log0log.c 80.9 KB
Newer Older
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
/******************************************************
Database log

(c) 1995-1997 Innobase Oy

Created 12/9/1995 Heikki Tuuri
*******************************************************/

#include "log0log.h"

#ifdef UNIV_NONINL
#include "log0log.ic"
#endif

#include "mem0mem.h"
#include "buf0buf.h"
#include "buf0flu.h"
#include "srv0srv.h"
#include "log0recv.h"
#include "fil0fil.h"
#include "dict0boot.h"
#include "srv0srv.h"
#include "srv0start.h"
#include "trx0sys.h"
#include "trx0trx.h"

/*
General philosophy of InnoDB redo-logs:

1) Every change to a contents of a data page must be done
through mtr, which in mtr_commit() writes log records
to the InnoDB redo log.

2) Normally these changes are performed using a mlog_write_ulint()
or similar function.

3) In some page level operations only a code number of a
c-function and its parameters are written to the log to
reduce the size of the log.

  3a) You should not add parameters to these kind of functions
  (e.g. trx_undo_header_create(), trx_undo_insert_header_reuse())

  3b) You should not add such functionality which either change
  working when compared with the old or are dependent on data
  outside of the page. These kind of functions should implement
  self-contained page transformation and it should be unchanged
  if you don't have very essential reasons to change log
  semantics or format.

*/

/* Current free limit of space 0; protected by the log sys mutex; 0 means
uninitialized */
UNIV_INTERN ulint	log_fsp_current_free_limit		= 0;

/* Global log system variable */
UNIV_INTERN log_t*	log_sys	= NULL;

#ifdef UNIV_DEBUG
UNIV_INTERN ibool	log_do_write = TRUE;

UNIV_INTERN ibool	log_debug_writes = FALSE;
#endif /* UNIV_DEBUG */

/* These control how often we print warnings if the last checkpoint is too
old */
UNIV_INTERN ibool	log_has_printed_chkp_warning = FALSE;
UNIV_INTERN time_t	log_last_warning_time;

#ifdef UNIV_LOG_ARCHIVE
/* Pointer to this variable is used as the i/o-message when we do i/o to an
archive */
UNIV_INTERN byte	log_archive_io;
#endif /* UNIV_LOG_ARCHIVE */

/* A margin for free space in the log buffer before a log entry is catenated */
#define LOG_BUF_WRITE_MARGIN	(4 * OS_FILE_LOG_BLOCK_SIZE)

/* Margins for free space in the log buffer after a log entry is catenated */
#define LOG_BUF_FLUSH_RATIO	2
#define LOG_BUF_FLUSH_MARGIN	(LOG_BUF_WRITE_MARGIN + 4 * UNIV_PAGE_SIZE)

/* Margin for the free space in the smallest log group, before a new query
step which modifies the database, is started */

#define LOG_CHECKPOINT_FREE_PER_THREAD	(4 * UNIV_PAGE_SIZE)
#define LOG_CHECKPOINT_EXTRA_FREE	(8 * UNIV_PAGE_SIZE)

/* This parameter controls asynchronous making of a new checkpoint; the value
should be bigger than LOG_POOL_PREFLUSH_RATIO_SYNC */

#define LOG_POOL_CHECKPOINT_RATIO_ASYNC	32

/* This parameter controls synchronous preflushing of modified buffer pages */
#define LOG_POOL_PREFLUSH_RATIO_SYNC	16

/* The same ratio for asynchronous preflushing; this value should be less than
the previous */
#define LOG_POOL_PREFLUSH_RATIO_ASYNC	8

/* Extra margin, in addition to one log file, used in archiving */
#define LOG_ARCHIVE_EXTRA_MARGIN	(4 * UNIV_PAGE_SIZE)

/* This parameter controls asynchronous writing to the archive */
#define LOG_ARCHIVE_RATIO_ASYNC		16

/* Codes used in unlocking flush latches */
#define LOG_UNLOCK_NONE_FLUSHED_LOCK	1
#define LOG_UNLOCK_FLUSH_LOCK		2

/* States of an archiving operation */
#define	LOG_ARCHIVE_READ	1
#define	LOG_ARCHIVE_WRITE	2

/**********************************************************
Completes a checkpoint write i/o to a log file. */
static
void
log_io_complete_checkpoint(void);
/*============================*/
#ifdef UNIV_LOG_ARCHIVE
/**********************************************************
Completes an archiving i/o. */
static
void
log_io_complete_archive(void);
/*=========================*/
#endif /* UNIV_LOG_ARCHIVE */

/********************************************************************
Sets the global variable log_fsp_current_free_limit. Also makes a checkpoint,
so that we know that the limit has been written to a log checkpoint field
on disk. */
UNIV_INTERN
void
log_fsp_current_free_limit_set_and_checkpoint(
/*==========================================*/
	ulint	limit)	/* in: limit to set */
{
	ibool	success;

	mutex_enter(&(log_sys->mutex));

	log_fsp_current_free_limit = limit;

	mutex_exit(&(log_sys->mutex));

	/* Try to make a synchronous checkpoint */

	success = FALSE;

	while (!success) {
		success = log_checkpoint(TRUE, TRUE);
	}
}

/********************************************************************
Returns the oldest modified block lsn in the pool, or log_sys->lsn if none
exists. */
static
ib_uint64_t
log_buf_pool_get_oldest_modification(void)
/*======================================*/
{
	ib_uint64_t	lsn;

	ut_ad(mutex_own(&(log_sys->mutex)));

	lsn = buf_pool_get_oldest_modification();

	if (!lsn) {

		lsn = log_sys->lsn;
	}

	return(lsn);
}

/****************************************************************
Opens the log for log_write_low. The log must be closed with log_close and
released with log_release. */
UNIV_INTERN
ib_uint64_t
log_reserve_and_open(
/*=================*/
			/* out: start lsn of the log record */
	ulint	len)	/* in: length of data to be catenated */
{
	log_t*	log			= log_sys;
	ulint	len_upper_limit;
#ifdef UNIV_LOG_ARCHIVE
	ulint	archived_lsn_age;
	ulint	dummy;
#endif /* UNIV_LOG_ARCHIVE */
#ifdef UNIV_DEBUG
	ulint	count			= 0;
#endif /* UNIV_DEBUG */

	ut_a(len < log->buf_size / 2);
loop:
	mutex_enter(&(log->mutex));

	/* Calculate an upper limit for the space the string may take in the
	log buffer */

	len_upper_limit = LOG_BUF_WRITE_MARGIN + (5 * len) / 4;

	if (log->buf_free + len_upper_limit > log->buf_size) {

		mutex_exit(&(log->mutex));

		/* Not enough free space, do a syncronous flush of the log
		buffer */

		log_buffer_flush_to_disk();

		srv_log_waits++;

		ut_ad(++count < 50);

		goto loop;
	}

#ifdef UNIV_LOG_ARCHIVE
	if (log->archiving_state != LOG_ARCH_OFF) {

		archived_lsn_age = log->lsn - log->archived_lsn;
		if (archived_lsn_age + len_upper_limit
		    > log->max_archived_lsn_age) {
			/* Not enough free archived space in log groups: do a
			synchronous archive write batch: */

			mutex_exit(&(log->mutex));

			ut_ad(len_upper_limit <= log->max_archived_lsn_age);

			log_archive_do(TRUE, &dummy);

			ut_ad(++count < 50);

			goto loop;
		}
	}
#endif /* UNIV_LOG_ARCHIVE */

#ifdef UNIV_LOG_DEBUG
	log->old_buf_free = log->buf_free;
	log->old_lsn = log->lsn;
#endif
	return(log->lsn);
}

/****************************************************************
Writes to the log the string given. It is assumed that the caller holds the
log mutex. */
UNIV_INTERN
void
log_write_low(
/*==========*/
	byte*	str,		/* in: string */
	ulint	str_len)	/* in: string length */
{
	log_t*	log	= log_sys;
	ulint	len;
	ulint	data_len;
	byte*	log_block;

	ut_ad(mutex_own(&(log->mutex)));
part_loop:
	/* Calculate a part length */

	data_len = (log->buf_free % OS_FILE_LOG_BLOCK_SIZE) + str_len;

	if (data_len <= OS_FILE_LOG_BLOCK_SIZE - LOG_BLOCK_TRL_SIZE) {

		/* The string fits within the current log block */

		len = str_len;
	} else {
		data_len = OS_FILE_LOG_BLOCK_SIZE - LOG_BLOCK_TRL_SIZE;

		len = OS_FILE_LOG_BLOCK_SIZE
			- (log->buf_free % OS_FILE_LOG_BLOCK_SIZE)
			- LOG_BLOCK_TRL_SIZE;
	}

	ut_memcpy(log->buf + log->buf_free, str, len);

	str_len -= len;
	str = str + len;

	log_block = ut_align_down(log->buf + log->buf_free,
				  OS_FILE_LOG_BLOCK_SIZE);
	log_block_set_data_len(log_block, data_len);

	if (data_len == OS_FILE_LOG_BLOCK_SIZE - LOG_BLOCK_TRL_SIZE) {
		/* This block became full */
		log_block_set_data_len(log_block, OS_FILE_LOG_BLOCK_SIZE);
		log_block_set_checkpoint_no(log_block,
					    log_sys->next_checkpoint_no);
		len += LOG_BLOCK_HDR_SIZE + LOG_BLOCK_TRL_SIZE;

		log->lsn += len;

		/* Initialize the next block header */
		log_block_init(log_block + OS_FILE_LOG_BLOCK_SIZE, log->lsn);
	} else {
		log->lsn += len;
	}

	log->buf_free += len;

	ut_ad(log->buf_free <= log->buf_size);

	if (str_len > 0) {
		goto part_loop;
	}

	srv_log_write_requests++;
}

/****************************************************************
Closes the log. */
UNIV_INTERN
ib_uint64_t
log_close(void)
/*===========*/
			/* out: lsn */
{
	byte*		log_block;
	ulint		first_rec_group;
	ib_uint64_t	oldest_lsn;
	ib_uint64_t	lsn;
	log_t*		log	= log_sys;
	ib_uint64_t	checkpoint_age;

	ut_ad(mutex_own(&(log->mutex)));

	lsn = log->lsn;

	log_block = ut_align_down(log->buf + log->buf_free,
				  OS_FILE_LOG_BLOCK_SIZE);
	first_rec_group = log_block_get_first_rec_group(log_block);

	if (first_rec_group == 0) {
		/* We initialized a new log block which was not written
		full by the current mtr: the next mtr log record group
		will start within this block at the offset data_len */

		log_block_set_first_rec_group(
			log_block, log_block_get_data_len(log_block));
	}

	if (log->buf_free > log->max_buf_free) {

		log->check_flush_or_checkpoint = TRUE;
	}

	checkpoint_age = lsn - log->last_checkpoint_lsn;

	if (checkpoint_age >= log->log_group_capacity) {
		/* TODO: split btr_store_big_rec_extern_fields() into small
		steps so that we can release all latches in the middle, and
		call log_free_check() to ensure we never write over log written
		after the latest checkpoint. In principle, we should split all
		big_rec operations, but other operations are smaller. */

		if (!log_has_printed_chkp_warning
		    || difftime(time(NULL), log_last_warning_time) > 15) {

			log_has_printed_chkp_warning = TRUE;
			log_last_warning_time = time(NULL);

			ut_print_timestamp(stderr);
			fprintf(stderr,
				"  InnoDB: ERROR: the age of the last"
				" checkpoint is %lu,\n"
				"InnoDB: which exceeds the log group"
				" capacity %lu.\n"
				"InnoDB: If you are using big"
				" BLOB or TEXT rows, you must set the\n"
				"InnoDB: combined size of log files"
				" at least 10 times bigger than the\n"
				"InnoDB: largest such row.\n",
				(ulong) checkpoint_age,
				(ulong) log->log_group_capacity);
		}
	}

	if (checkpoint_age <= log->max_modified_age_async) {

		goto function_exit;
	}

	oldest_lsn = buf_pool_get_oldest_modification();

	if (!oldest_lsn
	    || lsn - oldest_lsn > log->max_modified_age_async
	    || checkpoint_age > log->max_checkpoint_age_async) {

		log->check_flush_or_checkpoint = TRUE;
	}
function_exit:

#ifdef UNIV_LOG_DEBUG
	log_check_log_recs(log->buf + log->old_buf_free,
			   log->buf_free - log->old_buf_free, log->old_lsn);
#endif

	return(lsn);
}

#ifdef UNIV_LOG_ARCHIVE
/**********************************************************
Pads the current log block full with dummy log records. Used in producing
consistent archived log files. */
static
void
log_pad_current_log_block(void)
/*===========================*/
{
	byte		b		= MLOG_DUMMY_RECORD;
	ulint		pad_length;
	ulint		i;
	ib_uint64_t	lsn;

