aboutsummaryrefslogtreecommitdiffstats
path: root/drivers/block/rbd.c
blob: 8f428a8ab003d8c7a029036eea6a1666a96d7dd3 (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
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
/*
   rbd.c -- Export ceph rados objects as a Linux block device


   based on drivers/block/osdblk.c:

   Copyright 2009 Red Hat, Inc.

   This program is free software; you can redistribute it and/or modify
   it under the terms of the GNU General Public License as published by
   the Free Software Foundation.

   This program is distributed in the hope that it will be useful,
   but WITHOUT ANY WARRANTY; without even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   GNU General Public License for more details.

   You should have received a copy of the GNU General Public License
   along with this program; see the file COPYING.  If not, write to
   the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA.



   For usage instructions, please refer to:

                 Documentation/ABI/testing/sysfs-bus-rbd

 */

#include <linux/ceph/libceph.h>
#include <linux/ceph/osd_client.h>
#include <linux/ceph/mon_client.h>
#include <linux/ceph/decode.h>
#include <linux/parser.h>

#include <linux/kernel.h>
#include <linux/device.h>
#include <linux/module.h>
#include <linux/fs.h>
#include <linux/blkdev.h>

#include "rbd_types.h"

/*
 * The basic unit of block I/O is a sector.  It is interpreted in a
 * number of contexts in Linux (blk, bio, genhd), but the default is
 * universally 512 bytes.  These symbols are just slightly more
 * meaningful than the bare numbers they represent.
 */
#define	SECTOR_SHIFT	9
#define	SECTOR_SIZE	(1ULL << SECTOR_SHIFT)

#define RBD_DRV_NAME "rbd"
#define RBD_DRV_NAME_LONG "rbd (rados block device)"

#define RBD_MINORS_PER_MAJOR	256		/* max minors per blkdev */

#define RBD_MAX_MD_NAME_LEN	(RBD_MAX_OBJ_NAME_LEN + sizeof(RBD_SUFFIX))
#define RBD_MAX_POOL_NAME_LEN	64
#define RBD_MAX_SNAP_NAME_LEN	32
#define RBD_MAX_OPT_LEN		1024

#define RBD_SNAP_HEAD_NAME	"-"

/*
 * An RBD device name will be "rbd#", where the "rbd" comes from
 * RBD_DRV_NAME above, and # is a unique integer identifier.
 * MAX_INT_FORMAT_WIDTH is used in ensuring DEV_NAME_LEN is big
 * enough to hold all possible device names.
 */
#define DEV_NAME_LEN		32
#define MAX_INT_FORMAT_WIDTH	((5 * sizeof (int)) / 2 + 1)

#define RBD_NOTIFY_TIMEOUT_DEFAULT 10

/*
 * block device image metadata (in-memory version)
 */
struct rbd_image_header {
	u64 image_size;
	char block_name[32];
	__u8 obj_order;
	__u8 crypt_type;
	__u8 comp_type;
	struct ceph_snap_context *snapc;
	size_t snap_names_len;
	u64 snap_seq;
	u32 total_snaps;

	char *snap_names;
	u64 *snap_sizes;

	u64 obj_version;
};

struct rbd_options {
	int	notify_timeout;
};

/*
 * an instance of the client.  multiple devices may share an rbd client.
 */
struct rbd_client {
	struct ceph_client	*client;
	struct rbd_options	*rbd_opts;
	struct kref		kref;
	struct list_head	node;
};

/*
 * a request completion status
 */
struct rbd_req_status {
	int done;
	int rc;
	u64 bytes;
};

/*
 * a collection of requests
 */
struct rbd_req_coll {
	int			total;
	int			num_done;
	struct kref		kref;
	struct rbd_req_status	status[0];
};

/*
 * a single io request
 */
struct rbd_request {
	struct request		*rq;		/* blk layer request */
	struct bio		*bio;		/* cloned bio */
	struct page		**pages;	/* list of used pages */
	u64			len;
	int			coll_index;
	struct rbd_req_coll	*coll;
};

struct rbd_snap {
	struct	device		dev;
	const char		*name;
	u64			size;
	struct list_head	node;
	u64			id;
};

/*
 * a single device
 */
struct rbd_device {
	int			id;		/* blkdev unique id */

	int			major;		/* blkdev assigned major */
	struct gendisk		*disk;		/* blkdev's gendisk and rq */
	struct request_queue	*q;

	struct rbd_client	*rbd_client;

	char			name[DEV_NAME_LEN]; /* blkdev name, e.g. rbd3 */

	spinlock_t		lock;		/* queue lock */

	struct rbd_image_header	header;
	char			obj[RBD_MAX_OBJ_NAME_LEN]; /* rbd image name */
	int			obj_len;
	char			obj_md_name[RBD_MAX_MD_NAME_LEN]; /* hdr nm. */
	char			pool_name[RBD_MAX_POOL_NAME_LEN];
	int			poolid;

	struct ceph_osd_event   *watch_event;
	struct ceph_osd_request *watch_request;

	/* protects updating the header */
	struct rw_semaphore     header_rwsem;
	char                    snap_name[RBD_MAX_SNAP_NAME_LEN];
	u64                     snap_id;	/* current snapshot id */
	int read_only;

	struct list_head	node;

	/* list of snapshots */
	struct list_head	snaps;

	/* sysfs related */
	struct device		dev;
};

static DEFINE_MUTEX(ctl_mutex);	  /* Serialize open/close/setup/teardown */

static LIST_HEAD(rbd_dev_list);    /* devices */
static DEFINE_SPINLOCK(rbd_dev_list_lock);

static LIST_HEAD(rbd_client_list);		/* clients */
static DEFINE_SPINLOCK(rbd_client_list_lock);

static int __rbd_init_snaps_header(struct rbd_device *rbd_dev);
static void rbd_dev_release(struct device *dev);
static ssize_t rbd_snap_add(struct device *dev,
			    struct device_attribute *attr,
			    const char *buf,
			    size_t count);
static void __rbd_remove_snap_dev(struct rbd_device *rbd_dev,
				  struct rbd_snap *snap);

static ssize_t rbd_add(struct bus_type *bus, const char *buf,
		       size_t count);
static ssize_t rbd_remove(struct bus_type *bus, const char *buf,
			  size_t count);

static struct bus_attribute rbd_bus_attrs[] = {
	__ATTR(add, S_IWUSR, NULL, rbd_add),
	__ATTR(remove, S_IWUSR, NULL, rbd_remove),
	__ATTR_NULL
};

static struct bus_type rbd_bus_type = {
	.name		= "rbd",
	.bus_attrs	= rbd_bus_attrs,
};

static void rbd_root_dev_release(struct device *dev)
{
}

static struct device rbd_root_dev = {
	.init_name =    "rbd",
	.release =      rbd_root_dev_release,
};


static struct device *rbd_get_dev(struct rbd_device *rbd_dev)
{
	return get_device(&rbd_dev->dev);
}

static void rbd_put_dev(struct rbd_device *rbd_dev)
{
	put_device(&rbd_dev->dev);
}

static int __rbd_refresh_header(struct rbd_device *rbd_dev);

static int rbd_open(struct block_device *bdev, fmode_t mode)
{
	struct rbd_device *rbd_dev = bdev->bd_disk->private_data;

	rbd_get_dev(rbd_dev);

	set_device_ro(bdev, rbd_dev->read_only);

	if ((mode & FMODE_WRITE) && rbd_dev->read_only)
		return -EROFS;

	return 0;
}

static int rbd_release(struct gendisk *disk, fmode_t mode)
{
	struct rbd_device *rbd_dev = disk->private_data;

	rbd_put_dev(rbd_dev);

	return 0;
}

static const struct block_device_operations rbd_bd_ops = {
	.owner			= THIS_MODULE,
	.open			= rbd_open,
	.release		= rbd_release,
};

/*
 * Initialize an rbd client instance.
 * We own *opt.
 */
static struct rbd_client *rbd_client_create(struct ceph_options *opt,
					    struct rbd_options *rbd_opts)
{
	struct rbd_client *rbdc;
	int ret = -ENOMEM;

	dout("rbd_client_create\n");
	rbdc = kmalloc(sizeof(struct rbd_client), GFP_KERNEL);
	if (!rbdc)
		goto out_opt;

	kref_init(&rbdc->kref);
	INIT_LIST_HEAD(&rbdc->node);

	mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);

	rbdc->client = ceph_create_client(opt, rbdc, 0, 0);
	if (IS_ERR(rbdc->client))
		goto out_mutex;
	opt = NULL; /* Now rbdc->client is responsible for opt */

	ret = ceph_open_session(rbdc->client);
	if (ret < 0)
		goto out_err;

	rbdc->rbd_opts = rbd_opts;

	spin_lock(&rbd_client_list_lock);
	list_add_tail(&rbdc->node, &rbd_client_list);
	spin_unlock(&rbd_client_list_lock);

	mutex_unlock(&ctl_mutex);

	dout("rbd_client_create created %p\n", rbdc);
	return rbdc;

out_err:
	ceph_destroy_client(rbdc->client);
out_mutex:
	mutex_unlock(&ctl_mutex);
	kfree(rbdc);
out_opt:
	if (opt)
		ceph_destroy_options(opt);
	return ERR_PTR(ret);
}

/*
 * Find a ceph client with specific addr and configuration.
 */
static struct rbd_client *__rbd_client_find(struct ceph_options *opt)
{
	struct rbd_client *client_node;

	if (opt->flags & CEPH_OPT_NOSHARE)
		return NULL;

	list_for_each_entry(client_node, &rbd_client_list, node)
		if (ceph_compare_options(opt, client_node->client) == 0)
			return client_node;
	return NULL;
}

