summaryrefslogtreecommitdiff
path: root/drivers/md/dm-snap.c
blob: 36805b12661e173be5d2a845626f82baaf8de690 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
/*
 * dm-snapshot.c
 *
 * Copyright (C) 2001-2002 Sistina Software (UK) Limited.
 *
 * This file is released under the GPL.
 */

#include <linux/blkdev.h>
#include <linux/device-mapper.h>
#include <linux/delay.h>
#include <linux/fs.h>
#include <linux/init.h>
#include <linux/kdev_t.h>
#include <linux/list.h>
#include <linux/mempool.h>
#include <linux/module.h>
#include <linux/slab.h>
#include <linux/vmalloc.h>
#include <linux/log2.h>
#include <linux/dm-kcopyd.h>
#include <linux/semaphore.h>

#include "dm.h"

#include "dm-exception-store.h"

#define DM_MSG_PREFIX "snapshots"

static const char dm_snapshot_merge_target_name[] = "snapshot-merge";

#define dm_target_is_snapshot_merge(ti) \
	((ti)->type->name == dm_snapshot_merge_target_name)

/*
 * The size of the mempool used to track chunks in use.
 */
#define MIN_IOS 256

#define DM_TRACKED_CHUNK_HASH_SIZE	16
#define DM_TRACKED_CHUNK_HASH(x)	((unsigned long)(x) & \
					 (DM_TRACKED_CHUNK_HASH_SIZE - 1))

struct dm_exception_table {
	uint32_t hash_mask;
	unsigned hash_shift;
	struct list_head *table;
};

struct dm_snapshot {
	struct mutex lock;

	struct dm_dev *origin;
	struct dm_dev *cow;

	struct dm_target *ti;

	/* List of snapshots per Origin */
	struct list_head list;

	/*
	 * You can't use a snapshot if this is 0 (e.g. if full).
	 * A snapshot-merge target never clears this.
	 */
	int valid;

	/*
	 * The snapshot overflowed because of a write to the snapshot device.
	 * We don't have to invalidate the snapshot in this case, but we need
	 * to prevent further writes.
	 */
	int snapshot_overflowed;

	/* Origin writes don't trigger exceptions until this is set */
	int active;

	atomic_t pending_exceptions_count;

	/* Protected by "lock" */
	sector_t exception_start_sequence;

	/* Protected by kcopyd single-threaded callback */
	sector_t exception_complete_sequence;

	/*
	 * A list of pending exceptions that completed out of order.
	 * Protected by kcopyd single-threaded callback.
	 */
	struct rb_root out_of_order_tree;

	mempool_t pending_pool;

	struct dm_exception_table pending;
	struct dm_exception_table complete;

	/*
	 * pe_lock protects all pending_exception operations and access
	 * as well as the snapshot_bios list.
	 */
	spinlock_t pe_lock;

	/* Chunks with outstanding reads */
	spinlock_t tracked_chunk_lock;
	struct hlist_head tracked_chunk_hash[DM_TRACKED_CHUNK_HASH_SIZE];

	/* The on disk metadata handler */
	struct dm_exception_store *store;

	/* Maximum number of in-flight COW jobs. */
	struct semaphore cow_count;

	struct dm_kcopyd_client *kcopyd_client;

	/* Wait for events based on state_bits */
	unsigned long state_bits;

	/* Range of chunks currently being merged. */
	chunk_t first_merging_chunk;
	int num_merging_chunks;

	/*
	 * The merge operation failed if this flag is set.
	 * Failure modes are handled as follows:
	 * - I/O error reading the header
	 *   	=> don't load the target; abort.
	 * - Header does not have "valid" flag set
	 *   	=> use the origin; forget about the snapshot.
	 * - I/O error when reading exceptions
	 *   	=> don't load the target; abort.
	 *         (We can't use the intermediate origin state.)
	 * - I/O error while merging
	 *	=> stop merging; set merge_failed; process I/O normally.
	 */
	int merge_failed;

	/*
	 * Incoming bios that overlap with chunks being merged must wait
	 * for them to be committed.
	 */
	struct bio_list bios_queued_during_merge;
};

/*
 * state_bits:
 *   RUNNING_MERGE  - Merge operation is in progress.
 *   SHUTDOWN_MERGE - Set to signal that merge needs to be stopped;
 *                    cleared afterwards.
 */
#define RUNNING_MERGE          0
#define SHUTDOWN_MERGE         1

/*
 * Maximum number of chunks being copied on write.
 *
 * The value was decided experimentally as a trade-off between memory
 * consumption, stalling the kernel's workqueues and maintaining a high enough
 * throughput.
 */
#define DEFAULT_COW_THRESHOLD 2048

static int cow_threshold = DEFAULT_COW_THRESHOLD;
module_param_named(snapshot_cow_threshold, cow_threshold, int, 0644);
MODULE_PARM_DESC(snapshot_cow_threshold, "Maximum number of chunks being copied on write");

DECLARE_DM_KCOPYD_THROTTLE_WITH_MODULE_PARM(snapshot_copy_throttle,
		"A percentage of time allocated for copy on write");

struct dm_dev *dm_snap_origin(struct dm_snapshot *s)
{
	return s->origin;
}
EXPORT_SYMBOL(dm_snap_origin);

struct dm_dev *dm_snap_cow(struct dm_snapshot *s)
{
	return s->cow;
}
EXPORT_SYMBOL(dm_snap_cow);

static sector_t chunk_to_sector(struct dm_exception_store *store,
				chunk_t chunk)
{
	return chunk << store->chunk_shift;
}

static int bdev_equal(struct block_device *lhs, struct block_device *rhs)
{
	/*
	 * There is only ever one instance of a particular block
	 * device so we can compare pointers safely.
	 */
	return lhs == rhs;
}

struct dm_snap_pending_exception {
	struct dm_exception e;

	/*
	 * Origin buffers waiting for this to complete are held
	 * in a bio list
	 */
	struct bio_list origin_bios;
	struct bio_list snapshot_bios;

	/* Pointer back to snapshot context */
	struct dm_snapshot *snap;

	/*
	 * 1 indicates the exception has already been sent to
	 * kcopyd.
	 */
	int started;

	/* There was copying error. */
	int copy_error;

	/* A sequence number, it is used for in-order completion. */
	sector_t exception_sequence;

	struct rb_node out_of_order_node;

	/*
	 * For writing a complete chunk, bypassing the copy.
	 */
	struct bio *full_bio;
	bio_end_io_t *full_bio_end_io;
};

/*
 * Hash table mapping origin volumes to lists of snapshots and
 * a lock to protect it
 */
static struct kmem_cache *exception_cache;
static struct kmem_cache *pending_cache;

struct dm_snap_tracked_chunk {
	struct hlist_node node;
	chunk_t chunk;
};

static void init_tracked_chunk(struct bio *bio)
{
	struct dm_snap_tracked_chunk *c = dm_per_bio_data(bio, sizeof(struct dm_snap_tracked_chunk));
	INIT_HLIST_NODE(&c->node);
}

static bool is_bio_tracked(struct bio *bio)
{
	struct dm_snap_tracked_chunk *c = dm_per_bio_data(bio, sizeof(struct dm_snap_tracked_chunk));
	return !hlist_unhashed(&c->node);
}

static void track_chunk(struct dm_snapshot *s, struct bio *bio, chunk_t chunk)
{
	struct dm_snap_tracked_chunk *c = dm_per_bio_data(bio, sizeof(struct dm_snap_tracked_chunk));

	c->chunk = chunk;

	spin_lock_irq(&s->tracked_chunk_lock);
	hlist_add_head(&c->node,
		       &s->tracked_chunk_hash[DM_TRACKED_CHUNK_HASH(chunk)]);
	spin_unlock_irq(&s->tracked_chunk_lock);
}

static void stop_tracking_chunk(struct dm_snapshot *s, struct bio *bio)
{
	struct dm_snap_tracked_chunk *c = dm_per_bio_data(bio, sizeof(struct dm_snap_tracked_chunk));
	unsigned long flags;

	spin_lock_irqsave(&s->tracked_chunk_lock, flags);
	hlist_del(&c->node);
	spin_unlock_irqrestore(&s->tracked_chunk_lock, flags);
}

static int __chunk_is_tracked(struct dm_snapshot *s, chunk_t chunk)
{
	struct dm_snap_tracked_chunk *c;
	int found = 0;

	spin_lock_irq(&s->tracked_chunk_lock);

	hlist_for_each_entry(c,
	    &s->tracked_chunk_hash[DM_TRACKED_CHUNK_HASH(chunk)], node) {
		if (c->chunk == chunk) {
			found = 1;
			break;
		}
	}

	spin_unlock_irq(&s->tracked_chunk_lock);

	return found;
}

/*
 * This conflicting I/O is extremely improbable in the caller,
 * so msleep(1) is sufficient and there is no need for a wait queue.
 */
static void __check_for_conflicting_io(struct dm_snapshot *s, chunk_t chunk)
{
	while (__chunk_is_tracked(s, chunk))
		msleep(1);
}