	/* We retrieve lsn only because otherwise gcc crashed on HP-UX */
	lsn = log_reserve_and_open(OS_FILE_LOG_BLOCK_SIZE);

	pad_length = OS_FILE_LOG_BLOCK_SIZE
		- (log_sys->buf_free % OS_FILE_LOG_BLOCK_SIZE)
		- LOG_BLOCK_TRL_SIZE;

	for (i = 0; i < pad_length; i++) {
		log_write_low(&b, 1);
	}

	lsn = log_sys->lsn;

	log_close();
	log_release();

	ut_a(lsn % OS_FILE_LOG_BLOCK_SIZE == LOG_BLOCK_HDR_SIZE);
}
#endif /* UNIV_LOG_ARCHIVE */

/**********************************************************
Calculates the data capacity of a log group, when the log file headers are not
included. */
UNIV_INTERN
ulint
log_group_get_capacity(
/*===================*/
				/* out: capacity in bytes */
	log_group_t*	group)	/* in: log group */
{
	ut_ad(mutex_own(&(log_sys->mutex)));

	return((group->file_size - LOG_FILE_HDR_SIZE) * group->n_files);
}

/**********************************************************
Calculates the offset within a log group, when the log file headers are not
included. */
UNIV_INLINE
ulint
log_group_calc_size_offset(
/*=======================*/
				/* out: size offset (<= offset) */
	ulint		offset,	/* in: real offset within the log group */
	log_group_t*	group)	/* in: log group */
{
	ut_ad(mutex_own(&(log_sys->mutex)));

	return(offset - LOG_FILE_HDR_SIZE * (1 + offset / group->file_size));
}

/**********************************************************
Calculates the offset within a log group, when the log file headers are
included. */
UNIV_INLINE
ulint
log_group_calc_real_offset(
/*=======================*/
				/* out: real offset (>= offset) */
	ulint		offset,	/* in: size offset within the log group */
	log_group_t*	group)	/* in: log group */
{
	ut_ad(mutex_own(&(log_sys->mutex)));

	return(offset + LOG_FILE_HDR_SIZE
	       * (1 + offset / (group->file_size - LOG_FILE_HDR_SIZE)));
}

/**********************************************************
Calculates the offset of an lsn within a log group. */
static
ulint
log_group_calc_lsn_offset(
/*======================*/
				/* out: offset within the log group */
	ib_uint64_t	lsn,	/* in: lsn, must be within 4 GB of
				group->lsn */
	log_group_t*	group)	/* in: log group */
{
	ib_uint64_t	gr_lsn;
	ib_int64_t	gr_lsn_size_offset;
	ib_int64_t	difference;
	ib_int64_t	group_size;
	ib_int64_t	offset;

	ut_ad(mutex_own(&(log_sys->mutex)));

	/* If total log file size is > 2 GB we can easily get overflows
	with 32-bit integers. Use 64-bit integers instead. */

	gr_lsn = group->lsn;

	gr_lsn_size_offset = (ib_int64_t)
		log_group_calc_size_offset(group->lsn_offset, group);

	group_size = (ib_int64_t) log_group_get_capacity(group);

	if (lsn >= gr_lsn) {

		difference = (ib_int64_t) (lsn - gr_lsn);
	} else {
		difference = (ib_int64_t) (gr_lsn - lsn);

		difference = difference % group_size;

		difference = group_size - difference;
	}

	offset = (gr_lsn_size_offset + difference) % group_size;

	ut_a(offset < (((ib_int64_t) 1) << 32)); /* offset must be < 4 GB */

	/* fprintf(stderr,
	"Offset is %lu gr_lsn_offset is %lu difference is %lu\n",
	(ulint)offset,(ulint)gr_lsn_size_offset, (ulint)difference);
	*/

	return(log_group_calc_real_offset((ulint)offset, group));
}

/***********************************************************************
Calculates where in log files we find a specified lsn. */
UNIV_INTERN
ulint
log_calc_where_lsn_is(
/*==================*/
						/* out: log file number */
	ib_int64_t*	log_file_offset,	/* out: offset in that file
						(including the header) */
	ib_uint64_t	first_header_lsn,	/* in: first log file start
						lsn */
	ib_uint64_t	lsn,			/* in: lsn whose position to
						determine */
	ulint		n_log_files,		/* in: total number of log
						files */
	ib_int64_t	log_file_size)		/* in: log file size
						(including the header) */
{
	ib_int64_t	capacity	= log_file_size - LOG_FILE_HDR_SIZE;
	ulint		file_no;
	ib_int64_t	add_this_many;

	if (lsn < first_header_lsn) {
		add_this_many = 1 + (first_header_lsn - lsn)
			/ (capacity * (ib_int64_t)n_log_files);
		lsn += add_this_many
			* capacity * (ib_int64_t)n_log_files;
	}

	ut_a(lsn >= first_header_lsn);

	file_no = ((ulint)((lsn - first_header_lsn) / capacity))
		% n_log_files;
	*log_file_offset = (lsn - first_header_lsn) % capacity;

	*log_file_offset = *log_file_offset + LOG_FILE_HDR_SIZE;

	return(file_no);
}

/************************************************************
Sets the field values in group to correspond to a given lsn. For this function
to work, the values must already be correctly initialized to correspond to
some lsn, for instance, a checkpoint lsn. */
UNIV_INTERN
void
log_group_set_fields(
/*=================*/
	log_group_t*	group,	/* in: group */
	ib_uint64_t	lsn)	/* in: lsn for which the values should be
				set */
{
	group->lsn_offset = log_group_calc_lsn_offset(lsn, group);
	group->lsn = lsn;
}

/*********************************************************************
Calculates the recommended highest values for lsn - last_checkpoint_lsn,
lsn - buf_get_oldest_modification(), and lsn - max_archive_lsn_age. */
static
ibool
log_calc_max_ages(void)
/*===================*/
			/* out: error value FALSE if the smallest log group is
			too small to accommodate the number of OS threads in
			the database server */
{
	log_group_t*	group;
	ulint		margin;
	ulint		free;
	ibool		success		= TRUE;
	ulint		smallest_capacity;
	ulint		archive_margin;
	ulint		smallest_archive_margin;

	ut_ad(!mutex_own(&(log_sys->mutex)));

	mutex_enter(&(log_sys->mutex));

	group = UT_LIST_GET_FIRST(log_sys->log_groups);

	ut_ad(group);

	smallest_capacity = ULINT_MAX;
	smallest_archive_margin = ULINT_MAX;

	while (group) {
		if (log_group_get_capacity(group) < smallest_capacity) {

			smallest_capacity = log_group_get_capacity(group);
		}

		archive_margin = log_group_get_capacity(group)
			- (group->file_size - LOG_FILE_HDR_SIZE)
			- LOG_ARCHIVE_EXTRA_MARGIN;

		if (archive_margin < smallest_archive_margin) {

			smallest_archive_margin = archive_margin;
		}

		group = UT_LIST_GET_NEXT(log_groups, group);
	}

	/* Add extra safety */
	smallest_capacity = smallest_capacity - smallest_capacity / 10;

	/* For each OS thread we must reserve so much free space in the
	smallest log group that it can accommodate the log entries produced
	by single query steps: running out of free log space is a serious
	system error which requires rebooting the database. */

	free = LOG_CHECKPOINT_FREE_PER_THREAD * (10 + srv_thread_concurrency)
		+ LOG_CHECKPOINT_EXTRA_FREE;
	if (free >= smallest_capacity / 2) {
		success = FALSE;

		goto failure;
	} else {
		margin = smallest_capacity - free;
	}

	margin = ut_min(margin, log_sys->adm_checkpoint_interval);

	margin = margin - margin / 10;	/* Add still some extra safety */

	log_sys->log_group_capacity = smallest_capacity;

	log_sys->max_modified_age_async = margin
		- margin / LOG_POOL_PREFLUSH_RATIO_ASYNC;
	log_sys->max_modified_age_sync = margin
		- margin / LOG_POOL_PREFLUSH_RATIO_SYNC;

	log_sys->max_checkpoint_age_async = margin - margin
		/ LOG_POOL_CHECKPOINT_RATIO_ASYNC;
	log_sys->max_checkpoint_age = margin;

#ifdef UNIV_LOG_ARCHIVE
	log_sys->max_archived_lsn_age = smallest_archive_margin;

	log_sys->max_archived_lsn_age_async = smallest_archive_margin
		- smallest_archive_margin / LOG_ARCHIVE_RATIO_ASYNC;
#endif /* UNIV_LOG_ARCHIVE */
failure:
	mutex_exit(&(log_sys->mutex));

	if (!success) {
		fprintf(stderr,
			"InnoDB: Error: ib_logfiles are too small"
			" for innodb_thread_concurrency %lu.\n"
			"InnoDB: The combined size of ib_logfiles"
			" should be bigger than\n"
			"InnoDB: 200 kB * innodb_thread_concurrency.\n"
			"InnoDB: To get mysqld to start up, set"
			" innodb_thread_concurrency in my.cnf\n"
			"InnoDB: to a lower value, for example, to 8."
			" After an ERROR-FREE shutdown\n"
			"InnoDB: of mysqld you can adjust the size of"
			" ib_logfiles, as explained in\n"
			"InnoDB: http://dev.mysql.com/doc/refman/5.1/en/"
			"adding-and-removing.html\n"
			"InnoDB: Cannot continue operation."
			" Calling exit(1).\n",
			(ulong)srv_thread_concurrency);

		exit(1);
	}

	return(success);
}

/**********************************************************
Initializes the log. */
UNIV_INTERN
void
log_init(void)
/*==========*/
{
	byte*	buf;

	log_sys = mem_alloc(sizeof(log_t));

	mutex_create(&log_sys->mutex, SYNC_LOG);

	mutex_enter(&(log_sys->mutex));

	/* Start the lsn from one log block from zero: this way every
	log record has a start lsn != zero, a fact which we will use */

	log_sys->lsn = LOG_START_LSN;

	ut_a(LOG_BUFFER_SIZE >= 16 * OS_FILE_LOG_BLOCK_SIZE);
	ut_a(LOG_BUFFER_SIZE >= 4 * UNIV_PAGE_SIZE);

	buf = ut_malloc(LOG_BUFFER_SIZE + OS_FILE_LOG_BLOCK_SIZE);
	log_sys->buf = ut_align(buf, OS_FILE_LOG_BLOCK_SIZE);

	log_sys->buf_size = LOG_BUFFER_SIZE;

	memset(log_sys->buf, '\0', LOG_BUFFER_SIZE);

	log_sys->max_buf_free = log_sys->buf_size / LOG_BUF_FLUSH_RATIO
		- LOG_BUF_FLUSH_MARGIN;
	log_sys->check_flush_or_checkpoint = TRUE;
	UT_LIST_INIT(log_sys->log_groups);

	log_sys->n_log_ios = 0;

	log_sys->n_log_ios_old = log_sys->n_log_ios;
	log_sys->last_printout_time = time(NULL);
	/*----------------------------*/

	log_sys->buf_next_to_write = 0;

	log_sys->write_lsn = 0;
	log_sys->current_flush_lsn = 0;
	log_sys->flushed_to_disk_lsn = 0;

	log_sys->written_to_some_lsn = log_sys->lsn;
	log_sys->written_to_all_lsn = log_sys->lsn;

	log_sys->n_pending_writes = 0;

	log_sys->no_flush_event = os_event_create(NULL);

	os_event_set(log_sys->no_flush_event);

	log_sys->one_flushed_event = os_event_create(NULL);

	os_event_set(log_sys->one_flushed_event);

	/*----------------------------*/
	log_sys->adm_checkpoint_interval = ULINT_MAX;

	log_sys->next_checkpoint_no = 0;
	log_sys->last_checkpoint_lsn = log_sys->lsn;
	log_sys->n_pending_checkpoint_writes = 0;

	rw_lock_create(&log_sys->checkpoint_lock, SYNC_NO_ORDER_CHECK);

	log_sys->checkpoint_buf
		= ut_align(mem_alloc(2 * OS_FILE_LOG_BLOCK_SIZE),
			   OS_FILE_LOG_BLOCK_SIZE);
	memset(log_sys->checkpoint_buf, '\0', OS_FILE_LOG_BLOCK_SIZE);
	/*----------------------------*/

