summaryrefslogtreecommitdiff
path: root/daemons/lvmlockd/lvmlockd-core.c
blob: 7b5dd277caf289330d6ec47ff8927942daf0bcb0 (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
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
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
3522
3523
3524
3525
3526
3527
3528
3529
3530
3531
3532
3533
3534
3535
3536
3537
3538
3539
3540
3541
3542
3543
3544
3545
3546
3547
3548
3549
3550
3551
3552
3553
3554
3555
3556
3557
3558
3559
3560
3561
3562
3563
3564
3565
3566
3567
3568
3569
3570
3571
3572
3573
3574
3575
3576
3577
3578
3579
3580
3581
3582
3583
3584
3585
3586
3587
3588
3589
3590
3591
3592
3593
3594
3595
3596
3597
3598
3599
3600
3601
3602
3603
3604
3605
3606
3607
3608
3609
3610
3611
3612
3613
3614
3615
3616
3617
3618
3619
3620
3621
3622
3623
3624
3625
3626
3627
3628
3629
3630
3631
3632
3633
3634
3635
3636
3637
3638
3639
3640
3641
3642
3643
3644
3645
3646
3647
3648
3649
3650
3651
3652
3653
3654
3655
3656
3657
3658
3659
3660
3661
3662
3663
3664
3665
3666
3667
3668
3669
3670
3671
3672
3673
3674
3675
3676
3677
3678
3679
3680
3681
3682
3683
3684
3685
3686
3687
3688
3689
3690
3691
3692
3693
3694
3695
3696
3697
3698
3699
3700
3701
3702
3703
3704
3705
3706
3707
3708
3709
3710
3711
3712
3713
3714
3715
3716
3717
3718
3719
3720
3721
3722
3723
3724
3725
3726
3727
3728
3729
3730
3731
3732
3733
3734
3735
3736
3737
3738
3739
3740
3741
3742
3743
3744
3745
3746
3747
3748
3749
3750
3751
3752
3753
3754
3755
3756
3757
3758
3759
3760
3761
3762
3763
3764
3765
3766
3767
3768
3769
3770
3771
3772
3773
3774
3775
3776
3777
3778
3779
3780
3781
3782
3783
3784
3785
3786
3787
3788
3789
3790
3791
3792
3793
3794
3795
3796
3797
3798
3799
3800
3801
3802
3803
3804
3805
3806
3807
3808
3809
3810
3811
3812
3813
3814
3815
3816
3817
3818
3819
3820
3821
3822
3823
3824
3825
3826
3827
3828
3829
3830
3831
3832
3833
3834
3835
3836
3837
3838
3839
3840
3841
3842
3843
3844
3845
3846
3847
3848
3849
3850
3851
3852
3853
3854
3855
3856
3857
3858
3859
3860
3861
3862
3863
3864
3865
3866
3867
3868
3869
3870
3871
3872
3873
3874
3875
3876
3877
3878
3879
3880
3881
3882
3883
3884
3885
3886
3887
3888
3889
3890
3891
3892
3893
3894
3895
3896
3897
3898
3899
3900
3901
3902
3903
3904
3905
3906
3907
3908
3909
3910
3911
3912
3913
3914
3915
3916
3917
3918
3919
3920
3921
3922
3923
3924
3925
3926
3927
3928
3929
3930
3931
3932
3933
3934
3935
3936
3937
3938
3939
3940
3941
3942
3943
3944
3945
3946
3947
3948
3949
3950
3951
3952
3953
3954
3955
3956
3957
3958
3959
3960
3961
3962
3963
3964
3965
3966
3967
3968
3969
3970
3971
3972
3973
3974
3975
3976
3977
3978
3979
3980
3981
3982
3983
3984
3985
3986
3987
3988
3989
3990
3991
3992
3993
3994
3995
3996
3997
3998
3999
4000
4001
4002
4003
4004
4005
4006
4007
4008
4009
4010
4011
4012
4013
4014
4015
4016
4017
4018
4019
4020
4021
4022
4023
4024
4025
4026
4027
4028
4029
4030
4031
4032
4033
4034
4035
4036
4037
4038
4039
4040
4041
4042
4043
4044
4045
4046
4047
4048
4049
4050
4051
4052
4053
4054
4055
4056
4057
4058
4059
4060
4061
4062
4063
4064
4065
4066
4067
4068
4069
4070
4071
4072
4073
4074
4075
4076
4077
4078
4079
4080
4081
4082
4083
4084
4085
4086
4087
4088
4089
4090
4091
4092
4093
4094
4095
4096
4097
4098
4099
4100
4101
4102
4103
4104
4105
4106
4107
4108
4109
4110
4111
4112
4113
4114
4115
4116
4117
4118
4119
4120
4121
4122
4123
4124
4125
4126
4127
4128
4129
4130
4131
4132
4133
4134
4135
4136
4137
4138
4139
4140
4141
4142
4143
4144
4145
4146
4147
4148
4149
4150
4151
4152
4153
4154
4155
4156
4157
4158
4159
4160
4161
4162
4163
4164
4165
4166
4167
4168
4169
4170
4171
4172
4173
4174
4175
4176
4177
4178
4179
4180
4181
4182
4183
4184
4185
4186
4187
4188
4189
4190
4191
4192
4193
4194
4195
4196
4197
4198
4199
4200
4201
4202
4203
4204
4205
4206
4207
4208
4209
4210
4211
4212
4213
4214
4215
4216
4217
4218
4219
4220
4221
4222
4223
4224
4225
4226
4227
4228
4229
4230
4231
4232
4233
4234
4235
4236
4237
4238
4239
4240
4241
4242
4243
4244
4245
4246
4247
4248
4249
4250
4251
4252
4253
4254
4255
4256
4257
4258
4259
4260
4261
4262
4263
4264
4265
4266
4267
4268
4269
4270
4271
4272
4273
4274
4275
4276
4277
4278
4279
4280
4281
4282
4283
4284
4285
4286
4287
4288
4289
4290
4291
4292
4293
4294
4295
4296
4297
4298
4299
4300
4301
4302
4303
4304
4305
4306
4307
4308
4309
4310
4311
4312
4313
4314
4315
4316
4317
4318
4319
4320
4321
4322
4323
4324
4325
4326
4327
4328
4329
4330
4331
4332
4333
4334
4335
4336
4337
4338
4339
4340
4341
4342
4343
4344
4345
4346
4347
4348
4349
4350
4351
4352
4353
4354
4355
4356
4357
4358
4359
4360
4361
4362
4363
4364
4365
4366
4367
4368
4369
4370
4371
4372
4373
4374
4375
4376
4377
4378
4379
4380
4381
4382
4383
4384
4385
4386
4387
4388
4389
4390
4391
4392
4393
4394
4395
4396
4397
4398
4399
4400
4401
4402
4403
4404
4405
4406
4407
4408
4409
4410
4411
4412
4413
4414
4415
4416
4417
4418
4419
4420
4421
4422
4423
4424
4425
4426
4427
4428
4429
4430
4431
4432
4433
4434
4435
4436
4437
4438
4439
4440
4441
4442
4443
4444
4445
4446
4447
4448
4449
4450
4451
4452
4453
4454
4455
4456
4457
4458
4459
4460
4461
4462
4463
4464
4465
4466
4467
4468
4469
4470
4471
4472
4473
4474
4475
4476
4477
4478
4479
4480
4481
4482
4483
4484
4485
4486
4487
4488
4489
4490
4491
4492
4493
4494
4495
4496
4497
4498
4499
4500
4501
4502
4503
4504
4505
4506
4507
4508
4509
4510
4511
4512
4513
4514
4515
4516
4517
4518
4519
4520
4521
4522
4523
4524
4525
4526
4527
4528
4529
4530
4531
4532
4533
4534
4535
4536
4537
4538
4539
4540
4541
4542
4543
4544
4545
4546
4547
4548
4549
4550
4551
4552
4553
4554
4555
4556
4557
4558
4559
4560
4561
4562
4563
4564
4565
4566
4567
4568
4569
4570
4571
4572
4573
4574
4575
4576
4577
4578
4579
4580
4581
4582
4583
4584
4585
4586
4587
4588
4589
4590
4591
4592
4593
4594
4595
4596
4597
4598
4599
4600
4601
4602
4603
4604
4605
4606
4607
4608
4609
4610
4611
4612
4613
4614
4615
4616
4617
4618
4619
4620
4621
4622
4623
4624
4625
4626
4627
4628
4629
4630
4631
4632
4633
4634
4635
4636
4637
4638
4639
4640
4641
4642
4643
4644
4645
4646
4647
4648
4649
4650
4651
4652
4653
4654
4655
4656
4657
4658
4659
4660
4661
4662
4663
4664
4665
4666
4667
4668
4669
4670
4671
4672
4673
4674
4675
4676
4677
4678
4679
4680
4681
4682
4683
4684
4685
4686
4687
4688
4689
4690
4691
4692
4693
4694
4695
4696
4697
4698
4699
4700
4701
4702
4703
4704
4705
4706
4707
4708
4709
4710
4711
4712
4713
4714
4715
4716
4717
4718
4719
4720
4721
4722
4723
4724
4725
4726
4727
4728
4729
4730
4731
4732
4733
4734
4735
4736
4737
4738
4739
4740
4741
4742
4743
4744
4745
4746
4747
4748
4749
4750
4751
4752
4753
4754
4755
4756
4757
4758
4759
4760
4761
4762
4763
4764
4765
4766
4767
4768
4769
4770
4771
4772
4773
4774
4775
4776
4777
4778
4779
4780
4781
4782
4783
4784
4785
4786
4787
4788
4789
4790
4791
4792
4793
4794
4795
4796
4797
4798
4799
4800
4801
4802
4803
4804
4805
4806
4807
4808
4809
4810
4811
4812
4813
4814
4815
4816
4817
4818
4819
4820
4821
4822
4823
4824
4825
4826
4827
4828
4829
4830
4831
4832
4833
4834
4835
4836
4837
4838
4839
4840
4841
4842
4843
4844
4845
4846
4847
4848
4849
4850
4851
4852
4853
4854
4855
4856
4857
4858
4859
4860
4861
4862
4863
4864
4865
4866
4867
4868
4869
4870
4871
4872
4873
4874
4875
4876
4877
4878
4879
4880
4881
4882
4883
4884
4885
4886
4887
4888
4889
4890
4891
4892
4893
4894
4895
4896
4897
4898
4899
4900
4901
4902
4903
4904
4905
4906
4907
4908
4909
4910
4911
4912
4913
4914
4915
4916
4917
4918
4919
4920
4921
4922
4923
4924
4925
4926
4927
4928
4929
4930
4931
4932
4933
4934
4935
4936
4937
4938
4939
4940
4941
4942
4943
4944
4945
4946
4947
4948
4949
4950
4951
4952
4953
4954
4955
4956
4957
4958
4959
4960
4961
4962
4963
4964
4965
4966
4967
4968
4969
4970
4971
4972
4973
4974
4975
4976
4977
4978
4979
4980
4981
4982
4983
4984
4985
4986
4987
4988
4989
4990
4991
4992
4993
4994
4995
4996
4997
4998
4999
5000
5001
5002
5003
5004
5005
5006
5007
5008
5009
5010
5011
5012
5013
5014
5015
5016
5017
5018
5019
5020
5021
5022
5023
5024
5025
5026
5027
5028
5029
5030
5031
5032
5033
5034
5035
5036
5037
5038
5039
5040
5041
5042
5043
5044
5045
5046
5047
5048
5049
5050
5051
5052
5053
5054
5055
5056
5057
5058
5059
5060
5061
5062
5063
5064
5065
5066
5067
5068
5069
5070
5071
5072
5073
5074
5075
5076
5077
5078
5079
5080
5081
5082
5083
5084
5085
5086
5087
5088
5089
5090
5091
5092
5093
5094
5095
5096
5097
5098
5099
5100
5101
5102
5103
5104
5105
5106
5107
5108
5109
5110
5111
5112
5113
5114
5115
5116
5117
5118
5119
5120
5121
5122
5123
5124
5125
5126
5127
5128
5129
5130
5131
5132
5133
5134
5135
5136
5137
5138
5139
5140
5141
5142
5143
5144
5145
5146
5147
5148
5149
5150
5151
5152
5153
5154
5155
5156
5157
5158
5159
5160
5161
5162
5163
5164
5165
5166
5167
5168
5169
5170
5171
5172
5173
5174
5175
5176
5177
5178
5179
5180
5181
5182
5183
5184
5185
5186
5187
5188
5189
5190
5191
5192
5193
5194
5195
5196
5197
5198
5199
5200
5201
5202
5203
5204
5205
5206
5207
5208
5209
5210
5211
5212
5213
5214
5215
5216
5217
5218
5219
5220
5221
5222
5223
5224
5225
5226
5227
5228
5229
5230
5231
5232
5233
5234
5235
5236
5237
5238
5239
5240
5241
5242
5243
5244
5245
5246
5247
5248
5249
5250
5251
5252
5253
5254
5255
5256
5257
5258
5259
5260
5261
5262
5263
5264
5265
5266
5267
5268
5269
5270
5271
5272
5273
5274
5275
5276
5277
5278
5279
5280
5281
5282
5283
5284
5285
5286
5287
5288
5289
5290
5291
5292
5293
5294
5295
5296
5297
5298
5299
5300
5301
5302
5303
5304
5305
5306
5307
5308
5309
5310
5311
5312
5313
5314
5315
5316
5317
5318
5319
5320
5321
5322
5323
5324
5325
5326
5327
5328
5329
5330
5331
5332
5333
5334
5335
5336
5337
5338
5339
5340
5341
5342
5343
5344
5345
5346
5347
5348
5349
5350
5351
5352
5353
5354
5355
5356
5357
5358
5359
5360
5361
5362
5363
5364
5365
5366
5367
5368
5369
5370
5371
5372
5373
5374
5375
5376
5377
5378
5379
5380
5381
5382
5383
5384
5385
5386
5387
5388
5389
5390
5391
5392
5393
5394
5395
5396
5397
5398
5399
5400
5401
5402
5403
5404
5405
5406
5407
5408
5409
5410
5411
5412
5413
5414
5415
5416
5417
5418
5419
5420
5421
5422
5423
5424
5425
5426
5427
5428
5429
5430
5431
5432
5433
5434
5435
5436
5437
5438
5439
5440
5441
5442
5443
5444
5445
5446
5447
5448
5449
5450
5451
5452
5453
5454
5455
5456
5457
5458
5459
5460
5461
5462
/*
 * Copyright (C) 2013 Red Hat, Inc.
 *
 * This file is part of LVM2.
 *
 * This copyrighted material is made available to anyone wishing to use,
 * modify, copy, or redistribute it subject to the terms and conditions
 * of the GNU Lesser General Public License v.2.1.
 */

#define _XOPEN_SOURCE 500  /* pthread */
#define _ISOC99_SOURCE
#define _GNU_SOURCE

#include "configure.h"
#include "daemon-io.h"
#include "daemon-server.h"
#include "daemon-log.h"
#include "config-util.h"
#include "lvm-version.h"
#include "lvmetad-client.h"
#include "lvmlockd-client.h"

#include <assert.h>
#include <pthread.h>
#include <stdint.h>
#include <stddef.h>
#include <stdlib.h>
#include <unistd.h>
#include <poll.h>
#include <errno.h>
#include <signal.h>
#include <getopt.h>
#include <syslog.h>
#include <sys/types.h>
#include <sys/socket.h>
#include <sys/utsname.h>
#include <sys/un.h>

#define EXTERN
#include "lvmlockd-internal.h"

/*
 * Basic operation of lvmlockd
 *
 * lvmlockd main process runs main_loop() which uses poll().
 * poll listens for new connections from lvm commands and for
 * messages from existing connected lvm commands.
 *
 * lvm command starts and connects to lvmlockd.
 *
 * lvmlockd receives a connection request from command and adds a
 * 'struct client' to keep track of the connection to the command.
 * The client's fd is added to the set of fd's in poll().
 *
 * lvm command sends a lock request to lvmlockd.  The lock request
 * can be for the global lock, a vg lock, or an lv lock.
 *
 * lvmlockd main_loop/poll sees a message from an existing client.
 * It sets client.recv = 1, then wakes up client_thread_main.
 *
 * client_thread_main iterates through client structs (cl), looking
 * for any that need processing, finds the one with cl->recv set,
 * and calls client_recv_action(cl).
 *
 * client_recv_action(cl) reads the message/request from the client,
 * allocates a new 'struct action' (act) to represent the request,
 * sets the act with what is found in the request, then looks at
 * the specific operation in act->op (LD_OP_FOO) to decide what to
 * do with the action:
 *
 * . If the action is to start a lockspace, create a new thread
 *   to manage that lockspace: add_lockspace(act).
 *
 * . If the action is a lock request, pass the act to the thread
 *   that is managing that lockspace: add_lock_action(act).
 *
 * . If the action is to add/remove a local VG to the list of
 *   local VGs, do that directly: add_local_vg()/rem_local_vg().
 *
 * . Other misc actions are are passed to the worker_thread:
 *   add_work_action(act).
 *
 * Onec the client_thread has passed the action off to another
 * thread to process, it goes back to waiting for more client
 * handling work to do.
 *
 * The thread that was given the action by the client_thread
 * now processes that action according to the operation, act->op.
 * This is either a lockspace_thread (for lock ops or ops that
 * add/rem a lockspace), or the worker_thread.  See below for
 * how these ops are processed by these threads.  When the
 * given thread is done processing the action, the result is
 * set in act->result, and the act struct for the completed action
 * is passed back to the client_thread (client_results list).
 *
 * The client_thread takes completed actions (from client_results
 * list), and sends the result back to the client that sent the
 * request represented by the action.  The act struct is then freed.
 *
 * This completes the cycle of work between lvm commands (clients)
 * and lvmlockd.  In summary:
 *
 * - main process polls for new client connections and new requests
 *   from lvm commands
 * - client_thread reads requests from clients
 * - client_thread creates an action struct for each request
 * - client_thread passes the act to another thread for processing
 * - other threads pass completed act structs back to client_thread
 * - client_thread sends the act result back to the client and frees the act
 *
 *
 * Lockspace threads:
 * Each lockd VG has its own lockspace that contains locks for that VG.
 * Each 'struct lockspace' is managed by a separate lockspace_thread.
 * When the lockspace_thread is first created, the first thing it does
 * is join the lockspace in the lock manager.  This can take a long time.
 * If the join fails, the thread exits.  After the join, the thread
 * enters a loop waiting for lock actions to perform in the lockspace.
 *
 * The request to remove/leave a lockspace causes a flag to be set in
 * the lockspace struct.  When the lockspace_thread sees this flag
 * set, it leaves the lockspace, and exits.
 *
 * When the client_thread passes a new action to a lockspace_thread,
 * i.e. a new lock request, the lockspace_thread identifies which resource
 * is being locked (GL, VG, LV), and gets the 'struct resource' (r) for it.
 * r->type will be LD_RT_GL, LD_RT_VG, or LD_RT_LV.  r->name is the
 * resource name, and is fixed for GL and VG resources, but is based on
 * the LV name for LV resources.  The act is added to the resource's
 * list of actions: r->actions, i.e. outstanding lock requests on the
 * resource.
 *
 * The lockspace thread then iterates through each resource in the
 * lockspace, processing any outstanding actions on each: res_process(ls, r).
 *
 * res_process() compares the outstanding actions/requests in r->actions
 * against any existing locks on the resource in r->locks.  If the
 * action is blocked by existing locks, it's left on r->actions.  If not,
 * the action/request is passed to the lock manager.  If the result from
 * the lock manager is success, a new 'struct lock' is created for the
 * action and saved on r->locks.  The result is set in act->result and
 * the act is passed back to the client_thread to be returned to the client.
 */

static const char *lvmlockd_protocol = "lvmlockd";
static const int lvmlockd_protocol_version = 1;
static int daemon_quit;
static char *our_system_id;

static daemon_handle lvmetad_handle;
static pthread_mutex_t lvmetad_mutex;
static int lvmetad_connected;

/*
 * We use a separate socket for dumping daemon info.
 * This will not interfere with normal operations, and allows
 * free-form debug data to be dumped instead of the libdaemon
 * protocol that wants all data in the cft format.
 * 1MB should fit all the info we need to dump.
 */
#define DUMP_SOCKET_NAME "lvmlockd-dump.sock"
#define DUMP_BUF_SIZE (1024 * 1024)
static char dump_buf[DUMP_BUF_SIZE];
static struct sockaddr_un dump_addr;
static socklen_t dump_addrlen;

/*
 * Main program polls client connections, adds new clients,
 * adds work for client thread.
 *
 * pollfd_mutex is used for adding vs removing entries,
 * and for resume vs realloc.
 */
#define POLL_FD_UNUSED -1		/* slot if free */
#define POLL_FD_IGNORE -2		/* slot is used but ignore in poll */
#define ADD_POLL_SIZE 16		/* increment slots by this amount */

static pthread_mutex_t pollfd_mutex;
static struct pollfd *pollfd;
static int pollfd_size;
static int pollfd_maxi;
static int listen_pi;
static int listen_fd;
static int restart_pi;
static int restart_fds[2];

/*
 * Each lockspace has its own thread to do locking.
 * The lockspace thread makes synchronous lock requests to dlm/sanlock.
 * Every vg with a lockd type, i.e. "dlm", "sanlock", should be on this list.
 *
 * lockspaces_inactive holds old ls structs for vgs that have been
 * stopped, or for vgs that failed to start.  The old ls structs
 * are removed from the inactive list and freed when a new ls with
 * the same name is started and added to the standard lockspaces list.
 * Keeping this bit of "history" for the ls allows us to return a
 * more informative error message if a vg lock request is made for
 * an ls that has been stopped or failed to start.
 */
static pthread_mutex_t lockspaces_mutex;
static struct list_head lockspaces;
static struct list_head lockspaces_inactive;