/*
 * One of these per registered origin, held in the snapshot_origins hash
 */
struct origin {
	/* The origin device */
	struct block_device *bdev;

	struct list_head hash_list;

	/* List of snapshots for this origin */
	struct list_head snapshots;
};

/*
 * This structure is allocated for each origin target
 */
struct dm_origin {
	struct dm_dev *dev;
	struct dm_target *ti;
	unsigned split_boundary;
	struct list_head hash_list;
};

/*
 * Size of the hash table for origin volumes. If we make this
 * the size of the minors list then it should be nearly perfect
 */
#define ORIGIN_HASH_SIZE 256
#define ORIGIN_MASK      0xFF
static struct list_head *_origins;
static struct list_head *_dm_origins;
static struct rw_semaphore _origins_lock;

static DECLARE_WAIT_QUEUE_HEAD(_pending_exceptions_done);
static DEFINE_SPINLOCK(_pending_exceptions_done_spinlock);
static uint64_t _pending_exceptions_done_count;

static int init_origin_hash(void)
{
	int i;

	_origins = kmalloc_array(ORIGIN_HASH_SIZE, sizeof(struct list_head),
				 GFP_KERNEL);
	if (!_origins) {
		DMERR("unable to allocate memory for _origins");
		return -ENOMEM;
	}
	for (i = 0; i < ORIGIN_HASH_SIZE; i++)
		INIT_LIST_HEAD(_origins + i);

	_dm_origins = kmalloc_array(ORIGIN_HASH_SIZE,
				    sizeof(struct list_head),
				    GFP_KERNEL);
	if (!_dm_origins) {
		DMERR("unable to allocate memory for _dm_origins");
		kfree(_origins);
		return -ENOMEM;
	}
	for (i = 0; i < ORIGIN_HASH_SIZE; i++)
		INIT_LIST_HEAD(_dm_origins + i);

	init_rwsem(&_origins_lock);

	return 0;
}

static void exit_origin_hash(void)
{
	kfree(_origins);
	kfree(_dm_origins);
}

static unsigned origin_hash(struct block_device *bdev)
{
	return bdev->bd_dev & ORIGIN_MASK;
}

static struct origin *__lookup_origin(struct block_device *origin)
{
	struct list_head *ol;
	struct origin *o;

	ol = &_origins[origin_hash(origin)];
	list_for_each_entry (o, ol, hash_list)
		if (bdev_equal(o->bdev, origin))
			return o;

	return NULL;
}

static void __insert_origin(struct origin *o)
{
	struct list_head *sl = &_origins[origin_hash(o->bdev)];
	list_add_tail(&o->hash_list, sl);
}

static struct dm_origin *__lookup_dm_origin(struct block_device *origin)
{
	struct list_head *ol;
	struct dm_origin *o;

	ol = &_dm_origins[origin_hash(origin)];
	list_for_each_entry (o, ol, hash_list)
		if (bdev_equal(o->dev->bdev, origin))
			return o;

	return NULL;
}

static void __insert_dm_origin(struct dm_origin *o)
{
	struct list_head *sl = &_dm_origins[origin_hash(o->dev->bdev)];
	list_add_tail(&o->hash_list, sl);
}

static void __remove_dm_origin(struct dm_origin *o)
{
	list_del(&o->hash_list);
}

/*
 * _origins_lock must be held when calling this function.
 * Returns number of snapshots registered using the supplied cow device, plus:
 * snap_src - a snapshot suitable for use as a source of exception handover
 * snap_dest - a snapshot capable of receiving exception handover.
 * snap_merge - an existing snapshot-merge target linked to the same origin.
 *   There can be at most one snapshot-merge target. The parameter is optional.
 *
 * Possible return values and states of snap_src and snap_dest.
 *   0: NULL, NULL  - first new snapshot
 *   1: snap_src, NULL - normal snapshot
 *   2: snap_src, snap_dest  - waiting for handover
 *   2: snap_src, NULL - handed over, waiting for old to be deleted
 *   1: NULL, snap_dest - source got destroyed without handover
 */
static int __find_snapshots_sharing_cow(struct dm_snapshot *snap,
					struct dm_snapshot **snap_src,
					struct dm_snapshot **snap_dest,
					struct dm_snapshot **snap_merge)
{
	struct dm_snapshot *s;
	struct origin *o;
	int count = 0;
	int active;

	o = __lookup_origin(snap->origin->bdev);
	if (!o)
		goto out;

	list_for_each_entry(s, &o->snapshots, list) {
		if (dm_target_is_snapshot_merge(s->ti) && snap_merge)
			*snap_merge = s;
		if (!bdev_equal(s->cow->bdev, snap->cow->bdev))
			continue;

		mutex_lock(&s->lock);
		active = s->active;
		mutex_unlock(&s->lock);

		if (active) {
			if (snap_src)
				*snap_src = s;
		} else if (snap_dest)
			*snap_dest = s;

		count++;
	}

out:
	return count;
}

/*
 * On success, returns 1 if this snapshot is a handover destination,
 * otherwise returns 0.
 */
static int __validate_exception_handover(struct dm_snapshot *snap)
{
	struct dm_snapshot *snap_src = NULL, *snap_dest = NULL;
	struct dm_snapshot *snap_merge = NULL;

	/* Does snapshot need exceptions handed over to it? */
	if ((__find_snapshots_sharing_cow(snap, &snap_src, &snap_dest,
					  &snap_merge) == 2) ||
	    snap_dest) {
		snap->ti->error = "Snapshot cow pairing for exception "
				  "table handover failed";
		return -EINVAL;
	}

	/*
	 * If no snap_src was found, snap cannot become a handover
	 * destination.
	 */
	if (!snap_src)
		return 0;

	/*
	 * Non-snapshot-merge handover?
	 */
	if (!dm_target_is_snapshot_merge(snap->ti))
		return 1;

	/*
	 * Do not allow more than one merging snapshot.
	 */
	if (snap_merge) {
		snap->ti->error = "A snapshot is already merging.";
		return -EINVAL;
	}

	if (!snap_src->store->type->prepare_merge ||
	    !snap_src->store->type->commit_merge) {
		snap->ti->error = "Snapshot exception store does not "
				  "support snapshot-merge.";
		return -EINVAL;
	}

	return 1;
}

static void __insert_snapshot(struct origin *o, struct dm_snapshot *s)
{
	struct dm_snapshot *l;

	/* Sort the list according to chunk size, largest-first smallest-last */
	list_for_each_entry(l, &o->snapshots, list)
		if (l->store->chunk_size < s->store->chunk_size)
			break;
	list_add_tail(&s->list, &l->list);
}

/*
 * Make a note of the snapshot and its origin so we can look it
 * up when the origin has a write on it.
 *
 * Also validate snapshot exception store handovers.
 * On success, returns 1 if this registration is a handover destination,
 * otherwise returns 0.
 */
static int register_snapshot(struct dm_snapshot *snap)
{
	struct origin *o, *new_o = NULL;
	struct block_device *bdev = snap->origin->bdev;
	int r = 0;

	new_o = kmalloc(sizeof(*new_o), GFP_KERNEL);
	if (!new_o)
		return -ENOMEM;

	down_write(&_origins_lock);

	r = __validate_exception_handover(snap);
	if (r < 0) {
		kfree(new_o);
		goto out;
	}

	o = __lookup_origin(bdev);
	if (o)
		kfree(new_o);
	else {
		/* New origin */
		o = new_o;

		/* Initialise the struct */
		INIT_LIST_HEAD(&o->snapshots);
		o->bdev = bdev;

		__insert_origin(o);
	}

	__insert_snapshot(o, snap);

out:
	up_write(&_origins_lock);

	return r;
}

/*
 * Move snapshot to correct place in list according to chunk size.
 */
static void reregister_snapshot(struct dm_snapshot *s)
{
	struct block_device *bdev = s->origin->bdev;

	down_write(&_origins_lock);

	list_del(&s->list);
	__insert_snapshot(__lookup_origin(bdev), s);

	up_write(&_origins_lock);
}

static void unregister_snapshot(struct dm_snapshot *s)
{
	struct origin *o;

	down_write(&_origins_lock);
	o = __lookup_origin(s->origin->bdev);

	list_del(&s->list);
	if (o && list_empty(&o->snapshots)) {
		list_del(&o->hash_list);
		kfree(o);
	}

	up_write(&_origins_lock);
}

/*
 * Implementation of the exception hash tables.
 * The lowest hash_shift bits of the chunk number are ignored, allowing
 * some consecutive chunks to be grouped together.
 */
static int dm_exception_table_init(struct dm_exception_table *et,
				   uint32_t size, unsigned hash_shift)
{
	unsigned int i;

	et->hash_shift = hash_shift;
	et->hash_mask = size - 1;
	et->table = dm_vcalloc(size, sizeof(struct list_head));
	if (!et->table)
		return -ENOMEM;

	for (i = 0; i < size; i++)
		INIT_LIST_HEAD(et->table + i);

	return 0;
}

static void dm_exception_table_exit(struct dm_exception_table *et,
				    struct kmem_cache *mem)
{
	struct list_head *slot;
	struct dm_exception *ex, *next;
	int i, size;

	size = et->hash_mask + 1;
	for (i = 0; i < size; i++) {
		slot = et->table + i;

		list_for_each_entry_safe (ex, next, slot, hash_list)
			kmem_cache_free(mem, ex);
	}

	vfree(et->table);
}

static uint32_t exception_hash(struct dm_exception_table *et, chunk_t chunk)
{
	return (chunk >> et->hash_shift) & et->hash_mask;
}

static void dm_remove_exception(struct dm_exception *e)
{
	list_del(&e->hash_list);
}