#ifdef UNIV_LOG_ARCHIVE
	/* Under MySQL, log archiving is always off */
	log_sys->archiving_state = LOG_ARCH_OFF;
	log_sys->archived_lsn = log_sys->lsn;
	log_sys->next_archived_lsn = 0;

	log_sys->n_pending_archive_ios = 0;

	rw_lock_create(&log_sys->archive_lock, SYNC_NO_ORDER_CHECK);

	log_sys->archive_buf = NULL;

	/* ut_align(
	ut_malloc(LOG_ARCHIVE_BUF_SIZE
	+ OS_FILE_LOG_BLOCK_SIZE),
	OS_FILE_LOG_BLOCK_SIZE); */
	log_sys->archive_buf_size = 0;

	/* memset(log_sys->archive_buf, '\0', LOG_ARCHIVE_BUF_SIZE); */

	log_sys->archiving_on = os_event_create(NULL);
#endif /* UNIV_LOG_ARCHIVE */

	/*----------------------------*/

	log_block_init(log_sys->buf, log_sys->lsn);
	log_block_set_first_rec_group(log_sys->buf, LOG_BLOCK_HDR_SIZE);

	log_sys->buf_free = LOG_BLOCK_HDR_SIZE;
	log_sys->lsn = LOG_START_LSN + LOG_BLOCK_HDR_SIZE;

	mutex_exit(&(log_sys->mutex));

#ifdef UNIV_LOG_DEBUG
	recv_sys_create();
	recv_sys_init(FALSE, buf_pool_get_curr_size());

	recv_sys->parse_start_lsn = log_sys->lsn;
	recv_sys->scanned_lsn = log_sys->lsn;
	recv_sys->scanned_checkpoint_no = 0;
	recv_sys->recovered_lsn = log_sys->lsn;
	recv_sys->limit_lsn = IB_ULONGLONG_MAX;
#endif
}

/**********************************************************************
Inits a log group to the log system. */
UNIV_INTERN
void
log_group_init(
/*===========*/
	ulint	id,			/* in: group id */
	ulint	n_files,		/* in: number of log files */
	ulint	file_size,		/* in: log file size in bytes */
	ulint	space_id,		/* in: space id of the file space
					which contains the log files of this
					group */
	ulint	archive_space_id __attribute__((unused)))
					/* in: space id of the file space
					which contains some archived log
					files for this group; currently, only
					for the first log group this is
					used */
{
	ulint	i;

	log_group_t*	group;

	group = mem_alloc(sizeof(log_group_t));

	group->id = id;
	group->n_files = n_files;
	group->file_size = file_size;
	group->space_id = space_id;
	group->state = LOG_GROUP_OK;
	group->lsn = LOG_START_LSN;
	group->lsn_offset = LOG_FILE_HDR_SIZE;
	group->n_pending_writes = 0;

	group->file_header_bufs = mem_alloc(sizeof(byte*) * n_files);
#ifdef UNIV_LOG_ARCHIVE
	group->archive_file_header_bufs = mem_alloc(sizeof(byte*) * n_files);
#endif /* UNIV_LOG_ARCHIVE */

	for (i = 0; i < n_files; i++) {
		*(group->file_header_bufs + i) = ut_align(
			mem_alloc(LOG_FILE_HDR_SIZE + OS_FILE_LOG_BLOCK_SIZE),
			OS_FILE_LOG_BLOCK_SIZE);

		memset(*(group->file_header_bufs + i), '\0',
		       LOG_FILE_HDR_SIZE);

#ifdef UNIV_LOG_ARCHIVE
		*(group->archive_file_header_bufs + i) = ut_align(
			mem_alloc(LOG_FILE_HDR_SIZE + OS_FILE_LOG_BLOCK_SIZE),
			OS_FILE_LOG_BLOCK_SIZE);
		memset(*(group->archive_file_header_bufs + i), '\0',
		       LOG_FILE_HDR_SIZE);
#endif /* UNIV_LOG_ARCHIVE */
	}

#ifdef UNIV_LOG_ARCHIVE
	group->archive_space_id = archive_space_id;

	group->archived_file_no = 0;
	group->archived_offset = 0;
#endif /* UNIV_LOG_ARCHIVE */

	group->checkpoint_buf = ut_align(
		mem_alloc(2 * OS_FILE_LOG_BLOCK_SIZE), OS_FILE_LOG_BLOCK_SIZE);

	memset(group->checkpoint_buf, '\0', OS_FILE_LOG_BLOCK_SIZE);

	UT_LIST_ADD_LAST(log_groups, log_sys->log_groups, group);

	ut_a(log_calc_max_ages());
}

/**********************************************************************
Does the unlockings needed in flush i/o completion. */
UNIV_INLINE
void
log_flush_do_unlocks(
/*=================*/
	ulint	code)	/* in: any ORed combination of LOG_UNLOCK_FLUSH_LOCK
			and LOG_UNLOCK_NONE_FLUSHED_LOCK */
{
	ut_ad(mutex_own(&(log_sys->mutex)));

	/* NOTE that we must own the log mutex when doing the setting of the
	events: this is because transactions will wait for these events to
	be set, and at that moment the log flush they were waiting for must
	have ended. If the log mutex were not reserved here, the i/o-thread
	calling this function might be preempted for a while, and when it
	resumed execution, it might be that a new flush had been started, and
	this function would erroneously signal the NEW flush as completed.
	Thus, the changes in the state of these events are performed
	atomically in conjunction with the changes in the state of
	log_sys->n_pending_writes etc. */

	if (code & LOG_UNLOCK_NONE_FLUSHED_LOCK) {
		os_event_set(log_sys->one_flushed_event);
	}

	if (code & LOG_UNLOCK_FLUSH_LOCK) {
		os_event_set(log_sys->no_flush_event);
	}
}

/**********************************************************************
Checks if a flush is completed for a log group and does the completion
routine if yes. */
UNIV_INLINE
ulint
log_group_check_flush_completion(
/*=============================*/
				/* out: LOG_UNLOCK_NONE_FLUSHED_LOCK or 0 */
	log_group_t*	group)	/* in: log group */
{
	ut_ad(mutex_own(&(log_sys->mutex)));

	if (!log_sys->one_flushed && group->n_pending_writes == 0) {
#ifdef UNIV_DEBUG
		if (log_debug_writes) {
			fprintf(stderr,
				"Log flushed first to group %lu\n",
				(ulong) group->id);
		}
#endif /* UNIV_DEBUG */
		log_sys->written_to_some_lsn = log_sys->write_lsn;
		log_sys->one_flushed = TRUE;

		return(LOG_UNLOCK_NONE_FLUSHED_LOCK);
	}

#ifdef UNIV_DEBUG
	if (log_debug_writes && (group->n_pending_writes == 0)) {

		fprintf(stderr, "Log flushed to group %lu\n",
			(ulong) group->id);
	}
#endif /* UNIV_DEBUG */
	return(0);
}

/**********************************************************
Checks if a flush is completed and does the completion routine if yes. */
static
ulint
log_sys_check_flush_completion(void)
/*================================*/
			/* out: LOG_UNLOCK_FLUSH_LOCK or 0 */
{
	ulint	move_start;
	ulint	move_end;

	ut_ad(mutex_own(&(log_sys->mutex)));

	if (log_sys->n_pending_writes == 0) {

		log_sys->written_to_all_lsn = log_sys->write_lsn;
		log_sys->buf_next_to_write = log_sys->write_end_offset;

		if (log_sys->write_end_offset > log_sys->max_buf_free / 2) {
			/* Move the log buffer content to the start of the
			buffer */

			move_start = ut_calc_align_down(
				log_sys->write_end_offset,
				OS_FILE_LOG_BLOCK_SIZE);
			move_end = ut_calc_align(log_sys->buf_free,
						 OS_FILE_LOG_BLOCK_SIZE);

			ut_memmove(log_sys->buf, log_sys->buf + move_start,
				   move_end - move_start);
			log_sys->buf_free -= move_start;

			log_sys->buf_next_to_write -= move_start;
		}

		return(LOG_UNLOCK_FLUSH_LOCK);
	}

	return(0);
}

/**********************************************************
Completes an i/o to a log file. */
UNIV_INTERN
void
log_io_complete(
/*============*/
	log_group_t*	group)	/* in: log group or a dummy pointer */
{
	ulint	unlock;

#ifdef UNIV_LOG_ARCHIVE
	if ((byte*)group == &log_archive_io) {
		/* It was an archive write */

		log_io_complete_archive();

		return;
	}
#endif /* UNIV_LOG_ARCHIVE */

	if ((ulint)group & 0x1UL) {
		/* It was a checkpoint write */
		group = (log_group_t*)((ulint)group - 1);

		if (srv_unix_file_flush_method != SRV_UNIX_O_DSYNC
		    && srv_unix_file_flush_method != SRV_UNIX_NOSYNC) {

			fil_flush(group->space_id);
		}

#ifdef UNIV_DEBUG
		if (log_debug_writes) {
			fprintf(stderr,
				"Checkpoint info written to group %lu\n",
				group->id);
		}
#endif /* UNIV_DEBUG */
		log_io_complete_checkpoint();

		return;
	}

	ut_error;	/* We currently use synchronous writing of the
			logs and cannot end up here! */

	if (srv_unix_file_flush_method != SRV_UNIX_O_DSYNC
	    && srv_unix_file_flush_method != SRV_UNIX_NOSYNC
	    && srv_flush_log_at_trx_commit != 2) {

		fil_flush(group->space_id);
	}

	mutex_enter(&(log_sys->mutex));

	ut_a(group->n_pending_writes > 0);
	ut_a(log_sys->n_pending_writes > 0);

	group->n_pending_writes--;
	log_sys->n_pending_writes--;

	unlock = log_group_check_flush_completion(group);
	unlock = unlock | log_sys_check_flush_completion();

	log_flush_do_unlocks(unlock);

	mutex_exit(&(log_sys->mutex));
}

/**********************************************************
Writes a log file header to a log file space. */
static
void
log_group_file_header_flush(
/*========================*/
	log_group_t*	group,		/* in: log group */
	ulint		nth_file,	/* in: header to the nth file in the
					log file space */
	ib_uint64_t	start_lsn)	/* in: log file data starts at this
					lsn */
{
	byte*	buf;
	ulint	dest_offset;

	ut_ad(mutex_own(&(log_sys->mutex)));
	ut_a(nth_file < group->n_files);

	buf = *(group->file_header_bufs + nth_file);

	mach_write_to_4(buf + LOG_GROUP_ID, group->id);
	mach_write_ull(buf + LOG_FILE_START_LSN, start_lsn);

	/* Wipe over possible label of ibbackup --restore */
	memcpy(buf + LOG_FILE_WAS_CREATED_BY_HOT_BACKUP, "    ", 4);

	dest_offset = nth_file * group->file_size;

#ifdef UNIV_DEBUG
	if (log_debug_writes) {
		fprintf(stderr,
			"Writing log file header to group %lu file %lu\n",
			(ulong) group->id, (ulong) nth_file);
	}
#endif /* UNIV_DEBUG */
	if (log_do_write) {
		log_sys->n_log_ios++;

		srv_os_log_pending_writes++;

		fil_io(OS_FILE_WRITE | OS_FILE_LOG, TRUE, group->space_id, 0,
		       dest_offset / UNIV_PAGE_SIZE,
		       dest_offset % UNIV_PAGE_SIZE,
		       OS_FILE_LOG_BLOCK_SIZE,
		       buf, group);

		srv_os_log_pending_writes--;
	}
}

/**********************************************************
Stores a 4-byte checksum to the trailer checksum field of a log block
before writing it to a log file. This checksum is used in recovery to
check the consistency of a log block. */
static
void
log_block_store_checksum(
/*=====================*/
	byte*	block)	/* in/out: pointer to a log block */
{
	log_block_set_checksum(block, log_block_calc_checksum(block));
}

/**********************************************************
Writes a buffer to a log file group. */
UNIV_INTERN
void
log_group_write_buf(
/*================*/
	log_group_t*	group,		/* in: log group */
	byte*		buf,		/* in: buffer */
	ulint		len,		/* in: buffer len; must be divisible
					by OS_FILE_LOG_BLOCK_SIZE */
	ib_uint64_t	start_lsn,	/* in: start lsn of the buffer; must
					be divisible by
					OS_FILE_LOG_BLOCK_SIZE */
	ulint		new_data_offset)/* in: start offset of new data in
					buf: this parameter is used to decide
					if we have to write a new log file
					header */
{
	ulint	write_len;
	ibool	write_header;
	ulint	next_offset;
	ulint	i;

	ut_ad(mutex_own(&(log_sys->mutex)));
	ut_a(len % OS_FILE_LOG_BLOCK_SIZE == 0);
	ut_a(((ulint) start_lsn) % OS_FILE_LOG_BLOCK_SIZE == 0);

	if (new_data_offset == 0) {
		write_header = TRUE;
	} else {
		write_header = FALSE;
	}
loop:
	if (len == 0) {

		return;
	}

	next_offset = log_group_calc_lsn_offset(start_lsn, group);

	if ((next_offset % group->file_size == LOG_FILE_HDR_SIZE)
	    && write_header) {
		/* We start to write a new log file instance in the group */

		log_group_file_header_flush(group,
					    next_offset / group->file_size,
					    start_lsn);
		srv_os_log_written+= OS_FILE_LOG_BLOCK_SIZE;
		srv_log_writes++;
	}

	if ((next_offset % group->file_size) + len > group->file_size) {

		write_len = group->file_size
			- (next_offset % group->file_size);
	} else {
		write_len = len;
	}