/*
 * mount options
 */
enum {
	Opt_notify_timeout,
	Opt_last_int,
	/* int args above */
	Opt_last_string,
	/* string args above */
};

static match_table_t rbdopt_tokens = {
	{Opt_notify_timeout, "notify_timeout=%d"},
	/* int args above */
	/* string args above */
	{-1, NULL}
};

static int parse_rbd_opts_token(char *c, void *private)
{
	struct rbd_options *rbdopt = private;
	substring_t argstr[MAX_OPT_ARGS];
	int token, intval, ret;

	token = match_token(c, rbdopt_tokens, argstr);
	if (token < 0)
		return -EINVAL;

	if (token < Opt_last_int) {
		ret = match_int(&argstr[0], &intval);
		if (ret < 0) {
			pr_err("bad mount option arg (not int) "
			       "at '%s'\n", c);
			return ret;
		}
		dout("got int token %d val %d\n", token, intval);
	} else if (token > Opt_last_int && token < Opt_last_string) {
		dout("got string token %d val %s\n", token,
		     argstr[0].from);
	} else {
		dout("got token %d\n", token);
	}

	switch (token) {
	case Opt_notify_timeout:
		rbdopt->notify_timeout = intval;
		break;
	default:
		BUG_ON(token);
	}
	return 0;
}

/*
 * Get a ceph client with specific addr and configuration, if one does
 * not exist create it.
 */
static struct rbd_client *rbd_get_client(const char *mon_addr,
					 size_t mon_addr_len,
					 char *options)
{
	struct rbd_client *rbdc;
	struct ceph_options *opt;
	struct rbd_options *rbd_opts;

	rbd_opts = kzalloc(sizeof(*rbd_opts), GFP_KERNEL);
	if (!rbd_opts)
		return ERR_PTR(-ENOMEM);

	rbd_opts->notify_timeout = RBD_NOTIFY_TIMEOUT_DEFAULT;

	opt = ceph_parse_options(options, mon_addr,
				mon_addr + mon_addr_len,
				parse_rbd_opts_token, rbd_opts);
	if (IS_ERR(opt)) {
		kfree(rbd_opts);
		return ERR_CAST(opt);
	}

	spin_lock(&rbd_client_list_lock);
	rbdc = __rbd_client_find(opt);
	if (rbdc) {
		/* using an existing client */
		kref_get(&rbdc->kref);
		spin_unlock(&rbd_client_list_lock);

		ceph_destroy_options(opt);
		kfree(rbd_opts);

		return rbdc;
	}
	spin_unlock(&rbd_client_list_lock);

	rbdc = rbd_client_create(opt, rbd_opts);

	if (IS_ERR(rbdc))
		kfree(rbd_opts);

	return rbdc;
}

/*
 * Destroy ceph client
 *
 * Caller must hold rbd_client_list_lock.
 */
static void rbd_client_release(struct kref *kref)
{
	struct rbd_client *rbdc = container_of(kref, struct rbd_client, kref);

	dout("rbd_release_client %p\n", rbdc);
	spin_lock(&rbd_client_list_lock);
	list_del(&rbdc->node);
	spin_unlock(&rbd_client_list_lock);

	ceph_destroy_client(rbdc->client);
	kfree(rbdc->rbd_opts);
	kfree(rbdc);
}

/*
 * Drop reference to ceph client node. If it's not referenced anymore, release
 * it.
 */
static void rbd_put_client(struct rbd_device *rbd_dev)
{
	kref_put(&rbd_dev->rbd_client->kref, rbd_client_release);
	rbd_dev->rbd_client = NULL;
}

/*
 * Destroy requests collection
 */
static void rbd_coll_release(struct kref *kref)
{
	struct rbd_req_coll *coll =
		container_of(kref, struct rbd_req_coll, kref);

	dout("rbd_coll_release %p\n", coll);
	kfree(coll);
}

/*
 * Create a new header structure, translate header format from the on-disk
 * header.
 */
static int rbd_header_from_disk(struct rbd_image_header *header,
				 struct rbd_image_header_ondisk *ondisk,
				 u32 allocated_snaps,
				 gfp_t gfp_flags)
{
	u32 i, snap_count;

	if (memcmp(ondisk, RBD_HEADER_TEXT, sizeof(RBD_HEADER_TEXT)))
		return -ENXIO;

	snap_count = le32_to_cpu(ondisk->snap_count);
	if (snap_count > (UINT_MAX - sizeof(struct ceph_snap_context))
			 / sizeof (*ondisk))
		return -EINVAL;
	header->snapc = kmalloc(sizeof(struct ceph_snap_context) +
				snap_count * sizeof(u64),
				gfp_flags);
	if (!header->snapc)
		return -ENOMEM;

	header->snap_names_len = le64_to_cpu(ondisk->snap_names_len);
	if (snap_count) {
		header->snap_names = kmalloc(header->snap_names_len,
					     gfp_flags);
		if (!header->snap_names)
			goto err_snapc;
		header->snap_sizes = kmalloc(snap_count * sizeof(u64),
					     gfp_flags);
		if (!header->snap_sizes)
			goto err_names;
	} else {
		header->snap_names = NULL;
		header->snap_sizes = NULL;
	}
	memcpy(header->block_name, ondisk->block_name,
	       sizeof(ondisk->block_name));

	header->image_size = le64_to_cpu(ondisk->image_size);
	header->obj_order = ondisk->options.order;
	header->crypt_type = ondisk->options.crypt_type;
	header->comp_type = ondisk->options.comp_type;

	atomic_set(&header->snapc->nref, 1);
	header->snap_seq = le64_to_cpu(ondisk->snap_seq);
	header->snapc->num_snaps = snap_count;
	header->total_snaps = snap_count;

	if (snap_count && allocated_snaps == snap_count) {
		for (i = 0; i < snap_count; i++) {
			header->snapc->snaps[i] =
				le64_to_cpu(ondisk->snaps[i].id);
			header->snap_sizes[i] =
				le64_to_cpu(ondisk->snaps[i].image_size);
		}

		/* copy snapshot names */
		memcpy(header->snap_names, &ondisk->snaps[i],
			header->snap_names_len);
	}

	return 0;

err_names:
	kfree(header->snap_names);
err_snapc:
	kfree(header->snapc);
	return -ENOMEM;
}

static int snap_by_name(struct rbd_image_header *header, const char *snap_name,
			u64 *seq, u64 *size)
{
	int i;
	char *p = header->snap_names;

	for (i = 0; i < header->total_snaps; i++) {
		if (!strcmp(snap_name, p)) {

			/* Found it.  Pass back its id and/or size */

			if (seq)
				*seq = header->snapc->snaps[i];
			if (size)
				*size = header->snap_sizes[i];
			return i;
		}
		p += strlen(p) + 1;	/* Skip ahead to the next name */
	}
	return -ENOENT;
}

static int rbd_header_set_snap(struct rbd_device *dev, u64 *size)
{
	struct rbd_image_header *header = &dev->header;
	struct ceph_snap_context *snapc = header->snapc;
	int ret = -ENOENT;

	BUILD_BUG_ON(sizeof (dev->snap_name) < sizeof (RBD_SNAP_HEAD_NAME));

	down_write(&dev->header_rwsem);

	if (!memcmp(dev->snap_name, RBD_SNAP_HEAD_NAME,
		    sizeof (RBD_SNAP_HEAD_NAME))) {
		if (header->total_snaps)
			snapc->seq = header->snap_seq;
		else
			snapc->seq = 0;
		dev->snap_id = CEPH_NOSNAP;
		dev->read_only = 0;
		if (size)
			*size = header->image_size;
	} else {
		ret = snap_by_name(header, dev->snap_name, &snapc->seq, size);
		if (ret < 0)
			goto done;
		dev->snap_id = snapc->seq;
		dev->read_only = 1;
	}

	ret = 0;
done:
	up_write(&dev->header_rwsem);
	return ret;
}

static void rbd_header_free(struct rbd_image_header *header)
{
	kfree(header->snapc);
	kfree(header->snap_names);
	kfree(header->snap_sizes);
}

/*
 * get the actual striped segment name, offset and length
 */
static u64 rbd_get_segment(struct rbd_image_header *header,
			   const char *block_name,
			   u64 ofs, u64 len,
			   char *seg_name, u64 *segofs)
{
	u64 seg = ofs >> header->obj_order;

	if (seg_name)
		snprintf(seg_name, RBD_MAX_SEG_NAME_LEN,
			 "%s.%012llx", block_name, seg);

	ofs = ofs & ((1 << header->obj_order) - 1);
	len = min_t(u64, len, (1 << header->obj_order) - ofs);

	if (segofs)
		*segofs = ofs;

	return len;
}

static int rbd_get_num_segments(struct rbd_image_header *header,
				u64 ofs, u64 len)
{
	u64 start_seg = ofs >> header->obj_order;
	u64 end_seg = (ofs + len - 1) >> header->obj_order;
	return end_seg - start_seg + 1;
}

/*
 * returns the size of an object in the image
 */
static u64 rbd_obj_bytes(struct rbd_image_header *header)
{
	return 1 << header->obj_order;
}