/*
 * This flag is set to 1 if we see multiple vgs with the global
 * lock enabled.  While this is set, we return a special flag
 * with the vg lock result indicating to the lvm command that
 * there is a duplicate gl in the vg which should be resolved.
 * While this is set, find_lockspace_name has the side job of
 * counting the number of lockspaces with enabled gl's so that
 * this can be set back to zero when the duplicates are disabled.
 */
static int sanlock_gl_dup;

/*
 * VG's that do not have a lockd type are on the local_vgs list.
 * Every vg on the system should be in either the lockspaces
 * list or the local_vgs list.
 *
 * lvm commands send lock requests to lvmlockd for local vgs
 * because at the point locks are acquired in the command,
 * the vg has not been read, so the command does not know if
 * the vg's lock_type is local and the locks can be skipped.
 * So lvmlockd keeps track of which vg's are local so it can
 * quickly check if a vg lock request can be skipped.  (Rather
 * than having to look up the lock_type in lvmetad for every
 * operation on a local vg.)
 *
 * When local_thread_also is set, lvmlockd's local_thread is
 * used to manage locks for local pids on vgs from local_vgs.
 * (In addition to standard locking for lockd type vgs.)
 *
 * When local_thread_only is set, lvmlockd is only used to
 * manage locks for local pids on vgs from local_vgs, and
 * not to manage lockd type vgs.
 *
 * local locking:
 *
 * lock_gl: only do local_thread locking for gl when local_thread_only
 * is set.  local_thread_only means that no standard lockd lockspaces
 * are being used, and lvmlockd is used only for inter-pid locking.
 * When local_thread_only is not set (meaning both local and shared vgs
 * are expected), then the standard gl lockspace works for both local
 * (between local pids) and remote (between pids on different nodes).
 *
 * lock_vg: only do local_thread locking for local, non-lockd, vgs in
 * the local_vgs list.  When the vg is a lockd-type, then the standard
 * lockspace thread works for locking between pids also.
 *
 * local_thread_only=1 local_thread_also=1
 * Use lvmlockd for locking only between local pids, both gl and vg locks.
 * No shared disks or lockd type vgs should exist.
 *
 * local_thread_only=0 local_thread_also=1
 * Use lvmlockd for locking between local pids for local vgs,
 * and use lvmlockd for distributed locking for lockd-type vgs.
 * Use global lock from a lockd-type vgs.  A local-only gl does
 * not make sense here.
 *
 * local_thread_only=0 local_thread_also=0
 * Do not use lvmlockd for locking between local pids.
 * No shared disks or lockd type vgs should exist.
 * (lvmlockd should probably not be run at all in this case.)
 *
 * local_thread_only=1 local_thread_also=0
 * Not allowed.
 */
static pthread_t local_thread;
static pthread_mutex_t local_thread_mutex;
static pthread_cond_t local_thread_cond;
static struct list_head local_thread_actions;
static struct list_head local_vgs;
static struct lockspace *local_thread_gls;
static int local_thread_also;
static int local_thread_only;
static int local_thread_stop;
static int local_thread_work;

/*
 * Client thread reads client requests and writes client results.
 */
static pthread_t client_thread;
static pthread_mutex_t client_mutex;
static pthread_cond_t client_cond;
static struct list_head client_list;    /* connected clients */
static struct list_head client_results; /* actions to send back to clients */
static uint32_t client_ids;
static int client_stop;                 /* stop the thread */
static int client_work;                 /* a client on client_list has work to do */

/*
 * Worker thread performs misc non-locking actions, e.g. init/free.
 */
static pthread_t worker_thread;
static pthread_mutex_t worker_mutex;
static pthread_cond_t worker_cond;
static struct list_head worker_list;    /* actions for worker_thread */
static int worker_stop;                 /* stop the thread */
static int worker_wake;                 /* wake the thread without adding work */

static int add_lock_action(struct action *act);
static int str_to_lm(const char *str);
static void clear_lockspace_inactive(char *name);

/*
 * The content of every log_foo() statement is saved in the
 * circular buffer, which can be dumped to a client and printed.
 */
#define LOG_LINE_SIZE 256
#define LOG_DUMP_SIZE DUMP_BUF_SIZE
static char log_dump[LOG_DUMP_SIZE];
static unsigned int log_point;
static unsigned int log_wrap;
static pthread_mutex_t log_mutex;
static int syslog_priority = LOG_WARNING;

#define DO_STOP 1
#define NO_STOP 0
#define DO_FREE 1
#define NO_FREE 0
#define DO_FORCE 1
#define NO_FORCE 0

static uint64_t monotime(void)
{
	struct timespec ts;
	clock_gettime(CLOCK_MONOTONIC, &ts);
	return ts.tv_sec;
}

static void log_save_line(int len, char *line,
			  char *log_buf, unsigned int *point, unsigned int *wrap)
{
	unsigned int p = *point;
	unsigned int w = *wrap;
	int i;

	if (len < LOG_DUMP_SIZE - p) {
		memcpy(log_buf + p, line, len);
		p += len;

		if (p == LOG_DUMP_SIZE) {
			p = 0;
			w = 1;
		}
		goto out;
	}

	for (i = 0; i < len; i++) {
		log_buf[p++] = line[i];

		if (p == LOG_DUMP_SIZE) {
			p = 0;
			w = 1;
		}
	}
 out:
	*point = p;
	*wrap = w;
}

void log_level(int level, const char *fmt, ...)
{
	char line[LOG_LINE_SIZE];
	va_list ap;
	int len = LOG_LINE_SIZE - 1;
	int ret, pos = 0;

	memset(line, 0, sizeof(line));

	ret = snprintf(line, len, "%llu ", (unsigned long long)time(NULL));
	pos += ret;

	va_start(ap, fmt);
	ret = vsnprintf(line + pos, len - pos, fmt, ap);
	va_end(ap);

	if (ret >= len - pos)
		pos = len - 1;
	else
		pos += ret;

	line[pos++] = '\n';
	line[pos++] = '\0';

	pthread_mutex_lock(&log_mutex);
	log_save_line(pos - 1, line, log_dump, &log_point, &log_wrap);
	pthread_mutex_unlock(&log_mutex);

	if (level <= syslog_priority)
		syslog(level, "%s", line);

	if (daemon_debug)
		fprintf(stderr, "%s", line);
}

static int dump_log(int *dump_len)
{
	int tail_len;

	pthread_mutex_lock(&log_mutex);

	if (!log_wrap && !log_point) {
		*dump_len = 0;
	} else if (log_wrap) {
		tail_len = LOG_DUMP_SIZE - log_point;
		memcpy(dump_buf, log_dump+log_point, tail_len);
		if (log_point)
			memcpy(dump_buf+tail_len, log_dump, log_point);
		*dump_len = LOG_DUMP_SIZE;
	} else {
		memcpy(dump_buf, log_dump, log_point-1);
		*dump_len = log_point-1;
	}
	pthread_mutex_unlock(&log_mutex);

	return 0;
}

/*
 * List from kernel
 */
static inline void INIT_LIST_HEAD(struct list_head *list)
{
	list->next = list;
	list->prev = list;
}

static inline void __list_add(struct list_head *new,
                              struct list_head *prev,
                              struct list_head *next)
{
	next->prev = new;
	new->next = next;
	new->prev = prev;
	prev->next = new;
}

static inline void __list_del(struct list_head *prev, struct list_head *next)
{
	next->prev = prev;
	prev->next = next;
}

static inline void list_add(struct list_head *new, struct list_head *head)
{
	__list_add(new, head, head->next);
}

static inline void list_add_tail(struct list_head *new, struct list_head *head)
{
	__list_add(new, head->prev, head);
}

static inline void list_del(struct list_head *entry)
{
	__list_del(entry->prev, entry->next);
}

static inline int list_empty(const struct list_head *head)
{
        return head->next == head;
}

#define list_entry(ptr, type, member) \
        container_of(ptr, type, member)

#define list_first_entry(ptr, type, member) \
        list_entry((ptr)->next, type, member)

#define list_for_each_entry(pos, head, member)                          \
        for (pos = list_entry((head)->next, typeof(*pos), member);      \
             &pos->member != (head);    \
             pos = list_entry(pos->member.next, typeof(*pos), member))

#define list_for_each_entry_safe(pos, n, head, member)                  \
        for (pos = list_entry((head)->next, typeof(*pos), member),      \
                n = list_entry(pos->member.next, typeof(*pos), member); \
             &pos->member != (head);                                    \
             pos = n, n = list_entry(n->member.next, typeof(*n), member))

#define MAX_UNUSED_ACTION 32
#define MAX_UNUSED_CLIENT 32
#define MAX_UNUSED_RESOURCE 32
#define MAX_UNUSED_LOCK 32
static pthread_mutex_t unused_struct_mutex;
static struct list_head unused_action;
static struct list_head unused_client;
static struct list_head unused_resource;
static struct list_head unused_lock;
static int unused_action_count;
static int unused_client_count;
static int unused_resource_count;
static int unused_lock_count;

static struct action *alloc_action(void)
{
	struct action *act;

	pthread_mutex_lock(&unused_struct_mutex);
	if (!unused_action_count) {
		act = malloc(sizeof(struct action));
	} else {
		act = list_first_entry(&unused_action, struct action, list);
		list_del(&act->list);
		unused_action_count--;
	}
	pthread_mutex_unlock(&unused_struct_mutex);
	if (act)
		memset(act, 0, sizeof(struct action));
	return act;
}

static struct client *alloc_client(void)
{
	struct client *cl;

	pthread_mutex_lock(&unused_struct_mutex);
	if (!unused_client_count) {
		cl = malloc(sizeof(struct client));
	} else {
		cl = list_first_entry(&unused_client, struct client, list);
		list_del(&cl->list);
		unused_client_count--;
	}
	pthread_mutex_unlock(&unused_struct_mutex);
	if (cl)
		memset(cl, 0, sizeof(struct client));
	return cl;
}

static struct resource *alloc_resource(void)
{
	struct resource *r;

	pthread_mutex_lock(&unused_struct_mutex);
	if (!unused_resource_count) {
		r = malloc(sizeof(struct resource));
	} else {
		r = list_first_entry(&unused_resource, struct resource, list);
		list_del(&r->list);
		unused_resource_count--;
	}
	pthread_mutex_unlock(&unused_struct_mutex);
	if (r)
		memset(r, 0, sizeof(struct resource));
	return r;
}

static struct lock *alloc_lock(void)
{
	struct lock *lk;

	pthread_mutex_lock(&unused_struct_mutex);
	if (!unused_lock_count) {
		lk = malloc(sizeof(struct lock));
	} else {
		lk = list_first_entry(&unused_lock, struct lock, list);
		list_del(&lk->list);
		unused_lock_count--;
	}
	pthread_mutex_unlock(&unused_struct_mutex);
	if (lk)
		memset(lk, 0, sizeof(struct lock));
	return lk;
}

static void free_action(struct action *act)
{
	pthread_mutex_lock(&unused_struct_mutex);
	if (unused_action_count >= MAX_UNUSED_ACTION) {
		free(act);
	} else {
		list_add_tail(&act->list, &unused_action);
		unused_action_count++;
	}
	pthread_mutex_unlock(&unused_struct_mutex);
}

static void free_client(struct client *cl)
{
	pthread_mutex_lock(&unused_struct_mutex);
	if (unused_client_count >= MAX_UNUSED_CLIENT) {
		free(cl);
	} else {
		list_add_tail(&cl->list, &unused_client);
		unused_client_count++;
	}
	pthread_mutex_unlock(&unused_struct_mutex);
}

static void free_resource(struct resource *r)
{
	pthread_mutex_lock(&unused_struct_mutex);
	if (unused_resource_count >= MAX_UNUSED_RESOURCE) {
		free(r);
	} else {
		list_add_tail(&r->list, &unused_resource);
		unused_resource_count++;
	}
	pthread_mutex_unlock(&unused_struct_mutex);
}

static void free_lock(struct lock *lk)
{
	pthread_mutex_lock(&unused_struct_mutex);
	if (unused_lock_count >= MAX_UNUSED_LOCK) {
		free(lk);
	} else {
		list_add_tail(&lk->list, &unused_lock);
		unused_lock_count++;
	}
	pthread_mutex_unlock(&unused_struct_mutex);
}

static void setup_structs(void)
{
	struct action *act;
	struct client *cl;
	struct resource *r;
	struct lock *lk;
	int i;

	pthread_mutex_init(&unused_struct_mutex, NULL);
	INIT_LIST_HEAD(&unused_action);
	INIT_LIST_HEAD(&unused_client);
	INIT_LIST_HEAD(&unused_resource);
	INIT_LIST_HEAD(&unused_lock);

	for (i = 0; i < MAX_UNUSED_ACTION; i++) {
		act = alloc_action();
		free_action(act);
	}

	for (i = 0; i < MAX_UNUSED_CLIENT; i++) {
		cl = alloc_client();
		free_client(cl);
	}

	for (i = 0; i < MAX_UNUSED_RESOURCE; i++) {
		r = alloc_resource();
		free_resource(r);
	}

	for (i = 0; i < MAX_UNUSED_LOCK; i++) {
		lk = alloc_lock();
		free_lock(lk);
	}
}

static int add_pollfd(int fd)
{
	int i, new_size;

	pthread_mutex_lock(&pollfd_mutex);
	for (i = 0; i < pollfd_size; i++) {
		if (pollfd[i].fd != POLL_FD_UNUSED)
			continue;

		pollfd[i].fd = fd;
		pollfd[i].events = POLLIN;
		pollfd[i].revents = 0;

		if (i > pollfd_maxi)
			pollfd_maxi = i;

		pthread_mutex_unlock(&pollfd_mutex);
		return i;
	}

	new_size = pollfd_size + ADD_POLL_SIZE;

	pollfd = realloc(pollfd, new_size * sizeof(struct pollfd));
	if (!pollfd) {
		log_error("can't alloc new size %d for pollfd", new_size);
		return -ENOMEM;
	}

	for (i = pollfd_size; i < new_size; i++) {
		pollfd[i].fd = POLL_FD_UNUSED;
		pollfd[i].events = 0;
		pollfd[i].revents = 0;
	}

	i = pollfd_size;
	pollfd[i].fd = fd;
	pollfd[i].events = POLLIN;
	pollfd[i].revents = 0;
	pollfd_maxi = i;

	pollfd_size = new_size;

	pthread_mutex_unlock(&pollfd_mutex);
	return i;
}

static void rem_pollfd(int pi)
{
	if (pi < 0) {
		log_error("rem_pollfd %d", pi);
		return;
	}
	pthread_mutex_lock(&pollfd_mutex);
	pollfd[pi].fd = POLL_FD_UNUSED;
	pollfd[pi].events = 0;
	pollfd[pi].revents = 0;
	pthread_mutex_unlock(&pollfd_mutex);
}

static const char *lm_str(int x)
{
	switch (x) {
	case LD_LM_NONE:
		return "none";
	case LD_LM_DLM:
		return "dlm";
	case LD_LM_SANLOCK:
		return "sanlock";
	default:
		return "lm_unknown";
	}
}

static const char *rt_str(int x)
{
	switch (x) {
	case LD_RT_GL:
		return "gl";
	case LD_RT_VG:
		return "vg";
	case LD_RT_LV:
		return "lv";
	default:
		return ".";
	};
}

static const char *op_str(int x)
{
	switch (x) {
	case LD_OP_INIT:
		return "init";
	case LD_OP_FREE:
		return "free";
	case LD_OP_START:
		return "start";
	case LD_OP_STOP:
		return "stop";
	case LD_OP_LOCK:
		return "lock";
	case LD_OP_UPDATE:
		return "update";
	case LD_OP_CLOSE:
		return "close";
	case LD_OP_ENABLE:
		return "enable";
	case LD_OP_DISABLE:
		return "disable";
	case LD_OP_ADD_LOCAL:
		return "add_local";
	case LD_OP_REM_LOCAL:
		return "rem_local";
	case LD_OP_UPDATE_LOCAL:
		return "update_local";
	case LD_OP_START_WAIT:
		return "start_wait";
	case LD_OP_STOP_ALL:
		return "stop_all";
	case LD_OP_RENAME_BEFORE:
		return "rename_before";
	case LD_OP_RENAME_FINAL:
		return "rename_final";
	default:
		return "op_unknown";
	};
}

static const char *mode_str(int x)
{
	switch (x) {
	case LD_LK_IV:
		return "iv";
	case LD_LK_UN:
		return "un";
	case LD_LK_NL:
		return "nl";
	case LD_LK_SH:
		return "sh";
	case LD_LK_EX:
		return "ex";
	default:
		return ".";
	};
}

int last_string_from_args(char *args_in, char *last)
{
	const char *args = args_in;
	const char *colon, *str = NULL;

	while (1) {
		if (!args || (*args == '\0'))
			break;
		colon = strstr(args, ":");
		if (!colon)
			break;
		str = colon;
		args = colon + 1;
        }

	if (str) {
		snprintf(last, MAX_ARGS, "%s", str + 1);
		return 0;
	}
	return -1;
}

int version_from_args(char *args, unsigned int *major, unsigned int *minor, unsigned int *patch)
{
	char version[MAX_ARGS];
	char *major_str, *minor_str, *patch_str;
	char *n, *d1, *d2;

	strncpy(version, args, MAX_ARGS);

	n = strstr(version, ":");
	if (n)
		*n = '\0';

	d1 = strstr(version, ".");
	if (!d1)
		return -1;

	d2 = strstr(d1 + 1, ".");
	if (!d2)
		return -1;

	major_str = version;
	minor_str = d1 + 1;
	patch_str = d2 + 1;

	*d1 = '\0';
	*d2 = '\0';

	if (major)
		*major = atoi(major_str);
	if (minor)
		*minor = atoi(minor_str);
	if (patch)
		*patch = atoi(patch_str);

	return 0;
}

/*
 * These are few enough that arrays of function pointers can
 * be avoided.
 */

static int lm_add_lockspace(struct lockspace *ls, struct action *act)
{
	int rv;

	if (ls->lm_type == LD_LM_DLM)
		rv = lm_add_lockspace_dlm(ls);
	else if (ls->lm_type == LD_LM_SANLOCK)
		rv = lm_add_lockspace_sanlock(ls);
	else
		return -1;

	if (act)
		act->lm_rv = rv;
	return rv;
}

static int lm_rem_lockspace(struct lockspace *ls, struct action *act, int free_vg)
{
	int rv;

	if (ls->lm_type == LD_LM_DLM)
		rv = lm_rem_lockspace_dlm(ls, free_vg);
	else if (ls->lm_type == LD_LM_SANLOCK)
		rv = lm_rem_lockspace_sanlock(ls, free_vg);
	else
		return -1;

	if (act)
		act->lm_rv = rv;
	return rv;
}

static int lm_lock(struct lockspace *ls, struct resource *r, int mode, struct action *act,
		   uint32_t *r_version, uint32_t *n_version, int *retry)
{
	int rv;

	if (ls->lm_type == LD_LM_DLM)
		rv = lm_lock_dlm(ls, r, mode, r_version, n_version);
	else if (ls->lm_type == LD_LM_SANLOCK)
		rv = lm_lock_sanlock(ls, r, mode, act->lv_args, r_version, n_version, retry);
	else
		return -1;

	if (act)
		act->lm_rv = rv;
	return rv;
}

static int lm_convert(struct lockspace *ls, struct resource *r,
		      int mode, struct action *act, uint32_t r_version)
{
	int rv;

	if (ls->lm_type == LD_LM_DLM)
		rv = lm_convert_dlm(ls, r, mode, r_version);
	else if (ls->lm_type == LD_LM_SANLOCK)
		rv = lm_convert_sanlock(ls, r, mode, r_version);
	else
		return -1;

	if (act)
		act->lm_rv = rv;
	return rv;
}

static int lm_unlock(struct lockspace *ls, struct resource *r, struct action *act,
		     uint32_t r_version, uint32_t n_version, uint32_t lmu_flags)
{
	int rv;

	if (ls->lm_type == LD_LM_DLM)
		return lm_unlock_dlm(ls, r, r_version, n_version, lmu_flags);
	else if (ls->lm_type == LD_LM_SANLOCK)
		return lm_unlock_sanlock(ls, r, r_version, n_version, lmu_flags);
	else
		return -1;

	if (act)
		act->lm_rv = rv;
	return rv;
}

static int lm_hosts(struct lockspace *ls, int notify)
{
	if (ls->lm_type == LD_LM_DLM)
		return 0;
	else if (ls->lm_type == LD_LM_SANLOCK)
		return lm_hosts_sanlock(ls, notify);
	return -1;
}

static void lm_rem_resource(struct lockspace *ls, struct resource *r)
{
	if (ls->lm_type == LD_LM_DLM)
		lm_rem_resource_dlm(ls, r);
	else if (ls->lm_type == LD_LM_SANLOCK)
		lm_rem_resource_sanlock(ls, r);
}

static void add_client_result(struct action *act)
{
	pthread_mutex_lock(&client_mutex);
	list_add_tail(&act->list, &client_results);
	pthread_cond_signal(&client_cond);
	pthread_mutex_unlock(&client_mutex);
}

static struct lock *find_lock_client(struct resource *r, uint32_t client_id)
{
	struct lock *lk;

	list_for_each_entry(lk, &r->locks, list) {
		if (lk->client_id == client_id)
			return lk;
	}
	return NULL;
}

static struct lock *find_lock_persistent(struct resource *r)
{
	struct lock *lk;

	list_for_each_entry(lk, &r->locks, list) {
		if (lk->flags & LD_LF_PERSISTENT)
			return lk;
	}
	return NULL;
}

static struct action *find_action_client(struct resource *r, uint32_t client_id)
{
	struct action *act;

	list_for_each_entry(act, &r->actions, list) {
		if (act->client_id != client_id)
			continue;
		return act;
	}
	return NULL;
}

static void add_work_action(struct action *act)
{
	pthread_mutex_lock(&worker_mutex);
	if (!worker_stop) {
		list_add_tail(&act->list, &worker_list);
		pthread_cond_signal(&worker_cond);
	}
	pthread_mutex_unlock(&worker_mutex);
}

static void create_work_action(int op)
{
	struct action *act;

	act = alloc_action();
	if (!act)
		return;
	act->op = op;
	add_work_action(act);
}

static int res_lock(struct lockspace *ls, struct resource *r, struct action *act, int *retry)
{
	struct lock *lk;
	uint32_t r_version = 0;
	uint32_t n_version = 0;
	int rv;

	log_debug("S %s R %s res_lock mode %d", ls->name, r->name, act->mode);

	if (r->mode == LD_LK_SH && act->mode == LD_LK_SH)
		goto add_lk;

	rv = lm_lock(ls, r, act->mode, act, &r_version, &n_version, retry);
	if (rv == -EAGAIN)
		return rv;
	if (rv < 0) {
		log_error("S %s R %s res_lock lm error %d", ls->name, r->name, rv);
		return rv;
	}

	log_debug("S %s R %s res_lock lm done r_version %u n_version %u",
		  ls->name, r->name, r_version, n_version);