#ifdef UNIV_DEBUG
	if (log_debug_writes) {

		fprintf(stderr,
			"Writing log file segment to group %lu"
			" offset %lu len %lu\n"
			"start lsn %llu\n"
			"First block n:o %lu last block n:o %lu\n",
			(ulong) group->id, (ulong) next_offset,
			(ulong) write_len,
			start_lsn,
			(ulong) log_block_get_hdr_no(buf),
			(ulong) log_block_get_hdr_no(
				buf + write_len - OS_FILE_LOG_BLOCK_SIZE));
		ut_a(log_block_get_hdr_no(buf)
		     == log_block_convert_lsn_to_no(start_lsn));

		for (i = 0; i < write_len / OS_FILE_LOG_BLOCK_SIZE; i++) {

			ut_a(log_block_get_hdr_no(buf) + i
			     == log_block_get_hdr_no(
				     buf + i * OS_FILE_LOG_BLOCK_SIZE));
		}
	}
#endif /* UNIV_DEBUG */
	/* Calculate the checksums for each log block and write them to
	the trailer fields of the log blocks */

	for (i = 0; i < write_len / OS_FILE_LOG_BLOCK_SIZE; i++) {
		log_block_store_checksum(buf + i * OS_FILE_LOG_BLOCK_SIZE);
	}

	if (log_do_write) {
		log_sys->n_log_ios++;

		srv_os_log_pending_writes++;

		fil_io(OS_FILE_WRITE | OS_FILE_LOG, TRUE, group->space_id, 0,
		       next_offset / UNIV_PAGE_SIZE,
		       next_offset % UNIV_PAGE_SIZE, write_len, buf, group);

		srv_os_log_pending_writes--;

		srv_os_log_written+= write_len;
		srv_log_writes++;
	}

	if (write_len < len) {
		start_lsn += write_len;
		len -= write_len;
		buf += write_len;

		write_header = TRUE;

		goto loop;
	}
}

/**********************************************************
This function is called, e.g., when a transaction wants to commit. It checks
that the log has been written to the log file up to the last log entry written
by the transaction. If there is a flush running, it waits and checks if the
flush flushed enough. If not, starts a new flush. */
UNIV_INTERN
void
log_write_up_to(
/*============*/
	ib_uint64_t	lsn,	/* in: log sequence number up to which
				the log should be written,
				IB_ULONGLONG_MAX if not specified */
	ulint		wait,	/* in: LOG_NO_WAIT, LOG_WAIT_ONE_GROUP,
				or LOG_WAIT_ALL_GROUPS */
	ibool		flush_to_disk)
				/* in: TRUE if we want the written log
				also to be flushed to disk */
{
	log_group_t*	group;
	ulint		start_offset;
	ulint		end_offset;
	ulint		area_start;
	ulint		area_end;
#ifdef UNIV_DEBUG
	ulint		loop_count	= 0;
#endif /* UNIV_DEBUG */
	ulint		unlock;

	if (recv_no_ibuf_operations) {
		/* Recovery is running and no operations on the log files are
		allowed yet (the variable name .._no_ibuf_.. is misleading) */

		return;
	}

loop:
#ifdef UNIV_DEBUG
	loop_count++;

	ut_ad(loop_count < 5);

# if 0
	if (loop_count > 2) {
		fprintf(stderr, "Log loop count %lu\n", loop_count);
	}
# endif
#endif

	mutex_enter(&(log_sys->mutex));

	if (flush_to_disk
	    && log_sys->flushed_to_disk_lsn >= lsn) {

		mutex_exit(&(log_sys->mutex));

		return;
	}

	if (!flush_to_disk
	    && (log_sys->written_to_all_lsn >= lsn
		|| (log_sys->written_to_some_lsn >= lsn
		    && wait != LOG_WAIT_ALL_GROUPS))) {

		mutex_exit(&(log_sys->mutex));

		return;
	}

	if (log_sys->n_pending_writes > 0) {
		/* A write (+ possibly flush to disk) is running */

		if (flush_to_disk
		    && log_sys->current_flush_lsn >= lsn) {
			/* The write + flush will write enough: wait for it to
			complete  */

			goto do_waits;
		}

		if (!flush_to_disk
		    && log_sys->write_lsn >= lsn) {
			/* The write will write enough: wait for it to
			complete  */

			goto do_waits;
		}

		mutex_exit(&(log_sys->mutex));

		/* Wait for the write to complete and try to start a new
		write */

		os_event_wait(log_sys->no_flush_event);

		goto loop;
	}

	if (!flush_to_disk
	    && log_sys->buf_free == log_sys->buf_next_to_write) {
		/* Nothing to write and no flush to disk requested */

		mutex_exit(&(log_sys->mutex));

		return;
	}

#ifdef UNIV_DEBUG
	if (log_debug_writes) {
		fprintf(stderr,
			"Writing log from %llu up to lsn %llu\n",
			log_sys->written_to_all_lsn,
			log_sys->lsn);
	}
#endif /* UNIV_DEBUG */
	log_sys->n_pending_writes++;

	group = UT_LIST_GET_FIRST(log_sys->log_groups);
	group->n_pending_writes++;	/* We assume here that we have only
					one log group! */

	os_event_reset(log_sys->no_flush_event);
	os_event_reset(log_sys->one_flushed_event);

	start_offset = log_sys->buf_next_to_write;
	end_offset = log_sys->buf_free;

	area_start = ut_calc_align_down(start_offset, OS_FILE_LOG_BLOCK_SIZE);
	area_end = ut_calc_align(end_offset, OS_FILE_LOG_BLOCK_SIZE);

	ut_ad(area_end - area_start > 0);

	log_sys->write_lsn = log_sys->lsn;

	if (flush_to_disk) {
		log_sys->current_flush_lsn = log_sys->lsn;
	}

	log_sys->one_flushed = FALSE;

	log_block_set_flush_bit(log_sys->buf + area_start, TRUE);
	log_block_set_checkpoint_no(
		log_sys->buf + area_end - OS_FILE_LOG_BLOCK_SIZE,
		log_sys->next_checkpoint_no);

	/* Copy the last, incompletely written, log block a log block length
	up, so that when the flush operation writes from the log buffer, the
	segment to write will not be changed by writers to the log */

	ut_memcpy(log_sys->buf + area_end,
		  log_sys->buf + area_end - OS_FILE_LOG_BLOCK_SIZE,
		  OS_FILE_LOG_BLOCK_SIZE);

	log_sys->buf_free += OS_FILE_LOG_BLOCK_SIZE;
	log_sys->write_end_offset = log_sys->buf_free;

	group = UT_LIST_GET_FIRST(log_sys->log_groups);

	/* Do the write to the log files */

	while (group) {
		log_group_write_buf(
			group, log_sys->buf + area_start,
			area_end - area_start,
			ut_uint64_align_down(log_sys->written_to_all_lsn,
					     OS_FILE_LOG_BLOCK_SIZE),
			start_offset - area_start);

		log_group_set_fields(group, log_sys->write_lsn);

		group = UT_LIST_GET_NEXT(log_groups, group);
	}

	mutex_exit(&(log_sys->mutex));

	if (srv_unix_file_flush_method == SRV_UNIX_O_DSYNC) {
		/* O_DSYNC means the OS did not buffer the log file at all:
		so we have also flushed to disk what we have written */

		log_sys->flushed_to_disk_lsn = log_sys->write_lsn;

	} else if (flush_to_disk) {

		group = UT_LIST_GET_FIRST(log_sys->log_groups);

		fil_flush(group->space_id);
		log_sys->flushed_to_disk_lsn = log_sys->write_lsn;
	}

	mutex_enter(&(log_sys->mutex));

	group = UT_LIST_GET_FIRST(log_sys->log_groups);

	ut_a(group->n_pending_writes == 1);
	ut_a(log_sys->n_pending_writes == 1);

	group->n_pending_writes--;
	log_sys->n_pending_writes--;

	unlock = log_group_check_flush_completion(group);
	unlock = unlock | log_sys_check_flush_completion();

	log_flush_do_unlocks(unlock);

	mutex_exit(&(log_sys->mutex));

	return;

do_waits:
	mutex_exit(&(log_sys->mutex));

	switch (wait) {
	case LOG_WAIT_ONE_GROUP:
		os_event_wait(log_sys->one_flushed_event);
		break;
	case LOG_WAIT_ALL_GROUPS:
		os_event_wait(log_sys->no_flush_event);
		break;
#ifdef UNIV_DEBUG
	case LOG_NO_WAIT:
		break;
	default:
		ut_error;
#endif /* UNIV_DEBUG */
	}
}

/********************************************************************
Does a syncronous flush of the log buffer to disk. */
UNIV_INTERN
void
log_buffer_flush_to_disk(void)
/*==========================*/
{
	ib_uint64_t	lsn;

	mutex_enter(&(log_sys->mutex));

	lsn = log_sys->lsn;

	mutex_exit(&(log_sys->mutex));

	log_write_up_to(lsn, LOG_WAIT_ALL_GROUPS, TRUE);
}

/********************************************************************
Tries to establish a big enough margin of free space in the log buffer, such
that a new log entry can be catenated without an immediate need for a flush. */
static
void
log_flush_margin(void)
/*==================*/
{
	log_t*		log	= log_sys;
	ib_uint64_t	lsn	= 0;

	mutex_enter(&(log->mutex));

	if (log->buf_free > log->max_buf_free) {

		if (log->n_pending_writes > 0) {
			/* A flush is running: hope that it will provide enough
			free space */
		} else {
			lsn = log->lsn;
		}
	}

	mutex_exit(&(log->mutex));

	if (lsn) {
		log_write_up_to(lsn, LOG_NO_WAIT, FALSE);
	}
}

/********************************************************************
Advances the smallest lsn for which there are unflushed dirty blocks in the
buffer pool. NOTE: this function may only be called if the calling thread owns
no synchronization objects! */
UNIV_INTERN
ibool
log_preflush_pool_modified_pages(
/*=============================*/
					/* out: FALSE if there was a
					flush batch of the same type
					running, which means that we
					could not start this flush
					batch */
	ib_uint64_t	new_oldest,	/* in: try to advance
					oldest_modified_lsn at least
					to this lsn */
	ibool		sync)		/* in: TRUE if synchronous
					operation is desired */
{
	ulint	n_pages;

	if (recv_recovery_on) {
		/* If the recovery is running, we must first apply all
		log records to their respective file pages to get the
		right modify lsn values to these pages: otherwise, there
		might be pages on disk which are not yet recovered to the
		current lsn, and even after calling this function, we could
		not know how up-to-date the disk version of the database is,
		and we could not make a new checkpoint on the basis of the
		info on the buffer pool only. */

		recv_apply_hashed_log_recs(TRUE);
	}

	n_pages = buf_flush_batch(BUF_FLUSH_LIST, ULINT_MAX, new_oldest);

	if (sync) {
		buf_flush_wait_batch_end(BUF_FLUSH_LIST);
	}

	if (n_pages == ULINT_UNDEFINED) {

		return(FALSE);
	}

	return(TRUE);
}

/**********************************************************
Completes a checkpoint. */
static
void
log_complete_checkpoint(void)
/*=========================*/
{
	ut_ad(mutex_own(&(log_sys->mutex)));
	ut_ad(log_sys->n_pending_checkpoint_writes == 0);

	log_sys->next_checkpoint_no++;

	log_sys->last_checkpoint_lsn = log_sys->next_checkpoint_lsn;

	rw_lock_x_unlock_gen(&(log_sys->checkpoint_lock), LOG_CHECKPOINT);
}

/**********************************************************
Completes an asynchronous checkpoint info write i/o to a log file. */
static
void
log_io_complete_checkpoint(void)
/*============================*/
{
	mutex_enter(&(log_sys->mutex));

	ut_ad(log_sys->n_pending_checkpoint_writes > 0);

	log_sys->n_pending_checkpoint_writes--;

	if (log_sys->n_pending_checkpoint_writes == 0) {
		log_complete_checkpoint();
	}

	mutex_exit(&(log_sys->mutex));
}

/***********************************************************************
Writes info to a checkpoint about a log group. */
static
void
log_checkpoint_set_nth_group_info(
/*==============================*/
	byte*	buf,	/* in: buffer for checkpoint info */
	ulint	n,	/* in: nth slot */
	ulint	file_no,/* in: archived file number */
	ulint	offset)	/* in: archived file offset */
{
	ut_ad(n < LOG_MAX_N_GROUPS);

	mach_write_to_4(buf + LOG_CHECKPOINT_GROUP_ARRAY
			+ 8 * n + LOG_CHECKPOINT_ARCHIVED_FILE_NO, file_no);
	mach_write_to_4(buf + LOG_CHECKPOINT_GROUP_ARRAY
			+ 8 * n + LOG_CHECKPOINT_ARCHIVED_OFFSET, offset);
}