/*
 * bio helpers
 */

static void bio_chain_put(struct bio *chain)
{
	struct bio *tmp;

	while (chain) {
		tmp = chain;
		chain = chain->bi_next;
		bio_put(tmp);
	}
}

/*
 * zeros a bio chain, starting at specific offset
 */
static void zero_bio_chain(struct bio *chain, int start_ofs)
{
	struct bio_vec *bv;
	unsigned long flags;
	void *buf;
	int i;
	int pos = 0;

	while (chain) {
		bio_for_each_segment(bv, chain, i) {
			if (pos + bv->bv_len > start_ofs) {
				int remainder = max(start_ofs - pos, 0);
				buf = bvec_kmap_irq(bv, &flags);
				memset(buf + remainder, 0,
				       bv->bv_len - remainder);
				bvec_kunmap_irq(buf, &flags);
			}
			pos += bv->bv_len;
		}

		chain = chain->bi_next;
	}
}

/*
 * bio_chain_clone - clone a chain of bios up to a certain length.
 * might return a bio_pair that will need to be released.
 */
static struct bio *bio_chain_clone(struct bio **old, struct bio **next,
				   struct bio_pair **bp,
				   int len, gfp_t gfpmask)
{
	struct bio *tmp, *old_chain = *old, *new_chain = NULL, *tail = NULL;
	int total = 0;

	if (*bp) {
		bio_pair_release(*bp);
		*bp = NULL;
	}

	while (old_chain && (total < len)) {
		tmp = bio_kmalloc(gfpmask, old_chain->bi_max_vecs);
		if (!tmp)
			goto err_out;

		if (total + old_chain->bi_size > len) {
			struct bio_pair *bp;

			/*
			 * this split can only happen with a single paged bio,
			 * split_bio will BUG_ON if this is not the case
			 */
			dout("bio_chain_clone split! total=%d remaining=%d"
			     "bi_size=%d\n",
			     (int)total, (int)len-total,
			     (int)old_chain->bi_size);

			/* split the bio. We'll release it either in the next
			   call, or it will have to be released outside */
			bp = bio_split(old_chain, (len - total) / SECTOR_SIZE);
			if (!bp)
				goto err_out;

			__bio_clone(tmp, &bp->bio1);

			*next = &bp->bio2;
		} else {
			__bio_clone(tmp, old_chain);
			*next = old_chain->bi_next;
		}

		tmp->bi_bdev = NULL;
		gfpmask &= ~__GFP_WAIT;
		tmp->bi_next = NULL;

		if (!new_chain) {
			new_chain = tail = tmp;
		} else {
			tail->bi_next = tmp;
			tail = tmp;
		}
		old_chain = old_chain->bi_next;

		total += tmp->bi_size;
	}

	BUG_ON(total < len);

	if (tail)
		tail->bi_next = NULL;

	*old = old_chain;

	return new_chain;

err_out:
	dout("bio_chain_clone with err\n");
	bio_chain_put(new_chain);
	return NULL;
}

/*
 * helpers for osd request op vectors.
 */
static int rbd_create_rw_ops(struct ceph_osd_req_op **ops,
			    int num_ops,
			    int opcode,
			    u32 payload_len)
{
	*ops = kzalloc(sizeof(struct ceph_osd_req_op) * (num_ops + 1),
		       GFP_NOIO);
	if (!*ops)
		return -ENOMEM;
	(*ops)[0].op = opcode;
	/*
	 * op extent offset and length will be set later on
	 * in calc_raw_layout()
	 */
	(*ops)[0].payload_len = payload_len;
	return 0;
}

static void rbd_destroy_ops(struct ceph_osd_req_op *ops)
{
	kfree(ops);
}

static void rbd_coll_end_req_index(struct request *rq,
				   struct rbd_req_coll *coll,
				   int index,
				   int ret, u64 len)
{
	struct request_queue *q;
	int min, max, i;

	dout("rbd_coll_end_req_index %p index %d ret %d len %lld\n",
	     coll, index, ret, len);

	if (!rq)
		return;

	if (!coll) {
		blk_end_request(rq, ret, len);
		return;
	}

	q = rq->q;

	spin_lock_irq(q->queue_lock);
	coll->status[index].done = 1;
	coll->status[index].rc = ret;
	coll->status[index].bytes = len;
	max = min = coll->num_done;
	while (max < coll->total && coll->status[max].done)
		max++;

	for (i = min; i<max; i++) {
		__blk_end_request(rq, coll->status[i].rc,
				  coll->status[i].bytes);
		coll->num_done++;
		kref_put(&coll->kref, rbd_coll_release);
	}
	spin_unlock_irq(q->queue_lock);
}

static void rbd_coll_end_req(struct rbd_request *req,
			     int ret, u64 len)
{
	rbd_coll_end_req_index(req->rq, req->coll, req->coll_index, ret, len);
}

/*
 * Send ceph osd request
 */
static int rbd_do_request(struct request *rq,
			  struct rbd_device *dev,
			  struct ceph_snap_context *snapc,
			  u64 snapid,
			  const char *obj, u64 ofs, u64 len,
			  struct bio *bio,
			  struct page **pages,
			  int num_pages,
			  int flags,
			  struct ceph_osd_req_op *ops,
			  int num_reply,
			  struct rbd_req_coll *coll,
			  int coll_index,
			  void (*rbd_cb)(struct ceph_osd_request *req,
					 struct ceph_msg *msg),
			  struct ceph_osd_request **linger_req,
			  u64 *ver)
{
	struct ceph_osd_request *req;
	struct ceph_file_layout *layout;
	int ret;
	u64 bno;
	struct timespec mtime = CURRENT_TIME;
	struct rbd_request *req_data;
	struct ceph_osd_request_head *reqhead;
	struct ceph_osd_client *osdc;

	req_data = kzalloc(sizeof(*req_data), GFP_NOIO);
	if (!req_data) {
		if (coll)
			rbd_coll_end_req_index(rq, coll, coll_index,
					       -ENOMEM, len);
		return -ENOMEM;
	}

	if (coll) {
		req_data->coll = coll;
		req_data->coll_index = coll_index;
	}

	dout("rbd_do_request obj=%s ofs=%lld len=%lld\n", obj, len, ofs);

	down_read(&dev->header_rwsem);

	osdc = &dev->rbd_client->client->osdc;
	req = ceph_osdc_alloc_request(osdc, flags, snapc, ops,
					false, GFP_NOIO, pages, bio);
	if (!req) {
		up_read(&dev->header_rwsem);
		ret = -ENOMEM;
		goto done_pages;
	}

	req->r_callback = rbd_cb;

	req_data->rq = rq;
	req_data->bio = bio;
	req_data->pages = pages;
	req_data->len = len;

	req->r_priv = req_data;

	reqhead = req->r_request->front.iov_base;
	reqhead->snapid = cpu_to_le64(CEPH_NOSNAP);

	strncpy(req->r_oid, obj, sizeof(req->r_oid));
	req->r_oid_len = strlen(req->r_oid);

	layout = &req->r_file_layout;
	memset(layout, 0, sizeof(*layout));
	layout->fl_stripe_unit = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER);
	layout->fl_stripe_count = cpu_to_le32(1);
	layout->fl_object_size = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER);
	layout->fl_pg_pool = cpu_to_le32(dev->poolid);
	ceph_calc_raw_layout(osdc, layout, snapid, ofs, &len, &bno,
				req, ops);

	ceph_osdc_build_request(req, ofs, &len,
				ops,
				snapc,
				&mtime,
				req->r_oid, req->r_oid_len);
	up_read(&dev->header_rwsem);

	if (linger_req) {
		ceph_osdc_set_request_linger(osdc, req);
		*linger_req = req;
	}

	ret = ceph_osdc_start_request(osdc, req, false);
	if (ret < 0)
		goto done_err;

	if (!rbd_cb) {
		ret = ceph_osdc_wait_request(osdc, req);
		if (ver)
			*ver = le64_to_cpu(req->r_reassert_version.version);
		dout("reassert_ver=%lld\n",
		     le64_to_cpu(req->r_reassert_version.version));
		ceph_osdc_put_request(req);
	}
	return ret;

done_err:
	bio_chain_put(req_data->bio);
	ceph_osdc_put_request(req);
done_pages:
	rbd_coll_end_req(req_data, ret, len);
	kfree(req_data);
	return ret;
}

/*
 * Ceph osd op callback
 */
static void rbd_req_cb(struct ceph_osd_request *req, struct ceph_msg *msg)
{
	struct rbd_request *req_data = req->r_priv;
	struct ceph_osd_reply_head *replyhead;
	struct ceph_osd_op *op;
	__s32 rc;
	u64 bytes;
	int read_op;

	/* parse reply */
	replyhead = msg->front.iov_base;
	WARN_ON(le32_to_cpu(replyhead->num_ops) == 0);
	op = (void *)(replyhead + 1);
	rc = le32_to_cpu(replyhead->result);
	bytes = le64_to_cpu(op->extent.length);
	read_op = (le16_to_cpu(op->op) == CEPH_OSD_OP_READ);

	dout("rbd_req_cb bytes=%lld readop=%d rc=%d\n", bytes, read_op, rc);

	if (rc == -ENOENT && read_op) {
		zero_bio_chain(req_data->bio, 0);
		rc = 0;
	} else if (rc == 0 && read_op && bytes < req_data->len) {
		zero_bio_chain(req_data->bio, bytes);
		bytes = req_data->len;
	}

	rbd_coll_end_req(req_data, rc, bytes);

	if (req_data->bio)
		bio_chain_put(req_data->bio);

	ceph_osdc_put_request(req);
	kfree(req_data);
}

static void rbd_simple_req_cb(struct ceph_osd_request *req, struct ceph_msg *msg)
{
	ceph_osdc_put_request(req);
}

