1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
2959
2960
2961
2962
2963
2964
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
3015
3016
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
3329
3330
3331
3332
3333
3334
3335
3336
3337
3338
3339
3340
3341
3342
3343
3344
3345
3346
3347
3348
3349
3350
3351
3352
3353
3354
3355
3356
3357
3358
3359
3360
3361
3362
3363
3364
3365
3366
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
3378
3379
3380
3381
3382
3383
3384
3385
3386
3387
3388
3389
3390
3391
3392
3393
3394
3395
3396
3397
3398
3399
3400
3401
3402
3403
3404
3405
3406
3407
3408
3409
3410
3411
3412
3413
3414
3415
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437
3438
3439
3440
3441
3442
3443
3444
3445
3446
3447
3448
3449
3450
3451
3452
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
3489
3490
3491
3492
3493
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
3522
3523
3524
3525
3526
3527
3528
3529
3530
3531
3532
3533
3534
3535
3536
3537
3538
3539
3540
3541
3542
3543
3544
3545
3546
3547
3548
3549
3550
3551
3552
3553
3554
3555
3556
3557
3558
3559
3560
3561
3562
3563
3564
3565
3566
3567
3568
3569
3570
3571
3572
3573
3574
3575
3576
3577
3578
3579
3580
3581
3582
3583
3584
3585
3586
3587
3588
3589
3590
3591
3592
3593
3594
3595
3596
3597
3598
3599
3600
3601
3602
3603
3604
3605
3606
3607
3608
3609
3610
3611
3612
3613
3614
3615
3616
3617
3618
3619
3620
3621
3622
3623
3624
3625
3626
3627
3628
3629
3630
3631
3632
3633
3634
3635
3636
3637
3638
3639
3640
3641
3642
3643
3644
3645
3646
3647
3648
3649
3650
3651
3652
3653
3654
3655
3656
3657
3658
3659
3660
3661
3662
3663
3664
3665
3666
3667
3668
3669
3670
3671
3672
3673
3674
3675
3676
3677
3678
3679
3680
3681
3682
3683
3684
3685
3686
3687
3688
3689
3690
3691
3692
3693
3694
3695
3696
3697
3698
3699
3700
3701
3702
3703
3704
3705
3706
3707
3708
3709
3710
3711
3712
3713
3714
3715
3716
3717
3718
3719
3720
3721
3722
3723
3724
3725
3726
3727
3728
3729
3730
3731
3732
3733
3734
3735
3736
3737
3738
3739
3740
3741
3742
3743
3744
3745
3746
3747
3748
3749
3750
3751
3752
3753
3754
3755
3756
3757
3758
3759
3760
3761
3762
3763
3764
3765
3766
3767
3768
3769
3770
3771
3772
3773
3774
3775
3776
3777
3778
3779
3780
3781
3782
3783
3784
3785
3786
3787
3788
3789
3790
3791
3792
3793
3794
3795
3796
3797
3798
3799
3800
3801
3802
3803
3804
3805
3806
3807
3808
3809
3810
3811
3812
3813
3814
3815
3816
3817
3818
3819
3820
3821
3822
3823
3824
3825
3826
3827
3828
3829
3830
3831
3832
3833
3834
3835
3836
3837
3838
3839
3840
3841
3842
3843
3844
3845
3846
3847
3848
3849
3850
3851
3852
3853
3854
3855
3856
3857
3858
3859
3860
3861
3862
3863
3864
3865
3866
3867
3868
3869
3870
3871
3872
3873
3874
3875
3876
3877
3878
3879
3880
3881
3882
3883
3884
3885
3886
3887
3888
3889
3890
3891
3892
3893
3894
3895
3896
3897
3898
3899
3900
3901
3902
3903
3904
3905
3906
3907
3908
3909
3910
3911
3912
3913
3914
3915
3916
3917
3918
3919
3920
3921
3922
3923
3924
3925
3926
3927
3928
3929
3930
3931
3932
3933
3934
3935
3936
3937
3938
3939
3940
3941
3942
3943
3944
3945
3946
3947
3948
3949
3950
3951
3952
3953
3954
3955
3956
3957
3958
3959
3960
3961
3962
3963
3964
3965
3966