/***********************************************************************
Gets info from a checkpoint about a log group. */
UNIV_INTERN
void
log_checkpoint_get_nth_group_info(
/*==============================*/
	byte*	buf,	/* in: buffer containing checkpoint info */
	ulint	n,	/* in: nth slot */
	ulint*	file_no,/* out: archived file number */
	ulint*	offset)	/* out: archived file offset */
{
	ut_ad(n < LOG_MAX_N_GROUPS);

	*file_no = mach_read_from_4(buf + LOG_CHECKPOINT_GROUP_ARRAY
				    + 8 * n + LOG_CHECKPOINT_ARCHIVED_FILE_NO);
	*offset = mach_read_from_4(buf + LOG_CHECKPOINT_GROUP_ARRAY
				   + 8 * n + LOG_CHECKPOINT_ARCHIVED_OFFSET);
}

/**********************************************************
Writes the checkpoint info to a log group header. */
static
void
log_group_checkpoint(
/*=================*/
	log_group_t*	group)	/* in: log group */
{
	log_group_t*	group2;
#ifdef UNIV_LOG_ARCHIVE
	ib_uint64_t	archived_lsn;
	ib_uint64_t	next_archived_lsn;
#endif /* UNIV_LOG_ARCHIVE */
	ulint		write_offset;
	ulint		fold;
	byte*		buf;
	ulint		i;

	ut_ad(mutex_own(&(log_sys->mutex)));
#if LOG_CHECKPOINT_SIZE > OS_FILE_LOG_BLOCK_SIZE
# error "LOG_CHECKPOINT_SIZE > OS_FILE_LOG_BLOCK_SIZE"
#endif

	buf = group->checkpoint_buf;

	mach_write_ull(buf + LOG_CHECKPOINT_NO, log_sys->next_checkpoint_no);
	mach_write_ull(buf + LOG_CHECKPOINT_LSN, log_sys->next_checkpoint_lsn);

	mach_write_to_4(buf + LOG_CHECKPOINT_OFFSET,
			log_group_calc_lsn_offset(
				log_sys->next_checkpoint_lsn, group));

	mach_write_to_4(buf + LOG_CHECKPOINT_LOG_BUF_SIZE, log_sys->buf_size);

#ifdef UNIV_LOG_ARCHIVE
	if (log_sys->archiving_state == LOG_ARCH_OFF) {
		archived_lsn = IB_ULONGLONG_MAX;
	} else {
		archived_lsn = log_sys->archived_lsn;

		if (archived_lsn != log_sys->next_archived_lsn) {
			next_archived_lsn = log_sys->next_archived_lsn;
			/* For debugging only */
		}
	}

	mach_write_ull(buf + LOG_CHECKPOINT_ARCHIVED_LSN, archived_lsn);
#else /* UNIV_LOG_ARCHIVE */
	mach_write_ull(buf + LOG_CHECKPOINT_ARCHIVED_LSN, IB_ULONGLONG_MAX);
#endif /* UNIV_LOG_ARCHIVE */

	for (i = 0; i < LOG_MAX_N_GROUPS; i++) {
		log_checkpoint_set_nth_group_info(buf, i, 0, 0);
	}

	group2 = UT_LIST_GET_FIRST(log_sys->log_groups);

	while (group2) {
		log_checkpoint_set_nth_group_info(buf, group2->id,
#ifdef UNIV_LOG_ARCHIVE
						  group2->archived_file_no,
						  group2->archived_offset
#else /* UNIV_LOG_ARCHIVE */
						  0, 0
#endif /* UNIV_LOG_ARCHIVE */
						  );

		group2 = UT_LIST_GET_NEXT(log_groups, group2);
	}

	fold = ut_fold_binary(buf, LOG_CHECKPOINT_CHECKSUM_1);
	mach_write_to_4(buf + LOG_CHECKPOINT_CHECKSUM_1, fold);

	fold = ut_fold_binary(buf + LOG_CHECKPOINT_LSN,
			      LOG_CHECKPOINT_CHECKSUM_2 - LOG_CHECKPOINT_LSN);
	mach_write_to_4(buf + LOG_CHECKPOINT_CHECKSUM_2, fold);

	/* Starting from InnoDB-3.23.50, we also write info on allocated
	size in the tablespace */

	mach_write_to_4(buf + LOG_CHECKPOINT_FSP_FREE_LIMIT,
			log_fsp_current_free_limit);

	mach_write_to_4(buf + LOG_CHECKPOINT_FSP_MAGIC_N,
			LOG_CHECKPOINT_FSP_MAGIC_N_VAL);

	/* We alternate the physical place of the checkpoint info in the first
	log file */

	if ((log_sys->next_checkpoint_no & 1) == 0) {
		write_offset = LOG_CHECKPOINT_1;
	} else {
		write_offset = LOG_CHECKPOINT_2;
	}

	if (log_do_write) {
		if (log_sys->n_pending_checkpoint_writes == 0) {

			rw_lock_x_lock_gen(&(log_sys->checkpoint_lock),
					   LOG_CHECKPOINT);
		}

		log_sys->n_pending_checkpoint_writes++;

		log_sys->n_log_ios++;

		/* We send as the last parameter the group machine address
		added with 1, as we want to distinguish between a normal log
		file write and a checkpoint field write */

		fil_io(OS_FILE_WRITE | OS_FILE_LOG, FALSE, group->space_id, 0,
		       write_offset / UNIV_PAGE_SIZE,
		       write_offset % UNIV_PAGE_SIZE,
		       OS_FILE_LOG_BLOCK_SIZE,
		       buf, ((byte*)group + 1));

		ut_ad(((ulint)group & 0x1UL) == 0);
	}
}

/**********************************************************
Writes info to a buffer of a log group when log files are created in
backup restoration. */
UNIV_INTERN
void
log_reset_first_header_and_checkpoint(
/*==================================*/
	byte*		hdr_buf,/* in: buffer which will be written to the
				start of the first log file */
	ib_uint64_t	start)	/* in: lsn of the start of the first log file;
				we pretend that there is a checkpoint at
				start + LOG_BLOCK_HDR_SIZE */
{
	ulint		fold;
	byte*		buf;
	ib_uint64_t	lsn;

	mach_write_to_4(hdr_buf + LOG_GROUP_ID, 0);
	mach_write_ull(hdr_buf + LOG_FILE_START_LSN, start);

	lsn = start + LOG_BLOCK_HDR_SIZE;

	/* Write the label of ibbackup --restore */
	strcpy((char*) hdr_buf + LOG_FILE_WAS_CREATED_BY_HOT_BACKUP,
	       "ibbackup ");
	ut_sprintf_timestamp((char*) hdr_buf
			     + (LOG_FILE_WAS_CREATED_BY_HOT_BACKUP
				+ (sizeof "ibbackup ") - 1));
	buf = hdr_buf + LOG_CHECKPOINT_1;

	mach_write_ull(buf + LOG_CHECKPOINT_NO, 0);
	mach_write_ull(buf + LOG_CHECKPOINT_LSN, lsn);

	mach_write_to_4(buf + LOG_CHECKPOINT_OFFSET,
			LOG_FILE_HDR_SIZE + LOG_BLOCK_HDR_SIZE);

	mach_write_to_4(buf + LOG_CHECKPOINT_LOG_BUF_SIZE, 2 * 1024 * 1024);

	mach_write_ull(buf + LOG_CHECKPOINT_ARCHIVED_LSN, IB_ULONGLONG_MAX);

	fold = ut_fold_binary(buf, LOG_CHECKPOINT_CHECKSUM_1);
	mach_write_to_4(buf + LOG_CHECKPOINT_CHECKSUM_1, fold);

	fold = ut_fold_binary(buf + LOG_CHECKPOINT_LSN,
			      LOG_CHECKPOINT_CHECKSUM_2 - LOG_CHECKPOINT_LSN);
	mach_write_to_4(buf + LOG_CHECKPOINT_CHECKSUM_2, fold);

	/* Starting from InnoDB-3.23.50, we should also write info on
	allocated size in the tablespace, but unfortunately we do not
	know it here */
}

/**********************************************************
Reads a checkpoint info from a log group header to log_sys->checkpoint_buf. */
UNIV_INTERN
void
log_group_read_checkpoint_info(
/*===========================*/
	log_group_t*	group,	/* in: log group */
	ulint		field)	/* in: LOG_CHECKPOINT_1 or LOG_CHECKPOINT_2 */
{
	ut_ad(mutex_own(&(log_sys->mutex)));

	log_sys->n_log_ios++;

	fil_io(OS_FILE_READ | OS_FILE_LOG, TRUE, group->space_id, 0,
	       field / UNIV_PAGE_SIZE, field % UNIV_PAGE_SIZE,
	       OS_FILE_LOG_BLOCK_SIZE, log_sys->checkpoint_buf, NULL);
}

/**********************************************************
Writes checkpoint info to groups. */
UNIV_INTERN
void
log_groups_write_checkpoint_info(void)
/*==================================*/
{
	log_group_t*	group;

	ut_ad(mutex_own(&(log_sys->mutex)));

	group = UT_LIST_GET_FIRST(log_sys->log_groups);

	while (group) {
		log_group_checkpoint(group);

		group = UT_LIST_GET_NEXT(log_groups, group);
	}
}

/**********************************************************
Makes a checkpoint. Note that this function does not flush dirty
blocks from the buffer pool: it only checks what is lsn of the oldest
modification in the pool, and writes information about the lsn in
log files. Use log_make_checkpoint_at to flush also the pool. */
UNIV_INTERN
ibool
log_checkpoint(
/*===========*/
				/* out: TRUE if success, FALSE if a checkpoint
				write was already running */
	ibool	sync,		/* in: TRUE if synchronous operation is
				desired */
	ibool	write_always)	/* in: the function normally checks if the
				the new checkpoint would have a greater
				lsn than the previous one: if not, then no
				physical write is done; by setting this
				parameter TRUE, a physical write will always be
				made to log files */
{
	ib_uint64_t	oldest_lsn;

	if (recv_recovery_is_on()) {
		recv_apply_hashed_log_recs(TRUE);
	}

	if (srv_unix_file_flush_method != SRV_UNIX_NOSYNC) {
		fil_flush_file_spaces(FIL_TABLESPACE);
	}

	mutex_enter(&(log_sys->mutex));

	oldest_lsn = log_buf_pool_get_oldest_modification();

	mutex_exit(&(log_sys->mutex));

	/* Because log also contains headers and dummy log records,
	if the buffer pool contains no dirty buffers, oldest_lsn
	gets the value log_sys->lsn from the previous function,
	and we must make sure that the log is flushed up to that
	lsn. If there are dirty buffers in the buffer pool, then our
	write-ahead-logging algorithm ensures that the log has been flushed
	up to oldest_lsn. */

	log_write_up_to(oldest_lsn, LOG_WAIT_ALL_GROUPS, TRUE);

	mutex_enter(&(log_sys->mutex));

	if (!write_always
	    && log_sys->last_checkpoint_lsn >= oldest_lsn) {

		mutex_exit(&(log_sys->mutex));

		return(TRUE);
	}

	ut_ad(log_sys->written_to_all_lsn >= oldest_lsn);

	if (log_sys->n_pending_checkpoint_writes > 0) {
		/* A checkpoint write is running */

		mutex_exit(&(log_sys->mutex));

		if (sync) {
			/* Wait for the checkpoint write to complete */
			rw_lock_s_lock(&(log_sys->checkpoint_lock));
			rw_lock_s_unlock(&(log_sys->checkpoint_lock));
		}

		return(FALSE);
	}

	log_sys->next_checkpoint_lsn = oldest_lsn;