/*
 * Return the exception data for a sector, or NULL if not
 * remapped.
 */
static struct dm_exception *dm_lookup_exception(struct dm_exception_table *et,
						chunk_t chunk)
{
	struct list_head *slot;
	struct dm_exception *e;

	slot = &et->table[exception_hash(et, chunk)];
	list_for_each_entry (e, slot, hash_list)
		if (chunk >= e->old_chunk &&
		    chunk <= e->old_chunk + dm_consecutive_chunk_count(e))
			return e;

	return NULL;
}

static struct dm_exception *alloc_completed_exception(gfp_t gfp)
{
	struct dm_exception *e;

	e = kmem_cache_alloc(exception_cache, gfp);
	if (!e && gfp == GFP_NOIO)
		e = kmem_cache_alloc(exception_cache, GFP_ATOMIC);

	return e;
}

static void free_completed_exception(struct dm_exception *e)
{
	kmem_cache_free(exception_cache, e);
}

static struct dm_snap_pending_exception *alloc_pending_exception(struct dm_snapshot *s)
{
	struct dm_snap_pending_exception *pe = mempool_alloc(&s->pending_pool,
							     GFP_NOIO);

	atomic_inc(&s->pending_exceptions_count);
	pe->snap = s;

	return pe;
}

static void free_pending_exception(struct dm_snap_pending_exception *pe)
{
	struct dm_snapshot *s = pe->snap;

	mempool_free(pe, &s->pending_pool);
	smp_mb__before_atomic();
	atomic_dec(&s->pending_exceptions_count);
}

static void dm_insert_exception(struct dm_exception_table *eh,
				struct dm_exception *new_e)
{
	struct list_head *l;
	struct dm_exception *e = NULL;

	l = &eh->table[exception_hash(eh, new_e->old_chunk)];

	/* Add immediately if this table doesn't support consecutive chunks */
	if (!eh->hash_shift)
		goto out;

	/* List is ordered by old_chunk */
	list_for_each_entry_reverse(e, l, hash_list) {
		/* Insert after an existing chunk? */
		if (new_e->old_chunk == (e->old_chunk +
					 dm_consecutive_chunk_count(e) + 1) &&
		    new_e->new_chunk == (dm_chunk_number(e->new_chunk) +
					 dm_consecutive_chunk_count(e) + 1)) {
			dm_consecutive_chunk_count_inc(e);
			free_completed_exception(new_e);
			return;
		}

		/* Insert before an existing chunk? */
		if (new_e->old_chunk == (e->old_chunk - 1) &&
		    new_e->new_chunk == (dm_chunk_number(e->new_chunk) - 1)) {
			dm_consecutive_chunk_count_inc(e);
			e->old_chunk--;
			e->new_chunk--;
			free_completed_exception(new_e);
			return;
		}

		if (new_e->old_chunk > e->old_chunk)
			break;
	}

out:
	list_add(&new_e->hash_list, e ? &e->hash_list : l);
}

/*
 * Callback used by the exception stores to load exceptions when
 * initialising.
 */
static int dm_add_exception(void *context, chunk_t old, chunk_t new)
{
	struct dm_snapshot *s = context;
	struct dm_exception *e;

	e = alloc_completed_exception(GFP_KERNEL);
	if (!e)
		return -ENOMEM;

	e->old_chunk = old;

	/* Consecutive_count is implicitly initialised to zero */
	e->new_chunk = new;

	dm_insert_exception(&s->complete, e);

	return 0;
}

/*
 * Return a minimum chunk size of all snapshots that have the specified origin.
 * Return zero if the origin has no snapshots.
 */
static uint32_t __minimum_chunk_size(struct origin *o)
{
	struct dm_snapshot *snap;
	unsigned chunk_size = 0;

	if (o)
		list_for_each_entry(snap, &o->snapshots, list)
			chunk_size = min_not_zero(chunk_size,
						  snap->store->chunk_size);

	return (uint32_t) chunk_size;
}

/*
 * Hard coded magic.
 */
static int calc_max_buckets(void)
{
	/* use a fixed size of 2MB */
	unsigned long mem = 2 * 1024 * 1024;
	mem /= sizeof(struct list_head);

	return mem;
}

/*
 * Allocate room for a suitable hash table.
 */
static int init_hash_tables(struct dm_snapshot *s)
{
	sector_t hash_size, cow_dev_size, max_buckets;

	/*
	 * Calculate based on the size of the original volume or
	 * the COW volume...
	 */
	cow_dev_size = get_dev_size(s->cow->bdev);
	max_buckets = calc_max_buckets();

	hash_size = cow_dev_size >> s->store->chunk_shift;
	hash_size = min(hash_size, max_buckets);

	if (hash_size < 64)
		hash_size = 64;
	hash_size = rounddown_pow_of_two(hash_size);
	if (dm_exception_table_init(&s->complete, hash_size,
				    DM_CHUNK_CONSECUTIVE_BITS))
		return -ENOMEM;

	/*
	 * Allocate hash table for in-flight exceptions
	 * Make this smaller than the real hash table
	 */
	hash_size >>= 3;
	if (hash_size < 64)
		hash_size = 64;

	if (dm_exception_table_init(&s->pending, hash_size, 0)) {
		dm_exception_table_exit(&s->complete, exception_cache);
		return -ENOMEM;
	}

	return 0;
}

static void merge_shutdown(struct dm_snapshot *s)
{
	clear_bit_unlock(RUNNING_MERGE, &s->state_bits);
	smp_mb__after_atomic();
	wake_up_bit(&s->state_bits, RUNNING_MERGE);
}

static struct bio *__release_queued_bios_after_merge(struct dm_snapshot *s)
{
	s->first_merging_chunk = 0;
	s->num_merging_chunks = 0;

	return bio_list_get(&s->bios_queued_during_merge);
}

/*
 * Remove one chunk from the index of completed exceptions.
 */
static int __remove_single_exception_chunk(struct dm_snapshot *s,
					   chunk_t old_chunk)
{
	struct dm_exception *e;

	e = dm_lookup_exception(&s->complete, old_chunk);
	if (!e) {
		DMERR("Corruption detected: exception for block %llu is "
		      "on disk but not in memory",
		      (unsigned long long)old_chunk);
		return -EINVAL;
	}

	/*
	 * If this is the only chunk using this exception, remove exception.
	 */
	if (!dm_consecutive_chunk_count(e)) {
		dm_remove_exception(e);
		free_completed_exception(e);
		return 0;
	}

	/*
	 * The chunk may be either at the beginning or the end of a
	 * group of consecutive chunks - never in the middle.  We are
	 * removing chunks in the opposite order to that in which they
	 * were added, so this should always be true.
	 * Decrement the consecutive chunk counter and adjust the
	 * starting point if necessary.
	 */
	if (old_chunk == e->old_chunk) {
		e->old_chunk++;
		e->new_chunk++;
	} else if (old_chunk != e->old_chunk +
		   dm_consecutive_chunk_count(e)) {
		DMERR("Attempt to merge block %llu from the "
		      "middle of a chunk range [%llu - %llu]",
		      (unsigned long long)old_chunk,
		      (unsigned long long)e->old_chunk,
		      (unsigned long long)
		      e->old_chunk + dm_consecutive_chunk_count(e));
		return -EINVAL;
	}

	dm_consecutive_chunk_count_dec(e);

	return 0;
}

static void flush_bios(struct bio *bio);

static int remove_single_exception_chunk(struct dm_snapshot *s)
{
	struct bio *b = NULL;
	int r;
	chunk_t old_chunk = s->first_merging_chunk + s->num_merging_chunks - 1;

	mutex_lock(&s->lock);

	/*
	 * Process chunks (and associated exceptions) in reverse order
	 * so that dm_consecutive_chunk_count_dec() accounting works.
	 */
	do {
		r = __remove_single_exception_chunk(s, old_chunk);
		if (r)
			goto out;
	} while (old_chunk-- > s->first_merging_chunk);

	b = __release_queued_bios_after_merge(s);

out:
	mutex_unlock(&s->lock);
	if (b)
		flush_bios(b);

	return r;
}

static int origin_write_extent(struct dm_snapshot *merging_snap,
			       sector_t sector, unsigned chunk_size);

static void merge_callback(int read_err, unsigned long write_err,
			   void *context);

static uint64_t read_pending_exceptions_done_count(void)
{
	uint64_t pending_exceptions_done;

	spin_lock(&_pending_exceptions_done_spinlock);
	pending_exceptions_done = _pending_exceptions_done_count;
	spin_unlock(&_pending_exceptions_done_spinlock);

	return pending_exceptions_done;
}

static void increment_pending_exceptions_done_count(void)
{
	spin_lock(&_pending_exceptions_done_spinlock);
	_pending_exceptions_done_count++;
	spin_unlock(&_pending_exceptions_done_spinlock);

	wake_up_all(&_pending_exceptions_done);
}

static void snapshot_merge_next_chunks(struct dm_snapshot *s)
{
	int i, linear_chunks;
	chunk_t old_chunk, new_chunk;
	struct dm_io_region src, dest;
	sector_t io_size;
	uint64_t previous_count;

