Eric Lee / smarc-fsl-linux-kernel

rbd.c 185 KB
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506 3507 3508 3509 3510 3511 3512 3513 3514 3515 3516 3517 3518 3519 3520 3521 3522 3523 3524 3525 3526 3527 3528 3529 3530 3531 3532 3533 3534 3535 3536 3537 3538 3539 3540 3541 3542 3543 3544 3545 3546 3547 3548 3549 3550 3551 3552 3553 3554 3555 3556 3557 3558 3559 3560 3561 3562 3563 3564 3565 3566 3567 3568 3569 3570 3571 3572 3573 3574 3575 3576 3577 3578 3579 3580 3581 3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612 3613 3614 3615 3616 3617 3618 3619 3620 3621 3622 3623 3624 3625 3626 3627 3628 3629 3630 3631 3632 3633 3634 3635 3636 3637 3638 3639 3640 3641 3642 3643 3644 3645 3646 3647 3648 3649 3650 3651 3652 3653 3654 3655 3656 3657 3658 3659 3660 3661 3662 3663 3664 3665 3666 3667 3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678 3679 3680 3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691 3692 3693 3694 3695 3696 3697 3698 3699 3700 3701 3702 3703 3704 3705 3706 3707 3708 3709 3710 3711 3712 3713 3714 3715 3716 3717 3718 3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729 3730 3731 3732 3733 3734 3735 3736 3737 3738 3739 3740 3741 3742 3743 3744 3745 3746 3747 3748 3749 3750 3751 3752 3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 3763 3764 3765 3766 3767 3768 3769 3770 3771 3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 3787 3788 3789 3790 3791 3792 3793 3794 3795 3796 3797 3798 3799 3800 3801 3802 3803 3804 3805 3806 3807 3808 3809 3810 3811 3812 3813 3814 3815 3816 3817 3818 3819 3820 3821 3822 3823 3824 3825 3826 3827 3828 3829 3830 3831 3832 3833 3834 3835 3836 3837 3838 3839 3840 3841 3842 3843 3844 3845 3846 3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 3867 3868 3869 3870 3871 3872 3873 3874 3875 3876 3877 3878 3879 3880 3881 3882 3883 3884 3885 3886 3887 3888 3889 3890 3891 3892 3893 3894 3895 3896 3897 3898 3899 3900 3901 3902 3903 3904 3905 3906 3907 3908 3909 3910 3911 3912 3913 3914 3915 3916 3917 3918 3919 3920 3921 3922 3923 3924 3925 3926 3927 3928 3929 3930 3931 3932 3933 3934 3935 3936 3937 3938 3939 3940 3941 3942 3943 3944 3945 3946 3947 3948 3949 3950 3951 3952 3953 3954 3955 3956 3957 3958 3959 3960 3961 3962 3963 3964 3965 3966 3967 3968 3969 3970 3971 3972 3973 3974 3975 3976 3977 3978 3979 3980 3981 3982 3983 3984 3985 3986 3987 3988 3989 3990 3991 3992 3993 3994 3995 3996 3997 3998 3999 4000 4001 4002 4003 4004 4005 4006 4007 4008 4009 4010 4011 4012 4013 4014 4015 4016 4017 4018 4019 4020 4021 4022 4023 4024 4025 4026 4027 4028 4029 4030 4031 4032 4033 4034 4035 4036 4037 4038 4039 4040 4041 4042 4043 4044 4045 4046 4047 4048 4049 4050 4051 4052 4053 4054 4055 4056 4057 4058 4059 4060 4061 4062 4063 4064 4065 4066 4067 4068 4069 4070 4071 4072 4073 4074 4075 4076 4077 4078 4079 4080 4081 4082 4083 4084 4085 4086 4087 4088 4089 4090 4091 4092 4093 4094 4095 4096 4097 4098 4099 4100 4101 4102 4103 4104 4105 4106 4107 4108 4109 4110 4111 4112 4113 4114 4115 4116 4117 4118 4119 4120 4121 4122 4123 4124 4125 4126 4127 4128 4129 4130 4131 4132 4133 4134 4135 4136 4137 4138 4139 4140 4141 4142 4143 4144 4145 4146 4147 4148 4149 4150 4151 4152 4153 4154 4155 4156 4157 4158 4159 4160 4161 4162 4163 4164 4165 4166 4167 4168 4169 4170 4171 4172 4173 4174 4175 4176 4177 4178 4179 4180 4181 4182 4183 4184 4185 4186 4187 4188 4189 4190 4191 4192 4193 4194 4195 4196 4197 4198 4199 4200 4201 4202 4203 4204 4205 4206 4207 4208 4209 4210 4211 4212 4213 4214 4215 4216 4217 4218 4219 4220 4221 4222 4223 4224 4225 4226 4227 4228 4229 4230 4231 4232 4233 4234 4235 4236 4237 4238 4239 4240 4241 4242 4243 4244 4245 4246 4247 4248 4249 4250 4251 4252 4253 4254 4255 4256 4257 4258 4259 4260 4261 4262 4263 4264 4265 4266 4267 4268 4269 4270 4271 4272 4273 4274 4275 4276 4277 4278 4279 4280 4281 4282 4283 4284 4285 4286 4287 4288 4289 4290 4291 4292 4293 4294 4295 4296 4297 4298 4299 4300 4301 4302 4303 4304 4305 4306 4307 4308 4309 4310 4311 4312 4313 4314 4315 4316 4317 4318 4319 4320 4321 4322 4323 4324 4325 4326 4327 4328 4329 4330 4331 4332 4333 4334 4335 4336 4337 4338 4339 4340 4341 4342 4343 4344 4345 4346 4347 4348 4349 4350 4351 4352 4353 4354 4355 4356 4357 4358 4359 4360 4361 4362 4363 4364 4365 4366 4367 4368 4369 4370 4371 4372 4373 4374 4375 4376 4377 4378 4379 4380 4381 4382 4383 4384 4385 4386 4387 4388 4389 4390 4391 4392 4393 4394 4395 4396 4397 4398 4399 4400 4401 4402 4403 4404 4405 4406 4407 4408 4409 4410 4411 4412 4413 4414 4415 4416 4417 4418 4419 4420 4421 4422 4423 4424 4425 4426 4427 4428 4429 4430 4431 4432 4433 4434 4435 4436 4437 4438 4439 4440 4441 4442 4443 4444 4445 4446 4447 4448 4449 4450 4451 4452 4453 4454 4455 4456 4457 4458 4459 4460 4461 4462 4463 4464 4465 4466 4467 4468 4469 4470 4471 4472 4473 4474 4475 4476 4477 4478 4479 4480 4481 4482 4483 4484 4485 4486 4487 4488 4489 4490 4491 4492 4493 4494 4495 4496 4497 4498 4499 4500 4501 4502 4503 4504 4505 4506 4507 4508 4509 4510 4511 4512 4513 4514 4515 4516 4517 4518 4519 4520 4521 4522 4523 4524 4525 4526 4527 4528 4529 4530 4531 4532 4533 4534 4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 4545 4546 4547 4548 4549 4550 4551 4552 4553 4554 4555 4556 4557 4558 4559 4560 4561 4562 4563 4564 4565 4566 4567 4568 4569 4570 4571 4572 4573 4574 4575 4576 4577 4578 4579 4580 4581 4582 4583 4584 4585 4586 4587 4588 4589 4590 4591 4592 4593 4594 4595 4596 4597 4598 4599 4600 4601 4602 4603 4604 4605 4606 4607 4608 4609 4610 4611 4612 4613 4614 4615 4616 4617 4618 4619 4620 4621 4622 4623 4624 4625 4626 4627 4628 4629 4630 4631 4632 4633 4634 4635 4636 4637 4638 4639 4640 4641 4642 4643 4644 4645 4646 4647 4648 4649 4650 4651 4652 4653 4654 4655 4656 4657 4658 4659 4660 4661 4662 4663 4664 4665 4666 4667 4668 4669 4670 4671 4672 4673 4674 4675 4676 4677 4678 4679 4680 4681 4682 4683 4684 4685 4686 4687 4688 4689 4690 4691 4692 4693 4694 4695 4696 4697 4698 4699 4700 4701 4702 4703 4704 4705 4706 4707 4708 4709 4710 4711 4712 4713 4714 4715 4716 4717 4718 4719 4720 4721 4722 4723 4724 4725 4726 4727 4728 4729 4730 4731 4732 4733 4734 4735 4736 4737 4738 4739 4740 4741 4742 4743 4744 4745 4746 4747 4748 4749 4750 4751 4752 4753 4754 4755 4756 4757 4758 4759 4760 4761 4762 4763 4764 4765 4766 4767 4768 4769 4770 4771 4772 4773 4774 4775 4776 4777 4778 4779 4780 4781 4782 4783 4784 4785 4786 4787 4788 4789 4790 4791 4792 4793 4794 4795 4796 4797 4798 4799 4800 4801 4802 4803 4804 4805 4806 4807 4808 4809 4810 4811 4812 4813 4814 4815 4816 4817 4818 4819 4820 4821 4822 4823 4824 4825 4826 4827 4828 4829 4830 4831 4832 4833 4834 4835 4836 4837 4838 4839 4840 4841 4842 4843 4844 4845 4846 4847 4848 4849 4850 4851 4852 4853 4854 4855 4856 4857 4858 4859 4860 4861 4862 4863 4864 4865 4866 4867 4868 4869 4870 4871 4872 4873 4874 4875 4876 4877 4878 4879 4880 4881 4882 4883 4884 4885 4886 4887 4888 4889 4890 4891 4892 4893 4894 4895 4896 4897 4898 4899 4900 4901 4902 4903 4904 4905 4906 4907 4908 4909 4910 4911 4912 4913 4914 4915 4916 4917 4918 4919 4920 4921 4922 4923 4924 4925 4926 4927 4928 4929 4930 4931 4932 4933 4934 4935 4936 4937 4938 4939 4940 4941 4942 4943 4944 4945 4946 4947 4948 4949 4950 4951 4952 4953 4954 4955 4956 4957 4958 4959 4960 4961 4962 4963 4964 4965 4966 4967 4968 4969 4970 4971 4972 4973 4974 4975 4976 4977 4978 4979 4980 4981 4982 4983 4984 4985 4986 4987 4988 4989 4990 4991 4992 4993 4994 4995 4996 4997 4998 4999 5000 5001 5002 5003 5004 5005 5006 5007 5008 5009 5010 5011 5012 5013 5014 5015 5016 5017 5018 5019 5020 5021 5022 5023 5024 5025 5026 5027 5028 5029 5030 5031 5032 5033 5034 5035 5036 5037 5038 5039 5040 5041 5042 5043 5044 5045 5046 5047 5048 5049 5050 5051 5052 5053 5054 5055 5056 5057 5058 5059 5060 5061 5062 5063 5064 5065 5066 5067 5068 5069 5070 5071 5072 5073 5074 5075 5076 5077 5078 5079 5080 5081 5082 5083 5084 5085 5086 5087 5088 5089 5090 5091 5092 5093 5094 5095 5096 5097 5098 5099 5100 5101 5102 5103 5104 5105 5106 5107 5108 5109 5110 5111 5112 5113 5114 5115 5116 5117 5118 5119 5120 5121 5122 5123 5124 5125 5126 5127 5128 5129 5130 5131 5132 5133 5134 5135 5136 5137 5138 5139 5140 5141 5142 5143 5144 5145 5146 5147 5148 5149 5150 5151 5152 5153 5154 5155 5156 5157 5158 5159 5160 5161 5162 5163 5164 5165 5166 5167 5168 5169 5170 5171 5172 5173 5174 5175 5176 5177 5178 5179 5180 5181 5182 5183 5184 5185 5186 5187 5188 5189 5190 5191 5192 5193 5194 5195 5196 5197 5198 5199 5200 5201 5202 5203 5204 5205 5206 5207 5208 5209 5210 5211 5212 5213 5214 5215 5216 5217 5218 5219 5220 5221 5222 5223 5224 5225 5226 5227 5228 5229 5230 5231 5232 5233 5234 5235 5236 5237 5238 5239 5240 5241 5242 5243 5244 5245 5246 5247 5248 5249 5250 5251 5252 5253 5254 5255 5256 5257 5258 5259 5260 5261 5262 5263 5264 5265 5266 5267 5268 5269 5270 5271 5272 5273 5274 5275 5276 5277 5278 5279 5280 5281 5282 5283 5284 5285 5286 5287 5288 5289 5290 5291 5292 5293 5294 5295 5296 5297 5298 5299 5300 5301 5302 5303 5304 5305 5306 5307 5308 5309 5310 5311 5312 5313 5314 5315 5316 5317 5318 5319 5320 5321 5322 5323 5324 5325 5326 5327 5328 5329 5330 5331 5332 5333 5334 5335 5336 5337 5338 5339 5340 5341 5342 5343 5344 5345 5346 5347 5348 5349 5350 5351 5352 5353 5354 5355 5356 5357 5358 5359 5360 5361 5362 5363 5364 5365 5366 5367 5368 5369 5370 5371 5372 5373 5374 5375 5376 5377 5378 5379 5380 5381 5382 5383 5384 5385 5386 5387 5388 5389 5390 5391 5392 5393 5394 5395 5396 5397 5398 5399 5400 5401 5402 5403 5404 5405 5406 5407 5408 5409 5410 5411 5412 5413 5414 5415 5416 5417 5418 5419 5420 5421 5422 5423 5424 5425 5426 5427 5428 5429 5430 5431 5432 5433 5434 5435 5436 5437 5438 5439 5440 5441 5442 5443 5444 5445 5446 5447 5448 5449 5450 5451 5452 5453 5454 5455 5456 5457 5458 5459 5460 5461 5462 5463 5464 5465 5466 5467 5468 5469 5470 5471 5472 5473 5474 5475 5476 5477 5478 5479 5480 5481 5482 5483 5484 5485 5486 5487 5488 5489 5490 5491 5492 5493 5494 5495 5496 5497 5498 5499 5500 5501 5502 5503 5504 5505 5506 5507 5508 5509 5510 5511 5512 5513 5514 5515 5516 5517 5518 5519 5520 5521 5522 5523 5524 5525 5526 5527 5528 5529 5530 5531 5532 5533 5534 5535 5536 5537 5538 5539 5540 5541 5542 5543 5544 5545 5546 5547 5548 5549 5550 5551 5552 5553 5554 5555 5556 5557 5558 5559 5560 5561 5562 5563 5564 5565 5566 5567 5568 5569 5570 5571 5572 5573 5574 5575 5576 5577 5578 5579 5580 5581 5582 5583 5584 5585 5586 5587 5588 5589 5590 5591 5592 5593 5594 5595 5596 5597 5598 5599 5600 5601 5602 5603 5604 5605 5606 5607 5608 5609 5610 5611 5612 5613 5614 5615 5616 5617 5618 5619 5620 5621 5622 5623 5624 5625 5626 5627 5628 5629 5630 5631 5632 5633 5634 5635 5636 5637 5638 5639 5640 5641 5642 5643 5644 5645 5646 5647 5648 5649 5650 5651 5652 5653 5654 5655 5656 5657 5658 5659 5660 5661 5662 5663 5664 5665 5666 5667 5668 5669 5670 5671 5672 5673 5674 5675 5676 5677 5678 5679 5680 5681 5682 5683 5684 5685 5686 5687 5688 5689 5690 5691 5692 5693 5694 5695 5696 5697 5698 5699 5700 5701 5702 5703 5704 5705 5706 5707 5708 5709 5710 5711 5712 5713 5714 5715 5716 5717 5718 5719 5720 5721 5722 5723 5724 5725 5726 5727 5728 5729 5730 5731 5732 5733 5734 5735 5736 5737 5738 5739 5740 5741 5742 5743 5744 5745 5746 5747 5748 5749 5750 5751 5752 5753 5754 5755 5756 5757 5758 5759 5760 5761 5762 5763 5764 5765 5766 5767 5768 5769 5770 5771 5772 5773 5774 5775 5776 5777 5778 5779 5780 5781 5782 5783 5784 5785 5786 5787 5788 5789 5790 5791 5792 5793 5794 5795 5796 5797 5798 5799 5800 5801 5802 5803 5804 5805 5806 5807 5808 5809 5810 5811 5812 5813 5814 5815 5816 5817 5818 5819 5820 5821 5822 5823 5824 5825 5826 5827 5828 5829 5830 5831 5832 5833 5834 5835 5836 5837 5838 5839 5840 5841 5842 5843 5844 5845 5846 5847 5848 5849 5850 5851 5852 5853 5854 5855 5856 5857 5858 5859 5860 5861 5862 5863 5864 5865 5866 5867 5868 5869 5870 5871 5872 5873 5874 5875 5876 5877 5878 5879 5880 5881 5882 5883 5884 5885 5886 5887 5888 5889 5890 5891 5892 5893 5894 5895 5896 5897 5898 5899 5900 5901 5902 5903 5904 5905 5906 5907 5908 5909 5910 5911 5912 5913 5914 5915 5916 5917 5918 5919 5920 5921 5922 5923 5924 5925 5926 5927 5928 5929 5930 5931 5932 5933 5934 5935 5936 5937 5938 5939 5940 5941 5942 5943 5944 5945 5946 5947 5948 5949 5950 5951 5952 5953 5954 5955 5956 5957 5958 5959 5960 5961 5962 5963 5964 5965 5966 5967 5968 5969 5970 5971 5972 5973 5974 5975 5976 5977 5978 5979 5980 5981 5982 5983 5984 5985 5986 5987 5988 5989 5990 5991 5992 5993 5994 5995 5996 5997 5998 5999 6000 6001 6002 6003 6004 6005 6006 6007 6008 6009 6010 6011 6012 6013 6014 6015 6016 6017 6018 6019 6020 6021 6022 6023 6024 6025 6026 6027 6028 6029 6030 6031 6032 6033 6034 6035 6036 6037 6038 6039 6040 6041 6042 6043 6044 6045 6046 6047 6048 6049 6050 6051 6052 6053 6054 6055 6056 6057 6058 6059 6060 6061 6062 6063 6064 6065 6066 6067 6068 6069 6070 6071 6072 6073 6074 6075 6076 6077 6078 6079 6080 6081 6082 6083 6084 6085 6086 6087 6088 6089 6090 6091 6092 6093 6094 6095 6096 6097 6098 6099 6100 6101 6102 6103 6104 6105 6106 6107 6108 6109 6110 6111 6112 6113 6114 6115 6116 6117 6118 6119 6120 6121 6122 6123 6124 6125 6126 6127 6128 6129 6130 6131 6132 6133 6134 6135 6136 6137 6138 6139 6140 6141 6142 6143 6144 6145 6146 6147 6148 6149 6150 6151 6152 6153 6154 6155 6156 6157 6158 6159 6160 6161 6162 6163 6164 6165 6166 6167 6168 6169 6170 6171 6172 6173 6174 6175 6176 6177 6178 6179 6180 6181 6182 6183 6184 6185 6186 6187 6188 6189 6190 6191 6192 6193 6194 6195 6196 6197 6198 6199 6200 6201 6202 6203 6204 6205 6206 6207 6208 6209 6210 6211 6212 6213 6214 6215 6216 6217 6218 6219 6220 6221 6222 6223 6224 6225 6226 6227 6228 6229 6230 6231 6232 6233 6234 6235 6236 6237 6238 6239 6240 6241 6242 6243 6244 6245 6246 6247 6248 6249 6250 6251 6252 6253 6254 6255 6256 6257 6258 6259 6260 6261 6262 6263 6264 6265 6266 6267 6268 6269 6270 6271 6272 6273 6274 6275 6276 6277 6278 6279 6280 6281 6282 6283 6284 6285 6286 6287 6288 6289 6290 6291 6292 6293 6294 6295 6296 6297 6298 6299 6300 6301 6302 6303 6304 6305 6306 6307 6308 6309 6310 6311 6312 6313 6314 6315 6316 6317 6318 6319 6320 6321 6322 6323 6324 6325 6326 6327 6328 6329 6330 6331 6332 6333 6334 6335 6336 6337 6338 6339 6340 6341 6342 6343 6344 6345 6346 6347 6348 6349 6350 6351 6352 6353 6354 6355 6356 6357 6358 6359 6360 6361 6362 6363 6364 6365 6366 6367 6368 6369 6370 6371 6372 6373 6374 6375 6376 6377 6378 6379 6380 6381 6382 6383 6384 6385 6386 6387 6388 6389 6390 6391 6392 6393 6394 6395 6396 6397 6398 6399 6400 6401 6402 6403 6404 6405 6406 6407 6408 6409 6410 6411 6412 6413 6414 6415 6416 6417 6418 6419 6420 6421 6422 6423 6424 6425 6426 6427 6428 6429 6430 6431 6432 6433 6434 6435 6436 6437 6438 6439 6440 6441 6442 6443 6444 6445 6446 6447 6448 6449 6450 6451 6452 6453 6454 6455 6456 6457 6458 6459 6460 6461 6462 6463 6464 6465 6466 6467 6468 6469 6470 6471 6472 6473 6474 6475 6476 6477 6478 6479 6480 6481 6482 6483 6484 6485 6486 6487 6488 6489 6490 6491 6492 6493 6494 6495 6496 6497 6498 6499 6500 6501 6502 6503 6504 6505 6506 6507 6508 6509 6510 6511 6512 6513 6514 6515 6516 6517 6518 6519 6520 6521 6522 6523 6524 6525 6526 6527 6528 6529 6530 6531 6532 6533 6534 6535 6536 6537 6538 6539 6540 6541 6542 6543 6544 6545 6546 6547 6548 6549 6550 6551 6552 6553 6554 6555 6556 6557 6558 6559 6560 6561 6562 6563 6564 6565 6566 6567 6568 6569 6570 6571 6572 6573 6574 6575 6576 6577 6578 6579 6580 6581 6582 6583 6584 6585 6586 6587 6588 6589 6590 6591 6592 6593 6594 6595 6596 6597 6598 6599 6600 6601 6602 6603 6604 6605 6606 6607 6608 6609 6610 6611 6612 6613 6614 6615 6616 6617 6618 6619 6620 6621 6622 6623 6624 6625 6626 6627 6628 6629 6630 6631 6632 6633 6634 6635 6636 6637 6638 6639 6640 6641 6642 6643 6644 6645 6646 6647 6648 6649 6650 6651 6652 6653 6654 6655 6656 6657 6658 6659 6660 6661 6662 6663 6664 6665 6666 6667 6668 6669 6670 6671 6672 6673 6674 6675 6676 6677 6678 6679 6680 6681 6682 6683 6684 6685 6686 6687 6688 6689 6690 6691 6692 6693 6694 6695 6696 6697 6698 6699 6700 6701 6702 6703 6704 6705 6706 6707 6708 6709 6710 6711 6712 6713 6714 6715 6716 6717 6718 6719 6720 6721 6722 6723 6724 6725 6726 6727 6728 6729 6730 6731 6732 6733 6734 6735 6736 6737 6738 6739 6740 6741 6742 6743 6744 6745 6746 6747 6748 6749 6750 6751 6752 6753 6754 6755 6756 6757 6758 6759 6760 6761 6762 6763 6764 6765 6766 6767 6768 6769 6770 6771 6772 6773 6774 6775 6776 6777 6778 6779 6780 6781 6782 6783 6784 6785 6786 6787 6788 6789 6790 6791 6792 6793 6794 6795 6796 6797 6798 6799 6800 6801 6802 6803 6804 6805 6806 6807 6808 6809 6810 6811 6812 6813 6814 6815 6816 6817 6818 6819 6820 6821 6822 6823 6824 6825 6826 6827 6828 6829 6830 6831 6832 6833 6834 6835 6836 6837 6838 6839 6840 6841 6842 6843 6844 6845 6846 6847 6848 6849 6850 6851 6852 6853 6854 6855 6856 6857 6858 6859 6860 6861 6862 6863 6864 6865 6866 6867 6868 6869 6870 6871 6872 6873 6874 6875 6876 6877 6878 6879 6880 6881 6882 6883 6884 6885 6886 6887 6888 6889 6890 6891 6892 6893 6894 6895 6896 6897 6898 6899 6900 6901 6902 6903 6904 6905 6906 6907 6908 6909 6910 6911 6912 6913 6914 6915 6916 6917 6918 6919 6920 6921 6922 6923 6924 6925 6926 6927 6928 6929 6930 6931 6932 6933 6934 6935 6936 6937 6938 6939 6940 6941 6942 6943 6944 6945 6946 6947 6948 6949 6950 6951 6952 6953 6954 6955 6956 6957 6958 6959 6960 6961 6962 6963 6964 6965 6966 6967 6968 6969 6970 6971 6972 6973 6974 6975 6976 6977 6978 6979 6980 6981 6982 6983 6984 6985 6986 6987 6988 6989 6990 6991 6992 6993 6994 6995 6996 6997 6998 6999 7000 7001 7002 7003 7004 7005 7006 7007 7008 7009 7010 7011 7012 7013 7014 7015 7016 7017 7018 7019 7020 7021 7022 7023 7024 7025 7026 7027 7028 7029 7030 7031 7032 7033 7034 7035 7036 7037 7038 7039 7040 7041 7042 7043 7044 7045 7046 7047 7048 7049 7050 7051 7052 7053 7054 7055 7056 7057 7058 7059 7060 7061 7062 7063 7064 7065 7066 7067 7068 7069 7070 7071 7072 7073 7074 7075 7076 7077 7078 7079 7080 7081 7082 7083 7084 7085 7086 7087 7088 7089 7090 7091 7092 7093 7094 7095 7096 7097 7098 7099 7100 7101 7102 7103 7104 7105 7106 7107 7108 7109 7110 7111 7112 7113 7114 7115 7116 7117 7118 7119 7120 7121 7122 7123 7124 7125 7126 7127 7128 7129 7130 7131 7132 7133 7134 7135 7136 7137 7138 7139 7140 7141 7142 7143 7144 7145 7146 7147 7148 7149 7150 7151 7152 7153 7154 7155 7156 7157 7158 7159 7160 7161 7162 7163 7164 7165 7166 7167 7168 7169 7170 7171 7172 7173 7174 7175 7176 7177 7178 7179 7180 7181 7182 7183 7184 7185 7186 7187 7188 7189 7190 7191 7192 7193 7194 7195 7196 7197 7198 7199 7200 7201 7202 7203 7204 7205 7206 7207 7208 7209 7210 7211 7212 7213 7214 7215 7216 7217 7218 7219 7220 7221 7222 7223 7224 7225 7226 7227 7228 7229 7230 7231 7232 7233 7234 7235 7236 7237 7238 7239 7240 7241 7242 7243 7244 7245 7246 7247 7248 7249 7250 7251 7252 7253 7254 7255 7256 7257 7258 7259 7260 7261 7262 7263 7264 7265 7266 7267 7268 7269 7270 7271 7272 7273 7274 7275 7276 7277 7278 7279 7280 7281 7282 7283 7284 7285 7286 7287 7288 7289 7290 7291 7292 7293 7294 7295 7296 7297 7298 7299 7300 7301 7302 7303 7304 7305 7306 7307 7308 7309 7310 7311 7312 7313 7314 7315 7316 7317 7318 7319 7320 7321 7322 7323 7324 7325 7326 7327 7328 7329 7330 7331 7332 7333 7334 7335 7336 7337 7338 7339 7340 7341 7342 7343 7344 7345 7346 7347 7348 7349 7350 7351 7352 7353 7354 7355 7356 7357 7358 7359 7360 7361 7362 7363 7364 7365 7366 7367 7368 7369 7370 7371 7372 7373 7374 7375 7376 7377 7378 7379 7380 7381 7382 7383 7384 7385 7386 7387 7388 7389 7390 7391 7392 7393 7394 7395 7396 7397 7398 7399 7400 7401 7402 7403 7404 7405 7406 7407 7408 7409 7410 7411 7412 7413 7414 7415 7416 7417 7418 7419 7420 7421 7422 7423 7424 7425 7426 
/*
   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/cls_lock_client.h>
#include <linux/ceph/striper.h>
#include <linux/ceph/decode.h>
#include <linux/fs_parser.h>
#include <linux/bsearch.h>

#include <linux/kernel.h>
#include <linux/device.h>
#include <linux/module.h>
#include <linux/blk-mq.h>
#include <linux/fs.h>
#include <linux/blkdev.h>
#include <linux/slab.h>
#include <linux/idr.h>
#include <linux/workqueue.h>

#include "rbd_types.h"

#define RBD_DEBUG	/* Activate rbd_assert() calls */

/*
 * Increment the given counter and return its updated value.
 * If the counter is already 0 it will not be incremented.
 * If the counter is already at its maximum value returns
 * -EINVAL without updating it.
 */
static int atomic_inc_return_safe(atomic_t *v)
{
	unsigned int counter;

	counter = (unsigned int)atomic_fetch_add_unless(v, 1, 0);
	if (counter <= (unsigned int)INT_MAX)
		return (int)counter;

	atomic_dec(v);

	return -EINVAL;
}

/* Decrement the counter.  Return the resulting value, or -EINVAL */
static int atomic_dec_return_safe(atomic_t *v)
{
	int counter;

	counter = atomic_dec_return(v);
	if (counter >= 0)
		return counter;

	atomic_inc(v);

	return -EINVAL;
}

#define RBD_DRV_NAME "rbd"

#define RBD_MINORS_PER_MAJOR		256
#define RBD_SINGLE_MAJOR_PART_SHIFT	4

#define RBD_MAX_PARENT_CHAIN_LEN	16

#define RBD_SNAP_DEV_NAME_PREFIX	"snap_"
#define RBD_MAX_SNAP_NAME_LEN	\
			(NAME_MAX - (sizeof (RBD_SNAP_DEV_NAME_PREFIX) - 1))

#define RBD_MAX_SNAP_COUNT	510	/* allows max snapc to fit in 4KB */

#define RBD_SNAP_HEAD_NAME	"-"

#define	BAD_SNAP_INDEX	U32_MAX		/* invalid index into snap array */

/* This allows a single page to hold an image name sent by OSD */
#define RBD_IMAGE_NAME_LEN_MAX	(PAGE_SIZE - sizeof (__le32) - 1)
#define RBD_IMAGE_ID_LEN_MAX	64

#define RBD_OBJ_PREFIX_LEN_MAX	64

#define RBD_NOTIFY_TIMEOUT	5	/* seconds */
#define RBD_RETRY_DELAY		msecs_to_jiffies(1000)

/* Feature bits */

#define RBD_FEATURE_LAYERING		(1ULL<<0)
#define RBD_FEATURE_STRIPINGV2		(1ULL<<1)
#define RBD_FEATURE_EXCLUSIVE_LOCK	(1ULL<<2)
#define RBD_FEATURE_OBJECT_MAP		(1ULL<<3)
#define RBD_FEATURE_FAST_DIFF		(1ULL<<4)
#define RBD_FEATURE_DEEP_FLATTEN	(1ULL<<5)
#define RBD_FEATURE_DATA_POOL		(1ULL<<7)
#define RBD_FEATURE_OPERATIONS		(1ULL<<8)

#define RBD_FEATURES_ALL	(RBD_FEATURE_LAYERING |		\
				 RBD_FEATURE_STRIPINGV2 |	\
				 RBD_FEATURE_EXCLUSIVE_LOCK |	\
				 RBD_FEATURE_OBJECT_MAP |	\
				 RBD_FEATURE_FAST_DIFF |	\
				 RBD_FEATURE_DEEP_FLATTEN |	\
				 RBD_FEATURE_DATA_POOL |	\
				 RBD_FEATURE_OPERATIONS)

/* Features supported by this (client software) implementation. */

#define RBD_FEATURES_SUPPORTED	(RBD_FEATURES_ALL)

/*
 * An RBD device name will be "rbd#", where the "rbd" comes from
 * RBD_DRV_NAME above, and # is a unique integer identifier.
 */
#define DEV_NAME_LEN		32

/*
 * block device image metadata (in-memory version)
 */
struct rbd_image_header {
	/* These six fields never change for a given rbd image */
	char *object_prefix;
	__u8 obj_order;
	u64 stripe_unit;
	u64 stripe_count;
	s64 data_pool_id;
	u64 features;		/* Might be changeable someday? */

	/* The remaining fields need to be updated occasionally */
	u64 image_size;
	struct ceph_snap_context *snapc;
	char *snap_names;	/* format 1 only */
	u64 *snap_sizes;	/* format 1 only */
};