#ifdef UNIV_DEBUG
	if (log_debug_writes) {
		fprintf(stderr, "Making checkpoint no %lu at lsn %llu\n",
			(ulong) log_sys->next_checkpoint_no,
			oldest_lsn);
	}
#endif /* UNIV_DEBUG */

	log_groups_write_checkpoint_info();

	mutex_exit(&(log_sys->mutex));

	if (sync) {
		/* Wait for the checkpoint write to complete */
		rw_lock_s_lock(&(log_sys->checkpoint_lock));
		rw_lock_s_unlock(&(log_sys->checkpoint_lock));
	}

	return(TRUE);
}

/********************************************************************
Makes a checkpoint at a given lsn or later. */
UNIV_INTERN
void
log_make_checkpoint_at(
/*===================*/
	ib_uint64_t	lsn,		/* in: make a checkpoint at this or a
					later lsn, if IB_ULONGLONG_MAX, makes
					a checkpoint at the latest lsn */
	ibool		write_always)	/* in: the function normally checks if
					the the new checkpoint would have a
					greater lsn than the previous one: if
					not, then no physical write is done;
					by setting this parameter TRUE, a
					physical write will always be made to
					log files */
{
	/* Preflush pages synchronously */

	while (!log_preflush_pool_modified_pages(lsn, TRUE));

	while (!log_checkpoint(TRUE, write_always));
}

/********************************************************************
Tries to establish a big enough margin of free space in the log groups, such
that a new log entry can be catenated without an immediate need for a
checkpoint. NOTE: this function may only be called if the calling thread
owns no synchronization objects! */
static
void
log_checkpoint_margin(void)
/*=======================*/
{
	log_t*		log		= log_sys;
	ib_uint64_t	age;
	ib_uint64_t	checkpoint_age;
	ib_uint64_t	advance;
	ib_uint64_t	oldest_lsn;
	ibool		sync;
	ibool		checkpoint_sync;
	ibool		do_checkpoint;
	ibool		success;
loop:
	sync = FALSE;
	checkpoint_sync = FALSE;
	do_checkpoint = FALSE;

	mutex_enter(&(log->mutex));

	if (log->check_flush_or_checkpoint == FALSE) {
		mutex_exit(&(log->mutex));

		return;
	}

	oldest_lsn = log_buf_pool_get_oldest_modification();

	age = log->lsn - oldest_lsn;

	if (age > log->max_modified_age_sync) {

		/* A flush is urgent: we have to do a synchronous preflush */

		sync = TRUE;
		advance = 2 * (age - log->max_modified_age_sync);
	} else if (age > log->max_modified_age_async) {

		/* A flush is not urgent: we do an asynchronous preflush */
		advance = age - log->max_modified_age_async;
	} else {
		advance = 0;
	}

	checkpoint_age = log->lsn - log->last_checkpoint_lsn;

	if (checkpoint_age > log->max_checkpoint_age) {
		/* A checkpoint is urgent: we do it synchronously */

		checkpoint_sync = TRUE;

		do_checkpoint = TRUE;

	} else if (checkpoint_age > log->max_checkpoint_age_async) {
		/* A checkpoint is not urgent: do it asynchronously */

		do_checkpoint = TRUE;

		log->check_flush_or_checkpoint = FALSE;
	} else {
		log->check_flush_or_checkpoint = FALSE;
	}

	mutex_exit(&(log->mutex));

	if (advance) {
		ib_uint64_t	new_oldest = oldest_lsn + advance;

		success = log_preflush_pool_modified_pages(new_oldest, sync);

		/* If the flush succeeded, this thread has done its part
		and can proceed. If it did not succeed, there was another
		thread doing a flush at the same time. If sync was FALSE,
		the flush was not urgent, and we let this thread proceed.
		Otherwise, we let it start from the beginning again. */

		if (sync && !success) {
			mutex_enter(&(log->mutex));

			log->check_flush_or_checkpoint = TRUE;

			mutex_exit(&(log->mutex));
			goto loop;
		}
	}

	if (do_checkpoint) {
		log_checkpoint(checkpoint_sync, FALSE);

		if (checkpoint_sync) {

			goto loop;
		}
	}
}

/**********************************************************
Reads a specified log segment to a buffer. */
UNIV_INTERN
void
log_group_read_log_seg(
/*===================*/
	ulint		type,		/* in: LOG_ARCHIVE or LOG_RECOVER */
	byte*		buf,		/* in: buffer where to read */
	log_group_t*	group,		/* in: log group */
	ib_uint64_t	start_lsn,	/* in: read area start */
	ib_uint64_t	end_lsn)	/* in: read area end */
{
	ulint	len;
	ulint	source_offset;
	ibool	sync;

	ut_ad(mutex_own(&(log_sys->mutex)));

	sync = (type == LOG_RECOVER);
loop:
	source_offset = log_group_calc_lsn_offset(start_lsn, group);

	len = (ulint) (end_lsn - start_lsn);

	ut_ad(len != 0);

	if ((source_offset % group->file_size) + len > group->file_size) {

		len = group->file_size - (source_offset % group->file_size);
	}

#ifdef UNIV_LOG_ARCHIVE
	if (type == LOG_ARCHIVE) {

		log_sys->n_pending_archive_ios++;
	}
#endif /* UNIV_LOG_ARCHIVE */

	log_sys->n_log_ios++;

	fil_io(OS_FILE_READ | OS_FILE_LOG, sync, group->space_id, 0,
	       source_offset / UNIV_PAGE_SIZE, source_offset % UNIV_PAGE_SIZE,
	       len, buf, NULL);

	start_lsn += len;
	buf += len;

	if (start_lsn != end_lsn) {

		goto loop;
	}
}

#ifdef UNIV_LOG_ARCHIVE
/**********************************************************
Generates an archived log file name. */
UNIV_INTERN
void
log_archived_file_name_gen(
/*=======================*/
	char*	buf,	/* in: buffer where to write */
	ulint	id __attribute__((unused)),
			/* in: group id;
			currently we only archive the first group */
	ulint	file_no)/* in: file number */
{
	sprintf(buf, "%sib_arch_log_%010lu", srv_arch_dir, (ulong) file_no);
}

/**********************************************************
Writes a log file header to a log file space. */
static
void
log_group_archive_file_header_write(
/*================================*/
	log_group_t*	group,		/* in: log group */
	ulint		nth_file,	/* in: header to the nth file in the
					archive log file space */
	ulint		file_no,	/* in: archived file number */
	ib_uint64_t	start_lsn)	/* in: log file data starts at this
					lsn */
{
	byte*	buf;
	ulint	dest_offset;

	ut_ad(mutex_own(&(log_sys->mutex)));

	ut_a(nth_file < group->n_files);

	buf = *(group->archive_file_header_bufs + nth_file);

	mach_write_to_4(buf + LOG_GROUP_ID, group->id);
	mach_write_ull(buf + LOG_FILE_START_LSN, start_lsn);
	mach_write_to_4(buf + LOG_FILE_NO, file_no);

	mach_write_to_4(buf + LOG_FILE_ARCH_COMPLETED, FALSE);

	dest_offset = nth_file * group->file_size;

	log_sys->n_log_ios++;

	fil_io(OS_FILE_WRITE | OS_FILE_LOG, TRUE, group->archive_space_id,
	       dest_offset / UNIV_PAGE_SIZE,
	       dest_offset % UNIV_PAGE_SIZE,
	       2 * OS_FILE_LOG_BLOCK_SIZE,
	       buf, &log_archive_io);
}

/**********************************************************
Writes a log file header to a completed archived log file. */
static
void
log_group_archive_completed_header_write(
/*=====================================*/
	log_group_t*	group,		/* in: log group */
	ulint		nth_file,	/* in: header to the nth file in the
					archive log file space */
	ib_uint64_t	end_lsn)	/* in: end lsn of the file */
{
	byte*	buf;
	ulint	dest_offset;

	ut_ad(mutex_own(&(log_sys->mutex)));
	ut_a(nth_file < group->n_files);

	buf = *(group->archive_file_header_bufs + nth_file);

	mach_write_to_4(buf + LOG_FILE_ARCH_COMPLETED, TRUE);
	mach_write_ull(buf + LOG_FILE_END_LSN, end_lsn);

	dest_offset = nth_file * group->file_size + LOG_FILE_ARCH_COMPLETED;

	log_sys->n_log_ios++;

	fil_io(OS_FILE_WRITE | OS_FILE_LOG, TRUE, group->archive_space_id,
	       dest_offset / UNIV_PAGE_SIZE,
	       dest_offset % UNIV_PAGE_SIZE,
	       OS_FILE_LOG_BLOCK_SIZE,
	       buf + LOG_FILE_ARCH_COMPLETED,
	       &log_archive_io);
}

/**********************************************************
Does the archive writes for a single log group. */
static
void
log_group_archive(
/*==============*/
	log_group_t*	group)	/* in: log group */
{
	os_file_t	 file_handle;
	ib_uint64_t	start_lsn;
	ib_uint64_t	end_lsn;
	char		name[1024];
	byte*		buf;
	ulint		len;
	ibool		ret;
	ulint		next_offset;
	ulint		n_files;
	ulint		open_mode;

	ut_ad(mutex_own(&(log_sys->mutex)));

	start_lsn = log_sys->archived_lsn;

	ut_a(start_lsn % OS_FILE_LOG_BLOCK_SIZE == 0);

	end_lsn = log_sys->next_archived_lsn;

	ut_a(end_lsn % OS_FILE_LOG_BLOCK_SIZE == 0);

	buf = log_sys->archive_buf;

	n_files = 0;

	next_offset = group->archived_offset;
loop:
	if ((next_offset % group->file_size == 0)
	    || (fil_space_get_size(group->archive_space_id) == 0)) {

		/* Add the file to the archive file space; create or open the
		file */

		if (next_offset % group->file_size == 0) {
			open_mode = OS_FILE_CREATE;
		} else {
			open_mode = OS_FILE_OPEN;
		}

		log_archived_file_name_gen(name, group->id,
					   group->archived_file_no + n_files);

		file_handle = os_file_create(name, open_mode, OS_FILE_AIO,
					     OS_DATA_FILE, &ret);

		if (!ret && (open_mode == OS_FILE_CREATE)) {
			file_handle = os_file_create(
				name, OS_FILE_OPEN, OS_FILE_AIO,
				OS_DATA_FILE, &ret);
		}

		if (!ret) {
			fprintf(stderr,
				"InnoDB: Cannot create or open"
				" archive log file %s.\n"
				"InnoDB: Cannot continue operation.\n"
				"InnoDB: Check that the log archive"
				" directory exists,\n"
				"InnoDB: you have access rights to it, and\n"
				"InnoDB: there is space available.\n", name);
			exit(1);
		}

#ifdef UNIV_DEBUG
		if (log_debug_writes) {
			fprintf(stderr, "Created archive file %s\n", name);
		}
#endif /* UNIV_DEBUG */

		ret = os_file_close(file_handle);

		ut_a(ret);

		/* Add the archive file as a node to the space */

		fil_node_create(name, group->file_size / UNIV_PAGE_SIZE,
				group->archive_space_id, FALSE);

		if (next_offset % group->file_size == 0) {
			log_group_archive_file_header_write(
				group, n_files,
				group->archived_file_no + n_files,
				start_lsn);

			next_offset += LOG_FILE_HDR_SIZE;
		}
	}

	len = end_lsn - start_lsn;

	if (group->file_size < (next_offset % group->file_size) + len) {

		len = group->file_size - (next_offset % group->file_size);
	}

#ifdef UNIV_DEBUG
	if (log_debug_writes) {
		fprintf(stderr,
			"Archiving starting at lsn %llu, len %lu"
			" to group %lu\n",
			start_lsn,
			(ulong) len, (ulong) group->id);
	}
#endif /* UNIV_DEBUG */

	log_sys->n_pending_archive_ios++;

	log_sys->n_log_ios++;

	fil_io(OS_FILE_WRITE | OS_FILE_LOG, FALSE, group->archive_space_id,
	       next_offset / UNIV_PAGE_SIZE, next_offset % UNIV_PAGE_SIZE,
	       ut_calc_align(len, OS_FILE_LOG_BLOCK_SIZE), buf,
	       &log_archive_io);

	start_lsn += len;
	next_offset += len;
	buf += len;

	if (next_offset % group->file_size == 0) {
		n_files++;
	}

	if (end_lsn != start_lsn) {

		goto loop;
	}

	group->next_archived_file_no = group->archived_file_no + n_files;
	group->next_archived_offset = next_offset % group->file_size;

	ut_a(group->next_archived_offset % OS_FILE_LOG_BLOCK_SIZE == 0);
}

/*********************************************************
(Writes to the archive of each log group.) Currently, only the first
group is archived. */
static
void
log_archive_groups(void)
/*====================*/
{
	log_group_t*	group;

	ut_ad(mutex_own(&(log_sys->mutex)));

	group = UT_LIST_GET_FIRST(log_sys->log_groups);

	log_group_archive(group);
}

/*********************************************************
Completes the archiving write phase for (each log group), currently,
the first log group. */
static
void
log_archive_write_complete_groups(void)
/*===================================*/
{
	log_group_t*	group;
	ulint		end_offset;
	ulint		trunc_files;
	ulint		n_files;
	ib_uint64_t	start_lsn;
	ib_uint64_t	end_lsn;
	ulint		i;

	ut_ad(mutex_own(&(log_sys->mutex)));

	group = UT_LIST_GET_FIRST(log_sys->log_groups);

	group->archived_file_no = group->next_archived_file_no;
	group->archived_offset = group->next_archived_offset;

