summaryrefslogtreecommitdiffstats
path: root/lib/libzfs/libzfs_import.c
blob: f371d925bc59c4223089260bcb482fdc0700035b (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
/*
 * CDDL HEADER START
 *
 * The contents of this file are subject to the terms of the
 * Common Development and Distribution License (the "License").
 * You may not use this file except in compliance with the License.
 *
 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
 * or http://www.opensolaris.org/os/licensing.
 * See the License for the specific language governing permissions
 * and limitations under the License.
 *
 * When distributing Covered Code, include this CDDL HEADER in each
 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
 * If applicable, add the following below this CDDL HEADER, with the
 * fields enclosed by brackets "[]" replaced with your own identifying
 * information: Portions Copyright [yyyy] [name of copyright owner]
 *
 * CDDL HEADER END
 */
/*
 * Copyright 2015 Nexenta Systems, Inc. All rights reserved.
 * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
 * Copyright (c) 2012, 2016 by Delphix. All rights reserved.
 * Copyright 2015 RackTop Systems.
 * Copyright (c) 2016, Intel Corporation.
 */

/*
 * Pool import support functions.
 *
 * To import a pool, we rely on reading the configuration information from the
 * ZFS label of each device.  If we successfully read the label, then we
 * organize the configuration information in the following hierarchy:
 *
 * 	pool guid -> toplevel vdev guid -> label txg
 *
 * Duplicate entries matching this same tuple will be discarded.  Once we have
 * examined every device, we pick the best label txg config for each toplevel
 * vdev.  We then arrange these toplevel vdevs into a complete pool config, and
 * update any paths that have changed.  Finally, we attempt to import the pool
 * using our derived config, and record the results.
 */

#include <ctype.h>
#include <devid.h>
#include <dirent.h>
#include <errno.h>
#include <libintl.h>
#ifdef HAVE_LIBUDEV
#include <libudev.h>
#include <sched.h>
#endif
#include <stddef.h>
#include <stdlib.h>
#include <string.h>
#include <sys/stat.h>
#include <unistd.h>
#include <fcntl.h>
#include <sys/vtoc.h>
#include <sys/dktp/fdisk.h>
#include <sys/efi_partition.h>
#include <sys/vdev_impl.h>
#include <blkid/blkid.h>
#include "libzfs.h"
#include "libzfs_impl.h"
#include <libzfs.h>

/*
 * Intermediate structures used to gather configuration information.
 */
typedef struct config_entry {
	uint64_t		ce_txg;
	nvlist_t		*ce_config;
	struct config_entry	*ce_next;
} config_entry_t;

typedef struct vdev_entry {
	uint64_t		ve_guid;
	config_entry_t		*ve_configs;
	struct vdev_entry	*ve_next;
} vdev_entry_t;

typedef struct pool_entry {
	uint64_t		pe_guid;
	vdev_entry_t		*pe_vdevs;
	struct pool_entry	*pe_next;
} pool_entry_t;

typedef struct name_entry {
	char			*ne_name;
	uint64_t		ne_guid;
	uint64_t		ne_order;
	uint64_t		ne_num_labels;
	struct name_entry	*ne_next;
} name_entry_t;

typedef struct pool_list {
	pool_entry_t		*pools;
	name_entry_t		*names;
} pool_list_t;

#define	DEV_BYID_PATH	"/dev/disk/by-id/"

/*
 * Linux persistent device strings for vdev labels
 *
 * based on libudev for consistency with libudev disk add/remove events
 */
#ifdef HAVE_LIBUDEV

typedef struct vdev_dev_strs {
	char	vds_devid[128];
	char	vds_devphys[128];
} vdev_dev_strs_t;

/*
 * Obtain the persistent device id string (describes what)
 *
 * used by ZED vdev matching for auto-{online,expand,replace}
 */
int
zfs_device_get_devid(struct udev_device *dev, char *bufptr, size_t buflen)
{
	struct udev_list_entry *entry;
	const char *bus;
	char devbyid[MAXPATHLEN];

	/* The bus based by-id path is preferred */
	bus = udev_device_get_property_value(dev, "ID_BUS");

	if (bus == NULL) {
		const char *dm_uuid;

		/*
		 * For multipath nodes use the persistent uuid based identifier
		 *
		 * Example: /dev/disk/by-id/dm-uuid-mpath-35000c5006304de3f
		 */
		dm_uuid = udev_device_get_property_value(dev, "DM_UUID");
		if (dm_uuid != NULL) {
			(void) snprintf(bufptr, buflen, "dm-uuid-%s", dm_uuid);
			return (0);
		}
		return (ENODATA);
	}

	/*
	 * locate the bus specific by-id link
	 */
	(void) snprintf(devbyid, sizeof (devbyid), "%s%s-", DEV_BYID_PATH, bus);
	entry = udev_device_get_devlinks_list_entry(dev);
	while (entry != NULL) {
		const char *name;

		name = udev_list_entry_get_name(entry);
		if (strncmp(name, devbyid, strlen(devbyid)) == 0) {
			name += strlen(DEV_BYID_PATH);
			(void) strlcpy(bufptr, name, buflen);
			return (0);
		}
		entry = udev_list_entry_get_next(entry);
	}

	return (ENODATA);
}

/*
 * Obtain the persistent physical location string (describes where)
 *
 * used by ZED vdev matching for auto-{online,expand,replace}
 */
int
zfs_device_get_physical(struct udev_device *dev, char *bufptr, size_t buflen)
{
	const char *physpath = NULL;

	/*
	 * Normal disks use ID_PATH for their physical path.  Device mapper
	 * devices are virtual and don't have a physical path.  For them we
	 * use ID_VDEV instead, which is setup via the /etc/vdev_id.conf file.
	 * ID_VDEV provides a persistent path to a virtual device.  If you
	 * don't have vdev_id.conf setup, you cannot use multipath autoreplace.
	 */
	if (!((physpath = udev_device_get_property_value(dev, "ID_PATH")) &&
	    physpath[0])) {
		if (!((physpath =
		    udev_device_get_property_value(dev, "ID_VDEV")) &&
		    physpath[0])) {
			return (ENODATA);
		}
	}

	(void) strlcpy(bufptr, physpath, buflen);

	return (0);
}

boolean_t
udev_is_mpath(struct udev_device *dev)
{
	return udev_device_get_property_value(dev, "DM_UUID") &&
	    udev_device_get_property_value(dev, "MPATH_SBIN_PATH");
}

/*
 * A disk is considered a multipath whole disk when:
 *	DEVNAME key value has "dm-"
 *	DM_NAME key value has "mpath" prefix
 *	DM_UUID key exists
 *	ID_PART_TABLE_TYPE key does not exist or is not gpt
 */
static boolean_t
udev_mpath_whole_disk(struct udev_device *dev)
{
	const char *devname, *type, *uuid;

	devname = udev_device_get_property_value(dev, "DEVNAME");
	type = udev_device_get_property_value(dev, "ID_PART_TABLE_TYPE");
	uuid = udev_device_get_property_value(dev, "DM_UUID");

	if ((devname != NULL && strncmp(devname, "/dev/dm-", 8) == 0) &&
	    ((type == NULL) || (strcmp(type, "gpt") != 0)) &&
	    (uuid != NULL)) {
		return (B_TRUE);
	}

	return (B_FALSE);
}

/*
 * Check if a disk is effectively a multipath whole disk
 */
boolean_t
is_mpath_whole_disk(const char *path)
{
	struct udev *udev;
	struct udev_device *dev = NULL;
	char nodepath[MAXPATHLEN];
	char *sysname;
	boolean_t wholedisk = B_FALSE;

	if (realpath(path, nodepath) == NULL)
		return (B_FALSE);
	sysname = strrchr(nodepath, '/') + 1;
	if (strncmp(sysname, "dm-", 3) != 0)
		return (B_FALSE);
	if ((udev = udev_new()) == NULL)
		return (B_FALSE);
	if ((dev = udev_device_new_from_subsystem_sysname(udev, "block",
	    sysname)) == NULL) {
		udev_device_unref(dev);
		return (B_FALSE);
	}

	wholedisk = udev_mpath_whole_disk(dev);

	udev_device_unref(dev);
	return (wholedisk);
}

static int
udev_device_is_ready(struct udev_device *dev)
{
#ifdef HAVE_LIBUDEV_UDEV_DEVICE_GET_IS_INITIALIZED
	return (udev_device_get_is_initialized(dev));
#else
	/* wait for DEVLINKS property to be initialized */
	return (udev_device_get_property_value(dev, "DEVLINKS") != NULL);
#endif
}