3967
3968
3969
3970
3971
3972
3973
3974
3975
3976
3977
3978
3979
3980
3981
3982
3983
3984
3985
3986
3987
3988
3989
3990
3991
3992
3993
3994
3995
3996
3997
3998
3999
4000
4001
4002
4003
4004
4005
4006
4007
4008
4009
4010
4011
4012
4013
4014
4015
4016
4017
4018
4019
4020
4021
4022
4023
4024
4025
4026
4027
4028
4029
4030
4031
4032
4033
4034
4035
4036
4037
4038
4039
4040
4041
4042
4043
4044
4045
4046
4047
4048
4049
4050
4051
4052
4053
4054
4055
4056
4057
4058
4059
4060
4061
4062
4063
4064
4065
4066
4067
4068
4069
4070
4071
4072
4073
4074
4075
4076
4077
4078
4079
4080
4081
4082
4083
4084
4085
4086
4087
4088
4089
4090
4091
4092
4093
4094
4095
4096
4097
4098
4099
4100
4101
4102
4103
4104
4105
4106
4107
4108
4109
4110
4111
4112
4113
4114
4115
4116
4117
4118
4119
4120
4121
4122
4123
4124
4125
4126
4127
4128
4129
4130
4131
4132
4133
4134
4135
4136
4137
4138
4139
4140
4141
4142
4143
4144
4145
4146
4147
4148
4149
4150
4151
4152
4153
4154
4155
4156
4157
4158
4159
4160
4161
4162
4163
4164
4165
4166
4167
4168
4169
4170
4171
4172
4173
4174
4175
4176
4177
4178
4179
4180
4181
4182
4183
4184
4185
4186
4187
4188
4189
4190
4191
4192
4193
4194
4195
4196
4197
4198
4199
4200
4201
4202
4203
4204
4205
4206
4207
4208
4209
4210
4211
4212
4213
4214
4215
4216
4217
4218
4219
4220
4221
4222
4223
4224
4225
4226
4227
4228
4229
4230
4231
4232
4233
4234
4235
4236
4237
4238
4239
4240
4241
4242
4243
4244
4245
4246
4247
4248
4249
4250
4251
4252
4253
4254
4255
4256
4257
4258
4259
4260
4261
4262
4263
4264
4265
4266
4267
4268
4269
4270
4271
4272
4273
4274
4275
4276
4277
4278
4279
4280
4281
4282
4283
4284
4285
4286
4287
4288
4289
4290
4291
4292
4293
4294
4295
4296
4297
4298
4299
4300
4301
4302
4303
4304
4305
4306
4307
4308
4309
4310
4311
4312
4313
4314
4315
4316
4317
4318
4319
4320
4321
4322
4323
4324
4325
4326
4327
4328
4329
4330
4331
4332
4333
4334
4335
4336
4337
4338
4339
4340
4341
4342
4343
4344
4345
4346
4347
4348
4349
4350
4351
4352
4353
4354
4355
4356
4357
4358
4359
4360
4361
4362
4363
4364
4365
4366
4367
4368
4369
4370
4371
4372
4373
4374
4375
4376
4377
4378
4379
4380
4381
4382
4383
4384
4385
4386
4387
4388
4389
4390
4391
4392
4393
4394
4395
4396
4397
4398
4399
4400
4401
4402
4403
4404
4405
4406
4407
4408
4409
4410
4411
4412
4413
4414
4415
4416
4417
4418
4419
4420
4421
4422
4423
4424
4425
4426
4427
4428
4429
4430
4431
4432
4433
4434
4435
4436
4437
4438
4439
4440
4441
4442
4443
4444
4445
4446
4447
4448
4449
4450
4451
4452
4453
4454
4455
4456
4457
4458
4459
4460
4461
4462
4463
4464
4465
4466
4467
4468
4469
4470
4471
4472
4473
4474
4475
4476
4477
4478
4479
4480
4481
4482
4483
4484
4485
4486
4487
4488
4489
4490
4491
4492
4493
4494
4495
4496
4497
4498
4499
4500
4501
4502
4503
4504
4505
4506
4507
4508
4509
4510
4511
4512
4513
4514
4515
4516
4517
4518
4519
4520
4521
4522
4523
4524
4525
4526
4527
4528
4529
4530
4531
4532
4533
4534
4535
4536
4537
4538
4539
4540
4541
4542
4543
4544
4545
4546
4547
4548
4549
4550
4551
4552
4553
4554
4555
4556
4557
4558
4559
4560
4561
4562
4563
4564
4565
4566
4567
4568
4569
4570
4571
4572
4573
4574
4575
4576
4577
4578
4579
4580
4581
4582
4583
4584
4585
4586
4587
4588
4589
4590
4591