	/* lm_lock() reads new r_version and n_version */

	if (r_version > r->version) {
		/*
		 * New r_version of the lock: means that another
		 * host has changed data protected by this lock
		 * since the last time we acquired it.  We
		 * should invalidate any local cache of the data
		 * protected by this lock and reread it from disk.
		 */
		r->version = r_version;

		/*
		 * r is vglk: tell lvmetad to set the vg invalid
		 * flag, and provide the new r_version.  If lvmetad finds
		 * that its cached vg has seqno less than the value
		 * we send here, it will set the vg invalid flag.
		 * lvm commands that read the vg from lvmetad, will
		 * see the invalid flag returned, will reread the
		 * vg from disk, update the lvmetad copy, and go on.
		 *
		 * r is global: tell lvmetad to set the global invalid
		 * flag.  When commands see this flag returned from lvmetad,
		 * they will reread metadata from disk, update the lvmetad
		 * caches, and tell lvmetad to set global invalid to 0.
		 */

		if ((r->type == LD_RT_VG) && lvmetad_connected) {
			daemon_reply reply;
			char *uuid;

			log_debug("S %s R %s res_lock set lvmetad vg version %u",
				  ls->name, r->name, r_version);
	
			if (!ls->vg_uuid[0] || !strcmp(ls->vg_uuid, "none"))
				uuid = ls->name;
			else
				uuid = ls->vg_uuid;

			pthread_mutex_lock(&lvmetad_mutex);
			reply = daemon_send_simple(lvmetad_handle, "set_vg_info",
						   "token = %s", "skip",
						   "uuid = %s", uuid,
						   "version = %d", (int)r_version,
						   NULL);
			pthread_mutex_unlock(&lvmetad_mutex);
			/* TODO: check reply? */
			daemon_reply_destroy(reply);
		}

		if ((r->type == LD_RT_GL) && lvmetad_connected) {
			daemon_reply reply;

			log_debug("S %s R %s res_lock set lvmetad global invalid",
				  ls->name, r->name);

			pthread_mutex_lock(&lvmetad_mutex);
			reply = daemon_send_simple(lvmetad_handle, "set_global_info",
						   "token = %s", "skip",
						   "global_invalid = %d", 1,
						   NULL);
			pthread_mutex_unlock(&lvmetad_mutex);
			/* TODO: check reply? */
			daemon_reply_destroy(reply);
		}
	}

	if ((r->type == LD_RT_GL) && (n_version > ls->names_version)) {
		/*
		 * Set a flag that will cause update_local_vgs to be run
		 * when the gl is unlocked (by queueing an UPDATE_LOCK action).
		 * It needs to happen on unlock because lvmetad needs to be updated
		 * by the command before there is an updated vg list to be read.
		 */
		log_debug("S %s gl res_lock set update_local_vgs", ls->name);
		ls->update_local_vgs = 1;
		ls->names_version = n_version;
	}

	if ((r->type == LD_RT_GL) && (act->flags & LD_AF_UPDATE_NAMES_VERSION)) {
		/*
		 * Set a flag that will cause the ls->names_version to be
		 * incremented and written to the gl lvb n_version when
		 * the gl is unlocked.
		 * Other hosts will eventually take the gl lock, see the new
		 * n_version and run update_local_vgs.
		 */
		log_debug("S %s gl res_lock set update_names_version", ls->name);
		ls->update_names_version = 1;
	}

	r->mode = act->mode;

add_lk:
	if (r->mode == LD_LK_SH)
		r->sh_count++;

	lk = alloc_lock();
	if (!lk) {
		/* TODO */
		log_error("res_lock ENOMEM");
		return -ENOMEM;
	}

	lk->client_id = act->client_id;
	lk->mode = act->mode;

	if (act->flags & LD_AF_PERSISTENT) {
		lk->flags |= LD_LF_PERSISTENT;
		lk->client_id = 0;
	}

	list_add_tail(&lk->list, &r->locks);

	return 0;
}

static int res_convert(struct lockspace *ls, struct resource *r,
		       struct lock *lk, struct action *act)
{
	uint32_t r_version;
	int rv;

	log_debug("S %s R %s res_convert mode %d", ls->name, r->name, act->mode);

	if (act->mode == LD_LK_EX && lk->mode == LD_LK_SH && r->sh_count > 1)
		return -EAGAIN;

	/*
	 * lm_convert() writes new version (from ex)
	 * Same as lm_unlock()
	 */

        if ((r->type == LD_RT_GL) && (r->mode == LD_LK_EX)) {
		r->version++;
		lk->version = r->version;
		r_version = r->version;
		log_debug("S %s R %s res_convert r_version inc %u",
			  ls->name, r->name, r_version);

	} else if ((r->type == LD_RT_VG) && (r->mode == LD_LK_EX) && (lk->version > r->version)) {
		r->version = lk->version;
		r_version = r->version;
		log_debug("S %s R %s res_convert r_version new %u", ls->name, r->name, r_version);
	} else {
		r_version = 0;
	}

	rv = lm_convert(ls, r, act->mode, act, r_version);
	if (rv < 0) {
		log_error("S %s R %s res_convert lm error %d", ls->name, r->name, rv);
		return rv;
	}

	log_debug("S %s R %s res_convert lm done", ls->name, r->name);

	if (lk->mode == LD_LK_EX && act->mode == LD_LK_SH) {
		r->sh_count = 1;
	} else if (lk->mode == LD_LK_SH && act->mode == LD_LK_EX) {
		r->sh_count = 0;
	} else {
		/* should not be possible */
		log_error("S %s R %s res_convert invalid modes %d %d",
			  ls->name, r->name, lk->mode, act->mode);
		return -1;
	}

	r->mode = act->mode;
	lk->mode = act->mode;

	return 0;
}

static int res_cancel(struct lockspace *ls, struct resource *r,
		      struct action *act)
{
	struct action *cact;

	/*
	 * a client can cancel its own non-persistent lock requests,
	 * when could this happen?
	 *
	 * a client can cancel other client's persistent lock requests,
	 * when could this happen?
	 */

	if (act->flags & LD_AF_PERSISTENT) {
		list_for_each_entry(cact, &r->actions, list) {
			if (!(cact->flags & LD_AF_PERSISTENT))
				continue;
			goto do_cancel;
		}
	} else {
		cact = find_action_client(r, act->client_id);
		if (cact)
			goto do_cancel;
	}

	return -ENOENT;

do_cancel:
	log_debug("S %s R %s res_cancel client %d", ls->name, r->name, cact->client_id);
	cact->result = -ECANCELED;
	list_del(&cact->list);
	add_client_result(cact);

	return -ECANCELED;
}

/*
 * lm_unlock() writes new a r_version (from ex)
 *
 * The r_version of the vg resource is incremented if
 * an "update" was received for the vg lock.  The update
 * contains the new vg seqno from the vg metadata which is
 * used as the r_version.
 *
 * The r_version of the global resource is automatically
 * incremented when it is unlocked from ex mode.
 *
 * For the global resource, n_version is used in addition
 * to r_version:
 *
 * r_version is incremented every time a command releases
 * the global lock from ex.
 *
 * n_version is incremented every time a command that
 * changes the list of vg names releases the global lock from ex.
 *
 * Changes to n_version are used by hosts to detect that other
 * hosts have added/removed/renamed local (non-lockd) vgs which
 * can be seen by multiple hosts, so the local_vgs list probably
 * needs to be updated.  lvmlockd knows about changes to lockd-type
 * vgs through their locks, but local vgs do not have locks,
 * so the n_version change is the only way to know that the
 * local_vgs list should be updated.
 */

/*
 * persistent locks will not be unlocked for OP_CLOSE/act_close
 * because act_close->flags does not have the PERSISTENT flag
 * set, and a persistent lk->client_id is zero, which will not
 * match the client in act_close->client_id.
 */

static int res_unlock(struct lockspace *ls, struct resource *r,
		      struct action *act)
{
	struct lock *lk;
	uint32_t r_version;
	uint32_t n_version = 0;
	int rv;

	if (act->flags & LD_AF_PERSISTENT) {
		lk = find_lock_persistent(r);
		if (lk)
			goto do_unlock;
	} else {
		lk = find_lock_client(r, act->client_id);
		if (lk)
			goto do_unlock;
	}

	if (act->op != LD_OP_CLOSE)
		log_error("S %s R %s res_unlock no locks", ls->name, r->name);
	return -ENOENT;

do_unlock:
	log_debug("S %s R %s res_unlock %s", ls->name, r->name,
		  (act->op == LD_OP_CLOSE) ? "from close" : "");

	/* send unlock to lm when last sh lock is unlocked */
	if (lk->mode == LD_LK_SH) {
		r->sh_count--;
		if (r->sh_count > 0)
			goto rem_lk;
	}

	if ((r->type == LD_RT_GL) && (r->mode == LD_LK_EX)) {
		r->version++;
		lk->version = r->version;
		r_version = r->version;

		log_debug("S %s R %s res_unlock r_version inc %u", ls->name, r->name, r_version);

		if (ls->update_names_version) {
			ls->names_version++;
			n_version = ls->names_version;
			log_debug("S %s gl res_unlock got update_names_version %u",
				  ls->name, n_version);
		}

	} else if ((r->type == LD_RT_VG) && (r->mode == LD_LK_EX) && (lk->version > r->version)) {
		r->version = lk->version;
		r_version = r->version;

		log_debug("S %s R %s res_unlock r_version new %u",
			  ls->name, r->name, r_version);
	} else {
		r_version = 0;
	}

	rv = lm_unlock(ls, r, act, r_version, n_version, 0);
	if (rv < 0) {
		/* should never happen, retry? */
		log_error("S %s R %s res_unlock lm error %d", ls->name, r->name, rv);
		return rv;
	}

	log_debug("S %s R %s res_unlock lm done", ls->name, r->name);

	if ((r->type == LD_RT_GL) && (ls->update_local_vgs || ls->update_names_version)) {
		log_debug("S %s gl res_unlock got update_local_vgs %d update_names_version %d",
			  ls->name, ls->update_local_vgs, ls->update_names_version);
		ls->update_local_vgs = 0;
		ls->update_names_version = 0;
		create_work_action(LD_OP_UPDATE_LOCAL);
	}

rem_lk:
	list_del(&lk->list);
	free_lock(lk);

	/*
	 * TODO: if unlock isn't synchronous, and next lock runs into
	 * it, what will the effect be?
	 */

	if (list_empty(&r->locks))
		r->mode = LD_LK_UN;

	return 0;
}

static int res_update(struct lockspace *ls, struct resource *r,
		      struct action *act)
{
	struct lock *lk;

	lk = find_lock_client(r, act->client_id);
	if (!lk) {
		log_error("S %s R %s res_update client %u lock not found",
			  ls->name, r->name, act->client_id);
		return -ENOENT;
	}

	if (r->mode != LD_LK_EX) {
		log_error("S %s R %s res_update version on non-ex lock",
			  ls->name, r->name);
		return -EINVAL;
	}

	/* lk version will be written to lm by unlock */

	/* TODO: try to write it to lm here in some cases?
	 * when a SYNC flag is set for update? */

	if (act->flags & LD_AF_NEXT_VERSION)
		lk->version = r->version + 1;
	else
		lk->version = act->version;

	log_debug("S %s R %s res_update lk version to %u", ls->name, r->name, lk->version);

	return 0;
}

static int free_lv(struct lockspace *ls, struct resource *r)
{
	if (ls->lm_type == LD_LM_SANLOCK)
		return lm_free_lv_sanlock(ls, r);
	else if (ls->lm_type == LD_LM_DLM)
		return 0;
	else
		return -EINVAL;
}

/*
 * NB. we can't do this if sanlock is holding any locks on
 * the resource; we'd be rewriting the resource from under
 * sanlock and would confuse or break it badly.  We don't
 * know what another host is doing, so these must be used
 * very carefully.
 */

static int res_able(struct lockspace *ls, struct resource *r,
		    struct action *act)
{
	int rv;

	if (ls->lm_type != LD_LM_SANLOCK) {
		log_error("enable/disable only applies to sanlock");
		return -EINVAL;
	}

	if (r->type != LD_RT_GL) {
		log_error("enable/disable only applies to global lock");
		return -EINVAL;
	}

	if (r->mode != LD_LK_UN) {
		log_error("enable/disable only allowed on unlocked resource");
		return -EINVAL;
	}

	if (act->op == LD_OP_ENABLE && gl_lsname_sanlock[0]) {
		log_error("disable global lock in %s before enable in %s",
			  gl_lsname_sanlock, ls->name);
		return -EINVAL;
	}

	if ((act->op == LD_OP_DISABLE) && (act->flags & LD_AF_EX_DISABLE)) {
		rv = lm_ex_disable_gl_sanlock(ls);
		goto out;
	}

	rv = lm_able_gl_sanlock(ls, act->op == LD_OP_ENABLE);
out:
	return rv;
}

/*
 * Go through queued actions, and make lock/unlock calls on the resource
 * based on the actions and the existing lock state.
 *
 * All lock operations sent to the lock manager are non-blocking.
 * This is because sanlock does not support lock queueing.
 * Eventually we could enhance this to take advantage of lock
 * queueing when available (i.e. for the dlm).
 *
 * act_close_list: list of CLOSE actions, identifying clients that have
 * closed/terminated their lvmlockd connection, and whose locks should
 * be released.  Do not remove these actions from act_close_list.
 *
 * retry_out: set to 1 if the lock manager said we should retry,
 * meaning we should call res_process() again in a short while to retry.
 */

static void res_process(struct lockspace *ls, struct resource *r,
			struct list_head *act_close_list, int *retry_out)
{
	struct action *act, *safe, *act_close;
	struct lock *lk;
	int lm_retry;
	int rv;

	/*
	 * handle version updates for ex locks
	 * (new version will be written by unlock)
	 */

	list_for_each_entry_safe(act, safe, &r->actions, list) {
		if (act->op == LD_OP_UPDATE) {
			rv = res_update(ls, r, act);
			act->result = rv;
			list_del(&act->list);
			add_client_result(act);
		}
	}

	/*
	 * handle explicit unlock actions
	 */

	list_for_each_entry_safe(act, safe, &r->actions, list) {
		if ((act->op == LD_OP_LOCK) &&
		    (act->mode == LD_LK_IV || act->mode == LD_LK_NL)) {
			act->result = -EINVAL;
			list_del(&act->list);
			add_client_result(act);
		}

		if (act->op == LD_OP_LOCK && act->mode == LD_LK_UN) {
			rv = res_unlock(ls, r, act);

			if (rv == -ENOENT && (act->flags & LD_AF_UNLOCK_CANCEL))
				rv = res_cancel(ls, r, act);

			/*
			 * possible unlock results:
			 * 0: unlock succeeded
			 * -ECANCELED: cancel succeeded
			 * -ENOENT: nothing to unlock or cancel
			 */

			act->result = rv;
			list_del(&act->list);
			add_client_result(act);
		}
	}

	/*
	 * handle implicit unlocks due to client exit,
	 * also clear any outstanding actions for the client
	 */

	list_for_each_entry(act_close, act_close_list, list) {
		res_unlock(ls, r, act_close);
		res_cancel(ls, r, act_close);
	}

	/*
	 * handle freeing a lock for an lv that has been removed
	 */

	list_for_each_entry_safe(act, safe, &r->actions, list) {
		if (act->op == LD_OP_FREE && act->rt == LD_RT_LV) {
			log_debug("S %s R %s free_lv", ls->name, r->name);
			rv = free_lv(ls, r);
			act->result = rv;
			list_del(&act->list);
			add_client_result(act);
			goto r_free;

		}
	}

	/*
	 * handle enable/disable
	 */

	list_for_each_entry_safe(act, safe, &r->actions, list) {
		if (act->op == LD_OP_ENABLE || act->op == LD_OP_DISABLE) {
			rv = res_able(ls, r, act);
			act->result = rv;
			list_del(&act->list);
			add_client_result(act);
		}

		if (!rv && act->op == LD_OP_DISABLE) {
			log_debug("S %s R %s free disabled", ls->name, r->name);
			goto r_free;
		}
	}

	/*
	 * transient requests on existing transient locks
	 */

	list_for_each_entry_safe(act, safe, &r->actions, list) {
		if (act->flags & LD_AF_PERSISTENT)
			continue;

		lk = find_lock_client(r, act->client_id);
		if (!lk)
			continue;

		if (lk->mode != act->mode) {
			/* convert below */
			/*
			act->result = -EEXIST;
			list_del(&act->list);
			add_client_result(act);
			*/
			continue;
		} else {
			/* success */
			act->result = -EALREADY;
			list_del(&act->list);
			add_client_result(act);
		}
	}

	/*
	 * persistent requests on existing persistent locks
	 *
	 * persistent locks are not owned by a client, so any
	 * existing with matching mode satisfies a request.
	 * only one persistent lock is kept on a resource.
	 * a single "unowned" persistent lock satisfies
	 * any/multiple client requests for a persistent lock.
	 */

	list_for_each_entry_safe(act, safe, &r->actions, list) {
		if (!(act->flags & LD_AF_PERSISTENT))
			continue;

		lk = find_lock_persistent(r);
		if (!lk)
			continue;

		if (lk->mode != act->mode) {
			/* convert below */
			/*
			act->result = -EEXIST;
			list_del(&act->list);
			add_client_result(act);
			*/
			continue;
		} else {
			/* success */
			act->result = -EALREADY;
			list_del(&act->list);
			add_client_result(act);
		}
	}

	/*
	 * transient requests with existing persistent locks
	 *
	 * Just grant the transient request and do not
	 * keep a record of it.  Assume that the persistent
	 * lock will not go away while the transient lock
	 * is needed.
	 *
	 * TODO: define exactly when this can be used,
	 * because there are a number of cases where it
	 * will not work: updating version number (lv
	 * locks have none), ex locks from multiple
	 * clients will not conflict, explicit un of the
	 * transient lock will fail.
	 *
	 * This would be used when an ex, persistent lv lock
	 * exists from activation, and then something like
	 * lvextend asks for ex lock to change the lv.  The
	 * lv could not be unlocked by deactivation while
	 * the lvextend was running.
	 */

	list_for_each_entry_safe(act, safe, &r->actions, list) {
		if (act->flags & LD_AF_PERSISTENT)
			continue;

		lk = find_lock_persistent(r);
		if (!lk)
			continue;

		if ((lk->mode == LD_LK_EX) ||
		    (lk->mode == LD_LK_SH && act->mode == LD_LK_SH)) {
			act->result = 0;
			list_del(&act->list);
			add_client_result(act);
		} else {
			/* persistent lock is sh, transient request is ex */
			/* is this case needed? do a convert here? */
			log_debug("res_process %s existing persistent lock new transient", r->name);
			act->result = -EEXIST;
			list_del(&act->list);
			add_client_result(act);
		}
	}

	/*
	 * persistent requests with existing transient locks
	 *
	 * If a client requests a P lock for a T lock it already
	 * holds, we can just change T to P.  Fail if the same
	 * happens for locks from different clients.  Changing
	 * another client's lock from T to P may cause problems
	 * if that client tries to unlock or update version.
	 *
	 * This would be used in a case like vgchange --lock-vg ex vgname
	 * where a transient vg lock was acquired to read the vg,
	 * then the command wants to acquire a persistent lock.
	 * The command could instead unlock, then relock in the mode
	 * it wants, so this case may not be necessary.  Or, lvmlockd
	 * could itself attempt a lock conversion by unlock+relock.
	 */

	list_for_each_entry_safe(act, safe, &r->actions, list) {
		if (!(act->flags & LD_AF_PERSISTENT))
			continue;

		lk = find_lock_client(r, act->client_id);
		if (!lk)
			continue;

		if (lk->mode != act->mode) {
			/* TODO: convert and change to persistent? */
			log_debug("res_process %s existing transient lock new persistent", r->name);
			act->result = -EEXIST;
			list_del(&act->list);
			add_client_result(act);
		} else {
			lk->flags |= LD_LF_PERSISTENT;
			lk->client_id = 0;
			act->result = 0;
			list_del(&act->list);
			add_client_result(act);
		}
	}

	/*
	 * convert mode of existing locks
	 */

	list_for_each_entry_safe(act, safe, &r->actions, list) {
		if (act->flags & LD_AF_PERSISTENT)
			lk = find_lock_persistent(r);
		else
			lk = find_lock_client(r, act->client_id);
		if (!lk)
			continue;

		if (lk->mode == act->mode) {
			/* should never happen, should be found above */
			log_error("convert same mode");
			continue;
		}

		/* convert fails immediately, no EAGAIN retry */
		rv = res_convert(ls, r, lk, act);
		act->result = rv;
		list_del(&act->list);
		add_client_result(act);
	}