/*
 * Wait up to timeout_ms for udev to set up the device node.  The device is
 * considered ready when libudev determines it has been initialized, all of
 * the device links have been verified to exist, and it has been allowed to
 * settle.  At this point the device the device can be accessed reliably.
 * Depending on the complexity of the udev rules this process could take
 * several seconds.
 */
int
zpool_label_disk_wait(char *path, int timeout_ms)
{
	struct udev *udev;
	struct udev_device *dev = NULL;
	char nodepath[MAXPATHLEN];
	char *sysname = NULL;
	int ret = ENODEV;
	int settle_ms = 50;
	long sleep_ms = 10;
	hrtime_t start, settle;

	if ((udev = udev_new()) == NULL)
		return (ENXIO);

	start = gethrtime();
	settle = 0;

	do {
		if (sysname == NULL) {
			if (realpath(path, nodepath) != NULL) {
				sysname = strrchr(nodepath, '/') + 1;
			} else {
				(void) usleep(sleep_ms * MILLISEC);
				continue;
			}
		}

		dev = udev_device_new_from_subsystem_sysname(udev,
		    "block", sysname);
		if ((dev != NULL) && udev_device_is_ready(dev)) {
			struct udev_list_entry *links, *link;

			ret = 0;
			links = udev_device_get_devlinks_list_entry(dev);

			udev_list_entry_foreach(link, links) {
				struct stat64 statbuf;
				const char *name;

				name = udev_list_entry_get_name(link);
				errno = 0;
				if (stat64(name, &statbuf) == 0 && errno == 0)
					continue;

				settle = 0;
				ret = ENODEV;
				break;
			}

			if (ret == 0) {
				if (settle == 0) {
					settle = gethrtime();
				} else if (NSEC2MSEC(gethrtime() - settle) >=
				    settle_ms) {
					udev_device_unref(dev);
					break;
				}
			}
		}

		udev_device_unref(dev);
		(void) usleep(sleep_ms * MILLISEC);

	} while (NSEC2MSEC(gethrtime() - start) < timeout_ms);

	udev_unref(udev);

	return (ret);
}


/*
 * Encode the persistent devices strings
 * used for the vdev disk label
 */
static int
encode_device_strings(const char *path, vdev_dev_strs_t *ds,
    boolean_t wholedisk)
{
	struct udev *udev;
	struct udev_device *dev = NULL;
	char nodepath[MAXPATHLEN];
	char *sysname;
	int ret = ENODEV;
	hrtime_t start;

	if ((udev = udev_new()) == NULL)
		return (ENXIO);

	/* resolve path to a runtime device node instance */
	if (realpath(path, nodepath) == NULL)
		goto no_dev;

	sysname = strrchr(nodepath, '/') + 1;

	/*
	 * Wait up to 3 seconds for udev to set up the device node context
	 */
	start = gethrtime();
	do {
		dev = udev_device_new_from_subsystem_sysname(udev, "block",
		    sysname);
		if (dev == NULL)
			goto no_dev;
		if (udev_device_is_ready(dev))
			break;  /* udev ready */

		udev_device_unref(dev);
		dev = NULL;

		if (NSEC2MSEC(gethrtime() - start) < 10)
			(void) sched_yield();	/* yield/busy wait up to 10ms */
		else
			(void) usleep(10 * MILLISEC);

	} while (NSEC2MSEC(gethrtime() - start) < (3 * MILLISEC));

	if (dev == NULL)
		goto no_dev;

	/*
	 * Only whole disks require extra device strings
	 */
	if (!wholedisk && !udev_mpath_whole_disk(dev))
		goto no_dev;

	ret = zfs_device_get_devid(dev, ds->vds_devid, sizeof (ds->vds_devid));
	if (ret != 0)
		goto no_dev_ref;

	/* physical location string (optional) */
	if (zfs_device_get_physical(dev, ds->vds_devphys,
	    sizeof (ds->vds_devphys)) != 0) {
		ds->vds_devphys[0] = '\0'; /* empty string --> not available */
	}

no_dev_ref:
	udev_device_unref(dev);
no_dev:
	udev_unref(udev);

	return (ret);
}

/*
 * Update a leaf vdev's persistent device strings (Linux only)
 *
 * - only applies for a dedicated leaf vdev (aka whole disk)
 * - updated during pool create|add|attach|import
 * - used for matching device matching during auto-{online,expand,replace}
 * - stored in a leaf disk config label (i.e. alongside 'path' NVP)
 * - these strings are currently not used in kernel (i.e. for vdev_disk_open)
 *
 * single device node example:
 * 	devid:		'scsi-MG03SCA300_350000494a8cb3d67-part1'
 * 	phys_path:	'pci-0000:04:00.0-sas-0x50000394a8cb3d67-lun-0'
 *
 * multipath device node example:
 * 	devid:		'dm-uuid-mpath-35000c5006304de3f'
 *
 * We also store the enclosure sysfs path for turning on enclosure LEDs
 * (if applicable):
 *	vdev_enc_sysfs_path: '/sys/class/enclosure/11:0:1:0/SLOT 4'
 */
void
update_vdev_config_dev_strs(nvlist_t *nv)
{
	vdev_dev_strs_t vds;
	char *env, *type, *path;
	uint64_t wholedisk = 0;
	char *upath, *spath;

	/*
	 * For the benefit of legacy ZFS implementations, allow
	 * for opting out of devid strings in the vdev label.
	 *
	 * example use:
	 *	env ZFS_VDEV_DEVID_OPT_OUT=YES zpool import dozer
	 *
	 * explanation:
	 * Older ZFS on Linux implementations had issues when attempting to
	 * display pool config VDEV names if a "devid" NVP value is present
	 * in the pool's config.
	 *
	 * For example, a pool that originated on illumos platform would
	 * have a devid value in the config and "zpool status" would fail
	 * when listing the config.
	 *
	 * A pool can be stripped of any "devid" values on import or
	 * prevented from adding them on zpool create|add by setting
	 * ZFS_VDEV_DEVID_OPT_OUT.
	 */
	env = getenv("ZFS_VDEV_DEVID_OPT_OUT");
	if (env && (strtoul(env, NULL, 0) > 0 ||
	    !strncasecmp(env, "YES", 3) || !strncasecmp(env, "ON", 2))) {
		(void) nvlist_remove_all(nv, ZPOOL_CONFIG_DEVID);
		(void) nvlist_remove_all(nv, ZPOOL_CONFIG_PHYS_PATH);
		return;
	}

	if (nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) != 0 ||
	    strcmp(type, VDEV_TYPE_DISK) != 0) {
		return;
	}
	if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) != 0)
		return;
	(void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK, &wholedisk);

	/*
	 * Update device string values in config nvlist
	 */
	if (encode_device_strings(path, &vds, (boolean_t)wholedisk) == 0) {
		(void) nvlist_add_string(nv, ZPOOL_CONFIG_DEVID, vds.vds_devid);
		if (vds.vds_devphys[0] != '\0') {
			(void) nvlist_add_string(nv, ZPOOL_CONFIG_PHYS_PATH,
			    vds.vds_devphys);
		}

		/* Add enclosure sysfs path (if disk is in an enclosure) */
		upath = zfs_get_underlying_path(path);
		spath = zfs_get_enclosure_sysfs_path(upath);
		if (spath)
			nvlist_add_string(nv, ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH,
			    spath);
		else
			nvlist_remove_all(nv, ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH);

		free(upath);
		free(spath);
	} else {
		/* clear out any stale entries */
		(void) nvlist_remove_all(nv, ZPOOL_CONFIG_DEVID);
		(void) nvlist_remove_all(nv, ZPOOL_CONFIG_PHYS_PATH);
		(void) nvlist_remove_all(nv, ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH);
	}
}
#else

boolean_t
is_mpath_whole_disk(const char *path)
{
	return (B_FALSE);
}

/*
 * Wait up to timeout_ms for udev to set up the device node.  The device is
 * considered ready when the provided path have been verified to exist and
 * it has been allowed to settle.  At this point the device the device can
 * be accessed reliably.  Depending on the complexity of the udev rules thisi
 * process could take several seconds.
 */
int
zpool_label_disk_wait(char *path, int timeout_ms)
{
	int settle_ms = 50;
	long sleep_ms = 10;
	hrtime_t start, settle;
	struct stat64 statbuf;

	start = gethrtime();
	settle = 0;

	do {
		errno = 0;
		if ((stat64(path, &statbuf) == 0) && (errno == 0)) {
			if (settle == 0)
				settle = gethrtime();
			else if (NSEC2MSEC(gethrtime() - settle) >= settle_ms)
				return (0);
		} else if (errno != ENOENT) {
			return (errno);
		}

		usleep(sleep_ms * MILLISEC);
	} while (NSEC2MSEC(gethrtime() - start) < timeout_ms);

	return (ENODEV);
}

void
update_vdev_config_dev_strs(nvlist_t *nv)
{
}

#endif /* HAVE_LIBUDEV */

/*
 * Go through and fix up any path and/or devid information for the given vdev
 * configuration.
 */
static int
fix_paths(nvlist_t *nv, name_entry_t *names)
{
	nvlist_t **child;
	uint_t c, children;
	uint64_t guid;
	name_entry_t *ne, *best;
	char *path;

	if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
	    &child, &children) == 0) {
		for (c = 0; c < children; c++)
			if (fix_paths(child[c], names) != 0)
				return (-1);
		return (0);
	}

	/*
	 * This is a leaf (file or disk) vdev.  In either case, go through
	 * the name list and see if we find a matching guid.  If so, replace
	 * the path and see if we can calculate a new devid.
	 *
	 * There may be multiple names associated with a particular guid, in
	 * which case we have overlapping partitions or multiple paths to the
	 * same disk.  In this case we prefer to use the path name which
	 * matches the ZPOOL_CONFIG_PATH.  If no matching entry is found we
	 * use the lowest order device which corresponds to the first match
	 * while traversing the ZPOOL_IMPORT_PATH search path.
	 */
	verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) == 0);
	if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) != 0)
		path = NULL;

	best = NULL;
	for (ne = names; ne != NULL; ne = ne->ne_next) {
		if (ne->ne_guid == guid) {
			if (path == NULL) {
				best = ne;
				break;
			}

			if ((strlen(path) == strlen(ne->ne_name)) &&
			    strncmp(path, ne->ne_name, strlen(path)) == 0) {
				best = ne;
				break;
			}

			if (best == NULL) {
				best = ne;
				continue;
			}

			/* Prefer paths with move vdev labels. */
			if (ne->ne_num_labels > best->ne_num_labels) {
				best = ne;
				continue;
			}