4592
4593
4594
4595
4596
4597
4598
4599
4600
4601
4602
4603
4604
4605
4606
4607
4608
4609
4610
4611
4612
4613
4614
4615
4616
4617
4618
4619
4620
4621
4622
4623
4624
4625
4626
4627
4628
4629
4630
4631
4632
4633
4634
4635
4636
4637
4638
4639
4640
4641
4642
4643
4644
4645
4646
4647
4648
4649
4650
4651
4652
4653
4654
4655
4656
4657
4658
4659
4660
4661
4662
4663
4664
4665
4666
4667
4668
4669
4670
4671
4672
4673
4674
4675
4676
4677
4678
4679
4680
4681
4682
4683
4684
4685
4686
4687
4688
4689
4690
4691
4692
4693
4694
4695
4696
4697
4698
4699
4700
4701
4702
4703
4704
4705
4706
4707
4708
4709
4710
4711
4712
4713
4714
4715
4716
4717
4718
4719
4720
4721
4722
4723
4724
4725
4726
4727
4728
4729
4730
4731
4732
4733
4734
4735
4736
4737
4738
4739
4740
4741
4742
4743
4744
4745
4746
4747
4748
4749
4750
4751
4752
4753
4754
4755
4756
4757
4758
4759
4760
4761
4762
4763
4764
4765
4766
4767
4768
4769
4770
4771
4772
4773
4774
4775
4776
4777
4778
4779
4780
4781
4782
4783
4784
4785
4786
4787
4788
4789
4790
4791
4792
4793
4794
4795
4796
4797
4798
4799
4800
4801
4802
4803
4804
4805
4806
4807
4808
4809
4810
4811
4812
4813
4814
4815
4816
4817
4818
4819
4820
4821
4822
4823
4824
4825
4826
4827
4828
4829
4830
4831
4832
4833
4834
4835
4836
4837
4838
4839
4840
4841
4842
4843
4844
4845
4846
4847
4848
4849
4850
4851
4852
4853
4854
4855
4856
4857
4858
4859
4860
4861
4862
4863
4864
4865
4866
4867
4868
4869
4870
4871
4872
4873
4874
4875
4876
4877
4878
4879
4880
4881
4882
4883
4884
4885
4886
4887
4888
4889
4890
4891
4892
4893
4894
4895
4896
4897
4898
4899
4900
4901
4902
4903
4904
4905
4906
4907
4908
4909
4910
4911
4912
4913
4914
4915
4916
4917
4918
4919
4920
4921
4922
4923
4924
4925
4926
4927
4928
4929
4930
4931
4932
4933
4934
4935
4936
4937
4938
4939
4940
4941
4942
4943
4944
4945
4946
4947
4948
4949
4950
4951
4952
4953
4954
4955
4956
4957
4958
4959
4960
4961
4962
4963
4964
4965
4966
4967
4968
4969
4970
4971
4972
4973
4974
4975
4976
4977
4978
4979
4980
4981
4982
4983
4984
4985
4986
4987
4988
4989
4990
4991
4992
4993
4994
4995
4996
4997
4998
4999
5000
5001
5002
5003
5004
5005
5006
5007
5008
5009
5010
5011
5012
5013
5014
5015
5016
5017
5018
5019
5020
5021
5022
5023
5024
5025
5026
5027
5028
5029
5030
5031
5032
5033
5034
5035
5036
5037
5038
5039
5040
5041
5042
5043
5044
5045
5046
5047
5048
5049
5050
5051
5052
5053
5054
5055
5056
5057
5058
5059
5060
5061
5062
5063
5064
5065
5066
5067
5068
5069
5070
5071
5072
5073
5074
5075
5076
5077
5078
5079
5080
5081
5082
5083
5084
5085
5086
5087
5088
5089
5090
5091
5092
5093
5094
5095
5096
5097
5098
5099
5100
5101
5102
5103
5104
5105
5106
5107
5108
5109
5110
5111
5112
5113
5114
5115
5116
5117
5118
5119
5120
5121
5122
5123
5124
5125
5126
5127
5128
5129
5130
5131
5132
5133
5134
5135
5136
5137
5138
5139
5140
5141
5142
5143
5144
5145
5146
5147
5148
5149
5150
5151
5152
5153
5154
5155
5156
5157
5158
5159
5160
5161
5162
5163
5164
5165
5166
5167
5168
5169
5170
5171
5172
5173
5174
5175
5176
5177
5178
5179
5180
5181
5182
5183
5184
5185
5186
5187
5188
5189
5190
5191
5192
5193
5194
5195
5196
5197
5198
5199
5200
5201
5202
5203
5204
5205
5206
5207
5208
5209
5210
5211
5212
5213
5214
5215
5216