	/*
	 * Cases above are all requests addressed by existing locks.
	 * Below handles the rest.  Transient and persistent are
	 * handled the same, except
	 * - if mode of existing lock is incompat with requested,
	 *   leave the act on r->actions
	 * - if r mode is EX, any lock action is blocked, just quit
	 */

	if (r->mode == LD_LK_EX)
		return;

	/*
	 * Retry a lock request that fails due to a lock conflict (-EAGAIN):
	 * if we have not exceeded max retries and lm sets lm_retry (sanlock
	 * transient conflicts from shared lock implementation), or r type
	 * is gl or vg (transient real conflicts we want to hide from command).
	 * lv lock conflicts won't be transient so don't retry them.
	 */

	/*
	 * r mode is SH or UN, pass lock-sh actions to lm
	 */

	list_for_each_entry_safe(act, safe, &r->actions, list) {
		/* grant in order, so break here */
		if (act->op == LD_OP_LOCK && act->mode == LD_LK_EX)
			break;

		if (act->op == LD_OP_LOCK && act->mode == LD_LK_SH) {
			lm_retry = 0;

			rv = res_lock(ls, r, act, &lm_retry);
			if ((rv == -EAGAIN) &&
			    (act->retries <= act->max_retries) &&
			    (lm_retry || (r->type != LD_RT_LV))) {
				/* leave act on list */
				log_debug("S %s R %s res_lock EAGAIN retry", ls->name, r->name);
				act->retries++;
				*retry_out = 1;
			} else {
				act->result = rv;
				list_del(&act->list);
				add_client_result(act);
			}
			if (rv == -EUNATCH)
				goto r_free;
		}
	}

	/*
	 * r mode is SH, any ex lock action is blocked, just quit
	 */

	if (r->mode == LD_LK_SH)
		return;

	/*
	 * r mode is UN, pass lock-ex action to lm
	 */

	list_for_each_entry_safe(act, safe, &r->actions, list) {
		if (act->op == LD_OP_LOCK && act->mode == LD_LK_EX) {
			lm_retry = 0;

			rv = res_lock(ls, r, act, &lm_retry);
			if ((rv == -EAGAIN) &&
			    (act->retries <= act->max_retries) &&
			    (lm_retry || (r->type != LD_RT_LV))) {
				/* leave act on list */
				log_debug("S %s R %s res_lock EAGAIN retry", ls->name, r->name);
				act->retries++;
				*retry_out = 1;
			} else {
				act->result = rv;
				list_del(&act->list);
				add_client_result(act);
			}
			if (rv == -EUNATCH)
				goto r_free;
			break;
		}
	}

	return;

r_free:
	/* For the EUNATCH case it may be possible there are queued actions? */
	list_for_each_entry_safe(act, safe, &r->actions, list) {
		log_error("S %s R %s res_process r_free cancel %s client %d",
			  ls->name, r->name, op_str(act->op), act->client_id);
		act->result = -ECANCELED;
		list_del(&act->list);
		add_client_result(act);
	}
	log_debug("S %s R %s res_process free", ls->name, r->name);
	lm_rem_resource(ls, r);
	list_del(&r->list);
	free_resource(r);
}

#define LOCKS_EXIST_ANY 1
#define LOCKS_EXIST_GL  2
#define LOCKS_EXIST_VG  3
#define LOCKS_EXIST_LV  4

static int for_each_lock(struct lockspace *ls, int locks_do)
{
	struct resource *r;
	struct lock *lk;

	list_for_each_entry(r, &ls->resources, list) {
		list_for_each_entry(lk, &r->locks, list) {
			if (locks_do == LOCKS_EXIST_ANY)
				return 1;

			if (locks_do == LOCKS_EXIST_GL && r->type == LD_RT_GL)
				return 1;

			if (locks_do == LOCKS_EXIST_VG && r->type == LD_RT_VG)
				return 1;

			if (locks_do == LOCKS_EXIST_LV && r->type == LD_RT_LV)
				return 1;
		}
	}

	return 0;
}

static int clear_locks(struct lockspace *ls, int free_vg)
{
	struct resource *r, *r_safe;
	struct lock *lk, *lk_safe;
	struct action *act, *act_safe;
	uint32_t lk_version;
	uint32_t r_version;
	int lk_count = 0;
	int rv;

	list_for_each_entry_safe(r, r_safe, &ls->resources, list) {
		lk_version = 0;

		list_for_each_entry_safe(lk, lk_safe, &r->locks, list) {
			lk_count++;

			if (lk->flags & LD_LF_PERSISTENT)
				log_error("S %s R %s clear lock persistent", ls->name, r->name);
			else
				log_error("S %s R %s clear lock client %d", ls->name, r->name, lk->client_id);

			if (lk->version > lk_version)
				lk_version = lk->version;

			list_del(&lk->list);
			free_lock(lk);
		}

		if (r->mode == LD_LK_UN)
			goto r_free;

		if ((r->type == LD_RT_GL) && (r->mode == LD_LK_EX)) {
			r->version++;
			r_version = r->version;
			log_debug("S %s R %s clear_locks r_version inc %u",
				  ls->name, r->name, r_version);

		} else if ((r->type == LD_RT_VG) && (r->mode == LD_LK_EX) && (lk_version > r->version)) {
			r->version = lk_version;
			r_version = r->version;
			log_debug("S %s R %s clear_locks r_version new %u",
				  ls->name, r->name, r_version);

		} else {
			r_version = 0;
		}

		rv = lm_unlock(ls, r, NULL, r_version, 0, free_vg ? LMUF_FREE_VG : 0);
		if (rv < 0) {
			/* should never happen */
			log_error("S %s R %s clear_locks free %d lm unlock error %d",
				  ls->name, r->name, free_vg, rv);
		}

		list_for_each_entry_safe(act, act_safe, &r->actions, list) {
			log_error("S %s R %s clear_locks cancel %s client %d",
				  ls->name, r->name, op_str(act->op), act->client_id);
			act->result = -ECANCELED;
			list_del(&act->list);
			add_client_result(act);
		}
 r_free:
		log_debug("S %s R %s free", ls->name, r->name);
		lm_rem_resource(ls, r);
		list_del(&r->list);
		free_resource(r);
	}

	return lk_count;
}

/*
 * find and return the resource that is referenced by the action
 * - there is a single gl resource per lockspace
 * - there is a single vg resource per lockspace
 * - there can be many lv resources per lockspace, compare names
 */

static struct resource *find_resource_act(struct lockspace *ls,
					  struct action *act,
					  int nocreate)
{
	struct resource *r;

	list_for_each_entry(r, &ls->resources, list) {
		if (r->type != act->rt)
			continue;

		if (r->type == LD_RT_GL && act->rt == LD_RT_GL)
			return r;

		if (r->type == LD_RT_VG && act->rt == LD_RT_VG)
			return r;

		if (r->type == LD_RT_LV && act->rt == LD_RT_LV &&
		    !strcmp(r->name, act->lv_name))
			return r;
	}

	if (nocreate)
		return NULL;

	r = alloc_resource();
	if (!r)
		return NULL;

	r->type = act->rt;

	r->mode = LD_LK_UN;

	if (r->type == LD_RT_GL)
		strncpy(r->name, R_NAME_GL, MAX_NAME);
	else if (r->type == LD_RT_VG)
		strncpy(r->name, R_NAME_VG, MAX_NAME);
	else if (r->type == LD_RT_LV)
		strncpy(r->name, act->lv_name, MAX_NAME);

	INIT_LIST_HEAD(&r->locks);
	INIT_LIST_HEAD(&r->actions);

	list_add_tail(&r->list, &ls->resources);

	return r;
}

static void free_ls_resources(struct lockspace *ls)
{
	struct resource *r, *r_safe;

	list_for_each_entry_safe(r, r_safe, &ls->resources, list) {
		lm_rem_resource(ls, r);
		list_del(&r->list);
		free_resource(r);
	}
}

/*
 * Process actions queued for this lockspace by
 * client_recv_action / add_lock_action.
 *
 * The lockspace_thread can touch its own ls struct without holding
 * lockspaces_mutex until it sets ls->thread_done, after which it
 * cannot touch ls without holding lockspaces_mutex.
 */

#define LOCK_RETRY_MS 1000 /* milliseconds to delay between retry */

static void *lockspace_thread_main(void *arg_in)
{
	struct lockspace *ls = arg_in;
	struct resource *r, *r2;
	struct action *add_act, *act, *safe;
	struct list_head tmp_act;
	struct list_head act_close;
	int free_vg = 0;
	int error = 0;
	int retry;
	int rv;

	INIT_LIST_HEAD(&act_close);

	/* first action may be client add */
	pthread_mutex_lock(&ls->mutex);
	act = NULL;
	add_act = NULL;
	if (!list_empty(&ls->actions)) {
		act = list_first_entry(&ls->actions, struct action, list);
		if (act->op == LD_OP_START) {
			add_act = act;
			list_del(&add_act->list);
		}
	}
	pthread_mutex_unlock(&ls->mutex);

	log_debug("S %s lm_add_lockspace %s", ls->name, lm_str(ls->lm_type));

	if (add_act && !(add_act->flags & LD_AF_WAIT)) {
		/* send partial join result back to client */
		add_act->result = 0;
		pthread_mutex_lock(&client_mutex);
		list_add_tail(&add_act->list, &client_results);
		pthread_cond_signal(&client_cond);
		pthread_mutex_unlock(&client_mutex);
		add_act = NULL;
	}

	/* the lm join can take a while */

	error = lm_add_lockspace(ls, add_act);

	log_debug("S %s lm_add_lockspace done %d", ls->name, error);

	if (ls->sanlock_gl_enabled && gl_lsname_sanlock[0] &&
	    strcmp(ls->name, gl_lsname_sanlock))
		sanlock_gl_dup = 1;

	if (add_act) {
		/* send synchronous join result back to client */
		add_act->result = error;
		pthread_mutex_lock(&client_mutex);
		list_add_tail(&add_act->list, &client_results);
		pthread_cond_signal(&client_cond);
		pthread_mutex_unlock(&client_mutex);
	}

	pthread_mutex_lock(&ls->mutex);
	if (error) {
		ls->thread_stop = 1;
		ls->create_fail = 1;
	} else {
		ls->create_done = 1;
	}
	pthread_mutex_unlock(&ls->mutex);

	if (error)
		goto out_act;

	while (1) {
		pthread_mutex_lock(&ls->mutex);
		while (!ls->thread_work) {
			if (ls->thread_stop) {
				pthread_mutex_unlock(&ls->mutex);
				goto out_rem;
			}
			pthread_cond_wait(&ls->cond, &ls->mutex);
		}

		/* client thread queues actions on ls->actions, we move
		   ls->actions to r->actions, then process the resources */

		while (1) {
			if (list_empty(&ls->actions)) {
				ls->thread_work = 0;
				break;
			}

			act = list_first_entry(&ls->actions, struct action, list);

			if (sanlock_gl_dup && ls->sanlock_gl_enabled)
				act->flags |= LD_AF_DUP_GL_LS;

			if (act->op == LD_OP_STOP) {
				ls->thread_work = 0;
				break;
			}

			if (act->op == LD_OP_FREE && act->rt == LD_RT_VG) {
				/* vgremove */
				log_debug("S %s checking for lockspace hosts", ls->name);
				rv = lm_hosts(ls, 1);
				if (rv) {
					/*
					 * Checking for hosts here in addition to after the
					 * main loop allows vgremove to fail and be rerun
					 * after the ls is stopped on other hosts.
					 */
					log_error("S %s lockspace hosts %d", ls->name, rv);
					list_del(&act->list);
					act->result = -EBUSY;
					add_client_result(act);
					continue;
				}
				ls->thread_work = 0;
				ls->thread_stop = 1;
				free_vg = 1;
				break;
			}

			if (act->op == LD_OP_RENAME_BEFORE && act->rt == LD_RT_VG) {
				/* vgrename */
				log_debug("S %s checking for lockspace hosts", ls->name);
				rv = lm_hosts(ls, 1);
				if (rv) {
					log_error("S %s lockspace hosts %d", ls->name, rv);
					list_del(&act->list);
					act->result = -EBUSY;
					add_client_result(act);
					continue;
				}
				ls->thread_work = 0;
				ls->thread_stop = 1;
				/* Do we want to check hosts again below like vgremove? */
				break;
			}

			list_del(&act->list);

			/* applies to all resources */
			if (act->op == LD_OP_CLOSE) {
				list_add(&act->list, &act_close);
				continue;
			}

			/*
			 * Find the specific resource this action refers to;
			 * creates resource if not found.
			 */

			r = find_resource_act(ls, act, (act->op == LD_OP_FREE) ? 1 : 0);
			if (!r) {
				act->result = (act->op == LD_OP_FREE) ? -ENOENT : -ENOMEM;
				add_client_result(act);
				continue;
			}

			list_add_tail(&act->list, &r->actions);

			log_debug("S %s R %s action %s %s", ls->name, r->name,
				  op_str(act->op), mode_str(act->mode));
		}
		pthread_mutex_unlock(&ls->mutex);

		retry = 0;

		list_for_each_entry_safe(r, r2, &ls->resources, list)
			res_process(ls, r, &act_close, &retry);

		list_for_each_entry_safe(act, safe, &act_close, list) {
			list_del(&act->list);
			free_action(act);
		}

		if (retry) {
			ls->thread_work = 1;
			usleep(LOCK_RETRY_MS * 1000);
		}
	}

out_rem:
	log_debug("S %s stopping", ls->name);

	/*
	 * For sanlock, we need to unlock any existing locks
	 * before removing the lockspace, otherwise the sanlock
	 * daemon will kill us when the lockspace goes away.
	 * For dlm, we leave with force, so all locks will
	 * automatically be dropped when we leave the lockspace,
	 * so unlocking all before leaving could be skipped.
	 *
	 * Blindly dropping all existing locks must only be
	 * allowed in emergency/force situations, otherwise it's
	 * obviously dangerous, since the lock holders are still
	 * operating under the assumption that they hold the lock.
	 *
	 * For vgremove of a sanlock vg, the vg lock will be held,
	 * and possibly the gl lock if this vg holds the gl.
	 * sanlock vgremove wants to unlock-rename these locks.
	 */

	log_debug("S %s clearing locks", ls->name);

	rv = clear_locks(ls, free_vg);

	/*
	 * Tell any other hosts in the lockspace to leave it
	 * before we remove it (for vgremove).  We do this
	 * before leaving the lockspace ourself because we
	 * need to be in the lockspace to see others.
	 */

	if (free_vg) {
		log_debug("S %s checking for lockspace hosts", ls->name);
		rv = lm_hosts(ls, 1);
		if (rv)
			log_error("S %s other lockspace hosts %d", ls->name, rv);
	}

	/*
	 * Leave the lockspace.
	 */

	rv = lm_rem_lockspace(ls, NULL, free_vg);

	log_debug("S %s rem_lockspace done %d", ls->name, rv);

out_act:
	/*
	 * Move remaining actions to results; this will usually (always?)
	 * be only the stop action.
	 */
	INIT_LIST_HEAD(&tmp_act);

	pthread_mutex_lock(&ls->mutex);
	list_for_each_entry_safe(act, safe, &ls->actions, list) {
		if (act->op == LD_OP_FREE)
			act->result = 0;
		else if (act->op == LD_OP_STOP)
			act->result = 0;
		else if (act->op == LD_OP_RENAME_BEFORE)
			act->result = 0;
		else
			act->result = -ENOLS;
		list_del(&act->list);
		list_add_tail(&act->list, &tmp_act);
	}
	pthread_mutex_unlock(&ls->mutex);

	pthread_mutex_lock(&client_mutex);
	list_for_each_entry_safe(act, safe, &tmp_act, list) {
		list_del(&act->list);
		list_add_tail(&act->list, &client_results);
	}
	pthread_cond_signal(&client_cond);
	pthread_mutex_unlock(&client_mutex);

	pthread_mutex_lock(&lockspaces_mutex);
	ls->thread_done = 1;
	pthread_mutex_unlock(&lockspaces_mutex);

	/*
	 * worker_thread will join this thread, and move the
	 * ls struct from lockspaces list to lockspaces_inactive.
	 */
	pthread_mutex_lock(&worker_mutex);
	worker_wake = 1;
	pthread_cond_signal(&worker_cond);
	pthread_mutex_unlock(&worker_mutex);

	return NULL;
}

static void process_local_ls(struct lockspace *ls)
{
	struct resource *r = list_first_entry(&ls->resources, struct resource, list);
	struct action *act, *act_safe;
	struct lock *lk;
	int prev_mode;
	int result;

	list_for_each_entry_safe(act, act_safe, &ls->actions, list) {
		if (act->op != LD_OP_LOCK)
			continue;
		if (act->mode != LD_LK_UN)
			continue;

		result = -ENOENT;

		list_for_each_entry(lk, &r->locks, list) {
			if (lk->client_id != act->client_id)
				continue;
			list_del(&lk->list);
			free_lock(lk);
			result = 0;
			break;
		}

		act->result = result;
		list_del(&act->list);
		add_client_result(act);
	}

	prev_mode = LD_LK_UN;

	if (!list_empty(&r->locks)) {
		lk = list_first_entry(&r->locks, struct lock, list);
		if (lk->mode == LD_LK_EX)
			return;

		/* sanity check */
		if (lk->mode != LD_LK_SH) {
			log_error("process_local_ls bad lk mode %d", lk->mode);
			return;
		}

		prev_mode = LD_LK_SH;
	}

	/* grant lock requests until we reach one that's one not compat with prev_mode */

	list_for_each_entry_safe(act, act_safe, &ls->actions, list) {

		if (act->mode == LD_LK_EX && prev_mode == LD_LK_UN) {
			/* grant it and return because no more can be granted */

			lk = alloc_lock();
			if (!lk)
				return;

			lk->client_id = act->client_id;
			lk->mode = LD_LK_EX;
			list_add(&lk->list, &r->locks);

			act->result = 0;
			list_del(&act->list);
			add_client_result(act);
			return;

		} else if (act->mode == LD_LK_EX && prev_mode == LD_LK_SH) {

			/* we'll process this act and try to grant it the
			   next we come through here. */

			return;

		} else if (act->mode == LD_LK_SH) {
			prev_mode = LD_LK_SH;

			/* grant it and continue */

			lk = alloc_lock();
			if (!lk)
				return;

			lk->client_id = act->client_id;
			lk->mode = LD_LK_SH;
			list_add_tail(&lk->list, &r->locks);

			act->result = 0;
			list_del(&act->list);
			add_client_result(act);
		}
	}
}

static void purge_local_client(uint32_t client_id)
{
	struct lockspace *ls;
	struct resource *r;
	struct lock *lk, *lk_safe;
	struct action *act, *act_safe;

	list_for_each_entry(ls, &local_vgs, list) {
		r = list_first_entry(&ls->resources, struct resource, list);

		list_for_each_entry_safe(lk, lk_safe, &r->locks, list) {
			if (lk->client_id != client_id)
				continue;
			list_del(&lk->list);
			free_lock(lk);
		}

		list_for_each_entry_safe(act, act_safe, &ls->actions, list) {
			if (act->client_id != client_id)
				continue;
			list_del(&act->list);
			free_action(act);
		}
	}
}

static void *local_thread_main(void *arg_in)
{
	struct lockspace *ls;
	struct action *act, *act_safe;

	while (1) {
		pthread_mutex_lock(&local_thread_mutex);
		while (!local_thread_work) {
			if (local_thread_stop) {
				pthread_mutex_unlock(&local_thread_mutex);
				goto out;
			}
			pthread_cond_wait(&local_thread_cond, &local_thread_mutex);
		}

		/* close actions: clear all locks and actions in all lockspaces for client */
		list_for_each_entry_safe(act, act_safe, &local_thread_actions, list) {
			if (act->op != LD_OP_CLOSE)
				continue;
			purge_local_client(act->client_id);
			list_del(&act->list);
			free_action(act);
		}

		list_for_each_entry(ls, &local_vgs, list) {
			if (list_empty(&ls->actions))
				continue;
			process_local_ls(ls);
		}

		local_thread_work = 0;
		pthread_mutex_unlock(&local_thread_mutex);
	}
out:
	return NULL;
}

int lockspaces_empty(void)
{
	int rv;
	pthread_mutex_lock(&lockspaces_mutex);
	rv = list_empty(&lockspaces);
	pthread_mutex_unlock(&lockspaces_mutex);
	return rv;
}

/*
 * lockspaces_mutex is locked
 *
 * When duplicate sanlock global locks have been seen,
 * this function has a secondary job of counting the
 * number of lockspaces that exist with the gl enabled,
 * with the side effect of setting sanlock_gl_dup back to
 * zero when the duplicates have been removed/disabled.
 */

static struct lockspace *find_lockspace_name(char *ls_name)
{
	struct lockspace *ls_found = NULL;
	struct lockspace *ls;
	int gl_count = 0;

	list_for_each_entry(ls, &lockspaces, list) {
		if (!strcmp(ls->name, ls_name))
			ls_found = ls;

		if (!sanlock_gl_dup && ls_found)
			return ls_found;

		if (sanlock_gl_dup && ls->sanlock_gl_enabled)
			gl_count++;
	}

	/* this is the side effect we want from this function */
	if (sanlock_gl_dup && gl_count < 2)
		sanlock_gl_dup = 0;

	return ls_found;
}

/* local_thread_mutex is locked */
static struct lockspace *find_local_vg(const char *name, const char *uuid)
{
	struct lockspace *ls;

	list_for_each_entry(ls, &local_vgs, list) {
		if (name && name[0] && !strcmp(ls->vg_name, name))
			return ls;
		if (uuid && uuid[0] && !strcmp(ls->vg_uuid, uuid))
			return ls;
	}
	return NULL;
}