			/* Prefer paths earlier in the search order. */
			if (ne->ne_num_labels == best->ne_num_labels &&
			    ne->ne_order < best->ne_order) {
				best = ne;
				continue;
			}
		}
	}

	if (best == NULL)
		return (0);

	if (nvlist_add_string(nv, ZPOOL_CONFIG_PATH, best->ne_name) != 0)
		return (-1);

	/* Linux only - update ZPOOL_CONFIG_DEVID and ZPOOL_CONFIG_PHYS_PATH */
	update_vdev_config_dev_strs(nv);

	return (0);
}

/*
 * Add the given configuration to the list of known devices.
 */
static int
add_config(libzfs_handle_t *hdl, pool_list_t *pl, const char *path,
    int order, int num_labels, nvlist_t *config)
{
	uint64_t pool_guid, vdev_guid, top_guid, txg, state;
	pool_entry_t *pe;
	vdev_entry_t *ve;
	config_entry_t *ce;
	name_entry_t *ne;

	/*
	 * If this is a hot spare not currently in use or level 2 cache
	 * device, add it to the list of names to translate, but don't do
	 * anything else.
	 */
	if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE,
	    &state) == 0 &&
	    (state == POOL_STATE_SPARE || state == POOL_STATE_L2CACHE) &&
	    nvlist_lookup_uint64(config, ZPOOL_CONFIG_GUID, &vdev_guid) == 0) {
		if ((ne = zfs_alloc(hdl, sizeof (name_entry_t))) == NULL) {
			nvlist_free(config);
			return (-1);
		}

		if ((ne->ne_name = zfs_strdup(hdl, path)) == NULL) {
			free(ne);
			nvlist_free(config);
			return (-1);
		}
		ne->ne_guid = vdev_guid;
		ne->ne_order = order;
		ne->ne_num_labels = num_labels;
		ne->ne_next = pl->names;
		pl->names = ne;
		nvlist_free(config);
		return (0);
	}

	/*
	 * If we have a valid config but cannot read any of these fields, then
	 * it means we have a half-initialized label.  In vdev_label_init()
	 * we write a label with txg == 0 so that we can identify the device
	 * in case the user refers to the same disk later on.  If we fail to
	 * create the pool, we'll be left with a label in this state
	 * which should not be considered part of a valid pool.
	 */
	if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID,
	    &pool_guid) != 0 ||
	    nvlist_lookup_uint64(config, ZPOOL_CONFIG_GUID,
	    &vdev_guid) != 0 ||
	    nvlist_lookup_uint64(config, ZPOOL_CONFIG_TOP_GUID,
	    &top_guid) != 0 ||
	    nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_TXG,
	    &txg) != 0 || txg == 0) {
		nvlist_free(config);
		return (0);
	}

	/*
	 * First, see if we know about this pool.  If not, then add it to the
	 * list of known pools.
	 */
	for (pe = pl->pools; pe != NULL; pe = pe->pe_next) {
		if (pe->pe_guid == pool_guid)
			break;
	}

	if (pe == NULL) {
		if ((pe = zfs_alloc(hdl, sizeof (pool_entry_t))) == NULL) {
			nvlist_free(config);
			return (-1);
		}
		pe->pe_guid = pool_guid;
		pe->pe_next = pl->pools;
		pl->pools = pe;
	}

	/*
	 * Second, see if we know about this toplevel vdev.  Add it if its
	 * missing.
	 */
	for (ve = pe->pe_vdevs; ve != NULL; ve = ve->ve_next) {
		if (ve->ve_guid == top_guid)
			break;
	}

	if (ve == NULL) {
		if ((ve = zfs_alloc(hdl, sizeof (vdev_entry_t))) == NULL) {
			nvlist_free(config);
			return (-1);
		}
		ve->ve_guid = top_guid;
		ve->ve_next = pe->pe_vdevs;
		pe->pe_vdevs = ve;
	}

	/*
	 * Third, see if we have a config with a matching transaction group.  If
	 * so, then we do nothing.  Otherwise, add it to the list of known
	 * configs.
	 */
	for (ce = ve->ve_configs; ce != NULL; ce = ce->ce_next) {
		if (ce->ce_txg == txg)
			break;
	}

	if (ce == NULL) {
		if ((ce = zfs_alloc(hdl, sizeof (config_entry_t))) == NULL) {
			nvlist_free(config);
			return (-1);
		}
		ce->ce_txg = txg;
		ce->ce_config = config;
		ce->ce_next = ve->ve_configs;
		ve->ve_configs = ce;
	} else {
		nvlist_free(config);
	}

	/*
	 * At this point we've successfully added our config to the list of
	 * known configs.  The last thing to do is add the vdev guid -> path
	 * mappings so that we can fix up the configuration as necessary before
	 * doing the import.
	 */
	if ((ne = zfs_alloc(hdl, sizeof (name_entry_t))) == NULL)
		return (-1);

	if ((ne->ne_name = zfs_strdup(hdl, path)) == NULL) {
		free(ne);
		return (-1);
	}

	ne->ne_guid = vdev_guid;
	ne->ne_order = order;
	ne->ne_num_labels = num_labels;
	ne->ne_next = pl->names;
	pl->names = ne;

	return (0);
}

/*
 * Returns true if the named pool matches the given GUID.
 */
static int
pool_active(libzfs_handle_t *hdl, const char *name, uint64_t guid,
    boolean_t *isactive)
{
	zpool_handle_t *zhp;
	uint64_t theguid;

	if (zpool_open_silent(hdl, name, &zhp) != 0)
		return (-1);

	if (zhp == NULL) {
		*isactive = B_FALSE;
		return (0);
	}

	verify(nvlist_lookup_uint64(zhp->zpool_config, ZPOOL_CONFIG_POOL_GUID,
	    &theguid) == 0);

	zpool_close(zhp);

	*isactive = (theguid == guid);
	return (0);
}

static nvlist_t *
refresh_config(libzfs_handle_t *hdl, nvlist_t *config)
{
	nvlist_t *nvl;
	zfs_cmd_t zc = {"\0"};
	int err, dstbuf_size;

	if (zcmd_write_conf_nvlist(hdl, &zc, config) != 0)
		return (NULL);

	dstbuf_size = MAX(CONFIG_BUF_MINSIZE, zc.zc_nvlist_conf_size * 4);

	if (zcmd_alloc_dst_nvlist(hdl, &zc, dstbuf_size) != 0) {
		zcmd_free_nvlists(&zc);
		return (NULL);
	}

	while ((err = ioctl(hdl->libzfs_fd, ZFS_IOC_POOL_TRYIMPORT,
	    &zc)) != 0 && errno == ENOMEM) {
		if (zcmd_expand_dst_nvlist(hdl, &zc) != 0) {
			zcmd_free_nvlists(&zc);
			return (NULL);
		}
	}

	if (err) {
		zcmd_free_nvlists(&zc);
		return (NULL);
	}

	if (zcmd_read_dst_nvlist(hdl, &zc, &nvl) != 0) {
		zcmd_free_nvlists(&zc);
		return (NULL);
	}

	zcmd_free_nvlists(&zc);
	return (nvl);
}

/*
 * Determine if the vdev id is a hole in the namespace.
 */
boolean_t
vdev_is_hole(uint64_t *hole_array, uint_t holes, uint_t id)
{
	int c;

	for (c = 0; c < holes; c++) {

		/* Top-level is a hole */
		if (hole_array[c] == id)
			return (B_TRUE);
	}
	return (B_FALSE);
}

/*
 * Convert our list of pools into the definitive set of configurations.  We
 * start by picking the best config for each toplevel vdev.  Once that's done,
 * we assemble the toplevel vdevs into a full config for the pool.  We make a
 * pass to fix up any incorrect paths, and then add it to the main list to
 * return to the user.
 */
static nvlist_t *
get_configs(libzfs_handle_t *hdl, pool_list_t *pl, boolean_t active_ok)
{
	pool_entry_t *pe;
	vdev_entry_t *ve;
	config_entry_t *ce;
	nvlist_t *ret = NULL, *config = NULL, *tmp = NULL, *nvtop, *nvroot;
	nvlist_t **spares, **l2cache;
	uint_t i, nspares, nl2cache;
	boolean_t config_seen;
	uint64_t best_txg;
	char *name, *hostname = NULL;
	uint64_t guid;
	uint_t children = 0;
	nvlist_t **child = NULL;
	uint_t holes;
	uint64_t *hole_array, max_id;
	uint_t c;
	boolean_t isactive;
	uint64_t hostid;
	nvlist_t *nvl;
	boolean_t valid_top_config = B_FALSE;

	if (nvlist_alloc(&ret, 0, 0) != 0)
		goto nomem;

	for (pe = pl->pools; pe != NULL; pe = pe->pe_next) {
		uint64_t id, max_txg = 0;

		if (nvlist_alloc(&config, NV_UNIQUE_NAME, 0) != 0)
			goto nomem;
		config_seen = B_FALSE;

		/*
		 * Iterate over all toplevel vdevs.  Grab the pool configuration
		 * from the first one we find, and then go through the rest and
		 * add them as necessary to the 'vdevs' member of the config.
		 */
		for (ve = pe->pe_vdevs; ve != NULL; ve = ve->ve_next) {