/*
 * An rbd image specification.
 *
 * The tuple (pool_id, image_id, snap_id) is sufficient to uniquely
 * identify an image.  Each rbd_dev structure includes a pointer to
 * an rbd_spec structure that encapsulates this identity.
 *
 * Each of the id's in an rbd_spec has an associated name.  For a
 * user-mapped image, the names are supplied and the id's associated
 * with them are looked up.  For a layered image, a parent image is
 * defined by the tuple, and the names are looked up.
 *
 * An rbd_dev structure contains a parent_spec pointer which is
 * non-null if the image it represents is a child in a layered
 * image.  This pointer will refer to the rbd_spec structure used
 * by the parent rbd_dev for its own identity (i.e., the structure
 * is shared between the parent and child).
 *
 * Since these structures are populated once, during the discovery
 * phase of image construction, they are effectively immutable so
 * we make no effort to synchronize access to them.
 *
 * Note that code herein does not assume the image name is known (it
 * could be a null pointer).
 */
struct rbd_spec {
	u64		pool_id;
	const char	*pool_name;
	const char	*pool_ns;	/* NULL if default, never "" */

	const char	*image_id;
	const char	*image_name;

	u64		snap_id;
	const char	*snap_name;

	struct kref	kref;
};

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

struct pending_result {
	int			result;		/* first nonzero result */
	int			num_pending;
};

struct rbd_img_request;

enum obj_request_type {
	OBJ_REQUEST_NODATA = 1,
	OBJ_REQUEST_BIO,	/* pointer into provided bio (list) */
	OBJ_REQUEST_BVECS,	/* pointer into provided bio_vec array */
	OBJ_REQUEST_OWN_BVECS,	/* private bio_vec array, doesn't own pages */
};

enum obj_operation_type {
	OBJ_OP_READ = 1,
	OBJ_OP_WRITE,
	OBJ_OP_DISCARD,
	OBJ_OP_ZEROOUT,
};

#define RBD_OBJ_FLAG_DELETION			(1U << 0)
#define RBD_OBJ_FLAG_COPYUP_ENABLED		(1U << 1)
#define RBD_OBJ_FLAG_COPYUP_ZEROS		(1U << 2)
#define RBD_OBJ_FLAG_MAY_EXIST			(1U << 3)
#define RBD_OBJ_FLAG_NOOP_FOR_NONEXISTENT	(1U << 4)

enum rbd_obj_read_state {
	RBD_OBJ_READ_START = 1,
	RBD_OBJ_READ_OBJECT,
	RBD_OBJ_READ_PARENT,
};

/*
 * Writes go through the following state machine to deal with
 * layering:
 *
 *            . . . . . RBD_OBJ_WRITE_GUARD. . . . . . . . . . . . . .
 *            .                 |                                    .
 *            .                 v                                    .
 *            .    RBD_OBJ_WRITE_READ_FROM_PARENT. . .               .
 *            .                 |                    .               .
 *            .                 v                    v (deep-copyup  .
 *    (image  .   RBD_OBJ_WRITE_COPYUP_EMPTY_SNAPC   .  not needed)  .
 * flattened) v                 |                    .               .
 *            .                 v                    .               .
 *            . . . .RBD_OBJ_WRITE_COPYUP_OPS. . . . .      (copyup  .
 *                              |                        not needed) v
 *                              v                                    .
 *                            done . . . . . . . . . . . . . . . . . .
 *                              ^
 *                              |
 *                     RBD_OBJ_WRITE_FLAT
 *
 * Writes start in RBD_OBJ_WRITE_GUARD or _FLAT, depending on whether
 * assert_exists guard is needed or not (in some cases it's not needed
 * even if there is a parent).
 */
enum rbd_obj_write_state {
	RBD_OBJ_WRITE_START = 1,
	RBD_OBJ_WRITE_PRE_OBJECT_MAP,
	RBD_OBJ_WRITE_OBJECT,
	__RBD_OBJ_WRITE_COPYUP,
	RBD_OBJ_WRITE_COPYUP,
	RBD_OBJ_WRITE_POST_OBJECT_MAP,
};

enum rbd_obj_copyup_state {
	RBD_OBJ_COPYUP_START = 1,
	RBD_OBJ_COPYUP_READ_PARENT,
	__RBD_OBJ_COPYUP_OBJECT_MAPS,
	RBD_OBJ_COPYUP_OBJECT_MAPS,
	__RBD_OBJ_COPYUP_WRITE_OBJECT,
	RBD_OBJ_COPYUP_WRITE_OBJECT,
};

struct rbd_obj_request {
	struct ceph_object_extent ex;
	unsigned int		flags;	/* RBD_OBJ_FLAG_* */
	union {
		enum rbd_obj_read_state	 read_state;	/* for reads */
		enum rbd_obj_write_state write_state;	/* for writes */
	};

	struct rbd_img_request	*img_request;
	struct ceph_file_extent	*img_extents;
	u32			num_img_extents;

	union {
		struct ceph_bio_iter	bio_pos;
		struct {
			struct ceph_bvec_iter	bvec_pos;
			u32			bvec_count;
			u32			bvec_idx;
		};
	};

	enum rbd_obj_copyup_state copyup_state;
	struct bio_vec		*copyup_bvecs;
	u32			copyup_bvec_count;

	struct list_head	osd_reqs;	/* w/ r_private_item */

	struct mutex		state_mutex;
	struct pending_result	pending;
	struct kref		kref;
};

enum img_req_flags {
	IMG_REQ_CHILD,		/* initiator: block = 0, child image = 1 */
	IMG_REQ_LAYERED,	/* ENOENT handling: normal = 0, layered = 1 */
};

enum rbd_img_state {
	RBD_IMG_START = 1,
	RBD_IMG_EXCLUSIVE_LOCK,
	__RBD_IMG_OBJECT_REQUESTS,
	RBD_IMG_OBJECT_REQUESTS,
};

struct rbd_img_request {
	struct rbd_device	*rbd_dev;
	enum obj_operation_type	op_type;
	enum obj_request_type	data_type;
	unsigned long		flags;
	enum rbd_img_state	state;
	union {
		u64			snap_id;	/* for reads */
		struct ceph_snap_context *snapc;	/* for writes */
	};
	struct rbd_obj_request	*obj_request;	/* obj req initiator */

	struct list_head	lock_item;
	struct list_head	object_extents;	/* obj_req.ex structs */

	struct mutex		state_mutex;
	struct pending_result	pending;
	struct work_struct	work;
	int			work_result;
};

#define for_each_obj_request(ireq, oreq) \
	list_for_each_entry(oreq, &(ireq)->object_extents, ex.oe_item)
#define for_each_obj_request_safe(ireq, oreq, n) \
	list_for_each_entry_safe(oreq, n, &(ireq)->object_extents, ex.oe_item)

enum rbd_watch_state {
	RBD_WATCH_STATE_UNREGISTERED,
	RBD_WATCH_STATE_REGISTERED,
	RBD_WATCH_STATE_ERROR,
};

enum rbd_lock_state {
	RBD_LOCK_STATE_UNLOCKED,
	RBD_LOCK_STATE_LOCKED,
	RBD_LOCK_STATE_RELEASING,
};

/* WatchNotify::ClientId */
struct rbd_client_id {
	u64 gid;
	u64 handle;
};

struct rbd_mapping {
	u64                     size;
};

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

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

	u32			image_format;	/* Either 1 or 2 */
	struct rbd_client	*rbd_client;

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

	spinlock_t		lock;		/* queue, flags, open_count */

	struct rbd_image_header	header;
	unsigned long		flags;		/* possibly lock protected */
	struct rbd_spec		*spec;
	struct rbd_options	*opts;
	char			*config_info;	/* add{,_single_major} string */

	struct ceph_object_id	header_oid;
	struct ceph_object_locator header_oloc;

	struct ceph_file_layout	layout;		/* used for all rbd requests */

	struct mutex		watch_mutex;
	enum rbd_watch_state	watch_state;
	struct ceph_osd_linger_request *watch_handle;
	u64			watch_cookie;
	struct delayed_work	watch_dwork;

	struct rw_semaphore	lock_rwsem;
	enum rbd_lock_state	lock_state;
	char			lock_cookie[32];
	struct rbd_client_id	owner_cid;
	struct work_struct	acquired_lock_work;
	struct work_struct	released_lock_work;
	struct delayed_work	lock_dwork;
	struct work_struct	unlock_work;
	spinlock_t		lock_lists_lock;
	struct list_head	acquiring_list;
	struct list_head	running_list;
	struct completion	acquire_wait;
	int			acquire_err;
	struct completion	releasing_wait;

	spinlock_t		object_map_lock;
	u8			*object_map;
	u64			object_map_size;	/* in objects */
	u64			object_map_flags;

	struct workqueue_struct	*task_wq;

	struct rbd_spec		*parent_spec;
	u64			parent_overlap;
	atomic_t		parent_ref;
	struct rbd_device	*parent;

	/* Block layer tags. */
	struct blk_mq_tag_set	tag_set;

	/* protects updating the header */
	struct rw_semaphore     header_rwsem;

	struct rbd_mapping	mapping;

	struct list_head	node;

	/* sysfs related */
	struct device		dev;
	unsigned long		open_count;	/* protected by lock */
};

/*
 * Flag bits for rbd_dev->flags:
 * - REMOVING (which is coupled with rbd_dev->open_count) is protected
 *   by rbd_dev->lock
 */
enum rbd_dev_flags {
	RBD_DEV_FLAG_EXISTS,	/* rbd_dev_device_setup() ran */
	RBD_DEV_FLAG_REMOVING,	/* this mapping is being removed */
	RBD_DEV_FLAG_READONLY,  /* -o ro or snapshot */
};

static DEFINE_MUTEX(client_mutex);	/* Serialize client creation */

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);

/* Slab caches for frequently-allocated structures */

static struct kmem_cache	*rbd_img_request_cache;
static struct kmem_cache	*rbd_obj_request_cache;

static int rbd_major;
static DEFINE_IDA(rbd_dev_id_ida);

static struct workqueue_struct *rbd_wq;

static struct ceph_snap_context rbd_empty_snapc = {
	.nref = REFCOUNT_INIT(1),
};

/*
 * single-major requires >= 0.75 version of userspace rbd utility.
 */
static bool single_major = true;
module_param(single_major, bool, 0444);
MODULE_PARM_DESC(single_major, "Use a single major number for all rbd devices (default: true)");

static ssize_t add_store(struct bus_type *bus, const char *buf, size_t count);
static ssize_t remove_store(struct bus_type *bus, const char *buf,
			    size_t count);
static ssize_t add_single_major_store(struct bus_type *bus, const char *buf,
				      size_t count);
static ssize_t remove_single_major_store(struct bus_type *bus, const char *buf,
					 size_t count);
static int rbd_dev_image_probe(struct rbd_device *rbd_dev, int depth);

static int rbd_dev_id_to_minor(int dev_id)
{
	return dev_id << RBD_SINGLE_MAJOR_PART_SHIFT;
}

static int minor_to_rbd_dev_id(int minor)
{
	return minor >> RBD_SINGLE_MAJOR_PART_SHIFT;
}

static bool rbd_is_ro(struct rbd_device *rbd_dev)
{
	return test_bit(RBD_DEV_FLAG_READONLY, &rbd_dev->flags);
}

static bool rbd_is_snap(struct rbd_device *rbd_dev)
{
	return rbd_dev->spec->snap_id != CEPH_NOSNAP;
}

static bool __rbd_is_lock_owner(struct rbd_device *rbd_dev)
{
	lockdep_assert_held(&rbd_dev->lock_rwsem);

	return rbd_dev->lock_state == RBD_LOCK_STATE_LOCKED ||
	       rbd_dev->lock_state == RBD_LOCK_STATE_RELEASING;
}

static bool rbd_is_lock_owner(struct rbd_device *rbd_dev)
{
	bool is_lock_owner;

	down_read(&rbd_dev->lock_rwsem);
	is_lock_owner = __rbd_is_lock_owner(rbd_dev);
	up_read(&rbd_dev->lock_rwsem);
	return is_lock_owner;
}

static ssize_t supported_features_show(struct bus_type *bus, char *buf)
{
	return sprintf(buf, "0x%llx\n", RBD_FEATURES_SUPPORTED);
}

static BUS_ATTR_WO(add);
static BUS_ATTR_WO(remove);
static BUS_ATTR_WO(add_single_major);
static BUS_ATTR_WO(remove_single_major);
static BUS_ATTR_RO(supported_features);

static struct attribute *rbd_bus_attrs[] = {
	&bus_attr_add.attr,
	&bus_attr_remove.attr,
	&bus_attr_add_single_major.attr,
	&bus_attr_remove_single_major.attr,
	&bus_attr_supported_features.attr,
	NULL,
};

static umode_t rbd_bus_is_visible(struct kobject *kobj,
				  struct attribute *attr, int index)
{
	if (!single_major &&
	    (attr == &bus_attr_add_single_major.attr ||
	     attr == &bus_attr_remove_single_major.attr))
		return 0;

	return attr->mode;
}

static const struct attribute_group rbd_bus_group = {
	.attrs = rbd_bus_attrs,
	.is_visible = rbd_bus_is_visible,
};
__ATTRIBUTE_GROUPS(rbd_bus);

static struct bus_type rbd_bus_type = {
	.name		= "rbd",
	.bus_groups	= rbd_bus_groups,
};

static void rbd_root_dev_release(struct device *dev)
{
}

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

static __printf(2, 3)
void rbd_warn(struct rbd_device *rbd_dev, const char *fmt, ...)
{
	struct va_format vaf;
	va_list args;

	va_start(args, fmt);
	vaf.fmt = fmt;
	vaf.va = &args;

	if (!rbd_dev)
		printk(KERN_WARNING "%s: %pV\n", RBD_DRV_NAME, &vaf);
	else if (rbd_dev->disk)
		printk(KERN_WARNING "%s: %s: %pV\n",
			RBD_DRV_NAME, rbd_dev->disk->disk_name, &vaf);
	else if (rbd_dev->spec && rbd_dev->spec->image_name)
		printk(KERN_WARNING "%s: image %s: %pV\n",
			RBD_DRV_NAME, rbd_dev->spec->image_name, &vaf);
	else if (rbd_dev->spec && rbd_dev->spec->image_id)
		printk(KERN_WARNING "%s: id %s: %pV\n",
			RBD_DRV_NAME, rbd_dev->spec->image_id, &vaf);
	else	/* punt */
		printk(KERN_WARNING "%s: rbd_dev %p: %pV\n",
			RBD_DRV_NAME, rbd_dev, &vaf);
	va_end(args);
}

#ifdef RBD_DEBUG
#define rbd_assert(expr)						\
		if (unlikely(!(expr))) {				\
			printk(KERN_ERR "\nAssertion failure in %s() "	\
						"at line %d:\n\n"	\
					"\trbd_assert(%s);\n\n",	\
					__func__, __LINE__, #expr);	\
			BUG();						\
		}
#else /* !RBD_DEBUG */
#  define rbd_assert(expr)	((void) 0)
#endif /* !RBD_DEBUG */

static void rbd_dev_remove_parent(struct rbd_device *rbd_dev);

static int rbd_dev_refresh(struct rbd_device *rbd_dev);
static int rbd_dev_v2_header_onetime(struct rbd_device *rbd_dev);
static int rbd_dev_header_info(struct rbd_device *rbd_dev);
static int rbd_dev_v2_parent_info(struct rbd_device *rbd_dev);
static const char *rbd_dev_v2_snap_name(struct rbd_device *rbd_dev,
					u64 snap_id);
static int _rbd_dev_v2_snap_size(struct rbd_device *rbd_dev, u64 snap_id,
				u8 *order, u64 *snap_size);
static int rbd_dev_v2_get_flags(struct rbd_device *rbd_dev);

static void rbd_obj_handle_request(struct rbd_obj_request *obj_req, int result);
static void rbd_img_handle_request(struct rbd_img_request *img_req, int result);

/*
 * Return true if nothing else is pending.
 */
static bool pending_result_dec(struct pending_result *pending, int *result)
{
	rbd_assert(pending->num_pending > 0);

	if (*result && !pending->result)
		pending->result = *result;
	if (--pending->num_pending)
		return false;

	*result = pending->result;
	return true;
}

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

	spin_lock_irq(&rbd_dev->lock);
	if (test_bit(RBD_DEV_FLAG_REMOVING, &rbd_dev->flags))
		removing = true;
	else
		rbd_dev->open_count++;
	spin_unlock_irq(&rbd_dev->lock);
	if (removing)
		return -ENOENT;

	(void) get_device(&rbd_dev->dev);

	return 0;
}

static void rbd_release(struct gendisk *disk, fmode_t mode)
{
	struct rbd_device *rbd_dev = disk->private_data;
	unsigned long open_count_before;

	spin_lock_irq(&rbd_dev->lock);
	open_count_before = rbd_dev->open_count--;
	spin_unlock_irq(&rbd_dev->lock);
	rbd_assert(open_count_before > 0);

	put_device(&rbd_dev->dev);
}

static int rbd_ioctl_set_ro(struct rbd_device *rbd_dev, unsigned long arg)
{
	int ro;

	if (get_user(ro, (int __user *)arg))
		return -EFAULT;

	/*
	 * Both images mapped read-only and snapshots can't be marked
	 * read-write.
	 */
	if (!ro) {
		if (rbd_is_ro(rbd_dev))
			return -EROFS;

		rbd_assert(!rbd_is_snap(rbd_dev));
	}

	/* Let blkdev_roset() handle it */
	return -ENOTTY;
}

static int rbd_ioctl(struct block_device *bdev, fmode_t mode,
			unsigned int cmd, unsigned long arg)
{
	struct rbd_device *rbd_dev = bdev->bd_disk->private_data;
	int ret;

	switch (cmd) {
	case BLKROSET:
		ret = rbd_ioctl_set_ro(rbd_dev, arg);
		break;
	default:
		ret = -ENOTTY;
	}

	return ret;
}

#ifdef CONFIG_COMPAT
static int rbd_compat_ioctl(struct block_device *bdev, fmode_t mode,
				unsigned int cmd, unsigned long arg)
{
	return rbd_ioctl(bdev, mode, cmd, arg);
}
#endif /* CONFIG_COMPAT */

static const struct block_device_operations rbd_bd_ops = {
	.owner			= THIS_MODULE,
	.open			= rbd_open,
	.release		= rbd_release,
	.ioctl			= rbd_ioctl,
#ifdef CONFIG_COMPAT
	.compat_ioctl		= rbd_compat_ioctl,
#endif
};

/*
 * Initialize an rbd client instance.  Success or not, this function
 * consumes ceph_opts.  Caller holds client_mutex.
 */
static struct rbd_client *rbd_client_create(struct ceph_options *ceph_opts)
{
	struct rbd_client *rbdc;
	int ret = -ENOMEM;

	dout("%s:\n", __func__);
	rbdc = kmalloc(sizeof(struct rbd_client), GFP_KERNEL);
	if (!rbdc)
		goto out_opt;

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

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

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

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

	dout("%s: rbdc %p\n", __func__, rbdc);

	return rbdc;
out_client:
	ceph_destroy_client(rbdc->client);
out_rbdc:
	kfree(rbdc);
out_opt:
	if (ceph_opts)
		ceph_destroy_options(ceph_opts);
	dout("%s: error %d\n", __func__, ret);

	return ERR_PTR(ret);
}

static struct rbd_client *__rbd_get_client(struct rbd_client *rbdc)
{
	kref_get(&rbdc->kref);

	return rbdc;
}

/*
 * Find a ceph client with specific addr and configuration.  If
 * found, bump its reference count.
 */
static struct rbd_client *rbd_client_find(struct ceph_options *ceph_opts)
{
	struct rbd_client *client_node;
	bool found = false;

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

	spin_lock(&rbd_client_list_lock);
	list_for_each_entry(client_node, &rbd_client_list, node) {
		if (!ceph_compare_options(ceph_opts, client_node->client)) {
			__rbd_get_client(client_node);

			found = true;
			break;
		}
	}
	spin_unlock(&rbd_client_list_lock);

	return found ? client_node : NULL;
}

/*
 * (Per device) rbd map options
 */
enum {
	Opt_queue_depth,
	Opt_alloc_size,
	Opt_lock_timeout,
	/* int args above */
	Opt_pool_ns,
	Opt_compression_hint,
	/* string args above */
	Opt_read_only,
	Opt_read_write,
	Opt_lock_on_read,
	Opt_exclusive,
	Opt_notrim,
};

enum {
	Opt_compression_hint_none,
	Opt_compression_hint_compressible,
	Opt_compression_hint_incompressible,
};

static const struct constant_table rbd_param_compression_hint[] = {
	{"none",		Opt_compression_hint_none},
	{"compressible",	Opt_compression_hint_compressible},
	{"incompressible",	Opt_compression_hint_incompressible},
	{}
};

static const struct fs_parameter_spec rbd_parameters[] = {
	fsparam_u32	("alloc_size",			Opt_alloc_size),
	fsparam_enum	("compression_hint",		Opt_compression_hint,
			 rbd_param_compression_hint),
	fsparam_flag	("exclusive",			Opt_exclusive),
	fsparam_flag	("lock_on_read",		Opt_lock_on_read),
	fsparam_u32	("lock_timeout",		Opt_lock_timeout),
	fsparam_flag	("notrim",			Opt_notrim),
	fsparam_string	("_pool_ns",			Opt_pool_ns),
	fsparam_u32	("queue_depth",			Opt_queue_depth),
	fsparam_flag	("read_only",			Opt_read_only),
	fsparam_flag	("read_write",			Opt_read_write),
	fsparam_flag	("ro",				Opt_read_only),
	fsparam_flag	("rw",				Opt_read_write),
	{}
};

struct rbd_options {
	int	queue_depth;
	int	alloc_size;
	unsigned long	lock_timeout;
	bool	read_only;
	bool	lock_on_read;
	bool	exclusive;
	bool	trim;

	u32 alloc_hint_flags;  /* CEPH_OSD_OP_ALLOC_HINT_FLAG_* */
};

#define RBD_QUEUE_DEPTH_DEFAULT	BLKDEV_MAX_RQ
#define RBD_ALLOC_SIZE_DEFAULT	(64 * 1024)
#define RBD_LOCK_TIMEOUT_DEFAULT 0  /* no timeout */
#define RBD_READ_ONLY_DEFAULT	false
#define RBD_LOCK_ON_READ_DEFAULT false
#define RBD_EXCLUSIVE_DEFAULT	false
#define RBD_TRIM_DEFAULT	true

struct rbd_parse_opts_ctx {
	struct rbd_spec		*spec;
	struct ceph_options	*copts;
	struct rbd_options	*opts;
};

static char* obj_op_name(enum obj_operation_type op_type)
{
	switch (op_type) {
	case OBJ_OP_READ:
		return "read";
	case OBJ_OP_WRITE:
		return "write";
	case OBJ_OP_DISCARD:
		return "discard";
	case OBJ_OP_ZEROOUT:
		return "zeroout";
	default:
		return "???";
	}
}

/*
 * 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("%s: rbdc %p\n", __func__, rbdc);
	spin_lock(&rbd_client_list_lock);
	list_del(&rbdc->node);
	spin_unlock(&rbd_client_list_lock);

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

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

/*
 * Get a ceph client with specific addr and configuration, if one does
 * not exist create it.  Either way, ceph_opts is consumed by this
 * function.
 */
static struct rbd_client *rbd_get_client(struct ceph_options *ceph_opts)
{
	struct rbd_client *rbdc;
	int ret;

	mutex_lock(&client_mutex);
	rbdc = rbd_client_find(ceph_opts);
	if (rbdc) {
		ceph_destroy_options(ceph_opts);

		/*
		 * Using an existing client.  Make sure ->pg_pools is up to
		 * date before we look up the pool id in do_rbd_add().
		 */
		ret = ceph_wait_for_latest_osdmap(rbdc->client,
					rbdc->client->options->mount_timeout);
		if (ret) {
			rbd_warn(NULL, "failed to get latest osdmap: %d", ret);
			rbd_put_client(rbdc);
			rbdc = ERR_PTR(ret);
		}
	} else {
		rbdc = rbd_client_create(ceph_opts);
	}
	mutex_unlock(&client_mutex);

	return rbdc;
}

static bool rbd_image_format_valid(u32 image_format)
{
	return image_format == 1 || image_format == 2;
}

static bool rbd_dev_ondisk_valid(struct rbd_image_header_ondisk *ondisk)
{
	size_t size;
	u32 snap_count;

	/* The header has to start with the magic rbd header text */
	if (memcmp(&ondisk->text, RBD_HEADER_TEXT, sizeof (RBD_HEADER_TEXT)))
		return false;

	/* The bio layer requires at least sector-sized I/O */

	if (ondisk->options.order < SECTOR_SHIFT)
		return false;

	/* If we use u64 in a few spots we may be able to loosen this */

	if (ondisk->options.order > 8 * sizeof (int) - 1)
		return false;

	/*
	 * The size of a snapshot header has to fit in a size_t, and
	 * that limits the number of snapshots.
	 */
	snap_count = le32_to_cpu(ondisk->snap_count);
	size = SIZE_MAX - sizeof (struct ceph_snap_context);
	if (snap_count > size / sizeof (__le64))
		return false;

	/*
	 * Not only that, but the size of the entire the snapshot
	 * header must also be representable in a size_t.
	 */
	size -= snap_count * sizeof (__le64);
	if ((u64) size < le64_to_cpu(ondisk->snap_names_len))
		return false;

	return true;
}

/*
 * returns the size of an object in the image
 */
static u32 rbd_obj_bytes(struct rbd_image_header *header)
{
	return 1U << header->obj_order;
}

static void rbd_init_layout(struct rbd_device *rbd_dev)
{
	if (rbd_dev->header.stripe_unit == 0 ||
	    rbd_dev->header.stripe_count == 0) {
		rbd_dev->header.stripe_unit = rbd_obj_bytes(&rbd_dev->header);
		rbd_dev->header.stripe_count = 1;
	}

	rbd_dev->layout.stripe_unit = rbd_dev->header.stripe_unit;
	rbd_dev->layout.stripe_count = rbd_dev->header.stripe_count;
	rbd_dev->layout.object_size = rbd_obj_bytes(&rbd_dev->header);
	rbd_dev->layout.pool_id = rbd_dev->header.data_pool_id == CEPH_NOPOOL ?
			  rbd_dev->spec->pool_id : rbd_dev->header.data_pool_id;
	RCU_INIT_POINTER(rbd_dev->layout.pool_ns, NULL);
}

/*
 * Fill an rbd image header with information from the given format 1
 * on-disk header.
 */
static int rbd_header_from_disk(struct rbd_device *rbd_dev,
				 struct rbd_image_header_ondisk *ondisk)
{
	struct rbd_image_header *header = &rbd_dev->header;
	bool first_time = header->object_prefix == NULL;
	struct ceph_snap_context *snapc;
	char *object_prefix = NULL;
	char *snap_names = NULL;
	u64 *snap_sizes = NULL;
	u32 snap_count;
	int ret = -ENOMEM;
	u32 i;

	/* Allocate this now to avoid having to handle failure below */

	if (first_time) {
		object_prefix = kstrndup(ondisk->object_prefix,
					 sizeof(ondisk->object_prefix),
					 GFP_KERNEL);
		if (!object_prefix)
			return -ENOMEM;
	}

	/* Allocate the snapshot context and fill it in */

	snap_count = le32_to_cpu(ondisk->snap_count);
	snapc = ceph_create_snap_context(snap_count, GFP_KERNEL);
	if (!snapc)
		goto out_err;
	snapc->seq = le64_to_cpu(ondisk->snap_seq);
	if (snap_count) {
		struct rbd_image_snap_ondisk *snaps;
		u64 snap_names_len = le64_to_cpu(ondisk->snap_names_len);

		/* We'll keep a copy of the snapshot names... */

		if (snap_names_len > (u64)SIZE_MAX)
			goto out_2big;
		snap_names = kmalloc(snap_names_len, GFP_KERNEL);
		if (!snap_names)
			goto out_err;

		/* ...as well as the array of their sizes. */
		snap_sizes = kmalloc_array(snap_count,
					   sizeof(*header->snap_sizes),
					   GFP_KERNEL);
		if (!snap_sizes)
			goto out_err;

		/*
		 * Copy the names, and fill in each snapshot's id
		 * and size.
		 *
		 * Note that rbd_dev_v1_header_info() guarantees the
		 * ondisk buffer we're working with has
		 * snap_names_len bytes beyond the end of the
		 * snapshot id array, this memcpy() is safe.
		 */
		memcpy(snap_names, &ondisk->snaps[snap_count], snap_names_len);
		snaps = ondisk->snaps;
		for (i = 0; i < snap_count; i++) {
			snapc->snaps[i] = le64_to_cpu(snaps[i].id);
			snap_sizes[i] = le64_to_cpu(snaps[i].image_size);
		}
	}

	/* We won't fail any more, fill in the header */

	if (first_time) {
		header->object_prefix = object_prefix;
		header->obj_order = ondisk->options.order;
		rbd_init_layout(rbd_dev);
	} else {
		ceph_put_snap_context(header->snapc);
		kfree(header->snap_names);
		kfree(header->snap_sizes);
	}

	/* The remaining fields always get updated (when we refresh) */

	header->image_size = le64_to_cpu(ondisk->image_size);
	header->snapc = snapc;
	header->snap_names = snap_names;
	header->snap_sizes = snap_sizes;

	return 0;
out_2big:
	ret = -EIO;
out_err:
	kfree(snap_sizes);
	kfree(snap_names);
	ceph_put_snap_context(snapc);
	kfree(object_prefix);

	return ret;
}

static const char *_rbd_dev_v1_snap_name(struct rbd_device *rbd_dev, u32 which)
{
	const char *snap_name;

	rbd_assert(which < rbd_dev->header.snapc->num_snaps);

	/* Skip over names until we find the one we are looking for */

	snap_name = rbd_dev->header.snap_names;
	while (which--)
		snap_name += strlen(snap_name) + 1;

	return kstrdup(snap_name, GFP_KERNEL);
}

/*
 * Snapshot id comparison function for use with qsort()/bsearch().
 * Note that result is for snapshots in *descending* order.
 */
static int snapid_compare_reverse(const void *s1, const void *s2)
{
	u64 snap_id1 = *(u64 *)s1;
	u64 snap_id2 = *(u64 *)s2;

	if (snap_id1 < snap_id2)
		return 1;
	return snap_id1 == snap_id2 ? 0 : -1;
}

/*
 * Search a snapshot context to see if the given snapshot id is
 * present.
 *
 * Returns the position of the snapshot id in the array if it's found,
 * or BAD_SNAP_INDEX otherwise.
 *
 * Note: The snapshot array is in kept sorted (by the osd) in
 * reverse order, highest snapshot id first.
 */
static u32 rbd_dev_snap_index(struct rbd_device *rbd_dev, u64 snap_id)
{
	struct ceph_snap_context *snapc = rbd_dev->header.snapc;
	u64 *found;

	found = bsearch(&snap_id, &snapc->snaps, snapc->num_snaps,
				sizeof (snap_id), snapid_compare_reverse);

	return found ? (u32)(found - &snapc->snaps[0]) : BAD_SNAP_INDEX;
}

static const char *rbd_dev_v1_snap_name(struct rbd_device *rbd_dev,
					u64 snap_id)
{
	u32 which;
	const char *snap_name;

	which = rbd_dev_snap_index(rbd_dev, snap_id);
	if (which == BAD_SNAP_INDEX)
		return ERR_PTR(-ENOENT);

	snap_name = _rbd_dev_v1_snap_name(rbd_dev, which);
	return snap_name ? snap_name : ERR_PTR(-ENOMEM);
}

static const char *rbd_snap_name(struct rbd_device *rbd_dev, u64 snap_id)
{
	if (snap_id == CEPH_NOSNAP)
		return RBD_SNAP_HEAD_NAME;

	rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
	if (rbd_dev->image_format == 1)
		return rbd_dev_v1_snap_name(rbd_dev, snap_id);

	return rbd_dev_v2_snap_name(rbd_dev, snap_id);
}

static int rbd_snap_size(struct rbd_device *rbd_dev, u64 snap_id,
				u64 *snap_size)
{
	rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
	if (snap_id == CEPH_NOSNAP) {
		*snap_size = rbd_dev->header.image_size;
	} else if (rbd_dev->image_format == 1) {
		u32 which;

		which = rbd_dev_snap_index(rbd_dev, snap_id);
		if (which == BAD_SNAP_INDEX)
			return -ENOENT;

		*snap_size = rbd_dev->header.snap_sizes[which];
	} else {
		u64 size = 0;
		int ret;

		ret = _rbd_dev_v2_snap_size(rbd_dev, snap_id, NULL, &size);
		if (ret)
			return ret;