5217
5218
5219
5220
5221
5222
5223
5224
5225
5226
5227
5228
5229
5230
5231
5232
5233
5234
5235
5236
5237
5238
5239
5240
5241
5242
5243
5244
5245
5246
5247
5248
5249
5250
5251
5252
5253
5254
5255
5256
5257
5258
5259
5260
5261
5262
5263
5264
5265
5266
5267
5268
5269
5270
5271
5272
5273
5274
5275
5276
5277
5278
5279
5280
5281
5282
5283
5284
5285
5286
5287
5288
5289
5290
5291
5292
5293
5294
5295
5296
5297
5298
5299
5300
5301
5302
5303
5304
5305
5306
5307
5308
5309
5310
5311
5312
5313
5314
5315
5316
5317
5318
5319
5320
5321
5322
5323
5324
5325
5326
5327
5328
5329
5330
5331
5332
5333
5334
5335
5336
5337
5338
5339
5340
5341
5342
5343
5344
5345
5346
5347
5348
5349
5350
5351
5352
5353
5354
5355
5356
5357
5358
5359
5360
5361
5362
5363
5364
5365
5366
5367
5368
5369
5370
5371
5372
5373
5374
5375
5376
5377
5378
5379
5380
5381
5382
5383
5384
5385
5386
5387
5388
5389
5390
5391
5392
5393
5394
5395
5396
5397
5398
5399
5400
5401
5402
5403
5404
5405
5406
5407
5408
5409
5410
5411
5412
5413
5414
5415
5416
5417
5418
5419
5420
5421
5422
5423
5424
5425
5426
5427
5428
5429
5430
5431
5432
5433
5434
5435
5436
5437
5438
5439
5440
5441
5442
5443
5444
5445
5446
5447
5448
5449
5450
5451
5452
5453
5454
5455
5456
5457
5458
5459
5460
5461
5462
5463
5464
5465
5466
5467
5468
5469
5470
5471
5472
5473
5474
5475
5476
5477
5478
5479
5480
5481
5482
5483
5484
5485
5486
5487
5488
5489
5490
5491
5492
5493
5494
5495
5496
5497
5498
5499
5500
5501
5502
5503
5504
5505
5506
5507
5508
5509
5510
5511
5512
5513
5514
5515
5516
5517
5518
5519
5520
5521
5522
5523
5524
5525
5526
5527
5528
5529
5530
5531
5532
5533
5534
5535
5536
5537
5538
5539
5540
5541
5542
5543
5544
5545
5546
5547
5548
5549
5550
5551
5552
5553
5554
5555
5556
5557
5558
5559
5560
5561
5562
5563
5564
5565
5566
5567
5568
5569
5570
5571
5572
5573
5574
5575
5576
5577
5578
5579
5580
5581
5582
5583
5584
5585
5586
5587
5588
5589
5590
5591
5592
5593
5594
5595
5596
5597
5598
5599
5600
5601
5602
5603
5604
5605
5606
5607
5608
5609
5610
5611
5612
5613
5614
5615
5616
5617
5618
5619
5620
5621
5622
5623
5624
5625
5626
5627
5628
5629
5630
5631
5632
5633
5634
5635
5636
5637
5638
5639
5640
5641
5642
5643
5644
5645
5646
5647
5648
5649
5650
5651
5652
5653
5654
5655
5656
5657
5658
5659
5660
5661
5662
5663
5664
5665
5666
5667
5668
5669
5670
5671
5672
5673
5674
5675
5676
5677
5678
5679
5680
5681
5682
5683
5684
5685
5686
5687
5688
5689
5690
5691
5692
5693
5694
5695
5696
5697
5698
5699
5700
5701
5702
5703
5704
5705
5706
5707
5708
5709
5710
5711
5712
5713
5714
5715
5716
5717
5718
5719
5720
5721
5722
5723
5724
5725
5726
5727
5728
5729
5730
5731
5732
5733
5734
5735
5736
5737
5738
5739
5740
5741
5742
5743
5744
5745
5746
5747
5748
5749
5750
5751
5752
5753
5754
5755
5756
5757
5758
5759
5760
5761
5762
5763
5764
5765
5766
5767
5768
5769
5770
5771
5772
5773
5774
5775
5776
5777
5778
5779
5780
5781
5782
5783
5784
5785
5786
5787
5788
5789
5790
5791
5792
5793
5794
5795
5796
5797
5798
5799
5800
5801
5802
5803
5804
5805
5806
5807
5808
5809
5810
5811
5812
5813
5814
5815
5816
5817
5818
5819
5820
5821
5822
5823
5824
5825
5826
5827
5828
5829
5830
5831
5832
5833
5834
5835
5836
5837
5838
5839
5840
5841
