aboutsummaryrefslogtreecommitdiff
path: root/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_import.c
blob: c4e9adae98abbd140cc341a80a91216176453947 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
/*
 * 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) 2013 by Delphix. All rights reserved.
 */

/*
 * 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>
#include <stddef.h>
#include <stdlib.h>
#include <string.h>
#include <sys/stat.h>
#include <unistd.h>
#include <fcntl.h>
#include <thread_pool.h>
#include <libgeom.h>

#include <sys/vdev_impl.h>

#include "libzfs.h"
#include "libzfs_impl.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;
	struct name_entry	*ne_next;
} name_entry_t;

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

static char *
get_devid(const char *path)
{
#ifdef have_devid
	int fd;
	ddi_devid_t devid;
	char *minor, *ret;

	if ((fd = open(path, O_RDONLY)) < 0)
		return (NULL);

	minor = NULL;
	ret = NULL;
	if (devid_get(fd, &devid) == 0) {
		if (devid_get_minor_name(fd, &minor) == 0)
			ret = devid_str_encode(devid, minor);
		if (minor != NULL)
			devid_str_free(minor);
		devid_free(devid);
	}
	(void) close(fd);

	return (ret);
#else
	return (NULL);
#endif
}


/*
 * 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, *devid;
	int matched;

	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 slices or multiple paths to the same
	 * disk.  If this is the case, then we want to pick the path that is
	 * the most similar to the original, where "most similar" is the number
	 * of matching characters starting from the end of the path.  This will
	 * preserve slice numbers even if the disks have been reorganized, and
	 * will also catch preferred disk names if multiple paths exist.
	 */
	verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) == 0);
	if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) != 0)
		path = NULL;

	matched = 0;
	best = NULL;
	for (ne = names; ne != NULL; ne = ne->ne_next) {
		if (ne->ne_guid == guid) {
			const char *src, *dst;
			int count;

			if (path == NULL) {
				best = ne;
				break;
			}

			src = ne->ne_name + strlen(ne->ne_name) - 1;
			dst = path + strlen(path) - 1;
			for (count = 0; src >= ne->ne_name && dst >= path;
			    src--, dst--, count++)
				if (*src != *dst)
					break;

			/*
			 * At this point, 'count' is the number of characters
			 * matched from the end.
			 */
			if (count > matched || best == NULL) {
				best = ne;
				matched = count;
			}
		}
	}

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

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

	if ((devid = get_devid(best->ne_name)) == NULL) {
		(void) nvlist_remove_all(nv, ZPOOL_CONFIG_DEVID);
	} else {
		if (nvlist_add_string(nv, ZPOOL_CONFIG_DEVID, devid) != 0) {
			devid_str_free(devid);
			return (-1);
		}
		devid_str_free(devid);
	}

	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,
    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)
			return (-1);

		if ((ne->ne_name = zfs_strdup(hdl, path)) == NULL) {
			free(ne);
			return (-1);
		}
		ne->ne_guid = vdev_guid;
		ne->ne_next = pl->names;
		pl->names = ne;
		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_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;

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

	if (zcmd_alloc_dst_nvlist(hdl, &zc,
	    zc.zc_nvlist_conf_size * 2) != 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)
{
	for (int 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, *nvtop, *nvroot;
	nvlist_t **spares, **l2cache;
	uint_t i, nspares, nl2cache;
	boolean_t config_seen;
	uint64_t best_txg;
	char *name, *hostname;
	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 found_one = B_FALSE;
	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;

		found_one = B_TRUE;
		nvlist_free(config);
		config = NULL;
	}

	if (!found_one) {
		nvlist_free(ret);
		ret = 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.
 */
int
zpool_read_label(int fd, nvlist_t **config)
{
	struct stat64 statbuf;
	int l;
	vdev_label_t *label;
	uint64_t state, txg, size;