		*snap_size = size;
	}
	return 0;
}

static int rbd_dev_mapping_set(struct rbd_device *rbd_dev)
{
	u64 snap_id = rbd_dev->spec->snap_id;
	u64 size = 0;
	int ret;

	ret = rbd_snap_size(rbd_dev, snap_id, &size);
	if (ret)
		return ret;

	rbd_dev->mapping.size = size;
	return 0;
}

static void rbd_dev_mapping_clear(struct rbd_device *rbd_dev)
{
	rbd_dev->mapping.size = 0;
}

static void zero_bvec(struct bio_vec *bv)
{
	void *buf;
	unsigned long flags;

	buf = bvec_kmap_irq(bv, &flags);
	memset(buf, 0, bv->bv_len);
	flush_dcache_page(bv->bv_page);
	bvec_kunmap_irq(buf, &flags);
}

static void zero_bios(struct ceph_bio_iter *bio_pos, u32 off, u32 bytes)
{
	struct ceph_bio_iter it = *bio_pos;

	ceph_bio_iter_advance(&it, off);
	ceph_bio_iter_advance_step(&it, bytes, ({
		zero_bvec(&bv);
	}));
}

static void zero_bvecs(struct ceph_bvec_iter *bvec_pos, u32 off, u32 bytes)
{
	struct ceph_bvec_iter it = *bvec_pos;

	ceph_bvec_iter_advance(&it, off);
	ceph_bvec_iter_advance_step(&it, bytes, ({
		zero_bvec(&bv);
	}));
}

/*
 * Zero a range in @obj_req data buffer defined by a bio (list) or
 * (private) bio_vec array.
 *
 * @off is relative to the start of the data buffer.
 */
static void rbd_obj_zero_range(struct rbd_obj_request *obj_req, u32 off,
			       u32 bytes)
{
	dout("%s %p data buf %u~%u\n", __func__, obj_req, off, bytes);

	switch (obj_req->img_request->data_type) {
	case OBJ_REQUEST_BIO:
		zero_bios(&obj_req->bio_pos, off, bytes);
		break;
	case OBJ_REQUEST_BVECS:
	case OBJ_REQUEST_OWN_BVECS:
		zero_bvecs(&obj_req->bvec_pos, off, bytes);
		break;
	default:
		BUG();
	}
}

static void rbd_obj_request_destroy(struct kref *kref);
static void rbd_obj_request_put(struct rbd_obj_request *obj_request)
{
	rbd_assert(obj_request != NULL);
	dout("%s: obj %p (was %d)\n", __func__, obj_request,
		kref_read(&obj_request->kref));
	kref_put(&obj_request->kref, rbd_obj_request_destroy);
}

static inline void rbd_img_obj_request_add(struct rbd_img_request *img_request,
					struct rbd_obj_request *obj_request)
{
	rbd_assert(obj_request->img_request == NULL);

	/* Image request now owns object's original reference */
	obj_request->img_request = img_request;
	dout("%s: img %p obj %p\n", __func__, img_request, obj_request);
}

static inline void rbd_img_obj_request_del(struct rbd_img_request *img_request,
					struct rbd_obj_request *obj_request)
{
	dout("%s: img %p obj %p\n", __func__, img_request, obj_request);
	list_del(&obj_request->ex.oe_item);
	rbd_assert(obj_request->img_request == img_request);
	rbd_obj_request_put(obj_request);
}

static void rbd_osd_submit(struct ceph_osd_request *osd_req)
{
	struct rbd_obj_request *obj_req = osd_req->r_priv;

	dout("%s osd_req %p for obj_req %p objno %llu %llu~%llu\n",
	     __func__, osd_req, obj_req, obj_req->ex.oe_objno,
	     obj_req->ex.oe_off, obj_req->ex.oe_len);
	ceph_osdc_start_request(osd_req->r_osdc, osd_req, false);
}

/*
 * The default/initial value for all image request flags is 0.  Each
 * is conditionally set to 1 at image request initialization time
 * and currently never change thereafter.
 */
static void img_request_layered_set(struct rbd_img_request *img_request)
{
	set_bit(IMG_REQ_LAYERED, &img_request->flags);
}

static bool img_request_layered_test(struct rbd_img_request *img_request)
{
	return test_bit(IMG_REQ_LAYERED, &img_request->flags) != 0;
}

static bool rbd_obj_is_entire(struct rbd_obj_request *obj_req)
{
	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;

	return !obj_req->ex.oe_off &&
	       obj_req->ex.oe_len == rbd_dev->layout.object_size;
}

static bool rbd_obj_is_tail(struct rbd_obj_request *obj_req)
{
	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;

	return obj_req->ex.oe_off + obj_req->ex.oe_len ==
					rbd_dev->layout.object_size;
}

/*
 * Must be called after rbd_obj_calc_img_extents().
 */
static bool rbd_obj_copyup_enabled(struct rbd_obj_request *obj_req)
{
	if (!obj_req->num_img_extents ||
	    (rbd_obj_is_entire(obj_req) &&
	     !obj_req->img_request->snapc->num_snaps))
		return false;

	return true;
}

static u64 rbd_obj_img_extents_bytes(struct rbd_obj_request *obj_req)
{
	return ceph_file_extents_bytes(obj_req->img_extents,
				       obj_req->num_img_extents);
}

static bool rbd_img_is_write(struct rbd_img_request *img_req)
{
	switch (img_req->op_type) {
	case OBJ_OP_READ:
		return false;
	case OBJ_OP_WRITE:
	case OBJ_OP_DISCARD:
	case OBJ_OP_ZEROOUT:
		return true;
	default:
		BUG();
	}
}

static void rbd_osd_req_callback(struct ceph_osd_request *osd_req)
{
	struct rbd_obj_request *obj_req = osd_req->r_priv;
	int result;

	dout("%s osd_req %p result %d for obj_req %p\n", __func__, osd_req,
	     osd_req->r_result, obj_req);

	/*
	 * Writes aren't allowed to return a data payload.  In some
	 * guarded write cases (e.g. stat + zero on an empty object)
	 * a stat response makes it through, but we don't care.
	 */
	if (osd_req->r_result > 0 && rbd_img_is_write(obj_req->img_request))
		result = 0;
	else
		result = osd_req->r_result;

	rbd_obj_handle_request(obj_req, result);
}

static void rbd_osd_format_read(struct ceph_osd_request *osd_req)
{
	struct rbd_obj_request *obj_request = osd_req->r_priv;
	struct rbd_device *rbd_dev = obj_request->img_request->rbd_dev;
	struct ceph_options *opt = rbd_dev->rbd_client->client->options;

	osd_req->r_flags = CEPH_OSD_FLAG_READ | opt->read_from_replica;
	osd_req->r_snapid = obj_request->img_request->snap_id;
}

static void rbd_osd_format_write(struct ceph_osd_request *osd_req)
{
	struct rbd_obj_request *obj_request = osd_req->r_priv;

	osd_req->r_flags = CEPH_OSD_FLAG_WRITE;
	ktime_get_real_ts64(&osd_req->r_mtime);
	osd_req->r_data_offset = obj_request->ex.oe_off;
}

static struct ceph_osd_request *
__rbd_obj_add_osd_request(struct rbd_obj_request *obj_req,
			  struct ceph_snap_context *snapc, int num_ops)
{
	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
	struct ceph_osd_request *req;
	const char *name_format = rbd_dev->image_format == 1 ?
				      RBD_V1_DATA_FORMAT : RBD_V2_DATA_FORMAT;
	int ret;

	req = ceph_osdc_alloc_request(osdc, snapc, num_ops, false, GFP_NOIO);
	if (!req)
		return ERR_PTR(-ENOMEM);

	list_add_tail(&req->r_private_item, &obj_req->osd_reqs);
	req->r_callback = rbd_osd_req_callback;
	req->r_priv = obj_req;

	/*
	 * Data objects may be stored in a separate pool, but always in
	 * the same namespace in that pool as the header in its pool.
	 */
	ceph_oloc_copy(&req->r_base_oloc, &rbd_dev->header_oloc);
	req->r_base_oloc.pool = rbd_dev->layout.pool_id;

	ret = ceph_oid_aprintf(&req->r_base_oid, GFP_NOIO, name_format,
			       rbd_dev->header.object_prefix,
			       obj_req->ex.oe_objno);
	if (ret)
		return ERR_PTR(ret);

	return req;
}

static struct ceph_osd_request *
rbd_obj_add_osd_request(struct rbd_obj_request *obj_req, int num_ops)
{
	return __rbd_obj_add_osd_request(obj_req, obj_req->img_request->snapc,
					 num_ops);
}

static struct rbd_obj_request *rbd_obj_request_create(void)
{
	struct rbd_obj_request *obj_request;

	obj_request = kmem_cache_zalloc(rbd_obj_request_cache, GFP_NOIO);
	if (!obj_request)
		return NULL;

	ceph_object_extent_init(&obj_request->ex);
	INIT_LIST_HEAD(&obj_request->osd_reqs);
	mutex_init(&obj_request->state_mutex);
	kref_init(&obj_request->kref);

	dout("%s %p\n", __func__, obj_request);
	return obj_request;
}

static void rbd_obj_request_destroy(struct kref *kref)
{
	struct rbd_obj_request *obj_request;
	struct ceph_osd_request *osd_req;
	u32 i;

	obj_request = container_of(kref, struct rbd_obj_request, kref);

	dout("%s: obj %p\n", __func__, obj_request);

	while (!list_empty(&obj_request->osd_reqs)) {
		osd_req = list_first_entry(&obj_request->osd_reqs,
				    struct ceph_osd_request, r_private_item);
		list_del_init(&osd_req->r_private_item);
		ceph_osdc_put_request(osd_req);
	}

	switch (obj_request->img_request->data_type) {
	case OBJ_REQUEST_NODATA:
	case OBJ_REQUEST_BIO:
	case OBJ_REQUEST_BVECS:
		break;		/* Nothing to do */
	case OBJ_REQUEST_OWN_BVECS:
		kfree(obj_request->bvec_pos.bvecs);
		break;
	default:
		BUG();
	}

	kfree(obj_request->img_extents);
	if (obj_request->copyup_bvecs) {
		for (i = 0; i < obj_request->copyup_bvec_count; i++) {
			if (obj_request->copyup_bvecs[i].bv_page)
				__free_page(obj_request->copyup_bvecs[i].bv_page);
		}
		kfree(obj_request->copyup_bvecs);
	}

	kmem_cache_free(rbd_obj_request_cache, obj_request);
}

/* It's OK to call this for a device with no parent */

static void rbd_spec_put(struct rbd_spec *spec);
static void rbd_dev_unparent(struct rbd_device *rbd_dev)
{
	rbd_dev_remove_parent(rbd_dev);
	rbd_spec_put(rbd_dev->parent_spec);
	rbd_dev->parent_spec = NULL;
	rbd_dev->parent_overlap = 0;
}

/*
 * Parent image reference counting is used to determine when an
 * image's parent fields can be safely torn down--after there are no
 * more in-flight requests to the parent image.  When the last
 * reference is dropped, cleaning them up is safe.
 */
static void rbd_dev_parent_put(struct rbd_device *rbd_dev)
{
	int counter;

	if (!rbd_dev->parent_spec)
		return;

	counter = atomic_dec_return_safe(&rbd_dev->parent_ref);
	if (counter > 0)
		return;

	/* Last reference; clean up parent data structures */

	if (!counter)
		rbd_dev_unparent(rbd_dev);
	else
		rbd_warn(rbd_dev, "parent reference underflow");
}

/*
 * If an image has a non-zero parent overlap, get a reference to its
 * parent.
 *
 * Returns true if the rbd device has a parent with a non-zero
 * overlap and a reference for it was successfully taken, or
 * false otherwise.
 */
static bool rbd_dev_parent_get(struct rbd_device *rbd_dev)
{
	int counter = 0;

	if (!rbd_dev->parent_spec)
		return false;

	if (rbd_dev->parent_overlap)
		counter = atomic_inc_return_safe(&rbd_dev->parent_ref);

	if (counter < 0)
		rbd_warn(rbd_dev, "parent reference overflow");

	return counter > 0;
}

static void rbd_img_request_init(struct rbd_img_request *img_request,
				 struct rbd_device *rbd_dev,
				 enum obj_operation_type op_type)
{
	memset(img_request, 0, sizeof(*img_request));

	img_request->rbd_dev = rbd_dev;
	img_request->op_type = op_type;

	INIT_LIST_HEAD(&img_request->lock_item);
	INIT_LIST_HEAD(&img_request->object_extents);
	mutex_init(&img_request->state_mutex);
}

static void rbd_img_capture_header(struct rbd_img_request *img_req)
{
	struct rbd_device *rbd_dev = img_req->rbd_dev;

	lockdep_assert_held(&rbd_dev->header_rwsem);

	if (rbd_img_is_write(img_req))
		img_req->snapc = ceph_get_snap_context(rbd_dev->header.snapc);
	else
		img_req->snap_id = rbd_dev->spec->snap_id;

	if (rbd_dev_parent_get(rbd_dev))
		img_request_layered_set(img_req);
}

static void rbd_img_request_destroy(struct rbd_img_request *img_request)
{
	struct rbd_obj_request *obj_request;
	struct rbd_obj_request *next_obj_request;

	dout("%s: img %p\n", __func__, img_request);

	WARN_ON(!list_empty(&img_request->lock_item));
	for_each_obj_request_safe(img_request, obj_request, next_obj_request)
		rbd_img_obj_request_del(img_request, obj_request);

	if (img_request_layered_test(img_request))
		rbd_dev_parent_put(img_request->rbd_dev);

	if (rbd_img_is_write(img_request))
		ceph_put_snap_context(img_request->snapc);

	if (test_bit(IMG_REQ_CHILD, &img_request->flags))
		kmem_cache_free(rbd_img_request_cache, img_request);
}

#define BITS_PER_OBJ	2
#define OBJS_PER_BYTE	(BITS_PER_BYTE / BITS_PER_OBJ)
#define OBJ_MASK	((1 << BITS_PER_OBJ) - 1)

static void __rbd_object_map_index(struct rbd_device *rbd_dev, u64 objno,
				   u64 *index, u8 *shift)
{
	u32 off;

	rbd_assert(objno < rbd_dev->object_map_size);
	*index = div_u64_rem(objno, OBJS_PER_BYTE, &off);
	*shift = (OBJS_PER_BYTE - off - 1) * BITS_PER_OBJ;
}

static u8 __rbd_object_map_get(struct rbd_device *rbd_dev, u64 objno)
{
	u64 index;
	u8 shift;

	lockdep_assert_held(&rbd_dev->object_map_lock);
	__rbd_object_map_index(rbd_dev, objno, &index, &shift);
	return (rbd_dev->object_map[index] >> shift) & OBJ_MASK;
}

static void __rbd_object_map_set(struct rbd_device *rbd_dev, u64 objno, u8 val)
{
	u64 index;
	u8 shift;
	u8 *p;

	lockdep_assert_held(&rbd_dev->object_map_lock);
	rbd_assert(!(val & ~OBJ_MASK));

	__rbd_object_map_index(rbd_dev, objno, &index, &shift);
	p = &rbd_dev->object_map[index];
	*p = (*p & ~(OBJ_MASK << shift)) | (val << shift);
}

static u8 rbd_object_map_get(struct rbd_device *rbd_dev, u64 objno)
{
	u8 state;

	spin_lock(&rbd_dev->object_map_lock);
	state = __rbd_object_map_get(rbd_dev, objno);
	spin_unlock(&rbd_dev->object_map_lock);
	return state;
}

static bool use_object_map(struct rbd_device *rbd_dev)
{
	/*
	 * An image mapped read-only can't use the object map -- it isn't
	 * loaded because the header lock isn't acquired.  Someone else can
	 * write to the image and update the object map behind our back.
	 *
	 * A snapshot can't be written to, so using the object map is always
	 * safe.
	 */
	if (!rbd_is_snap(rbd_dev) && rbd_is_ro(rbd_dev))
		return false;

	return ((rbd_dev->header.features & RBD_FEATURE_OBJECT_MAP) &&
		!(rbd_dev->object_map_flags & RBD_FLAG_OBJECT_MAP_INVALID));
}

static bool rbd_object_map_may_exist(struct rbd_device *rbd_dev, u64 objno)
{
	u8 state;

	/* fall back to default logic if object map is disabled or invalid */
	if (!use_object_map(rbd_dev))
		return true;

	state = rbd_object_map_get(rbd_dev, objno);
	return state != OBJECT_NONEXISTENT;
}

static void rbd_object_map_name(struct rbd_device *rbd_dev, u64 snap_id,
				struct ceph_object_id *oid)
{
	if (snap_id == CEPH_NOSNAP)
		ceph_oid_printf(oid, "%s%s", RBD_OBJECT_MAP_PREFIX,
				rbd_dev->spec->image_id);
	else
		ceph_oid_printf(oid, "%s%s.%016llx", RBD_OBJECT_MAP_PREFIX,
				rbd_dev->spec->image_id, snap_id);
}

static int rbd_object_map_lock(struct rbd_device *rbd_dev)
{
	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
	CEPH_DEFINE_OID_ONSTACK(oid);
	u8 lock_type;
	char *lock_tag;
	struct ceph_locker *lockers;
	u32 num_lockers;
	bool broke_lock = false;
	int ret;

	rbd_object_map_name(rbd_dev, CEPH_NOSNAP, &oid);

again:
	ret = ceph_cls_lock(osdc, &oid, &rbd_dev->header_oloc, RBD_LOCK_NAME,
			    CEPH_CLS_LOCK_EXCLUSIVE, "", "", "", 0);
	if (ret != -EBUSY || broke_lock) {
		if (ret == -EEXIST)
			ret = 0; /* already locked by myself */
		if (ret)
			rbd_warn(rbd_dev, "failed to lock object map: %d", ret);
		return ret;
	}

	ret = ceph_cls_lock_info(osdc, &oid, &rbd_dev->header_oloc,
				 RBD_LOCK_NAME, &lock_type, &lock_tag,
				 &lockers, &num_lockers);
	if (ret) {
		if (ret == -ENOENT)
			goto again;

		rbd_warn(rbd_dev, "failed to get object map lockers: %d", ret);
		return ret;
	}

	kfree(lock_tag);
	if (num_lockers == 0)
		goto again;

	rbd_warn(rbd_dev, "breaking object map lock owned by %s%llu",
		 ENTITY_NAME(lockers[0].id.name));

	ret = ceph_cls_break_lock(osdc, &oid, &rbd_dev->header_oloc,
				  RBD_LOCK_NAME, lockers[0].id.cookie,
				  &lockers[0].id.name);
	ceph_free_lockers(lockers, num_lockers);
	if (ret) {
		if (ret == -ENOENT)
			goto again;

		rbd_warn(rbd_dev, "failed to break object map lock: %d", ret);
		return ret;
	}

	broke_lock = true;
	goto again;
}

static void rbd_object_map_unlock(struct rbd_device *rbd_dev)
{
	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
	CEPH_DEFINE_OID_ONSTACK(oid);
	int ret;

	rbd_object_map_name(rbd_dev, CEPH_NOSNAP, &oid);

	ret = ceph_cls_unlock(osdc, &oid, &rbd_dev->header_oloc, RBD_LOCK_NAME,
			      "");
	if (ret && ret != -ENOENT)
		rbd_warn(rbd_dev, "failed to unlock object map: %d", ret);
}

static int decode_object_map_header(void **p, void *end, u64 *object_map_size)
{
	u8 struct_v;
	u32 struct_len;
	u32 header_len;
	void *header_end;
	int ret;

	ceph_decode_32_safe(p, end, header_len, e_inval);
	header_end = *p + header_len;

	ret = ceph_start_decoding(p, end, 1, "BitVector header", &struct_v,
				  &struct_len);
	if (ret)
		return ret;

	ceph_decode_64_safe(p, end, *object_map_size, e_inval);

	*p = header_end;
	return 0;

e_inval:
	return -EINVAL;
}

static int __rbd_object_map_load(struct rbd_device *rbd_dev)
{
	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
	CEPH_DEFINE_OID_ONSTACK(oid);
	struct page **pages;
	void *p, *end;
	size_t reply_len;
	u64 num_objects;
	u64 object_map_bytes;
	u64 object_map_size;
	int num_pages;
	int ret;

	rbd_assert(!rbd_dev->object_map && !rbd_dev->object_map_size);

	num_objects = ceph_get_num_objects(&rbd_dev->layout,
					   rbd_dev->mapping.size);
	object_map_bytes = DIV_ROUND_UP_ULL(num_objects * BITS_PER_OBJ,
					    BITS_PER_BYTE);
	num_pages = calc_pages_for(0, object_map_bytes) + 1;
	pages = ceph_alloc_page_vector(num_pages, GFP_KERNEL);
	if (IS_ERR(pages))
		return PTR_ERR(pages);

	reply_len = num_pages * PAGE_SIZE;
	rbd_object_map_name(rbd_dev, rbd_dev->spec->snap_id, &oid);
	ret = ceph_osdc_call(osdc, &oid, &rbd_dev->header_oloc,
			     "rbd", "object_map_load", CEPH_OSD_FLAG_READ,
			     NULL, 0, pages, &reply_len);
	if (ret)
		goto out;

	p = page_address(pages[0]);
	end = p + min(reply_len, (size_t)PAGE_SIZE);
	ret = decode_object_map_header(&p, end, &object_map_size);
	if (ret)
		goto out;

	if (object_map_size != num_objects) {
		rbd_warn(rbd_dev, "object map size mismatch: %llu vs %llu",
			 object_map_size, num_objects);
		ret = -EINVAL;
		goto out;
	}

	if (offset_in_page(p) + object_map_bytes > reply_len) {
		ret = -EINVAL;
		goto out;
	}

	rbd_dev->object_map = kvmalloc(object_map_bytes, GFP_KERNEL);
	if (!rbd_dev->object_map) {
		ret = -ENOMEM;
		goto out;
	}

	rbd_dev->object_map_size = object_map_size;
	ceph_copy_from_page_vector(pages, rbd_dev->object_map,
				   offset_in_page(p), object_map_bytes);

out:
	ceph_release_page_vector(pages, num_pages);
	return ret;
}

static void rbd_object_map_free(struct rbd_device *rbd_dev)
{
	kvfree(rbd_dev->object_map);
	rbd_dev->object_map = NULL;
	rbd_dev->object_map_size = 0;
}

static int rbd_object_map_load(struct rbd_device *rbd_dev)
{
	int ret;

	ret = __rbd_object_map_load(rbd_dev);
	if (ret)
		return ret;

	ret = rbd_dev_v2_get_flags(rbd_dev);
	if (ret) {
		rbd_object_map_free(rbd_dev);
		return ret;
	}

	if (rbd_dev->object_map_flags & RBD_FLAG_OBJECT_MAP_INVALID)
		rbd_warn(rbd_dev, "object map is invalid");

	return 0;
}

static int rbd_object_map_open(struct rbd_device *rbd_dev)
{
	int ret;

	ret = rbd_object_map_lock(rbd_dev);
	if (ret)
		return ret;

	ret = rbd_object_map_load(rbd_dev);
	if (ret) {
		rbd_object_map_unlock(rbd_dev);
		return ret;
	}

	return 0;
}

static void rbd_object_map_close(struct rbd_device *rbd_dev)
{
	rbd_object_map_free(rbd_dev);
	rbd_object_map_unlock(rbd_dev);
}

/*
 * This function needs snap_id (or more precisely just something to
 * distinguish between HEAD and snapshot object maps), new_state and
 * current_state that were passed to rbd_object_map_update().
 *
 * To avoid allocating and stashing a context we piggyback on the OSD
 * request.  A HEAD update has two ops (assert_locked).  For new_state
 * and current_state we decode our own object_map_update op, encoded in
 * rbd_cls_object_map_update().
 */
static int rbd_object_map_update_finish(struct rbd_obj_request *obj_req,
					struct ceph_osd_request *osd_req)
{
	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
	struct ceph_osd_data *osd_data;
	u64 objno;
	u8 state, new_state, current_state;
	bool has_current_state;
	void *p;

	if (osd_req->r_result)
		return osd_req->r_result;

	/*
	 * Nothing to do for a snapshot object map.
	 */
	if (osd_req->r_num_ops == 1)
		return 0;

	/*
	 * Update in-memory HEAD object map.
	 */
	rbd_assert(osd_req->r_num_ops == 2);
	osd_data = osd_req_op_data(osd_req, 1, cls, request_data);
	rbd_assert(osd_data->type == CEPH_OSD_DATA_TYPE_PAGES);

	p = page_address(osd_data->pages[0]);
	objno = ceph_decode_64(&p);
	rbd_assert(objno == obj_req->ex.oe_objno);
	rbd_assert(ceph_decode_64(&p) == objno + 1);
	new_state = ceph_decode_8(&p);
	has_current_state = ceph_decode_8(&p);
	if (has_current_state)
		current_state = ceph_decode_8(&p);

	spin_lock(&rbd_dev->object_map_lock);
	state = __rbd_object_map_get(rbd_dev, objno);
	if (!has_current_state || current_state == state ||
	    (current_state == OBJECT_EXISTS && state == OBJECT_EXISTS_CLEAN))
		__rbd_object_map_set(rbd_dev, objno, new_state);
	spin_unlock(&rbd_dev->object_map_lock);

	return 0;
}

static void rbd_object_map_callback(struct ceph_osd_request *osd_req)
{
	struct rbd_obj_request *obj_req = osd_req->r_priv;
	int result;

	dout("%s osd_req %p result %d for obj_req %p\n", __func__, osd_req,
	     osd_req->r_result, obj_req);

	result = rbd_object_map_update_finish(obj_req, osd_req);
	rbd_obj_handle_request(obj_req, result);
}

static bool update_needed(struct rbd_device *rbd_dev, u64 objno, u8 new_state)
{
	u8 state = rbd_object_map_get(rbd_dev, objno);

	if (state == new_state ||
	    (new_state == OBJECT_PENDING && state == OBJECT_NONEXISTENT) ||
	    (new_state == OBJECT_NONEXISTENT && state != OBJECT_PENDING))
		return false;

	return true;
}

static int rbd_cls_object_map_update(struct ceph_osd_request *req,
				     int which, u64 objno, u8 new_state,
				     const u8 *current_state)
{
	struct page **pages;
	void *p, *start;
	int ret;

	ret = osd_req_op_cls_init(req, which, "rbd", "object_map_update");
	if (ret)
		return ret;

	pages = ceph_alloc_page_vector(1, GFP_NOIO);
	if (IS_ERR(pages))
		return PTR_ERR(pages);

	p = start = page_address(pages[0]);
	ceph_encode_64(&p, objno);
	ceph_encode_64(&p, objno + 1);
	ceph_encode_8(&p, new_state);
	if (current_state) {
		ceph_encode_8(&p, 1);
		ceph_encode_8(&p, *current_state);
	} else {
		ceph_encode_8(&p, 0);
	}

	osd_req_op_cls_request_data_pages(req, which, pages, p - start, 0,
					  false, true);
	return 0;
}

/*
 * Return:
 *   0 - object map update sent
 *   1 - object map update isn't needed
 *  <0 - error
 */
static int rbd_object_map_update(struct rbd_obj_request *obj_req, u64 snap_id,
				 u8 new_state, const u8 *current_state)
{
	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
	struct ceph_osd_request *req;
	int num_ops = 1;
	int which = 0;
	int ret;

	if (snap_id == CEPH_NOSNAP) {
		if (!update_needed(rbd_dev, obj_req->ex.oe_objno, new_state))
			return 1;

		num_ops++; /* assert_locked */
	}

	req = ceph_osdc_alloc_request(osdc, NULL, num_ops, false, GFP_NOIO);
	if (!req)
		return -ENOMEM;

	list_add_tail(&req->r_private_item, &obj_req->osd_reqs);
	req->r_callback = rbd_object_map_callback;
	req->r_priv = obj_req;

	rbd_object_map_name(rbd_dev, snap_id, &req->r_base_oid);
	ceph_oloc_copy(&req->r_base_oloc, &rbd_dev->header_oloc);
	req->r_flags = CEPH_OSD_FLAG_WRITE;
	ktime_get_real_ts64(&req->r_mtime);

	if (snap_id == CEPH_NOSNAP) {
		/*
		 * Protect against possible race conditions during lock
		 * ownership transitions.
		 */
		ret = ceph_cls_assert_locked(req, which++, RBD_LOCK_NAME,
					     CEPH_CLS_LOCK_EXCLUSIVE, "", "");
		if (ret)
			return ret;
	}

	ret = rbd_cls_object_map_update(req, which, obj_req->ex.oe_objno,
					new_state, current_state);
	if (ret)
		return ret;

	ret = ceph_osdc_alloc_messages(req, GFP_NOIO);
	if (ret)
		return ret;

	ceph_osdc_start_request(osdc, req, false);
	return 0;
}

static void prune_extents(struct ceph_file_extent *img_extents,
			  u32 *num_img_extents, u64 overlap)
{
	u32 cnt = *num_img_extents;

	/* drop extents completely beyond the overlap */
	while (cnt && img_extents[cnt - 1].fe_off >= overlap)
		cnt--;

	if (cnt) {
		struct ceph_file_extent *ex = &img_extents[cnt - 1];

		/* trim final overlapping extent */
		if (ex->fe_off + ex->fe_len > overlap)
			ex->fe_len = overlap - ex->fe_off;
	}

	*num_img_extents = cnt;
}

/*
 * Determine the byte range(s) covered by either just the object extent
 * or the entire object in the parent image.
 */
static int rbd_obj_calc_img_extents(struct rbd_obj_request *obj_req,
				    bool entire)
{
	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
	int ret;

	if (!rbd_dev->parent_overlap)
		return 0;

	ret = ceph_extent_to_file(&rbd_dev->layout, obj_req->ex.oe_objno,
				  entire ? 0 : obj_req->ex.oe_off,
				  entire ? rbd_dev->layout.object_size :
							obj_req->ex.oe_len,
				  &obj_req->img_extents,
				  &obj_req->num_img_extents);
	if (ret)
		return ret;

	prune_extents(obj_req->img_extents, &obj_req->num_img_extents,
		      rbd_dev->parent_overlap);
	return 0;
}

static void rbd_osd_setup_data(struct ceph_osd_request *osd_req, int which)
{
	struct rbd_obj_request *obj_req = osd_req->r_priv;

	switch (obj_req->img_request->data_type) {
	case OBJ_REQUEST_BIO:
		osd_req_op_extent_osd_data_bio(osd_req, which,
					       &obj_req->bio_pos,
					       obj_req->ex.oe_len);
		break;
	case OBJ_REQUEST_BVECS:
	case OBJ_REQUEST_OWN_BVECS:
		rbd_assert(obj_req->bvec_pos.iter.bi_size ==
							obj_req->ex.oe_len);
		rbd_assert(obj_req->bvec_idx == obj_req->bvec_count);
		osd_req_op_extent_osd_data_bvec_pos(osd_req, which,
						    &obj_req->bvec_pos);
		break;
	default:
		BUG();
	}
}

static int rbd_osd_setup_stat(struct ceph_osd_request *osd_req, int which)
{
	struct page **pages;

	/*
	 * The response data for a STAT call consists of:
	 *     le64 length;
	 *     struct {
	 *         le32 tv_sec;
	 *         le32 tv_nsec;
	 *     } mtime;
	 */
	pages = ceph_alloc_page_vector(1, GFP_NOIO);
	if (IS_ERR(pages))
		return PTR_ERR(pages);

	osd_req_op_init(osd_req, which, CEPH_OSD_OP_STAT, 0);
	osd_req_op_raw_data_in_pages(osd_req, which, pages,
				     8 + sizeof(struct ceph_timespec),
				     0, false, true);
	return 0;
}

static int rbd_osd_setup_copyup(struct ceph_osd_request *osd_req, int which,
				u32 bytes)
{
	struct rbd_obj_request *obj_req = osd_req->r_priv;
	int ret;

	ret = osd_req_op_cls_init(osd_req, which, "rbd", "copyup");
	if (ret)
		return ret;

	osd_req_op_cls_request_data_bvecs(osd_req, which, obj_req->copyup_bvecs,
					  obj_req->copyup_bvec_count, bytes);
	return 0;
}

static int rbd_obj_init_read(struct rbd_obj_request *obj_req)
{
	obj_req->read_state = RBD_OBJ_READ_START;
	return 0;
}

static void __rbd_osd_setup_write_ops(struct ceph_osd_request *osd_req,
				      int which)
{
	struct rbd_obj_request *obj_req = osd_req->r_priv;
	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
	u16 opcode;

	if (!use_object_map(rbd_dev) ||
	    !(obj_req->flags & RBD_OBJ_FLAG_MAY_EXIST)) {
		osd_req_op_alloc_hint_init(osd_req, which++,
					   rbd_dev->layout.object_size,
					   rbd_dev->layout.object_size,
					   rbd_dev->opts->alloc_hint_flags);
	}

	if (rbd_obj_is_entire(obj_req))
		opcode = CEPH_OSD_OP_WRITEFULL;
	else
		opcode = CEPH_OSD_OP_WRITE;

	osd_req_op_extent_init(osd_req, which, opcode,
			       obj_req->ex.oe_off, obj_req->ex.oe_len, 0, 0);
	rbd_osd_setup_data(osd_req, which);
}

static int rbd_obj_init_write(struct rbd_obj_request *obj_req)
{
	int ret;