/*
 * Do a synchronous ceph osd operation
 */
static int rbd_req_sync_op(struct rbd_device *dev,
			   struct ceph_snap_context *snapc,
			   u64 snapid,
			   int opcode,
			   int flags,
			   struct ceph_osd_req_op *orig_ops,
			   int num_reply,
			   const char *obj,
			   u64 ofs, u64 len,
			   char *buf,
			   struct ceph_osd_request **linger_req,
			   u64 *ver)
{
	int ret;
	struct page **pages;
	int num_pages;
	struct ceph_osd_req_op *ops = orig_ops;
	u32 payload_len;

	num_pages = calc_pages_for(ofs , len);
	pages = ceph_alloc_page_vector(num_pages, GFP_KERNEL);
	if (IS_ERR(pages))
		return PTR_ERR(pages);

	if (!orig_ops) {
		payload_len = (flags & CEPH_OSD_FLAG_WRITE ? len : 0);
		ret = rbd_create_rw_ops(&ops, 1, opcode, payload_len);
		if (ret < 0)
			goto done;

		if ((flags & CEPH_OSD_FLAG_WRITE) && buf) {
			ret = ceph_copy_to_page_vector(pages, buf, ofs, len);
			if (ret < 0)
				goto done_ops;
		}
	}

	ret = rbd_do_request(NULL, dev, snapc, snapid,
			  obj, ofs, len, NULL,
			  pages, num_pages,
			  flags,
			  ops,
			  2,
			  NULL, 0,
			  NULL,
			  linger_req, ver);
	if (ret < 0)
		goto done_ops;

	if ((flags & CEPH_OSD_FLAG_READ) && buf)
		ret = ceph_copy_from_page_vector(pages, buf, ofs, ret);

done_ops:
	if (!orig_ops)
		rbd_destroy_ops(ops);
done:
	ceph_release_page_vector(pages, num_pages);
	return ret;
}

/*
 * Do an asynchronous ceph osd operation
 */
static int rbd_do_op(struct request *rq,
		     struct rbd_device *rbd_dev ,
		     struct ceph_snap_context *snapc,
		     u64 snapid,
		     int opcode, int flags, int num_reply,
		     u64 ofs, u64 len,
		     struct bio *bio,
		     struct rbd_req_coll *coll,
		     int coll_index)
{
	char *seg_name;
	u64 seg_ofs;
	u64 seg_len;
	int ret;
	struct ceph_osd_req_op *ops;
	u32 payload_len;

	seg_name = kmalloc(RBD_MAX_SEG_NAME_LEN + 1, GFP_NOIO);
	if (!seg_name)
		return -ENOMEM;

	seg_len = rbd_get_segment(&rbd_dev->header,
				  rbd_dev->header.block_name,
				  ofs, len,
				  seg_name, &seg_ofs);

	payload_len = (flags & CEPH_OSD_FLAG_WRITE ? seg_len : 0);

	ret = rbd_create_rw_ops(&ops, 1, opcode, payload_len);
	if (ret < 0)
		goto done;

	/* we've taken care of segment sizes earlier when we
	   cloned the bios. We should never have a segment
	   truncated at this point */
	BUG_ON(seg_len < len);

	ret = rbd_do_request(rq, rbd_dev, snapc, snapid,
			     seg_name, seg_ofs, seg_len,
			     bio,
			     NULL, 0,
			     flags,
			     ops,
			     num_reply,
			     coll, coll_index,
			     rbd_req_cb, 0, NULL);

	rbd_destroy_ops(ops);
done:
	kfree(seg_name);
	return ret;
}

/*
 * Request async osd write
 */
static int rbd_req_write(struct request *rq,
			 struct rbd_device *rbd_dev,
			 struct ceph_snap_context *snapc,
			 u64 ofs, u64 len,
			 struct bio *bio,
			 struct rbd_req_coll *coll,
			 int coll_index)
{
	return rbd_do_op(rq, rbd_dev, snapc, CEPH_NOSNAP,
			 CEPH_OSD_OP_WRITE,
			 CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK,
			 2,
			 ofs, len, bio, coll, coll_index);
}

/*
 * Request async osd read
 */
static int rbd_req_read(struct request *rq,
			 struct rbd_device *rbd_dev,
			 u64 snapid,
			 u64 ofs, u64 len,
			 struct bio *bio,
			 struct rbd_req_coll *coll,
			 int coll_index)
{
	return rbd_do_op(rq, rbd_dev, NULL,
			 snapid,
			 CEPH_OSD_OP_READ,
			 CEPH_OSD_FLAG_READ,
			 2,
			 ofs, len, bio, coll, coll_index);
}

/*
 * Request sync osd read
 */
static int rbd_req_sync_read(struct rbd_device *dev,
			  struct ceph_snap_context *snapc,
			  u64 snapid,
			  const char *obj,
			  u64 ofs, u64 len,
			  char *buf,
			  u64 *ver)
{
	return rbd_req_sync_op(dev, NULL,
			       snapid,
			       CEPH_OSD_OP_READ,
			       CEPH_OSD_FLAG_READ,
			       NULL,
			       1, obj, ofs, len, buf, NULL, ver);
}

/*
 * Request sync osd watch
 */
static int rbd_req_sync_notify_ack(struct rbd_device *dev,
				   u64 ver,
				   u64 notify_id,
				   const char *obj)
{
	struct ceph_osd_req_op *ops;
	struct page **pages = NULL;
	int ret;

	ret = rbd_create_rw_ops(&ops, 1, CEPH_OSD_OP_NOTIFY_ACK, 0);
	if (ret < 0)
		return ret;

	ops[0].watch.ver = cpu_to_le64(dev->header.obj_version);
	ops[0].watch.cookie = notify_id;
	ops[0].watch.flag = 0;

	ret = rbd_do_request(NULL, dev, NULL, CEPH_NOSNAP,
			  obj, 0, 0, NULL,
			  pages, 0,
			  CEPH_OSD_FLAG_READ,
			  ops,
			  1,
			  NULL, 0,
			  rbd_simple_req_cb, 0, NULL);

	rbd_destroy_ops(ops);
	return ret;
}

static void rbd_watch_cb(u64 ver, u64 notify_id, u8 opcode, void *data)
{
	struct rbd_device *dev = (struct rbd_device *)data;
	int rc;

	if (!dev)
		return;

	dout("rbd_watch_cb %s notify_id=%lld opcode=%d\n", dev->obj_md_name,
		notify_id, (int)opcode);
	mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
	rc = __rbd_refresh_header(dev);
	mutex_unlock(&ctl_mutex);
	if (rc)
		pr_warning(RBD_DRV_NAME "%d got notification but failed to "
			   " update snaps: %d\n", dev->major, rc);

	rbd_req_sync_notify_ack(dev, ver, notify_id, dev->obj_md_name);
}

/*
 * Request sync osd watch
 */
static int rbd_req_sync_watch(struct rbd_device *dev,
			      const char *obj,
			      u64 ver)
{
	struct ceph_osd_req_op *ops;
	struct ceph_osd_client *osdc = &dev->rbd_client->client->osdc;

	int ret = rbd_create_rw_ops(&ops, 1, CEPH_OSD_OP_WATCH, 0);
	if (ret < 0)
		return ret;

	ret = ceph_osdc_create_event(osdc, rbd_watch_cb, 0,
				     (void *)dev, &dev->watch_event);
	if (ret < 0)
		goto fail;

	ops[0].watch.ver = cpu_to_le64(ver);
	ops[0].watch.cookie = cpu_to_le64(dev->watch_event->cookie);
	ops[0].watch.flag = 1;

	ret = rbd_req_sync_op(dev, NULL,
			      CEPH_NOSNAP,
			      0,
			      CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK,
			      ops,
			      1, obj, 0, 0, NULL,
			      &dev->watch_request, NULL);

	if (ret < 0)
		goto fail_event;

	rbd_destroy_ops(ops);
	return 0;

fail_event:
	ceph_osdc_cancel_event(dev->watch_event);
	dev->watch_event = NULL;
fail:
	rbd_destroy_ops(ops);
	return ret;
}

/*
 * Request sync osd unwatch
 */
static int rbd_req_sync_unwatch(struct rbd_device *dev,
				const char *obj)
{
	struct ceph_osd_req_op *ops;

	int ret = rbd_create_rw_ops(&ops, 1, CEPH_OSD_OP_WATCH, 0);
	if (ret < 0)
		return ret;

	ops[0].watch.ver = 0;
	ops[0].watch.cookie = cpu_to_le64(dev->watch_event->cookie);
	ops[0].watch.flag = 0;

	ret = rbd_req_sync_op(dev, NULL,
			      CEPH_NOSNAP,
			      0,
			      CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK,
			      ops,
			      1, obj, 0, 0, NULL, NULL, NULL);

	rbd_destroy_ops(ops);
	ceph_osdc_cancel_event(dev->watch_event);
	dev->watch_event = NULL;
	return ret;
}

struct rbd_notify_info {
	struct rbd_device *dev;
};

static void rbd_notify_cb(u64 ver, u64 notify_id, u8 opcode, void *data)
{
	struct rbd_device *dev = (struct rbd_device *)data;
	if (!dev)
		return;

	dout("rbd_notify_cb %s notify_id=%lld opcode=%d\n", dev->obj_md_name,
		notify_id, (int)opcode);
}