	*config = NULL;

	if (fstat64(fd, &statbuf) == -1)
		return (0);
	size = P2ALIGN_TYPED(statbuf.st_size, sizeof (vdev_label_t), uint64_t);

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

	for (l = 0; l < VDEV_LABELS; l++) {
		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_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;
		}

		free(label);
		return (0);
	}

	free(label);
	*config = NULL;
	return (0);
}

typedef struct rdsk_node {
	char *rn_name;
	int rn_dfd;
	libzfs_handle_t *rn_hdl;
	nvlist_t *rn_config;
	avl_tree_t *rn_avl;
	avl_node_t rn_node;
	boolean_t rn_nozpool;
} rdsk_node_t;

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;
	char *nm1slice, *nm2slice;
	int rv;

	/*
	 * slices zero and two are the most likely to provide results,
	 * so put those first
	 */
	nm1slice = strstr(nm1, "s0");
	nm2slice = strstr(nm2, "s0");
	if (nm1slice && !nm2slice) {
		return (-1);
	}
	if (!nm1slice && nm2slice) {
		return (1);
	}
	nm1slice = strstr(nm1, "s2");
	nm2slice = strstr(nm2, "s2");
	if (nm1slice && !nm2slice) {
		return (-1);
	}
	if (!nm1slice && nm2slice) {
		return (1);
	}

	rv = strcmp(nm1, nm2);
	if (rv == 0)
		return (0);
	return (rv > 0 ? 1 : -1);
}

#ifdef illumos
static void
check_one_slice(avl_tree_t *r, char *diskname, uint_t partno,
    diskaddr_t size, uint_t blksz)
{
	rdsk_node_t tmpnode;
	rdsk_node_t *node;
	char sname[MAXNAMELEN];

	tmpnode.rn_name = &sname[0];
	(void) snprintf(tmpnode.rn_name, MAXNAMELEN, "%s%u",
	    diskname, partno);
	/*
	 * protect against division by zero for disk labels that
	 * contain a bogus sector size
	 */
	if (blksz == 0)
		blksz = DEV_BSIZE;
	/* too small to contain a zpool? */
	if ((size < (SPA_MINDEVSIZE / blksz)) &&
	    (node = avl_find(r, &tmpnode, NULL)))
		node->rn_nozpool = B_TRUE;
}
#endif	/* illumos */

static void
nozpool_all_slices(avl_tree_t *r, const char *sname)
{
#ifdef illumos
	char diskname[MAXNAMELEN];
	char *ptr;
	int i;

	(void) strncpy(diskname, sname, MAXNAMELEN);
	if (((ptr = strrchr(diskname, 's')) == NULL) &&
	    ((ptr = strrchr(diskname, 'p')) == NULL))
		return;
	ptr[0] = 's';
	ptr[1] = '\0';
	for (i = 0; i < NDKMAP; i++)
		check_one_slice(r, diskname, i, 0, 1);
	ptr[0] = 'p';
	for (i = 0; i <= FD_NUMPART; i++)
		check_one_slice(r, diskname, i, 0, 1);
#endif	/* illumos */
}

#ifdef illumos
static void
check_slices(avl_tree_t *r, int fd, const char *sname)
{
	struct extvtoc vtoc;
	struct dk_gpt *gpt;
	char diskname[MAXNAMELEN];
	char *ptr;
	int i;

	(void) strncpy(diskname, sname, MAXNAMELEN);
	if ((ptr = strrchr(diskname, 's')) == NULL || !isdigit(ptr[1]))
		return;
	ptr[1] = '\0';

	if (read_extvtoc(fd, &vtoc) >= 0) {
		for (i = 0; i < NDKMAP; i++)
			check_one_slice(r, diskname, i,
			    vtoc.v_part[i].p_size, vtoc.v_sectorsz);
	} else if (efi_alloc_and_read(fd, &gpt) >= 0) {
		/*
		 * on x86 we'll still have leftover links that point
		 * to slices s[9-15], so use NDKMAP instead
		 */
		for (i = 0; i < NDKMAP; i++)
			check_one_slice(r, diskname, i,
			    gpt->efi_parts[i].p_size, gpt->efi_lbasize);
		/* nodes p[1-4] are never used with EFI labels */
		ptr[0] = 'p';
		for (i = 1; i <= FD_NUMPART; i++)
			check_one_slice(r, diskname, i, 0, 1);
		efi_free(gpt);
	}
}
#endif	/* illumos */