	/* Truncate from the archive file space all but the last
	file, or if it has been written full, all files */

	n_files = (UNIV_PAGE_SIZE
		   * fil_space_get_size(group->archive_space_id))
		/ group->file_size;
	ut_ad(n_files > 0);

	end_offset = group->archived_offset;

	if (end_offset % group->file_size == 0) {

		trunc_files = n_files;
	} else {
		trunc_files = n_files - 1;
	}

#ifdef UNIV_DEBUG
	if (log_debug_writes && trunc_files) {
		fprintf(stderr,
			"Complete file(s) archived to group %lu\n",
			(ulong) group->id);
	}
#endif /* UNIV_DEBUG */

	/* Calculate the archive file space start lsn */
	start_lsn = log_sys->next_archived_lsn
		- (end_offset - LOG_FILE_HDR_SIZE + trunc_files
		   * (group->file_size - LOG_FILE_HDR_SIZE));
	end_lsn = start_lsn;

	for (i = 0; i < trunc_files; i++) {

		end_lsn += group->file_size - LOG_FILE_HDR_SIZE;

		/* Write a notice to the headers of archived log
		files that the file write has been completed */

		log_group_archive_completed_header_write(group, i, end_lsn);
	}

	fil_space_truncate_start(group->archive_space_id,
				 trunc_files * group->file_size);

#ifdef UNIV_DEBUG
	if (log_debug_writes) {
		fputs("Archiving writes completed\n", stderr);
	}
#endif /* UNIV_DEBUG */
}

/**********************************************************
Completes an archiving i/o. */
static
void
log_archive_check_completion_low(void)
/*==================================*/
{
	ut_ad(mutex_own(&(log_sys->mutex)));

	if (log_sys->n_pending_archive_ios == 0
	    && log_sys->archiving_phase == LOG_ARCHIVE_READ) {

#ifdef UNIV_DEBUG
		if (log_debug_writes) {
			fputs("Archiving read completed\n", stderr);
		}
#endif /* UNIV_DEBUG */

		/* Archive buffer has now been read in: start archive writes */

		log_sys->archiving_phase = LOG_ARCHIVE_WRITE;

		log_archive_groups();
	}

	if (log_sys->n_pending_archive_ios == 0
	    && log_sys->archiving_phase == LOG_ARCHIVE_WRITE) {

		log_archive_write_complete_groups();

		log_sys->archived_lsn = log_sys->next_archived_lsn;

		rw_lock_x_unlock_gen(&(log_sys->archive_lock), LOG_ARCHIVE);
	}
}

/**********************************************************
Completes an archiving i/o. */
static
void
log_io_complete_archive(void)
/*=========================*/
{
	log_group_t*	group;

	mutex_enter(&(log_sys->mutex));

	group = UT_LIST_GET_FIRST(log_sys->log_groups);

	mutex_exit(&(log_sys->mutex));

	fil_flush(group->archive_space_id);

	mutex_enter(&(log_sys->mutex));

	ut_ad(log_sys->n_pending_archive_ios > 0);

	log_sys->n_pending_archive_ios--;

	log_archive_check_completion_low();

	mutex_exit(&(log_sys->mutex));
}

/************************************************************************
Starts an archiving operation. */
UNIV_INTERN
ibool
log_archive_do(
/*===========*/
			/* out: TRUE if succeed, FALSE if an archiving
			operation was already running */
	ibool	sync,	/* in: TRUE if synchronous operation is desired */
	ulint*	n_bytes)/* out: archive log buffer size, 0 if nothing to
			archive */
{
	ibool		calc_new_limit;
	ib_uint64_t	start_lsn;
	ib_uint64_t	limit_lsn;

	calc_new_limit = TRUE;
loop:
	mutex_enter(&(log_sys->mutex));

	switch (log_sys->archiving_state) {
	case LOG_ARCH_OFF:
arch_none:
		mutex_exit(&(log_sys->mutex));

		*n_bytes = 0;

		return(TRUE);
	case LOG_ARCH_STOPPED:
	case LOG_ARCH_STOPPING2:
		mutex_exit(&(log_sys->mutex));

		os_event_wait(log_sys->archiving_on);

		goto loop;
	}

	start_lsn = log_sys->archived_lsn;

	if (calc_new_limit) {
		ut_a(log_sys->archive_buf_size % OS_FILE_LOG_BLOCK_SIZE == 0);
		limit_lsn = start_lsn + log_sys->archive_buf_size;

		*n_bytes = log_sys->archive_buf_size;

		if (limit_lsn >= log_sys->lsn) {

			limit_lsn = ut_uint64_align_down(
				log_sys->lsn, OS_FILE_LOG_BLOCK_SIZE);
		}
	}

	if (log_sys->archived_lsn >= limit_lsn) {

		goto arch_none;
	}

	if (log_sys->written_to_all_lsn < limit_lsn) {

		mutex_exit(&(log_sys->mutex));

		log_write_up_to(limit_lsn, LOG_WAIT_ALL_GROUPS, TRUE);

		calc_new_limit = FALSE;

		goto loop;
	}

	if (log_sys->n_pending_archive_ios > 0) {
		/* An archiving operation is running */

		mutex_exit(&(log_sys->mutex));

		if (sync) {
			rw_lock_s_lock(&(log_sys->archive_lock));
			rw_lock_s_unlock(&(log_sys->archive_lock));
		}

		*n_bytes = log_sys->archive_buf_size;

		return(FALSE);
	}

	rw_lock_x_lock_gen(&(log_sys->archive_lock), LOG_ARCHIVE);

	log_sys->archiving_phase = LOG_ARCHIVE_READ;

	log_sys->next_archived_lsn = limit_lsn;

#ifdef UNIV_DEBUG
	if (log_debug_writes) {
		fprintf(stderr,
			"Archiving from lsn %llu to lsn %llu\n",
			log_sys->archived_lsn, limit_lsn);
	}
#endif /* UNIV_DEBUG */

	/* Read the log segment to the archive buffer */

	log_group_read_log_seg(LOG_ARCHIVE, log_sys->archive_buf,
			       UT_LIST_GET_FIRST(log_sys->log_groups),
			       start_lsn, limit_lsn);

	mutex_exit(&(log_sys->mutex));

	if (sync) {
		rw_lock_s_lock(&(log_sys->archive_lock));
		rw_lock_s_unlock(&(log_sys->archive_lock));
	}

	*n_bytes = log_sys->archive_buf_size;

	return(TRUE);
}

/********************************************************************
Writes the log contents to the archive at least up to the lsn when this
function was called. */
static
void
log_archive_all(void)
/*=================*/
{
	ib_uint64_t	present_lsn;
	ulint		dummy;

	mutex_enter(&(log_sys->mutex));

	if (log_sys->archiving_state == LOG_ARCH_OFF) {
		mutex_exit(&(log_sys->mutex));

		return;
	}

	present_lsn = log_sys->lsn;

	mutex_exit(&(log_sys->mutex));

	log_pad_current_log_block();

	for (;;) {
		mutex_enter(&(log_sys->mutex));

		if (present_lsn <= log_sys->archived_lsn) {

			mutex_exit(&(log_sys->mutex));

			return;
		}

		mutex_exit(&(log_sys->mutex));

		log_archive_do(TRUE, &dummy);
	}
}

/*********************************************************
Closes the possible open archive log file (for each group) the first group,
and if it was open, increments the group file count by 2, if desired. */
static
void
log_archive_close_groups(
/*=====================*/
	ibool	increment_file_count)	/* in: TRUE if we want to increment
					the file count */
{
	log_group_t*	group;
	ulint		trunc_len;

	ut_ad(mutex_own(&(log_sys->mutex)));

	if (log_sys->archiving_state == LOG_ARCH_OFF) {

		return;
	}

	group = UT_LIST_GET_FIRST(log_sys->log_groups);

	trunc_len = UNIV_PAGE_SIZE
		* fil_space_get_size(group->archive_space_id);
	if (trunc_len > 0) {
		ut_a(trunc_len == group->file_size);

		/* Write a notice to the headers of archived log
		files that the file write has been completed */

		log_group_archive_completed_header_write(
			group, 0, log_sys->archived_lsn);

		fil_space_truncate_start(group->archive_space_id,
					 trunc_len);
		if (increment_file_count) {
			group->archived_offset = 0;
			group->archived_file_no += 2;
		}

#ifdef UNIV_DEBUG
		if (log_debug_writes) {
			fprintf(stderr,
				"Incrementing arch file no to %lu"
				" in log group %lu\n",
				(ulong) group->archived_file_no + 2,
				(ulong) group->id);
		}
#endif /* UNIV_DEBUG */
	}
}

/********************************************************************
Writes the log contents to the archive up to the lsn when this function was
called, and stops the archiving. When archiving is started again, the archived
log file numbers start from 2 higher, so that the archiving will not write
again to the archived log files which exist when this function returns. */
UNIV_INTERN
ulint
log_archive_stop(void)
/*==================*/
			/* out: DB_SUCCESS or DB_ERROR */
{
	ibool	success;

	mutex_enter(&(log_sys->mutex));

	if (log_sys->archiving_state != LOG_ARCH_ON) {

		mutex_exit(&(log_sys->mutex));

		return(DB_ERROR);
	}

	log_sys->archiving_state = LOG_ARCH_STOPPING;

	mutex_exit(&(log_sys->mutex));

	log_archive_all();

	mutex_enter(&(log_sys->mutex));

	log_sys->archiving_state = LOG_ARCH_STOPPING2;
	os_event_reset(log_sys->archiving_on);

	mutex_exit(&(log_sys->mutex));

	/* Wait for a possible archiving operation to end */

	rw_lock_s_lock(&(log_sys->archive_lock));
	rw_lock_s_unlock(&(log_sys->archive_lock));

	mutex_enter(&(log_sys->mutex));

	/* Close all archived log files, incrementing the file count by 2,
	if appropriate */

	log_archive_close_groups(TRUE);

	mutex_exit(&(log_sys->mutex));

	/* Make a checkpoint, so that if recovery is needed, the file numbers
	of new archived log files will start from the right value */

	success = FALSE;

	while (!success) {
		success = log_checkpoint(TRUE, TRUE);
	}

	mutex_enter(&(log_sys->mutex));

	log_sys->archiving_state = LOG_ARCH_STOPPED;

	mutex_exit(&(log_sys->mutex));

	return(DB_SUCCESS);
}

/********************************************************************
Starts again archiving which has been stopped. */
UNIV_INTERN
ulint
log_archive_start(void)
/*===================*/
			/* out: DB_SUCCESS or DB_ERROR */
{
	mutex_enter(&(log_sys->mutex));

	if (log_sys->archiving_state != LOG_ARCH_STOPPED) {

		mutex_exit(&(log_sys->mutex));

		return(DB_ERROR);
	}

	log_sys->archiving_state = LOG_ARCH_ON;

	os_event_set(log_sys->archiving_on);

	mutex_exit(&(log_sys->mutex));

	return(DB_SUCCESS);
}

/********************************************************************
Stop archiving the log so that a gap may occur in the archived log files. */
UNIV_INTERN
ulint
log_archive_noarchivelog(void)
/*==========================*/
			/* out: DB_SUCCESS or DB_ERROR */
{
loop:
	mutex_enter(&(log_sys->mutex));

	if (log_sys->archiving_state == LOG_ARCH_STOPPED
	    || log_sys->archiving_state == LOG_ARCH_OFF) {

		log_sys->archiving_state = LOG_ARCH_OFF;

		os_event_set(log_sys->archiving_on);

		mutex_exit(&(log_sys->mutex));

		return(DB_SUCCESS);
	}

	mutex_exit(&(log_sys->mutex));

	log_archive_stop();

	os_thread_sleep(500000);

	goto loop;
}

/********************************************************************
Start archiving the log so that a gap may occur in the archived log files. */
UNIV_INTERN
ulint
log_archive_archivelog(void)
/*========================*/
			/* out: DB_SUCCESS or DB_ERROR */
{
	mutex_enter(&(log_sys->mutex));

	if (log_sys->archiving_state == LOG_ARCH_OFF) {

		log_sys->archiving_state = LOG_ARCH_ON;

		log_sys->archived_lsn
			= ut_uint64_align_down(log_sys->lsn,
					       OS_FILE_LOG_BLOCK_SIZE);
		mutex_exit(&(log_sys->mutex));

		return(DB_SUCCESS);
	}

	mutex_exit(&(log_sys->mutex));

	return(DB_ERROR);
}

/********************************************************************
Tries to establish a big enough margin of free space in the log groups, such
that a new log entry can be catenated without an immediate need for
archiving. */
static
void
log_archive_margin(void)
/*====================*/
{
	log_t*	log		= log_sys;
	ulint	age;
	ibool	sync;
	ulint	dummy;
loop:
	mutex_enter(&(log->mutex));

	if (log->archiving_state == LOG_ARCH_OFF) {
		mutex_exit(&(log->mutex));

		return;
	}

	age = log->lsn - log->archived_lsn;

	if (age > log->max_archived_lsn_age) {

		/* An archiving is urgent: we have to do synchronous i/o */

		sync = TRUE;