5842
5843
5844
5845
5846
5847
5848
5849
5850
5851
5852
5853
5854
5855
5856
5857
5858
5859
5860
5861
5862
5863
5864
5865
5866
5867
5868
5869
5870
5871
5872
5873
5874
5875
5876
5877
5878
5879
5880
5881
5882
5883
5884
5885
5886
5887
5888
5889
5890
5891
5892
5893
5894
5895
5896
5897
5898
5899
5900
5901
5902
5903
5904
5905
5906
5907
5908
5909
5910
5911
5912
5913
5914
5915
5916
5917
5918
5919
5920
5921
5922
5923
5924
5925
5926
5927
5928
5929
5930
5931
5932
5933
5934
5935
5936
5937
5938
5939
5940
5941
5942
5943
5944
5945
5946
5947
5948
5949
5950
5951
5952
5953
5954
5955
5956
5957
5958
5959
5960
5961
5962
5963
5964
5965
5966
5967
5968
5969
5970
5971
5972
5973
5974
5975
5976
5977
5978
5979
5980
5981
5982
5983
5984
5985
5986
5987
5988
5989
5990
5991
5992
5993
5994
5995
5996
5997
5998
5999
6000
6001
6002
6003
6004
6005
6006
6007
6008
6009
6010
6011
6012
6013
6014
6015
6016
6017
6018
6019
6020
6021
6022
6023
6024
6025
6026
6027
6028
6029
6030
6031
6032
6033
6034
6035
6036
6037
6038
6039
6040
6041
6042
6043
6044
6045
6046
6047
6048
6049
6050
6051
6052
6053
6054
6055
6056
6057
6058
6059
6060
6061
6062
6063
6064
6065
6066
6067
6068
6069
6070
6071
6072
6073
6074
6075
6076
6077
6078
6079
6080
6081
6082
6083
6084
6085
6086
6087
6088
6089
6090
6091
6092
6093
6094
6095
6096
6097
6098
6099
6100
6101
6102
6103
6104
6105
6106
6107
6108
6109
6110
6111
6112
6113
6114
6115
6116
6117
6118
6119
6120
6121
6122
6123
6124
6125
6126
6127
6128
6129
6130
6131
6132
6133
6134
6135
6136
6137
6138
6139
6140
6141
6142
6143
6144
6145
6146
6147
6148
6149
6150
6151
6152
6153
6154
6155
6156
6157
6158
6159
6160
6161
6162
6163
6164
6165
6166
6167
6168
6169
6170
6171
6172
6173
6174
6175
6176
6177
6178
6179
6180
6181
6182
6183
6184
6185
6186
6187
6188
6189
6190
6191
6192
6193
6194
6195
6196
6197
6198
6199
6200
6201
6202
6203
6204
6205
6206
6207
6208
6209
6210
6211
6212
6213
6214
6215
6216
6217
6218
6219
6220
6221
6222
6223
6224
6225
6226
6227
6228
6229
6230
6231
6232
6233
6234
6235
6236
6237
6238
6239
6240
6241
6242
6243
6244
6245
6246
6247
6248
6249
6250
6251
6252
6253
6254
6255
6256
6257
6258
6259
6260
6261
6262
6263
6264
6265
6266
6267
6268
6269
6270
6271
6272
6273
6274
6275
6276
6277
6278
6279
6280
6281
6282
6283
6284
6285
6286
6287
6288
6289
6290
6291
6292
6293
6294
6295
6296
6297
6298
6299
6300
6301
6302
6303
6304
6305
6306
6307
6308
6309
6310
6311
6312
6313
6314
6315
6316
6317
6318
6319
6320
6321
6322
6323
6324
6325
6326
6327
6328
6329
6330
6331
6332
6333
6334
6335
6336
6337
6338
6339
6340
6341
6342
6343
6344
6345
6346
6347
6348
6349
6350
6351
6352
6353
6354
6355
6356
6357
6358
6359
6360
6361
6362
6363
6364
6365
6366
6367
6368
6369
6370
6371
6372
6373
6374
6375
6376
6377
6378
6379
6380
6381
6382
6383
6384
6385
6386
6387
6388
6389
6390
6391
6392
6393
6394
6395
6396
6397
6398
6399
6400
6401
6402
6403
6404
6405
6406
6407
6408
6409
6410
6411
6412
6413
6414
6415
6416
6417
6418
6419
6420
6421
6422
6423
6424
6425
6426
6427
6428
6429
6430
6431
6432
6433
6434
6435
6436
6437
6438
6439
6440
6441
6442
6443
6444
6445
6446
6447
6448
6449
6450
6451
6452
6453
6454
6455
6456
6457
6458
6459
6460
6461
6462
6463
6464
6465
6466