static void
zpool_open_func(void *arg)
{
	rdsk_node_t *rn = arg;
	struct stat64 statbuf;
	nvlist_t *config;
	int fd;

	if (rn->rn_nozpool)
		return;
	if ((fd = openat64(rn->rn_dfd, rn->rn_name, O_RDONLY)) < 0) {
		/* symlink to a device that's no longer there */
		if (errno == ENOENT)
			nozpool_all_slices(rn->rn_avl, rn->rn_name);
		return;
	}
	/*
	 * Ignore failed stats.  We only want regular
	 * files, character devs and block devs.
	 */
	if (fstat64(fd, &statbuf) != 0 ||
	    (!S_ISREG(statbuf.st_mode) &&
	    !S_ISCHR(statbuf.st_mode) &&
	    !S_ISBLK(statbuf.st_mode))) {
		(void) close(fd);
		return;
	}
	/* this file is too small to hold a zpool */
#ifdef illumos
	if (S_ISREG(statbuf.st_mode) &&
	    statbuf.st_size < SPA_MINDEVSIZE) {
		(void) close(fd);
		return;
	} else if (!S_ISREG(statbuf.st_mode)) {
		/*
		 * Try to read the disk label first so we don't have to
		 * open a bunch of minor nodes that can't have a zpool.
		 */
		check_slices(rn->rn_avl, fd, rn->rn_name);
	}
#else	/* !illumos */
	if (statbuf.st_size < SPA_MINDEVSIZE) {
		(void) close(fd);
		return;
	}
#endif	/* illumos */

	if ((zpool_read_label(fd, &config)) != 0) {
		(void) close(fd);
		(void) no_memory(rn->rn_hdl);
		return;
	}
	(void) close(fd);


	rn->rn_config = config;
	if (config != NULL) {
		assert(rn->rn_nozpool == B_FALSE);
	}
}

/*
 * 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(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);
}

/*
 * 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)
{
	int i, dirs = iarg->paths;
	struct dirent64 *dp;
	char path[MAXPATHLEN];
	char *end, **dir = iarg->path;
	size_t pathleft;
	nvlist_t *ret = NULL;
	static char *default_dir = "/dev";
	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;
	avl_tree_t slice_cache;
	rdsk_node_t *slice;
	void *cookie;

	if (dirs == 0) {
		dirs = 1;
		dir = &default_dir;
	}

	/*
	 * Go through and read the label configuration information from every
	 * possible device, organizing the information according to pool GUID
	 * and toplevel GUID.
	 */
	for (i = 0; i < dirs; i++) {
		tpool_t *t;
		char *rdsk;
		int dfd;
		boolean_t config_failed = B_FALSE;
		DIR *dirp;

		/* use realpath to normalize the path */
		if (realpath(dir[i], path) == 0) {
			(void) zfs_error_fmt(hdl, EZFS_BADPATH,
			    dgettext(TEXT_DOMAIN, "cannot open '%s'"), dir[i]);
			goto error;
		}
		end = &path[strlen(path)];
		*end++ = '/';
		*end = 0;
		pathleft = &path[sizeof (path)] - end;