/*
 * vgcreate/vgremove of local vgs do add_local/rem_local which
 * updates local_vgs on the local host.  Other hosts' local_vgs
 * are updated with these changes asynchronously when they see
 * the n_version change in the global lock lvb, and do
 * update_local_vgs.
 *
 * So, the global lock n_version and update_local_vgs is about
 * asyncronous propagation of add_local/rem_local to other hosts.
 * Because these are local vgs, they are not used concurrently
 * by multiple hosts, but will be used only by the host in the
 * vg's system_id, which is doing the add_local/rem_local.
 *
 * A local vg created on host1 does not need to be immediately
 * usable on host2, and is not locked between hosts anyway.
 * So, returning a not found error on host2 for a while will
 * be ok.  Once node2 asynchronously updates local_vgs, it
 * would know about a new local vg created on host1.  Then
 * lockd_vg on this vg would change from "not found" ENOLS
 * (as above) to -EOTHERVG (or ELOCALVG if no sysid is set,
 * but hosts shouldn't be actively sharing a vg with no
 * lock_type, so an async delay in this case is not a problem.)
 */

/* local_thread_mutex is locked */
static void add_local_vg(const char *vg_name, const char *vg_uuid, const char *vg_sysid)
{
	struct lockspace *ls;
	struct resource *r;

	/* not really a lockspace, we're just reusing the struct */

	if (!vg_name || !vg_uuid || !vg_name[0] || !vg_uuid[0]) {
		log_error("add_local_vg incomplete %s %s",
			  vg_name ? vg_name : "no-name",
			  vg_uuid ? vg_uuid : "no-uuid");
			  
		return;
	}

	ls = find_local_vg(vg_name, vg_uuid);
	if (ls) {
		if (vg_sysid && ls->vg_sysid[0] && !strcmp(vg_sysid, "none")) {
			log_debug("add_local_vg %s %s clear sysid", vg_name, vg_uuid);
			memset(&ls->vg_sysid, 0, MAX_NAME);
		} else if (vg_sysid && strcmp(ls->vg_sysid, vg_sysid)) {
			log_debug("add_local_vg %s %s update %s", vg_name, vg_uuid, vg_sysid);
			strncpy(ls->vg_sysid, vg_sysid, MAX_NAME);
		}
		return;
	}

	ls = malloc(sizeof(struct lockspace));
	if (!ls)
		return;

	r = alloc_resource();
	if (!r) {
		free(ls);
		return;
	}

	memset(ls, 0, sizeof(struct lockspace));
	strncpy(ls->vg_name, vg_name, MAX_NAME);
	strncpy(ls->vg_uuid, vg_uuid, 64);
	strncpy(ls->vg_sysid, vg_sysid, MAX_NAME);
	INIT_LIST_HEAD(&ls->actions);
	INIT_LIST_HEAD(&ls->resources);

	r->type = LD_RT_VG;
	r->mode = LD_LK_UN;
	strncpy(r->name, R_NAME_VG, MAX_NAME);
	INIT_LIST_HEAD(&r->locks);
	INIT_LIST_HEAD(&r->actions);
	list_add_tail(&r->list, &ls->resources);

	list_add(&ls->list, &local_vgs);

	log_debug("add_local_vg %s %s %s", vg_name, vg_uuid, vg_sysid ?: "");
}

/* local_thread_mutex is locked */
static void rem_local_vg(const char *vg_name, const char *vg_uuid)
{
	struct lockspace *ls;
	struct resource *r;
	struct lock *lk, *lk_safe;
	struct action *act, *act_safe;

	log_debug("rem_local_vg %s %s", vg_name, vg_uuid);

	ls = find_local_vg(vg_name, vg_uuid);
	if (!ls)
		return;

	r = list_first_entry(&ls->resources, struct resource, list);

	list_for_each_entry_safe(lk, lk_safe, &r->locks, list) {
		list_del(&lk->list);
		free_lock(lk);
	}

	list_del(&r->list);
	free_resource(r);

	list_for_each_entry_safe(act, act_safe, &ls->actions, list) {
		list_del(&act->list);
		free_action(act);
	}

	list_del(&ls->list);
	free(ls);
}

static struct lockspace *find_update_vg(struct list_head *head, const char *name, const char *uuid)
{
	struct lockspace *ls;

	list_for_each_entry(ls, head, list) {
		if (!strcmp(ls->vg_name, name) && !strcmp(ls->vg_uuid, uuid))
			return ls;
	}
	return NULL;
}

/*
 * called by worker_thread. the work action is queued when we see that another
 * host has changed the global lock n_version, which means they have changed the
 * global vg name list, so our local_vgs list may need updating.
 *
 * Handle the issue where a lot of devices all appear together,
 * pvscan is run for each of them to populate lvmetad, each pvscan
 * triggers an update_local, and we end up calling this function many
 * times in a row.  We only really need/want one update_local when all
 * the pvscans are done, and this is a rough approximation of that.
 * If we're asked to do update_local within one second of the previous run,
 * then push it off to the delayed work list, so it will be called in a
 * couple seconds.  Ignore more update_local actions while a delayed
 * update_local action exists.  IOW, if we see two quick back to back
 * update_local actions, delay the second one for a couple seconds in
 * an attempt to buffer more of them which can be eliminated.
 */

static uint64_t last_update_local;

static int work_update_local_vgs(void)
{
	struct list_head update_vgs;
	daemon_reply reply;
	struct dm_config_node *cn;
	struct dm_config_node *metadata;
	struct lockspace *lls, *uls, *safe;
	const char *vg_name;
	const char *vg_uuid;
	const char *lock_type;
	const char *system_id;
	int mutex_unlocked = 0;

	INIT_LIST_HEAD(&update_vgs);

	if (monotime() - last_update_local <= 1)
		return -EAGAIN;

	last_update_local = monotime();

	/* get a list of all vg uuids from lvmetad */

	pthread_mutex_lock(&lvmetad_mutex);
	reply = daemon_send_simple(lvmetad_handle, "vg_list",
				   "token = %s", "skip",
				   NULL);

	if (!(cn = dm_config_find_node(reply.cft->root, "volume_groups"))) {
		log_error("work_update_local no vgs");
		goto out;
	}

	/* create an update_vgs list of all vg uuids */

	for (cn = cn->child; cn; cn = cn->sib) {
		vg_uuid = cn->key;

		uls = malloc(sizeof(struct lockspace));
		if (!uls)
			goto out;

		memset(uls, 0, sizeof(struct lockspace));
		strncpy(uls->vg_uuid, vg_uuid, 64);
		list_add_tail(&uls->list, &update_vgs);
		log_debug("work_update_local %s", vg_uuid);
	}

	daemon_reply_destroy(reply);

	/* get vg_name and system_id for each vg uuid entry in update_vgs */

	list_for_each_entry(uls, &update_vgs, list) {
		reply = daemon_send_simple(lvmetad_handle, "vg_lookup",
					   "token = %s", "skip",
					   "uuid = %s", uls->vg_uuid,
					   NULL);

		vg_name = daemon_reply_str(reply, "name", NULL);
		if (!vg_name) {
			log_error("work_update_local %s no name", uls->vg_uuid);
			goto next;
		}

		strncpy(uls->vg_name, vg_name, MAX_NAME);

		metadata = dm_config_find_node(reply.cft->root, "metadata");
		if (!metadata) {
			log_error("work_update_local %s name %s no metadata",
				  uls->vg_uuid, uls->vg_name);
			goto next;
		}

		lock_type = dm_config_find_str(metadata, "metadata/lock_type", NULL);
		uls->lm_type = str_to_lm(lock_type);

		system_id = dm_config_find_str(metadata, "metadata/system_id", NULL);
		if (system_id)
			strncpy(uls->vg_sysid, system_id, MAX_NAME);
next:
		daemon_reply_destroy(reply);

		log_debug("work_update_local %s lock_type %s %d sysid %s %s",
			  uls->vg_name, lock_type ?: "NULL", uls->lm_type, uls->vg_sysid, uls->vg_uuid);

		if (!vg_name || !metadata)
			goto out;
	}
	pthread_mutex_unlock(&lvmetad_mutex);
	mutex_unlocked = 1;

	/* remove local_vgs entries that no longer exist in update_vgs */

	pthread_mutex_lock(&local_thread_mutex);

	list_for_each_entry_safe(lls, safe, &local_vgs, list) {
		uls = find_update_vg(&update_vgs, lls->vg_name, lls->vg_uuid);
		if (!uls) {
			log_debug("work_update_local remove local_vg %s %s",
				  lls->vg_name, lls->vg_uuid);
			list_del(&lls->list);
			free(lls);

		} else if (uls->lm_type != LD_LM_NONE) {
			log_debug("work_update_local remove local_vg %s %s new lm_type %d",
				  lls->vg_name, lls->vg_uuid, uls->lm_type);
			list_del(&lls->list);
			free(lls);
		}
	}

	/* add local_vgs entries for any new non-lockd entries in update_vgs */

	list_for_each_entry_safe(uls, safe, &update_vgs, list) {
		if (uls->lm_type != LD_LM_NONE)
			continue;
		/* add_local_vg doesn't add any that already exist, it may update sysid */
		add_local_vg(uls->vg_name, uls->vg_uuid, uls->vg_sysid);
	}
	pthread_mutex_unlock(&local_thread_mutex);
out:
	list_for_each_entry_safe(uls, safe, &update_vgs, list) {
		list_del(&uls->list);
		free(uls);
	}

	if (!mutex_unlocked)
		pthread_mutex_unlock(&lvmetad_mutex);

	return 0;
}

/*
 * TODO: we don't use the reply here, so it would be more
 * efficient to send without waiting for a reply.
 */

static void invalidate_lvmetad_vg(struct lockspace *ls)
{
	daemon_reply reply;

	pthread_mutex_lock(&lvmetad_mutex);
	reply = daemon_send_simple(lvmetad_handle, "set_vg_info",
				   "token = %s", "skip",
				   "uuid = %s", ls->vg_uuid,
				   "version = %d", 0,
				   NULL);
	pthread_mutex_unlock(&lvmetad_mutex);
	daemon_reply_destroy(reply);
}

/* TODO: handle lvm_<vg_name> longer than max lockspace name?  Use vg uuid? */

static int vg_ls_name(const char *vg_name, char *ls_name)
{
	if (strlen(vg_name) + 4 > MAX_NAME) {
		log_error("vg name too long %s", vg_name);
		return -1;
	}

	snprintf(ls_name, MAX_NAME, "lvm_%s", vg_name);
	return 0;
}

/* TODO: add mutex for gl_lsname_ ? */

static int gl_ls_name(char *ls_name)
{
	if (gl_use_dlm)
		memcpy(ls_name, gl_lsname_dlm, MAX_NAME);
	else if (gl_use_sanlock)
		memcpy(ls_name, gl_lsname_sanlock, MAX_NAME);
	else {
		log_error("gl_ls_name: global lockspace type unknown");
		return -1;
	}
	return 0;
}

/*
 * When this function returns an error, the caller needs to deal
 * with act (in the cases where act exists).
 */

static int add_lockspace_thread(const char *ls_name,
				const char *vg_name,
				const char *vg_uuid,
				int lm_type, const char *vg_args,
				struct action *act)
{
	struct lockspace *ls, *ls2;
	struct resource *r;
	uint32_t version = 0;
	int rv;

	if (act)
		version = act->version;

	log_debug("add_lockspace_thread %s %s version %u",
		  lm_str(lm_type), ls_name, version);

	ls = malloc(sizeof(struct lockspace));
	if (!ls)
		return -ENOMEM;

	memset(ls, 0, sizeof(struct lockspace));

	strncpy(ls->name, ls_name, MAX_NAME);
	ls->lm_type = lm_type;

	if (act)
		ls->start_client_id = act->client_id;

	if (vg_uuid)
		strncpy(ls->vg_uuid, vg_uuid, 64);

	if (vg_name)
		strncpy(ls->vg_name, vg_name, MAX_NAME);

	if (vg_args)
		strncpy(ls->vg_args, vg_args, MAX_ARGS);

	if (act)
		ls->host_id = act->host_id;

	pthread_mutex_init(&ls->mutex, NULL);
	pthread_cond_init(&ls->cond, NULL);
	INIT_LIST_HEAD(&ls->actions);
	INIT_LIST_HEAD(&ls->resources);

	r = alloc_resource();
	if (!r) {
		free(ls);
		return -ENOMEM;
	}

	r->type = LD_RT_VG;
	r->mode = LD_LK_UN;
	r->version = version;
	strncpy(r->name, R_NAME_VG, MAX_NAME);
	INIT_LIST_HEAD(&r->locks);
	INIT_LIST_HEAD(&r->actions);
	list_add_tail(&r->list, &ls->resources);

	pthread_mutex_lock(&lockspaces_mutex);
	ls2 = find_lockspace_name(ls->name);
	if (ls2) {
		if (ls2->thread_stop)
			rv = -EAGAIN;
		else
			rv = -EEXIST;
		pthread_mutex_unlock(&lockspaces_mutex);
		free_resource(r);
		free(ls);
		return rv;
	}

	/*
	 * act will be null when this lockspace is added automatically/internally
	 * and not by an explicit client action that wants a result.
	 */
	if (act)
		list_add(&act->list, &ls->actions);

	clear_lockspace_inactive(ls->name);

	list_add_tail(&ls->list, &lockspaces);
	pthread_mutex_unlock(&lockspaces_mutex);

	rv = pthread_create(&ls->thread, NULL, lockspace_thread_main, ls);
	if (rv < 0) {
		pthread_mutex_lock(&lockspaces_mutex);
		list_del(&ls->list);
		pthread_mutex_unlock(&lockspaces_mutex);
		free_resource(r);
		free(ls);
		return rv;
	}

	return 0;
}

/*
 * There is no add_sanlock_global_lockspace or
 * rem_sanlock_global_lockspace because with sanlock,
 * the global lockspace is one of the vg lockspaces.
 */

static int add_dlm_global_lockspace(struct action *act)
{
	int rv;

	if (gl_running_dlm)
		return -EEXIST;

	gl_running_dlm = 1;

	/* Keep track of whether we automatically added
	   the global ls, so we know to automatically
	   remove it. */

	if (act)
		gl_auto_dlm = 0;
	else
		gl_auto_dlm = 1;

	/*
	 * There's a short period after which a previous gl lockspace thread
	 * has set gl_running_dlm = 0, but before its ls struct has been
	 * deleted, during which this add_lockspace_thread() can fail with
	 * -EAGAIN.
	 */

	rv = add_lockspace_thread(gl_lsname_dlm, NULL, NULL, LD_LM_DLM, NULL, act);

	if (rv < 0) {
		log_error("add_dlm_global_lockspace add_lockspace_thread %d", rv);
		gl_running_dlm = 0;
		gl_auto_dlm = 0;
	}

	return rv;
}

/*
 * If dlm gl lockspace is the only one left, then stop it.
 * This is not used for an explicit rem_lockspace action from
 * the client, only for auto remove.
 */

static int rem_dlm_global_lockspace(void)
{
	struct lockspace *ls, *ls_gl = NULL;
	int others = 0;
	int rv = 0;

	pthread_mutex_lock(&lockspaces_mutex);
	list_for_each_entry(ls, &lockspaces, list) {
		if (!strcmp(ls->name, gl_lsname_dlm)) {
			ls_gl = ls;
			continue;
		}
		if (ls->thread_stop)
			continue;
		others++;
		break;
	}

	if (others) {
		rv = -EAGAIN;
		goto out;
	}

	if (!ls_gl) {
		rv = -ENOENT;
		goto out;
	}

	ls = ls_gl;
	pthread_mutex_lock(&ls->mutex);
	ls->thread_stop = 1;
	ls->thread_work = 1;
	pthread_cond_signal(&ls->cond);
	pthread_mutex_unlock(&ls->mutex);
	rv = 0;
out:
	pthread_mutex_unlock(&lockspaces_mutex);
	return rv;
}

/*
 * When the first dlm lockspace is added for a vg,
 * automatically add a separate dlm lockspace for the
 * global lock if it hasn't been done explicitly.
 * This is to make the dlm global lockspace work similarly to
 * the sanlock global lockspace, which is "automatic" by
 * nature of being one of the vg lockspaces.
 *
 * For sanlock, a separate lockspace is not used for
 * the global lock, but the gl lock lives in a vg
 * lockspace, (although it's recommended to create a
 * special vg dedicated to holding the gl).
 *
 * N.B. for dlm, if this is an add+WAIT action for a vg
 * lockspace, and this triggered the automatic addition
 * of the global lockspace, then the action may complete
 * for the vg ls add, while the gl ls add is still in
 * progress.  If the caller wants to ensure that the
 * gl ls add is complete, they should explicitly add+WAIT
 * the gl ls.
 *
 * If this function returns and error, the caller
 * will queue the act with that error for the client.
 */

static int add_lockspace(struct action *act)
{
	struct lockspace *ls;
	char ls_name[MAX_NAME+1];
	int rv;

	if (local_thread_only) {
		log_error("add_lockspace not allowed local_thread_only");
		return -EINVAL;
	}

	/*
	 * This should not generally happen, but does happen when a vg
	 * lock_type is changed from none to sanlock.
	 */
	pthread_mutex_lock(&local_thread_mutex);
	ls = find_local_vg(act->vg_name, NULL);
	if (ls) {
		log_error("add_lockspace vg %s remove matching local_vg", act->vg_name);
		list_del(&ls->list);
		free_ls_resources(ls);
		free(ls);
	}
	pthread_mutex_unlock(&local_thread_mutex);

	memset(ls_name, 0, sizeof(ls_name));

	if (act->rt == LD_RT_GL) {
		if (gl_use_dlm) {
			rv = add_dlm_global_lockspace(act);
			return rv;
		} else {
			return -EINVAL;
		}
	}

	if (act->rt == LD_RT_VG) {
		if (gl_use_dlm) {
			rv = add_dlm_global_lockspace(NULL);
			if (rv < 0 && rv != -EEXIST)
				return rv;
		}

		vg_ls_name(act->vg_name, ls_name);

		rv = add_lockspace_thread(ls_name, act->vg_name, act->vg_uuid,
					  act->lm_type, act->vg_args,
					  act);

		if (rv)
			log_error("add_lockspace %s add_lockspace_thread %d", ls_name, rv);
		return rv;
	}

	log_error("add_lockspace bad type %d", act->rt);
	return -1;
}

/*
 * vgchange --lock-stop vgname will lock the vg ex, then send a stop,
 * so we exect to find the ex vg lock held here, and will automatically
 * unlock it when stopping.
 *
 * TODO: if the vg contains the gl lock, should we also automatically
 * unlock that when other lockspaces exist?  Or return an error about
 * stopping the vg with the gl lock while other lockspaces are running,
 * and require a force to do that?
 */

static int rem_lockspace(struct action *act)
{
	struct lockspace *ls;
	char ls_name[MAX_NAME+1];
	int force = act->flags & LD_AF_FORCE;
	int rt = act->rt;

	if (act->rt == LD_RT_GL && act->lm_type != LD_LM_DLM)
		return -EINVAL;

	memset(ls_name, 0, sizeof(ls_name));

	if (act->rt == LD_RT_GL)
		gl_ls_name(ls_name);
	else
		vg_ls_name(act->vg_name, ls_name);

	pthread_mutex_lock(&lockspaces_mutex);
	ls = find_lockspace_name(ls_name);
	if (!ls) {
		pthread_mutex_unlock(&lockspaces_mutex);
		return -ENOLS;
	}

	pthread_mutex_lock(&ls->mutex);
	if (ls->thread_stop) {
		pthread_mutex_unlock(&ls->mutex);
		pthread_mutex_unlock(&lockspaces_mutex);
		return -ESTALE;
	}

	if (!force && for_each_lock(ls, LOCKS_EXIST_LV)) {
		pthread_mutex_unlock(&ls->mutex);
		pthread_mutex_unlock(&lockspaces_mutex);
		return -EBUSY;
	}
	ls->thread_work = 1;
	ls->thread_stop = 1;
	if (act)
		list_add_tail(&act->list, &ls->actions);
	pthread_cond_signal(&ls->cond);
	pthread_mutex_unlock(&ls->mutex);
	pthread_mutex_unlock(&lockspaces_mutex);

	/*
	 * If the dlm global lockspace was automatically added when
	 * the first dlm vg lockspace was added, then reverse that
	 * by automatically removing the dlm global lockspace when
	 * the last dlm vg lockspace is removed.
	 */

	if (rt == LD_RT_VG && gl_use_dlm && gl_auto_dlm)
		rem_dlm_global_lockspace();

	return 0;
}

/*
 * count how many lockspaces started by this client are still starting;
 * the client will use this to wait for all its start operations to finish
 * (START_WAIT).
 */

static int count_lockspace_starting(uint32_t client_id)
{
	struct lockspace *ls;
	int count = 0;
	int done = 0;
	int fail = 0;

	pthread_mutex_lock(&lockspaces_mutex);
	list_for_each_entry(ls, &lockspaces, list) {
		if (ls->start_client_id != client_id)
			continue;

		if (!ls->create_done && !ls->create_fail) {
			count++;
			continue;
		}

		if (ls->create_done)
			done++;
		if (ls->create_fail)
			fail++;
	}
	pthread_mutex_unlock(&lockspaces_mutex);

	log_debug("count_lockspace_starting client %u count %d done %d fail %d",
		  client_id, count, done, fail);

	return count;
}

/* lockspaces_mutex is held */
static struct lockspace *find_lockspace_inactive(char *ls_name)
{
	struct lockspace *ls;

	list_for_each_entry(ls, &lockspaces_inactive, list) {
		if (!strcmp(ls->name, ls_name))
			return ls;
	}

	return NULL;
}

/* lockspaces_mutex is held */
static void clear_lockspace_inactive(char *ls_name)
{
	struct lockspace *ls;

	ls = find_lockspace_inactive(ls_name);
	if (ls) {
		list_del(&ls->list);
		free(ls);
	}
}

static void free_lockspaces_inactive(void)
{
	struct lockspace *ls, *safe;

	pthread_mutex_lock(&lockspaces_mutex);
	list_for_each_entry_safe(ls, safe, &lockspaces_inactive, list) {
		list_del(&ls->list);
		free(ls);
	}
	pthread_mutex_unlock(&lockspaces_mutex);
}