	/* reverse map the entire object onto the parent */
	ret = rbd_obj_calc_img_extents(obj_req, true);
	if (ret)
		return ret;

	if (rbd_obj_copyup_enabled(obj_req))
		obj_req->flags |= RBD_OBJ_FLAG_COPYUP_ENABLED;

	obj_req->write_state = RBD_OBJ_WRITE_START;
	return 0;
}

static u16 truncate_or_zero_opcode(struct rbd_obj_request *obj_req)
{
	return rbd_obj_is_tail(obj_req) ? CEPH_OSD_OP_TRUNCATE :
					  CEPH_OSD_OP_ZERO;
}

static void __rbd_osd_setup_discard_ops(struct ceph_osd_request *osd_req,
					int which)
{
	struct rbd_obj_request *obj_req = osd_req->r_priv;

	if (rbd_obj_is_entire(obj_req) && !obj_req->num_img_extents) {
		rbd_assert(obj_req->flags & RBD_OBJ_FLAG_DELETION);
		osd_req_op_init(osd_req, which, CEPH_OSD_OP_DELETE, 0);
	} else {
		osd_req_op_extent_init(osd_req, which,
				       truncate_or_zero_opcode(obj_req),
				       obj_req->ex.oe_off, obj_req->ex.oe_len,
				       0, 0);
	}
}

static int rbd_obj_init_discard(struct rbd_obj_request *obj_req)
{
	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
	u64 off, next_off;
	int ret;

	/*
	 * Align the range to alloc_size boundary and punt on discards
	 * that are too small to free up any space.
	 *
	 * alloc_size == object_size && is_tail() is a special case for
	 * filestore with filestore_punch_hole = false, needed to allow
	 * truncate (in addition to delete).
	 */
	if (rbd_dev->opts->alloc_size != rbd_dev->layout.object_size ||
	    !rbd_obj_is_tail(obj_req)) {
		off = round_up(obj_req->ex.oe_off, rbd_dev->opts->alloc_size);
		next_off = round_down(obj_req->ex.oe_off + obj_req->ex.oe_len,
				      rbd_dev->opts->alloc_size);
		if (off >= next_off)
			return 1;

		dout("%s %p %llu~%llu -> %llu~%llu\n", __func__,
		     obj_req, obj_req->ex.oe_off, obj_req->ex.oe_len,
		     off, next_off - off);
		obj_req->ex.oe_off = off;
		obj_req->ex.oe_len = next_off - off;
	}

	/* reverse map the entire object onto the parent */
	ret = rbd_obj_calc_img_extents(obj_req, true);
	if (ret)
		return ret;

	obj_req->flags |= RBD_OBJ_FLAG_NOOP_FOR_NONEXISTENT;
	if (rbd_obj_is_entire(obj_req) && !obj_req->num_img_extents)
		obj_req->flags |= RBD_OBJ_FLAG_DELETION;

	obj_req->write_state = RBD_OBJ_WRITE_START;
	return 0;
}

static void __rbd_osd_setup_zeroout_ops(struct ceph_osd_request *osd_req,
					int which)
{
	struct rbd_obj_request *obj_req = osd_req->r_priv;
	u16 opcode;

	if (rbd_obj_is_entire(obj_req)) {
		if (obj_req->num_img_extents) {
			if (!(obj_req->flags & RBD_OBJ_FLAG_COPYUP_ENABLED))
				osd_req_op_init(osd_req, which++,
						CEPH_OSD_OP_CREATE, 0);
			opcode = CEPH_OSD_OP_TRUNCATE;
		} else {
			rbd_assert(obj_req->flags & RBD_OBJ_FLAG_DELETION);
			osd_req_op_init(osd_req, which++,
					CEPH_OSD_OP_DELETE, 0);
			opcode = 0;
		}
	} else {
		opcode = truncate_or_zero_opcode(obj_req);
	}

	if (opcode)
		osd_req_op_extent_init(osd_req, which, opcode,
				       obj_req->ex.oe_off, obj_req->ex.oe_len,
				       0, 0);
}

static int rbd_obj_init_zeroout(struct rbd_obj_request *obj_req)
{
	int ret;

	/* reverse map the entire object onto the parent */
	ret = rbd_obj_calc_img_extents(obj_req, true);
	if (ret)
		return ret;

	if (rbd_obj_copyup_enabled(obj_req))
		obj_req->flags |= RBD_OBJ_FLAG_COPYUP_ENABLED;
	if (!obj_req->num_img_extents) {
		obj_req->flags |= RBD_OBJ_FLAG_NOOP_FOR_NONEXISTENT;
		if (rbd_obj_is_entire(obj_req))
			obj_req->flags |= RBD_OBJ_FLAG_DELETION;
	}

	obj_req->write_state = RBD_OBJ_WRITE_START;
	return 0;
}

static int count_write_ops(struct rbd_obj_request *obj_req)
{
	struct rbd_img_request *img_req = obj_req->img_request;

	switch (img_req->op_type) {
	case OBJ_OP_WRITE:
		if (!use_object_map(img_req->rbd_dev) ||
		    !(obj_req->flags & RBD_OBJ_FLAG_MAY_EXIST))
			return 2; /* setallochint + write/writefull */

		return 1; /* write/writefull */
	case OBJ_OP_DISCARD:
		return 1; /* delete/truncate/zero */
	case OBJ_OP_ZEROOUT:
		if (rbd_obj_is_entire(obj_req) && obj_req->num_img_extents &&
		    !(obj_req->flags & RBD_OBJ_FLAG_COPYUP_ENABLED))
			return 2; /* create + truncate */

		return 1; /* delete/truncate/zero */
	default:
		BUG();
	}
}

static void rbd_osd_setup_write_ops(struct ceph_osd_request *osd_req,
				    int which)
{
	struct rbd_obj_request *obj_req = osd_req->r_priv;

	switch (obj_req->img_request->op_type) {
	case OBJ_OP_WRITE:
		__rbd_osd_setup_write_ops(osd_req, which);
		break;
	case OBJ_OP_DISCARD:
		__rbd_osd_setup_discard_ops(osd_req, which);
		break;
	case OBJ_OP_ZEROOUT:
		__rbd_osd_setup_zeroout_ops(osd_req, which);
		break;
	default:
		BUG();
	}
}

/*
 * Prune the list of object requests (adjust offset and/or length, drop
 * redundant requests).  Prepare object request state machines and image
 * request state machine for execution.
 */
static int __rbd_img_fill_request(struct rbd_img_request *img_req)
{
	struct rbd_obj_request *obj_req, *next_obj_req;
	int ret;

	for_each_obj_request_safe(img_req, obj_req, next_obj_req) {
		switch (img_req->op_type) {
		case OBJ_OP_READ:
			ret = rbd_obj_init_read(obj_req);
			break;
		case OBJ_OP_WRITE:
			ret = rbd_obj_init_write(obj_req);
			break;
		case OBJ_OP_DISCARD:
			ret = rbd_obj_init_discard(obj_req);
			break;
		case OBJ_OP_ZEROOUT:
			ret = rbd_obj_init_zeroout(obj_req);
			break;
		default:
			BUG();
		}
		if (ret < 0)
			return ret;
		if (ret > 0) {
			rbd_img_obj_request_del(img_req, obj_req);
			continue;
		}
	}

	img_req->state = RBD_IMG_START;
	return 0;
}

union rbd_img_fill_iter {
	struct ceph_bio_iter	bio_iter;
	struct ceph_bvec_iter	bvec_iter;
};

struct rbd_img_fill_ctx {
	enum obj_request_type	pos_type;
	union rbd_img_fill_iter	*pos;
	union rbd_img_fill_iter	iter;
	ceph_object_extent_fn_t	set_pos_fn;
	ceph_object_extent_fn_t	count_fn;
	ceph_object_extent_fn_t	copy_fn;
};

static struct ceph_object_extent *alloc_object_extent(void *arg)
{
	struct rbd_img_request *img_req = arg;
	struct rbd_obj_request *obj_req;

	obj_req = rbd_obj_request_create();
	if (!obj_req)
		return NULL;

	rbd_img_obj_request_add(img_req, obj_req);
	return &obj_req->ex;
}

/*
 * While su != os && sc == 1 is technically not fancy (it's the same
 * layout as su == os && sc == 1), we can't use the nocopy path for it
 * because ->set_pos_fn() should be called only once per object.
 * ceph_file_to_extents() invokes action_fn once per stripe unit, so
 * treat su != os && sc == 1 as fancy.
 */
static bool rbd_layout_is_fancy(struct ceph_file_layout *l)
{
	return l->stripe_unit != l->object_size;
}

static int rbd_img_fill_request_nocopy(struct rbd_img_request *img_req,
				       struct ceph_file_extent *img_extents,
				       u32 num_img_extents,
				       struct rbd_img_fill_ctx *fctx)
{
	u32 i;
	int ret;

	img_req->data_type = fctx->pos_type;

	/*
	 * Create object requests and set each object request's starting
	 * position in the provided bio (list) or bio_vec array.
	 */
	fctx->iter = *fctx->pos;
	for (i = 0; i < num_img_extents; i++) {
		ret = ceph_file_to_extents(&img_req->rbd_dev->layout,
					   img_extents[i].fe_off,
					   img_extents[i].fe_len,
					   &img_req->object_extents,
					   alloc_object_extent, img_req,
					   fctx->set_pos_fn, &fctx->iter);
		if (ret)
			return ret;
	}

	return __rbd_img_fill_request(img_req);
}

/*
 * Map a list of image extents to a list of object extents, create the
 * corresponding object requests (normally each to a different object,
 * but not always) and add them to @img_req.  For each object request,
 * set up its data descriptor to point to the corresponding chunk(s) of
 * @fctx->pos data buffer.
 *
 * Because ceph_file_to_extents() will merge adjacent object extents
 * together, each object request's data descriptor may point to multiple
 * different chunks of @fctx->pos data buffer.
 *
 * @fctx->pos data buffer is assumed to be large enough.
 */
static int rbd_img_fill_request(struct rbd_img_request *img_req,
				struct ceph_file_extent *img_extents,
				u32 num_img_extents,
				struct rbd_img_fill_ctx *fctx)
{
	struct rbd_device *rbd_dev = img_req->rbd_dev;
	struct rbd_obj_request *obj_req;
	u32 i;
	int ret;

	if (fctx->pos_type == OBJ_REQUEST_NODATA ||
	    !rbd_layout_is_fancy(&rbd_dev->layout))
		return rbd_img_fill_request_nocopy(img_req, img_extents,
						   num_img_extents, fctx);

	img_req->data_type = OBJ_REQUEST_OWN_BVECS;

	/*
	 * Create object requests and determine ->bvec_count for each object
	 * request.  Note that ->bvec_count sum over all object requests may
	 * be greater than the number of bio_vecs in the provided bio (list)
	 * or bio_vec array because when mapped, those bio_vecs can straddle
	 * stripe unit boundaries.
	 */
	fctx->iter = *fctx->pos;
	for (i = 0; i < num_img_extents; i++) {
		ret = ceph_file_to_extents(&rbd_dev->layout,
					   img_extents[i].fe_off,
					   img_extents[i].fe_len,
					   &img_req->object_extents,
					   alloc_object_extent, img_req,
					   fctx->count_fn, &fctx->iter);
		if (ret)
			return ret;
	}

	for_each_obj_request(img_req, obj_req) {
		obj_req->bvec_pos.bvecs = kmalloc_array(obj_req->bvec_count,
					      sizeof(*obj_req->bvec_pos.bvecs),
					      GFP_NOIO);
		if (!obj_req->bvec_pos.bvecs)
			return -ENOMEM;
	}

	/*
	 * Fill in each object request's private bio_vec array, splitting and
	 * rearranging the provided bio_vecs in stripe unit chunks as needed.
	 */
	fctx->iter = *fctx->pos;
	for (i = 0; i < num_img_extents; i++) {
		ret = ceph_iterate_extents(&rbd_dev->layout,
					   img_extents[i].fe_off,
					   img_extents[i].fe_len,
					   &img_req->object_extents,
					   fctx->copy_fn, &fctx->iter);
		if (ret)
			return ret;
	}

	return __rbd_img_fill_request(img_req);
}

static int rbd_img_fill_nodata(struct rbd_img_request *img_req,
			       u64 off, u64 len)
{
	struct ceph_file_extent ex = { off, len };
	union rbd_img_fill_iter dummy = {};
	struct rbd_img_fill_ctx fctx = {
		.pos_type = OBJ_REQUEST_NODATA,
		.pos = &dummy,
	};

	return rbd_img_fill_request(img_req, &ex, 1, &fctx);
}

static void set_bio_pos(struct ceph_object_extent *ex, u32 bytes, void *arg)
{
	struct rbd_obj_request *obj_req =
	    container_of(ex, struct rbd_obj_request, ex);
	struct ceph_bio_iter *it = arg;

	dout("%s objno %llu bytes %u\n", __func__, ex->oe_objno, bytes);
	obj_req->bio_pos = *it;
	ceph_bio_iter_advance(it, bytes);
}

static void count_bio_bvecs(struct ceph_object_extent *ex, u32 bytes, void *arg)
{
	struct rbd_obj_request *obj_req =
	    container_of(ex, struct rbd_obj_request, ex);
	struct ceph_bio_iter *it = arg;

	dout("%s objno %llu bytes %u\n", __func__, ex->oe_objno, bytes);
	ceph_bio_iter_advance_step(it, bytes, ({
		obj_req->bvec_count++;
	}));

}

static void copy_bio_bvecs(struct ceph_object_extent *ex, u32 bytes, void *arg)
{
	struct rbd_obj_request *obj_req =
	    container_of(ex, struct rbd_obj_request, ex);
	struct ceph_bio_iter *it = arg;

	dout("%s objno %llu bytes %u\n", __func__, ex->oe_objno, bytes);
	ceph_bio_iter_advance_step(it, bytes, ({
		obj_req->bvec_pos.bvecs[obj_req->bvec_idx++] = bv;
		obj_req->bvec_pos.iter.bi_size += bv.bv_len;
	}));
}

static int __rbd_img_fill_from_bio(struct rbd_img_request *img_req,
				   struct ceph_file_extent *img_extents,
				   u32 num_img_extents,
				   struct ceph_bio_iter *bio_pos)
{
	struct rbd_img_fill_ctx fctx = {
		.pos_type = OBJ_REQUEST_BIO,
		.pos = (union rbd_img_fill_iter *)bio_pos,
		.set_pos_fn = set_bio_pos,
		.count_fn = count_bio_bvecs,
		.copy_fn = copy_bio_bvecs,
	};

	return rbd_img_fill_request(img_req, img_extents, num_img_extents,
				    &fctx);
}

static int rbd_img_fill_from_bio(struct rbd_img_request *img_req,
				 u64 off, u64 len, struct bio *bio)
{
	struct ceph_file_extent ex = { off, len };
	struct ceph_bio_iter it = { .bio = bio, .iter = bio->bi_iter };

	return __rbd_img_fill_from_bio(img_req, &ex, 1, &it);
}

static void set_bvec_pos(struct ceph_object_extent *ex, u32 bytes, void *arg)
{
	struct rbd_obj_request *obj_req =
	    container_of(ex, struct rbd_obj_request, ex);
	struct ceph_bvec_iter *it = arg;

	obj_req->bvec_pos = *it;
	ceph_bvec_iter_shorten(&obj_req->bvec_pos, bytes);
	ceph_bvec_iter_advance(it, bytes);
}

static void count_bvecs(struct ceph_object_extent *ex, u32 bytes, void *arg)
{
	struct rbd_obj_request *obj_req =
	    container_of(ex, struct rbd_obj_request, ex);
	struct ceph_bvec_iter *it = arg;

	ceph_bvec_iter_advance_step(it, bytes, ({
		obj_req->bvec_count++;
	}));
}

static void copy_bvecs(struct ceph_object_extent *ex, u32 bytes, void *arg)
{
	struct rbd_obj_request *obj_req =
	    container_of(ex, struct rbd_obj_request, ex);
	struct ceph_bvec_iter *it = arg;

	ceph_bvec_iter_advance_step(it, bytes, ({
		obj_req->bvec_pos.bvecs[obj_req->bvec_idx++] = bv;
		obj_req->bvec_pos.iter.bi_size += bv.bv_len;
	}));
}

static int __rbd_img_fill_from_bvecs(struct rbd_img_request *img_req,
				     struct ceph_file_extent *img_extents,
				     u32 num_img_extents,
				     struct ceph_bvec_iter *bvec_pos)
{
	struct rbd_img_fill_ctx fctx = {
		.pos_type = OBJ_REQUEST_BVECS,
		.pos = (union rbd_img_fill_iter *)bvec_pos,
		.set_pos_fn = set_bvec_pos,
		.count_fn = count_bvecs,
		.copy_fn = copy_bvecs,
	};

	return rbd_img_fill_request(img_req, img_extents, num_img_extents,
				    &fctx);
}

static int rbd_img_fill_from_bvecs(struct rbd_img_request *img_req,
				   struct ceph_file_extent *img_extents,
				   u32 num_img_extents,
				   struct bio_vec *bvecs)
{
	struct ceph_bvec_iter it = {
		.bvecs = bvecs,
		.iter = { .bi_size = ceph_file_extents_bytes(img_extents,
							     num_img_extents) },
	};

	return __rbd_img_fill_from_bvecs(img_req, img_extents, num_img_extents,
					 &it);
}

static void rbd_img_handle_request_work(struct work_struct *work)
{
	struct rbd_img_request *img_req =
	    container_of(work, struct rbd_img_request, work);

	rbd_img_handle_request(img_req, img_req->work_result);
}

static void rbd_img_schedule(struct rbd_img_request *img_req, int result)
{
	INIT_WORK(&img_req->work, rbd_img_handle_request_work);
	img_req->work_result = result;
	queue_work(rbd_wq, &img_req->work);
}

static bool rbd_obj_may_exist(struct rbd_obj_request *obj_req)
{
	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;

	if (rbd_object_map_may_exist(rbd_dev, obj_req->ex.oe_objno)) {
		obj_req->flags |= RBD_OBJ_FLAG_MAY_EXIST;
		return true;
	}

	dout("%s %p objno %llu assuming dne\n", __func__, obj_req,
	     obj_req->ex.oe_objno);
	return false;
}

static int rbd_obj_read_object(struct rbd_obj_request *obj_req)
{
	struct ceph_osd_request *osd_req;
	int ret;

	osd_req = __rbd_obj_add_osd_request(obj_req, NULL, 1);
	if (IS_ERR(osd_req))
		return PTR_ERR(osd_req);

	osd_req_op_extent_init(osd_req, 0, CEPH_OSD_OP_READ,
			       obj_req->ex.oe_off, obj_req->ex.oe_len, 0, 0);
	rbd_osd_setup_data(osd_req, 0);
	rbd_osd_format_read(osd_req);

	ret = ceph_osdc_alloc_messages(osd_req, GFP_NOIO);
	if (ret)
		return ret;

	rbd_osd_submit(osd_req);
	return 0;
}

static int rbd_obj_read_from_parent(struct rbd_obj_request *obj_req)
{
	struct rbd_img_request *img_req = obj_req->img_request;
	struct rbd_device *parent = img_req->rbd_dev->parent;
	struct rbd_img_request *child_img_req;
	int ret;

	child_img_req = kmem_cache_alloc(rbd_img_request_cache, GFP_NOIO);
	if (!child_img_req)
		return -ENOMEM;

	rbd_img_request_init(child_img_req, parent, OBJ_OP_READ);
	__set_bit(IMG_REQ_CHILD, &child_img_req->flags);
	child_img_req->obj_request = obj_req;

	down_read(&parent->header_rwsem);
	rbd_img_capture_header(child_img_req);
	up_read(&parent->header_rwsem);

	dout("%s child_img_req %p for obj_req %p\n", __func__, child_img_req,
	     obj_req);

	if (!rbd_img_is_write(img_req)) {
		switch (img_req->data_type) {
		case OBJ_REQUEST_BIO:
			ret = __rbd_img_fill_from_bio(child_img_req,
						      obj_req->img_extents,
						      obj_req->num_img_extents,
						      &obj_req->bio_pos);
			break;
		case OBJ_REQUEST_BVECS:
		case OBJ_REQUEST_OWN_BVECS:
			ret = __rbd_img_fill_from_bvecs(child_img_req,
						      obj_req->img_extents,
						      obj_req->num_img_extents,
						      &obj_req->bvec_pos);
			break;
		default:
			BUG();
		}
	} else {
		ret = rbd_img_fill_from_bvecs(child_img_req,
					      obj_req->img_extents,
					      obj_req->num_img_extents,
					      obj_req->copyup_bvecs);
	}
	if (ret) {
		rbd_img_request_destroy(child_img_req);
		return ret;
	}

	/* avoid parent chain recursion */
	rbd_img_schedule(child_img_req, 0);
	return 0;
}

static bool rbd_obj_advance_read(struct rbd_obj_request *obj_req, int *result)
{
	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
	int ret;

again:
	switch (obj_req->read_state) {
	case RBD_OBJ_READ_START:
		rbd_assert(!*result);

		if (!rbd_obj_may_exist(obj_req)) {
			*result = -ENOENT;
			obj_req->read_state = RBD_OBJ_READ_OBJECT;
			goto again;
		}

		ret = rbd_obj_read_object(obj_req);
		if (ret) {
			*result = ret;
			return true;
		}
		obj_req->read_state = RBD_OBJ_READ_OBJECT;
		return false;
	case RBD_OBJ_READ_OBJECT:
		if (*result == -ENOENT && rbd_dev->parent_overlap) {
			/* reverse map this object extent onto the parent */
			ret = rbd_obj_calc_img_extents(obj_req, false);
			if (ret) {
				*result = ret;
				return true;
			}
			if (obj_req->num_img_extents) {
				ret = rbd_obj_read_from_parent(obj_req);
				if (ret) {
					*result = ret;
					return true;
				}
				obj_req->read_state = RBD_OBJ_READ_PARENT;
				return false;
			}
		}

		/*
		 * -ENOENT means a hole in the image -- zero-fill the entire
		 * length of the request.  A short read also implies zero-fill
		 * to the end of the request.
		 */
		if (*result == -ENOENT) {
			rbd_obj_zero_range(obj_req, 0, obj_req->ex.oe_len);
			*result = 0;
		} else if (*result >= 0) {
			if (*result < obj_req->ex.oe_len)
				rbd_obj_zero_range(obj_req, *result,
						obj_req->ex.oe_len - *result);
			else
				rbd_assert(*result == obj_req->ex.oe_len);
			*result = 0;
		}
		return true;
	case RBD_OBJ_READ_PARENT:
		/*
		 * The parent image is read only up to the overlap -- zero-fill
		 * from the overlap to the end of the request.
		 */
		if (!*result) {
			u32 obj_overlap = rbd_obj_img_extents_bytes(obj_req);

			if (obj_overlap < obj_req->ex.oe_len)
				rbd_obj_zero_range(obj_req, obj_overlap,
					    obj_req->ex.oe_len - obj_overlap);
		}
		return true;
	default:
		BUG();
	}
}

static bool rbd_obj_write_is_noop(struct rbd_obj_request *obj_req)
{
	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;

	if (rbd_object_map_may_exist(rbd_dev, obj_req->ex.oe_objno))
		obj_req->flags |= RBD_OBJ_FLAG_MAY_EXIST;

	if (!(obj_req->flags & RBD_OBJ_FLAG_MAY_EXIST) &&
	    (obj_req->flags & RBD_OBJ_FLAG_NOOP_FOR_NONEXISTENT)) {
		dout("%s %p noop for nonexistent\n", __func__, obj_req);
		return true;
	}

	return false;
}

/*
 * Return:
 *   0 - object map update sent
 *   1 - object map update isn't needed
 *  <0 - error
 */
static int rbd_obj_write_pre_object_map(struct rbd_obj_request *obj_req)
{
	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
	u8 new_state;

	if (!(rbd_dev->header.features & RBD_FEATURE_OBJECT_MAP))
		return 1;

	if (obj_req->flags & RBD_OBJ_FLAG_DELETION)
		new_state = OBJECT_PENDING;
	else
		new_state = OBJECT_EXISTS;

	return rbd_object_map_update(obj_req, CEPH_NOSNAP, new_state, NULL);
}

static int rbd_obj_write_object(struct rbd_obj_request *obj_req)
{
	struct ceph_osd_request *osd_req;
	int num_ops = count_write_ops(obj_req);
	int which = 0;
	int ret;

	if (obj_req->flags & RBD_OBJ_FLAG_COPYUP_ENABLED)
		num_ops++; /* stat */

	osd_req = rbd_obj_add_osd_request(obj_req, num_ops);
	if (IS_ERR(osd_req))
		return PTR_ERR(osd_req);

	if (obj_req->flags & RBD_OBJ_FLAG_COPYUP_ENABLED) {
		ret = rbd_osd_setup_stat(osd_req, which++);
		if (ret)
			return ret;
	}

	rbd_osd_setup_write_ops(osd_req, which);
	rbd_osd_format_write(osd_req);

	ret = ceph_osdc_alloc_messages(osd_req, GFP_NOIO);
	if (ret)
		return ret;

	rbd_osd_submit(osd_req);
	return 0;
}

/*
 * copyup_bvecs pages are never highmem pages
 */
static bool is_zero_bvecs(struct bio_vec *bvecs, u32 bytes)
{
	struct ceph_bvec_iter it = {
		.bvecs = bvecs,
		.iter = { .bi_size = bytes },
	};

	ceph_bvec_iter_advance_step(&it, bytes, ({
		if (memchr_inv(page_address(bv.bv_page) + bv.bv_offset, 0,
			       bv.bv_len))
			return false;
	}));
	return true;
}

#define MODS_ONLY	U32_MAX

static int rbd_obj_copyup_empty_snapc(struct rbd_obj_request *obj_req,
				      u32 bytes)
{
	struct ceph_osd_request *osd_req;
	int ret;

	dout("%s obj_req %p bytes %u\n", __func__, obj_req, bytes);
	rbd_assert(bytes > 0 && bytes != MODS_ONLY);

	osd_req = __rbd_obj_add_osd_request(obj_req, &rbd_empty_snapc, 1);
	if (IS_ERR(osd_req))
		return PTR_ERR(osd_req);

	ret = rbd_osd_setup_copyup(osd_req, 0, bytes);
	if (ret)
		return ret;

	rbd_osd_format_write(osd_req);

	ret = ceph_osdc_alloc_messages(osd_req, GFP_NOIO);
	if (ret)
		return ret;

	rbd_osd_submit(osd_req);
	return 0;
}

static int rbd_obj_copyup_current_snapc(struct rbd_obj_request *obj_req,
					u32 bytes)
{
	struct ceph_osd_request *osd_req;
	int num_ops = count_write_ops(obj_req);
	int which = 0;
	int ret;

	dout("%s obj_req %p bytes %u\n", __func__, obj_req, bytes);

	if (bytes != MODS_ONLY)
		num_ops++; /* copyup */

	osd_req = rbd_obj_add_osd_request(obj_req, num_ops);
	if (IS_ERR(osd_req))
		return PTR_ERR(osd_req);

	if (bytes != MODS_ONLY) {
		ret = rbd_osd_setup_copyup(osd_req, which++, bytes);
		if (ret)
			return ret;
	}

	rbd_osd_setup_write_ops(osd_req, which);
	rbd_osd_format_write(osd_req);

	ret = ceph_osdc_alloc_messages(osd_req, GFP_NOIO);
	if (ret)
		return ret;

	rbd_osd_submit(osd_req);
	return 0;
}

static int setup_copyup_bvecs(struct rbd_obj_request *obj_req, u64 obj_overlap)
{
	u32 i;

	rbd_assert(!obj_req->copyup_bvecs);
	obj_req->copyup_bvec_count = calc_pages_for(0, obj_overlap);
	obj_req->copyup_bvecs = kcalloc(obj_req->copyup_bvec_count,
					sizeof(*obj_req->copyup_bvecs),
					GFP_NOIO);
	if (!obj_req->copyup_bvecs)
		return -ENOMEM;

	for (i = 0; i < obj_req->copyup_bvec_count; i++) {
		unsigned int len = min(obj_overlap, (u64)PAGE_SIZE);

		obj_req->copyup_bvecs[i].bv_page = alloc_page(GFP_NOIO);
		if (!obj_req->copyup_bvecs[i].bv_page)
			return -ENOMEM;

		obj_req->copyup_bvecs[i].bv_offset = 0;
		obj_req->copyup_bvecs[i].bv_len = len;
		obj_overlap -= len;
	}

	rbd_assert(!obj_overlap);
	return 0;
}

/*
 * The target object doesn't exist.  Read the data for the entire
 * target object up to the overlap point (if any) from the parent,
 * so we can use it for a copyup.
 */
static int rbd_obj_copyup_read_parent(struct rbd_obj_request *obj_req)
{
	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
	int ret;

	rbd_assert(obj_req->num_img_extents);
	prune_extents(obj_req->img_extents, &obj_req->num_img_extents,
		      rbd_dev->parent_overlap);
	if (!obj_req->num_img_extents) {
		/*
		 * The overlap has become 0 (most likely because the
		 * image has been flattened).  Re-submit the original write
		 * request -- pass MODS_ONLY since the copyup isn't needed
		 * anymore.
		 */
		return rbd_obj_copyup_current_snapc(obj_req, MODS_ONLY);
	}

	ret = setup_copyup_bvecs(obj_req, rbd_obj_img_extents_bytes(obj_req));
	if (ret)
		return ret;

	return rbd_obj_read_from_parent(obj_req);
}

static void rbd_obj_copyup_object_maps(struct rbd_obj_request *obj_req)
{
	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
	struct ceph_snap_context *snapc = obj_req->img_request->snapc;
	u8 new_state;
	u32 i;
	int ret;

	rbd_assert(!obj_req->pending.result && !obj_req->pending.num_pending);

	if (!(rbd_dev->header.features & RBD_FEATURE_OBJECT_MAP))
		return;

	if (obj_req->flags & RBD_OBJ_FLAG_COPYUP_ZEROS)
		return;

	for (i = 0; i < snapc->num_snaps; i++) {
		if ((rbd_dev->header.features & RBD_FEATURE_FAST_DIFF) &&
		    i + 1 < snapc->num_snaps)
			new_state = OBJECT_EXISTS_CLEAN;
		else
			new_state = OBJECT_EXISTS;

		ret = rbd_object_map_update(obj_req, snapc->snaps[i],
					    new_state, NULL);
		if (ret < 0) {
			obj_req->pending.result = ret;
			return;
		}

		rbd_assert(!ret);
		obj_req->pending.num_pending++;
	}
}

static void rbd_obj_copyup_write_object(struct rbd_obj_request *obj_req)
{
	u32 bytes = rbd_obj_img_extents_bytes(obj_req);
	int ret;

	rbd_assert(!obj_req->pending.result && !obj_req->pending.num_pending);