6467
6468
6469
6470
6471
6472
6473
6474
6475
6476
6477
6478
6479
6480
6481
6482
6483
6484
6485
6486
6487
6488
6489
6490
6491
6492
6493
6494
6495
6496
6497
6498
6499
6500
6501
6502
6503
6504
6505
6506
6507
6508
6509
6510
6511
6512
6513
6514
6515
6516
6517
6518
6519
6520
6521
6522
6523
6524
6525
6526
6527
6528
6529
6530
6531
6532
6533
6534
6535
6536
6537
6538
6539
6540
6541
6542
6543
6544
6545
6546
6547
6548
6549
6550
6551
6552
6553
6554
6555
6556
6557
6558
6559
6560
6561
6562
6563
6564
6565
6566
6567
6568
6569
6570
6571
6572
6573
6574
6575
6576
6577
6578
6579
6580
6581
6582
6583
6584
6585
6586
6587
6588
6589
6590
6591
6592
6593
6594
6595
6596
6597
6598
6599
6600
6601
6602
6603
6604
6605
6606
6607
6608
6609
6610
6611
6612
6613
6614
6615
6616
6617
6618
6619
6620
6621
6622
6623
6624
6625
6626
6627
6628
6629
6630
6631
6632
6633
6634
6635
6636
6637
6638
6639
6640
6641
6642
6643
6644
6645
6646
6647
6648
6649
6650
6651
6652
6653
6654
6655
6656
6657
6658
6659
6660
6661
6662
6663
6664
6665
6666
6667
6668
6669
6670
6671
6672
6673
6674
6675
6676
6677
6678
6679
6680
6681
6682
6683
6684
6685
6686
6687
6688
6689
6690
6691
6692
6693
6694
6695
6696
6697
6698
6699
6700
6701
6702
6703
6704
6705
6706
6707
6708
6709
6710
6711
6712
6713
6714
6715
6716
6717
6718
6719
6720
6721
6722
6723
6724
6725
6726
6727
6728
6729
6730
6731
6732
6733
6734
6735
6736
6737
6738
6739
6740
6741
6742
6743
6744
6745
6746
6747
6748
6749
6750
6751
6752
6753
6754
6755
6756
6757
6758
6759
6760
6761
6762
6763
6764
6765
6766
6767
6768
6769
6770
6771
6772
6773
6774
6775
6776
6777
6778
6779
6780
6781
6782
6783
6784
6785
6786
6787
6788
6789
6790
6791
6792
6793
6794
6795
6796
6797
6798
6799
6800
6801
6802
6803
6804
6805
6806
6807
6808
6809
6810
6811
6812
6813
6814
6815
6816
6817
6818
6819
6820
6821
6822
6823
6824
6825
6826
6827
6828
6829
6830
6831
6832
6833
6834
6835
6836
6837
6838
6839
6840
6841
6842
6843
6844
6845
6846
6847
6848
6849
6850
6851
6852
6853
6854
6855
6856
6857
6858
6859
6860
6861
6862
6863
6864
6865
6866
6867
6868
6869
6870
6871
6872
6873
6874
6875
6876
6877
6878
6879
6880
6881
6882
6883
6884
6885
6886
6887
6888
6889
6890
6891
6892
6893
6894
6895
6896
6897
6898
6899
6900
6901
6902
6903
6904
6905
6906
6907
6908
6909
6910
6911
6912
6913
6914
6915
6916
6917
6918
6919
6920
6921
6922
6923
6924
6925
6926
6927
6928
6929
6930
6931
6932
6933
6934
6935
6936
6937
6938
6939
6940
6941
6942
6943
6944
6945
6946
6947
6948
6949
6950
6951
6952
6953
6954
6955
6956
6957
6958
6959
6960
6961
6962
6963
6964
6965
6966
6967
6968
6969
6970
6971
6972
6973
6974
6975
6976
6977
6978
6979
6980
6981
6982
6983
6984
6985
6986
6987
6988
6989
6990
6991
6992
6993
6994
6995
6996
6997
6998
6999
7000
7001
7002
7003
7004
7005
7006
7007
7008
7009
7010
7011
7012
7013
7014
7015
7016
7017
7018
7019
7020
7021
7022
7023
7024
7025
7026
7027
7028
7029
7030
7031
7032
7033
7034
7035
7036
7037
7038
7039
7040
7041
7042
7043
7044
7045
7046
7047
7048
7049
7050
7051
7052
7053
7054
7055
7056
7057
7058
7059
7060
7061
7062
7063
7064
7065
7066
7067
7068
7069
7070
7071
7072
7073
7074
7075
7076
7077
7078
7079
7080
7081
7082
7083
7084
7085
7086
7087
7088
7089
7090
7091