	BUG_ON(!test_bit(RUNNING_MERGE, &s->state_bits));
	if (unlikely(test_bit(SHUTDOWN_MERGE, &s->state_bits)))
		goto shut;

	/*
	 * valid flag never changes during merge, so no lock required.
	 */
	if (!s->valid) {
		DMERR("Snapshot is invalid: can't merge");
		goto shut;
	}

	linear_chunks = s->store->type->prepare_merge(s->store, &old_chunk,
						      &new_chunk);
	if (linear_chunks <= 0) {
		if (linear_chunks < 0) {
			DMERR("Read error in exception store: "
			      "shutting down merge");
			mutex_lock(&s->lock);
			s->merge_failed = 1;
			mutex_unlock(&s->lock);
		}
		goto shut;
	}

	/* Adjust old_chunk and new_chunk to reflect start of linear region */
	old_chunk = old_chunk + 1 - linear_chunks;
	new_chunk = new_chunk + 1 - linear_chunks;

	/*
	 * Use one (potentially large) I/O to copy all 'linear_chunks'
	 * from the exception store to the origin
	 */
	io_size = linear_chunks * s->store->chunk_size;

	dest.bdev = s->origin->bdev;
	dest.sector = chunk_to_sector(s->store, old_chunk);
	dest.count = min(io_size, get_dev_size(dest.bdev) - dest.sector);

	src.bdev = s->cow->bdev;
	src.sector = chunk_to_sector(s->store, new_chunk);
	src.count = dest.count;

	/*
	 * Reallocate any exceptions needed in other snapshots then
	 * wait for the pending exceptions to complete.
	 * Each time any pending exception (globally on the system)
	 * completes we are woken and repeat the process to find out
	 * if we can proceed.  While this may not seem a particularly
	 * efficient algorithm, it is not expected to have any
	 * significant impact on performance.
	 */
	previous_count = read_pending_exceptions_done_count();
	while (origin_write_extent(s, dest.sector, io_size)) {
		wait_event(_pending_exceptions_done,
			   (read_pending_exceptions_done_count() !=
			    previous_count));
		/* Retry after the wait, until all exceptions are done. */
		previous_count = read_pending_exceptions_done_count();
	}

	mutex_lock(&s->lock);
	s->first_merging_chunk = old_chunk;
	s->num_merging_chunks = linear_chunks;
	mutex_unlock(&s->lock);

	/* Wait until writes to all 'linear_chunks' drain */
	for (i = 0; i < linear_chunks; i++)
		__check_for_conflicting_io(s, old_chunk + i);

	dm_kcopyd_copy(s->kcopyd_client, &src, 1, &dest, 0, merge_callback, s);
	return;

shut:
	merge_shutdown(s);
}

static void error_bios(struct bio *bio);

static void merge_callback(int read_err, unsigned long write_err, void *context)
{
	struct dm_snapshot *s = context;
	struct bio *b = NULL;

	if (read_err || write_err) {
		if (read_err)
			DMERR("Read error: shutting down merge.");
		else
			DMERR("Write error: shutting down merge.");
		goto shut;
	}

	if (s->store->type->commit_merge(s->store,
					 s->num_merging_chunks) < 0) {
		DMERR("Write error in exception store: shutting down merge");
		goto shut;
	}

	if (remove_single_exception_chunk(s) < 0)
		goto shut;

	snapshot_merge_next_chunks(s);

	return;

shut:
	mutex_lock(&s->lock);
	s->merge_failed = 1;
	b = __release_queued_bios_after_merge(s);
	mutex_unlock(&s->lock);
	error_bios(b);

	merge_shutdown(s);
}

static void start_merge(struct dm_snapshot *s)
{
	if (!test_and_set_bit(RUNNING_MERGE, &s->state_bits))
		snapshot_merge_next_chunks(s);
}

/*
 * Stop the merging process and wait until it finishes.
 */
static void stop_merge(struct dm_snapshot *s)
{
	set_bit(SHUTDOWN_MERGE, &s->state_bits);
	wait_on_bit(&s->state_bits, RUNNING_MERGE, TASK_UNINTERRUPTIBLE);
	clear_bit(SHUTDOWN_MERGE, &s->state_bits);
}

/*
 * Construct a snapshot mapping: <origin_dev> <COW-dev> <p|po|n> <chunk-size>
 */
static int snapshot_ctr(struct dm_target *ti, unsigned int argc, char **argv)
{
	struct dm_snapshot *s;
	int i;
	int r = -EINVAL;
	char *origin_path, *cow_path;
	dev_t origin_dev, cow_dev;
	unsigned args_used, num_flush_bios = 1;
	fmode_t origin_mode = FMODE_READ;

	if (argc != 4) {
		ti->error = "requires exactly 4 arguments";
		r = -EINVAL;
		goto bad;
	}

	if (dm_target_is_snapshot_merge(ti)) {
		num_flush_bios = 2;
		origin_mode = FMODE_WRITE;
	}

	s = kzalloc(sizeof(*s), GFP_KERNEL);
	if (!s) {
		ti->error = "Cannot allocate private snapshot structure";
		r = -ENOMEM;
		goto bad;
	}

	origin_path = argv[0];
	argv++;
	argc--;

	r = dm_get_device(ti, origin_path, origin_mode, &s->origin);
	if (r) {
		ti->error = "Cannot get origin device";
		goto bad_origin;
	}
	origin_dev = s->origin->bdev->bd_dev;

	cow_path = argv[0];
	argv++;
	argc--;

	cow_dev = dm_get_dev_t(cow_path);
	if (cow_dev && cow_dev == origin_dev) {
		ti->error = "COW device cannot be the same as origin device";
		r = -EINVAL;
		goto bad_cow;
	}

	r = dm_get_device(ti, cow_path, dm_table_get_mode(ti->table), &s->cow);
	if (r) {
		ti->error = "Cannot get COW device";
		goto bad_cow;
	}

	r = dm_exception_store_create(ti, argc, argv, s, &args_used, &s->store);
	if (r) {
		ti->error = "Couldn't create exception store";
		r = -EINVAL;
		goto bad_store;
	}

	argv += args_used;
	argc -= args_used;

	s->ti = ti;
	s->valid = 1;
	s->snapshot_overflowed = 0;
	s->active = 0;
	atomic_set(&s->pending_exceptions_count, 0);
	s->exception_start_sequence = 0;
	s->exception_complete_sequence = 0;
	s->out_of_order_tree = RB_ROOT;
	mutex_init(&s->lock);
	INIT_LIST_HEAD(&s->list);
	spin_lock_init(&s->pe_lock);
	s->state_bits = 0;
	s->merge_failed = 0;
	s->first_merging_chunk = 0;
	s->num_merging_chunks = 0;
	bio_list_init(&s->bios_queued_during_merge);

	/* Allocate hash table for COW data */
	if (init_hash_tables(s)) {
		ti->error = "Unable to allocate hash table space";
		r = -ENOMEM;
		goto bad_hash_tables;
	}

	sema_init(&s->cow_count, (cow_threshold > 0) ? cow_threshold : INT_MAX);

	s->kcopyd_client = dm_kcopyd_client_create(&dm_kcopyd_throttle);
	if (IS_ERR(s->kcopyd_client)) {
		r = PTR_ERR(s->kcopyd_client);
		ti->error = "Could not create kcopyd client";
		goto bad_kcopyd;
	}

	r = mempool_init_slab_pool(&s->pending_pool, MIN_IOS, pending_cache);
	if (r) {
		ti->error = "Could not allocate mempool for pending exceptions";
		goto bad_pending_pool;
	}

	for (i = 0; i < DM_TRACKED_CHUNK_HASH_SIZE; i++)
		INIT_HLIST_HEAD(&s->tracked_chunk_hash[i]);

	spin_lock_init(&s->tracked_chunk_lock);

	ti->private = s;
	ti->num_flush_bios = num_flush_bios;
	ti->per_io_data_size = sizeof(struct dm_snap_tracked_chunk);

	/* Add snapshot to the list of snapshots for this origin */
	/* Exceptions aren't triggered till snapshot_resume() is called */
	r = register_snapshot(s);
	if (r == -ENOMEM) {
		ti->error = "Snapshot origin struct allocation failed";
		goto bad_load_and_register;
	} else if (r < 0) {
		/* invalid handover, register_snapshot has set ti->error */
		goto bad_load_and_register;
	}