/*
 * Request sync osd notify
 */
static int rbd_req_sync_notify(struct rbd_device *dev,
		          const char *obj)
{
	struct ceph_osd_req_op *ops;
	struct ceph_osd_client *osdc = &dev->rbd_client->client->osdc;
	struct ceph_osd_event *event;
	struct rbd_notify_info info;
	int payload_len = sizeof(u32) + sizeof(u32);
	int ret;

	ret = rbd_create_rw_ops(&ops, 1, CEPH_OSD_OP_NOTIFY, payload_len);
	if (ret < 0)
		return ret;

	info.dev = dev;

	ret = ceph_osdc_create_event(osdc, rbd_notify_cb, 1,
				     (void *)&info, &event);
	if (ret < 0)
		goto fail;

	ops[0].watch.ver = 1;
	ops[0].watch.flag = 1;
	ops[0].watch.cookie = event->cookie;
	ops[0].watch.prot_ver = RADOS_NOTIFY_VER;
	ops[0].watch.timeout = 12;

	ret = rbd_req_sync_op(dev, NULL,
			       CEPH_NOSNAP,
			       0,
			       CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK,
			       ops,
			       1, obj, 0, 0, NULL, NULL, NULL);
	if (ret < 0)
		goto fail_event;

	ret = ceph_osdc_wait_event(event, CEPH_OSD_TIMEOUT_DEFAULT);
	dout("ceph_osdc_wait_event returned %d\n", ret);
	rbd_destroy_ops(ops);
	return 0;

fail_event:
	ceph_osdc_cancel_event(event);
fail:
	rbd_destroy_ops(ops);
	return ret;
}

/*
 * Request sync osd read
 */
static int rbd_req_sync_exec(struct rbd_device *dev,
			     const char *obj,
			     const char *cls,
			     const char *method,
			     const char *data,
			     int len,
			     u64 *ver)
{
	struct ceph_osd_req_op *ops;
	int cls_len = strlen(cls);
	int method_len = strlen(method);
	int ret = rbd_create_rw_ops(&ops, 1, CEPH_OSD_OP_CALL,
				    cls_len + method_len + len);
	if (ret < 0)
		return ret;

	ops[0].cls.class_name = cls;
	ops[0].cls.class_len = (__u8)cls_len;
	ops[0].cls.method_name = method;
	ops[0].cls.method_len = (__u8)method_len;
	ops[0].cls.argc = 0;
	ops[0].cls.indata = data;
	ops[0].cls.indata_len = len;

	ret = rbd_req_sync_op(dev, NULL,
			       CEPH_NOSNAP,
			       0,
			       CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK,
			       ops,
			       1, obj, 0, 0, NULL, NULL, ver);

	rbd_destroy_ops(ops);

	dout("cls_exec returned %d\n", ret);
	return ret;
}

static struct rbd_req_coll *rbd_alloc_coll(int num_reqs)
{
	struct rbd_req_coll *coll =
			kzalloc(sizeof(struct rbd_req_coll) +
			        sizeof(struct rbd_req_status) * num_reqs,
				GFP_ATOMIC);

	if (!coll)
		return NULL;
	coll->total = num_reqs;
	kref_init(&coll->kref);
	return coll;
}

/*
 * block device queue callback
 */
static void rbd_rq_fn(struct request_queue *q)
{
	struct rbd_device *rbd_dev = q->queuedata;
	struct request *rq;
	struct bio_pair *bp = NULL;

	while ((rq = blk_fetch_request(q))) {
		struct bio *bio;
		struct bio *rq_bio, *next_bio = NULL;
		bool do_write;
		int size, op_size = 0;
		u64 ofs;
		int num_segs, cur_seg = 0;
		struct rbd_req_coll *coll;

		/* peek at request from block layer */
		if (!rq)
			break;

		dout("fetched request\n");

		/* filter out block requests we don't understand */
		if ((rq->cmd_type != REQ_TYPE_FS)) {
			__blk_end_request_all(rq, 0);
			continue;
		}

		/* deduce our operation (read, write) */
		do_write = (rq_data_dir(rq) == WRITE);

		size = blk_rq_bytes(rq);
		ofs = blk_rq_pos(rq) * SECTOR_SIZE;
		rq_bio = rq->bio;
		if (do_write && rbd_dev->read_only) {
			__blk_end_request_all(rq, -EROFS);
			continue;
		}

		spin_unlock_irq(q->queue_lock);

		dout("%s 0x%x bytes at 0x%llx\n",
		     do_write ? "write" : "read",
		     size, blk_rq_pos(rq) * SECTOR_SIZE);

		num_segs = rbd_get_num_segments(&rbd_dev->header, ofs, size);
		coll = rbd_alloc_coll(num_segs);
		if (!coll) {
			spin_lock_irq(q->queue_lock);
			__blk_end_request_all(rq, -ENOMEM);
			continue;
		}

		do {
			/* a bio clone to be passed down to OSD req */
			dout("rq->bio->bi_vcnt=%d\n", rq->bio->bi_vcnt);
			op_size = rbd_get_segment(&rbd_dev->header,
						  rbd_dev->header.block_name,
						  ofs, size,
						  NULL, NULL);
			kref_get(&coll->kref);
			bio = bio_chain_clone(&rq_bio, &next_bio, &bp,
					      op_size, GFP_ATOMIC);
			if (!bio) {
				rbd_coll_end_req_index(rq, coll, cur_seg,
						       -ENOMEM, op_size);
				goto next_seg;
			}


			/* init OSD command: write or read */
			if (do_write)
				rbd_req_write(rq, rbd_dev,
					      rbd_dev->header.snapc,
					      ofs,
					      op_size, bio,
					      coll, cur_seg);
			else
				rbd_req_read(rq, rbd_dev,
					     rbd_dev->snap_id,
					     ofs,
					     op_size, bio,
					     coll, cur_seg);

next_seg:
			size -= op_size;
			ofs += op_size;

			cur_seg++;
			rq_bio = next_bio;
		} while (size > 0);
		kref_put(&coll->kref, rbd_coll_release);

		if (bp)
			bio_pair_release(bp);
		spin_lock_irq(q->queue_lock);
	}
}

/*
 * a queue callback. Makes sure that we don't create a bio that spans across
 * multiple osd objects. One exception would be with a single page bios,
 * which we handle later at bio_chain_clone
 */
static int rbd_merge_bvec(struct request_queue *q, struct bvec_merge_data *bmd,
			  struct bio_vec *bvec)
{
	struct rbd_device *rbd_dev = q->queuedata;
	unsigned int chunk_sectors;
	sector_t sector;
	unsigned int bio_sectors;
	int max;

	chunk_sectors = 1 << (rbd_dev->header.obj_order - SECTOR_SHIFT);
	sector = bmd->bi_sector + get_start_sect(bmd->bi_bdev);
	bio_sectors = bmd->bi_size >> SECTOR_SHIFT;

	max =  (chunk_sectors - ((sector & (chunk_sectors - 1))
				 + bio_sectors)) << SECTOR_SHIFT;
	if (max < 0)
		max = 0; /* bio_add cannot handle a negative return */
	if (max <= bvec->bv_len && bio_sectors == 0)
		return bvec->bv_len;
	return max;
}

static void rbd_free_disk(struct rbd_device *rbd_dev)
{
	struct gendisk *disk = rbd_dev->disk;

	if (!disk)
		return;

	rbd_header_free(&rbd_dev->header);

	if (disk->flags & GENHD_FL_UP)
		del_gendisk(disk);
	if (disk->queue)
		blk_cleanup_queue(disk->queue);
	put_disk(disk);
}

/*
 * reload the ondisk the header 
 */
static int rbd_read_header(struct rbd_device *rbd_dev,
			   struct rbd_image_header *header)
{
	ssize_t rc;
	struct rbd_image_header_ondisk *dh;
	u32 snap_count = 0;
	u64 ver;
	size_t len;

	/*
	 * First reads the fixed-size header to determine the number
	 * of snapshots, then re-reads it, along with all snapshot
	 * records as well as their stored names.
	 */
	len = sizeof (*dh);
	while (1) {
		dh = kmalloc(len, GFP_KERNEL);
		if (!dh)
			return -ENOMEM;

		rc = rbd_req_sync_read(rbd_dev,
				       NULL, CEPH_NOSNAP,
				       rbd_dev->obj_md_name,
				       0, len,
				       (char *)dh, &ver);
		if (rc < 0)
			goto out_dh;

		rc = rbd_header_from_disk(header, dh, snap_count, GFP_KERNEL);
		if (rc < 0) {
			if (rc == -ENXIO)
				pr_warning("unrecognized header format"
					   " for image %s", rbd_dev->obj);
			goto out_dh;
		}

		if (snap_count == header->total_snaps)
			break;

		snap_count = header->total_snaps;
		len = sizeof (*dh) +
			snap_count * sizeof(struct rbd_image_snap_ondisk) +
			header->snap_names_len;

		rbd_header_free(header);
		kfree(dh);
	}
	header->obj_version = ver;

out_dh:
	kfree(dh);
	return rc;
}

/*
 * create a snapshot
 */
static int rbd_header_add_snap(struct rbd_device *dev,
			       const char *snap_name,
			       gfp_t gfp_flags)
{
	int name_len = strlen(snap_name);
	u64 new_snapid;
	int ret;
	void *data, *p, *e;
	u64 ver;
	struct ceph_mon_client *monc;