	} else if (age > log->max_archived_lsn_age_async) {

		/* An archiving is not urgent: we do asynchronous i/o */

		sync = FALSE;
	} else {
		/* No archiving required yet */

		mutex_exit(&(log->mutex));

		return;
	}

	mutex_exit(&(log->mutex));

	log_archive_do(sync, &dummy);

	if (sync == TRUE) {
		/* Check again that enough was written to the archive */

		goto loop;
	}
}
#endif /* UNIV_LOG_ARCHIVE */

/************************************************************************
Checks that there is enough free space in the log to start a new query step.
Flushes the log buffer or makes a new checkpoint if necessary. NOTE: this
function may only be called if the calling thread owns no synchronization
objects! */
UNIV_INTERN
void
log_check_margins(void)
/*===================*/
{
loop:
	log_flush_margin();

	log_checkpoint_margin();

#ifdef UNIV_LOG_ARCHIVE
	log_archive_margin();
#endif /* UNIV_LOG_ARCHIVE */

	mutex_enter(&(log_sys->mutex));

	if (log_sys->check_flush_or_checkpoint) {

		mutex_exit(&(log_sys->mutex));

		goto loop;
	}

	mutex_exit(&(log_sys->mutex));
}

/********************************************************************
Makes a checkpoint at the latest lsn and writes it to first page of each
data file in the database, so that we know that the file spaces contain
all modifications up to that lsn. This can only be called at database
shutdown. This function also writes all log in log files to the log archive. */
UNIV_INTERN
void
logs_empty_and_mark_files_at_shutdown(void)
/*=======================================*/
{
	ib_uint64_t	lsn;
	ulint		arch_log_no;

	if (srv_print_verbose_log) {
		ut_print_timestamp(stderr);
		fprintf(stderr, "  InnoDB: Starting shutdown...\n");
	}
	/* Wait until the master thread and all other operations are idle: our
	algorithm only works if the server is idle at shutdown */

	srv_shutdown_state = SRV_SHUTDOWN_CLEANUP;
loop:
	os_thread_sleep(100000);

	mutex_enter(&kernel_mutex);

	/* We need the monitor threads to stop before we proceed with a
	normal shutdown. In case of very fast shutdown, however, we can
	proceed without waiting for monitor threads. */

	if (srv_fast_shutdown < 2
	   && (srv_error_monitor_active
	      || srv_lock_timeout_and_monitor_active)) {

		mutex_exit(&kernel_mutex);

		goto loop;
	}

	/* Check that there are no longer transactions. We need this wait even
	for the 'very fast' shutdown, because the InnoDB layer may have
	committed or prepared transactions and we don't want to lose them. */

	if (trx_n_mysql_transactions > 0
	    || UT_LIST_GET_LEN(trx_sys->trx_list) > 0) {

		mutex_exit(&kernel_mutex);

		goto loop;
	}

	if (srv_fast_shutdown == 2) {
		/* In this fastest shutdown we do not flush the buffer pool:
		it is essentially a 'crash' of the InnoDB server. Make sure
		that the log is all flushed to disk, so that we can recover
		all committed transactions in a crash recovery. We must not
		write the lsn stamps to the data files, since at a startup
		InnoDB deduces from the stamps if the previous shutdown was
		clean. */

		log_buffer_flush_to_disk();

		return; /* We SKIP ALL THE REST !! */
	}

	/* Check that the master thread is suspended */

	if (srv_n_threads_active[SRV_MASTER] != 0) {

		mutex_exit(&kernel_mutex);

		goto loop;
	}

	mutex_exit(&kernel_mutex);

	mutex_enter(&(log_sys->mutex));

	if (log_sys->n_pending_checkpoint_writes
#ifdef UNIV_LOG_ARCHIVE
	    || log_sys->n_pending_archive_ios
#endif /* UNIV_LOG_ARCHIVE */
	    || log_sys->n_pending_writes) {

		mutex_exit(&(log_sys->mutex));

		goto loop;
	}

	mutex_exit(&(log_sys->mutex));

	if (!buf_pool_check_no_pending_io()) {

		goto loop;
	}

#ifdef UNIV_LOG_ARCHIVE
	log_archive_all();
#endif /* UNIV_LOG_ARCHIVE */

	log_make_checkpoint_at(IB_ULONGLONG_MAX, TRUE);

	mutex_enter(&(log_sys->mutex));

	lsn = log_sys->lsn;

	if (lsn != log_sys->last_checkpoint_lsn
#ifdef UNIV_LOG_ARCHIVE
	    || (srv_log_archive_on
		&& lsn != log_sys->archived_lsn + LOG_BLOCK_HDR_SIZE)
#endif /* UNIV_LOG_ARCHIVE */
	    ) {

		mutex_exit(&(log_sys->mutex));

		goto loop;
	}

	arch_log_no = 0;

#ifdef UNIV_LOG_ARCHIVE
	UT_LIST_GET_FIRST(log_sys->log_groups)->archived_file_no;

	if (0 == UT_LIST_GET_FIRST(log_sys->log_groups)->archived_offset) {

		arch_log_no--;
	}

	log_archive_close_groups(TRUE);
#endif /* UNIV_LOG_ARCHIVE */

	mutex_exit(&(log_sys->mutex));

	mutex_enter(&kernel_mutex);
	/* Check that the master thread has stayed suspended */
	if (srv_n_threads_active[SRV_MASTER] != 0) {
		fprintf(stderr,
			"InnoDB: Warning: the master thread woke up"
			" during shutdown\n");

		mutex_exit(&kernel_mutex);

		goto loop;
	}
	mutex_exit(&kernel_mutex);

	fil_flush_file_spaces(FIL_TABLESPACE);
	fil_flush_file_spaces(FIL_LOG);

	/* The call fil_write_flushed_lsn_to_data_files() will pass the buffer
	pool: therefore it is essential that the buffer pool has been
	completely flushed to disk! (We do not call fil_write... if the
	'very fast' shutdown is enabled.) */

	if (!buf_all_freed()) {

		goto loop;
	}

	srv_shutdown_state = SRV_SHUTDOWN_LAST_PHASE;

	/* Make some checks that the server really is quiet */
	ut_a(srv_n_threads_active[SRV_MASTER] == 0);
	ut_a(buf_all_freed());
	ut_a(lsn == log_sys->lsn);

	if (lsn < srv_start_lsn) {
		fprintf(stderr,
			"InnoDB: Error: log sequence number"
			" at shutdown %llu\n"
			"InnoDB: is lower than at startup %llu!\n",
			lsn, srv_start_lsn);
	}

	srv_shutdown_lsn = lsn;

	fil_write_flushed_lsn_to_data_files(lsn, arch_log_no);

	fil_flush_file_spaces(FIL_TABLESPACE);

	fil_close_all_files();

	/* Make some checks that the server really is quiet */
	ut_a(srv_n_threads_active[SRV_MASTER] == 0);
	ut_a(buf_all_freed());
	ut_a(lsn == log_sys->lsn);
}

/**********************************************************
Checks by parsing that the catenated log segment for a single mtr is
consistent. */
UNIV_INTERN
ibool
log_check_log_recs(
/*===============*/
	byte*		buf,		/* in: pointer to the start of
					the log segment in the
					log_sys->buf log buffer */
	ulint		len,		/* in: segment length in bytes */
	ib_uint64_t	buf_start_lsn)	/* in: buffer start lsn */
{
	ib_uint64_t	contiguous_lsn;
	ib_uint64_t	scanned_lsn;
	byte*		start;
	byte*		end;
	byte*		buf1;
	byte*		scan_buf;

	ut_ad(mutex_own(&(log_sys->mutex)));

	if (len == 0) {

		return(TRUE);
	}

	start = ut_align_down(buf, OS_FILE_LOG_BLOCK_SIZE);
	end = ut_align(buf + len, OS_FILE_LOG_BLOCK_SIZE);

	buf1 = mem_alloc((end - start) + OS_FILE_LOG_BLOCK_SIZE);
	scan_buf = ut_align(buf1, OS_FILE_LOG_BLOCK_SIZE);

	ut_memcpy(scan_buf, start, end - start);

	recv_scan_log_recs(TRUE,
			   (buf_pool->curr_size
			    - recv_n_pool_free_frames) * UNIV_PAGE_SIZE,
			   FALSE, scan_buf, end - start,
			   ut_uint64_align_down(buf_start_lsn,
						OS_FILE_LOG_BLOCK_SIZE),
			   &contiguous_lsn, &scanned_lsn);

	ut_a(scanned_lsn == buf_start_lsn + len);
	ut_a(recv_sys->recovered_lsn == scanned_lsn);

	mem_free(buf1);

	return(TRUE);
}

/**********************************************************
Peeks the current lsn. */
UNIV_INTERN
ibool
log_peek_lsn(
/*=========*/
				/* out: TRUE if success, FALSE if
				could not get the log system mutex */
	ib_uint64_t*	lsn)	/* out: if returns TRUE, current lsn is here */
{
	if (0 == mutex_enter_nowait(&(log_sys->mutex))) {
		*lsn = log_sys->lsn;

		mutex_exit(&(log_sys->mutex));

		return(TRUE);
	}

	return(FALSE);
}

/**********************************************************
Prints info of the log. */
UNIV_INTERN
void
log_print(
/*======*/
	FILE*	file)	/* in: file where to print */
{
	double	time_elapsed;
	time_t	current_time;

	mutex_enter(&(log_sys->mutex));

	fprintf(file,
		"Log sequence number %llu\n"
		"Log flushed up to   %llu\n"
		"Last checkpoint at  %llu\n",
		log_sys->lsn,
		log_sys->flushed_to_disk_lsn,
		log_sys->last_checkpoint_lsn);

3261 3262 3263 3264 3265 3266 3267 3268 3269
	fprintf(file,
		"Max checkpoint age  %lu\n"
		"Modified age        %lu\n"
		"Checkpoint age      %lu\n",
			(ulong) log_sys->max_checkpoint_age,
			(ulong) ut_dulint_minus(log_sys->lsn,
					log_buf_pool_get_oldest_modification()),
			(ulong) ut_dulint_minus(log_sys->lsn, log_sys->last_checkpoint_lsn));

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
	current_time = time(NULL);

	time_elapsed = 0.001 + difftime(current_time,
					log_sys->last_printout_time);
	fprintf(file,
		"%lu pending log writes, %lu pending chkp writes\n"
		"%lu log i/o's done, %.2f log i/o's/second\n",
		(ulong) log_sys->n_pending_writes,
		(ulong) log_sys->n_pending_checkpoint_writes,
		(ulong) log_sys->n_log_ios,
		((log_sys->n_log_ios - log_sys->n_log_ios_old)
		 / time_elapsed));

	log_sys->n_log_ios_old = log_sys->n_log_ios;
	log_sys->last_printout_time = current_time;

	mutex_exit(&(log_sys->mutex));
}

/**************************************************************************
Refreshes the statistics used to print per-second averages. */
UNIV_INTERN
void
log_refresh_stats(void)
/*===================*/
{
	log_sys->n_log_ios_old = log_sys->n_log_ios;
	log_sys->last_printout_time = time(NULL);
}