	/*
	 * Only send non-zero copyup data to save some I/O and network
	 * bandwidth -- zero copyup data is equivalent to the object not
	 * existing.
	 */
	if (obj_req->flags & RBD_OBJ_FLAG_COPYUP_ZEROS)
		bytes = 0;

	if (obj_req->img_request->snapc->num_snaps && bytes > 0) {
		/*
		 * Send a copyup request with an empty snapshot context to
		 * deep-copyup the object through all existing snapshots.
		 * A second request with the current snapshot context will be
		 * sent for the actual modification.
		 */
		ret = rbd_obj_copyup_empty_snapc(obj_req, bytes);
		if (ret) {
			obj_req->pending.result = ret;
			return;
		}

		obj_req->pending.num_pending++;
		bytes = MODS_ONLY;
	}

	ret = rbd_obj_copyup_current_snapc(obj_req, bytes);
	if (ret) {
		obj_req->pending.result = ret;
		return;
	}

	obj_req->pending.num_pending++;
}

static bool rbd_obj_advance_copyup(struct rbd_obj_request *obj_req, int *result)
{
	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
	int ret;

again:
	switch (obj_req->copyup_state) {
	case RBD_OBJ_COPYUP_START:
		rbd_assert(!*result);

		ret = rbd_obj_copyup_read_parent(obj_req);
		if (ret) {
			*result = ret;
			return true;
		}
		if (obj_req->num_img_extents)
			obj_req->copyup_state = RBD_OBJ_COPYUP_READ_PARENT;
		else
			obj_req->copyup_state = RBD_OBJ_COPYUP_WRITE_OBJECT;
		return false;
	case RBD_OBJ_COPYUP_READ_PARENT:
		if (*result)
			return true;

		if (is_zero_bvecs(obj_req->copyup_bvecs,
				  rbd_obj_img_extents_bytes(obj_req))) {
			dout("%s %p detected zeros\n", __func__, obj_req);
			obj_req->flags |= RBD_OBJ_FLAG_COPYUP_ZEROS;
		}

		rbd_obj_copyup_object_maps(obj_req);
		if (!obj_req->pending.num_pending) {
			*result = obj_req->pending.result;
			obj_req->copyup_state = RBD_OBJ_COPYUP_OBJECT_MAPS;
			goto again;
		}
		obj_req->copyup_state = __RBD_OBJ_COPYUP_OBJECT_MAPS;
		return false;
	case __RBD_OBJ_COPYUP_OBJECT_MAPS:
		if (!pending_result_dec(&obj_req->pending, result))
			return false;
		fallthrough;
	case RBD_OBJ_COPYUP_OBJECT_MAPS:
		if (*result) {
			rbd_warn(rbd_dev, "snap object map update failed: %d",
				 *result);
			return true;
		}

		rbd_obj_copyup_write_object(obj_req);
		if (!obj_req->pending.num_pending) {
			*result = obj_req->pending.result;
			obj_req->copyup_state = RBD_OBJ_COPYUP_WRITE_OBJECT;
			goto again;
		}
		obj_req->copyup_state = __RBD_OBJ_COPYUP_WRITE_OBJECT;
		return false;
	case __RBD_OBJ_COPYUP_WRITE_OBJECT:
		if (!pending_result_dec(&obj_req->pending, result))
			return false;
		fallthrough;
	case RBD_OBJ_COPYUP_WRITE_OBJECT:
		return true;
	default:
		BUG();
	}
}

/*
 * Return:
 *   0 - object map update sent
 *   1 - object map update isn't needed
 *  <0 - error
 */
static int rbd_obj_write_post_object_map(struct rbd_obj_request *obj_req)
{
	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
	u8 current_state = OBJECT_PENDING;

	if (!(rbd_dev->header.features & RBD_FEATURE_OBJECT_MAP))
		return 1;

	if (!(obj_req->flags & RBD_OBJ_FLAG_DELETION))
		return 1;

	return rbd_object_map_update(obj_req, CEPH_NOSNAP, OBJECT_NONEXISTENT,
				     &current_state);
}

static bool rbd_obj_advance_write(struct rbd_obj_request *obj_req, int *result)
{
	struct rbd_device *rbd_dev = obj_req->img_request->rbd_dev;
	int ret;

again:
	switch (obj_req->write_state) {
	case RBD_OBJ_WRITE_START:
		rbd_assert(!*result);

		if (rbd_obj_write_is_noop(obj_req))
			return true;

		ret = rbd_obj_write_pre_object_map(obj_req);
		if (ret < 0) {
			*result = ret;
			return true;
		}
		obj_req->write_state = RBD_OBJ_WRITE_PRE_OBJECT_MAP;
		if (ret > 0)
			goto again;
		return false;
	case RBD_OBJ_WRITE_PRE_OBJECT_MAP:
		if (*result) {
			rbd_warn(rbd_dev, "pre object map update failed: %d",
				 *result);
			return true;
		}
		ret = rbd_obj_write_object(obj_req);
		if (ret) {
			*result = ret;
			return true;
		}
		obj_req->write_state = RBD_OBJ_WRITE_OBJECT;
		return false;
	case RBD_OBJ_WRITE_OBJECT:
		if (*result == -ENOENT) {
			if (obj_req->flags & RBD_OBJ_FLAG_COPYUP_ENABLED) {
				*result = 0;
				obj_req->copyup_state = RBD_OBJ_COPYUP_START;
				obj_req->write_state = __RBD_OBJ_WRITE_COPYUP;
				goto again;
			}
			/*
			 * On a non-existent object:
			 *   delete - -ENOENT, truncate/zero - 0
			 */
			if (obj_req->flags & RBD_OBJ_FLAG_DELETION)
				*result = 0;
		}
		if (*result)
			return true;

		obj_req->write_state = RBD_OBJ_WRITE_COPYUP;
		goto again;
	case __RBD_OBJ_WRITE_COPYUP:
		if (!rbd_obj_advance_copyup(obj_req, result))
			return false;
		fallthrough;
	case RBD_OBJ_WRITE_COPYUP:
		if (*result) {
			rbd_warn(rbd_dev, "copyup failed: %d", *result);
			return true;
		}
		ret = rbd_obj_write_post_object_map(obj_req);
		if (ret < 0) {
			*result = ret;
			return true;
		}
		obj_req->write_state = RBD_OBJ_WRITE_POST_OBJECT_MAP;
		if (ret > 0)
			goto again;
		return false;
	case RBD_OBJ_WRITE_POST_OBJECT_MAP:
		if (*result)
			rbd_warn(rbd_dev, "post object map update failed: %d",
				 *result);
		return true;
	default:
		BUG();
	}
}

/*
 * Return true if @obj_req is completed.
 */
static bool __rbd_obj_handle_request(struct rbd_obj_request *obj_req,
				     int *result)
{
	struct rbd_img_request *img_req = obj_req->img_request;
	struct rbd_device *rbd_dev = img_req->rbd_dev;
	bool done;

	mutex_lock(&obj_req->state_mutex);
	if (!rbd_img_is_write(img_req))
		done = rbd_obj_advance_read(obj_req, result);
	else
		done = rbd_obj_advance_write(obj_req, result);
	mutex_unlock(&obj_req->state_mutex);

	if (done && *result) {
		rbd_assert(*result < 0);
		rbd_warn(rbd_dev, "%s at objno %llu %llu~%llu result %d",
			 obj_op_name(img_req->op_type), obj_req->ex.oe_objno,
			 obj_req->ex.oe_off, obj_req->ex.oe_len, *result);
	}
	return done;
}

/*
 * This is open-coded in rbd_img_handle_request() to avoid parent chain
 * recursion.
 */
static void rbd_obj_handle_request(struct rbd_obj_request *obj_req, int result)
{
	if (__rbd_obj_handle_request(obj_req, &result))
		rbd_img_handle_request(obj_req->img_request, result);
}

static bool need_exclusive_lock(struct rbd_img_request *img_req)
{
	struct rbd_device *rbd_dev = img_req->rbd_dev;

	if (!(rbd_dev->header.features & RBD_FEATURE_EXCLUSIVE_LOCK))
		return false;

	if (rbd_is_ro(rbd_dev))
		return false;

	rbd_assert(!test_bit(IMG_REQ_CHILD, &img_req->flags));
	if (rbd_dev->opts->lock_on_read ||
	    (rbd_dev->header.features & RBD_FEATURE_OBJECT_MAP))
		return true;

	return rbd_img_is_write(img_req);
}

static bool rbd_lock_add_request(struct rbd_img_request *img_req)
{
	struct rbd_device *rbd_dev = img_req->rbd_dev;
	bool locked;

	lockdep_assert_held(&rbd_dev->lock_rwsem);
	locked = rbd_dev->lock_state == RBD_LOCK_STATE_LOCKED;
	spin_lock(&rbd_dev->lock_lists_lock);
	rbd_assert(list_empty(&img_req->lock_item));
	if (!locked)
		list_add_tail(&img_req->lock_item, &rbd_dev->acquiring_list);
	else
		list_add_tail(&img_req->lock_item, &rbd_dev->running_list);
	spin_unlock(&rbd_dev->lock_lists_lock);
	return locked;
}

static void rbd_lock_del_request(struct rbd_img_request *img_req)
{
	struct rbd_device *rbd_dev = img_req->rbd_dev;
	bool need_wakeup;

	lockdep_assert_held(&rbd_dev->lock_rwsem);
	spin_lock(&rbd_dev->lock_lists_lock);
	rbd_assert(!list_empty(&img_req->lock_item));
	list_del_init(&img_req->lock_item);
	need_wakeup = (rbd_dev->lock_state == RBD_LOCK_STATE_RELEASING &&
		       list_empty(&rbd_dev->running_list));
	spin_unlock(&rbd_dev->lock_lists_lock);
	if (need_wakeup)
		complete(&rbd_dev->releasing_wait);
}

static int rbd_img_exclusive_lock(struct rbd_img_request *img_req)
{
	struct rbd_device *rbd_dev = img_req->rbd_dev;

	if (!need_exclusive_lock(img_req))
		return 1;

	if (rbd_lock_add_request(img_req))
		return 1;

	if (rbd_dev->opts->exclusive) {
		WARN_ON(1); /* lock got released? */
		return -EROFS;
	}

	/*
	 * Note the use of mod_delayed_work() in rbd_acquire_lock()
	 * and cancel_delayed_work() in wake_lock_waiters().
	 */
	dout("%s rbd_dev %p queueing lock_dwork\n", __func__, rbd_dev);
	queue_delayed_work(rbd_dev->task_wq, &rbd_dev->lock_dwork, 0);
	return 0;
}

static void rbd_img_object_requests(struct rbd_img_request *img_req)
{
	struct rbd_obj_request *obj_req;

	rbd_assert(!img_req->pending.result && !img_req->pending.num_pending);

	for_each_obj_request(img_req, obj_req) {
		int result = 0;

		if (__rbd_obj_handle_request(obj_req, &result)) {
			if (result) {
				img_req->pending.result = result;
				return;
			}
		} else {
			img_req->pending.num_pending++;
		}
	}
}

static bool rbd_img_advance(struct rbd_img_request *img_req, int *result)
{
	struct rbd_device *rbd_dev = img_req->rbd_dev;
	int ret;

again:
	switch (img_req->state) {
	case RBD_IMG_START:
		rbd_assert(!*result);

		ret = rbd_img_exclusive_lock(img_req);
		if (ret < 0) {
			*result = ret;
			return true;
		}
		img_req->state = RBD_IMG_EXCLUSIVE_LOCK;
		if (ret > 0)
			goto again;
		return false;
	case RBD_IMG_EXCLUSIVE_LOCK:
		if (*result)
			return true;

		rbd_assert(!need_exclusive_lock(img_req) ||
			   __rbd_is_lock_owner(rbd_dev));

		rbd_img_object_requests(img_req);
		if (!img_req->pending.num_pending) {
			*result = img_req->pending.result;
			img_req->state = RBD_IMG_OBJECT_REQUESTS;
			goto again;
		}
		img_req->state = __RBD_IMG_OBJECT_REQUESTS;
		return false;
	case __RBD_IMG_OBJECT_REQUESTS:
		if (!pending_result_dec(&img_req->pending, result))
			return false;
		fallthrough;
	case RBD_IMG_OBJECT_REQUESTS:
		return true;
	default:
		BUG();
	}
}

/*
 * Return true if @img_req is completed.
 */
static bool __rbd_img_handle_request(struct rbd_img_request *img_req,
				     int *result)
{
	struct rbd_device *rbd_dev = img_req->rbd_dev;
	bool done;

	if (need_exclusive_lock(img_req)) {
		down_read(&rbd_dev->lock_rwsem);
		mutex_lock(&img_req->state_mutex);
		done = rbd_img_advance(img_req, result);
		if (done)
			rbd_lock_del_request(img_req);
		mutex_unlock(&img_req->state_mutex);
		up_read(&rbd_dev->lock_rwsem);
	} else {
		mutex_lock(&img_req->state_mutex);
		done = rbd_img_advance(img_req, result);
		mutex_unlock(&img_req->state_mutex);
	}

	if (done && *result) {
		rbd_assert(*result < 0);
		rbd_warn(rbd_dev, "%s%s result %d",
		      test_bit(IMG_REQ_CHILD, &img_req->flags) ? "child " : "",
		      obj_op_name(img_req->op_type), *result);
	}
	return done;
}

static void rbd_img_handle_request(struct rbd_img_request *img_req, int result)
{
again:
	if (!__rbd_img_handle_request(img_req, &result))
		return;

	if (test_bit(IMG_REQ_CHILD, &img_req->flags)) {
		struct rbd_obj_request *obj_req = img_req->obj_request;

		rbd_img_request_destroy(img_req);
		if (__rbd_obj_handle_request(obj_req, &result)) {
			img_req = obj_req->img_request;
			goto again;
		}
	} else {
		struct request *rq = blk_mq_rq_from_pdu(img_req);

		rbd_img_request_destroy(img_req);
		blk_mq_end_request(rq, errno_to_blk_status(result));
	}
}

static const struct rbd_client_id rbd_empty_cid;

static bool rbd_cid_equal(const struct rbd_client_id *lhs,
			  const struct rbd_client_id *rhs)
{
	return lhs->gid == rhs->gid && lhs->handle == rhs->handle;
}

static struct rbd_client_id rbd_get_cid(struct rbd_device *rbd_dev)
{
	struct rbd_client_id cid;

	mutex_lock(&rbd_dev->watch_mutex);
	cid.gid = ceph_client_gid(rbd_dev->rbd_client->client);
	cid.handle = rbd_dev->watch_cookie;
	mutex_unlock(&rbd_dev->watch_mutex);
	return cid;
}

/*
 * lock_rwsem must be held for write
 */
static void rbd_set_owner_cid(struct rbd_device *rbd_dev,
			      const struct rbd_client_id *cid)
{
	dout("%s rbd_dev %p %llu-%llu -> %llu-%llu\n", __func__, rbd_dev,
	     rbd_dev->owner_cid.gid, rbd_dev->owner_cid.handle,
	     cid->gid, cid->handle);
	rbd_dev->owner_cid = *cid; /* struct */
}

static void format_lock_cookie(struct rbd_device *rbd_dev, char *buf)
{
	mutex_lock(&rbd_dev->watch_mutex);
	sprintf(buf, "%s %llu", RBD_LOCK_COOKIE_PREFIX, rbd_dev->watch_cookie);
	mutex_unlock(&rbd_dev->watch_mutex);
}

static void __rbd_lock(struct rbd_device *rbd_dev, const char *cookie)
{
	struct rbd_client_id cid = rbd_get_cid(rbd_dev);

	rbd_dev->lock_state = RBD_LOCK_STATE_LOCKED;
	strcpy(rbd_dev->lock_cookie, cookie);
	rbd_set_owner_cid(rbd_dev, &cid);
	queue_work(rbd_dev->task_wq, &rbd_dev->acquired_lock_work);
}

/*
 * lock_rwsem must be held for write
 */
static int rbd_lock(struct rbd_device *rbd_dev)
{
	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
	char cookie[32];
	int ret;

	WARN_ON(__rbd_is_lock_owner(rbd_dev) ||
		rbd_dev->lock_cookie[0] != '\0');

	format_lock_cookie(rbd_dev, cookie);
	ret = ceph_cls_lock(osdc, &rbd_dev->header_oid, &rbd_dev->header_oloc,
			    RBD_LOCK_NAME, CEPH_CLS_LOCK_EXCLUSIVE, cookie,
			    RBD_LOCK_TAG, "", 0);
	if (ret)
		return ret;

	__rbd_lock(rbd_dev, cookie);
	return 0;
}

/*
 * lock_rwsem must be held for write
 */
static void rbd_unlock(struct rbd_device *rbd_dev)
{
	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
	int ret;

	WARN_ON(!__rbd_is_lock_owner(rbd_dev) ||
		rbd_dev->lock_cookie[0] == '\0');

	ret = ceph_cls_unlock(osdc, &rbd_dev->header_oid, &rbd_dev->header_oloc,
			      RBD_LOCK_NAME, rbd_dev->lock_cookie);
	if (ret && ret != -ENOENT)
		rbd_warn(rbd_dev, "failed to unlock header: %d", ret);

	/* treat errors as the image is unlocked */
	rbd_dev->lock_state = RBD_LOCK_STATE_UNLOCKED;
	rbd_dev->lock_cookie[0] = '\0';
	rbd_set_owner_cid(rbd_dev, &rbd_empty_cid);
	queue_work(rbd_dev->task_wq, &rbd_dev->released_lock_work);
}

static int __rbd_notify_op_lock(struct rbd_device *rbd_dev,
				enum rbd_notify_op notify_op,
				struct page ***preply_pages,
				size_t *preply_len)
{
	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
	struct rbd_client_id cid = rbd_get_cid(rbd_dev);
	char buf[4 + 8 + 8 + CEPH_ENCODING_START_BLK_LEN];
	int buf_size = sizeof(buf);
	void *p = buf;

	dout("%s rbd_dev %p notify_op %d\n", __func__, rbd_dev, notify_op);

	/* encode *LockPayload NotifyMessage (op + ClientId) */
	ceph_start_encoding(&p, 2, 1, buf_size - CEPH_ENCODING_START_BLK_LEN);
	ceph_encode_32(&p, notify_op);
	ceph_encode_64(&p, cid.gid);
	ceph_encode_64(&p, cid.handle);

	return ceph_osdc_notify(osdc, &rbd_dev->header_oid,
				&rbd_dev->header_oloc, buf, buf_size,
				RBD_NOTIFY_TIMEOUT, preply_pages, preply_len);
}

static void rbd_notify_op_lock(struct rbd_device *rbd_dev,
			       enum rbd_notify_op notify_op)
{
	__rbd_notify_op_lock(rbd_dev, notify_op, NULL, NULL);
}

static void rbd_notify_acquired_lock(struct work_struct *work)
{
	struct rbd_device *rbd_dev = container_of(work, struct rbd_device,
						  acquired_lock_work);

	rbd_notify_op_lock(rbd_dev, RBD_NOTIFY_OP_ACQUIRED_LOCK);
}

static void rbd_notify_released_lock(struct work_struct *work)
{
	struct rbd_device *rbd_dev = container_of(work, struct rbd_device,
						  released_lock_work);

	rbd_notify_op_lock(rbd_dev, RBD_NOTIFY_OP_RELEASED_LOCK);
}

static int rbd_request_lock(struct rbd_device *rbd_dev)
{
	struct page **reply_pages;
	size_t reply_len;
	bool lock_owner_responded = false;
	int ret;

	dout("%s rbd_dev %p\n", __func__, rbd_dev);

	ret = __rbd_notify_op_lock(rbd_dev, RBD_NOTIFY_OP_REQUEST_LOCK,
				   &reply_pages, &reply_len);
	if (ret && ret != -ETIMEDOUT) {
		rbd_warn(rbd_dev, "failed to request lock: %d", ret);
		goto out;
	}

	if (reply_len > 0 && reply_len <= PAGE_SIZE) {
		void *p = page_address(reply_pages[0]);
		void *const end = p + reply_len;
		u32 n;

		ceph_decode_32_safe(&p, end, n, e_inval); /* num_acks */
		while (n--) {
			u8 struct_v;
			u32 len;

			ceph_decode_need(&p, end, 8 + 8, e_inval);
			p += 8 + 8; /* skip gid and cookie */

			ceph_decode_32_safe(&p, end, len, e_inval);
			if (!len)
				continue;

			if (lock_owner_responded) {
				rbd_warn(rbd_dev,
					 "duplicate lock owners detected");
				ret = -EIO;
				goto out;
			}

			lock_owner_responded = true;
			ret = ceph_start_decoding(&p, end, 1, "ResponseMessage",
						  &struct_v, &len);
			if (ret) {
				rbd_warn(rbd_dev,
					 "failed to decode ResponseMessage: %d",
					 ret);
				goto e_inval;
			}

			ret = ceph_decode_32(&p);
		}
	}

	if (!lock_owner_responded) {
		rbd_warn(rbd_dev, "no lock owners detected");
		ret = -ETIMEDOUT;
	}

out:
	ceph_release_page_vector(reply_pages, calc_pages_for(0, reply_len));
	return ret;

e_inval:
	ret = -EINVAL;
	goto out;
}

/*
 * Either image request state machine(s) or rbd_add_acquire_lock()
 * (i.e. "rbd map").
 */
static void wake_lock_waiters(struct rbd_device *rbd_dev, int result)
{
	struct rbd_img_request *img_req;

	dout("%s rbd_dev %p result %d\n", __func__, rbd_dev, result);
	lockdep_assert_held_write(&rbd_dev->lock_rwsem);

	cancel_delayed_work(&rbd_dev->lock_dwork);
	if (!completion_done(&rbd_dev->acquire_wait)) {
		rbd_assert(list_empty(&rbd_dev->acquiring_list) &&
			   list_empty(&rbd_dev->running_list));
		rbd_dev->acquire_err = result;
		complete_all(&rbd_dev->acquire_wait);
		return;
	}

	list_for_each_entry(img_req, &rbd_dev->acquiring_list, lock_item) {
		mutex_lock(&img_req->state_mutex);
		rbd_assert(img_req->state == RBD_IMG_EXCLUSIVE_LOCK);
		rbd_img_schedule(img_req, result);
		mutex_unlock(&img_req->state_mutex);
	}

	list_splice_tail_init(&rbd_dev->acquiring_list, &rbd_dev->running_list);
}

static int get_lock_owner_info(struct rbd_device *rbd_dev,
			       struct ceph_locker **lockers, u32 *num_lockers)
{
	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
	u8 lock_type;
	char *lock_tag;
	int ret;

	dout("%s rbd_dev %p\n", __func__, rbd_dev);

	ret = ceph_cls_lock_info(osdc, &rbd_dev->header_oid,
				 &rbd_dev->header_oloc, RBD_LOCK_NAME,
				 &lock_type, &lock_tag, lockers, num_lockers);
	if (ret)
		return ret;

	if (*num_lockers == 0) {
		dout("%s rbd_dev %p no lockers detected\n", __func__, rbd_dev);
		goto out;
	}

	if (strcmp(lock_tag, RBD_LOCK_TAG)) {
		rbd_warn(rbd_dev, "locked by external mechanism, tag %s",
			 lock_tag);
		ret = -EBUSY;
		goto out;
	}

	if (lock_type == CEPH_CLS_LOCK_SHARED) {
		rbd_warn(rbd_dev, "shared lock type detected");
		ret = -EBUSY;
		goto out;
	}

	if (strncmp((*lockers)[0].id.cookie, RBD_LOCK_COOKIE_PREFIX,
		    strlen(RBD_LOCK_COOKIE_PREFIX))) {
		rbd_warn(rbd_dev, "locked by external mechanism, cookie %s",
			 (*lockers)[0].id.cookie);
		ret = -EBUSY;
		goto out;
	}

out:
	kfree(lock_tag);
	return ret;
}

static int find_watcher(struct rbd_device *rbd_dev,
			const struct ceph_locker *locker)
{
	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
	struct ceph_watch_item *watchers;
	u32 num_watchers;
	u64 cookie;
	int i;
	int ret;

	ret = ceph_osdc_list_watchers(osdc, &rbd_dev->header_oid,
				      &rbd_dev->header_oloc, &watchers,
				      &num_watchers);
	if (ret)
		return ret;

	sscanf(locker->id.cookie, RBD_LOCK_COOKIE_PREFIX " %llu", &cookie);
	for (i = 0; i < num_watchers; i++) {
		if (!memcmp(&watchers[i].addr, &locker->info.addr,
			    sizeof(locker->info.addr)) &&
		    watchers[i].cookie == cookie) {
			struct rbd_client_id cid = {
				.gid = le64_to_cpu(watchers[i].name.num),
				.handle = cookie,
			};

			dout("%s rbd_dev %p found cid %llu-%llu\n", __func__,
			     rbd_dev, cid.gid, cid.handle);
			rbd_set_owner_cid(rbd_dev, &cid);
			ret = 1;
			goto out;
		}
	}

	dout("%s rbd_dev %p no watchers\n", __func__, rbd_dev);
	ret = 0;
out:
	kfree(watchers);
	return ret;
}

/*
 * lock_rwsem must be held for write
 */
static int rbd_try_lock(struct rbd_device *rbd_dev)
{
	struct ceph_client *client = rbd_dev->rbd_client->client;
	struct ceph_locker *lockers;
	u32 num_lockers;
	int ret;

	for (;;) {
		ret = rbd_lock(rbd_dev);
		if (ret != -EBUSY)
			return ret;

		/* determine if the current lock holder is still alive */
		ret = get_lock_owner_info(rbd_dev, &lockers, &num_lockers);
		if (ret)
			return ret;

		if (num_lockers == 0)
			goto again;

		ret = find_watcher(rbd_dev, lockers);
		if (ret)
			goto out; /* request lock or error */

		rbd_warn(rbd_dev, "breaking header lock owned by %s%llu",
			 ENTITY_NAME(lockers[0].id.name));

		ret = ceph_monc_blocklist_add(&client->monc,
					      &lockers[0].info.addr);
		if (ret) {
			rbd_warn(rbd_dev, "blocklist of %s%llu failed: %d",
				 ENTITY_NAME(lockers[0].id.name), ret);
			goto out;
		}

		ret = ceph_cls_break_lock(&client->osdc, &rbd_dev->header_oid,
					  &rbd_dev->header_oloc, RBD_LOCK_NAME,
					  lockers[0].id.cookie,
					  &lockers[0].id.name);
		if (ret && ret != -ENOENT)
			goto out;

again:
		ceph_free_lockers(lockers, num_lockers);
	}

out:
	ceph_free_lockers(lockers, num_lockers);
	return ret;
}

static int rbd_post_acquire_action(struct rbd_device *rbd_dev)
{
	int ret;

	if (rbd_dev->header.features & RBD_FEATURE_OBJECT_MAP) {
		ret = rbd_object_map_open(rbd_dev);
		if (ret)
			return ret;
	}

	return 0;
}

/*
 * Return:
 *   0 - lock acquired
 *   1 - caller should call rbd_request_lock()
 *  <0 - error
 */
static int rbd_try_acquire_lock(struct rbd_device *rbd_dev)
{
	int ret;

	down_read(&rbd_dev->lock_rwsem);
	dout("%s rbd_dev %p read lock_state %d\n", __func__, rbd_dev,
	     rbd_dev->lock_state);
	if (__rbd_is_lock_owner(rbd_dev)) {
		up_read(&rbd_dev->lock_rwsem);
		return 0;
	}

	up_read(&rbd_dev->lock_rwsem);
	down_write(&rbd_dev->lock_rwsem);
	dout("%s rbd_dev %p write lock_state %d\n", __func__, rbd_dev,
	     rbd_dev->lock_state);
	if (__rbd_is_lock_owner(rbd_dev)) {
		up_write(&rbd_dev->lock_rwsem);
		return 0;
	}

	ret = rbd_try_lock(rbd_dev);
	if (ret < 0) {
		rbd_warn(rbd_dev, "failed to lock header: %d", ret);
		if (ret == -EBLOCKLISTED)
			goto out;

		ret = 1; /* request lock anyway */
	}
	if (ret > 0) {
		up_write(&rbd_dev->lock_rwsem);
		return ret;
	}

	rbd_assert(rbd_dev->lock_state == RBD_LOCK_STATE_LOCKED);
	rbd_assert(list_empty(&rbd_dev->running_list));

	ret = rbd_post_acquire_action(rbd_dev);
	if (ret) {
		rbd_warn(rbd_dev, "post-acquire action failed: %d", ret);
		/*
		 * Can't stay in RBD_LOCK_STATE_LOCKED because
		 * rbd_lock_add_request() would let the request through,
		 * assuming that e.g. object map is locked and loaded.
		 */
		rbd_unlock(rbd_dev);
	}

out:
	wake_lock_waiters(rbd_dev, ret);
	up_write(&rbd_dev->lock_rwsem);
	return ret;
}

static void rbd_acquire_lock(struct work_struct *work)
{
	struct rbd_device *rbd_dev = container_of(to_delayed_work(work),
					    struct rbd_device, lock_dwork);
	int ret;

	dout("%s rbd_dev %p\n", __func__, rbd_dev);
again:
	ret = rbd_try_acquire_lock(rbd_dev);
	if (ret <= 0) {
		dout("%s rbd_dev %p ret %d - done\n", __func__, rbd_dev, ret);
		return;
	}

	ret = rbd_request_lock(rbd_dev);
	if (ret == -ETIMEDOUT) {
		goto again; /* treat this as a dead client */
	} else if (ret == -EROFS) {
		rbd_warn(rbd_dev, "peer will not release lock");
		down_write(&rbd_dev->lock_rwsem);
		wake_lock_waiters(rbd_dev, ret);
		up_write(&rbd_dev->lock_rwsem);
	} else if (ret < 0) {
		rbd_warn(rbd_dev, "error requesting lock: %d", ret);
		mod_delayed_work(rbd_dev->task_wq, &rbd_dev->lock_dwork,
				 RBD_RETRY_DELAY);
	} else {
		/*
		 * lock owner acked, but resend if we don't see them
		 * release the lock
		 */
		dout("%s rbd_dev %p requeuing lock_dwork\n", __func__,
		     rbd_dev);
		mod_delayed_work(rbd_dev->task_wq, &rbd_dev->lock_dwork,
		    msecs_to_jiffies(2 * RBD_NOTIFY_TIMEOUT * MSEC_PER_SEC));
	}
}

static bool rbd_quiesce_lock(struct rbd_device *rbd_dev)
{
	bool need_wait;

	dout("%s rbd_dev %p\n", __func__, rbd_dev);
	lockdep_assert_held_write(&rbd_dev->lock_rwsem);

	if (rbd_dev->lock_state != RBD_LOCK_STATE_LOCKED)
		return false;

	/*
	 * Ensure that all in-flight IO is flushed.
	 */
	rbd_dev->lock_state = RBD_LOCK_STATE_RELEASING;
	rbd_assert(!completion_done(&rbd_dev->releasing_wait));
	need_wait = !list_empty(&rbd_dev->running_list);
	downgrade_write(&rbd_dev->lock_rwsem);
	if (need_wait)
		wait_for_completion(&rbd_dev->releasing_wait);
	up_read(&rbd_dev->lock_rwsem);

	down_write(&rbd_dev->lock_rwsem);
	if (rbd_dev->lock_state != RBD_LOCK_STATE_RELEASING)
		return false;

	rbd_assert(list_empty(&rbd_dev->running_list));
	return true;
}

static void rbd_pre_release_action(struct rbd_device *rbd_dev)
{
	if (rbd_dev->header.features & RBD_FEATURE_OBJECT_MAP)
		rbd_object_map_close(rbd_dev);
}

static void __rbd_release_lock(struct rbd_device *rbd_dev)
{
	rbd_assert(list_empty(&rbd_dev->running_list));

	rbd_pre_release_action(rbd_dev);
	rbd_unlock(rbd_dev);
}

/*
 * lock_rwsem must be held for write
 */
static void rbd_release_lock(struct rbd_device *rbd_dev)
{
	if (!rbd_quiesce_lock(rbd_dev))
		return;