7092
7093
7094
7095
7096
7097
7098
7099
7100
7101
7102
7103
7104
7105
7106
7107
7108
7109
7110
7111
7112
7113
7114
7115
7116
7117
7118
7119
7120
7121
7122
7123
7124
7125
7126
7127
7128
7129
7130
7131
7132
7133
7134
7135
7136
7137
7138
7139
7140
7141
7142
7143
7144
7145
7146
7147
7148
7149
7150
7151
7152
7153
7154
7155
7156
7157
7158
7159
7160
7161
7162
7163
7164
7165
7166
7167
7168
7169
7170
7171
7172
7173
7174
7175
7176
7177
7178
7179
7180
7181
7182
7183
7184
7185
7186
7187
7188
7189
7190
7191
7192
7193
7194
7195
7196
7197
7198
7199
7200
7201
7202
7203
7204
7205
7206
7207
7208
7209
7210
7211
7212
7213
7214
7215
7216
7217
7218
7219
7220
7221
7222
7223
7224
7225
7226
7227
7228
7229
7230
7231
7232
7233
7234
7235
7236
7237
7238
7239
7240
7241
7242
7243
7244
7245
7246
7247
7248
7249
7250
7251
7252
7253
7254
7255
7256
7257
7258
7259
7260
7261
7262
7263
7264
7265
7266
7267
7268
7269
7270
7271
7272
7273
7274
7275
7276
7277
7278
7279
7280
7281
7282
7283
7284
7285
7286
7287
7288
7289
7290
7291
7292
7293
7294
7295
7296
7297
7298
7299
7300
7301
7302
7303
7304
7305
7306
7307
7308
7309
7310
7311
7312
7313
7314
7315
7316
7317
7318
7319
7320
7321
7322
7323
7324
7325
7326
7327
7328
7329
7330
7331
7332
7333
7334
7335
7336
7337
7338
7339
7340
7341
7342
7343
7344
7345
7346
7347
7348
7349
7350
7351
7352
7353
7354
7355
7356
7357
7358
7359
7360
7361
7362
7363
7364
7365
7366
7367
7368
7369
7370
7371
7372
7373
7374
7375
7376
7377
7378
7379
7380
7381
7382
7383
7384
7385
7386
7387
7388
7389
7390
7391
7392
7393
7394
7395
7396
7397
7398
7399
7400
7401
7402
7403
7404
7405
7406
7407
7408
7409
7410
7411
7412
7413
7414
7415
7416
7417
7418
7419
7420
7421
7422
7423
7424
7425
7426
7427
7428
7429
7430
7431
7432
7433
7434
7435
7436
7437
7438
7439
7440
7441
7442
7443
7444
7445
7446
7447
7448
7449
7450
7451
7452
7453
7454
7455
7456
7457
7458
7459
7460
7461
7462
7463
7464
7465
7466
7467
7468
7469
7470
7471
7472
7473
7474
7475
7476
7477
7478
7479
7480
7481
7482
7483
7484
7485
7486
7487
7488
7489
7490
7491
7492
7493
7494
7495
7496
7497
7498
7499
7500
7501
7502
7503
7504
7505
7506
7507
7508
7509
7510
7511
7512
7513
7514
7515
7516
7517
7518
7519
7520
7521
7522
7523
7524
7525
7526
7527
7528
7529
7530
7531
7532
7533
7534
7535
7536
7537
7538
7539
7540
7541
7542
7543
7544
7545
7546
7547
7548
7549
7550
7551
7552
7553
7554
7555
7556
7557
7558
7559
7560
7561
7562
7563
7564
7565
7566
7567
7568
7569
7570
7571
7572
7573
7574
7575
7576
7577
7578
7579
7580
7581
7582
7583
7584
7585
7586
7587
7588
7589
7590
7591
7592
7593
7594
7595
7596
7597
7598
7599
7600
7601
7602
7603
7604
7605
7606
7607
7608
7609
7610
7611
7612
7613
7614
7615
7616
7617
7618
7619
7620
7621
7622
7623
7624
7625
7626
7627
7628
7629
7630
7631
7632
7633
7634
7635
7636
7637
7638
7639
7640
7641
7642
7643
7644
7645
7646
7647
7648
7649
7650
7651
7652
7653
7654
7655
7656
7657
7658
7659
7660
7661
7662
7663
7664
7665
7666
7667
7668
7669
7670
7671
7672
7673
7674
7675
7676
7677
7678
7679
7680
7681
7682
7683
7684
7685
7686
7687
7688
7689
7690
7691
7692
7693
7694
7695
7696
7697
7698
7699
7700
7701
7702
7703
7704
7705
7706
7707
7708
7709
7710
7711
7712
7713
7714
7715
7716