			/*
			 * Determine the best configuration for this vdev by
			 * selecting the config with the latest transaction
			 * group.
			 */
			best_txg = 0;
			for (ce = ve->ve_configs; ce != NULL;
			    ce = ce->ce_next) {

				if (ce->ce_txg > best_txg) {
					tmp = ce->ce_config;
					best_txg = ce->ce_txg;
				}
			}

			/*
			 * We rely on the fact that the max txg for the
			 * pool will contain the most up-to-date information
			 * about the valid top-levels in the vdev namespace.
			 */
			if (best_txg > max_txg) {
				(void) nvlist_remove(config,
				    ZPOOL_CONFIG_VDEV_CHILDREN,
				    DATA_TYPE_UINT64);
				(void) nvlist_remove(config,
				    ZPOOL_CONFIG_HOLE_ARRAY,
				    DATA_TYPE_UINT64_ARRAY);

				max_txg = best_txg;
				hole_array = NULL;
				holes = 0;
				max_id = 0;
				valid_top_config = B_FALSE;

				if (nvlist_lookup_uint64(tmp,
				    ZPOOL_CONFIG_VDEV_CHILDREN, &max_id) == 0) {
					verify(nvlist_add_uint64(config,
					    ZPOOL_CONFIG_VDEV_CHILDREN,
					    max_id) == 0);
					valid_top_config = B_TRUE;
				}

				if (nvlist_lookup_uint64_array(tmp,
				    ZPOOL_CONFIG_HOLE_ARRAY, &hole_array,
				    &holes) == 0) {
					verify(nvlist_add_uint64_array(config,
					    ZPOOL_CONFIG_HOLE_ARRAY,
					    hole_array, holes) == 0);
				}
			}

			if (!config_seen) {
				/*
				 * Copy the relevant pieces of data to the pool
				 * configuration:
				 *
				 *	version
				 *	pool guid
				 *	name
				 *	comment (if available)
				 *	pool state
				 *	hostid (if available)
				 *	hostname (if available)
				 */
				uint64_t state, version;
				char *comment = NULL;

				version = fnvlist_lookup_uint64(tmp,
				    ZPOOL_CONFIG_VERSION);
				fnvlist_add_uint64(config,
				    ZPOOL_CONFIG_VERSION, version);
				guid = fnvlist_lookup_uint64(tmp,
				    ZPOOL_CONFIG_POOL_GUID);
				fnvlist_add_uint64(config,
				    ZPOOL_CONFIG_POOL_GUID, guid);
				name = fnvlist_lookup_string(tmp,
				    ZPOOL_CONFIG_POOL_NAME);
				fnvlist_add_string(config,
				    ZPOOL_CONFIG_POOL_NAME, name);

				if (nvlist_lookup_string(tmp,
				    ZPOOL_CONFIG_COMMENT, &comment) == 0)
					fnvlist_add_string(config,
					    ZPOOL_CONFIG_COMMENT, comment);

				state = fnvlist_lookup_uint64(tmp,
				    ZPOOL_CONFIG_POOL_STATE);
				fnvlist_add_uint64(config,
				    ZPOOL_CONFIG_POOL_STATE, state);

				hostid = 0;
				if (nvlist_lookup_uint64(tmp,
				    ZPOOL_CONFIG_HOSTID, &hostid) == 0) {
					fnvlist_add_uint64(config,
					    ZPOOL_CONFIG_HOSTID, hostid);
					hostname = fnvlist_lookup_string(tmp,
					    ZPOOL_CONFIG_HOSTNAME);
					fnvlist_add_string(config,
					    ZPOOL_CONFIG_HOSTNAME, hostname);
				}

				config_seen = B_TRUE;
			}

			/*
			 * Add this top-level vdev to the child array.
			 */
			verify(nvlist_lookup_nvlist(tmp,
			    ZPOOL_CONFIG_VDEV_TREE, &nvtop) == 0);
			verify(nvlist_lookup_uint64(nvtop, ZPOOL_CONFIG_ID,
			    &id) == 0);

			if (id >= children) {
				nvlist_t **newchild;

				newchild = zfs_alloc(hdl, (id + 1) *
				    sizeof (nvlist_t *));
				if (newchild == NULL)
					goto nomem;

				for (c = 0; c < children; c++)
					newchild[c] = child[c];

				free(child);
				child = newchild;
				children = id + 1;
			}
			if (nvlist_dup(nvtop, &child[id], 0) != 0)
				goto nomem;

		}

		/*
		 * If we have information about all the top-levels then
		 * clean up the nvlist which we've constructed. This
		 * means removing any extraneous devices that are
		 * beyond the valid range or adding devices to the end
		 * of our array which appear to be missing.
		 */
		if (valid_top_config) {
			if (max_id < children) {
				for (c = max_id; c < children; c++)
					nvlist_free(child[c]);
				children = max_id;
			} else if (max_id > children) {
				nvlist_t **newchild;

				newchild = zfs_alloc(hdl, (max_id) *
				    sizeof (nvlist_t *));
				if (newchild == NULL)
					goto nomem;

				for (c = 0; c < children; c++)
					newchild[c] = child[c];

				free(child);
				child = newchild;
				children = max_id;
			}
		}

		verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID,
		    &guid) == 0);

		/*
		 * The vdev namespace may contain holes as a result of
		 * device removal. We must add them back into the vdev
		 * tree before we process any missing devices.
		 */
		if (holes > 0) {
			ASSERT(valid_top_config);

			for (c = 0; c < children; c++) {
				nvlist_t *holey;

				if (child[c] != NULL ||
				    !vdev_is_hole(hole_array, holes, c))
					continue;

				if (nvlist_alloc(&holey, NV_UNIQUE_NAME,
				    0) != 0)
					goto nomem;

				/*
				 * Holes in the namespace are treated as
				 * "hole" top-level vdevs and have a
				 * special flag set on them.
				 */
				if (nvlist_add_string(holey,
				    ZPOOL_CONFIG_TYPE,
				    VDEV_TYPE_HOLE) != 0 ||
				    nvlist_add_uint64(holey,
				    ZPOOL_CONFIG_ID, c) != 0 ||
				    nvlist_add_uint64(holey,
				    ZPOOL_CONFIG_GUID, 0ULL) != 0) {
					nvlist_free(holey);
					goto nomem;
				}
				child[c] = holey;
			}
		}

		/*
		 * Look for any missing top-level vdevs.  If this is the case,
		 * create a faked up 'missing' vdev as a placeholder.  We cannot
		 * simply compress the child array, because the kernel performs
		 * certain checks to make sure the vdev IDs match their location
		 * in the configuration.
		 */
		for (c = 0; c < children; c++) {
			if (child[c] == NULL) {
				nvlist_t *missing;
				if (nvlist_alloc(&missing, NV_UNIQUE_NAME,
				    0) != 0)
					goto nomem;
				if (nvlist_add_string(missing,
				    ZPOOL_CONFIG_TYPE,
				    VDEV_TYPE_MISSING) != 0 ||
				    nvlist_add_uint64(missing,
				    ZPOOL_CONFIG_ID, c) != 0 ||
				    nvlist_add_uint64(missing,
				    ZPOOL_CONFIG_GUID, 0ULL) != 0) {
					nvlist_free(missing);
					goto nomem;
				}
				child[c] = missing;
			}
		}

		/*
		 * Put all of this pool's top-level vdevs into a root vdev.
		 */
		if (nvlist_alloc(&nvroot, NV_UNIQUE_NAME, 0) != 0)
			goto nomem;
		if (nvlist_add_string(nvroot, ZPOOL_CONFIG_TYPE,
		    VDEV_TYPE_ROOT) != 0 ||
		    nvlist_add_uint64(nvroot, ZPOOL_CONFIG_ID, 0ULL) != 0 ||
		    nvlist_add_uint64(nvroot, ZPOOL_CONFIG_GUID, guid) != 0 ||
		    nvlist_add_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
		    child, children) != 0) {
			nvlist_free(nvroot);
			goto nomem;
		}

		for (c = 0; c < children; c++)
			nvlist_free(child[c]);
		free(child);
		children = 0;
		child = NULL;

		/*
		 * Go through and fix up any paths and/or devids based on our
		 * known list of vdev GUID -> path mappings.
		 */
		if (fix_paths(nvroot, pl->names) != 0) {
			nvlist_free(nvroot);
			goto nomem;
		}

		/*
		 * Add the root vdev to this pool's configuration.
		 */
		if (nvlist_add_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
		    nvroot) != 0) {
			nvlist_free(nvroot);
			goto nomem;
		}
		nvlist_free(nvroot);

		/*
		 * zdb uses this path to report on active pools that were
		 * imported or created using -R.
		 */
		if (active_ok)
			goto add_pool;

		/*
		 * Determine if this pool is currently active, in which case we
		 * can't actually import it.
		 */
		verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME,
		    &name) == 0);
		verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID,
		    &guid) == 0);

		if (pool_active(hdl, name, guid, &isactive) != 0)
			goto error;

		if (isactive) {
			nvlist_free(config);
			config = NULL;
			continue;
		}

		if ((nvl = refresh_config(hdl, config)) == NULL) {
			nvlist_free(config);
			config = NULL;
			continue;
		}

		nvlist_free(config);
		config = nvl;

		/*
		 * Go through and update the paths for spares, now that we have
		 * them.
		 */
		verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE,
		    &nvroot) == 0);
		if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES,
		    &spares, &nspares) == 0) {
			for (i = 0; i < nspares; i++) {
				if (fix_paths(spares[i], pl->names) != 0)
					goto nomem;
			}
		}