/*
 * Loop through all lockspaces, and:
 * - if do_stop is set, stop any that are not stopped
 * - if do_free is set, join any that are done stopping (and free ls)
 *
 * do_stop will not stop an ls with lv locks unless force is set.
 *
 * This function does not block or wait for anything.
 *
 * do_stop (no do_free):
 * returns count of lockspaces that need stop (have locks and no force)
 *
 * do_free (no do_stop):
 * returns count of lockspaces that are stopped and need freeing
 *
 * do_stop and do_free:
 * returns sum of the previous two
 */

static int for_each_lockspace(int do_stop, int do_free, int do_force)
{
	struct lockspace *ls, *safe;
	int need_stop = 0;
	int need_free = 0;
	int stop_count = 0;
	int free_count = 0;
	int done;
	int stop;

	pthread_mutex_lock(&lockspaces_mutex);

	if (do_stop) {
		list_for_each_entry(ls, &lockspaces, list) {

			pthread_mutex_lock(&ls->mutex);
			if (ls->thread_stop) {
				pthread_mutex_unlock(&ls->mutex);
				continue;
			}

			if (!do_force && for_each_lock(ls, LOCKS_EXIST_ANY)) {
				need_stop++;
			} else {
				ls->thread_work = 1;
				ls->thread_stop = 1;
				pthread_cond_signal(&ls->cond);
				stop_count++;
			}
			pthread_mutex_unlock(&ls->mutex);
		}
	}

	if (do_free) {
		list_for_each_entry_safe(ls, safe, &lockspaces, list) {

			pthread_mutex_lock(&ls->mutex);
			done = ls->thread_done;
			stop = ls->thread_stop;
			pthread_mutex_unlock(&ls->mutex);

			/* This ls has locks and force is not set. */
			if (!stop)
				continue;

			/*
			 * Once thread_done is set, we know that the lockspace_thread
			 * will not be using/touching the ls struct.  Any other
			 * thread touches the ls struct under lockspaces_mutex.
			 */
			if (done) {
				pthread_join(ls->thread, NULL);
				list_del(&ls->list);

				/* TODO: remove this if unneeded */
				if (!list_empty(&ls->actions))
					log_error("TODO: free ls actions");

				free_ls_resources(ls);
				list_add(&ls->list, &lockspaces_inactive);
				free_count++;
			} else {
				need_free++;
			}
		}
	}

	if (list_empty(&lockspaces)) {
		if (!gl_type_static) {
			gl_use_dlm = 0;
			gl_use_sanlock = 0;
		}
	}
	pthread_mutex_unlock(&lockspaces_mutex);

	if (stop_count || free_count || need_stop || need_free) {
		log_debug("for_each_lockspace do_stop %d do_free %d "
			  "stop_count %d free_count %d need_stop %d need_free %d",
			  do_stop, do_free, stop_count, free_count, need_stop, need_free);
	}

	return need_stop + need_free;
}

/*
 * This is only called when the daemon is exiting so the sleep/retry
 * loop doesn't have any adverse impact.
 */

static void for_each_lockspace_retry(int do_stop, int do_free, int do_force)
{
	int count;

	while (1) {
		count = for_each_lockspace(do_stop, do_free, do_force);
		if (!count)
			break;

		log_debug("for_each_lockspace_retry remaining %d", count);
		sleep(1);
	}
}

static int work_init_vg(struct action *act)
{
	char ls_name[MAX_NAME+1];
	int rv = 0;

	memset(ls_name, 0, sizeof(ls_name));

	vg_ls_name(act->vg_name, ls_name);

	if (act->lm_type == LD_LM_SANLOCK)
		rv = lm_init_vg_sanlock(ls_name, act->vg_name, act->flags, act->vg_args);
	else if (act->lm_type == LD_LM_DLM)
		rv = lm_init_vg_dlm(ls_name, act->vg_name, act->flags, act->vg_args);
	else
		rv = -EINVAL;

	return rv;
}

static int work_rename_vg(struct action *act)
{
	char ls_name[MAX_NAME+1];
	int rv = 0;

	memset(ls_name, 0, sizeof(ls_name));

	vg_ls_name(act->vg_name, ls_name);

	if (act->lm_type == LD_LM_SANLOCK)
		rv = lm_rename_vg_sanlock(ls_name, act->vg_name, act->flags, act->vg_args);
	else if (act->lm_type == LD_LM_DLM)
		return 0;
	else
		rv = -EINVAL;

	return rv;
}

static void work_test_gl(void)
{
	struct lockspace *ls;
	int is_enabled = 0;

	pthread_mutex_lock(&lockspaces_mutex);
	list_for_each_entry(ls, &lockspaces, list) {
		if (ls->lm_type != LD_LM_SANLOCK)
			continue;

		pthread_mutex_lock(&ls->mutex);
		if (ls->create_done && !ls->thread_stop) {
			is_enabled = lm_gl_is_enabled(ls);
			if (is_enabled) {
				log_debug("S %s worker found gl_is_enabled", ls->name);
				strncpy(gl_lsname_sanlock, ls->name, MAX_NAME);
			}
		}
		pthread_mutex_unlock(&ls->mutex);

		if (is_enabled)
			break;
	}

	if (!is_enabled)
		log_debug("worker found no gl_is_enabled");
	pthread_mutex_unlock(&lockspaces_mutex);
}

static int work_init_lv(struct action *act)
{
	struct lockspace *ls;
	char ls_name[MAX_NAME+1];
	char vg_args[MAX_ARGS];
	char lv_args[MAX_ARGS];
	int lm_type = 0;
	int rv = 0;

	memset(ls_name, 0, sizeof(ls_name));
	memset(vg_args, 0, MAX_ARGS);
	memset(lv_args, 0, MAX_ARGS);

	vg_ls_name(act->vg_name, ls_name);

	pthread_mutex_lock(&lockspaces_mutex);
	ls = find_lockspace_name(ls_name);
	if (ls) {
		lm_type = ls->lm_type;
		memcpy(vg_args, ls->vg_args, MAX_ARGS);
	}
	pthread_mutex_unlock(&lockspaces_mutex);

	if (!ls) {
		lm_type = act->lm_type;
		memcpy(vg_args, act->vg_args, MAX_ARGS);
	}

	if (act->lm_type != lm_type) {
		log_error("init_lv ls_name %s wrong lm_type %d %d",
			  ls_name, act->lm_type, lm_type);
		return -EINVAL;
	}

	if (lm_type == LD_LM_SANLOCK) {
		rv = lm_init_lv_sanlock(ls_name, act->vg_name, act->lv_name,
					vg_args, lv_args);

		memcpy(act->lv_args, lv_args, MAX_ARGS);
		return rv;

	} else if (act->lm_type == LD_LM_DLM) {
		return 0;
	} else {
		log_error("init_lv ls_name %s bad lm_type %d", ls_name, act->lm_type);
		return -EINVAL;
	}
}

/*
 * When an action is queued for the worker_thread, it is processed right away.
 * After processing, some actions need to be retried again in a short while.
 * These actions are put on the delayed_list, and the worker_thread will
 * process these delayed actions again in SHORT_DELAY_PERIOD.
 */

#define SHORT_DELAY_PERIOD 2
#define LONG_DELAY_PERIOD 60

static void *worker_thread_main(void *arg_in)
{
	struct list_head delayed_list;
	struct timespec ts;
	struct action *act, *safe;
	uint64_t last_delayed_time = 0;
	int delayed_update_local = 0;
	int delay_sec = LONG_DELAY_PERIOD;
	int rv;

	INIT_LIST_HEAD(&delayed_list);

	while (1) {
		pthread_mutex_lock(&worker_mutex);
		clock_gettime(CLOCK_REALTIME, &ts);
		ts.tv_sec += delay_sec;
		rv = 0;
		act = NULL;

		while (list_empty(&worker_list) && !worker_stop && !worker_wake && !rv) {
			rv = pthread_cond_timedwait(&worker_cond, &worker_mutex, &ts);
		}
		worker_wake = 0;

		if (worker_stop) {
			pthread_mutex_unlock(&worker_mutex);
			goto out;
		}

		if (!list_empty(&worker_list)) {
			act = list_first_entry(&worker_list, struct action, list);
			list_del(&act->list);
		}
		pthread_mutex_unlock(&worker_mutex);

		/*
		 * Do new work actions before processing delayed work actions.
		 */

		if (!act)
			goto delayed_work;

		if ((act->op == LD_OP_LOCK) && (act->flags & LD_AF_SEARCH_LS)) {
			/*
			 * worker_thread used as a helper to search existing
			 * sanlock vgs for an enabled gl.
			 */
			log_debug("work search for gl");
			work_test_gl();

			/* try again to find a gl lockspace for this act */
			rv = add_lock_action(act);
			if (rv < 0) {
				act->result = rv;
				add_client_result(act);
			}

		} else if ((act->op == LD_OP_INIT) && (act->rt == LD_RT_VG)) {
			log_debug("work init_vg %s", act->vg_name);
			act->result = work_init_vg(act);
			add_client_result(act);

		} else if ((act->op == LD_OP_INIT) && (act->rt == LD_RT_LV)) {
			log_debug("work init_lv %s/%s", act->vg_name, act->lv_name);
			act->result = work_init_lv(act);
			add_client_result(act);

		} else if ((act->op == LD_OP_RENAME_FINAL) && (act->rt == LD_RT_VG)) {
			log_debug("work rename_vg %s", act->vg_name);
			act->result = work_rename_vg(act);
			add_client_result(act);

		} else if (act->op == LD_OP_UPDATE_LOCAL) {
			if (delayed_update_local) {
				log_debug("work update_local ignore repeat");
				free_action(act);
			} else {
				log_debug("work update_local");
				rv = work_update_local_vgs();
				if (rv == -EAGAIN) {
					delayed_update_local = 1;
					list_add(&act->list, &delayed_list);
				} else {
					free_action(act);
				}
			}

		} else if (act->op == LD_OP_START_WAIT) {
			act->result = count_lockspace_starting(act->client_id);
			if (!act->result)
				add_client_result(act);
			else
				list_add(&act->list, &delayed_list);

		} else if (act->op == LD_OP_STOP_ALL) {
			act->result = for_each_lockspace(DO_STOP, DO_FREE, (act->flags & LD_AF_FORCE) ? DO_FORCE : NO_FORCE);
			if (!act->result || !(act->flags & LD_AF_WAIT))
				add_client_result(act);
			else
				list_add(&act->list, &delayed_list);

		} else {
			log_error("work unknown op %d", act->op);
			act->result = -EINVAL;
			add_client_result(act);
		}

 delayed_work:
		/*
		 * We may want to track retry times per action so that
		 * we can delay different actions by different amounts.
		 */

		if (monotime() - last_delayed_time < SHORT_DELAY_PERIOD) {
			delay_sec = 1;
			continue;
		}
		last_delayed_time = monotime();

		list_for_each_entry_safe(act, safe, &delayed_list, list) {
			if (act->op == LD_OP_START_WAIT) {
				log_debug("work delayed start_wait for client %u", act->client_id);
				act->result = count_lockspace_starting(act->client_id);
				if (!act->result) {
					list_del(&act->list);
					add_client_result(act);
				}

			} else if (act->op == LD_OP_UPDATE_LOCAL) {
				log_debug("work delayed update_local");
				rv = work_update_local_vgs();
				if (rv == -EAGAIN)
					continue;
				list_del(&act->list);
				free_action(act);
				delayed_update_local = 0;

			} else if (act->op == LD_OP_STOP_ALL) {
				log_debug("work delayed stop_all");
				act->result = for_each_lockspace(DO_STOP, DO_FREE, (act->flags & LD_AF_FORCE) ? DO_FORCE : NO_FORCE);
				if (!act->result) {
					list_del(&act->list);
					act->result = 0;
					add_client_result(act);
				}
			}
		}

		/*
		 * This is not explicitly queued work, and not delayed work,
		 * but lockspace thread cleanup that's needed when a
		 * lockspace has been stopped/removed or failed to start.
		 */

		for_each_lockspace(NO_STOP, DO_FREE, NO_FORCE);

		if (list_empty(&delayed_list))
			delay_sec = LONG_DELAY_PERIOD;
		else
			delay_sec = 1;
	}
out:
	list_for_each_entry_safe(act, safe, &delayed_list, list) {
		list_del(&act->list);
		free_action(act);
	}

	pthread_mutex_lock(&worker_mutex);
	list_for_each_entry_safe(act, safe, &worker_list, list) {
		list_del(&act->list);
		free_action(act);
	}
	pthread_mutex_unlock(&worker_mutex);
	return NULL;
}

static int setup_worker_thread(void)
{
	int rv;

	INIT_LIST_HEAD(&worker_list);

	pthread_mutex_init(&worker_mutex, NULL);
	pthread_cond_init(&worker_cond, NULL);

	rv = pthread_create(&worker_thread, NULL, worker_thread_main, NULL);
	if (rv)
		return -1;
	return 0;
}

static void close_worker_thread(void)
{
	pthread_mutex_lock(&worker_mutex);
	worker_stop = 1;
	pthread_cond_signal(&worker_cond);
	pthread_mutex_unlock(&worker_mutex);
	pthread_join(worker_thread, NULL);
}

/* client_mutex is locked */
static struct client *find_client_work(void)
{
	struct client *cl;

	list_for_each_entry(cl, &client_list, list) {
		if (cl->recv || cl->dead)
			return cl;
	}
	return NULL;
}

/* client_mutex is locked */
static struct client *find_client_id(uint32_t id)
{
	struct client *cl;

	list_for_each_entry(cl, &client_list, list) {
		if (cl->id == id)
			return cl;
	}
	return NULL;
}

/* client_mutex is locked */
static struct client *find_client_pi(int pi)
{
	struct client *cl;

	list_for_each_entry(cl, &client_list, list) {
		if (cl->pi == pi)
			return cl;
	}
	return NULL;
}

/*
 * wake up poll() because we have added an fd
 * back into pollfd and poll() needs to be restarted
 * to recognize it.
 */
static void restart_poll(void)
{
	write(restart_fds[1], "w", 1);
}

/* poll will take requests from client again, cl->mutex must be held */
static void client_resume(struct client *cl)
{
	if (cl->dead)
		return;

	if (!cl->poll_ignore || cl->fd == -1 || cl->pi == -1) {
		/* shouldn't happen */
		log_error("client_resume %d bad state ig %d fd %d pi %d",
			  cl->id, cl->poll_ignore, cl->fd, cl->pi);
		return;
	}

	pthread_mutex_lock(&pollfd_mutex);
	if (pollfd[cl->pi].fd != POLL_FD_IGNORE) {
		log_error("client_resume %d pi %d fd %d not IGNORE",
			  cl->id, cl->pi, cl->fd);
	}
	pollfd[cl->pi].fd = cl->fd;
	pollfd[cl->pi].events = POLLIN;
	pthread_mutex_unlock(&pollfd_mutex);

	restart_poll();
}

/* called from client_thread, cl->mutex is held */
static void client_send_result(struct client *cl, struct action *act)
{
	response res;
	char result_flags[128];

	if (cl->dead) {
		log_debug("client send %d skip dead", cl->id);
		return;
	}

	memset(result_flags, 0, sizeof(result_flags));

	buffer_init(&res.buffer);

	/*
	 * EUNATCH is returned when the global lock existed,
	 * but had been disabled when we tried to lock it,
	 * so we removed it, and no longer have a gl to lock.
	 */

	if (act->result == -EUNATCH)
		act->result = -ENOLS;

	/*
	 * init_vg with dlm|sanlock returns vg_args
	 * init_lv with sanlock returns lv_args
	 */

	if (act->result == -ENOLS) {
		/*
		 * The lockspace could not be found, in which case
		 * the caller may want to know if any lockspaces exist
		 * or if lockspaces exist, but not one with the global lock.
		 * Given this detail, it may be able to procede without
		 * the lock.
		 */
		pthread_mutex_lock(&lockspaces_mutex);
		if (list_empty(&lockspaces))
			strcat(result_flags, "NO_LOCKSPACES,");
		pthread_mutex_unlock(&lockspaces_mutex);

		if (gl_use_sanlock && !gl_lsname_sanlock[0])
			strcat(result_flags, "NO_GL_LS,");
		else if (gl_use_dlm && !gl_lsname_dlm[0])
			strcat(result_flags, "NO_GL_LS,");
		else
			strcat(result_flags, "NO_GL_LS,");
	}

	if (act->flags & LD_AF_LOCAL_LS)
		strcat(result_flags, "LOCAL_LS,");

	if (act->flags & LD_AF_DUP_GL_LS)
		strcat(result_flags, "DUP_GL_LS,");

	if (act->flags & LD_AF_INACTIVE_LS)
		strcat(result_flags, "INACTIVE_LS,");

	if (act->flags & LD_AF_ADD_LS_ERROR)
		strcat(result_flags, "ADD_LS_ERROR,");
	
	if (act->op == LD_OP_INIT) {
		/*
		 * init is a special case where lock args need
		 * to be passed back to the client.
		 */
		const char *vg_args = "none";
		const char *lv_args = "none";

		if (act->vg_args[0])
			vg_args = act->vg_args;

		if (act->lv_args[0])
			lv_args = act->lv_args;

		log_debug("send %s[%d.%u] %s %s rv %d vg_args %s lv_args %s",
			  cl->name[0] ? cl->name : "client", cl->pid, cl->id,
			  op_str(act->op), rt_str(act->rt),
			  act->result, vg_args ? vg_args : "", lv_args ? lv_args : "");

		res = daemon_reply_simple("OK",
					  "op = %d", act->op,
					  "op_result = %d", act->result,
					  "lm_result = %d", act->lm_rv,
					  "vg_lock_args = %s", vg_args,
					  "lv_lock_args = %s", lv_args,
					  "result_flags = %s", result_flags[0] ? result_flags : "none",
					  NULL);
	} else {
		/*
		 * A normal reply.
		 */

		log_debug("send %s[%d.%u] %s %s rv %d %s %s",
			  cl->name[0] ? cl->name : "client", cl->pid, cl->id,
			  op_str(act->op), rt_str(act->rt),
			  act->result, (act->result == -ENOLS) ? "ENOLS" : "", result_flags);

		res = daemon_reply_simple("OK",
					  "op = %d", act->op,
					  "lock_type = %s", lm_str(act->lm_type),
					  "op_result = %d", act->result,
					  "lm_result = %d", act->lm_rv,
					  "result_flags = %s", result_flags[0] ? result_flags : "none",
					  NULL);
	}

	buffer_write(cl->fd, &res.buffer);
	buffer_destroy(&res.buffer);

	client_resume(cl);
}

/*
 * TODO: optimize common case where a client has only locked
 * a couple of lockspaces.  Keep two lockspace ids in cl->ls_ids
 * and queue OP_CLOSE in only these lockspaces.  If more than two
 * are used, then clear ls_ids and queue for all.
 */

/* called from client_thread */
static void client_purge(struct client *cl)
{
	struct lockspace *ls;
	struct action *act;

	pthread_mutex_lock(&lockspaces_mutex);
	list_for_each_entry(ls, &lockspaces, list) {
		act = alloc_action();
		if (!act)
			continue;

		act->op = LD_OP_CLOSE;
		act->client_id = cl->id;
		act->flags |= LD_AF_CLIENT_DEAD;

		pthread_mutex_lock(&ls->mutex);
		if (!ls->thread_stop) {
			list_add_tail(&act->list, &ls->actions);
			ls->thread_work = 1;
			pthread_cond_signal(&ls->cond);
		} else {
			free_action(act);
		}
		pthread_mutex_unlock(&ls->mutex);
	}
	pthread_mutex_unlock(&lockspaces_mutex);

	if (local_thread_also) {
		act = alloc_action();
		if (!act)
			return;

		act->op = LD_OP_CLOSE;
		act->client_id = cl->id;
		act->flags |= LD_AF_CLIENT_DEAD;

		pthread_mutex_lock(&local_thread_mutex);
		list_add_tail(&act->list, &local_thread_actions);
		local_thread_work = 1;
		pthread_cond_signal(&local_thread_cond);
		pthread_mutex_unlock(&local_thread_mutex);
	}
}

static int add_lock_action(struct action *act)
{
	struct lockspace *ls = NULL;
	char ls_name[MAX_NAME+1];

	memset(ls_name, 0, sizeof(ls_name));

	/* Determine which lockspace this action is for, and set ls_name. */

	if (act->rt == LD_RT_GL && gl_use_sanlock &&
	    (act->op == LD_OP_ENABLE || act->op == LD_OP_DISABLE))
		vg_ls_name(act->vg_name, ls_name);
	else if (act->rt == LD_RT_GL)
		gl_ls_name(ls_name);
	else
		vg_ls_name(act->vg_name, ls_name);

 retry:
	pthread_mutex_lock(&lockspaces_mutex);
	if (ls_name[0])
		ls = find_lockspace_name(ls_name);
	if (!ls) {
		int ls_inactive = 0;
		int ls_create_fail = 0;

		ls = find_lockspace_inactive(ls_name);
		if (ls) {
			ls_inactive = 1;
			ls_create_fail = ls->create_fail;
			ls = NULL;
		}
		pthread_mutex_unlock(&lockspaces_mutex);

		if (act->op == LD_OP_UPDATE && act->rt == LD_RT_VG) {
			log_debug("lockspace not found ignored for vg update");
			return -ENOLS;

		} else if (act->flags & LD_AF_SEARCH_LS) {
			/* fail if we've already tried searching for the ls */
			log_error("lockspace search repeated %s", ls_name);
			return -ENOLS;