	__rbd_release_lock(rbd_dev);

	/*
	 * Give others a chance to grab the lock - we would re-acquire
	 * almost immediately if we got new IO while draining the running
	 * list otherwise.  We need to ack our own notifications, so this
	 * lock_dwork will be requeued from rbd_handle_released_lock() by
	 * way of maybe_kick_acquire().
	 */
	cancel_delayed_work(&rbd_dev->lock_dwork);
}

static void rbd_release_lock_work(struct work_struct *work)
{
	struct rbd_device *rbd_dev = container_of(work, struct rbd_device,
						  unlock_work);

	down_write(&rbd_dev->lock_rwsem);
	rbd_release_lock(rbd_dev);
	up_write(&rbd_dev->lock_rwsem);
}

static void maybe_kick_acquire(struct rbd_device *rbd_dev)
{
	bool have_requests;

	dout("%s rbd_dev %p\n", __func__, rbd_dev);
	if (__rbd_is_lock_owner(rbd_dev))
		return;

	spin_lock(&rbd_dev->lock_lists_lock);
	have_requests = !list_empty(&rbd_dev->acquiring_list);
	spin_unlock(&rbd_dev->lock_lists_lock);
	if (have_requests || delayed_work_pending(&rbd_dev->lock_dwork)) {
		dout("%s rbd_dev %p kicking lock_dwork\n", __func__, rbd_dev);
		mod_delayed_work(rbd_dev->task_wq, &rbd_dev->lock_dwork, 0);
	}
}

static void rbd_handle_acquired_lock(struct rbd_device *rbd_dev, u8 struct_v,
				     void **p)
{
	struct rbd_client_id cid = { 0 };

	if (struct_v >= 2) {
		cid.gid = ceph_decode_64(p);
		cid.handle = ceph_decode_64(p);
	}

	dout("%s rbd_dev %p cid %llu-%llu\n", __func__, rbd_dev, cid.gid,
	     cid.handle);
	if (!rbd_cid_equal(&cid, &rbd_empty_cid)) {
		down_write(&rbd_dev->lock_rwsem);
		if (rbd_cid_equal(&cid, &rbd_dev->owner_cid)) {
			/*
			 * we already know that the remote client is
			 * the owner
			 */
			up_write(&rbd_dev->lock_rwsem);
			return;
		}

		rbd_set_owner_cid(rbd_dev, &cid);
		downgrade_write(&rbd_dev->lock_rwsem);
	} else {
		down_read(&rbd_dev->lock_rwsem);
	}

	maybe_kick_acquire(rbd_dev);
	up_read(&rbd_dev->lock_rwsem);
}

static void rbd_handle_released_lock(struct rbd_device *rbd_dev, u8 struct_v,
				     void **p)
{
	struct rbd_client_id cid = { 0 };

	if (struct_v >= 2) {
		cid.gid = ceph_decode_64(p);
		cid.handle = ceph_decode_64(p);
	}

	dout("%s rbd_dev %p cid %llu-%llu\n", __func__, rbd_dev, cid.gid,
	     cid.handle);
	if (!rbd_cid_equal(&cid, &rbd_empty_cid)) {
		down_write(&rbd_dev->lock_rwsem);
		if (!rbd_cid_equal(&cid, &rbd_dev->owner_cid)) {
			dout("%s rbd_dev %p unexpected owner, cid %llu-%llu != owner_cid %llu-%llu\n",
			     __func__, rbd_dev, cid.gid, cid.handle,
			     rbd_dev->owner_cid.gid, rbd_dev->owner_cid.handle);
			up_write(&rbd_dev->lock_rwsem);
			return;
		}

		rbd_set_owner_cid(rbd_dev, &rbd_empty_cid);
		downgrade_write(&rbd_dev->lock_rwsem);
	} else {
		down_read(&rbd_dev->lock_rwsem);
	}

	maybe_kick_acquire(rbd_dev);
	up_read(&rbd_dev->lock_rwsem);
}

/*
 * Returns result for ResponseMessage to be encoded (<= 0), or 1 if no
 * ResponseMessage is needed.
 */
static int rbd_handle_request_lock(struct rbd_device *rbd_dev, u8 struct_v,
				   void **p)
{
	struct rbd_client_id my_cid = rbd_get_cid(rbd_dev);
	struct rbd_client_id cid = { 0 };
	int result = 1;

	if (struct_v >= 2) {
		cid.gid = ceph_decode_64(p);
		cid.handle = ceph_decode_64(p);
	}

	dout("%s rbd_dev %p cid %llu-%llu\n", __func__, rbd_dev, cid.gid,
	     cid.handle);
	if (rbd_cid_equal(&cid, &my_cid))
		return result;

	down_read(&rbd_dev->lock_rwsem);
	if (__rbd_is_lock_owner(rbd_dev)) {
		if (rbd_dev->lock_state == RBD_LOCK_STATE_LOCKED &&
		    rbd_cid_equal(&rbd_dev->owner_cid, &rbd_empty_cid))
			goto out_unlock;

		/*
		 * encode ResponseMessage(0) so the peer can detect
		 * a missing owner
		 */
		result = 0;

		if (rbd_dev->lock_state == RBD_LOCK_STATE_LOCKED) {
			if (!rbd_dev->opts->exclusive) {
				dout("%s rbd_dev %p queueing unlock_work\n",
				     __func__, rbd_dev);
				queue_work(rbd_dev->task_wq,
					   &rbd_dev->unlock_work);
			} else {
				/* refuse to release the lock */
				result = -EROFS;
			}
		}
	}

out_unlock:
	up_read(&rbd_dev->lock_rwsem);
	return result;
}

static void __rbd_acknowledge_notify(struct rbd_device *rbd_dev,
				     u64 notify_id, u64 cookie, s32 *result)
{
	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
	char buf[4 + CEPH_ENCODING_START_BLK_LEN];
	int buf_size = sizeof(buf);
	int ret;

	if (result) {
		void *p = buf;

		/* encode ResponseMessage */
		ceph_start_encoding(&p, 1, 1,
				    buf_size - CEPH_ENCODING_START_BLK_LEN);
		ceph_encode_32(&p, *result);
	} else {
		buf_size = 0;
	}

	ret = ceph_osdc_notify_ack(osdc, &rbd_dev->header_oid,
				   &rbd_dev->header_oloc, notify_id, cookie,
				   buf, buf_size);
	if (ret)
		rbd_warn(rbd_dev, "acknowledge_notify failed: %d", ret);
}

static void rbd_acknowledge_notify(struct rbd_device *rbd_dev, u64 notify_id,
				   u64 cookie)
{
	dout("%s rbd_dev %p\n", __func__, rbd_dev);
	__rbd_acknowledge_notify(rbd_dev, notify_id, cookie, NULL);
}

static void rbd_acknowledge_notify_result(struct rbd_device *rbd_dev,
					  u64 notify_id, u64 cookie, s32 result)
{
	dout("%s rbd_dev %p result %d\n", __func__, rbd_dev, result);
	__rbd_acknowledge_notify(rbd_dev, notify_id, cookie, &result);
}

static void rbd_watch_cb(void *arg, u64 notify_id, u64 cookie,
			 u64 notifier_id, void *data, size_t data_len)
{
	struct rbd_device *rbd_dev = arg;
	void *p = data;
	void *const end = p + data_len;
	u8 struct_v = 0;
	u32 len;
	u32 notify_op;
	int ret;

	dout("%s rbd_dev %p cookie %llu notify_id %llu data_len %zu\n",
	     __func__, rbd_dev, cookie, notify_id, data_len);
	if (data_len) {
		ret = ceph_start_decoding(&p, end, 1, "NotifyMessage",
					  &struct_v, &len);
		if (ret) {
			rbd_warn(rbd_dev, "failed to decode NotifyMessage: %d",
				 ret);
			return;
		}

		notify_op = ceph_decode_32(&p);
	} else {
		/* legacy notification for header updates */
		notify_op = RBD_NOTIFY_OP_HEADER_UPDATE;
		len = 0;
	}

	dout("%s rbd_dev %p notify_op %u\n", __func__, rbd_dev, notify_op);
	switch (notify_op) {
	case RBD_NOTIFY_OP_ACQUIRED_LOCK:
		rbd_handle_acquired_lock(rbd_dev, struct_v, &p);
		rbd_acknowledge_notify(rbd_dev, notify_id, cookie);
		break;
	case RBD_NOTIFY_OP_RELEASED_LOCK:
		rbd_handle_released_lock(rbd_dev, struct_v, &p);
		rbd_acknowledge_notify(rbd_dev, notify_id, cookie);
		break;
	case RBD_NOTIFY_OP_REQUEST_LOCK:
		ret = rbd_handle_request_lock(rbd_dev, struct_v, &p);
		if (ret <= 0)
			rbd_acknowledge_notify_result(rbd_dev, notify_id,
						      cookie, ret);
		else
			rbd_acknowledge_notify(rbd_dev, notify_id, cookie);
		break;
	case RBD_NOTIFY_OP_HEADER_UPDATE:
		ret = rbd_dev_refresh(rbd_dev);
		if (ret)
			rbd_warn(rbd_dev, "refresh failed: %d", ret);

		rbd_acknowledge_notify(rbd_dev, notify_id, cookie);
		break;
	default:
		if (rbd_is_lock_owner(rbd_dev))
			rbd_acknowledge_notify_result(rbd_dev, notify_id,
						      cookie, -EOPNOTSUPP);
		else
			rbd_acknowledge_notify(rbd_dev, notify_id, cookie);
		break;
	}
}

static void __rbd_unregister_watch(struct rbd_device *rbd_dev);

static void rbd_watch_errcb(void *arg, u64 cookie, int err)
{
	struct rbd_device *rbd_dev = arg;

	rbd_warn(rbd_dev, "encountered watch error: %d", err);

	down_write(&rbd_dev->lock_rwsem);
	rbd_set_owner_cid(rbd_dev, &rbd_empty_cid);
	up_write(&rbd_dev->lock_rwsem);

	mutex_lock(&rbd_dev->watch_mutex);
	if (rbd_dev->watch_state == RBD_WATCH_STATE_REGISTERED) {
		__rbd_unregister_watch(rbd_dev);
		rbd_dev->watch_state = RBD_WATCH_STATE_ERROR;

		queue_delayed_work(rbd_dev->task_wq, &rbd_dev->watch_dwork, 0);
	}
	mutex_unlock(&rbd_dev->watch_mutex);
}

/*
 * watch_mutex must be locked
 */
static int __rbd_register_watch(struct rbd_device *rbd_dev)
{
	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
	struct ceph_osd_linger_request *handle;

	rbd_assert(!rbd_dev->watch_handle);
	dout("%s rbd_dev %p\n", __func__, rbd_dev);

	handle = ceph_osdc_watch(osdc, &rbd_dev->header_oid,
				 &rbd_dev->header_oloc, rbd_watch_cb,
				 rbd_watch_errcb, rbd_dev);
	if (IS_ERR(handle))
		return PTR_ERR(handle);

	rbd_dev->watch_handle = handle;
	return 0;
}

/*
 * watch_mutex must be locked
 */
static void __rbd_unregister_watch(struct rbd_device *rbd_dev)
{
	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
	int ret;

	rbd_assert(rbd_dev->watch_handle);
	dout("%s rbd_dev %p\n", __func__, rbd_dev);

	ret = ceph_osdc_unwatch(osdc, rbd_dev->watch_handle);
	if (ret)
		rbd_warn(rbd_dev, "failed to unwatch: %d", ret);

	rbd_dev->watch_handle = NULL;
}

static int rbd_register_watch(struct rbd_device *rbd_dev)
{
	int ret;

	mutex_lock(&rbd_dev->watch_mutex);
	rbd_assert(rbd_dev->watch_state == RBD_WATCH_STATE_UNREGISTERED);
	ret = __rbd_register_watch(rbd_dev);
	if (ret)
		goto out;

	rbd_dev->watch_state = RBD_WATCH_STATE_REGISTERED;
	rbd_dev->watch_cookie = rbd_dev->watch_handle->linger_id;

out:
	mutex_unlock(&rbd_dev->watch_mutex);
	return ret;
}

static void cancel_tasks_sync(struct rbd_device *rbd_dev)
{
	dout("%s rbd_dev %p\n", __func__, rbd_dev);

	cancel_work_sync(&rbd_dev->acquired_lock_work);
	cancel_work_sync(&rbd_dev->released_lock_work);
	cancel_delayed_work_sync(&rbd_dev->lock_dwork);
	cancel_work_sync(&rbd_dev->unlock_work);
}

/*
 * header_rwsem must not be held to avoid a deadlock with
 * rbd_dev_refresh() when flushing notifies.
 */
static void rbd_unregister_watch(struct rbd_device *rbd_dev)
{
	cancel_tasks_sync(rbd_dev);

	mutex_lock(&rbd_dev->watch_mutex);
	if (rbd_dev->watch_state == RBD_WATCH_STATE_REGISTERED)
		__rbd_unregister_watch(rbd_dev);
	rbd_dev->watch_state = RBD_WATCH_STATE_UNREGISTERED;
	mutex_unlock(&rbd_dev->watch_mutex);

	cancel_delayed_work_sync(&rbd_dev->watch_dwork);
	ceph_osdc_flush_notifies(&rbd_dev->rbd_client->client->osdc);
}

/*
 * lock_rwsem must be held for write
 */
static void rbd_reacquire_lock(struct rbd_device *rbd_dev)
{
	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
	char cookie[32];
	int ret;

	if (!rbd_quiesce_lock(rbd_dev))
		return;

	format_lock_cookie(rbd_dev, cookie);
	ret = ceph_cls_set_cookie(osdc, &rbd_dev->header_oid,
				  &rbd_dev->header_oloc, RBD_LOCK_NAME,
				  CEPH_CLS_LOCK_EXCLUSIVE, rbd_dev->lock_cookie,
				  RBD_LOCK_TAG, cookie);
	if (ret) {
		if (ret != -EOPNOTSUPP)
			rbd_warn(rbd_dev, "failed to update lock cookie: %d",
				 ret);

		/*
		 * Lock cookie cannot be updated on older OSDs, so do
		 * a manual release and queue an acquire.
		 */
		__rbd_release_lock(rbd_dev);
		queue_delayed_work(rbd_dev->task_wq, &rbd_dev->lock_dwork, 0);
	} else {
		__rbd_lock(rbd_dev, cookie);
		wake_lock_waiters(rbd_dev, 0);
	}
}

static void rbd_reregister_watch(struct work_struct *work)
{
	struct rbd_device *rbd_dev = container_of(to_delayed_work(work),
					    struct rbd_device, watch_dwork);
	int ret;

	dout("%s rbd_dev %p\n", __func__, rbd_dev);

	mutex_lock(&rbd_dev->watch_mutex);
	if (rbd_dev->watch_state != RBD_WATCH_STATE_ERROR) {
		mutex_unlock(&rbd_dev->watch_mutex);
		return;
	}

	ret = __rbd_register_watch(rbd_dev);
	if (ret) {
		rbd_warn(rbd_dev, "failed to reregister watch: %d", ret);
		if (ret != -EBLOCKLISTED && ret != -ENOENT) {
			queue_delayed_work(rbd_dev->task_wq,
					   &rbd_dev->watch_dwork,
					   RBD_RETRY_DELAY);
			mutex_unlock(&rbd_dev->watch_mutex);
			return;
		}

		mutex_unlock(&rbd_dev->watch_mutex);
		down_write(&rbd_dev->lock_rwsem);
		wake_lock_waiters(rbd_dev, ret);
		up_write(&rbd_dev->lock_rwsem);
		return;
	}

	rbd_dev->watch_state = RBD_WATCH_STATE_REGISTERED;
	rbd_dev->watch_cookie = rbd_dev->watch_handle->linger_id;
	mutex_unlock(&rbd_dev->watch_mutex);

	down_write(&rbd_dev->lock_rwsem);
	if (rbd_dev->lock_state == RBD_LOCK_STATE_LOCKED)
		rbd_reacquire_lock(rbd_dev);
	up_write(&rbd_dev->lock_rwsem);

	ret = rbd_dev_refresh(rbd_dev);
	if (ret)
		rbd_warn(rbd_dev, "reregistration refresh failed: %d", ret);
}

/*
 * Synchronous osd object method call.  Returns the number of bytes
 * returned in the outbound buffer, or a negative error code.
 */
static int rbd_obj_method_sync(struct rbd_device *rbd_dev,
			     struct ceph_object_id *oid,
			     struct ceph_object_locator *oloc,
			     const char *method_name,
			     const void *outbound,
			     size_t outbound_size,
			     void *inbound,
			     size_t inbound_size)
{
	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
	struct page *req_page = NULL;
	struct page *reply_page;
	int ret;

	/*
	 * Method calls are ultimately read operations.  The result
	 * should placed into the inbound buffer provided.  They
	 * also supply outbound data--parameters for the object
	 * method.  Currently if this is present it will be a
	 * snapshot id.
	 */
	if (outbound) {
		if (outbound_size > PAGE_SIZE)
			return -E2BIG;

		req_page = alloc_page(GFP_KERNEL);
		if (!req_page)
			return -ENOMEM;

		memcpy(page_address(req_page), outbound, outbound_size);
	}

	reply_page = alloc_page(GFP_KERNEL);
	if (!reply_page) {
		if (req_page)
			__free_page(req_page);
		return -ENOMEM;
	}

	ret = ceph_osdc_call(osdc, oid, oloc, RBD_DRV_NAME, method_name,
			     CEPH_OSD_FLAG_READ, req_page, outbound_size,
			     &reply_page, &inbound_size);
	if (!ret) {
		memcpy(inbound, page_address(reply_page), inbound_size);
		ret = inbound_size;
	}

	if (req_page)
		__free_page(req_page);
	__free_page(reply_page);
	return ret;
}

static void rbd_queue_workfn(struct work_struct *work)
{
	struct rbd_img_request *img_request =
	    container_of(work, struct rbd_img_request, work);
	struct rbd_device *rbd_dev = img_request->rbd_dev;
	enum obj_operation_type op_type = img_request->op_type;
	struct request *rq = blk_mq_rq_from_pdu(img_request);
	u64 offset = (u64)blk_rq_pos(rq) << SECTOR_SHIFT;
	u64 length = blk_rq_bytes(rq);
	u64 mapping_size;
	int result;

	/* Ignore/skip any zero-length requests */
	if (!length) {
		dout("%s: zero-length request\n", __func__);
		result = 0;
		goto err_img_request;
	}

	blk_mq_start_request(rq);

	down_read(&rbd_dev->header_rwsem);
	mapping_size = rbd_dev->mapping.size;
	rbd_img_capture_header(img_request);
	up_read(&rbd_dev->header_rwsem);

	if (offset + length > mapping_size) {
		rbd_warn(rbd_dev, "beyond EOD (%llu~%llu > %llu)", offset,
			 length, mapping_size);
		result = -EIO;
		goto err_img_request;
	}

	dout("%s rbd_dev %p img_req %p %s %llu~%llu\n", __func__, rbd_dev,
	     img_request, obj_op_name(op_type), offset, length);

	if (op_type == OBJ_OP_DISCARD || op_type == OBJ_OP_ZEROOUT)
		result = rbd_img_fill_nodata(img_request, offset, length);
	else
		result = rbd_img_fill_from_bio(img_request, offset, length,
					       rq->bio);
	if (result)
		goto err_img_request;

	rbd_img_handle_request(img_request, 0);
	return;

err_img_request:
	rbd_img_request_destroy(img_request);
	if (result)
		rbd_warn(rbd_dev, "%s %llx at %llx result %d",
			 obj_op_name(op_type), length, offset, result);
	blk_mq_end_request(rq, errno_to_blk_status(result));
}

static blk_status_t rbd_queue_rq(struct blk_mq_hw_ctx *hctx,
		const struct blk_mq_queue_data *bd)
{
	struct rbd_device *rbd_dev = hctx->queue->queuedata;
	struct rbd_img_request *img_req = blk_mq_rq_to_pdu(bd->rq);
	enum obj_operation_type op_type;

	switch (req_op(bd->rq)) {
	case REQ_OP_DISCARD:
		op_type = OBJ_OP_DISCARD;
		break;
	case REQ_OP_WRITE_ZEROES:
		op_type = OBJ_OP_ZEROOUT;
		break;
	case REQ_OP_WRITE:
		op_type = OBJ_OP_WRITE;
		break;
	case REQ_OP_READ:
		op_type = OBJ_OP_READ;
		break;
	default:
		rbd_warn(rbd_dev, "unknown req_op %d", req_op(bd->rq));
		return BLK_STS_IOERR;
	}

	rbd_img_request_init(img_req, rbd_dev, op_type);

	if (rbd_img_is_write(img_req)) {
		if (rbd_is_ro(rbd_dev)) {
			rbd_warn(rbd_dev, "%s on read-only mapping",
				 obj_op_name(img_req->op_type));
			return BLK_STS_IOERR;
		}
		rbd_assert(!rbd_is_snap(rbd_dev));
	}

	INIT_WORK(&img_req->work, rbd_queue_workfn);
	queue_work(rbd_wq, &img_req->work);
	return BLK_STS_OK;
}

static void rbd_free_disk(struct rbd_device *rbd_dev)
{
	blk_cleanup_queue(rbd_dev->disk->queue);
	blk_mq_free_tag_set(&rbd_dev->tag_set);
	put_disk(rbd_dev->disk);
	rbd_dev->disk = NULL;
}

static int rbd_obj_read_sync(struct rbd_device *rbd_dev,
			     struct ceph_object_id *oid,
			     struct ceph_object_locator *oloc,
			     void *buf, int buf_len)

{
	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
	struct ceph_osd_request *req;
	struct page **pages;
	int num_pages = calc_pages_for(0, buf_len);
	int ret;

	req = ceph_osdc_alloc_request(osdc, NULL, 1, false, GFP_KERNEL);
	if (!req)
		return -ENOMEM;

	ceph_oid_copy(&req->r_base_oid, oid);
	ceph_oloc_copy(&req->r_base_oloc, oloc);
	req->r_flags = CEPH_OSD_FLAG_READ;

	pages = ceph_alloc_page_vector(num_pages, GFP_KERNEL);
	if (IS_ERR(pages)) {
		ret = PTR_ERR(pages);
		goto out_req;
	}

	osd_req_op_extent_init(req, 0, CEPH_OSD_OP_READ, 0, buf_len, 0, 0);
	osd_req_op_extent_osd_data_pages(req, 0, pages, buf_len, 0, false,
					 true);

	ret = ceph_osdc_alloc_messages(req, GFP_KERNEL);
	if (ret)
		goto out_req;

	ceph_osdc_start_request(osdc, req, false);
	ret = ceph_osdc_wait_request(osdc, req);
	if (ret >= 0)
		ceph_copy_from_page_vector(pages, buf, 0, ret);

out_req:
	ceph_osdc_put_request(req);
	return ret;
}

/*
 * Read the complete header for the given rbd device.  On successful
 * return, the rbd_dev->header field will contain up-to-date
 * information about the image.
 */
static int rbd_dev_v1_header_info(struct rbd_device *rbd_dev)
{
	struct rbd_image_header_ondisk *ondisk = NULL;
	u32 snap_count = 0;
	u64 names_size = 0;
	u32 want_count;
	int ret;

	/*
	 * The complete header will include an array of its 64-bit
	 * snapshot ids, followed by the names of those snapshots as
	 * a contiguous block of NUL-terminated strings.  Note that
	 * the number of snapshots could change by the time we read
	 * it in, in which case we re-read it.
	 */
	do {
		size_t size;

		kfree(ondisk);

		size = sizeof (*ondisk);
		size += snap_count * sizeof (struct rbd_image_snap_ondisk);
		size += names_size;
		ondisk = kmalloc(size, GFP_KERNEL);
		if (!ondisk)
			return -ENOMEM;

		ret = rbd_obj_read_sync(rbd_dev, &rbd_dev->header_oid,
					&rbd_dev->header_oloc, ondisk, size);
		if (ret < 0)
			goto out;
		if ((size_t)ret < size) {
			ret = -ENXIO;
			rbd_warn(rbd_dev, "short header read (want %zd got %d)",
				size, ret);
			goto out;
		}
		if (!rbd_dev_ondisk_valid(ondisk)) {
			ret = -ENXIO;
			rbd_warn(rbd_dev, "invalid header");
			goto out;
		}

		names_size = le64_to_cpu(ondisk->snap_names_len);
		want_count = snap_count;
		snap_count = le32_to_cpu(ondisk->snap_count);
	} while (snap_count != want_count);

	ret = rbd_header_from_disk(rbd_dev, ondisk);
out:
	kfree(ondisk);

	return ret;
}

static void rbd_dev_update_size(struct rbd_device *rbd_dev)
{
	sector_t size;

	/*
	 * If EXISTS is not set, rbd_dev->disk may be NULL, so don't
	 * try to update its size.  If REMOVING is set, updating size
	 * is just useless work since the device can't be opened.
	 */
	if (test_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags) &&
	    !test_bit(RBD_DEV_FLAG_REMOVING, &rbd_dev->flags)) {
		size = (sector_t)rbd_dev->mapping.size / SECTOR_SIZE;
		dout("setting size to %llu sectors", (unsigned long long)size);
		set_capacity(rbd_dev->disk, size);
		revalidate_disk_size(rbd_dev->disk, true);
	}
}

static int rbd_dev_refresh(struct rbd_device *rbd_dev)
{
	u64 mapping_size;
	int ret;

	down_write(&rbd_dev->header_rwsem);
	mapping_size = rbd_dev->mapping.size;

	ret = rbd_dev_header_info(rbd_dev);
	if (ret)
		goto out;

	/*
	 * If there is a parent, see if it has disappeared due to the
	 * mapped image getting flattened.
	 */
	if (rbd_dev->parent) {
		ret = rbd_dev_v2_parent_info(rbd_dev);
		if (ret)
			goto out;
	}

	rbd_assert(!rbd_is_snap(rbd_dev));
	rbd_dev->mapping.size = rbd_dev->header.image_size;

out:
	up_write(&rbd_dev->header_rwsem);
	if (!ret && mapping_size != rbd_dev->mapping.size)
		rbd_dev_update_size(rbd_dev);

	return ret;
}

static const struct blk_mq_ops rbd_mq_ops = {
	.queue_rq	= rbd_queue_rq,
};

static int rbd_init_disk(struct rbd_device *rbd_dev)
{
	struct gendisk *disk;
	struct request_queue *q;
	unsigned int objset_bytes =
	    rbd_dev->layout.object_size * rbd_dev->layout.stripe_count;
	int err;

	/* create gendisk info */
	disk = alloc_disk(single_major ?
			  (1 << RBD_SINGLE_MAJOR_PART_SHIFT) :
			  RBD_MINORS_PER_MAJOR);
	if (!disk)
		return -ENOMEM;

	snprintf(disk->disk_name, sizeof(disk->disk_name), RBD_DRV_NAME "%d",
		 rbd_dev->dev_id);
	disk->major = rbd_dev->major;
	disk->first_minor = rbd_dev->minor;
	if (single_major)
		disk->flags |= GENHD_FL_EXT_DEVT;
	disk->fops = &rbd_bd_ops;
	disk->private_data = rbd_dev;

	memset(&rbd_dev->tag_set, 0, sizeof(rbd_dev->tag_set));
	rbd_dev->tag_set.ops = &rbd_mq_ops;
	rbd_dev->tag_set.queue_depth = rbd_dev->opts->queue_depth;
	rbd_dev->tag_set.numa_node = NUMA_NO_NODE;
	rbd_dev->tag_set.flags = BLK_MQ_F_SHOULD_MERGE;
	rbd_dev->tag_set.nr_hw_queues = num_present_cpus();
	rbd_dev->tag_set.cmd_size = sizeof(struct rbd_img_request);

	err = blk_mq_alloc_tag_set(&rbd_dev->tag_set);
	if (err)
		goto out_disk;

	q = blk_mq_init_queue(&rbd_dev->tag_set);
	if (IS_ERR(q)) {
		err = PTR_ERR(q);
		goto out_tag_set;
	}

	blk_queue_flag_set(QUEUE_FLAG_NONROT, q);
	/* QUEUE_FLAG_ADD_RANDOM is off by default for blk-mq */

	blk_queue_max_hw_sectors(q, objset_bytes >> SECTOR_SHIFT);
	q->limits.max_sectors = queue_max_hw_sectors(q);
	blk_queue_max_segments(q, USHRT_MAX);
	blk_queue_max_segment_size(q, UINT_MAX);
	blk_queue_io_min(q, rbd_dev->opts->alloc_size);
	blk_queue_io_opt(q, rbd_dev->opts->alloc_size);

	if (rbd_dev->opts->trim) {
		blk_queue_flag_set(QUEUE_FLAG_DISCARD, q);
		q->limits.discard_granularity = rbd_dev->opts->alloc_size;
		blk_queue_max_discard_sectors(q, objset_bytes >> SECTOR_SHIFT);
		blk_queue_max_write_zeroes_sectors(q, objset_bytes >> SECTOR_SHIFT);
	}

	if (!ceph_test_opt(rbd_dev->rbd_client->client, NOCRC))
		blk_queue_flag_set(QUEUE_FLAG_STABLE_WRITES, q);

	/*
	 * disk_release() expects a queue ref from add_disk() and will
	 * put it.  Hold an extra ref until add_disk() is called.
	 */
	WARN_ON(!blk_get_queue(q));
	disk->queue = q;
	q->queuedata = rbd_dev;

	rbd_dev->disk = disk;

	return 0;
out_tag_set:
	blk_mq_free_tag_set(&rbd_dev->tag_set);
out_disk:
	put_disk(disk);
	return err;
}

/*
  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->mapping.size);
}

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

	return sprintf(buf, "0x%016llx\n", rbd_dev->header.features);
}

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);

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

	return sprintf(buf, "(none)\n");
}

static ssize_t rbd_minor_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->minor);
}

static ssize_t rbd_client_addr_show(struct device *dev,
				    struct device_attribute *attr, char *buf)
{
	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
	struct ceph_entity_addr *client_addr =
	    ceph_client_addr(rbd_dev->rbd_client->client);

	return sprintf(buf, "%pISpc/%u\n", &client_addr->in_addr,
		       le32_to_cpu(client_addr->nonce));
}

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_gid(rbd_dev->rbd_client->client));
}

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

	return sprintf(buf, "%pU\n", &rbd_dev->rbd_client->client->fsid);
}

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

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

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

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->spec->pool_name);
}

static ssize_t rbd_pool_id_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->spec->pool_id);
}

static ssize_t rbd_pool_ns_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->spec->pool_ns ?: "");
}

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);

	if (rbd_dev->spec->image_name)
		return sprintf(buf, "%s\n", rbd_dev->spec->image_name);

	return sprintf(buf, "(unknown)\n");
}

static ssize_t rbd_image_id_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->spec->image_id);
}

/*
 * Shows the name of the currently-mapped snapshot (or
 * RBD_SNAP_HEAD_NAME for the base image).
 */
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->spec->snap_name);
}

static ssize_t rbd_snap_id_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", rbd_dev->spec->snap_id);
}

/*
 * For a v2 image, shows the chain of parent images, separated by empty
 * lines.  For v1 images or if there is no parent, shows "(no parent
 * image)".
 */
static ssize_t rbd_parent_show(struct device *dev,
			       struct device_attribute *attr,
			       char *buf)
{
	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
	ssize_t count = 0;

	if (!rbd_dev->parent)
		return sprintf(buf, "(no parent image)\n");

	for ( ; rbd_dev->parent; rbd_dev = rbd_dev->parent) {
		struct rbd_spec *spec = rbd_dev->parent_spec;

		count += sprintf(&buf[count], "%s"
			    "pool_id %llu\npool_name %s\n"
			    "pool_ns %s\n"
			    "image_id %s\nimage_name %s\n"
			    "snap_id %llu\nsnap_name %s\n"
			    "overlap %llu\n",
			    !count ? "" : "\n", /* first? */
			    spec->pool_id, spec->pool_name,
			    spec->pool_ns ?: "",
			    spec->image_id, spec->image_name ?: "(unknown)",
			    spec->snap_id, spec->snap_name,
			    rbd_dev->parent_overlap);
	}

	return count;
}

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 ret;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	ret = rbd_dev_refresh(rbd_dev);
	if (ret)
		return ret;

	return size;
}

static DEVICE_ATTR(size, 0444, rbd_size_show, NULL);
static DEVICE_ATTR(features, 0444, rbd_features_show, NULL);
static DEVICE_ATTR(major, 0444, rbd_major_show, NULL);
static DEVICE_ATTR(minor, 0444, rbd_minor_show, NULL);
static DEVICE_ATTR(client_addr, 0444, rbd_client_addr_show, NULL);
static DEVICE_ATTR(client_id, 0444, rbd_client_id_show, NULL);
static DEVICE_ATTR(cluster_fsid, 0444, rbd_cluster_fsid_show, NULL);
static DEVICE_ATTR(config_info, 0400, rbd_config_info_show, NULL);
static DEVICE_ATTR(pool, 0444, rbd_pool_show, NULL);
static DEVICE_ATTR(pool_id, 0444, rbd_pool_id_show, NULL);
static DEVICE_ATTR(pool_ns, 0444, rbd_pool_ns_show, NULL);
static DEVICE_ATTR(name, 0444, rbd_name_show, NULL);
static DEVICE_ATTR(image_id, 0444, rbd_image_id_show, NULL);
static DEVICE_ATTR(refresh, 0200, NULL, rbd_image_refresh);
static DEVICE_ATTR(current_snap, 0444, rbd_snap_show, NULL);
static DEVICE_ATTR(snap_id, 0444, rbd_snap_id_show, NULL);
static DEVICE_ATTR(parent, 0444, rbd_parent_show, NULL);

static struct attribute *rbd_attrs[] = {
	&dev_attr_size.attr,
	&dev_attr_features.attr,
	&dev_attr_major.attr,
	&dev_attr_minor.attr,
	&dev_attr_client_addr.attr,
	&dev_attr_client_id.attr,
	&dev_attr_cluster_fsid.attr,
	&dev_attr_config_info.attr,
	&dev_attr_pool.attr,
	&dev_attr_pool_id.attr,
	&dev_attr_pool_ns.attr,
	&dev_attr_name.attr,
	&dev_attr_image_id.attr,
	&dev_attr_current_snap.attr,
	&dev_attr_snap_id.attr,
	&dev_attr_parent.attr,
	&dev_attr_refresh.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_dev_release(struct device *dev);