	/*
	 * Metadata must only be loaded into one table at once, so skip this
	 * if metadata will be handed over during resume.
	 * Chunk size will be set during the handover - set it to zero to
	 * ensure it's ignored.
	 */
	if (r > 0) {
		s->store->chunk_size = 0;
		return 0;
	}

	r = s->store->type->read_metadata(s->store, dm_add_exception,
					  (void *)s);
	if (r < 0) {
		ti->error = "Failed to read snapshot metadata";
		goto bad_read_metadata;
	} else if (r > 0) {
		s->valid = 0;
		DMWARN("Snapshot is marked invalid.");
	}

	if (!s->store->chunk_size) {
		ti->error = "Chunk size not set";
		goto bad_read_metadata;
	}

	r = dm_set_target_max_io_len(ti, s->store->chunk_size);
	if (r)
		goto bad_read_metadata;

	return 0;

bad_read_metadata:
	unregister_snapshot(s);

bad_load_and_register:
	mempool_exit(&s->pending_pool);

bad_pending_pool:
	dm_kcopyd_client_destroy(s->kcopyd_client);

bad_kcopyd:
	dm_exception_table_exit(&s->pending, pending_cache);
	dm_exception_table_exit(&s->complete, exception_cache);

bad_hash_tables:
	dm_exception_store_destroy(s->store);

bad_store:
	dm_put_device(ti, s->cow);

bad_cow:
	dm_put_device(ti, s->origin);

bad_origin:
	kfree(s);

bad:
	return r;
}

static void __free_exceptions(struct dm_snapshot *s)
{
	dm_kcopyd_client_destroy(s->kcopyd_client);
	s->kcopyd_client = NULL;

	dm_exception_table_exit(&s->pending, pending_cache);
	dm_exception_table_exit(&s->complete, exception_cache);
}

static void __handover_exceptions(struct dm_snapshot *snap_src,
				  struct dm_snapshot *snap_dest)
{
	union {
		struct dm_exception_table table_swap;
		struct dm_exception_store *store_swap;
	} u;

	/*
	 * Swap all snapshot context information between the two instances.
	 */
	u.table_swap = snap_dest->complete;
	snap_dest->complete = snap_src->complete;
	snap_src->complete = u.table_swap;

	u.store_swap = snap_dest->store;
	snap_dest->store = snap_src->store;
	snap_dest->store->userspace_supports_overflow = u.store_swap->userspace_supports_overflow;
	snap_src->store = u.store_swap;

	snap_dest->store->snap = snap_dest;
	snap_src->store->snap = snap_src;

	snap_dest->ti->max_io_len = snap_dest->store->chunk_size;
	snap_dest->valid = snap_src->valid;
	snap_dest->snapshot_overflowed = snap_src->snapshot_overflowed;

	/*
	 * Set source invalid to ensure it receives no further I/O.
	 */
	snap_src->valid = 0;
}

static void snapshot_dtr(struct dm_target *ti)
{
#ifdef CONFIG_DM_DEBUG
	int i;
#endif
	struct dm_snapshot *s = ti->private;
	struct dm_snapshot *snap_src = NULL, *snap_dest = NULL;

	down_read(&_origins_lock);
	/* Check whether exception handover must be cancelled */
	(void) __find_snapshots_sharing_cow(s, &snap_src, &snap_dest, NULL);
	if (snap_src && snap_dest && (s == snap_src)) {
		mutex_lock(&snap_dest->lock);
		snap_dest->valid = 0;
		mutex_unlock(&snap_dest->lock);
		DMERR("Cancelling snapshot handover.");
	}
	up_read(&_origins_lock);

	if (dm_target_is_snapshot_merge(ti))
		stop_merge(s);

	/* Prevent further origin writes from using this snapshot. */
	/* After this returns there can be no new kcopyd jobs. */
	unregister_snapshot(s);

	while (atomic_read(&s->pending_exceptions_count))
		msleep(1);
	/*
	 * Ensure instructions in mempool_exit aren't reordered
	 * before atomic_read.
	 */
	smp_mb();

#ifdef CONFIG_DM_DEBUG
	for (i = 0; i < DM_TRACKED_CHUNK_HASH_SIZE; i++)
		BUG_ON(!hlist_empty(&s->tracked_chunk_hash[i]));
#endif

	__free_exceptions(s);

	mempool_exit(&s->pending_pool);

	dm_exception_store_destroy(s->store);

	mutex_destroy(&s->lock);

	dm_put_device(ti, s->cow);

	dm_put_device(ti, s->origin);

	kfree(s);
}

/*
 * Flush a list of buffers.
 */
static void flush_bios(struct bio *bio)
{
	struct bio *n;

	while (bio) {
		n = bio->bi_next;
		bio->bi_next = NULL;
		generic_make_request(bio);
		bio = n;
	}
}

static int do_origin(struct dm_dev *origin, struct bio *bio);

/*
 * Flush a list of buffers.
 */
static void retry_origin_bios(struct dm_snapshot *s, struct bio *bio)
{
	struct bio *n;
	int r;

	while (bio) {
		n = bio->bi_next;
		bio->bi_next = NULL;
		r = do_origin(s->origin, bio);
		if (r == DM_MAPIO_REMAPPED)
			generic_make_request(bio);
		bio = n;
	}
}

/*
 * Error a list of buffers.
 */
static void error_bios(struct bio *bio)
{
	struct bio *n;

	while (bio) {
		n = bio->bi_next;
		bio->bi_next = NULL;
		bio_io_error(bio);
		bio = n;
	}
}

static void __invalidate_snapshot(struct dm_snapshot *s, int err)
{
	if (!s->valid)
		return;

	if (err == -EIO)
		DMERR("Invalidating snapshot: Error reading/writing.");
	else if (err == -ENOMEM)
		DMERR("Invalidating snapshot: Unable to allocate exception.");

	if (s->store->type->drop_snapshot)
		s->store->type->drop_snapshot(s->store);

	s->valid = 0;

	dm_table_event(s->ti->table);
}

static void pending_complete(void *context, int success)
{
	struct dm_snap_pending_exception *pe = context;
	struct dm_exception *e;
	struct dm_snapshot *s = pe->snap;
	struct bio *origin_bios = NULL;
	struct bio *snapshot_bios = NULL;
	struct bio *full_bio = NULL;
	int error = 0;

	if (!success) {
		/* Read/write error - snapshot is unusable */
		mutex_lock(&s->lock);
		__invalidate_snapshot(s, -EIO);
		error = 1;
		goto out;
	}

	e = alloc_completed_exception(GFP_NOIO);
	if (!e) {
		mutex_lock(&s->lock);
		__invalidate_snapshot(s, -ENOMEM);
		error = 1;
		goto out;
	}
	*e = pe->e;

	mutex_lock(&s->lock);
	if (!s->valid) {
		free_completed_exception(e);
		error = 1;
		goto out;
	}

	/* Check for conflicting reads */
	__check_for_conflicting_io(s, pe->e.old_chunk);

	/*
	 * Add a proper exception, and remove the
	 * in-flight exception from the list.
	 */
	dm_insert_exception(&s->complete, e);

out:
	dm_remove_exception(&pe->e);
	snapshot_bios = bio_list_get(&pe->snapshot_bios);
	origin_bios = bio_list_get(&pe->origin_bios);
	full_bio = pe->full_bio;
	if (full_bio)
		full_bio->bi_end_io = pe->full_bio_end_io;
	increment_pending_exceptions_done_count();

	mutex_unlock(&s->lock);

	/* Submit any pending write bios */
	if (error) {
		if (full_bio)
			bio_io_error(full_bio);
		error_bios(snapshot_bios);
	} else {
		if (full_bio)
			bio_endio(full_bio);
		flush_bios(snapshot_bios);
	}

	retry_origin_bios(s, origin_bios);

	free_pending_exception(pe);
}

static void complete_exception(struct dm_snap_pending_exception *pe)
{
	struct dm_snapshot *s = pe->snap;

	/* Update the metadata if we are persistent */
	s->store->type->commit_exception(s->store, &pe->e, !pe->copy_error,
					 pending_complete, pe);
}

/*
 * Called when the copy I/O has finished.  kcopyd actually runs
 * this code so don't block.
 */
static void copy_callback(int read_err, unsigned long write_err, void *context)
{
	struct dm_snap_pending_exception *pe = context;
	struct dm_snapshot *s = pe->snap;

	pe->copy_error = read_err || write_err;

	if (pe->exception_sequence == s->exception_complete_sequence) {
		struct rb_node *next;

		s->exception_complete_sequence++;
		complete_exception(pe);

		next = rb_first(&s->out_of_order_tree);
		while (next) {
			pe = rb_entry(next, struct dm_snap_pending_exception,
					out_of_order_node);
			if (pe->exception_sequence != s->exception_complete_sequence)
				break;
			next = rb_next(next);
			s->exception_complete_sequence++;
			rb_erase(&pe->out_of_order_node, &s->out_of_order_tree);
			complete_exception(pe);
			cond_resched();
		}
	} else {
		struct rb_node *parent = NULL;
		struct rb_node **p = &s->out_of_order_tree.rb_node;
		struct dm_snap_pending_exception *pe2;

		while (*p) {
			pe2 = rb_entry(*p, struct dm_snap_pending_exception, out_of_order_node);
			parent = *p;