		} else if (act->op == LD_OP_LOCK && act->rt == LD_RT_GL && gl_use_sanlock) {
			/* gl may have been enabled in an existing vg */
			log_debug("gl lockspace not found check sanlock vgs");
			act->flags |= LD_AF_SEARCH_LS;
			add_work_action(act);
			return 0;

		} else if (act->op == LD_OP_LOCK && act->rt == LD_RT_GL && gl_use_dlm) {
			log_debug("gl lockspace not found add dlm global");
			act->flags |= LD_AF_SEARCH_LS;
			act->flags |= LD_AF_WAIT_STARTING;
			add_dlm_global_lockspace(NULL);
			gl_ls_name(ls_name);
			goto retry;

		} else if (act->op == LD_OP_LOCK && act->mode == LD_LK_UN) {
			log_debug("lockspace not found ignored for unlock");
			return -ENOLS;

		} else if (act->op == LD_OP_LOCK && act->rt == LD_RT_VG && ls_inactive) {
			/* ls has been stopped or previously failed to start */
			log_debug("lockspace inactive create_fail %d %s",
				  ls_create_fail, ls_name);
			act->flags |= LD_AF_INACTIVE_LS;
			if (ls_create_fail)
				act->flags |= LD_AF_ADD_LS_ERROR;
			return -ENOLS;

		} else {
			log_error("lockspace not found %s", ls_name);
			return -ENOLS;
		}
	}

	if (act->lm_type == LD_LM_NONE) {
		/* return to the command the type we are using */
		act->lm_type = ls->lm_type;
	} else if (act->lm_type != ls->lm_type) {
		/* should not happen */
		log_error("S %s add_lock_action bad lm_type %d ls %d",
			  ls_name, act->lm_type, ls->lm_type);
		return -EINVAL;
	}

	pthread_mutex_lock(&ls->mutex);
	if (ls->thread_stop) {
		pthread_mutex_unlock(&ls->mutex);
		pthread_mutex_unlock(&lockspaces_mutex);
		log_error("lockspace is stopping %s", ls_name);
		return -ESTALE;
	}

	if (!ls->create_fail && !ls->create_done && !(act->flags & LD_AF_WAIT_STARTING)) {
		pthread_mutex_unlock(&ls->mutex);
		pthread_mutex_unlock(&lockspaces_mutex);
		log_debug("lockspace is starting %s", ls_name);
		return -ESTARTING;
	}

	list_add_tail(&act->list, &ls->actions);
	ls->thread_work = 1;
	pthread_cond_signal(&ls->cond);
	pthread_mutex_unlock(&ls->mutex);
	pthread_mutex_unlock(&lockspaces_mutex);

	/* lockspace_thread_main / res_process take it from here */

	return 0;
}

static int add_local_lock_action(struct lockspace *ls, struct action *act)
{
	act->flags |= LD_AF_LOCAL_LS;
	pthread_mutex_lock(&local_thread_mutex);
	if (!ls && local_thread_only)
		list_add_tail(&act->list, &local_thread_gls->actions);
	else if (ls)
		list_add_tail(&act->list, &ls->actions);
	local_thread_work = 1;
	pthread_cond_signal(&local_thread_cond);
	pthread_mutex_unlock(&local_thread_mutex);
	return 0;
}

static int str_to_op_rt(const char *req_name, int *op, int *rt)
{
	if (!req_name)
		goto out;

	if (!strcmp(req_name, "hello")) {
		*op = LD_OP_HELLO;
		*rt = 0;
		return 0;
	}
	if (!strcmp(req_name, "quit")) {
		*op = LD_OP_QUIT;
		*rt = 0;
		return 0;
	}
	if (!strcmp(req_name, "info")) {
		*op = LD_OP_DUMP_INFO;
		*rt = 0;
		return 0;
	}
	if (!strcmp(req_name, "dump")) {
		*op = LD_OP_DUMP_LOG;
		*rt = 0;
		return 0;
	}
	if (!strcmp(req_name, "init_vg")) {
		*op = LD_OP_INIT;
		*rt = LD_RT_VG;
		return 0;
	}
	if (!strcmp(req_name, "init_lv")) {
		*op = LD_OP_INIT;
		*rt = LD_RT_LV;
		return 0;
	}
	if (!strcmp(req_name, "free_vg")) {
		*op = LD_OP_FREE;
		*rt = LD_RT_VG;
		return 0;
	}
	if (!strcmp(req_name, "free_lv")) {
		*op = LD_OP_FREE;
		*rt = LD_RT_LV;
		return 0;
	}
	if (!strcmp(req_name, "start_vg")) {
		*op = LD_OP_START;
		*rt = LD_RT_VG;
		return 0;
	}
	if (!strcmp(req_name, "stop_vg")) {
		*op = LD_OP_STOP;
		*rt = LD_RT_VG;
		return 0;
	}
	if (!strcmp(req_name, "start_wait")) {
		*op = LD_OP_START_WAIT;
		*rt = 0;
		return 0;
	}
	if (!strcmp(req_name, "stop_all")) {
		*op = LD_OP_STOP_ALL;
		*rt = 0;
		return 0;
	}
	if (!strcmp(req_name, "lock_gl")) {
		*op = LD_OP_LOCK;
		*rt = LD_RT_GL;
		return 0;
	}
	if (!strcmp(req_name, "lock_vg")) {
		*op = LD_OP_LOCK;
		*rt = LD_RT_VG;
		return 0;
	}
	if (!strcmp(req_name, "lock_lv")) {
		*op = LD_OP_LOCK;
		*rt = LD_RT_LV;
		return 0;
	}
	if (!strcmp(req_name, "vg_update")) {
		*op = LD_OP_UPDATE;
		*rt = LD_RT_VG;
		return 0;
	}
	if (!strcmp(req_name, "enable_gl")) {
		*op = LD_OP_ENABLE;
		*rt = LD_RT_GL;
		return 0;
	}
	if (!strcmp(req_name, "disable_gl")) {
		*op = LD_OP_DISABLE;
		*rt = LD_RT_GL;
		return 0;
	}
	if (!strcmp(req_name, "add_local")) {
		*op = LD_OP_ADD_LOCAL;
		return 0;
	}
	if (!strcmp(req_name, "rem_local")) {
		*op = LD_OP_REM_LOCAL;
		return 0;
	}
	if (!strcmp(req_name, "update_local")) {
		*op = LD_OP_UPDATE_LOCAL;
		return 0;
	}
	if (!strcmp(req_name, "rename_vg_before")) {
		*op = LD_OP_RENAME_BEFORE;
		*rt = LD_RT_VG;
		return 0;
	}
	if (!strcmp(req_name, "rename_vg_final")) {
		*op = LD_OP_RENAME_FINAL;
		*rt = LD_RT_VG;
		return 0;
	}
out:
	return -1;
}

static int str_to_mode(const char *str)
{
	if (!str)
		goto out;
	if (!strcmp(str, "un"))
		return LD_LK_UN;
	if (!strcmp(str, "nl"))
		return LD_LK_NL;
	if (!strcmp(str, "sh"))
		return LD_LK_SH;
	if (!strcmp(str, "ex"))
		return LD_LK_EX;
out:
	return LD_LK_IV;
}

static int str_to_lm(const char *str)
{
	if (!str || !strcmp(str, "none"))
		return LD_LM_NONE;
	if (!strcmp(str, "sanlock"))
		return LD_LM_SANLOCK;
	if (!strcmp(str, "dlm"))
		return LD_LM_DLM;
	return -2; 
}

static uint32_t str_to_opts(const char *str)
{
	uint32_t flags = 0;

	if (!str)
		goto out;
	if (strstr(str, "persistent"))
		flags |= LD_AF_PERSISTENT;
	if (strstr(str, "unlock_cancel"))
		flags |= LD_AF_UNLOCK_CANCEL;
	if (strstr(str, "next_version"))
		flags |= LD_AF_NEXT_VERSION;
	if (strstr(str, "wait"))
		flags |= LD_AF_WAIT;
	if (strstr(str, "force"))
		flags |= LD_AF_FORCE;
	if (strstr(str, "ex_disable"))
		flags |= LD_AF_EX_DISABLE;
	if (strstr(str, "enable"))
		flags |= LD_AF_ENABLE;
	if (strstr(str, "disable"))
		flags |= LD_AF_DISABLE;
	if (strstr(str, "update_names"))
		flags |= LD_AF_UPDATE_NAMES_VERSION;
out:
	return flags;
}

static int is_other_sysid(struct lockspace *lls)
{
	if (!our_system_id || !lls->vg_sysid[0])
		return 0;
	if (!strcmp(lls->vg_sysid, our_system_id))
		return 0;
	return 1;
}


/*
 * dump info
 * client_list: each client struct
 * local_vgs: each lockspace struct (representing a local vg)
 * lockspaces: each lockspace struct
 * lockspace actions: each action struct
 * lockspace resources: each resource struct
 * lockspace resource actions: each action struct
 * lockspace resource locks: each lock struct
 */

static int setup_dump_socket(void)
{
	int s;

	s = socket(AF_LOCAL, SOCK_DGRAM, 0);
	if (s < 0)
		return s;

	memset(&dump_addr, 0, sizeof(dump_addr));
	dump_addr.sun_family = AF_LOCAL;
	strcpy(&dump_addr.sun_path[1], DUMP_SOCKET_NAME);
	dump_addrlen = sizeof(sa_family_t) + strlen(dump_addr.sun_path+1) + 1;

	return s;
}

static int send_dump_buf(int fd, int dump_len)
{
	int pos = 0;
	int ret;

retry:
	ret = sendto(fd, dump_buf + pos, dump_len - pos, MSG_DONTWAIT | MSG_NOSIGNAL,
		     (struct sockaddr *)&dump_addr, dump_addrlen);
	if (ret <= 0)
		return ret;

	pos += ret;

	if (pos < dump_len)
		goto retry;

	return 0;
}

static int print_client(struct client *cl, const char *prefix, int pos, int len)
{
	return snprintf(dump_buf + pos, len - pos,
			"info=%s "
			"pid=%d "
			"fd=%d "
			"pi=%d "
			"id=%u "
			"name=%s\n",
			prefix,
			cl->pid,
			cl->fd,
			cl->pi,
			cl->id,
			cl->name[0] ? cl->name : ".");
}

static int print_local_vg(struct lockspace *ls, const char *prefix, int pos, int len)
{
	return snprintf(dump_buf + pos, len - pos,
			"info=%s "
			"vg_name=%s "
			"vg_uuid=%s "
			"vg_sysid=%s\n",
			prefix,
			ls->vg_name,
			ls->vg_uuid,
			ls->vg_sysid[0] ? ls->vg_sysid : ".");
}

static int print_lockspace(struct lockspace *ls, const char *prefix, int pos, int len)
{
	return snprintf(dump_buf + pos, len - pos,
			"info=%s "
			"ls_name=%s "
			"vg_name=%s "
			"vg_uuid=%s "
			"vg_sysid=%s "
			"vg_args=%s "
			"lm_type=%s "
			"host_id=%llu "
			"names_version=%u "
			"create_fail=%d "
			"create_done=%d "
			"thread_work=%d "
			"thread_stop=%d "
			"thread_done=%d "
			"update_local_vgs=%d "
			"update_names_version=%d "
			"sanlock_gl_enabled=%d "
			"sanlock_gl_dup=%d\n",
			prefix,
			ls->name,
			ls->vg_name,
			ls->vg_uuid,
			ls->vg_sysid[0] ? ls->vg_sysid : ".",
			ls->vg_args,
			lm_str(ls->lm_type),
			(unsigned long long)ls->host_id,
			ls->names_version,
			ls->create_fail ? 1 : 0,
			ls->create_done ? 1 : 0,
			ls->thread_work ? 1 : 0,
			ls->thread_stop ? 1 : 0,
			ls->thread_done ? 1 : 0,
			ls->update_local_vgs ? 1 : 0,
			ls->update_names_version ? 1 : 0,
			ls->sanlock_gl_enabled ? 1 : 0,
			ls->sanlock_gl_dup ? 1 : 0);
}

static int print_action(struct action *act, const char *prefix, int pos, int len)
{
	return snprintf(dump_buf + pos, len - pos,
			"info=%s "
			"client_id=%u "
			"flags=0x%x "
			"version=%u "
			"op=%s "
			"rt=%s "
			"mode=%s "
			"lm_type=%s "
			"result=%d "
			"lm_rv=%d\n",
			prefix,
			act->client_id,
			act->flags,
			act->version,
			op_str(act->op),
			rt_str(act->rt),
			mode_str(act->mode),
			lm_str(act->lm_type),
			act->result,
			act->lm_rv);
}

static int print_resource(struct resource *r, const char *prefix, int pos, int len)
{
	return snprintf(dump_buf + pos, len - pos,
			"info=%s "
			"name=%s "
			"type=%s "
			"mode=%s "
			"sh_count=%d "
			"version=%u\n",
			prefix,
			r->name,
			rt_str(r->type),
			mode_str(r->mode),
			r->sh_count,
			r->version);
}

static int print_lock(struct lock *lk, const char *prefix, int pos, int len)
{
	return snprintf(dump_buf + pos, len - pos,
			"info=%s "
			"mode=%s "
			"version=%u "
			"flags=0x%x "
			"client_id=%u\n",
			prefix,
			mode_str(lk->mode),
			lk->version,
			lk->flags,
			lk->client_id);
}

static int dump_info(int *dump_len)
{
	struct client *cl;
	struct lockspace *ls;
	struct resource *r;
	struct lock *lk;
	struct action *act;
	int len, pos, ret;
	int rv = 0;

	memset(dump_buf, 0, sizeof(dump_buf));
	len = sizeof(dump_buf);
	pos = 0;

	/*
	 * clients
	 */

	pthread_mutex_lock(&client_mutex);
	list_for_each_entry(cl, &client_list, list) {
		ret = print_client(cl, "client", pos, len);
		if (ret >= len - pos) {
			 rv = -ENOSPC;
			 break;
		}
		pos += ret;
	}
	pthread_mutex_unlock(&client_mutex);

	if (rv < 0)
		return rv;

	/*
	 * local vgs
	 */

	pthread_mutex_lock(&lockspaces_mutex);
	list_for_each_entry(ls, &local_vgs, list) {
		ret = print_local_vg(ls, "local_vg", pos, len);
		if (ret >= len - pos) {
			 rv = -ENOSPC;
			 break;
		}
		pos += ret;
	}
	pthread_mutex_unlock(&lockspaces_mutex);

	if (rv < 0)
		return rv;

	/*
	 * lockspaces with their action/resource/lock info
	 */

	pthread_mutex_lock(&lockspaces_mutex);
	list_for_each_entry(ls, &lockspaces, list) {

		ret = print_lockspace(ls, "ls", pos, len);
		if (ret >= len - pos) {
			 rv = -ENOSPC;
			 goto out;
		}
		pos += ret;

		list_for_each_entry(act, &ls->actions, list) {
			ret = print_action(act, "ls_action", pos, len);
			if (ret >= len - pos) {
				rv = -ENOSPC;
				goto out;
			}
			pos += ret;
		}

		list_for_each_entry(r, &ls->resources, list) {
			ret = print_resource(r, "r", pos, len);
			if (ret >= len - pos) {
				rv = -ENOSPC;
				goto out;
			}
			pos += ret;

			list_for_each_entry(lk, &r->locks, list) {
				ret = print_lock(lk, "lk", pos, len);
				if (ret >= len - pos) {
					rv = -ENOSPC;
					goto out;
				}
				pos += ret;
			}

			list_for_each_entry(act, &r->actions, list) {
				ret = print_action(act, "r_action", pos, len);
				if (ret >= len - pos) {
					rv = -ENOSPC;
					goto out;
				}
				pos += ret;
			}
		}
	}
out:
	pthread_mutex_unlock(&lockspaces_mutex);

	*dump_len = pos;

	return rv;
}

/* called from client_thread, cl->mutex is held */
static void client_recv_action(struct client *cl)
{
	request req;
	response res;
	struct lockspace *lls = NULL;
	struct action *act;
	const char *cl_name;
	const char *vg_name;
	const char *vg_uuid;
	const char *vg_sysid;
	const char *str;
	int64_t val;
	uint32_t opts = 0;
	int result = 0;
	int cl_pid;
	int op, rt, lm, mode;
	int rv;

	buffer_init(&req.buffer);

	rv = buffer_read(cl->fd, &req.buffer);
	if (!rv) {
		if (errno == ECONNRESET) {
			log_debug("client recv %d ECONNRESET", cl->id);
			cl->dead = 1;
		} else {
			log_error("client recv %d buffer_read error %d", cl->id, errno);
		}
		buffer_destroy(&req.buffer);
		client_resume(cl);
		return;
	}

	req.cft = dm_config_from_string(req.buffer.mem);
	if (!req.cft) {
		log_error("client recv %d config_from_string error", cl->id);
		buffer_destroy(&req.buffer);
		client_resume(cl);
		return;
	}

	str = daemon_request_str(req, "request", NULL);
	rv = str_to_op_rt(str, &op, &rt);
	if (rv < 0) {
		log_error("client recv %d bad request name \"%s\"", cl->id, str ? str : "");
		dm_config_destroy(req.cft);
		buffer_destroy(&req.buffer);
		client_resume(cl);
		return;
	}

	if (op == LD_OP_HELLO || op == LD_OP_QUIT ||
	    op == LD_OP_DUMP_INFO || op == LD_OP_DUMP_LOG) {

		/*
		 * TODO: add the client command name to the hello messages
		 * so it can be saved in cl->name here.
		 */

		result = 0;

		if (op == LD_OP_QUIT) {
			log_debug("op quit");
			pthread_mutex_lock(&lockspaces_mutex);
		       	if (list_empty(&lockspaces)) {
				daemon_quit = 1;
			} else {
				result = -EBUSY;
			}
			pthread_mutex_unlock(&lockspaces_mutex);
		}

		buffer_init(&res.buffer);

		if (op == LD_OP_DUMP_INFO || op == LD_OP_DUMP_LOG) {
			int dump_len = 0;
			int fd;

			fd = setup_dump_socket();
			if (fd < 0)
				result = fd;
			else if (op == LD_OP_DUMP_INFO)
				result = dump_info(&dump_len);
			else if (op == LD_OP_DUMP_LOG)
				result = dump_log(&dump_len);
			else
				result = -EINVAL;

			res = daemon_reply_simple("OK",
					  "result = %d", result,
					  "dump_len = %d", dump_len,
					  NULL);
			if (fd >= 0) {
				send_dump_buf(fd, dump_len);
				close(fd);
			}

		} else {
			res = daemon_reply_simple("OK",
					  "result = %d", result,
					  "protocol = %s", lvmlockd_protocol,
					  "version = %d", lvmlockd_protocol_version,
					  NULL);
		}

		buffer_write(cl->fd, &res.buffer);
		buffer_destroy(&res.buffer);
		dm_config_destroy(req.cft);
		buffer_destroy(&req.buffer);
		client_resume(cl);
		return;
	}

	cl_name = daemon_request_str(req, "cmd", NULL);
	cl_pid = daemon_request_int(req, "pid", 0);
	vg_name = daemon_request_str(req, "vg_name", NULL);
	vg_uuid = daemon_request_str(req, "vg_uuid", NULL);
	vg_sysid = daemon_request_str(req, "vg_sysid", NULL);
	str = daemon_request_str(req, "mode", NULL);
	mode = str_to_mode(str);
	str = daemon_request_str(req, "opts", NULL);
	opts = str_to_opts(str);
	str = daemon_request_str(req, "vg_lock_type", NULL);
	lm = str_to_lm(str);

	if (cl_pid && cl_pid != cl->pid)
		log_error("client recv bad message pid %d client %d", cl_pid, cl->pid);

	/* TODO: do this in hello message instead */
	if (!cl->name[0] && cl_name)
		strncpy(cl->name, cl_name, MAX_NAME-1);

	if (!our_system_id) {
		str = daemon_request_str(req, "our_system_id", NULL);
		if (str && strcmp(str, "none"))
			our_system_id = strdup(str);
	}

	/*
	 * Detect the common case of a lock op on a local vg and queue
	 * a reply immediately without going through a thread.
	 */

	if (rt == LD_RT_VG && op == LD_OP_LOCK) {
		pthread_mutex_lock(&local_thread_mutex);
		lls = find_local_vg(vg_name, vg_uuid);
		pthread_mutex_unlock(&local_thread_mutex);
		if (lls)
			result = is_other_sysid(lls) ? -EOTHERVG : -ELOCALVG;
	}