	/* we should create a snapshot only if we're pointing at the head */
	if (dev->snap_id != CEPH_NOSNAP)
		return -EINVAL;

	monc = &dev->rbd_client->client->monc;
	ret = ceph_monc_create_snapid(monc, dev->poolid, &new_snapid);
	dout("created snapid=%lld\n", new_snapid);
	if (ret < 0)
		return ret;

	data = kmalloc(name_len + 16, gfp_flags);
	if (!data)
		return -ENOMEM;

	p = data;
	e = data + name_len + 16;

	ceph_encode_string_safe(&p, e, snap_name, name_len, bad);
	ceph_encode_64_safe(&p, e, new_snapid, bad);

	ret = rbd_req_sync_exec(dev, dev->obj_md_name, "rbd", "snap_add",
				data, p - data, &ver);

	kfree(data);

	if (ret < 0)
		return ret;

	down_write(&dev->header_rwsem);
	dev->header.snapc->seq = new_snapid;
	up_write(&dev->header_rwsem);

	return 0;
bad:
	return -ERANGE;
}

static void __rbd_remove_all_snaps(struct rbd_device *rbd_dev)
{
	struct rbd_snap *snap;

	while (!list_empty(&rbd_dev->snaps)) {
		snap = list_first_entry(&rbd_dev->snaps, struct rbd_snap, node);
		__rbd_remove_snap_dev(rbd_dev, snap);
	}
}

/*
 * only read the first part of the ondisk header, without the snaps info
 */
static int __rbd_refresh_header(struct rbd_device *rbd_dev)
{
	int ret;
	struct rbd_image_header h;
	u64 snap_seq;
	int follow_seq = 0;

	ret = rbd_read_header(rbd_dev, &h);
	if (ret < 0)
		return ret;

	/* resized? */
	set_capacity(rbd_dev->disk, h.image_size / SECTOR_SIZE);

	down_write(&rbd_dev->header_rwsem);

	snap_seq = rbd_dev->header.snapc->seq;
	if (rbd_dev->header.total_snaps &&
	    rbd_dev->header.snapc->snaps[0] == snap_seq)
		/* pointing at the head, will need to follow that
		   if head moves */
		follow_seq = 1;

	kfree(rbd_dev->header.snapc);
	kfree(rbd_dev->header.snap_names);
	kfree(rbd_dev->header.snap_sizes);

	rbd_dev->header.total_snaps = h.total_snaps;
	rbd_dev->header.snapc = h.snapc;
	rbd_dev->header.snap_names = h.snap_names;
	rbd_dev->header.snap_names_len = h.snap_names_len;
	rbd_dev->header.snap_sizes = h.snap_sizes;
	if (follow_seq)
		rbd_dev->header.snapc->seq = rbd_dev->header.snapc->snaps[0];
	else
		rbd_dev->header.snapc->seq = snap_seq;

	ret = __rbd_init_snaps_header(rbd_dev);

	up_write(&rbd_dev->header_rwsem);

	return ret;
}

static int rbd_init_disk(struct rbd_device *rbd_dev)
{
	struct gendisk *disk;
	struct request_queue *q;
	int rc;
	u64 segment_size;
	u64 total_size = 0;

	/* contact OSD, request size info about the object being mapped */
	rc = rbd_read_header(rbd_dev, &rbd_dev->header);
	if (rc)
		return rc;

	/* no need to lock here, as rbd_dev is not registered yet */
	rc = __rbd_init_snaps_header(rbd_dev);
	if (rc)
		return rc;

	rc = rbd_header_set_snap(rbd_dev, &total_size);
	if (rc)
		return rc;

	/* create gendisk info */
	rc = -ENOMEM;
	disk = alloc_disk(RBD_MINORS_PER_MAJOR);
	if (!disk)
		goto out;

	snprintf(disk->disk_name, sizeof(disk->disk_name), RBD_DRV_NAME "%d",
		 rbd_dev->id);
	disk->major = rbd_dev->major;
	disk->first_minor = 0;
	disk->fops = &rbd_bd_ops;
	disk->private_data = rbd_dev;

	/* init rq */
	rc = -ENOMEM;
	q = blk_init_queue(rbd_rq_fn, &rbd_dev->lock);
	if (!q)
		goto out_disk;

	/* We use the default size, but let's be explicit about it. */
	blk_queue_physical_block_size(q, SECTOR_SIZE);

	/* set io sizes to object size */
	segment_size = rbd_obj_bytes(&rbd_dev->header);
	blk_queue_max_hw_sectors(q, segment_size / SECTOR_SIZE);
	blk_queue_max_segment_size(q, segment_size);
	blk_queue_io_min(q, segment_size);
	blk_queue_io_opt(q, segment_size);

	blk_queue_merge_bvec(q, rbd_merge_bvec);
	disk->queue = q;

	q->queuedata = rbd_dev;

	rbd_dev->disk = disk;
	rbd_dev->q = q;

	/* finally, announce the disk to the world */
	set_capacity(disk, total_size / SECTOR_SIZE);
	add_disk(disk);

	pr_info("%s: added with size 0x%llx\n",
		disk->disk_name, (unsigned long long)total_size);
	return 0;

out_disk:
	put_disk(disk);
out:
	return rc;
}

/*
  sysfs
*/

static struct rbd_device *dev_to_rbd_dev(struct device *dev)
{
	return container_of(dev, struct rbd_device, dev);
}

static ssize_t rbd_size_show(struct device *dev,
			     struct device_attribute *attr, char *buf)
{
	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);

	return sprintf(buf, "%llu\n", (unsigned long long)rbd_dev->header.image_size);
}

static ssize_t rbd_major_show(struct device *dev,
			      struct device_attribute *attr, char *buf)
{
	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);

	return sprintf(buf, "%d\n", rbd_dev->major);
}

static ssize_t rbd_client_id_show(struct device *dev,
				  struct device_attribute *attr, char *buf)
{
	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);

	return sprintf(buf, "client%lld\n",
			ceph_client_id(rbd_dev->rbd_client->client));
}

static ssize_t rbd_pool_show(struct device *dev,
			     struct device_attribute *attr, char *buf)
{
	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);

	return sprintf(buf, "%s\n", rbd_dev->pool_name);
}

static ssize_t rbd_name_show(struct device *dev,
			     struct device_attribute *attr, char *buf)
{
	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);

	return sprintf(buf, "%s\n", rbd_dev->obj);
}

static ssize_t rbd_snap_show(struct device *dev,
			     struct device_attribute *attr,
			     char *buf)
{
	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);

	return sprintf(buf, "%s\n", rbd_dev->snap_name);
}

static ssize_t rbd_image_refresh(struct device *dev,
				 struct device_attribute *attr,
				 const char *buf,
				 size_t size)
{
	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
	int rc;
	int ret = size;

	mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);

	rc = __rbd_refresh_header(rbd_dev);
	if (rc < 0)
		ret = rc;

	mutex_unlock(&ctl_mutex);
	return ret;
}

static DEVICE_ATTR(size, S_IRUGO, rbd_size_show, NULL);
static DEVICE_ATTR(major, S_IRUGO, rbd_major_show, NULL);
static DEVICE_ATTR(client_id, S_IRUGO, rbd_client_id_show, NULL);
static DEVICE_ATTR(pool, S_IRUGO, rbd_pool_show, NULL);
static DEVICE_ATTR(name, S_IRUGO, rbd_name_show, NULL);
static DEVICE_ATTR(refresh, S_IWUSR, NULL, rbd_image_refresh);
static DEVICE_ATTR(current_snap, S_IRUGO, rbd_snap_show, NULL);
static DEVICE_ATTR(create_snap, S_IWUSR, NULL, rbd_snap_add);

static struct attribute *rbd_attrs[] = {
	&dev_attr_size.attr,
	&dev_attr_major.attr,
	&dev_attr_client_id.attr,
	&dev_attr_pool.attr,
	&dev_attr_name.attr,
	&dev_attr_current_snap.attr,
	&dev_attr_refresh.attr,
	&dev_attr_create_snap.attr,
	NULL
};

static struct attribute_group rbd_attr_group = {
	.attrs = rbd_attrs,
};

static const struct attribute_group *rbd_attr_groups[] = {
	&rbd_attr_group,
	NULL
};

static void rbd_sysfs_dev_release(struct device *dev)
{
}

static struct device_type rbd_device_type = {
	.name		= "rbd",
	.groups		= rbd_attr_groups,
	.release	= rbd_sysfs_dev_release,
};