			BUG_ON(pe->exception_sequence == pe2->exception_sequence);
			if (pe->exception_sequence < pe2->exception_sequence)
				p = &((*p)->rb_left);
			else
				p = &((*p)->rb_right);
		}

		rb_link_node(&pe->out_of_order_node, parent, p);
		rb_insert_color(&pe->out_of_order_node, &s->out_of_order_tree);
	}
	up(&s->cow_count);
}

/*
 * Dispatches the copy operation to kcopyd.
 */
static void start_copy(struct dm_snap_pending_exception *pe)
{
	struct dm_snapshot *s = pe->snap;
	struct dm_io_region src, dest;
	struct block_device *bdev = s->origin->bdev;
	sector_t dev_size;

	dev_size = get_dev_size(bdev);

	src.bdev = bdev;
	src.sector = chunk_to_sector(s->store, pe->e.old_chunk);
	src.count = min((sector_t)s->store->chunk_size, dev_size - src.sector);

	dest.bdev = s->cow->bdev;
	dest.sector = chunk_to_sector(s->store, pe->e.new_chunk);
	dest.count = src.count;

	/* Hand over to kcopyd */
	down(&s->cow_count);
	dm_kcopyd_copy(s->kcopyd_client, &src, 1, &dest, 0, copy_callback, pe);
}

static void full_bio_end_io(struct bio *bio)
{
	void *callback_data = bio->bi_private;

	dm_kcopyd_do_callback(callback_data, 0, bio->bi_status ? 1 : 0);
}

static void start_full_bio(struct dm_snap_pending_exception *pe,
			   struct bio *bio)
{
	struct dm_snapshot *s = pe->snap;
	void *callback_data;

	pe->full_bio = bio;
	pe->full_bio_end_io = bio->bi_end_io;

	down(&s->cow_count);
	callback_data = dm_kcopyd_prepare_callback(s->kcopyd_client,
						   copy_callback, pe);

	bio->bi_end_io = full_bio_end_io;
	bio->bi_private = callback_data;

	generic_make_request(bio);
}

static struct dm_snap_pending_exception *
__lookup_pending_exception(struct dm_snapshot *s, chunk_t chunk)
{
	struct dm_exception *e = dm_lookup_exception(&s->pending, chunk);

	if (!e)
		return NULL;

	return container_of(e, struct dm_snap_pending_exception, e);
}

/*
 * Looks to see if this snapshot already has a pending exception
 * for this chunk, otherwise it allocates a new one and inserts
 * it into the pending table.
 *
 * NOTE: a write lock must be held on snap->lock before calling
 * this.
 */
static struct dm_snap_pending_exception *
__find_pending_exception(struct dm_snapshot *s,
			 struct dm_snap_pending_exception *pe, chunk_t chunk)
{
	struct dm_snap_pending_exception *pe2;

	pe2 = __lookup_pending_exception(s, chunk);
	if (pe2) {
		free_pending_exception(pe);
		return pe2;
	}

	pe->e.old_chunk = chunk;
	bio_list_init(&pe->origin_bios);
	bio_list_init(&pe->snapshot_bios);
	pe->started = 0;
	pe->full_bio = NULL;

	if (s->store->type->prepare_exception(s->store, &pe->e)) {
		free_pending_exception(pe);
		return NULL;
	}

	pe->exception_sequence = s->exception_start_sequence++;

	dm_insert_exception(&s->pending, &pe->e);

	return pe;
}

static void remap_exception(struct dm_snapshot *s, struct dm_exception *e,
			    struct bio *bio, chunk_t chunk)
{
	bio_set_dev(bio, s->cow->bdev);
	bio->bi_iter.bi_sector =
		chunk_to_sector(s->store, dm_chunk_number(e->new_chunk) +
				(chunk - e->old_chunk)) +
		(bio->bi_iter.bi_sector & s->store->chunk_mask);
}

static int snapshot_map(struct dm_target *ti, struct bio *bio)
{
	struct dm_exception *e;
	struct dm_snapshot *s = ti->private;
	int r = DM_MAPIO_REMAPPED;
	chunk_t chunk;
	struct dm_snap_pending_exception *pe = NULL;

	init_tracked_chunk(bio);

	if (bio->bi_opf & REQ_PREFLUSH) {
		bio_set_dev(bio, s->cow->bdev);
		return DM_MAPIO_REMAPPED;
	}

	chunk = sector_to_chunk(s->store, bio->bi_iter.bi_sector);

	/* Full snapshots are not usable */
	/* To get here the table must be live so s->active is always set. */
	if (!s->valid)
		return DM_MAPIO_KILL;

	mutex_lock(&s->lock);

	if (!s->valid || (unlikely(s->snapshot_overflowed) &&
	    bio_data_dir(bio) == WRITE)) {
		r = DM_MAPIO_KILL;
		goto out_unlock;
	}

	/* If the block is already remapped - use that, else remap it */
	e = dm_lookup_exception(&s->complete, chunk);
	if (e) {
		remap_exception(s, e, bio, chunk);
		goto out_unlock;
	}

	/*
	 * Write to snapshot - higher level takes care of RW/RO
	 * flags so we should only get this if we are
	 * writeable.
	 */
	if (bio_data_dir(bio) == WRITE) {
		pe = __lookup_pending_exception(s, chunk);
		if (!pe) {
			mutex_unlock(&s->lock);
			pe = alloc_pending_exception(s);
			mutex_lock(&s->lock);

			if (!s->valid || s->snapshot_overflowed) {
				free_pending_exception(pe);
				r = DM_MAPIO_KILL;
				goto out_unlock;
			}

			e = dm_lookup_exception(&s->complete, chunk);
			if (e) {
				free_pending_exception(pe);
				remap_exception(s, e, bio, chunk);
				goto out_unlock;
			}

			pe = __find_pending_exception(s, pe, chunk);
			if (!pe) {
				if (s->store->userspace_supports_overflow) {
					s->snapshot_overflowed = 1;
					DMERR("Snapshot overflowed: Unable to allocate exception.");
				} else
					__invalidate_snapshot(s, -ENOMEM);
				r = DM_MAPIO_KILL;
				goto out_unlock;
			}
		}

		remap_exception(s, &pe->e, bio, chunk);

		r = DM_MAPIO_SUBMITTED;

		if (!pe->started &&
		    bio->bi_iter.bi_size ==
		    (s->store->chunk_size << SECTOR_SHIFT)) {
			pe->started = 1;
			mutex_unlock(&s->lock);
			start_full_bio(pe, bio);
			goto out;
		}

		bio_list_add(&pe->snapshot_bios, bio);

		if (!pe->started) {
			/* this is protected by snap->lock */
			pe->started = 1;
			mutex_unlock(&s->lock);
			start_copy(pe);
			goto out;
		}
	} else {
		bio_set_dev(bio, s->origin->bdev);
		track_chunk(s, bio, chunk);
	}

out_unlock:
	mutex_unlock(&s->lock);
out:
	return r;
}

/*
 * A snapshot-merge target behaves like a combination of a snapshot
 * target and a snapshot-origin target.  It only generates new
 * exceptions in other snapshots and not in the one that is being
 * merged.
 *
 * For each chunk, if there is an existing exception, it is used to
 * redirect I/O to the cow device.  Otherwise I/O is sent to the origin,
 * which in turn might generate exceptions in other snapshots.
 * If merging is currently taking place on the chunk in question, the
 * I/O is deferred by adding it to s->bios_queued_during_merge.
 */
static int snapshot_merge_map(struct dm_target *ti, struct bio *bio)
{
	struct dm_exception *e;
	struct dm_snapshot *s = ti->private;
	int r = DM_MAPIO_REMAPPED;
	chunk_t chunk;

	init_tracked_chunk(bio);

	if (bio->bi_opf & REQ_PREFLUSH) {
		if (!dm_bio_get_target_bio_nr(bio))
			bio_set_dev(bio, s->origin->bdev);
		else
			bio_set_dev(bio, s->cow->bdev);
		return DM_MAPIO_REMAPPED;
	}

	chunk = sector_to_chunk(s->store, bio->bi_iter.bi_sector);

	mutex_lock(&s->lock);

	/* Full merging snapshots are redirected to the origin */
	if (!s->valid)
		goto redirect_to_origin;