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

static struct rbd_spec *rbd_spec_get(struct rbd_spec *spec)
{
	kref_get(&spec->kref);

	return spec;
}

static void rbd_spec_free(struct kref *kref);
static void rbd_spec_put(struct rbd_spec *spec)
{
	if (spec)
		kref_put(&spec->kref, rbd_spec_free);
}

static struct rbd_spec *rbd_spec_alloc(void)
{
	struct rbd_spec *spec;

	spec = kzalloc(sizeof (*spec), GFP_KERNEL);
	if (!spec)
		return NULL;

	spec->pool_id = CEPH_NOPOOL;
	spec->snap_id = CEPH_NOSNAP;
	kref_init(&spec->kref);

	return spec;
}

static void rbd_spec_free(struct kref *kref)
{
	struct rbd_spec *spec = container_of(kref, struct rbd_spec, kref);

	kfree(spec->pool_name);
	kfree(spec->pool_ns);
	kfree(spec->image_id);
	kfree(spec->image_name);
	kfree(spec->snap_name);
	kfree(spec);
}

static void rbd_dev_free(struct rbd_device *rbd_dev)
{
	WARN_ON(rbd_dev->watch_state != RBD_WATCH_STATE_UNREGISTERED);
	WARN_ON(rbd_dev->lock_state != RBD_LOCK_STATE_UNLOCKED);

	ceph_oid_destroy(&rbd_dev->header_oid);
	ceph_oloc_destroy(&rbd_dev->header_oloc);
	kfree(rbd_dev->config_info);

	rbd_put_client(rbd_dev->rbd_client);
	rbd_spec_put(rbd_dev->spec);
	kfree(rbd_dev->opts);
	kfree(rbd_dev);
}

static void rbd_dev_release(struct device *dev)
{
	struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
	bool need_put = !!rbd_dev->opts;

	if (need_put) {
		destroy_workqueue(rbd_dev->task_wq);
		ida_simple_remove(&rbd_dev_id_ida, rbd_dev->dev_id);
	}

	rbd_dev_free(rbd_dev);

	/*
	 * This is racy, but way better than putting module outside of
	 * the release callback.  The race window is pretty small, so
	 * doing something similar to dm (dm-builtin.c) is overkill.
	 */
	if (need_put)
		module_put(THIS_MODULE);
}

static struct rbd_device *__rbd_dev_create(struct rbd_client *rbdc,
					   struct rbd_spec *spec)
{
	struct rbd_device *rbd_dev;

	rbd_dev = kzalloc(sizeof(*rbd_dev), GFP_KERNEL);
	if (!rbd_dev)
		return NULL;

	spin_lock_init(&rbd_dev->lock);
	INIT_LIST_HEAD(&rbd_dev->node);
	init_rwsem(&rbd_dev->header_rwsem);

	rbd_dev->header.data_pool_id = CEPH_NOPOOL;
	ceph_oid_init(&rbd_dev->header_oid);
	rbd_dev->header_oloc.pool = spec->pool_id;
	if (spec->pool_ns) {
		WARN_ON(!*spec->pool_ns);
		rbd_dev->header_oloc.pool_ns =
		    ceph_find_or_create_string(spec->pool_ns,
					       strlen(spec->pool_ns));
	}

	mutex_init(&rbd_dev->watch_mutex);
	rbd_dev->watch_state = RBD_WATCH_STATE_UNREGISTERED;
	INIT_DELAYED_WORK(&rbd_dev->watch_dwork, rbd_reregister_watch);

	init_rwsem(&rbd_dev->lock_rwsem);
	rbd_dev->lock_state = RBD_LOCK_STATE_UNLOCKED;
	INIT_WORK(&rbd_dev->acquired_lock_work, rbd_notify_acquired_lock);
	INIT_WORK(&rbd_dev->released_lock_work, rbd_notify_released_lock);
	INIT_DELAYED_WORK(&rbd_dev->lock_dwork, rbd_acquire_lock);
	INIT_WORK(&rbd_dev->unlock_work, rbd_release_lock_work);
	spin_lock_init(&rbd_dev->lock_lists_lock);
	INIT_LIST_HEAD(&rbd_dev->acquiring_list);
	INIT_LIST_HEAD(&rbd_dev->running_list);
	init_completion(&rbd_dev->acquire_wait);
	init_completion(&rbd_dev->releasing_wait);

	spin_lock_init(&rbd_dev->object_map_lock);

	rbd_dev->dev.bus = &rbd_bus_type;
	rbd_dev->dev.type = &rbd_device_type;
	rbd_dev->dev.parent = &rbd_root_dev;
	device_initialize(&rbd_dev->dev);

	rbd_dev->rbd_client = rbdc;
	rbd_dev->spec = spec;

	return rbd_dev;
}

/*
 * Create a mapping rbd_dev.
 */
static struct rbd_device *rbd_dev_create(struct rbd_client *rbdc,
					 struct rbd_spec *spec,
					 struct rbd_options *opts)
{
	struct rbd_device *rbd_dev;

	rbd_dev = __rbd_dev_create(rbdc, spec);
	if (!rbd_dev)
		return NULL;

	rbd_dev->opts = opts;

	/* get an id and fill in device name */
	rbd_dev->dev_id = ida_simple_get(&rbd_dev_id_ida, 0,
					 minor_to_rbd_dev_id(1 << MINORBITS),
					 GFP_KERNEL);
	if (rbd_dev->dev_id < 0)
		goto fail_rbd_dev;

	sprintf(rbd_dev->name, RBD_DRV_NAME "%d", rbd_dev->dev_id);
	rbd_dev->task_wq = alloc_ordered_workqueue("%s-tasks", WQ_MEM_RECLAIM,
						   rbd_dev->name);
	if (!rbd_dev->task_wq)
		goto fail_dev_id;

	/* we have a ref from do_rbd_add() */
	__module_get(THIS_MODULE);

	dout("%s rbd_dev %p dev_id %d\n", __func__, rbd_dev, rbd_dev->dev_id);
	return rbd_dev;

fail_dev_id:
	ida_simple_remove(&rbd_dev_id_ida, rbd_dev->dev_id);
fail_rbd_dev:
	rbd_dev_free(rbd_dev);
	return NULL;
}

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

/*
 * Get the size and object order for an image snapshot, or if
 * snap_id is CEPH_NOSNAP, gets this information for the base
 * image.
 */
static int _rbd_dev_v2_snap_size(struct rbd_device *rbd_dev, u64 snap_id,
				u8 *order, u64 *snap_size)
{
	__le64 snapid = cpu_to_le64(snap_id);
	int ret;
	struct {
		u8 order;
		__le64 size;
	} __attribute__ ((packed)) size_buf = { 0 };

	ret = rbd_obj_method_sync(rbd_dev, &rbd_dev->header_oid,
				  &rbd_dev->header_oloc, "get_size",
				  &snapid, sizeof(snapid),
				  &size_buf, sizeof(size_buf));
	dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
	if (ret < 0)
		return ret;
	if (ret < sizeof (size_buf))
		return -ERANGE;

	if (order) {
		*order = size_buf.order;
		dout("  order %u", (unsigned int)*order);
	}
	*snap_size = le64_to_cpu(size_buf.size);

	dout("  snap_id 0x%016llx snap_size = %llu\n",
		(unsigned long long)snap_id,
		(unsigned long long)*snap_size);

	return 0;
}

static int rbd_dev_v2_image_size(struct rbd_device *rbd_dev)
{
	return _rbd_dev_v2_snap_size(rbd_dev, CEPH_NOSNAP,
					&rbd_dev->header.obj_order,
					&rbd_dev->header.image_size);
}

static int rbd_dev_v2_object_prefix(struct rbd_device *rbd_dev)
{
	size_t size;
	void *reply_buf;
	int ret;
	void *p;

	/* Response will be an encoded string, which includes a length */
	size = sizeof(__le32) + RBD_OBJ_PREFIX_LEN_MAX;
	reply_buf = kzalloc(size, GFP_KERNEL);
	if (!reply_buf)
		return -ENOMEM;

	ret = rbd_obj_method_sync(rbd_dev, &rbd_dev->header_oid,
				  &rbd_dev->header_oloc, "get_object_prefix",
				  NULL, 0, reply_buf, size);
	dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
	if (ret < 0)
		goto out;

	p = reply_buf;
	rbd_dev->header.object_prefix = ceph_extract_encoded_string(&p,
						p + ret, NULL, GFP_NOIO);
	ret = 0;

	if (IS_ERR(rbd_dev->header.object_prefix)) {
		ret = PTR_ERR(rbd_dev->header.object_prefix);
		rbd_dev->header.object_prefix = NULL;
	} else {
		dout("  object_prefix = %s\n", rbd_dev->header.object_prefix);
	}
out:
	kfree(reply_buf);

	return ret;
}

static int _rbd_dev_v2_snap_features(struct rbd_device *rbd_dev, u64 snap_id,
				     bool read_only, u64 *snap_features)
{
	struct {
		__le64 snap_id;
		u8 read_only;
	} features_in;
	struct {
		__le64 features;
		__le64 incompat;
	} __attribute__ ((packed)) features_buf = { 0 };
	u64 unsup;
	int ret;

	features_in.snap_id = cpu_to_le64(snap_id);
	features_in.read_only = read_only;

	ret = rbd_obj_method_sync(rbd_dev, &rbd_dev->header_oid,
				  &rbd_dev->header_oloc, "get_features",
				  &features_in, sizeof(features_in),
				  &features_buf, sizeof(features_buf));
	dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
	if (ret < 0)
		return ret;
	if (ret < sizeof (features_buf))
		return -ERANGE;

	unsup = le64_to_cpu(features_buf.incompat) & ~RBD_FEATURES_SUPPORTED;
	if (unsup) {
		rbd_warn(rbd_dev, "image uses unsupported features: 0x%llx",
			 unsup);
		return -ENXIO;
	}

	*snap_features = le64_to_cpu(features_buf.features);

	dout("  snap_id 0x%016llx features = 0x%016llx incompat = 0x%016llx\n",
		(unsigned long long)snap_id,
		(unsigned long long)*snap_features,
		(unsigned long long)le64_to_cpu(features_buf.incompat));

	return 0;
}

static int rbd_dev_v2_features(struct rbd_device *rbd_dev)
{
	return _rbd_dev_v2_snap_features(rbd_dev, CEPH_NOSNAP,
					 rbd_is_ro(rbd_dev),
					 &rbd_dev->header.features);
}

/*
 * These are generic image flags, but since they are used only for
 * object map, store them in rbd_dev->object_map_flags.
 *
 * For the same reason, this function is called only on object map
 * (re)load and not on header refresh.
 */
static int rbd_dev_v2_get_flags(struct rbd_device *rbd_dev)
{
	__le64 snapid = cpu_to_le64(rbd_dev->spec->snap_id);
	__le64 flags;
	int ret;

	ret = rbd_obj_method_sync(rbd_dev, &rbd_dev->header_oid,
				  &rbd_dev->header_oloc, "get_flags",
				  &snapid, sizeof(snapid),
				  &flags, sizeof(flags));
	if (ret < 0)
		return ret;
	if (ret < sizeof(flags))
		return -EBADMSG;

	rbd_dev->object_map_flags = le64_to_cpu(flags);
	return 0;
}

struct parent_image_info {
	u64		pool_id;
	const char	*pool_ns;
	const char	*image_id;
	u64		snap_id;

	bool		has_overlap;
	u64		overlap;
};

/*
 * The caller is responsible for @pii.
 */
static int decode_parent_image_spec(void **p, void *end,
				    struct parent_image_info *pii)
{
	u8 struct_v;
	u32 struct_len;
	int ret;

	ret = ceph_start_decoding(p, end, 1, "ParentImageSpec",
				  &struct_v, &struct_len);
	if (ret)
		return ret;

	ceph_decode_64_safe(p, end, pii->pool_id, e_inval);
	pii->pool_ns = ceph_extract_encoded_string(p, end, NULL, GFP_KERNEL);
	if (IS_ERR(pii->pool_ns)) {
		ret = PTR_ERR(pii->pool_ns);
		pii->pool_ns = NULL;
		return ret;
	}
	pii->image_id = ceph_extract_encoded_string(p, end, NULL, GFP_KERNEL);
	if (IS_ERR(pii->image_id)) {
		ret = PTR_ERR(pii->image_id);
		pii->image_id = NULL;
		return ret;
	}
	ceph_decode_64_safe(p, end, pii->snap_id, e_inval);
	return 0;

e_inval:
	return -EINVAL;
}

static int __get_parent_info(struct rbd_device *rbd_dev,
			     struct page *req_page,
			     struct page *reply_page,
			     struct parent_image_info *pii)
{
	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
	size_t reply_len = PAGE_SIZE;
	void *p, *end;
	int ret;

	ret = ceph_osdc_call(osdc, &rbd_dev->header_oid, &rbd_dev->header_oloc,
			     "rbd", "parent_get", CEPH_OSD_FLAG_READ,
			     req_page, sizeof(u64), &reply_page, &reply_len);
	if (ret)
		return ret == -EOPNOTSUPP ? 1 : ret;

	p = page_address(reply_page);
	end = p + reply_len;
	ret = decode_parent_image_spec(&p, end, pii);
	if (ret)
		return ret;

	ret = ceph_osdc_call(osdc, &rbd_dev->header_oid, &rbd_dev->header_oloc,
			     "rbd", "parent_overlap_get", CEPH_OSD_FLAG_READ,
			     req_page, sizeof(u64), &reply_page, &reply_len);
	if (ret)
		return ret;

	p = page_address(reply_page);
	end = p + reply_len;
	ceph_decode_8_safe(&p, end, pii->has_overlap, e_inval);
	if (pii->has_overlap)
		ceph_decode_64_safe(&p, end, pii->overlap, e_inval);

	return 0;

e_inval:
	return -EINVAL;
}

/*
 * The caller is responsible for @pii.
 */
static int __get_parent_info_legacy(struct rbd_device *rbd_dev,
				    struct page *req_page,
				    struct page *reply_page,
				    struct parent_image_info *pii)
{
	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
	size_t reply_len = PAGE_SIZE;
	void *p, *end;
	int ret;

	ret = ceph_osdc_call(osdc, &rbd_dev->header_oid, &rbd_dev->header_oloc,
			     "rbd", "get_parent", CEPH_OSD_FLAG_READ,
			     req_page, sizeof(u64), &reply_page, &reply_len);
	if (ret)
		return ret;

	p = page_address(reply_page);
	end = p + reply_len;
	ceph_decode_64_safe(&p, end, pii->pool_id, e_inval);
	pii->image_id = ceph_extract_encoded_string(&p, end, NULL, GFP_KERNEL);
	if (IS_ERR(pii->image_id)) {
		ret = PTR_ERR(pii->image_id);
		pii->image_id = NULL;
		return ret;
	}
	ceph_decode_64_safe(&p, end, pii->snap_id, e_inval);
	pii->has_overlap = true;
	ceph_decode_64_safe(&p, end, pii->overlap, e_inval);

	return 0;

e_inval:
	return -EINVAL;
}

static int get_parent_info(struct rbd_device *rbd_dev,
			   struct parent_image_info *pii)
{
	struct page *req_page, *reply_page;
	void *p;
	int ret;

	req_page = alloc_page(GFP_KERNEL);
	if (!req_page)
		return -ENOMEM;

	reply_page = alloc_page(GFP_KERNEL);
	if (!reply_page) {
		__free_page(req_page);
		return -ENOMEM;
	}

	p = page_address(req_page);
	ceph_encode_64(&p, rbd_dev->spec->snap_id);
	ret = __get_parent_info(rbd_dev, req_page, reply_page, pii);
	if (ret > 0)
		ret = __get_parent_info_legacy(rbd_dev, req_page, reply_page,
					       pii);

	__free_page(req_page);
	__free_page(reply_page);
	return ret;
}

static int rbd_dev_v2_parent_info(struct rbd_device *rbd_dev)
{
	struct rbd_spec *parent_spec;
	struct parent_image_info pii = { 0 };
	int ret;

	parent_spec = rbd_spec_alloc();
	if (!parent_spec)
		return -ENOMEM;

	ret = get_parent_info(rbd_dev, &pii);
	if (ret)
		goto out_err;

	dout("%s pool_id %llu pool_ns %s image_id %s snap_id %llu has_overlap %d overlap %llu\n",
	     __func__, pii.pool_id, pii.pool_ns, pii.image_id, pii.snap_id,
	     pii.has_overlap, pii.overlap);

	if (pii.pool_id == CEPH_NOPOOL || !pii.has_overlap) {
		/*
		 * Either the parent never existed, or we have
		 * record of it but the image got flattened so it no
		 * longer has a parent.  When the parent of a
		 * layered image disappears we immediately set the
		 * overlap to 0.  The effect of this is that all new
		 * requests will be treated as if the image had no
		 * parent.
		 *
		 * If !pii.has_overlap, the parent image spec is not
		 * applicable.  It's there to avoid duplication in each
		 * snapshot record.
		 */
		if (rbd_dev->parent_overlap) {
			rbd_dev->parent_overlap = 0;
			rbd_dev_parent_put(rbd_dev);
			pr_info("%s: clone image has been flattened\n",
				rbd_dev->disk->disk_name);
		}

		goto out;	/* No parent?  No problem. */
	}

	/* The ceph file layout needs to fit pool id in 32 bits */

	ret = -EIO;
	if (pii.pool_id > (u64)U32_MAX) {
		rbd_warn(NULL, "parent pool id too large (%llu > %u)",
			(unsigned long long)pii.pool_id, U32_MAX);
		goto out_err;
	}

	/*
	 * The parent won't change (except when the clone is
	 * flattened, already handled that).  So we only need to
	 * record the parent spec we have not already done so.
	 */
	if (!rbd_dev->parent_spec) {
		parent_spec->pool_id = pii.pool_id;
		if (pii.pool_ns && *pii.pool_ns) {
			parent_spec->pool_ns = pii.pool_ns;
			pii.pool_ns = NULL;
		}
		parent_spec->image_id = pii.image_id;
		pii.image_id = NULL;
		parent_spec->snap_id = pii.snap_id;

		rbd_dev->parent_spec = parent_spec;
		parent_spec = NULL;	/* rbd_dev now owns this */
	}

	/*
	 * We always update the parent overlap.  If it's zero we issue
	 * a warning, as we will proceed as if there was no parent.
	 */
	if (!pii.overlap) {
		if (parent_spec) {
			/* refresh, careful to warn just once */
			if (rbd_dev->parent_overlap)
				rbd_warn(rbd_dev,
				    "clone now standalone (overlap became 0)");
		} else {
			/* initial probe */
			rbd_warn(rbd_dev, "clone is standalone (overlap 0)");
		}
	}
	rbd_dev->parent_overlap = pii.overlap;

out:
	ret = 0;
out_err:
	kfree(pii.pool_ns);
	kfree(pii.image_id);
	rbd_spec_put(parent_spec);
	return ret;
}

static int rbd_dev_v2_striping_info(struct rbd_device *rbd_dev)
{
	struct {
		__le64 stripe_unit;
		__le64 stripe_count;
	} __attribute__ ((packed)) striping_info_buf = { 0 };
	size_t size = sizeof (striping_info_buf);
	void *p;
	int ret;

	ret = rbd_obj_method_sync(rbd_dev, &rbd_dev->header_oid,
				&rbd_dev->header_oloc, "get_stripe_unit_count",
				NULL, 0, &striping_info_buf, size);
	dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
	if (ret < 0)
		return ret;
	if (ret < size)
		return -ERANGE;

	p = &striping_info_buf;
	rbd_dev->header.stripe_unit = ceph_decode_64(&p);
	rbd_dev->header.stripe_count = ceph_decode_64(&p);
	return 0;
}

static int rbd_dev_v2_data_pool(struct rbd_device *rbd_dev)
{
	__le64 data_pool_id;
	int ret;

	ret = rbd_obj_method_sync(rbd_dev, &rbd_dev->header_oid,
				  &rbd_dev->header_oloc, "get_data_pool",
				  NULL, 0, &data_pool_id, sizeof(data_pool_id));
	if (ret < 0)
		return ret;
	if (ret < sizeof(data_pool_id))
		return -EBADMSG;

	rbd_dev->header.data_pool_id = le64_to_cpu(data_pool_id);
	WARN_ON(rbd_dev->header.data_pool_id == CEPH_NOPOOL);
	return 0;
}

static char *rbd_dev_image_name(struct rbd_device *rbd_dev)
{
	CEPH_DEFINE_OID_ONSTACK(oid);
	size_t image_id_size;
	char *image_id;
	void *p;
	void *end;
	size_t size;
	void *reply_buf = NULL;
	size_t len = 0;
	char *image_name = NULL;
	int ret;

	rbd_assert(!rbd_dev->spec->image_name);

	len = strlen(rbd_dev->spec->image_id);
	image_id_size = sizeof (__le32) + len;
	image_id = kmalloc(image_id_size, GFP_KERNEL);
	if (!image_id)
		return NULL;

	p = image_id;
	end = image_id + image_id_size;
	ceph_encode_string(&p, end, rbd_dev->spec->image_id, (u32)len);

	size = sizeof (__le32) + RBD_IMAGE_NAME_LEN_MAX;
	reply_buf = kmalloc(size, GFP_KERNEL);
	if (!reply_buf)
		goto out;

	ceph_oid_printf(&oid, "%s", RBD_DIRECTORY);
	ret = rbd_obj_method_sync(rbd_dev, &oid, &rbd_dev->header_oloc,
				  "dir_get_name", image_id, image_id_size,
				  reply_buf, size);
	if (ret < 0)
		goto out;
	p = reply_buf;
	end = reply_buf + ret;

	image_name = ceph_extract_encoded_string(&p, end, &len, GFP_KERNEL);
	if (IS_ERR(image_name))
		image_name = NULL;
	else
		dout("%s: name is %s len is %zd\n", __func__, image_name, len);
out:
	kfree(reply_buf);
	kfree(image_id);

	return image_name;
}

static u64 rbd_v1_snap_id_by_name(struct rbd_device *rbd_dev, const char *name)
{
	struct ceph_snap_context *snapc = rbd_dev->header.snapc;
	const char *snap_name;
	u32 which = 0;

	/* Skip over names until we find the one we are looking for */

	snap_name = rbd_dev->header.snap_names;
	while (which < snapc->num_snaps) {
		if (!strcmp(name, snap_name))
			return snapc->snaps[which];
		snap_name += strlen(snap_name) + 1;
		which++;
	}
	return CEPH_NOSNAP;
}

static u64 rbd_v2_snap_id_by_name(struct rbd_device *rbd_dev, const char *name)
{
	struct ceph_snap_context *snapc = rbd_dev->header.snapc;
	u32 which;
	bool found = false;
	u64 snap_id;

	for (which = 0; !found && which < snapc->num_snaps; which++) {
		const char *snap_name;

		snap_id = snapc->snaps[which];
		snap_name = rbd_dev_v2_snap_name(rbd_dev, snap_id);
		if (IS_ERR(snap_name)) {
			/* ignore no-longer existing snapshots */
			if (PTR_ERR(snap_name) == -ENOENT)
				continue;
			else
				break;
		}
		found = !strcmp(name, snap_name);
		kfree(snap_name);
	}
	return found ? snap_id : CEPH_NOSNAP;
}

/*
 * Assumes name is never RBD_SNAP_HEAD_NAME; returns CEPH_NOSNAP if
 * no snapshot by that name is found, or if an error occurs.
 */
static u64 rbd_snap_id_by_name(struct rbd_device *rbd_dev, const char *name)
{
	if (rbd_dev->image_format == 1)
		return rbd_v1_snap_id_by_name(rbd_dev, name);

	return rbd_v2_snap_id_by_name(rbd_dev, name);
}

/*
 * An image being mapped will have everything but the snap id.
 */
static int rbd_spec_fill_snap_id(struct rbd_device *rbd_dev)
{
	struct rbd_spec *spec = rbd_dev->spec;

	rbd_assert(spec->pool_id != CEPH_NOPOOL && spec->pool_name);
	rbd_assert(spec->image_id && spec->image_name);
	rbd_assert(spec->snap_name);

	if (strcmp(spec->snap_name, RBD_SNAP_HEAD_NAME)) {
		u64 snap_id;

		snap_id = rbd_snap_id_by_name(rbd_dev, spec->snap_name);
		if (snap_id == CEPH_NOSNAP)
			return -ENOENT;

		spec->snap_id = snap_id;
	} else {
		spec->snap_id = CEPH_NOSNAP;
	}

	return 0;
}

/*
 * A parent image will have all ids but none of the names.
 *
 * All names in an rbd spec are dynamically allocated.  It's OK if we
 * can't figure out the name for an image id.
 */
static int rbd_spec_fill_names(struct rbd_device *rbd_dev)
{
	struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
	struct rbd_spec *spec = rbd_dev->spec;
	const char *pool_name;
	const char *image_name;
	const char *snap_name;
	int ret;

	rbd_assert(spec->pool_id != CEPH_NOPOOL);
	rbd_assert(spec->image_id);
	rbd_assert(spec->snap_id != CEPH_NOSNAP);

	/* Get the pool name; we have to make our own copy of this */

	pool_name = ceph_pg_pool_name_by_id(osdc->osdmap, spec->pool_id);
	if (!pool_name) {
		rbd_warn(rbd_dev, "no pool with id %llu", spec->pool_id);
		return -EIO;
	}
	pool_name = kstrdup(pool_name, GFP_KERNEL);
	if (!pool_name)
		return -ENOMEM;

	/* Fetch the image name; tolerate failure here */

	image_name = rbd_dev_image_name(rbd_dev);
	if (!image_name)
		rbd_warn(rbd_dev, "unable to get image name");

	/* Fetch the snapshot name */

	snap_name = rbd_snap_name(rbd_dev, spec->snap_id);
	if (IS_ERR(snap_name)) {
		ret = PTR_ERR(snap_name);
		goto out_err;
	}

	spec->pool_name = pool_name;
	spec->image_name = image_name;
	spec->snap_name = snap_name;

	return 0;

out_err:
	kfree(image_name);
	kfree(pool_name);
	return ret;
}

static int rbd_dev_v2_snap_context(struct rbd_device *rbd_dev)
{
	size_t size;
	int ret;
	void *reply_buf;
	void *p;
	void *end;
	u64 seq;
	u32 snap_count;
	struct ceph_snap_context *snapc;
	u32 i;

	/*
	 * We'll need room for the seq value (maximum snapshot id),
	 * snapshot count, and array of that many snapshot ids.
	 * For now we have a fixed upper limit on the number we're
	 * prepared to receive.
	 */
	size = sizeof (__le64) + sizeof (__le32) +
			RBD_MAX_SNAP_COUNT * sizeof (__le64);
	reply_buf = kzalloc(size, GFP_KERNEL);
	if (!reply_buf)
		return -ENOMEM;

	ret = rbd_obj_method_sync(rbd_dev, &rbd_dev->header_oid,
				  &rbd_dev->header_oloc, "get_snapcontext",
				  NULL, 0, reply_buf, size);
	dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
	if (ret < 0)
		goto out;

	p = reply_buf;
	end = reply_buf + ret;
	ret = -ERANGE;
	ceph_decode_64_safe(&p, end, seq, out);
	ceph_decode_32_safe(&p, end, snap_count, out);

	/*
	 * Make sure the reported number of snapshot ids wouldn't go
	 * beyond the end of our buffer.  But before checking that,
	 * make sure the computed size of the snapshot context we
	 * allocate is representable in a size_t.
	 */
	if (snap_count > (SIZE_MAX - sizeof (struct ceph_snap_context))
				 / sizeof (u64)) {
		ret = -EINVAL;
		goto out;
	}
	if (!ceph_has_room(&p, end, snap_count * sizeof (__le64)))
		goto out;
	ret = 0;

	snapc = ceph_create_snap_context(snap_count, GFP_KERNEL);
	if (!snapc) {
		ret = -ENOMEM;
		goto out;
	}
	snapc->seq = seq;
	for (i = 0; i < snap_count; i++)
		snapc->snaps[i] = ceph_decode_64(&p);

	ceph_put_snap_context(rbd_dev->header.snapc);
	rbd_dev->header.snapc = snapc;

	dout("  snap context seq = %llu, snap_count = %u\n",
		(unsigned long long)seq, (unsigned int)snap_count);
out:
	kfree(reply_buf);

	return ret;
}

static const char *rbd_dev_v2_snap_name(struct rbd_device *rbd_dev,
					u64 snap_id)
{
	size_t size;
	void *reply_buf;
	__le64 snapid;
	int ret;
	void *p;
	void *end;
	char *snap_name;

	size = sizeof (__le32) + RBD_MAX_SNAP_NAME_LEN;
	reply_buf = kmalloc(size, GFP_KERNEL);
	if (!reply_buf)
		return ERR_PTR(-ENOMEM);

	snapid = cpu_to_le64(snap_id);
	ret = rbd_obj_method_sync(rbd_dev, &rbd_dev->header_oid,
				  &rbd_dev->header_oloc, "get_snapshot_name",
				  &snapid, sizeof(snapid), reply_buf, size);
	dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
	if (ret < 0) {
		snap_name = ERR_PTR(ret);
		goto out;
	}

	p = reply_buf;
	end = reply_buf + ret;
	snap_name = ceph_extract_encoded_string(&p, end, NULL, GFP_KERNEL);
	if (IS_ERR(snap_name))
		goto out;

	dout("  snap_id 0x%016llx snap_name = %s\n",
		(unsigned long long)snap_id, snap_name);
out:
	kfree(reply_buf);

	return snap_name;
}

static int rbd_dev_v2_header_info(struct rbd_device *rbd_dev)
{
	bool first_time = rbd_dev->header.object_prefix == NULL;
	int ret;

	ret = rbd_dev_v2_image_size(rbd_dev);
	if (ret)
		return ret;

	if (first_time) {
		ret = rbd_dev_v2_header_onetime(rbd_dev);
		if (ret)
			return ret;
	}

	ret = rbd_dev_v2_snap_context(rbd_dev);
	if (ret && first_time) {
		kfree(rbd_dev->header.object_prefix);
		rbd_dev->header.object_prefix = NULL;
	}

	return ret;
}

static int rbd_dev_header_info(struct rbd_device *rbd_dev)
{
	rbd_assert(rbd_image_format_valid(rbd_dev->image_format));

	if (rbd_dev->image_format == 1)
		return rbd_dev_v1_header_info(rbd_dev);

	return rbd_dev_v2_header_info(rbd_dev);
}

/*
 * 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, dynamically allocates a buffer big
 * enough to hold a copy of it, and copies the token into the new
 * buffer.  The copy is guaranteed to be terminated with '\0'.  Note
 * that a duplicate buffer is created even for a zero-length token.
 *
 * Returns a pointer to the newly-allocated duplicate, or a null
 * pointer if memory for the duplicate was not available.  If
 * the lenp argument is a non-null pointer, the length of the token
 * (not including the '\0') is returned in *lenp.
 *
 * If successful, the *buf pointer will be updated to point beyond
 * the end of the found token.
 *
 * Note: uses GFP_KERNEL for allocation.
 */
static inline char *dup_token(const char **buf, size_t *lenp)
{
	char *dup;
	size_t len;

	len = next_token(buf);
	dup = kmemdup(*buf, len + 1, GFP_KERNEL);
	if (!dup)
		return NULL;
	*(dup + len) = '\0';
	*buf += len;

	if (lenp)
		*lenp = len;

	return dup;
}

static int rbd_parse_param(struct fs_parameter *param,
			    struct rbd_parse_opts_ctx *pctx)
{
	struct rbd_options *opt = pctx->opts;
	struct fs_parse_result result;
	struct p_log log = {.prefix = "rbd"};
	int token, ret;

	ret = ceph_parse_param(param, pctx->copts, NULL);
	if (ret != -ENOPARAM)
		return ret;

	token = __fs_parse(&log, rbd_parameters, param, &result);
	dout("%s fs_parse '%s' token %d\n", __func__, param->key, token);
	if (token < 0) {
		if (token == -ENOPARAM)
			return inval_plog(&log, "Unknown parameter '%s'",
					  param->key);
		return token;
	}