		/*
		 * Update the paths for l2cache devices.
		 */
		if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE,
		    &l2cache, &nl2cache) == 0) {
			for (i = 0; i < nl2cache; i++) {
				if (fix_paths(l2cache[i], pl->names) != 0)
					goto nomem;
			}
		}

		/*
		 * Restore the original information read from the actual label.
		 */
		(void) nvlist_remove(config, ZPOOL_CONFIG_HOSTID,
		    DATA_TYPE_UINT64);
		(void) nvlist_remove(config, ZPOOL_CONFIG_HOSTNAME,
		    DATA_TYPE_STRING);
		if (hostid != 0) {
			verify(nvlist_add_uint64(config, ZPOOL_CONFIG_HOSTID,
			    hostid) == 0);
			verify(nvlist_add_string(config, ZPOOL_CONFIG_HOSTNAME,
			    hostname) == 0);
		}

add_pool:
		/*
		 * Add this pool to the list of configs.
		 */
		verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME,
		    &name) == 0);
		if (nvlist_add_nvlist(ret, name, config) != 0)
			goto nomem;

		nvlist_free(config);
		config = NULL;
	}

	return (ret);

nomem:
	(void) no_memory(hdl);
error:
	nvlist_free(config);
	nvlist_free(ret);
	for (c = 0; c < children; c++)
		nvlist_free(child[c]);
	free(child);

	return (NULL);
}

/*
 * Return the offset of the given label.
 */
static uint64_t
label_offset(uint64_t size, int l)
{
	ASSERT(P2PHASE_TYPED(size, sizeof (vdev_label_t), uint64_t) == 0);
	return (l * sizeof (vdev_label_t) + (l < VDEV_LABELS / 2 ?
	    0 : size - VDEV_LABELS * sizeof (vdev_label_t)));
}

/*
 * Given a file descriptor, read the label information and return an nvlist
 * describing the configuration, if there is one.  The number of valid
 * labels found will be returned in num_labels when non-NULL.
 */
int
zpool_read_label(int fd, nvlist_t **config, int *num_labels)
{
	struct stat64 statbuf;
	int l, count = 0;
	vdev_label_t *label;
	nvlist_t *expected_config = NULL;
	uint64_t expected_guid = 0, size;
	int error;

	*config = NULL;

	if (fstat64_blk(fd, &statbuf) == -1)
		return (0);
	size = P2ALIGN_TYPED(statbuf.st_size, sizeof (vdev_label_t), uint64_t);

	error = posix_memalign((void **)&label, PAGESIZE, sizeof (*label));
	if (error)
		return (-1);

	for (l = 0; l < VDEV_LABELS; l++) {
		uint64_t state, guid, txg;

		if (pread64(fd, label, sizeof (vdev_label_t),
		    label_offset(size, l)) != sizeof (vdev_label_t))
			continue;

		if (nvlist_unpack(label->vl_vdev_phys.vp_nvlist,
		    sizeof (label->vl_vdev_phys.vp_nvlist), config, 0) != 0)
			continue;

		if (nvlist_lookup_uint64(*config, ZPOOL_CONFIG_GUID,
		    &guid) != 0 || guid == 0) {
			nvlist_free(*config);
			continue;
		}

		if (nvlist_lookup_uint64(*config, ZPOOL_CONFIG_POOL_STATE,
		    &state) != 0 || state > POOL_STATE_L2CACHE) {
			nvlist_free(*config);
			continue;
		}

		if (state != POOL_STATE_SPARE && state != POOL_STATE_L2CACHE &&
		    (nvlist_lookup_uint64(*config, ZPOOL_CONFIG_POOL_TXG,
		    &txg) != 0 || txg == 0)) {
			nvlist_free(*config);
			continue;
		}

		if (expected_guid) {
			if (expected_guid == guid)
				count++;

			nvlist_free(*config);
		} else {
			expected_config = *config;
			expected_guid = guid;
			count++;
		}
	}

	if (num_labels != NULL)
		*num_labels = count;

	free(label);
	*config = expected_config;

	return (0);
}

typedef struct rdsk_node {
	char *rn_name;			/* Full path to device */
	int rn_order;			/* Preferred order (low to high) */
	int rn_num_labels;		/* Number of valid labels */
	uint64_t rn_vdev_guid;		/* Expected vdev guid when set */
	libzfs_handle_t *rn_hdl;
	nvlist_t *rn_config;		/* Label config */
	avl_tree_t *rn_avl;
	avl_node_t rn_node;
	kmutex_t *rn_lock;
	boolean_t rn_labelpaths;
} rdsk_node_t;

/*
 * Sorted by vdev guid and full path to allow for multiple entries with
 * the same full path name.  This is required because it's possible to
 * have multiple block devices with labels that refer to the same
 * ZPOOL_CONFIG_PATH yet have different vdev guids.  In this case both
 * entries need to be added to the cache.  Scenarios where this can occur
 * include overwritten pool labels, devices which are visible from multiple
 * hosts and multipath devices.
 */
static int
slice_cache_compare(const void *arg1, const void *arg2)
{
	const char  *nm1 = ((rdsk_node_t *)arg1)->rn_name;
	const char  *nm2 = ((rdsk_node_t *)arg2)->rn_name;
	uint64_t guid1 = ((rdsk_node_t *)arg1)->rn_vdev_guid;
	uint64_t guid2 = ((rdsk_node_t *)arg2)->rn_vdev_guid;
	int rv;

	rv = AVL_CMP(guid1, guid2);
	if (rv)
		return (rv);

	return (AVL_ISIGN(strcmp(nm1, nm2)));
}

static boolean_t
is_watchdog_dev(char *dev)
{
	/* For 'watchdog' dev */
	if (strcmp(dev, "watchdog") == 0)
		return (B_TRUE);

	/* For 'watchdog<digit><whatever> */
	if (strstr(dev, "watchdog") == dev && isdigit(dev[8]))
		return (B_TRUE);

	return (B_FALSE);
}

static int
label_paths_impl(libzfs_handle_t *hdl, nvlist_t *nvroot, uint64_t pool_guid,
    uint64_t vdev_guid, char **path, char **devid)
{
	nvlist_t **child;
	uint_t c, children;
	uint64_t guid;
	char *val;
	int error;

	if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN,
	    &child, &children) == 0) {
		for (c = 0; c < children; c++) {
			error  = label_paths_impl(hdl, child[c],
			    pool_guid, vdev_guid, path, devid);
			if (error)
				return (error);
		}
		return (0);
	}

	if (nvroot == NULL)
		return (0);

	error = nvlist_lookup_uint64(nvroot, ZPOOL_CONFIG_GUID, &guid);
	if ((error != 0) || (guid != vdev_guid))
		return (0);

	error = nvlist_lookup_string(nvroot, ZPOOL_CONFIG_PATH, &val);
	if (error == 0)
		*path = val;

	error = nvlist_lookup_string(nvroot, ZPOOL_CONFIG_DEVID, &val);
	if (error == 0)
		*devid = val;

	return (0);
}

/*
 * Given a disk label fetch the ZPOOL_CONFIG_PATH and ZPOOL_CONFIG_DEVID
 * and store these strings as config_path and devid_path respectively.
 * The returned pointers are only valid as long as label remains valid.
 */
static int
label_paths(libzfs_handle_t *hdl, nvlist_t *label, char **path, char **devid)
{
	nvlist_t *nvroot;
	uint64_t pool_guid;
	uint64_t vdev_guid;

	*path = NULL;
	*devid = NULL;

	if (nvlist_lookup_nvlist(label, ZPOOL_CONFIG_VDEV_TREE, &nvroot) ||
	    nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_GUID, &pool_guid) ||
	    nvlist_lookup_uint64(label, ZPOOL_CONFIG_GUID, &vdev_guid))
		return (ENOENT);

	return (label_paths_impl(hdl, nvroot, pool_guid, vdev_guid, path,
	    devid));
}

static void
zpool_open_func(void *arg)
{
	rdsk_node_t *rn = arg;
	libzfs_handle_t *hdl = rn->rn_hdl;
	struct stat64 statbuf;
	nvlist_t *config;
	char *bname, *dupname;
	uint64_t vdev_guid = 0;
	int error;
	int num_labels;
	int fd;

	/*
	 * Skip devices with well known prefixes there can be side effects
	 * when opening devices which need to be avoided.
	 *
	 * hpet     - High Precision Event Timer
	 * watchdog - Watchdog must be closed in a special way.
	 */
	dupname = zfs_strdup(hdl, rn->rn_name);
	bname = basename(dupname);
	error = ((strcmp(bname, "hpet") == 0) || is_watchdog_dev(bname));
	free(dupname);
	if (error)
		return;

	/*
	 * Ignore failed stats.  We only want regular files and block devices.
	 */
	if (stat64(rn->rn_name, &statbuf) != 0 ||
	    (!S_ISREG(statbuf.st_mode) && !S_ISBLK(statbuf.st_mode)))
		return;

	/*
	 * Preferentially open using O_DIRECT to bypass the block device
	 * cache which may be stale for multipath devices.  An EINVAL errno
	 * indicates O_DIRECT is unsupported so fallback to just O_RDONLY.
	 */
	fd = open(rn->rn_name, O_RDONLY | O_DIRECT);
	if ((fd < 0) && (errno == EINVAL))
		fd = open(rn->rn_name, O_RDONLY);

	if (fd < 0)
		return;