	/* If the block is already remapped - use that */
	e = dm_lookup_exception(&s->complete, chunk);
	if (e) {
		/* Queue writes overlapping with chunks being merged */
		if (bio_data_dir(bio) == WRITE &&
		    chunk >= s->first_merging_chunk &&
		    chunk < (s->first_merging_chunk +
			     s->num_merging_chunks)) {
			bio_set_dev(bio, s->origin->bdev);
			bio_list_add(&s->bios_queued_during_merge, bio);
			r = DM_MAPIO_SUBMITTED;
			goto out_unlock;
		}

		remap_exception(s, e, bio, chunk);

		if (bio_data_dir(bio) == WRITE)
			track_chunk(s, bio, chunk);
		goto out_unlock;
	}

redirect_to_origin:
	bio_set_dev(bio, s->origin->bdev);

	if (bio_data_dir(bio) == WRITE) {
		mutex_unlock(&s->lock);
		return do_origin(s->origin, bio);
	}

out_unlock:
	mutex_unlock(&s->lock);

	return r;
}

static int snapshot_end_io(struct dm_target *ti, struct bio *bio,
		blk_status_t *error)
{
	struct dm_snapshot *s = ti->private;

	if (is_bio_tracked(bio))
		stop_tracking_chunk(s, bio);

	return DM_ENDIO_DONE;
}

static void snapshot_merge_presuspend(struct dm_target *ti)
{
	struct dm_snapshot *s = ti->private;

	stop_merge(s);
}

static int snapshot_preresume(struct dm_target *ti)
{
	int r = 0;
	struct dm_snapshot *s = ti->private;
	struct dm_snapshot *snap_src = NULL, *snap_dest = NULL;

	down_read(&_origins_lock);
	(void) __find_snapshots_sharing_cow(s, &snap_src, &snap_dest, NULL);
	if (snap_src && snap_dest) {
		mutex_lock(&snap_src->lock);
		if (s == snap_src) {
			DMERR("Unable to resume snapshot source until "
			      "handover completes.");
			r = -EINVAL;
		} else if (!dm_suspended(snap_src->ti)) {
			DMERR("Unable to perform snapshot handover until "
			      "source is suspended.");
			r = -EINVAL;
		}
		mutex_unlock(&snap_src->lock);
	}
	up_read(&_origins_lock);

	return r;
}

static void snapshot_resume(struct dm_target *ti)
{
	struct dm_snapshot *s = ti->private;
	struct dm_snapshot *snap_src = NULL, *snap_dest = NULL, *snap_merging = NULL;
	struct dm_origin *o;
	struct mapped_device *origin_md = NULL;
	bool must_restart_merging = false;

	down_read(&_origins_lock);

	o = __lookup_dm_origin(s->origin->bdev);
	if (o)
		origin_md = dm_table_get_md(o->ti->table);
	if (!origin_md) {
		(void) __find_snapshots_sharing_cow(s, NULL, NULL, &snap_merging);
		if (snap_merging)
			origin_md = dm_table_get_md(snap_merging->ti->table);
	}
	if (origin_md == dm_table_get_md(ti->table))
		origin_md = NULL;
	if (origin_md) {
		if (dm_hold(origin_md))
			origin_md = NULL;
	}

	up_read(&_origins_lock);

	if (origin_md) {
		dm_internal_suspend_fast(origin_md);
		if (snap_merging && test_bit(RUNNING_MERGE, &snap_merging->state_bits)) {
			must_restart_merging = true;
			stop_merge(snap_merging);
		}
	}

	down_read(&_origins_lock);

	(void) __find_snapshots_sharing_cow(s, &snap_src, &snap_dest, NULL);
	if (snap_src && snap_dest) {
		mutex_lock(&snap_src->lock);
		mutex_lock_nested(&snap_dest->lock, SINGLE_DEPTH_NESTING);
		__handover_exceptions(snap_src, snap_dest);
		mutex_unlock(&snap_dest->lock);
		mutex_unlock(&snap_src->lock);
	}

	up_read(&_origins_lock);

	if (origin_md) {
		if (must_restart_merging)
			start_merge(snap_merging);
		dm_internal_resume_fast(origin_md);
		dm_put(origin_md);
	}

	/* Now we have correct chunk size, reregister */
	reregister_snapshot(s);

	mutex_lock(&s->lock);
	s->active = 1;
	mutex_unlock(&s->lock);
}

static uint32_t get_origin_minimum_chunksize(struct block_device *bdev)
{
	uint32_t min_chunksize;

	down_read(&_origins_lock);
	min_chunksize = __minimum_chunk_size(__lookup_origin(bdev));
	up_read(&_origins_lock);

	return min_chunksize;
}

static void snapshot_merge_resume(struct dm_target *ti)
{
	struct dm_snapshot *s = ti->private;

	/*
	 * Handover exceptions from existing snapshot.
	 */
	snapshot_resume(ti);

	/*
	 * snapshot-merge acts as an origin, so set ti->max_io_len
	 */
	ti->max_io_len = get_origin_minimum_chunksize(s->origin->bdev);

	start_merge(s);
}

static void snapshot_status(struct dm_target *ti, status_type_t type,
			    unsigned status_flags, char *result, unsigned maxlen)
{
	unsigned sz = 0;
	struct dm_snapshot *snap = ti->private;

	switch (type) {
	case STATUSTYPE_INFO:

		mutex_lock(&snap->lock);

		if (!snap->valid)
			DMEMIT("Invalid");
		else if (snap->merge_failed)
			DMEMIT("Merge failed");
		else if (snap->snapshot_overflowed)
			DMEMIT("Overflow");
		else {
			if (snap->store->type->usage) {
				sector_t total_sectors, sectors_allocated,
					 metadata_sectors;
				snap->store->type->usage(snap->store,
							 &total_sectors,
							 &sectors_allocated,
							 &metadata_sectors);
				DMEMIT("%llu/%llu %llu",
				       (unsigned long long)sectors_allocated,
				       (unsigned long long)total_sectors,
				       (unsigned long long)metadata_sectors);
			}
			else
				DMEMIT("Unknown");
		}

		mutex_unlock(&snap->lock);

		break;

	case STATUSTYPE_TABLE:
		/*
		 * kdevname returns a static pointer so we need
		 * to make private copies if the output is to
		 * make sense.
		 */
		DMEMIT("%s %s", snap->origin->name, snap->cow->name);
		snap->store->type->status(snap->store, type, result + sz,
					  maxlen - sz);
		break;
	}
}

static int snapshot_iterate_devices(struct dm_target *ti,
				    iterate_devices_callout_fn fn, void *data)
{
	struct dm_snapshot *snap = ti->private;
	int r;

	r = fn(ti, snap->origin, 0, ti->len, data);

	if (!r)
		r = fn(ti, snap->cow, 0, get_dev_size(snap->cow->bdev), data);

	return r;
}


/*-----------------------------------------------------------------
 * Origin methods
 *---------------------------------------------------------------*/

/*
 * If no exceptions need creating, DM_MAPIO_REMAPPED is returned and any
 * supplied bio was ignored.  The caller may submit it immediately.
 * (No remapping actually occurs as the origin is always a direct linear
 * map.)
 *
 * If further exceptions are required, DM_MAPIO_SUBMITTED is returned
 * and any supplied bio is added to a list to be submitted once all
 * the necessary exceptions exist.
 */
static int __origin_write(struct list_head *snapshots, sector_t sector,
			  struct bio *bio)
{
	int r = DM_MAPIO_REMAPPED;
	struct dm_snapshot *snap;
	struct dm_exception *e;
	struct dm_snap_pending_exception *pe;
	struct dm_snap_pending_exception *pe_to_start_now = NULL;
	struct dm_snap_pending_exception *pe_to_start_last = NULL;
	chunk_t chunk;

	/* Do all the snapshots on this origin */
	list_for_each_entry (snap, snapshots, list) {
		/*
		 * Don't make new exceptions in a merging snapshot
		 * because it has effectively been deleted
		 */
		if (dm_target_is_snapshot_merge(snap->ti))
			continue;

		mutex_lock(&snap->lock);

		/* Only deal with valid and active snapshots */
		if (!snap->valid || !snap->active)
			goto next_snapshot;

		/* Nothing to do if writing beyond end of snapshot */
		if (sector >= dm_table_get_size(snap->ti->table))
			goto next_snapshot;

		/*
		 * Remember, different snapshots can have
		 * different chunk sizes.
		 */
		chunk = sector_to_chunk(snap->store, sector);

		/*
		 * Check exception table to see if block
		 * is already remapped in this snapshot
		 * and trigger an exception if not.
		 */
		e = dm_lookup_exception(&snap->complete, chunk);
		if (e)
			goto next_snapshot;

		pe = __lookup_pending_exception(snap, chunk);
		if (!pe) {
			mutex_unlock(&snap->lock);
			pe = alloc_pending_exception(snap);
			mutex_lock(&snap->lock);

			if (!snap->valid) {
				free_pending_exception(pe);
				goto next_snapshot;
			}

			e = dm_lookup_exception(&snap->complete, chunk);
			if (e) {
				free_pending_exception(pe);
				goto next_snapshot;
			}

			pe = __find_pending_exception(snap, pe, chunk);
			if (!pe) {
				__invalidate_snapshot(snap, -ENOMEM);
				goto next_snapshot;
			}
		}

		r = DM_MAPIO_SUBMITTED;

		/*
		 * If an origin bio was supplied, queue it to wait for the
		 * completion of this exception, and start this one last,
		 * at the end of the function.
		 */
		if (bio) {
			bio_list_add(&pe->origin_bios, bio);
			bio = NULL;

			if (!pe->started) {
				pe->started = 1;
				pe_to_start_last = pe;
			}
		}

		if (!pe->started) {
			pe->started = 1;
			pe_to_start_now = pe;
		}

next_snapshot:
		mutex_unlock(&snap->lock);

		if (pe_to_start_now) {
			start_copy(pe_to_start_now);
			pe_to_start_now = NULL;
		}
	}