	/*
	 * A local vg with no sysid, accessible from multiple hosts, can be
	 * modified without coordination if a user is not careful.  The best we
	 * can do is disable the lvmetad cache for these vgs so any problems are
	 * detected earlier, and not masked by lvmetad caching.
	 */

	if (lls && (result == -ELOCALVG) && !lls->vg_sysid[0])
		invalidate_lvmetad_vg(lls);

	if ((result == -EOTHERVG) || (result == -ELOCALVG && !local_thread_also)) {
		const char *sysid = lls->vg_sysid[0] ? lls->vg_sysid : "none";

		log_debug("local vg %s result %d %s sysid %s", vg_name, result,
			  (result == -EOTHERVG) ? "other" : "local", sysid);

		buffer_init(&res.buffer);
		res = daemon_reply_simple("OK",
					  "op_result = %d", result,
					  "vg_sysid = %s", sysid,
					  "lock_type = %s", "none",
					  "result_flags = %s", "LOCAL_LS",
					  NULL);
		buffer_write(cl->fd, &res.buffer);
		buffer_destroy(&res.buffer);
		dm_config_destroy(req.cft);
		buffer_destroy(&req.buffer);
		client_resume(cl);
		return;
	}

	if (!gl_use_dlm && !gl_use_sanlock && (lm > 0)) {
		if (lm == LD_LM_DLM)
			gl_use_dlm = 1;
		else if (lm == LD_LM_SANLOCK)
			gl_use_sanlock = 1;

		log_debug("set gl_use_%s", lm_str(lm));
	}

	act = alloc_action();
	if (!act) {
		/* TODO: return an error to the client */
		dm_config_destroy(req.cft);
		buffer_destroy(&req.buffer);
		client_resume(cl);
		return;
	}

	act->client_id = cl->id;
	act->op = op;
	act->rt = rt;
	act->mode = mode;
	act->flags = opts;
	act->lm_type = lm;

	if (vg_name && strcmp(vg_name, "none"))
		strncpy(act->vg_name, vg_name, MAX_NAME);

	if (vg_uuid && strcmp(vg_uuid, "none"))
		strncpy(act->vg_uuid, vg_uuid, 64);

	if (vg_sysid && strcmp(vg_sysid, "none"))
		strncpy(act->vg_sysid, vg_sysid, MAX_NAME);

	str = daemon_request_str(req, "lv_name", NULL);
	if (str && strcmp(str, "none"))
		strncpy(act->lv_name, str, MAX_NAME);

	val = daemon_request_int(req, "version", 0);
	if (val)
		act->version = (uint32_t)val;

	str = daemon_request_str(req, "vg_lock_args", NULL);
	if (str && strcmp(str, "none"))
		strncpy(act->vg_args, str, MAX_ARGS);

	str = daemon_request_str(req, "lv_lock_args", NULL);
	if (str && strcmp(str, "none"))
		strncpy(act->lv_args, str, MAX_ARGS);

	/* start_vg will include lvmlocal.conf local/host_id here */
	val = daemon_request_int(req, "host_id", 0);
	if (val)
		act->host_id = val;

	act->max_retries = daemon_request_int(req, "max_retries", DEFAULT_MAX_RETRIES);

	dm_config_destroy(req.cft);
	buffer_destroy(&req.buffer);

	log_debug("recv %s[%d.%u] %s %s \"%s\" mode %s flags %x",
		  cl->name[0] ? cl->name : "client", cl->pid, cl->id,
		  op_str(act->op), rt_str(act->rt), act->vg_name, mode_str(act->mode), opts);

	/*
	 * local lock on local vg (lls) is done when local locking is enabled.
	 * local lock on gl is done when local locking is enabled and lockd is not.
	 */
	if ((local_thread_also && lls) ||
	    (local_thread_only && rt == LD_RT_GL && op == LD_OP_LOCK)) {
		add_local_lock_action(lls, act);
		return;
	}

	switch (act->op) {
	case LD_OP_START:
		rv = add_lockspace(act);
		break;
	case LD_OP_STOP:
		rv = rem_lockspace(act);
		break;
	case LD_OP_INIT:
	case LD_OP_UPDATE_LOCAL:
	case LD_OP_START_WAIT:
	case LD_OP_STOP_ALL:
	case LD_OP_RENAME_FINAL:
		add_work_action(act);
		rv = 0;
		break;
	case LD_OP_LOCK:
	case LD_OP_UPDATE:
	case LD_OP_ENABLE:
	case LD_OP_DISABLE:
	case LD_OP_FREE:
	case LD_OP_RENAME_BEFORE:
		rv = add_lock_action(act);
		break;
	case LD_OP_ADD_LOCAL:
		pthread_mutex_lock(&local_thread_mutex);
		add_local_vg(act->vg_name, act->vg_uuid, act->vg_sysid);
		pthread_mutex_unlock(&local_thread_mutex);
		act->result = 0;
		add_client_result(act);
		rv = 0;
		break;
	case LD_OP_REM_LOCAL:
		pthread_mutex_lock(&local_thread_mutex);
		rem_local_vg(act->vg_name, act->vg_uuid);
		pthread_mutex_unlock(&local_thread_mutex);
		act->result = 0;
		add_client_result(act);
		rv = 0;
		break;
	default:
		rv = -EINVAL;
	};

	if (rv < 0) {
		act->result = rv;
		add_client_result(act);
	}
}

static void *client_thread_main(void *arg_in)
{
	struct client *cl;
	struct action *act;

	while (1) {
		pthread_mutex_lock(&client_mutex);
		while (!client_work && list_empty(&client_results)) {
			if (client_stop) {
				pthread_mutex_unlock(&client_mutex);
				goto out;
			}
			pthread_cond_wait(&client_cond, &client_mutex);
		}

		/*
		 * Send outgoing results back to clients
		 */

		if (!list_empty(&client_results)) {
			act = list_first_entry(&client_results, struct action, list);
			list_del(&act->list);
			cl = find_client_id(act->client_id);
			pthread_mutex_unlock(&client_mutex);

			if (cl) {
				pthread_mutex_lock(&cl->mutex);
				client_send_result(cl, act);
				pthread_mutex_unlock(&cl->mutex);
			} else {
				log_debug("no client for result");
			}
			free_action(act);
			continue;
		}

		/*
		 * Queue incoming actions for lockspace threads
		 */

		if (client_work) {
			cl = find_client_work();
			if (!cl)
				client_work = 0;
			pthread_mutex_unlock(&client_mutex);

			if (!cl)
				continue;

			pthread_mutex_lock(&cl->mutex);

			if (cl->recv) {
				cl->recv = 0;
				client_recv_action(cl);
			}

			if (cl->dead) {
				/*
				log_debug("client rem %d pi %d fd %d ig %d",
					  cl->id, cl->pi, cl->fd, cl->poll_ignore);
				*/
				/*
				 * If cl->dead was set in main_loop, then the
				 * fd has already been closed and the pollfd
				 * entry is already unused.
				 * main_loop set dead=1, ignore=0, pi=-1, fd=-1
				 *
				 * if cl->dead was not set in main_loop, but
				 * set in client_recv_action, then the main_loop
				 * should be ignoring this client fd.
				 * main_loop set ignore=1
				 */

				if (cl->poll_ignore) {
					log_debug("client close %d pi %d fd %d",
						  cl->id, cl->pi, cl->fd);
					/* assert cl->pi != -1 */
					/* assert pollfd[pi].fd == FD_IGNORE */
					close(cl->fd);
					rem_pollfd(cl->pi);
					cl->pi = -1;
					cl->fd = -1;
					cl->poll_ignore = 0;
				} else {
					/* main thread should have closed */
					if (cl->pi != -1 || cl->fd != -1) {
						log_error("client %d bad state pi %d fd %d",
							  cl->id, cl->pi, cl->fd);
					}
				}
				pthread_mutex_unlock(&cl->mutex);

				pthread_mutex_lock(&client_mutex);
				list_del(&cl->list);
				pthread_mutex_unlock(&client_mutex);

				client_purge(cl);

				free_client(cl);
			} else {
				pthread_mutex_unlock(&cl->mutex);
			}
		}
		pthread_mutex_unlock(&client_mutex);
	}
out:
	return NULL;
}

static int setup_client_thread(void)
{
	int rv;

	INIT_LIST_HEAD(&client_list);
	INIT_LIST_HEAD(&client_results);

	pthread_mutex_init(&client_mutex, NULL);
	pthread_cond_init(&client_cond, NULL);

	rv = pthread_create(&client_thread, NULL, client_thread_main, NULL);
	if (rv)
		return -1;
	return 0;
}

static void close_client_thread(void)
{
	pthread_mutex_lock(&client_mutex);
	client_stop = 1;
	pthread_cond_signal(&client_cond);
	pthread_mutex_unlock(&client_mutex);
	pthread_join(client_thread, NULL);
}

static int setup_local_thread(void)
{
	struct lockspace *ls;
	struct resource *r;
	int rv;

	if (!local_thread_also)
		return 0;

	if (local_thread_only) {
		ls = malloc(sizeof(struct lockspace));
		if (!ls)
			return -ENOMEM;

		r = alloc_resource();
		if (!r) {
			free(ls);
			return -ENOMEM;
		}

		memset(ls, 0, sizeof(struct lockspace));
		strcpy(ls->name, "local_thread_gls");
		INIT_LIST_HEAD(&ls->actions);
		INIT_LIST_HEAD(&ls->resources);

		r->type = LD_RT_GL;
		r->mode = LD_LK_UN;
		strncpy(r->name, R_NAME_GL, MAX_NAME);
		INIT_LIST_HEAD(&r->locks);
		INIT_LIST_HEAD(&r->actions);
		list_add_tail(&r->list, &ls->resources);

		list_add(&ls->list, &local_vgs);
		local_thread_gls = ls;
	}

	rv = pthread_create(&local_thread, NULL, local_thread_main, NULL);
	if (rv)
		return -1;

	return 0;
}

static void close_local_thread(void)
{
	if (!local_thread_also)
		return;

	pthread_mutex_lock(&local_thread_mutex);
	local_thread_stop = 1;
	pthread_cond_signal(&local_thread_cond);
	pthread_mutex_unlock(&local_thread_mutex);
	pthread_join(local_thread, NULL);
}

#if 0
static void setup_listener(void)
{
	struct sockaddr_un addr;
	int rv, fd, ci;

	rv = lvmlockd_socket_address(&addr);
	if (rv < 0)
		return rv;

	fd = socket(AF_LOCAL, SOCK_STREAM, 0);
	if (fd < 0)
		return fd;

	unlink(addr.sun_path);
	rv = bind(fd, (struct sockaddr *) &addr, sizeof(struct sockaddr_un));
	if (rv < 0)
		goto exit_fail;

	rv = chmod(addr.sun_path, DEFAULT_SOCKET_MODE);
	if (rv < 0)
		goto exit_fail;

	rv = chown(addr.sun_path, com.uid, com.gid);
	if (rv < 0) {
		log_error("could not set socket %s permissions: %s",
			  addr.sun_path, strerror(errno));
		goto exit_fail;
	}

	rv = listen(fd, 5);
	if (rv < 0)
		goto exit_fail;

	fcntl(fd, F_SETFL, fcntl(fd, F_GETFL, 0) | O_NONBLOCK);

	listen_pi = add_pollfd(fd);

	return 0;

exit_fail:
	close(fd);
	return -1;
}
#endif

static int get_peer_pid(int fd)
{
	struct ucred cred;
	unsigned int len = sizeof(cred);

	if (getsockopt(fd, SOL_SOCKET, SO_PEERCRED, &cred, &len) != 0)
		return -1;

	return cred.pid;
}

static void process_listener(int poll_fd)
{
	struct client *cl;
	int fd, pi;

	/* assert poll_fd == listen_fd */

	fd = accept(listen_fd, NULL, NULL);
	if (fd < 0)
		return;

	cl = alloc_client();
	if (!cl)
		return;
	memset(cl, 0, sizeof(struct client));

	pi = add_pollfd(fd);
	if (pi < 0) {
		log_error("process_listener add_pollfd error %d", pi);
		free_client(cl);
		return;
	}

	cl->pi = pi;
	cl->fd = fd;
	cl->pid = get_peer_pid(fd);

	pthread_mutex_init(&cl->mutex, NULL);

	pthread_mutex_lock(&client_mutex);
	cl->id = ++client_ids;
	list_add_tail(&cl->list, &client_list);
	pthread_mutex_unlock(&client_mutex);

	/* log_debug("client add %d pi %d fd %d", cl->id, cl->pi, cl->fd); */
}

/*
 * main loop polls on pipe[0] so that a thread can
 * restart the poll by writing to pipe[1].
 */
static int setup_restart(void)
{
	if (pipe(restart_fds)) {
		log_error("setup_restart pipe error %d", errno);
		return -1;
	}

	restart_pi = add_pollfd(restart_fds[0]);
	if (restart_pi < 0)
		return restart_pi;

	return 0;
}

/*
 * thread wrote 'w' to restart_fds[1] to restart poll()
 * after adding an fd back into pollfd.
 */
static void process_restart(int fd)
{
	char wake[1];
	/* assert fd == restart_fds[0] */
	read(restart_fds[0], wake, 1);
}

static void sigterm_handler(int sig __attribute__((unused)))
{
	daemon_quit = 1;
}

static int main_loop(daemon_state *ds_arg)
{
	struct client *cl;
	int i, rv, is_recv, is_dead;

	signal(SIGTERM, &sigterm_handler);

	setup_structs();

	/* TODO: avoid possible vg name collision */
	strcpy(gl_lsname_dlm, S_NAME_GL_DLM);

	INIT_LIST_HEAD(&local_vgs);
	INIT_LIST_HEAD(&local_thread_actions);
	pthread_mutex_init(&local_thread_mutex, NULL);
	pthread_cond_init(&local_thread_cond, NULL);
	INIT_LIST_HEAD(&lockspaces);
	INIT_LIST_HEAD(&lockspaces_inactive);
	pthread_mutex_init(&lockspaces_mutex, NULL);
	pthread_mutex_init(&pollfd_mutex, NULL);
	pthread_mutex_init(&log_mutex, NULL);

	openlog("lvmlockd", LOG_CONS | LOG_PID, LOG_DAEMON);
	log_warn("lvmlockd started");

	listen_fd = ds_arg->socket_fd;
	listen_pi = add_pollfd(listen_fd);

	setup_client_thread();
	setup_worker_thread();
	setup_local_thread();
	setup_restart();

	pthread_mutex_init(&lvmetad_mutex, NULL);
	lvmetad_handle = lvmetad_open(NULL);
	if (lvmetad_handle.error || lvmetad_handle.socket_fd < 0)
		log_error("lvmetad_open error %d", lvmetad_handle.error);
	else
		lvmetad_connected = 1;

	/* add entries to local_vgs */
	create_work_action(LD_OP_UPDATE_LOCAL);

	while (1) {
		rv = poll(pollfd, pollfd_maxi + 1, -1);
		if (rv == -1 && errno == EINTR) {
			if (daemon_quit) {
				int count;
				/* first sigterm would trigger stops, and
				   second sigterm may finish the joins. */
				count = for_each_lockspace(DO_STOP, DO_FREE, NO_FORCE);
				if (!count)
					break;
				log_debug("ignore shutdown for %d lockspaces", count);
				daemon_quit = 0;
			}
			continue;
		}
		if (rv < 0) {
			log_error("poll errno %d", errno);
			break;
		}

		for (i = 0; i <= pollfd_maxi; i++) {
			if (pollfd[i].fd < 0)
				continue;

			is_recv = 0;
			is_dead = 0;

			if (pollfd[i].revents & POLLIN)
				is_recv = 1;
			if (pollfd[i].revents & (POLLERR | POLLHUP | POLLNVAL))
				is_dead = 1;

			if (!is_recv && !is_dead)
				continue;

			if (i == listen_pi) {
				process_listener(pollfd[i].fd);
				continue;
			}

			if (i == restart_pi) {
				process_restart(pollfd[i].fd);
				continue;
			}

			/*
			log_debug("poll pi %d fd %d revents %x",
				  i, pollfd[i].fd, pollfd[i].revents);
			*/

			pthread_mutex_lock(&client_mutex);
			cl = find_client_pi(i);
			if (cl) {
				pthread_mutex_lock(&cl->mutex);

				if (cl->recv) {
					/* should not happen */
					log_error("main client %d already recv", cl->id);

				} else if (cl->dead) {
					/* should not happen */
					log_error("main client %d already dead", cl->id);

				} else if (is_dead) {
					log_debug("close %s[%d.%u] fd %d",
						  cl->name[0] ? cl->name : "client",
						  cl->pid, cl->id, cl->fd);
					cl->dead = 1;
					cl->pi = -1;
					cl->fd = -1;
					cl->poll_ignore = 0;
					close(pollfd[i].fd);
					pollfd[i].fd = POLL_FD_UNUSED;
					pollfd[i].events = 0;
					pollfd[i].revents = 0;

				} else if (is_recv) {
					cl->recv = 1;
					cl->poll_ignore = 1;
					pollfd[i].fd = POLL_FD_IGNORE;
					pollfd[i].events = 0;
					pollfd[i].revents = 0;
				}

				pthread_mutex_unlock(&cl->mutex);

				client_work = 1;
				pthread_cond_signal(&client_cond);

				/* client_thread will pick up and work on any
				   client with cl->recv or cl->dead set */

			} else {
				/* don't think this can happen */
				log_error("no client for index %d fd %d",
					  i, pollfd[i].fd);
				close(pollfd[i].fd);
				pollfd[i].fd = POLL_FD_UNUSED;
				pollfd[i].events = 0;
				pollfd[i].revents = 0;
			}
			pthread_mutex_unlock(&client_mutex);

			/* TODO?: after set_dead, scan pollfd for last unused
			   slot and reduce pollfd_maxi */
		}
	}

	for_each_lockspace_retry(DO_STOP, DO_FREE, DO_FORCE);
	free_lockspaces_inactive();
	close_worker_thread();
	close_client_thread();
	close_local_thread();
	closelog();
	daemon_close(lvmetad_handle);
	return 0;
}

static void usage(char *prog, FILE *file)
{
	fprintf(file, "Usage:\n");
	fprintf(file, "%s [options]\n\n", prog);
	fprintf(file, "  --help | -h\n");
	fprintf(file, "        Show this help information.\n");
	fprintf(file, "  --version | -V\n");
	fprintf(file, "        Show version of lvmlockd.\n");
	fprintf(file, "  --test | -T\n");
	fprintf(file, "        Test mode, do not call lock manager.\n");
	fprintf(file, "  --foreground | -f\n");
	fprintf(file, "        Don't fork.\n");
	fprintf(file, "  --daemon-debug | -D\n");
	fprintf(file, "        Don't fork and print debugging to stdout.\n");
	fprintf(file, "  --pid-file | -p <path>\n");
	fprintf(file, "        Set path to the pid file. [%s]\n", LVMLOCKD_PIDFILE);
	fprintf(file, "  --socket-path | -s <path>\n");
	fprintf(file, "        Set path to the socket to listen on. [%s]\n", LVMLOCKD_SOCKET);
	fprintf(file, "  --log-config | -l <str>\n");
	fprintf(file, "        Set log config.\n");
	fprintf(file, "  --local-also | -a\n");
	fprintf(file, "        Manage locks between pids for local vgs.\n");
	fprintf(file, "  --local-only | -o\n");
	fprintf(file, "        Only manage locks for local vgs, not dlm|sanlock vgs.\n");
	fprintf(file, "  --gl-type | -g <str>\n");
	fprintf(file, "        Set global lock type to be dlm|sanlock.\n");
	fprintf(file, "  --system-id | -y <str>\n");
	fprintf(file, "        Set the local system id.\n");
	fprintf(file, "  --host-id | -i <num>\n");
	fprintf(file, "        Set the local sanlock host id.\n");
	fprintf(file, "  --host-id-file | -F <path>\n");
	fprintf(file, "        A file containing the local sanlock host_id.\n");
}

int main(int argc, char *argv[])
{
	daemon_state ds;

	ds.daemon_main = main_loop;
	ds.daemon_init = NULL;
	ds.daemon_fini = NULL;
	ds.pidfile = getenv("LVM_LVMLOCKD_PIDFILE");
	ds.socket_path = getenv("LVM_LVMLOCKD_SOCKET");
	ds.protocol = lvmlockd_protocol;
	ds.protocol_version = lvmlockd_protocol_version;
	ds.name = "lvmlockd";

	static struct option long_options[] = {
		{"help",        no_argument,       0, 'h' },
		{"version",     no_argument,       0, 'V' },
		{"test",        no_argument,       0, 'T' },
		{"foreground",  no_argument,       0, 'f' },
		{"daemon-debug",no_argument,       0, 'D' },
		{"pid-file",    required_argument, 0, 'p' },
		{"socket-path", required_argument, 0, 's' },
		{"local-also",  no_argument,       0, 'a' },
		{"local-only",  no_argument,       0, 'o' },
		{"gl-type",     required_argument, 0, 'g' },
		{"system-id",   required_argument, 0, 'y' },
		{"host-id",     required_argument, 0, 'i' },
		{"host-id-file",required_argument, 0, 'F' },
		{0, 0, 0, 0 }
	};

	while (1) {
		int c;
		int lm;
		int option_index = 0;

		c = getopt_long(argc, argv, "hVTfDp:s:l:aog:S:I:",
				long_options, &option_index);
		if (c == -1)
			break;

		switch (c) {
		case '0':
			break;
		case 'h':
			usage(argv[0], stdout);
			exit(EXIT_SUCCESS);
		case 'V':
			printf("lvmlockd version: " LVM_VERSION "\n");
			exit(EXIT_SUCCESS);
		case 'T':
			daemon_test = 1;
			break;
		case 'f':
			ds.foreground = 1;
			break;
		case 'D':
			ds.foreground = 1;
			daemon_debug = 1;
			break;
		case 'p':
			ds.pidfile = strdup(optarg);
			break;
		case 's':
			ds.socket_path = strdup(optarg);
			break;
		case 'a':
			local_thread_also = 1;
			break;
		case 'o':
			local_thread_also = 1;
			local_thread_only = 1;
			break;
		case 'g':
			lm = str_to_lm(optarg);
			if (lm == LD_LM_DLM)
				gl_use_dlm = 1;
			else if (lm == LD_LM_SANLOCK)
				gl_use_sanlock = 1;
			else {
				fprintf(stderr, "invalid gl-type option");
				exit(EXIT_FAILURE);
			}
			break;
		case 'y':
			our_system_id = strdup(optarg);
			break;
		case 'i':
			daemon_host_id = atoi(optarg);
			break;
		case 'F':
			daemon_host_id_file = strdup(optarg);
			break;
		case '?':
		default:
			usage(argv[0], stdout);
			exit(EXIT_FAILURE);
		}
	}

	if (!ds.pidfile)
		ds.pidfile = LVMLOCKD_PIDFILE;

	if (!ds.socket_path)
		ds.socket_path = LVMLOCKD_SOCKET;

	/* runs daemon_main/main_loop */
	daemon_start(ds);

	return 0;
}