	/*
	 * This file is too small to hold a zpool
	 */
	if (S_ISREG(statbuf.st_mode) && statbuf.st_size < SPA_MINDEVSIZE) {
		(void) close(fd);
		return;
	}

	error = zpool_read_label(fd, &config, &num_labels);
	if (error != 0) {
		(void) close(fd);
		return;
	}

	if (num_labels == 0) {
		(void) close(fd);
		nvlist_free(config);
		return;
	}

	/*
	 * Check that the vdev is for the expected guid.  Additional entries
	 * are speculatively added based on the paths stored in the labels.
	 * Entries with valid paths but incorrect guids must be removed.
	 */
	error = nvlist_lookup_uint64(config, ZPOOL_CONFIG_GUID, &vdev_guid);
	if (error || (rn->rn_vdev_guid && rn->rn_vdev_guid != vdev_guid)) {
		(void) close(fd);
		nvlist_free(config);
		return;
	}

	(void) close(fd);

	rn->rn_config = config;
	rn->rn_num_labels = num_labels;

	/*
	 * Add additional entries for paths described by this label.
	 */
	if (rn->rn_labelpaths) {
		char *path = NULL;
		char *devid = NULL;
		rdsk_node_t *slice;
		avl_index_t where;
		int error;

		if (label_paths(rn->rn_hdl, rn->rn_config, &path, &devid))
			return;

		/*
		 * Allow devlinks to stabilize so all paths are available.
		 */
		zpool_label_disk_wait(rn->rn_name, DISK_LABEL_WAIT);

		if (path != NULL) {
			slice = zfs_alloc(hdl, sizeof (rdsk_node_t));
			slice->rn_name = zfs_strdup(hdl, path);
			slice->rn_vdev_guid = vdev_guid;
			slice->rn_avl = rn->rn_avl;
			slice->rn_hdl = hdl;
			slice->rn_order = IMPORT_ORDER_PREFERRED_1;
			slice->rn_labelpaths = B_FALSE;
			mutex_enter(rn->rn_lock);
			if (avl_find(rn->rn_avl, slice, &where)) {
			mutex_exit(rn->rn_lock);
				free(slice->rn_name);
				free(slice);
			} else {
				avl_insert(rn->rn_avl, slice, where);
				mutex_exit(rn->rn_lock);
				zpool_open_func(slice);
			}
		}

		if (devid != NULL) {
			slice = zfs_alloc(hdl, sizeof (rdsk_node_t));
			error = asprintf(&slice->rn_name, "%s%s",
			    DEV_BYID_PATH, devid);
			if (error == -1) {
				free(slice);
				return;
			}

			slice->rn_vdev_guid = vdev_guid;
			slice->rn_avl = rn->rn_avl;
			slice->rn_hdl = hdl;
			slice->rn_order = IMPORT_ORDER_PREFERRED_2;
			slice->rn_labelpaths = B_FALSE;
			mutex_enter(rn->rn_lock);
			if (avl_find(rn->rn_avl, slice, &where)) {
				mutex_exit(rn->rn_lock);
				free(slice->rn_name);
				free(slice);
			} else {
				avl_insert(rn->rn_avl, slice, where);
				mutex_exit(rn->rn_lock);
				zpool_open_func(slice);
			}
		}
	}
}

/*
 * Given a file descriptor, clear (zero) the label information.  This function
 * is used in the appliance stack as part of the ZFS sysevent module and
 * to implement the "zpool labelclear" command.
 */
int
zpool_clear_label(int fd)
{
	struct stat64 statbuf;
	int l;
	vdev_label_t *label;
	uint64_t size;

	if (fstat64_blk(fd, &statbuf) == -1)
		return (0);
	size = P2ALIGN_TYPED(statbuf.st_size, sizeof (vdev_label_t), uint64_t);

	if ((label = calloc(sizeof (vdev_label_t), 1)) == NULL)
		return (-1);

	for (l = 0; l < VDEV_LABELS; l++) {
		if (pwrite64(fd, label, sizeof (vdev_label_t),
		    label_offset(size, l)) != sizeof (vdev_label_t)) {
			free(label);
			return (-1);
		}
	}

	free(label);
	return (0);
}

/*
 * Scan a list of directories for zfs devices.
 */
static int
zpool_find_import_scan(libzfs_handle_t *hdl, kmutex_t *lock,
    avl_tree_t **slice_cache, char **dir, int dirs)
{
	avl_tree_t *cache;
	rdsk_node_t *slice;
	void *cookie;
	int i, error;

	*slice_cache = NULL;
	cache = zfs_alloc(hdl, sizeof (avl_tree_t));
	avl_create(cache, slice_cache_compare, sizeof (rdsk_node_t),
	    offsetof(rdsk_node_t, rn_node));

	for (i = 0; i < dirs; i++) {
		char path[MAXPATHLEN];
		struct dirent64 *dp;
		DIR *dirp;

		if (realpath(dir[i], path) == NULL) {
			error = errno;
			if (error == ENOENT)
				continue;

			zfs_error_aux(hdl, strerror(error));
			(void) zfs_error_fmt(hdl, EZFS_BADPATH, dgettext(
			    TEXT_DOMAIN, "cannot resolve path '%s'"), dir[i]);
			goto error;
		}

		dirp = opendir(path);
		if (dirp == NULL) {
			error = errno;
			zfs_error_aux(hdl, strerror(error));
			(void) zfs_error_fmt(hdl, EZFS_BADPATH,
			    dgettext(TEXT_DOMAIN, "cannot open '%s'"), path);
			goto error;
		}

		while ((dp = readdir64(dirp)) != NULL) {
			const char *name = dp->d_name;
			if (name[0] == '.' &&
			    (name[1] == 0 || (name[1] == '.' && name[2] == 0)))
				continue;

			slice = zfs_alloc(hdl, sizeof (rdsk_node_t));
			error = asprintf(&slice->rn_name, "%s/%s", path, name);
			if (error == -1) {
				free(slice);
				continue;
			}
			slice->rn_vdev_guid = 0;
			slice->rn_lock = lock;
			slice->rn_avl = cache;
			slice->rn_hdl = hdl;
			slice->rn_order = i + IMPORT_ORDER_SCAN_OFFSET;
			slice->rn_labelpaths = B_FALSE;
			mutex_enter(lock);
			avl_add(cache, slice);
			mutex_exit(lock);
		}

		(void) closedir(dirp);
	}

	*slice_cache = cache;
	return (0);

error:
	cookie = NULL;
	while ((slice = avl_destroy_nodes(cache, &cookie)) != NULL) {
		free(slice->rn_name);
		free(slice);
	}
	free(cache);

	return (error);
}

/*
 * Use libblkid to quickly enumerate all known zfs devices.
 */
static int
zpool_find_import_blkid(libzfs_handle_t *hdl, kmutex_t *lock,
    avl_tree_t **slice_cache)
{
	rdsk_node_t *slice;
	blkid_cache cache;
	blkid_dev_iterate iter;
	blkid_dev dev;
	avl_index_t where;
	int error;

	*slice_cache = NULL;

	error = blkid_get_cache(&cache, NULL);
	if (error != 0)
		return (error);

	error = blkid_probe_all_new(cache);
	if (error != 0) {
		blkid_put_cache(cache);
		return (error);
	}

	iter = blkid_dev_iterate_begin(cache);
	if (iter == NULL) {
		blkid_put_cache(cache);
		return (EINVAL);
	}

	error = blkid_dev_set_search(iter, "TYPE", "zfs_member");
	if (error != 0) {
		blkid_dev_iterate_end(iter);
		blkid_put_cache(cache);
		return (error);
	}

	*slice_cache = zfs_alloc(hdl, sizeof (avl_tree_t));
	avl_create(*slice_cache, slice_cache_compare, sizeof (rdsk_node_t),
	    offsetof(rdsk_node_t, rn_node));

	while (blkid_dev_next(iter, &dev) == 0) {
		slice = zfs_alloc(hdl, sizeof (rdsk_node_t));
		slice->rn_name = zfs_strdup(hdl, blkid_dev_devname(dev));
		slice->rn_vdev_guid = 0;
		slice->rn_lock = lock;
		slice->rn_avl = *slice_cache;
		slice->rn_hdl = hdl;
		slice->rn_labelpaths = B_TRUE;

		error = zfs_path_order(slice->rn_name, &slice->rn_order);
		if (error == 0)
			slice->rn_order += IMPORT_ORDER_SCAN_OFFSET;
		else
			slice->rn_order = IMPORT_ORDER_DEFAULT;

		mutex_enter(lock);
		if (avl_find(*slice_cache, slice, &where)) {
			free(slice->rn_name);
			free(slice);
		} else {
			avl_insert(*slice_cache, slice, where);
		}
		mutex_exit(lock);
	}

	blkid_dev_iterate_end(iter);
	blkid_put_cache(cache);

	return (0);
}

char *
zpool_default_import_path[DEFAULT_IMPORT_PATH_SIZE] = {
	"/dev/disk/by-vdev",	/* Custom rules, use first if they exist */
	"/dev/mapper",		/* Use multipath devices before components */
	"/dev/disk/by-partlabel", /* Single unique entry set by user */
	"/dev/disk/by-partuuid", /* Generated partition uuid */
	"/dev/disk/by-label",	/* Custom persistent labels */
	"/dev/disk/by-uuid",	/* Single unique entry and persistent */
	"/dev/disk/by-id",	/* May be multiple entries and persistent */
	"/dev/disk/by-path",	/* Encodes physical location and persistent */
	"/dev"			/* UNSAFE device names will change */
};