/*
  sysfs - snapshots
*/

static ssize_t rbd_snap_size_show(struct device *dev,
				  struct device_attribute *attr,
				  char *buf)
{
	struct rbd_snap *snap = container_of(dev, struct rbd_snap, dev);

	return sprintf(buf, "%llu\n", (unsigned long long)snap->size);
}

static ssize_t rbd_snap_id_show(struct device *dev,
				struct device_attribute *attr,
				char *buf)
{
	struct rbd_snap *snap = container_of(dev, struct rbd_snap, dev);

	return sprintf(buf, "%llu\n", (unsigned long long)snap->id);
}

static DEVICE_ATTR(snap_size, S_IRUGO, rbd_snap_size_show, NULL);
static DEVICE_ATTR(snap_id, S_IRUGO, rbd_snap_id_show, NULL);

static struct attribute *rbd_snap_attrs[] = {
	&dev_attr_snap_size.attr,
	&dev_attr_snap_id.attr,
	NULL,
};

static struct attribute_group rbd_snap_attr_group = {
	.attrs = rbd_snap_attrs,
};

static void rbd_snap_dev_release(struct device *dev)
{
	struct rbd_snap *snap = container_of(dev, struct rbd_snap, dev);
	kfree(snap->name);
	kfree(snap);
}

static const struct attribute_group *rbd_snap_attr_groups[] = {
	&rbd_snap_attr_group,
	NULL
};

static struct device_type rbd_snap_device_type = {
	.groups		= rbd_snap_attr_groups,
	.release	= rbd_snap_dev_release,
};

static void __rbd_remove_snap_dev(struct rbd_device *rbd_dev,
				  struct rbd_snap *snap)
{
	list_del(&snap->node);
	device_unregister(&snap->dev);
}

static int rbd_register_snap_dev(struct rbd_device *rbd_dev,
				  struct rbd_snap *snap,
				  struct device *parent)
{
	struct device *dev = &snap->dev;
	int ret;

	dev->type = &rbd_snap_device_type;
	dev->parent = parent;
	dev->release = rbd_snap_dev_release;
	dev_set_name(dev, "snap_%s", snap->name);
	ret = device_register(dev);

	return ret;
}

static int __rbd_add_snap_dev(struct rbd_device *rbd_dev,
			      int i, const char *name,
			      struct rbd_snap **snapp)
{
	int ret;
	struct rbd_snap *snap = kzalloc(sizeof(*snap), GFP_KERNEL);
	if (!snap)
		return -ENOMEM;
	snap->name = kstrdup(name, GFP_KERNEL);
	snap->size = rbd_dev->header.snap_sizes[i];
	snap->id = rbd_dev->header.snapc->snaps[i];
	if (device_is_registered(&rbd_dev->dev)) {
		ret = rbd_register_snap_dev(rbd_dev, snap,
					     &rbd_dev->dev);
		if (ret < 0)
			goto err;
	}
	*snapp = snap;
	return 0;
err:
	kfree(snap->name);
	kfree(snap);
	return ret;
}

/*
 * search for the previous snap in a null delimited string list
 */
const char *rbd_prev_snap_name(const char *name, const char *start)
{
	if (name < start + 2)
		return NULL;

	name -= 2;
	while (*name) {
		if (name == start)
			return start;
		name--;
	}
	return name + 1;
}

/*
 * compare the old list of snapshots that we have to what's in the header
 * and update it accordingly. Note that the header holds the snapshots
 * in a reverse order (from newest to oldest) and we need to go from
 * older to new so that we don't get a duplicate snap name when
 * doing the process (e.g., removed snapshot and recreated a new
 * one with the same name.
 */
static int __rbd_init_snaps_header(struct rbd_device *rbd_dev)
{
	const char *name, *first_name;
	int i = rbd_dev->header.total_snaps;
	struct rbd_snap *snap, *old_snap = NULL;
	int ret;
	struct list_head *p, *n;

	first_name = rbd_dev->header.snap_names;
	name = first_name + rbd_dev->header.snap_names_len;

	list_for_each_prev_safe(p, n, &rbd_dev->snaps) {
		u64 cur_id;

		old_snap = list_entry(p, struct rbd_snap, node);

		if (i)
			cur_id = rbd_dev->header.snapc->snaps[i - 1];

		if (!i || old_snap->id < cur_id) {
			/* old_snap->id was skipped, thus was removed */
			__rbd_remove_snap_dev(rbd_dev, old_snap);
			continue;
		}
		if (old_snap->id == cur_id) {
			/* we have this snapshot already */
			i--;
			name = rbd_prev_snap_name(name, first_name);
			continue;
		}
		for (; i > 0;
		     i--, name = rbd_prev_snap_name(name, first_name)) {
			if (!name) {
				WARN_ON(1);
				return -EINVAL;
			}
			cur_id = rbd_dev->header.snapc->snaps[i];
			/* snapshot removal? handle it above */
			if (cur_id >= old_snap->id)
				break;
			/* a new snapshot */
			ret = __rbd_add_snap_dev(rbd_dev, i - 1, name, &snap);
			if (ret < 0)
				return ret;

			/* note that we add it backward so using n and not p */
			list_add(&snap->node, n);
			p = &snap->node;
		}
	}
	/* we're done going over the old snap list, just add what's left */
	for (; i > 0; i--) {
		name = rbd_prev_snap_name(name, first_name);
		if (!name) {
			WARN_ON(1);
			return -EINVAL;
		}
		ret = __rbd_add_snap_dev(rbd_dev, i - 1, name, &snap);
		if (ret < 0)
			return ret;
		list_add(&snap->node, &rbd_dev->snaps);
	}

	return 0;
}

static int rbd_bus_add_dev(struct rbd_device *rbd_dev)
{
	int ret;
	struct device *dev;
	struct rbd_snap *snap;

	mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
	dev = &rbd_dev->dev;

	dev->bus = &rbd_bus_type;
	dev->type = &rbd_device_type;
	dev->parent = &rbd_root_dev;
	dev->release = rbd_dev_release;
	dev_set_name(dev, "%d", rbd_dev->id);
	ret = device_register(dev);
	if (ret < 0)
		goto out;

	list_for_each_entry(snap, &rbd_dev->snaps, node) {
		ret = rbd_register_snap_dev(rbd_dev, snap,
					     &rbd_dev->dev);
		if (ret < 0)
			break;
	}
out:
	mutex_unlock(&ctl_mutex);
	return ret;
}

static void rbd_bus_del_dev(struct rbd_device *rbd_dev)
{
	device_unregister(&rbd_dev->dev);
}

static int rbd_init_watch_dev(struct rbd_device *rbd_dev)
{
	int ret, rc;

	do {
		ret = rbd_req_sync_watch(rbd_dev, rbd_dev->obj_md_name,
					 rbd_dev->header.obj_version);
		if (ret == -ERANGE) {
			mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
			rc = __rbd_refresh_header(rbd_dev);
			mutex_unlock(&ctl_mutex);
			if (rc < 0)
				return rc;
		}
	} while (ret == -ERANGE);

	return ret;
}

static atomic64_t rbd_id_max = ATOMIC64_INIT(0);

/*
 * Get a unique rbd identifier for the given new rbd_dev, and add
 * the rbd_dev to the global list.  The minimum rbd id is 1.
 */
static void rbd_id_get(struct rbd_device *rbd_dev)
{
	rbd_dev->id = atomic64_inc_return(&rbd_id_max);

	spin_lock(&rbd_dev_list_lock);
	list_add_tail(&rbd_dev->node, &rbd_dev_list);
	spin_unlock(&rbd_dev_list_lock);
}

/*
 * Remove an rbd_dev from the global list, and record that its
 * identifier is no longer in use.
 */
static void rbd_id_put(struct rbd_device *rbd_dev)
{
	struct list_head *tmp;
	int rbd_id = rbd_dev->id;
	int max_id;

	BUG_ON(rbd_id < 1);

	spin_lock(&rbd_dev_list_lock);
	list_del_init(&rbd_dev->node);

	/*
	 * If the id being "put" is not the current maximum, there
	 * is nothing special we need to do.
	 */
	if (rbd_id != atomic64_read(&rbd_id_max)) {
		spin_unlock(&rbd_dev_list_lock);
		return;
	}

	/*
	 * We need to update the current maximum id.  Search the
	 * list to find out what it is.  We're more likely to find
	 * the maximum at the end, so search the list backward.
	 */
	max_id = 0;
	list_for_each_prev(tmp, &rbd_dev_list) {
		struct rbd_device *rbd_dev;

		rbd_dev = list_entry(tmp, struct rbd_device, node);
		if (rbd_id > max_id)
			max_id = rbd_id;
	}
	spin_unlock(&rbd_dev_list_lock);

	/*
	 * The max id could have been updated by rbd_id_get(), in
	 * which case it now accurately reflects the new maximum.
	 * Be careful not to overwrite the maximum value in that
	 * case.
	 */
	atomic64_cmpxchg(&rbd_id_max, rbd_id, max_id);
}

/*
 * Skips over white space at *buf, and updates *buf to point to the
 * first found non-space character (if any). Returns the length of
 * the token (string of non-white space characters) found.  Note
 * that *buf must be terminated with '\0'.
 */
static inline size_t next_token(const char **buf)
{
        /*
        * These are the characters that produce nonzero for
        * isspace() in the "C" and "POSIX" locales.
        */
        const char *spaces = " \f\n\r\t\v";

        *buf += strspn(*buf, spaces);	/* Find start of token */

	return strcspn(*buf, spaces);   /* Return token length */
}

/*
 * Finds the next token in *buf, and if the provided token buffer is
 * big enough, copies the found token into it.  The result, if
 * copied, is guaranteed to be terminated with '\0'.  Note that *buf
 * must be terminated with '\0' on entry.
 *
 * Returns the length of the token found (not including the '\0').
 * Return value will be 0 if no token is found, and it will be >=
 * token_size if the token would not fit.
 *
 * The *buf pointer will be updated to point beyond the end of the
 * found token.  Note that this occurs even if the token buffer is
 * too small to hold it.
 */
static inline size_t copy_token(const char **buf,
				char *token,
				size_t token_size)
{
        size_t len;

	len = next_token(buf);
	if (len < token_size) {
		memcpy(token, *buf, len);
		*(token + len) = '\0';
	}
	*buf += len;

        return len;
}

/*
 * This fills in the pool_name, obj, obj_len, snap_name, obj_len,
 * rbd_dev, rbd_md_name, and name fields of the given rbd_dev, based
 * on the list of monitor addresses and other options provided via
 * /sys/bus/rbd/add.
 */
static int rbd_add_parse_args(struct rbd_device *rbd_dev,
			      const char *buf,
			      const char **mon_addrs,
			      size_t *mon_addrs_size,
			      char *options,
			      size_t options_size)
{
	size_t	len;