	/*
	 * Submit the exception against which the bio is queued last,
	 * to give the other exceptions a head start.
	 */
	if (pe_to_start_last)
		start_copy(pe_to_start_last);

	return r;
}

/*
 * Called on a write from the origin driver.
 */
static int do_origin(struct dm_dev *origin, struct bio *bio)
{
	struct origin *o;
	int r = DM_MAPIO_REMAPPED;

	down_read(&_origins_lock);
	o = __lookup_origin(origin->bdev);
	if (o)
		r = __origin_write(&o->snapshots, bio->bi_iter.bi_sector, bio);
	up_read(&_origins_lock);

	return r;
}

/*
 * Trigger exceptions in all non-merging snapshots.
 *
 * The chunk size of the merging snapshot may be larger than the chunk
 * size of some other snapshot so we may need to reallocate multiple
 * chunks in other snapshots.
 *
 * We scan all the overlapping exceptions in the other snapshots.
 * Returns 1 if anything was reallocated and must be waited for,
 * otherwise returns 0.
 *
 * size must be a multiple of merging_snap's chunk_size.
 */
static int origin_write_extent(struct dm_snapshot *merging_snap,
			       sector_t sector, unsigned size)
{
	int must_wait = 0;
	sector_t n;
	struct origin *o;

	/*
	 * The origin's __minimum_chunk_size() got stored in max_io_len
	 * by snapshot_merge_resume().
	 */
	down_read(&_origins_lock);
	o = __lookup_origin(merging_snap->origin->bdev);
	for (n = 0; n < size; n += merging_snap->ti->max_io_len)
		if (__origin_write(&o->snapshots, sector + n, NULL) ==
		    DM_MAPIO_SUBMITTED)
			must_wait = 1;
	up_read(&_origins_lock);

	return must_wait;
}

/*
 * Origin: maps a linear range of a device, with hooks for snapshotting.
 */

/*
 * Construct an origin mapping: <dev_path>
 * The context for an origin is merely a 'struct dm_dev *'
 * pointing to the real device.
 */
static int origin_ctr(struct dm_target *ti, unsigned int argc, char **argv)
{
	int r;
	struct dm_origin *o;

	if (argc != 1) {
		ti->error = "origin: incorrect number of arguments";
		return -EINVAL;
	}

	o = kmalloc(sizeof(struct dm_origin), GFP_KERNEL);
	if (!o) {
		ti->error = "Cannot allocate private origin structure";
		r = -ENOMEM;
		goto bad_alloc;
	}

	r = dm_get_device(ti, argv[0], dm_table_get_mode(ti->table), &o->dev);
	if (r) {
		ti->error = "Cannot get target device";
		goto bad_open;
	}

	o->ti = ti;
	ti->private = o;
	ti->num_flush_bios = 1;

	return 0;

bad_open:
	kfree(o);
bad_alloc:
	return r;
}

static void origin_dtr(struct dm_target *ti)
{
	struct dm_origin *o = ti->private;

	dm_put_device(ti, o->dev);
	kfree(o);
}

static int origin_map(struct dm_target *ti, struct bio *bio)
{
	struct dm_origin *o = ti->private;
	unsigned available_sectors;

	bio_set_dev(bio, o->dev->bdev);

	if (unlikely(bio->bi_opf & REQ_PREFLUSH))
		return DM_MAPIO_REMAPPED;

	if (bio_data_dir(bio) != WRITE)
		return DM_MAPIO_REMAPPED;

	available_sectors = o->split_boundary -
		((unsigned)bio->bi_iter.bi_sector & (o->split_boundary - 1));

	if (bio_sectors(bio) > available_sectors)
		dm_accept_partial_bio(bio, available_sectors);

	/* Only tell snapshots if this is a write */
	return do_origin(o->dev, bio);
}

static long origin_dax_direct_access(struct dm_target *ti, pgoff_t pgoff,
		long nr_pages, void **kaddr, pfn_t *pfn)
{
	DMWARN("device does not support dax.");
	return -EIO;
}

/*
 * Set the target "max_io_len" field to the minimum of all the snapshots'
 * chunk sizes.
 */
static void origin_resume(struct dm_target *ti)
{
	struct dm_origin *o = ti->private;

	o->split_boundary = get_origin_minimum_chunksize(o->dev->bdev);

	down_write(&_origins_lock);
	__insert_dm_origin(o);
	up_write(&_origins_lock);
}

static void origin_postsuspend(struct dm_target *ti)
{
	struct dm_origin *o = ti->private;

	down_write(&_origins_lock);
	__remove_dm_origin(o);
	up_write(&_origins_lock);
}

static void origin_status(struct dm_target *ti, status_type_t type,
			  unsigned status_flags, char *result, unsigned maxlen)
{
	struct dm_origin *o = ti->private;

	switch (type) {
	case STATUSTYPE_INFO:
		result[0] = '\0';
		break;

	case STATUSTYPE_TABLE:
		snprintf(result, maxlen, "%s", o->dev->name);
		break;
	}
}

static int origin_iterate_devices(struct dm_target *ti,
				  iterate_devices_callout_fn fn, void *data)
{
	struct dm_origin *o = ti->private;

	return fn(ti, o->dev, 0, ti->len, data);
}

static struct target_type origin_target = {
	.name    = "snapshot-origin",
	.version = {1, 9, 0},
	.module  = THIS_MODULE,
	.ctr     = origin_ctr,
	.dtr     = origin_dtr,
	.map     = origin_map,
	.resume  = origin_resume,
	.postsuspend = origin_postsuspend,
	.status  = origin_status,
	.iterate_devices = origin_iterate_devices,
	.direct_access = origin_dax_direct_access,
};

static struct target_type snapshot_target = {
	.name    = "snapshot",
	.version = {1, 15, 0},
	.module  = THIS_MODULE,
	.ctr     = snapshot_ctr,
	.dtr     = snapshot_dtr,
	.map     = snapshot_map,
	.end_io  = snapshot_end_io,
	.preresume  = snapshot_preresume,
	.resume  = snapshot_resume,
	.status  = snapshot_status,
	.iterate_devices = snapshot_iterate_devices,
};

static struct target_type merge_target = {
	.name    = dm_snapshot_merge_target_name,
	.version = {1, 4, 0},
	.module  = THIS_MODULE,
	.ctr     = snapshot_ctr,
	.dtr     = snapshot_dtr,
	.map     = snapshot_merge_map,
	.end_io  = snapshot_end_io,
	.presuspend = snapshot_merge_presuspend,
	.preresume  = snapshot_preresume,
	.resume  = snapshot_merge_resume,
	.status  = snapshot_status,
	.iterate_devices = snapshot_iterate_devices,
};

static int __init dm_snapshot_init(void)
{
	int r;

	r = dm_exception_store_init();
	if (r) {
		DMERR("Failed to initialize exception stores");
		return r;
	}

	r = init_origin_hash();
	if (r) {
		DMERR("init_origin_hash failed.");
		goto bad_origin_hash;
	}

	exception_cache = KMEM_CACHE(dm_exception, 0);
	if (!exception_cache) {
		DMERR("Couldn't create exception cache.");
		r = -ENOMEM;
		goto bad_exception_cache;
	}

	pending_cache = KMEM_CACHE(dm_snap_pending_exception, 0);
	if (!pending_cache) {
		DMERR("Couldn't create pending cache.");
		r = -ENOMEM;
		goto bad_pending_cache;
	}

	r = dm_register_target(&snapshot_target);
	if (r < 0) {
		DMERR("snapshot target register failed %d", r);
		goto bad_register_snapshot_target;
	}

	r = dm_register_target(&origin_target);
	if (r < 0) {
		DMERR("Origin target register failed %d", r);
		goto bad_register_origin_target;
	}

	r = dm_register_target(&merge_target);
	if (r < 0) {
		DMERR("Merge target register failed %d", r);
		goto bad_register_merge_target;
	}

	return 0;

bad_register_merge_target:
	dm_unregister_target(&origin_target);
bad_register_origin_target:
	dm_unregister_target(&snapshot_target);
bad_register_snapshot_target:
	kmem_cache_destroy(pending_cache);
bad_pending_cache:
	kmem_cache_destroy(exception_cache);
bad_exception_cache:
	exit_origin_hash();
bad_origin_hash:
	dm_exception_store_exit();

	return r;
}

static void __exit dm_snapshot_exit(void)
{
	dm_unregister_target(&snapshot_target);
	dm_unregister_target(&origin_target);
	dm_unregister_target(&merge_target);

	exit_origin_hash();
	kmem_cache_destroy(pending_cache);
	kmem_cache_destroy(exception_cache);

	dm_exception_store_exit();
}

/* Module hooks */
module_init(dm_snapshot_init);
module_exit(dm_snapshot_exit);

MODULE_DESCRIPTION(DM_NAME " snapshot target");
MODULE_AUTHOR("Joe Thornber");
MODULE_LICENSE("GPL");
MODULE_ALIAS("dm-snapshot-origin");
MODULE_ALIAS("dm-snapshot-merge");