/*
 * Given a list of directories to search, find all pools stored on disk.  This
 * includes partial pools which are not available to import.  If no args are
 * given (argc is 0), then the default directory (/dev/dsk) is searched.
 * poolname or guid (but not both) are provided by the caller when trying
 * to import a specific pool.
 */
static nvlist_t *
zpool_find_import_impl(libzfs_handle_t *hdl, importargs_t *iarg)
{
	nvlist_t *ret = NULL;
	pool_list_t pools = { 0 };
	pool_entry_t *pe, *penext;
	vdev_entry_t *ve, *venext;
	config_entry_t *ce, *cenext;
	name_entry_t *ne, *nenext;
	kmutex_t lock;
	avl_tree_t *cache;
	rdsk_node_t *slice;
	void *cookie;
	taskq_t *t;

	verify(iarg->poolname == NULL || iarg->guid == 0);
	mutex_init(&lock, NULL, MUTEX_DEFAULT, NULL);

	/*
	 * Locate pool member vdevs using libblkid or by directory scanning.
	 * On success a newly allocated AVL tree which is populated with an
	 * entry for each discovered vdev will be returned as the cache.
	 * It's the callers responsibility to consume and destroy this tree.
	 */
	if (iarg->scan || iarg->paths != 0) {
		int dirs = iarg->paths;
		char **dir = iarg->path;

		if (dirs == 0) {
			dir = zpool_default_import_path;
			dirs = DEFAULT_IMPORT_PATH_SIZE;
		}

		if (zpool_find_import_scan(hdl, &lock, &cache, dir,  dirs) != 0)
			return (NULL);
	} else {
		if (zpool_find_import_blkid(hdl, &lock, &cache) != 0)
			return (NULL);
	}

	/*
	 * Create a thread pool to parallelize the process of reading and
	 * validating labels, a large number of threads can be used due to
	 * minimal contention.
	 */
	t = taskq_create("z_import", 2 * boot_ncpus, defclsyspri,
	    2 * boot_ncpus, INT_MAX, TASKQ_PREPOPULATE);

	for (slice = avl_first(cache); slice;
	    (slice = avl_walk(cache, slice, AVL_AFTER)))
		(void) taskq_dispatch(t, zpool_open_func, slice, TQ_SLEEP);

	taskq_wait(t);
	taskq_destroy(t);

	/*
	 * Process the cache filtering out any entries which are not
	 * for the specificed pool then adding matching label configs.
	 */
	cookie = NULL;
	while ((slice = avl_destroy_nodes(cache, &cookie)) != NULL) {
		if (slice->rn_config != NULL) {
			nvlist_t *config = slice->rn_config;
			boolean_t matched = B_TRUE;
			boolean_t aux = B_FALSE;
			int fd;

			/*
			 * Check if it's a spare or l2cache device. If it is,
			 * we need to skip the name and guid check since they
			 * don't exist on aux device label.
			 */
			if (iarg->poolname != NULL || iarg->guid != 0) {
				uint64_t state;
				aux = nvlist_lookup_uint64(config,
				    ZPOOL_CONFIG_POOL_STATE, &state) == 0 &&
				    (state == POOL_STATE_SPARE ||
				    state == POOL_STATE_L2CACHE);
			}

			if (iarg->poolname != NULL && !aux) {
				char *pname;

				matched = nvlist_lookup_string(config,
				    ZPOOL_CONFIG_POOL_NAME, &pname) == 0 &&
				    strcmp(iarg->poolname, pname) == 0;
			} else if (iarg->guid != 0 && !aux) {
				uint64_t this_guid;

				matched = nvlist_lookup_uint64(config,
				    ZPOOL_CONFIG_POOL_GUID, &this_guid) == 0 &&
				    iarg->guid == this_guid;
			}
			if (!matched) {
				nvlist_free(config);
			} else {
				/*
				 * Verify all remaining entries can be opened
				 * exclusively. This will prune all underlying
				 * multipath devices which otherwise could
				 * result in the vdev appearing as UNAVAIL.
				 *
				 * Under zdb, this step isn't required and
				 * would prevent a zdb -e of active pools with
				 * no cachefile.
				 */
				fd = open(slice->rn_name, O_RDONLY | O_EXCL);
				if (fd >= 0 || iarg->can_be_active) {
					if (fd >= 0)
						close(fd);
					add_config(hdl, &pools,
					    slice->rn_name, slice->rn_order,
					    slice->rn_num_labels, config);
				} else {
					nvlist_free(config);
				}
			}
		}
		free(slice->rn_name);
		free(slice);
	}
	avl_destroy(cache);
	free(cache);
	mutex_destroy(&lock);

	ret = get_configs(hdl, &pools, iarg->can_be_active);

	for (pe = pools.pools; pe != NULL; pe = penext) {
		penext = pe->pe_next;
		for (ve = pe->pe_vdevs; ve != NULL; ve = venext) {
			venext = ve->ve_next;
			for (ce = ve->ve_configs; ce != NULL; ce = cenext) {
				cenext = ce->ce_next;
				nvlist_free(ce->ce_config);
				free(ce);
			}
			free(ve);
		}
		free(pe);
	}

	for (ne = pools.names; ne != NULL; ne = nenext) {
		nenext = ne->ne_next;
		free(ne->ne_name);
		free(ne);
	}

	return (ret);
}

nvlist_t *
zpool_find_import(libzfs_handle_t *hdl, int argc, char **argv)
{
	importargs_t iarg = { 0 };

	iarg.paths = argc;
	iarg.path = argv;

	return (zpool_find_import_impl(hdl, &iarg));
}

/*
 * Given a cache file, return the contents as a list of importable pools.
 * poolname or guid (but not both) are provided by the caller when trying
 * to import a specific pool.
 */
nvlist_t *
zpool_find_import_cached(libzfs_handle_t *hdl, const char *cachefile,
    char *poolname, uint64_t guid)
{
	char *buf;
	int fd;
	struct stat64 statbuf;
	nvlist_t *raw, *src, *dst;
	nvlist_t *pools;
	nvpair_t *elem;
	char *name;
	uint64_t this_guid;
	boolean_t active;

	verify(poolname == NULL || guid == 0);

	if ((fd = open(cachefile, O_RDONLY)) < 0) {
		zfs_error_aux(hdl, "%s", strerror(errno));
		(void) zfs_error(hdl, EZFS_BADCACHE,
		    dgettext(TEXT_DOMAIN, "failed to open cache file"));
		return (NULL);
	}

	if (fstat64(fd, &statbuf) != 0) {
		zfs_error_aux(hdl, "%s", strerror(errno));
		(void) close(fd);
		(void) zfs_error(hdl, EZFS_BADCACHE,
		    dgettext(TEXT_DOMAIN, "failed to get size of cache file"));
		return (NULL);
	}

	if ((buf = zfs_alloc(hdl, statbuf.st_size)) == NULL) {
		(void) close(fd);
		return (NULL);
	}

	if (read(fd, buf, statbuf.st_size) != statbuf.st_size) {
		(void) close(fd);
		free(buf);
		(void) zfs_error(hdl, EZFS_BADCACHE,
		    dgettext(TEXT_DOMAIN,
		    "failed to read cache file contents"));
		return (NULL);
	}

	(void) close(fd);

	if (nvlist_unpack(buf, statbuf.st_size, &raw, 0) != 0) {
		free(buf);
		(void) zfs_error(hdl, EZFS_BADCACHE,
		    dgettext(TEXT_DOMAIN,
		    "invalid or corrupt cache file contents"));
		return (NULL);
	}

	free(buf);

	/*
	 * Go through and get the current state of the pools and refresh their
	 * state.
	 */
	if (nvlist_alloc(&pools, 0, 0) != 0) {
		(void) no_memory(hdl);
		nvlist_free(raw);
		return (NULL);
	}

	elem = NULL;
	while ((elem = nvlist_next_nvpair(raw, elem)) != NULL) {
		src = fnvpair_value_nvlist(elem);

		name = fnvlist_lookup_string(src, ZPOOL_CONFIG_POOL_NAME);
		if (poolname != NULL && strcmp(poolname, name) != 0)
			continue;

		this_guid = fnvlist_lookup_uint64(src, ZPOOL_CONFIG_POOL_GUID);
		if (guid != 0 && guid != this_guid)
			continue;

		if (pool_active(hdl, name, this_guid, &active) != 0) {
			nvlist_free(raw);
			nvlist_free(pools);
			return (NULL);
		}

		if (active)
			continue;

		if ((dst = refresh_config(hdl, src)) == NULL) {
			nvlist_free(raw);
			nvlist_free(pools);
			return (NULL);
		}

		if (nvlist_add_nvlist(pools, nvpair_name(elem), dst) != 0) {
			(void) no_memory(hdl);
			nvlist_free(dst);
			nvlist_free(raw);
			nvlist_free(pools);
			return (NULL);
		}
		nvlist_free(dst);
	}

	nvlist_free(raw);
	return (pools);
}

static int
name_or_guid_exists(zpool_handle_t *zhp, void *data)
{
	importargs_t *import = data;
	int found = 0;

	if (import->poolname != NULL) {
		char *pool_name;

		verify(nvlist_lookup_string(zhp->zpool_config,
		    ZPOOL_CONFIG_POOL_NAME, &pool_name) == 0);
		if (strcmp(pool_name, import->poolname) == 0)
			found = 1;
	} else {
		uint64_t pool_guid;