	switch (token) {
	case Opt_queue_depth:
		if (result.uint_32 < 1)
			goto out_of_range;
		opt->queue_depth = result.uint_32;
		break;
	case Opt_alloc_size:
		if (result.uint_32 < SECTOR_SIZE)
			goto out_of_range;
		if (!is_power_of_2(result.uint_32))
			return inval_plog(&log, "alloc_size must be a power of 2");
		opt->alloc_size = result.uint_32;
		break;
	case Opt_lock_timeout:
		/* 0 is "wait forever" (i.e. infinite timeout) */
		if (result.uint_32 > INT_MAX / 1000)
			goto out_of_range;
		opt->lock_timeout = msecs_to_jiffies(result.uint_32 * 1000);
		break;
	case Opt_pool_ns:
		kfree(pctx->spec->pool_ns);
		pctx->spec->pool_ns = param->string;
		param->string = NULL;
		break;
	case Opt_compression_hint:
		switch (result.uint_32) {
		case Opt_compression_hint_none:
			opt->alloc_hint_flags &=
			    ~(CEPH_OSD_ALLOC_HINT_FLAG_COMPRESSIBLE |
			      CEPH_OSD_ALLOC_HINT_FLAG_INCOMPRESSIBLE);
			break;
		case Opt_compression_hint_compressible:
			opt->alloc_hint_flags |=
			    CEPH_OSD_ALLOC_HINT_FLAG_COMPRESSIBLE;
			opt->alloc_hint_flags &=
			    ~CEPH_OSD_ALLOC_HINT_FLAG_INCOMPRESSIBLE;
			break;
		case Opt_compression_hint_incompressible:
			opt->alloc_hint_flags |=
			    CEPH_OSD_ALLOC_HINT_FLAG_INCOMPRESSIBLE;
			opt->alloc_hint_flags &=
			    ~CEPH_OSD_ALLOC_HINT_FLAG_COMPRESSIBLE;
			break;
		default:
			BUG();
		}
		break;
	case Opt_read_only:
		opt->read_only = true;
		break;
	case Opt_read_write:
		opt->read_only = false;
		break;
	case Opt_lock_on_read:
		opt->lock_on_read = true;
		break;
	case Opt_exclusive:
		opt->exclusive = true;
		break;
	case Opt_notrim:
		opt->trim = false;
		break;
	default:
		BUG();
	}

	return 0;

out_of_range:
	return inval_plog(&log, "%s out of range", param->key);
}

/*
 * This duplicates most of generic_parse_monolithic(), untying it from
 * fs_context and skipping standard superblock and security options.
 */
static int rbd_parse_options(char *options, struct rbd_parse_opts_ctx *pctx)
{
	char *key;
	int ret = 0;

	dout("%s '%s'\n", __func__, options);
	while ((key = strsep(&options, ",")) != NULL) {
		if (*key) {
			struct fs_parameter param = {
				.key	= key,
				.type	= fs_value_is_flag,
			};
			char *value = strchr(key, '=');
			size_t v_len = 0;

			if (value) {
				if (value == key)
					continue;
				*value++ = 0;
				v_len = strlen(value);
				param.string = kmemdup_nul(value, v_len,
							   GFP_KERNEL);
				if (!param.string)
					return -ENOMEM;
				param.type = fs_value_is_string;
			}
			param.size = v_len;

			ret = rbd_parse_param(&param, pctx);
			kfree(param.string);
			if (ret)
				break;
		}
	}

	return ret;
}

/*
 * Parse the options provided for an "rbd add" (i.e., rbd image
 * mapping) request.  These arrive via a write to /sys/bus/rbd/add,
 * and the data written is passed here via a NUL-terminated buffer.
 * Returns 0 if successful or an error code otherwise.
 *
 * The information extracted from these options is recorded in
 * the other parameters which return dynamically-allocated
 * structures:
 *  ceph_opts
 *      The address of a pointer that will refer to a ceph options
 *      structure.  Caller must release the returned pointer using
 *      ceph_destroy_options() when it is no longer needed.
 *  rbd_opts
 *	Address of an rbd options pointer.  Fully initialized by
 *	this function; caller must release with kfree().
 *  spec
 *	Address of an rbd image specification pointer.  Fully
 *	initialized by this function based on parsed options.
 *	Caller must release with rbd_spec_put().
 *
 * The options passed take this form:
 *  <mon_addrs> <options> <pool_name> <image_name> [<snap_id>]
 * where:
 *  <mon_addrs>
 *      A comma-separated list of one or more monitor addresses.
 *      A monitor address is an ip address, optionally followed
 *      by a port number (separated by a colon).
 *        I.e.:  ip1[:port1][,ip2[:port2]...]
 *  <options>
 *      A comma-separated list of ceph and/or rbd options.
 *  <pool_name>
 *      The name of the rados pool containing the rbd image.
 *  <image_name>
 *      The name of the image in that pool to map.
 *  <snap_id>
 *      An optional snapshot id.  If provided, the mapping will
 *      present data from the image at the time that snapshot was
 *      created.  The image head is used if no snapshot id is
 *      provided.  Snapshot mappings are always read-only.
 */
static int rbd_add_parse_args(const char *buf,
				struct ceph_options **ceph_opts,
				struct rbd_options **opts,
				struct rbd_spec **rbd_spec)
{
	size_t len;
	char *options;
	const char *mon_addrs;
	char *snap_name;
	size_t mon_addrs_size;
	struct rbd_parse_opts_ctx pctx = { 0 };
	int ret;

	/* The first four tokens are required */

	len = next_token(&buf);
	if (!len) {
		rbd_warn(NULL, "no monitor address(es) provided");
		return -EINVAL;
	}
	mon_addrs = buf;
	mon_addrs_size = len;
	buf += len;

	ret = -EINVAL;
	options = dup_token(&buf, NULL);
	if (!options)
		return -ENOMEM;
	if (!*options) {
		rbd_warn(NULL, "no options provided");
		goto out_err;
	}

	pctx.spec = rbd_spec_alloc();
	if (!pctx.spec)
		goto out_mem;

	pctx.spec->pool_name = dup_token(&buf, NULL);
	if (!pctx.spec->pool_name)
		goto out_mem;
	if (!*pctx.spec->pool_name) {
		rbd_warn(NULL, "no pool name provided");
		goto out_err;
	}

	pctx.spec->image_name = dup_token(&buf, NULL);
	if (!pctx.spec->image_name)
		goto out_mem;
	if (!*pctx.spec->image_name) {
		rbd_warn(NULL, "no image name provided");
		goto out_err;
	}

	/*
	 * Snapshot name is optional; default is to use "-"
	 * (indicating the head/no snapshot).
	 */
	len = next_token(&buf);
	if (!len) {
		buf = RBD_SNAP_HEAD_NAME; /* No snapshot supplied */
		len = sizeof (RBD_SNAP_HEAD_NAME) - 1;
	} else if (len > RBD_MAX_SNAP_NAME_LEN) {
		ret = -ENAMETOOLONG;
		goto out_err;
	}
	snap_name = kmemdup(buf, len + 1, GFP_KERNEL);
	if (!snap_name)
		goto out_mem;
	*(snap_name + len) = '\0';
	pctx.spec->snap_name = snap_name;

	pctx.copts = ceph_alloc_options();
	if (!pctx.copts)
		goto out_mem;

	/* Initialize all rbd options to the defaults */

	pctx.opts = kzalloc(sizeof(*pctx.opts), GFP_KERNEL);
	if (!pctx.opts)
		goto out_mem;

	pctx.opts->read_only = RBD_READ_ONLY_DEFAULT;
	pctx.opts->queue_depth = RBD_QUEUE_DEPTH_DEFAULT;
	pctx.opts->alloc_size = RBD_ALLOC_SIZE_DEFAULT;
	pctx.opts->lock_timeout = RBD_LOCK_TIMEOUT_DEFAULT;
	pctx.opts->lock_on_read = RBD_LOCK_ON_READ_DEFAULT;
	pctx.opts->exclusive = RBD_EXCLUSIVE_DEFAULT;
	pctx.opts->trim = RBD_TRIM_DEFAULT;

	ret = ceph_parse_mon_ips(mon_addrs, mon_addrs_size, pctx.copts, NULL);
	if (ret)
		goto out_err;

	ret = rbd_parse_options(options, &pctx);
	if (ret)
		goto out_err;

	*ceph_opts = pctx.copts;
	*opts = pctx.opts;
	*rbd_spec = pctx.spec;
	kfree(options);
	return 0;

out_mem:
	ret = -ENOMEM;
out_err:
	kfree(pctx.opts);
	ceph_destroy_options(pctx.copts);
	rbd_spec_put(pctx.spec);
	kfree(options);
	return ret;
}

static void rbd_dev_image_unlock(struct rbd_device *rbd_dev)
{
	down_write(&rbd_dev->lock_rwsem);
	if (__rbd_is_lock_owner(rbd_dev))
		__rbd_release_lock(rbd_dev);
	up_write(&rbd_dev->lock_rwsem);
}

/*
 * If the wait is interrupted, an error is returned even if the lock
 * was successfully acquired.  rbd_dev_image_unlock() will release it
 * if needed.
 */
static int rbd_add_acquire_lock(struct rbd_device *rbd_dev)
{
	long ret;

	if (!(rbd_dev->header.features & RBD_FEATURE_EXCLUSIVE_LOCK)) {
		if (!rbd_dev->opts->exclusive && !rbd_dev->opts->lock_on_read)
			return 0;

		rbd_warn(rbd_dev, "exclusive-lock feature is not enabled");
		return -EINVAL;
	}

	if (rbd_is_ro(rbd_dev))
		return 0;

	rbd_assert(!rbd_is_lock_owner(rbd_dev));
	queue_delayed_work(rbd_dev->task_wq, &rbd_dev->lock_dwork, 0);
	ret = wait_for_completion_killable_timeout(&rbd_dev->acquire_wait,
			    ceph_timeout_jiffies(rbd_dev->opts->lock_timeout));
	if (ret > 0) {
		ret = rbd_dev->acquire_err;
	} else {
		cancel_delayed_work_sync(&rbd_dev->lock_dwork);
		if (!ret)
			ret = -ETIMEDOUT;
	}

	if (ret) {
		rbd_warn(rbd_dev, "failed to acquire exclusive lock: %ld", ret);
		return ret;
	}

	/*
	 * The lock may have been released by now, unless automatic lock
	 * transitions are disabled.
	 */
	rbd_assert(!rbd_dev->opts->exclusive || rbd_is_lock_owner(rbd_dev));
	return 0;
}

/*
 * An rbd format 2 image has a unique identifier, distinct from the
 * name given to it by the user.  Internally, that identifier is
 * what's used to specify the names of objects related to the image.
 *
 * A special "rbd id" object is used to map an rbd image name to its
 * id.  If that object doesn't exist, then there is no v2 rbd image
 * with the supplied name.
 *
 * This function will record the given rbd_dev's image_id field if
 * it can be determined, and in that case will return 0.  If any
 * errors occur a negative errno will be returned and the rbd_dev's
 * image_id field will be unchanged (and should be NULL).
 */
static int rbd_dev_image_id(struct rbd_device *rbd_dev)
{
	int ret;
	size_t size;
	CEPH_DEFINE_OID_ONSTACK(oid);
	void *response;
	char *image_id;

	/*
	 * When probing a parent image, the image id is already
	 * known (and the image name likely is not).  There's no
	 * need to fetch the image id again in this case.  We
	 * do still need to set the image format though.
	 */
	if (rbd_dev->spec->image_id) {
		rbd_dev->image_format = *rbd_dev->spec->image_id ? 2 : 1;

		return 0;
	}

	/*
	 * First, see if the format 2 image id file exists, and if
	 * so, get the image's persistent id from it.
	 */
	ret = ceph_oid_aprintf(&oid, GFP_KERNEL, "%s%s", RBD_ID_PREFIX,
			       rbd_dev->spec->image_name);
	if (ret)
		return ret;

	dout("rbd id object name is %s\n", oid.name);

	/* Response will be an encoded string, which includes a length */
	size = sizeof (__le32) + RBD_IMAGE_ID_LEN_MAX;
	response = kzalloc(size, GFP_NOIO);
	if (!response) {
		ret = -ENOMEM;
		goto out;
	}

	/* If it doesn't exist we'll assume it's a format 1 image */

	ret = rbd_obj_method_sync(rbd_dev, &oid, &rbd_dev->header_oloc,
				  "get_id", NULL, 0,
				  response, size);
	dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
	if (ret == -ENOENT) {
		image_id = kstrdup("", GFP_KERNEL);
		ret = image_id ? 0 : -ENOMEM;
		if (!ret)
			rbd_dev->image_format = 1;
	} else if (ret >= 0) {
		void *p = response;

		image_id = ceph_extract_encoded_string(&p, p + ret,
						NULL, GFP_NOIO);
		ret = PTR_ERR_OR_ZERO(image_id);
		if (!ret)
			rbd_dev->image_format = 2;
	}

	if (!ret) {
		rbd_dev->spec->image_id = image_id;
		dout("image_id is %s\n", image_id);
	}
out:
	kfree(response);
	ceph_oid_destroy(&oid);
	return ret;
}

/*
 * Undo whatever state changes are made by v1 or v2 header info
 * call.
 */
static void rbd_dev_unprobe(struct rbd_device *rbd_dev)
{
	struct rbd_image_header	*header;

	rbd_dev_parent_put(rbd_dev);
	rbd_object_map_free(rbd_dev);
	rbd_dev_mapping_clear(rbd_dev);

	/* Free dynamic fields from the header, then zero it out */

	header = &rbd_dev->header;
	ceph_put_snap_context(header->snapc);
	kfree(header->snap_sizes);
	kfree(header->snap_names);
	kfree(header->object_prefix);
	memset(header, 0, sizeof (*header));
}

static int rbd_dev_v2_header_onetime(struct rbd_device *rbd_dev)
{
	int ret;

	ret = rbd_dev_v2_object_prefix(rbd_dev);
	if (ret)
		goto out_err;

	/*
	 * Get the and check features for the image.  Currently the
	 * features are assumed to never change.
	 */
	ret = rbd_dev_v2_features(rbd_dev);
	if (ret)
		goto out_err;

	/* If the image supports fancy striping, get its parameters */

	if (rbd_dev->header.features & RBD_FEATURE_STRIPINGV2) {
		ret = rbd_dev_v2_striping_info(rbd_dev);
		if (ret < 0)
			goto out_err;
	}

	if (rbd_dev->header.features & RBD_FEATURE_DATA_POOL) {
		ret = rbd_dev_v2_data_pool(rbd_dev);
		if (ret)
			goto out_err;
	}

	rbd_init_layout(rbd_dev);
	return 0;

out_err:
	rbd_dev->header.features = 0;
	kfree(rbd_dev->header.object_prefix);
	rbd_dev->header.object_prefix = NULL;
	return ret;
}

/*
 * @depth is rbd_dev_image_probe() -> rbd_dev_probe_parent() ->
 * rbd_dev_image_probe() recursion depth, which means it's also the
 * length of the already discovered part of the parent chain.
 */
static int rbd_dev_probe_parent(struct rbd_device *rbd_dev, int depth)
{
	struct rbd_device *parent = NULL;
	int ret;

	if (!rbd_dev->parent_spec)
		return 0;

	if (++depth > RBD_MAX_PARENT_CHAIN_LEN) {
		pr_info("parent chain is too long (%d)\n", depth);
		ret = -EINVAL;
		goto out_err;
	}

	parent = __rbd_dev_create(rbd_dev->rbd_client, rbd_dev->parent_spec);
	if (!parent) {
		ret = -ENOMEM;
		goto out_err;
	}

	/*
	 * Images related by parent/child relationships always share
	 * rbd_client and spec/parent_spec, so bump their refcounts.
	 */
	__rbd_get_client(rbd_dev->rbd_client);
	rbd_spec_get(rbd_dev->parent_spec);

	__set_bit(RBD_DEV_FLAG_READONLY, &parent->flags);

	ret = rbd_dev_image_probe(parent, depth);
	if (ret < 0)
		goto out_err;

	rbd_dev->parent = parent;
	atomic_set(&rbd_dev->parent_ref, 1);
	return 0;

out_err:
	rbd_dev_unparent(rbd_dev);
	rbd_dev_destroy(parent);
	return ret;
}

static void rbd_dev_device_release(struct rbd_device *rbd_dev)
{
	clear_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags);
	rbd_free_disk(rbd_dev);
	if (!single_major)
		unregister_blkdev(rbd_dev->major, rbd_dev->name);
}

/*
 * rbd_dev->header_rwsem must be locked for write and will be unlocked
 * upon return.
 */
static int rbd_dev_device_setup(struct rbd_device *rbd_dev)
{
	int ret;

	/* Record our major and minor device numbers. */

	if (!single_major) {
		ret = register_blkdev(0, rbd_dev->name);
		if (ret < 0)
			goto err_out_unlock;

		rbd_dev->major = ret;
		rbd_dev->minor = 0;
	} else {
		rbd_dev->major = rbd_major;
		rbd_dev->minor = rbd_dev_id_to_minor(rbd_dev->dev_id);
	}

	/* Set up the blkdev mapping. */

	ret = rbd_init_disk(rbd_dev);
	if (ret)
		goto err_out_blkdev;

	set_capacity(rbd_dev->disk, rbd_dev->mapping.size / SECTOR_SIZE);
	set_disk_ro(rbd_dev->disk, rbd_is_ro(rbd_dev));

	ret = dev_set_name(&rbd_dev->dev, "%d", rbd_dev->dev_id);
	if (ret)
		goto err_out_disk;

	set_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags);
	up_write(&rbd_dev->header_rwsem);
	return 0;

err_out_disk:
	rbd_free_disk(rbd_dev);
err_out_blkdev:
	if (!single_major)
		unregister_blkdev(rbd_dev->major, rbd_dev->name);
err_out_unlock:
	up_write(&rbd_dev->header_rwsem);
	return ret;
}

static int rbd_dev_header_name(struct rbd_device *rbd_dev)
{
	struct rbd_spec *spec = rbd_dev->spec;
	int ret;

	/* Record the header object name for this rbd image. */

	rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
	if (rbd_dev->image_format == 1)
		ret = ceph_oid_aprintf(&rbd_dev->header_oid, GFP_KERNEL, "%s%s",
				       spec->image_name, RBD_SUFFIX);
	else
		ret = ceph_oid_aprintf(&rbd_dev->header_oid, GFP_KERNEL, "%s%s",
				       RBD_HEADER_PREFIX, spec->image_id);

	return ret;
}

static void rbd_print_dne(struct rbd_device *rbd_dev, bool is_snap)
{
	if (!is_snap) {
		pr_info("image %s/%s%s%s does not exist\n",
			rbd_dev->spec->pool_name,
			rbd_dev->spec->pool_ns ?: "",
			rbd_dev->spec->pool_ns ? "/" : "",
			rbd_dev->spec->image_name);
	} else {
		pr_info("snap %s/%s%s%s@%s does not exist\n",
			rbd_dev->spec->pool_name,
			rbd_dev->spec->pool_ns ?: "",
			rbd_dev->spec->pool_ns ? "/" : "",
			rbd_dev->spec->image_name,
			rbd_dev->spec->snap_name);
	}
}

static void rbd_dev_image_release(struct rbd_device *rbd_dev)
{
	if (!rbd_is_ro(rbd_dev))
		rbd_unregister_watch(rbd_dev);

	rbd_dev_unprobe(rbd_dev);
	rbd_dev->image_format = 0;
	kfree(rbd_dev->spec->image_id);
	rbd_dev->spec->image_id = NULL;
}

/*
 * Probe for the existence of the header object for the given rbd
 * device.  If this image is the one being mapped (i.e., not a
 * parent), initiate a watch on its header object before using that
 * object to get detailed information about the rbd image.
 *
 * On success, returns with header_rwsem held for write if called
 * with @depth == 0.
 */
static int rbd_dev_image_probe(struct rbd_device *rbd_dev, int depth)
{
	bool need_watch = !rbd_is_ro(rbd_dev);
	int ret;

	/*
	 * Get the id from the image id object.  Unless there's an
	 * error, rbd_dev->spec->image_id will be filled in with
	 * a dynamically-allocated string, and rbd_dev->image_format
	 * will be set to either 1 or 2.
	 */
	ret = rbd_dev_image_id(rbd_dev);
	if (ret)
		return ret;

	ret = rbd_dev_header_name(rbd_dev);
	if (ret)
		goto err_out_format;

	if (need_watch) {
		ret = rbd_register_watch(rbd_dev);
		if (ret) {
			if (ret == -ENOENT)
				rbd_print_dne(rbd_dev, false);
			goto err_out_format;
		}
	}

	if (!depth)
		down_write(&rbd_dev->header_rwsem);

	ret = rbd_dev_header_info(rbd_dev);
	if (ret) {
		if (ret == -ENOENT && !need_watch)
			rbd_print_dne(rbd_dev, false);
		goto err_out_probe;
	}

	/*
	 * If this image is the one being mapped, we have pool name and
	 * id, image name and id, and snap name - need to fill snap id.
	 * Otherwise this is a parent image, identified by pool, image
	 * and snap ids - need to fill in names for those ids.
	 */
	if (!depth)
		ret = rbd_spec_fill_snap_id(rbd_dev);
	else
		ret = rbd_spec_fill_names(rbd_dev);
	if (ret) {
		if (ret == -ENOENT)
			rbd_print_dne(rbd_dev, true);
		goto err_out_probe;
	}

	ret = rbd_dev_mapping_set(rbd_dev);
	if (ret)
		goto err_out_probe;

	if (rbd_is_snap(rbd_dev) &&
	    (rbd_dev->header.features & RBD_FEATURE_OBJECT_MAP)) {
		ret = rbd_object_map_load(rbd_dev);
		if (ret)
			goto err_out_probe;
	}

	if (rbd_dev->header.features & RBD_FEATURE_LAYERING) {
		ret = rbd_dev_v2_parent_info(rbd_dev);
		if (ret)
			goto err_out_probe;
	}

	ret = rbd_dev_probe_parent(rbd_dev, depth);
	if (ret)
		goto err_out_probe;

	dout("discovered format %u image, header name is %s\n",
		rbd_dev->image_format, rbd_dev->header_oid.name);
	return 0;

err_out_probe:
	if (!depth)
		up_write(&rbd_dev->header_rwsem);
	if (need_watch)
		rbd_unregister_watch(rbd_dev);
	rbd_dev_unprobe(rbd_dev);
err_out_format:
	rbd_dev->image_format = 0;
	kfree(rbd_dev->spec->image_id);
	rbd_dev->spec->image_id = NULL;
	return ret;
}

static ssize_t do_rbd_add(struct bus_type *bus,
			  const char *buf,
			  size_t count)
{
	struct rbd_device *rbd_dev = NULL;
	struct ceph_options *ceph_opts = NULL;
	struct rbd_options *rbd_opts = NULL;
	struct rbd_spec *spec = NULL;
	struct rbd_client *rbdc;
	int rc;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

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

	/* parse add command */
	rc = rbd_add_parse_args(buf, &ceph_opts, &rbd_opts, &spec);
	if (rc < 0)
		goto out;

	rbdc = rbd_get_client(ceph_opts);
	if (IS_ERR(rbdc)) {
		rc = PTR_ERR(rbdc);
		goto err_out_args;
	}

	/* pick the pool */
	rc = ceph_pg_poolid_by_name(rbdc->client->osdc.osdmap, spec->pool_name);
	if (rc < 0) {
		if (rc == -ENOENT)
			pr_info("pool %s does not exist\n", spec->pool_name);
		goto err_out_client;
	}
	spec->pool_id = (u64)rc;

	rbd_dev = rbd_dev_create(rbdc, spec, rbd_opts);
	if (!rbd_dev) {
		rc = -ENOMEM;
		goto err_out_client;
	}
	rbdc = NULL;		/* rbd_dev now owns this */
	spec = NULL;		/* rbd_dev now owns this */
	rbd_opts = NULL;	/* rbd_dev now owns this */

	/* if we are mapping a snapshot it will be a read-only mapping */
	if (rbd_dev->opts->read_only ||
	    strcmp(rbd_dev->spec->snap_name, RBD_SNAP_HEAD_NAME))
		__set_bit(RBD_DEV_FLAG_READONLY, &rbd_dev->flags);

	rbd_dev->config_info = kstrdup(buf, GFP_KERNEL);
	if (!rbd_dev->config_info) {
		rc = -ENOMEM;
		goto err_out_rbd_dev;
	}

	rc = rbd_dev_image_probe(rbd_dev, 0);
	if (rc < 0)
		goto err_out_rbd_dev;

	if (rbd_dev->opts->alloc_size > rbd_dev->layout.object_size) {
		rbd_warn(rbd_dev, "alloc_size adjusted to %u",
			 rbd_dev->layout.object_size);
		rbd_dev->opts->alloc_size = rbd_dev->layout.object_size;
	}

	rc = rbd_dev_device_setup(rbd_dev);
	if (rc)
		goto err_out_image_probe;

	rc = rbd_add_acquire_lock(rbd_dev);
	if (rc)
		goto err_out_image_lock;

	/* Everything's ready.  Announce the disk to the world. */

	rc = device_add(&rbd_dev->dev);
	if (rc)
		goto err_out_image_lock;

	device_add_disk(&rbd_dev->dev, rbd_dev->disk, NULL);
	/* see rbd_init_disk() */
	blk_put_queue(rbd_dev->disk->queue);

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

	pr_info("%s: capacity %llu features 0x%llx\n", rbd_dev->disk->disk_name,
		(unsigned long long)get_capacity(rbd_dev->disk) << SECTOR_SHIFT,
		rbd_dev->header.features);
	rc = count;
out:
	module_put(THIS_MODULE);
	return rc;

err_out_image_lock:
	rbd_dev_image_unlock(rbd_dev);
	rbd_dev_device_release(rbd_dev);
err_out_image_probe:
	rbd_dev_image_release(rbd_dev);
err_out_rbd_dev:
	rbd_dev_destroy(rbd_dev);
err_out_client:
	rbd_put_client(rbdc);
err_out_args:
	rbd_spec_put(spec);
	kfree(rbd_opts);
	goto out;
}

static ssize_t add_store(struct bus_type *bus, const char *buf, size_t count)
{
	if (single_major)
		return -EINVAL;

	return do_rbd_add(bus, buf, count);
}

static ssize_t add_single_major_store(struct bus_type *bus, const char *buf,
				      size_t count)
{
	return do_rbd_add(bus, buf, count);
}

static void rbd_dev_remove_parent(struct rbd_device *rbd_dev)
{
	while (rbd_dev->parent) {
		struct rbd_device *first = rbd_dev;
		struct rbd_device *second = first->parent;
		struct rbd_device *third;

		/*
		 * Follow to the parent with no grandparent and
		 * remove it.
		 */
		while (second && (third = second->parent)) {
			first = second;
			second = third;
		}
		rbd_assert(second);
		rbd_dev_image_release(second);
		rbd_dev_destroy(second);
		first->parent = NULL;
		first->parent_overlap = 0;

		rbd_assert(first->parent_spec);
		rbd_spec_put(first->parent_spec);
		first->parent_spec = NULL;
	}
}

static ssize_t do_rbd_remove(struct bus_type *bus,
			     const char *buf,
			     size_t count)
{
	struct rbd_device *rbd_dev = NULL;
	struct list_head *tmp;
	int dev_id;
	char opt_buf[6];
	bool force = false;
	int ret;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	dev_id = -1;
	opt_buf[0] = '\0';
	sscanf(buf, "%d %5s", &dev_id, opt_buf);
	if (dev_id < 0) {
		pr_err("dev_id out of range\n");
		return -EINVAL;
	}
	if (opt_buf[0] != '\0') {
		if (!strcmp(opt_buf, "force")) {
			force = true;
		} else {
			pr_err("bad remove option at '%s'\n", opt_buf);
			return -EINVAL;
		}
	}

	ret = -ENOENT;
	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->dev_id == dev_id) {
			ret = 0;
			break;
		}
	}
	if (!ret) {
		spin_lock_irq(&rbd_dev->lock);
		if (rbd_dev->open_count && !force)
			ret = -EBUSY;
		else if (test_and_set_bit(RBD_DEV_FLAG_REMOVING,
					  &rbd_dev->flags))
			ret = -EINPROGRESS;
		spin_unlock_irq(&rbd_dev->lock);
	}
	spin_unlock(&rbd_dev_list_lock);
	if (ret)
		return ret;

	if (force) {
		/*
		 * Prevent new IO from being queued and wait for existing
		 * IO to complete/fail.
		 */
		blk_mq_freeze_queue(rbd_dev->disk->queue);
		blk_set_queue_dying(rbd_dev->disk->queue);
	}

	del_gendisk(rbd_dev->disk);
	spin_lock(&rbd_dev_list_lock);
	list_del_init(&rbd_dev->node);
	spin_unlock(&rbd_dev_list_lock);
	device_del(&rbd_dev->dev);

	rbd_dev_image_unlock(rbd_dev);
	rbd_dev_device_release(rbd_dev);
	rbd_dev_image_release(rbd_dev);
	rbd_dev_destroy(rbd_dev);
	return count;
}

static ssize_t remove_store(struct bus_type *bus, const char *buf, size_t count)
{
	if (single_major)
		return -EINVAL;

	return do_rbd_remove(bus, buf, count);
}

static ssize_t remove_single_major_store(struct bus_type *bus, const char *buf,
					 size_t count)
{
	return do_rbd_remove(bus, buf, count);
}

/*
 * create control files in sysfs
 * /sys/bus/rbd/...
 */
static int __init 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 __exit rbd_sysfs_cleanup(void)
{
	bus_unregister(&rbd_bus_type);
	device_unregister(&rbd_root_dev);
}

static int __init rbd_slab_init(void)
{
	rbd_assert(!rbd_img_request_cache);
	rbd_img_request_cache = KMEM_CACHE(rbd_img_request, 0);
	if (!rbd_img_request_cache)
		return -ENOMEM;

	rbd_assert(!rbd_obj_request_cache);
	rbd_obj_request_cache = KMEM_CACHE(rbd_obj_request, 0);
	if (!rbd_obj_request_cache)
		goto out_err;

	return 0;

out_err:
	kmem_cache_destroy(rbd_img_request_cache);
	rbd_img_request_cache = NULL;
	return -ENOMEM;
}

static void rbd_slab_exit(void)
{
	rbd_assert(rbd_obj_request_cache);
	kmem_cache_destroy(rbd_obj_request_cache);
	rbd_obj_request_cache = NULL;

	rbd_assert(rbd_img_request_cache);
	kmem_cache_destroy(rbd_img_request_cache);
	rbd_img_request_cache = NULL;
}

static int __init rbd_init(void)
{
	int rc;

	if (!libceph_compatible(NULL)) {
		rbd_warn(NULL, "libceph incompatibility (quitting)");
		return -EINVAL;
	}

	rc = rbd_slab_init();
	if (rc)
		return rc;

	/*
	 * The number of active work items is limited by the number of
	 * rbd devices * queue depth, so leave @max_active at default.
	 */
	rbd_wq = alloc_workqueue(RBD_DRV_NAME, WQ_MEM_RECLAIM, 0);
	if (!rbd_wq) {
		rc = -ENOMEM;
		goto err_out_slab;
	}

	if (single_major) {
		rbd_major = register_blkdev(0, RBD_DRV_NAME);
		if (rbd_major < 0) {
			rc = rbd_major;
			goto err_out_wq;
		}
	}

	rc = rbd_sysfs_init();
	if (rc)
		goto err_out_blkdev;

	if (single_major)
		pr_info("loaded (major %d)\n", rbd_major);
	else
		pr_info("loaded\n");

	return 0;

err_out_blkdev:
	if (single_major)
		unregister_blkdev(rbd_major, RBD_DRV_NAME);
err_out_wq:
	destroy_workqueue(rbd_wq);
err_out_slab:
	rbd_slab_exit();
	return rc;
}

static void __exit rbd_exit(void)
{
	ida_destroy(&rbd_dev_id_ida);
	rbd_sysfs_cleanup();
	if (single_major)
		unregister_blkdev(rbd_major, RBD_DRV_NAME);
	destroy_workqueue(rbd_wq);
	rbd_slab_exit();
}

module_init(rbd_init);
module_exit(rbd_exit);

MODULE_AUTHOR("Alex Elder <elder@inktank.com>");
MODULE_AUTHOR("Sage Weil <sage@newdream.net>");
MODULE_AUTHOR("Yehuda Sadeh <yehuda@hq.newdream.net>");
/* following authorship retained from original osdblk.c */
MODULE_AUTHOR("Jeff Garzik <jeff@garzik.org>");

MODULE_DESCRIPTION("RADOS Block Device (RBD) driver");
MODULE_LICENSE("GPL");