		/*
		 * Using raw devices instead of block devices when we're
		 * reading the labels skips a bunch of slow operations during
		 * close(2) processing, so we replace /dev/dsk with /dev/rdsk.
		 */
		if (strcmp(path, "/dev/dsk/") == 0)
			rdsk = "/dev/";
		else
			rdsk = path;

		if ((dfd = open64(rdsk, O_RDONLY)) < 0 ||
		    (dirp = fdopendir(dfd)) == NULL) {
			if (dfd >= 0)
				(void) close(dfd);
			zfs_error_aux(hdl, strerror(errno));
			(void) zfs_error_fmt(hdl, EZFS_BADPATH,
			    dgettext(TEXT_DOMAIN, "cannot open '%s'"),
			    rdsk);
			goto error;
		}

		avl_create(&slice_cache, slice_cache_compare,
		    sizeof (rdsk_node_t), offsetof(rdsk_node_t, rn_node));

		if (strcmp(rdsk, "/dev/") == 0) {
			struct gmesh mesh;
			struct gclass *mp;
			struct ggeom *gp;
			struct gprovider *pp;

			errno = geom_gettree(&mesh);
			if (errno != 0) {
				zfs_error_aux(hdl, strerror(errno));
				(void) zfs_error_fmt(hdl, EZFS_BADPATH,
				    dgettext(TEXT_DOMAIN, "cannot get GEOM tree"));
				goto error;
			}

			LIST_FOREACH(mp, &mesh.lg_class, lg_class) {
		        	LIST_FOREACH(gp, &mp->lg_geom, lg_geom) {
					LIST_FOREACH(pp, &gp->lg_provider, lg_provider) {
						slice = zfs_alloc(hdl, sizeof (rdsk_node_t));
						slice->rn_name = zfs_strdup(hdl, pp->lg_name);
						slice->rn_avl = &slice_cache;
						slice->rn_dfd = dfd;
						slice->rn_hdl = hdl;
						slice->rn_nozpool = B_FALSE;
						avl_add(&slice_cache, slice);
					}
				}
			}

			geom_deletetree(&mesh);
			goto skipdir;
		}

		/*
		 * This is not MT-safe, but we have no MT consumers of libzfs
		 */
		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));
			slice->rn_name = zfs_strdup(hdl, name);
			slice->rn_avl = &slice_cache;
			slice->rn_dfd = dfd;
			slice->rn_hdl = hdl;
			slice->rn_nozpool = B_FALSE;
			avl_add(&slice_cache, slice);
		}
skipdir:
		/*
		 * create a thread pool to do all of this in parallel;
		 * rn_nozpool is not protected, so this is racy in that
		 * multiple tasks could decide that the same slice can
		 * not hold a zpool, which is benign.  Also choose
		 * double the number of processors; we hold a lot of
		 * locks in the kernel, so going beyond this doesn't
		 * buy us much.
		 */
		t = tpool_create(1, 2 * sysconf(_SC_NPROCESSORS_ONLN),
		    0, NULL);
		for (slice = avl_first(&slice_cache); slice;
		    (slice = avl_walk(&slice_cache, slice,
		    AVL_AFTER)))
			(void) tpool_dispatch(t, zpool_open_func, slice);
		tpool_wait(t);
		tpool_destroy(t);

		cookie = NULL;
		while ((slice = avl_destroy_nodes(&slice_cache,
		    &cookie)) != NULL) {
			if (slice->rn_config != NULL && !config_failed) {
				nvlist_t *config = slice->rn_config;
				boolean_t matched = B_TRUE;

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

					matched = nvlist_lookup_string(config,
					    ZPOOL_CONFIG_POOL_NAME,
					    &pname) == 0 &&
					    strcmp(iarg->poolname, pname) == 0;
				} else if (iarg->guid != 0) {
					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 {
					/*
					 * use the non-raw path for the config
					 */
					(void) strlcpy(end, slice->rn_name,
					    pathleft);
					if (add_config(hdl, &pools, path,
					    config) != 0)
						config_failed = B_TRUE;
				}
			}
			free(slice->rn_name);
			free(slice);
		}
		avl_destroy(&slice_cache);

		(void) closedir(dirp);

		if (config_failed)
			goto error;
	}

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

error:
	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;
				if (ce->ce_config)
					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));
}

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) != 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);
}