		verify(nvlist_lookup_uint64(zhp->zpool_config,
		    ZPOOL_CONFIG_POOL_GUID, &pool_guid) == 0);
		if (pool_guid == import->guid)
			found = 1;
	}

	zpool_close(zhp);
	return (found);
}

nvlist_t *
zpool_search_import(libzfs_handle_t *hdl, importargs_t *import)
{
	verify(import->poolname == NULL || import->guid == 0);

	if (import->unique)
		import->exists = zpool_iter(hdl, name_or_guid_exists, import);

	if (import->cachefile != NULL)
		return (zpool_find_import_cached(hdl, import->cachefile,
		    import->poolname, import->guid));

	return (zpool_find_import_impl(hdl, import));
}

static boolean_t
pool_match(nvlist_t *cfg, char *tgt)
{
	uint64_t v, guid = strtoull(tgt, NULL, 0);
	char *s;

	if (guid != 0) {
		if (nvlist_lookup_uint64(cfg, ZPOOL_CONFIG_POOL_GUID, &v) == 0)
			return (v == guid);
	} else {
		if (nvlist_lookup_string(cfg, ZPOOL_CONFIG_POOL_NAME, &s) == 0)
			return (strcmp(s, tgt) == 0);
	}
	return (B_FALSE);
}

int
zpool_tryimport(libzfs_handle_t *hdl, char *target, nvlist_t **configp,
    importargs_t *args)
{
	nvlist_t *pools;
	nvlist_t *match = NULL;
	nvlist_t *config = NULL;
	char *name = NULL, *sepp = NULL;
	char sep = '\0';
	int count = 0;
	char *targetdup = strdup(target);

	*configp = NULL;

	if ((sepp = strpbrk(targetdup, "/@")) != NULL) {
		sep = *sepp;
		*sepp = '\0';
	}

	pools = zpool_search_import(hdl, args);

	if (pools != NULL) {
		nvpair_t *elem = NULL;
		while ((elem = nvlist_next_nvpair(pools, elem)) != NULL) {
			VERIFY0(nvpair_value_nvlist(elem, &config));
			if (pool_match(config, targetdup)) {
				count++;
				if (match != NULL) {
					/* multiple matches found */
					continue;
				} else {
					match = config;
					name = nvpair_name(elem);
				}
			}
		}
	}

	if (count == 0) {
		(void) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
		    "no pools found"));
		free(targetdup);
		return (ENOENT);
	}

	if (count > 1) {
		(void) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
		    "%d pools found, use pool GUID\n"), count);
		free(targetdup);
		return (EINVAL);
	}

	*configp = match;
	free(targetdup);

	return (0);
}

boolean_t
find_guid(nvlist_t *nv, uint64_t guid)
{
	uint64_t tmp;
	nvlist_t **child;
	uint_t c, children;

	verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &tmp) == 0);
	if (tmp == guid)
		return (B_TRUE);

	if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
	    &child, &children) == 0) {
		for (c = 0; c < children; c++)
			if (find_guid(child[c], guid))
				return (B_TRUE);
	}

	return (B_FALSE);
}

typedef struct aux_cbdata {
	const char	*cb_type;
	uint64_t	cb_guid;
	zpool_handle_t	*cb_zhp;
} aux_cbdata_t;

static int
find_aux(zpool_handle_t *zhp, void *data)
{
	aux_cbdata_t *cbp = data;
	nvlist_t **list;
	uint_t i, count;
	uint64_t guid;
	nvlist_t *nvroot;

	verify(nvlist_lookup_nvlist(zhp->zpool_config, ZPOOL_CONFIG_VDEV_TREE,
	    &nvroot) == 0);

	if (nvlist_lookup_nvlist_array(nvroot, cbp->cb_type,
	    &list, &count) == 0) {
		for (i = 0; i < count; i++) {
			verify(nvlist_lookup_uint64(list[i],
			    ZPOOL_CONFIG_GUID, &guid) == 0);
			if (guid == cbp->cb_guid) {
				cbp->cb_zhp = zhp;
				return (1);
			}
		}
	}

	zpool_close(zhp);
	return (0);
}

/*
 * Determines if the pool is in use.  If so, it returns true and the state of
 * the pool as well as the name of the pool.  Both strings are allocated and
 * must be freed by the caller.
 */
int
zpool_in_use(libzfs_handle_t *hdl, int fd, pool_state_t *state, char **namestr,
    boolean_t *inuse)
{
	nvlist_t *config;
	char *name;
	boolean_t ret;
	uint64_t guid, vdev_guid;
	zpool_handle_t *zhp;
	nvlist_t *pool_config;
	uint64_t stateval, isspare;
	aux_cbdata_t cb = { 0 };
	boolean_t isactive;

	*inuse = B_FALSE;

	if (zpool_read_label(fd, &config, NULL) != 0) {
		(void) no_memory(hdl);
		return (-1);
	}

	if (config == NULL)
		return (0);

	verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE,
	    &stateval) == 0);
	verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_GUID,
	    &vdev_guid) == 0);

	if (stateval != POOL_STATE_SPARE && stateval != POOL_STATE_L2CACHE) {
		verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME,
		    &name) == 0);
		verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID,
		    &guid) == 0);
	}

	switch (stateval) {
	case POOL_STATE_EXPORTED:
		/*
		 * A pool with an exported state may in fact be imported
		 * read-only, so check the in-core state to see if it's
		 * active and imported read-only.  If it is, set
		 * its state to active.
		 */
		if (pool_active(hdl, name, guid, &isactive) == 0 && isactive &&
		    (zhp = zpool_open_canfail(hdl, name)) != NULL) {
			if (zpool_get_prop_int(zhp, ZPOOL_PROP_READONLY, NULL))
				stateval = POOL_STATE_ACTIVE;

			/*
			 * All we needed the zpool handle for is the
			 * readonly prop check.
			 */
			zpool_close(zhp);
		}

		ret = B_TRUE;
		break;

	case POOL_STATE_ACTIVE:
		/*
		 * For an active pool, we have to determine if it's really part
		 * of a currently active pool (in which case the pool will exist
		 * and the guid will be the same), or whether it's part of an
		 * active pool that was disconnected without being explicitly
		 * exported.
		 */
		if (pool_active(hdl, name, guid, &isactive) != 0) {
			nvlist_free(config);
			return (-1);
		}

		if (isactive) {
			/*
			 * Because the device may have been removed while
			 * offlined, we only report it as active if the vdev is
			 * still present in the config.  Otherwise, pretend like
			 * it's not in use.
			 */
			if ((zhp = zpool_open_canfail(hdl, name)) != NULL &&
			    (pool_config = zpool_get_config(zhp, NULL))
			    != NULL) {
				nvlist_t *nvroot;

				verify(nvlist_lookup_nvlist(pool_config,
				    ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0);
				ret = find_guid(nvroot, vdev_guid);
			} else {
				ret = B_FALSE;
			}

			/*
			 * If this is an active spare within another pool, we
			 * treat it like an unused hot spare.  This allows the
			 * user to create a pool with a hot spare that currently
			 * in use within another pool.  Since we return B_TRUE,
			 * libdiskmgt will continue to prevent generic consumers
			 * from using the device.
			 */
			if (ret && nvlist_lookup_uint64(config,
			    ZPOOL_CONFIG_IS_SPARE, &isspare) == 0 && isspare)
				stateval = POOL_STATE_SPARE;

			if (zhp != NULL)
				zpool_close(zhp);
		} else {
			stateval = POOL_STATE_POTENTIALLY_ACTIVE;
			ret = B_TRUE;
		}
		break;

	case POOL_STATE_SPARE:
		/*
		 * For a hot spare, it can be either definitively in use, or
		 * potentially active.  To determine if it's in use, we iterate
		 * over all pools in the system and search for one with a spare
		 * with a matching guid.
		 *
		 * Due to the shared nature of spares, we don't actually report
		 * the potentially active case as in use.  This means the user
		 * can freely create pools on the hot spares of exported pools,
		 * but to do otherwise makes the resulting code complicated, and
		 * we end up having to deal with this case anyway.
		 */
		cb.cb_zhp = NULL;
		cb.cb_guid = vdev_guid;
		cb.cb_type = ZPOOL_CONFIG_SPARES;
		if (zpool_iter(hdl, find_aux, &cb) == 1) {
			name = (char *)zpool_get_name(cb.cb_zhp);
			ret = B_TRUE;
		} else {
			ret = B_FALSE;
		}
		break;

	case POOL_STATE_L2CACHE:

		/*
		 * Check if any pool is currently using this l2cache device.
		 */
		cb.cb_zhp = NULL;
		cb.cb_guid = vdev_guid;
		cb.cb_type = ZPOOL_CONFIG_L2CACHE;
		if (zpool_iter(hdl, find_aux, &cb) == 1) {
			name = (char *)zpool_get_name(cb.cb_zhp);
			ret = B_TRUE;
		} else {
			ret = B_FALSE;
		}
		break;

	default:
		ret = B_FALSE;
	}


	if (ret) {
		if ((*namestr = zfs_strdup(hdl, name)) == NULL) {
			if (cb.cb_zhp)
				zpool_close(cb.cb_zhp);
			nvlist_free(config);
			return (-1);
		}
		*state = (pool_state_t)stateval;
	}

	if (cb.cb_zhp)
		zpool_close(cb.cb_zhp);

	nvlist_free(config);
	*inuse = ret;
	return (0);
}