	/* The first four tokens are required */

	len = next_token(&buf);
	if (!len)
		return -EINVAL;
	*mon_addrs_size = len + 1;
	*mon_addrs = buf;

	buf += len;

	len = copy_token(&buf, options, options_size);
	if (!len || len >= options_size)
		return -EINVAL;

	len = copy_token(&buf, rbd_dev->pool_name, sizeof (rbd_dev->pool_name));
	if (!len || len >= sizeof (rbd_dev->pool_name))
		return -EINVAL;

	len = copy_token(&buf, rbd_dev->obj, sizeof (rbd_dev->obj));
	if (!len || len >= sizeof (rbd_dev->obj))
		return -EINVAL;

	/* We have the object length in hand, save it. */

	rbd_dev->obj_len = len;

	BUILD_BUG_ON(RBD_MAX_MD_NAME_LEN
				< RBD_MAX_OBJ_NAME_LEN + sizeof (RBD_SUFFIX));
	sprintf(rbd_dev->obj_md_name, "%s%s", rbd_dev->obj, RBD_SUFFIX);

	/*
	 * The snapshot name is optional, but it's an error if it's
	 * too long.  If no snapshot is supplied, fill in the default.
	 */
	len = copy_token(&buf, rbd_dev->snap_name, sizeof (rbd_dev->snap_name));
	if (!len)
		memcpy(rbd_dev->snap_name, RBD_SNAP_HEAD_NAME,
			sizeof (RBD_SNAP_HEAD_NAME));
	else if (len >= sizeof (rbd_dev->snap_name))
		return -EINVAL;

	return 0;
}

static ssize_t rbd_add(struct bus_type *bus,
		       const char *buf,
		       size_t count)
{
	struct rbd_device *rbd_dev;
	const char *mon_addrs = NULL;
	size_t mon_addrs_size = 0;
	char *options = NULL;
	struct ceph_osd_client *osdc;
	int rc = -ENOMEM;

	if (!try_module_get(THIS_MODULE))
		return -ENODEV;

	rbd_dev = kzalloc(sizeof(*rbd_dev), GFP_KERNEL);
	if (!rbd_dev)
		goto err_nomem;
	options = kmalloc(count, GFP_KERNEL);
	if (!options)
		goto err_nomem;

	/* static rbd_device initialization */
	spin_lock_init(&rbd_dev->lock);
	INIT_LIST_HEAD(&rbd_dev->node);
	INIT_LIST_HEAD(&rbd_dev->snaps);
	init_rwsem(&rbd_dev->header_rwsem);

	init_rwsem(&rbd_dev->header_rwsem);

	/* generate unique id: find highest unique id, add one */
	rbd_id_get(rbd_dev);

	/* Fill in the device name, now that we have its id. */
	BUILD_BUG_ON(DEV_NAME_LEN
			< sizeof (RBD_DRV_NAME) + MAX_INT_FORMAT_WIDTH);
	sprintf(rbd_dev->name, "%s%d", RBD_DRV_NAME, rbd_dev->id);

	/* parse add command */
	rc = rbd_add_parse_args(rbd_dev, buf, &mon_addrs, &mon_addrs_size,
				options, count);
	if (rc)
		goto err_put_id;

	rbd_dev->rbd_client = rbd_get_client(mon_addrs, mon_addrs_size - 1,
						options);
	if (IS_ERR(rbd_dev->rbd_client)) {
		rc = PTR_ERR(rbd_dev->rbd_client);
		goto err_put_id;
	}

	/* pick the pool */
	osdc = &rbd_dev->rbd_client->client->osdc;
	rc = ceph_pg_poolid_by_name(osdc->osdmap, rbd_dev->pool_name);
	if (rc < 0)
		goto err_out_client;
	rbd_dev->poolid = rc;

	/* register our block device */
	rc = register_blkdev(0, rbd_dev->name);
	if (rc < 0)
		goto err_out_client;
	rbd_dev->major = rc;

	rc = rbd_bus_add_dev(rbd_dev);
	if (rc)
		goto err_out_blkdev;

	/*
	 * At this point cleanup in the event of an error is the job
	 * of the sysfs code (initiated by rbd_bus_del_dev()).
	 *
	 * Set up and announce blkdev mapping.
	 */
	rc = rbd_init_disk(rbd_dev);
	if (rc)
		goto err_out_bus;

	rc = rbd_init_watch_dev(rbd_dev);
	if (rc)
		goto err_out_bus;

	return count;

err_out_bus:
	/* this will also clean up rest of rbd_dev stuff */

	rbd_bus_del_dev(rbd_dev);
	kfree(options);
	return rc;

err_out_blkdev:
	unregister_blkdev(rbd_dev->major, rbd_dev->name);
err_out_client:
	rbd_put_client(rbd_dev);
err_put_id:
	rbd_id_put(rbd_dev);
err_nomem:
	kfree(options);
	kfree(rbd_dev);

	dout("Error adding device %s\n", buf);
	module_put(THIS_MODULE);

	return (ssize_t) rc;
}

static struct rbd_device *__rbd_get_dev(unsigned long id)
{
	struct list_head *tmp;
	struct rbd_device *rbd_dev;

	spin_lock(&rbd_dev_list_lock);
	list_for_each(tmp, &rbd_dev_list) {
		rbd_dev = list_entry(tmp, struct rbd_device, node);
		if (rbd_dev->id == id) {
			spin_unlock(&rbd_dev_list_lock);
			return rbd_dev;
		}
	}
	spin_unlock(&rbd_dev_list_lock);
	return NULL;
}

static void rbd_dev_release(struct device *dev)
{
	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);

	if (rbd_dev->watch_request) {
		struct ceph_client *client = rbd_dev->rbd_client->client;

		ceph_osdc_unregister_linger_request(&client->osdc,
						    rbd_dev->watch_request);
	}
	if (rbd_dev->watch_event)
		rbd_req_sync_unwatch(rbd_dev, rbd_dev->obj_md_name);

	rbd_put_client(rbd_dev);

	/* clean up and free blkdev */
	rbd_free_disk(rbd_dev);
	unregister_blkdev(rbd_dev->major, rbd_dev->name);

	/* done with the id, and with the rbd_dev */
	rbd_id_put(rbd_dev);
	kfree(rbd_dev);

	/* release module ref */
	module_put(THIS_MODULE);
}

static ssize_t rbd_remove(struct bus_type *bus,
			  const char *buf,
			  size_t count)
{
	struct rbd_device *rbd_dev = NULL;
	int target_id, rc;
	unsigned long ul;
	int ret = count;

	rc = strict_strtoul(buf, 10, &ul);
	if (rc)
		return rc;

	/* convert to int; abort if we lost anything in the conversion */
	target_id = (int) ul;
	if (target_id != ul)
		return -EINVAL;

	mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);

	rbd_dev = __rbd_get_dev(target_id);
	if (!rbd_dev) {
		ret = -ENOENT;
		goto done;
	}

	__rbd_remove_all_snaps(rbd_dev);
	rbd_bus_del_dev(rbd_dev);

done:
	mutex_unlock(&ctl_mutex);
	return ret;
}

static ssize_t rbd_snap_add(struct device *dev,
			    struct device_attribute *attr,
			    const char *buf,
			    size_t count)
{
	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
	int ret;
	char *name = kmalloc(count + 1, GFP_KERNEL);
	if (!name)
		return -ENOMEM;

	snprintf(name, count, "%s", buf);

	mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);

	ret = rbd_header_add_snap(rbd_dev,
				  name, GFP_KERNEL);
	if (ret < 0)
		goto err_unlock;

	ret = __rbd_refresh_header(rbd_dev);
	if (ret < 0)
		goto err_unlock;

	/* shouldn't hold ctl_mutex when notifying.. notify might
	   trigger a watch callback that would need to get that mutex */
	mutex_unlock(&ctl_mutex);

	/* make a best effort, don't error if failed */
	rbd_req_sync_notify(rbd_dev, rbd_dev->obj_md_name);

	ret = count;
	kfree(name);
	return ret;

err_unlock:
	mutex_unlock(&ctl_mutex);
	kfree(name);
	return ret;
}

/*
 * create control files in sysfs
 * /sys/bus/rbd/...
 */
static int rbd_sysfs_init(void)
{
	int ret;

	ret = device_register(&rbd_root_dev);
	if (ret < 0)
		return ret;

	ret = bus_register(&rbd_bus_type);
	if (ret < 0)
		device_unregister(&rbd_root_dev);

	return ret;
}

static void rbd_sysfs_cleanup(void)
{
	bus_unregister(&rbd_bus_type);
	device_unregister(&rbd_root_dev);
}

int __init rbd_init(void)
{
	int rc;

	rc = rbd_sysfs_init();
	if (rc)
		return rc;
	pr_info("loaded " RBD_DRV_NAME_LONG "\n");
	return 0;
}

void __exit rbd_exit(void)
{
	rbd_sysfs_cleanup();
}

module_init(rbd_init);
module_exit(rbd_exit);

MODULE_AUTHOR("Sage Weil <sage@newdream.net>");
MODULE_AUTHOR("Yehuda Sadeh <yehuda@hq.newdream.net>");
MODULE_DESCRIPTION("rados block device");

/* following authorship retained from original osdblk.c */
MODULE_AUTHOR("Jeff Garzik <jeff@garzik.org>");

MODULE_LICENSE("GPL");