace period.
	 */
	result->stat.stat = nlm4_granted;

out:
	nlm_release_vfs_state(&vs);
	if (rpcp)
		*rpcp = nlm_host_get_rpc(host, TRUE);
	nlm_host_release(host);
	return (0);
}

int
nlm_do_granted(nlm4_testargs *argp, nlm4_res *result, struct svc_req *rqstp,

    CLIENT **rpcp)
{
	struct nlm_host *host;
	struct nlm_waiting_lock *nw;
	
	memset(result, 0, sizeof(*result));

	host = nlm_find_host_by_addr(svc_getrpccaller(rqstp), rqstp->rq_vers);
	if (!host) {
		result->stat.stat = nlm4_denied_nolocks;
		return (ENOMEM);
	}

	nlm_copy_netobj(&result->cookie, &argp->cookie, M_RPC);
	result->stat.stat = nlm4_denied;
	KFAIL_POINT_CODE(DEBUG_FP, nlm_deny_grant, goto out);

	mtx_lock(&nlm_global_lock);
	TAILQ_FOREACH(nw, &nlm_waiting_locks, nw_link) {
		if (!nw->nw_waiting)
			continue;
		if (argp->alock.svid == nw->nw_lock.svid
		    && argp->alock.l_offset == nw->nw_lock.l_offset
		    && argp->alock.l_len == nw->nw_lock.l_len
		    && argp->alock.fh.n_len == nw->nw_lock.fh.n_len
		    && !memcmp(argp->alock.fh.n_bytes, nw->nw_lock.fh.n_bytes,
			nw->nw_lock.fh.n_len)) {
			nw->nw_waiting = FALSE;
			wakeup(nw);
			result->stat.stat = nlm4_granted;
			break;
		}
	}
	mtx_unlock(&nlm_global_lock);

out:
	if (rpcp)
		*rpcp = nlm_host_get_rpc(host, TRUE);
	nlm_host_release(host);
	return (0);
}

void
nlm_do_granted_res(nlm4_res *argp, struct svc_req *rqstp)
{
	struct nlm_host *host = NULL;
	struct nlm_async_lock *af = NULL;
	int error;

	if (argp->cookie.n_len != sizeof(struct nlm_grantcookie)) {
		NLM_DEBUG(1, "NLM: bogus grant cookie");
		goto out;
	}

	host = nlm_find_host_by_sysid(ng_sysid(&argp->cookie));
	if (!host) {
		NLM_DEBUG(1, "NLM: Unknown host rejected our grant");
		goto out;
	}

	mtx_lock(&host->nh_lock);
	TAILQ_FOREACH(af, &host->nh_granted, af_link)
	    if (ng_cookie(&argp->cookie) ==
		ng_cookie(&af->af_granted.cookie))
		    break;
	if (af)
		TAILQ_REMOVE(&host->nh_granted, af, af_link);
	mtx_unlock(&host->nh_lock);

	if (!af) {
		NLM_DEBUG(1, "NLM: host %s (sysid %d) replied to our grant "
		    "with unrecognized cookie %d:%d", host->nh_caller_name,
		    host->nh_sysid, ng_sysid(&argp->cookie),
		    ng_cookie(&argp->cookie));
		goto out;
	}

	if (argp->stat.stat != nlm4_granted) {
		af->af_fl.l_type = F_UNLCK;
		error = VOP_ADVLOCK(af->af_vp, NULL, F_UNLCK, &af->af_fl, F_REMOTE);
		if (error) {
			NLM_DEBUG(1, "NLM: host %s (sysid %d) rejected our grant "
			    "and we failed to unlock (%d)", host->nh_caller_name,
			    host->nh_sysid, error);
			goto out;
		}

		NLM_DEBUG(5, "NLM: async lock %p rejected by host %s (sysid %d)",
		    af, host->nh_caller_name, host->nh_sysid);
	} else {
		NLM_DEBUG(5, "NLM: async lock %p accepted by host %s (sysid %d)",
		    af, host->nh_caller_name, host->nh_sysid);
	}

 out:
	if (af)
		nlm_free_async_lock(af);
	if (host)
		nlm_host_release(host);
}

void
nlm_do_free_all(nlm4_notify *argp)
{
	struct nlm_host *host, *thost;

	TAILQ_FOREACH_SAFE(host, &nlm_hosts, nh_link, thost) {
		if (!strcmp(host->nh_caller_name, argp->name))
			nlm_host_notify(host, argp->state);
	}
}

/*
 * Kernel module glue
 */
static int
nfslockd_modevent(module_t mod, int type, void *data)
{

	switch (type) {
	case MOD_LOAD:
		return (0);
	case MOD_UNLOAD:
		/* The NLM module cannot be safely unloaded. */
		/* FALLTHROUGH */
	default:
		return (EOPNOTSUPP);
	}
}
static moduledata_t nfslockd_mod = {
	"nfslockd",
	nfslockd_modevent,
	NULL,
};
DECLARE_MODULE(nfslockd, nfslockd_mod, SI_SUB_VFS, SI_ORDER_ANY);

/* So that loader and kldload(2) can find us, wherever we are.. */
MODULE_DEPEND(nfslockd, krpc, 1, 1, 1);
MODULE_DEPEND(nfslockd, nfslock, 1, 1, 1);
MODULE_VERSION(nfslockd, 1);
OpenPOWER on IntegriCloud