7717
7718
7719
7720
7721
7722
7723
7724
7725
7726
7727
7728
7729
7730
7731
7732
7733
7734
7735
7736
7737
7738
7739
7740
7741
7742
7743
7744
7745
7746
7747
7748
7749
7750
7751
7752
7753
7754
7755
7756
7757
7758
7759
7760
7761
7762
7763
7764
7765
7766
7767
7768
7769
7770
7771
7772
7773
7774
7775
7776
7777
7778
7779
7780
7781
7782
7783
7784
7785
7786
7787
7788
7789
7790
7791
7792
7793
7794
7795
7796
7797
7798
7799
7800
7801
7802
7803
7804
7805
7806
7807
7808
7809
7810
7811
7812
7813
7814
7815
7816
7817
7818
7819
7820
7821
7822
7823
7824
7825
7826
7827
7828
7829
7830
7831
7832
7833
7834
7835
7836
7837
7838
7839
7840
7841
7842
7843
7844
7845
7846
7847
7848
7849
7850
7851
7852
7853
7854
7855
7856
7857
7858
7859
7860
7861
7862
7863
7864
7865
7866
7867
7868
7869
7870
7871
7872
7873
7874
7875
7876
7877
7878
7879
7880
7881
7882
7883
7884
7885
7886
7887
7888
7889
7890
7891
7892
7893
7894
7895
7896
7897
7898
7899
7900
7901
7902
7903
7904
7905
7906
7907
7908
7909
7910
7911
7912
7913
7914
7915
7916
7917
7918
7919
7920
7921
7922
7923
7924
7925
7926
7927
7928
7929
7930
7931
7932
7933
7934
7935
7936
7937
7938
7939
7940
7941
7942
7943
7944
7945
7946
7947
7948
7949
7950
7951
7952
7953
7954
7955
7956
7957
7958
7959
7960
7961
7962
7963
7964
7965
7966
7967
7968
7969
7970
7971
7972
7973
7974
7975
7976
7977
7978
7979
7980
7981
7982
7983
7984
7985
7986
7987
7988
7989
7990
7991
7992
7993
7994
7995
7996
7997
7998
7999
8000
8001
8002
8003
8004
8005
8006
8007
8008
8009
8010
8011
8012
8013
8014
8015
8016
8017
8018
8019
8020
8021
8022
8023
8024
8025
8026
8027
8028
8029
8030
8031
8032
8033
8034
8035
8036
8037
8038
8039
8040
8041
8042
8043
8044
8045
8046
8047
8048
8049
8050
8051
8052
8053
8054
8055
8056
8057
8058
8059
8060
8061
8062
8063
8064
8065
8066
8067
8068
8069
8070
8071
8072
8073
8074
8075
8076
8077
8078
8079
8080
8081
8082
8083
8084
8085
8086
8087
8088
8089
8090
8091
8092
8093
8094
8095
8096
8097
8098
8099
8100
8101
8102
8103
8104
8105
8106
8107
8108
8109
8110
8111
8112
8113
8114
8115
8116
8117
8118
8119
8120
8121
8122
8123
8124
8125
8126
8127
8128
8129
8130
8131
8132
8133
8134
8135
8136
8137
8138
8139
8140
8141
8142
8143
8144
8145
8146
8147
8148
8149
8150
8151
8152
8153
8154
8155
8156
8157
8158
8159
8160
8161
8162
8163
8164
8165
8166
8167
8168
8169
8170
8171
8172
8173
8174
8175
8176
8177
8178
8179
8180
8181
8182
8183
8184
8185
8186
8187
8188
8189
8190
8191
8192
8193
8194
8195
8196
8197
8198
8199
8200
8201
8202
8203
8204
8205
8206
8207
8208
8209
8210
8211
8212
8213
8214
8215
8216
8217
8218
8219
8220
8221
8222
8223
8224
8225
8226
8227
8228
8229
8230
8231
8232
8233
8234
8235
8236
8237
8238
8239
8240
8241
8242
8243
8244
8245
8246
8247
8248
8249
8250
8251
8252
8253
8254
8255
8256
8257
8258
8259
8260
8261
8262
8263
8264
8265
8266
8267
8268
8269
8270
8271
8272
8273
8274
8275
8276
8277
8278
8279
8280
8281
8282
8283
8284
8285
8286
8287
8288
8289
8290
8291
8292
8293
8294
8295
8296
8297
8298
8299
8300
8301
8302
8303
8304
8305
8306
8307
8308
8309
8310
8311
8312
8313
8314
8315
8316
8317
8318
8319
8320
8321
8322
8323
8324
8325
8326
8327
8328
8329
8330
8331
8332
8333
8334
8335
8336
8337
8338
8339
8340
8341
8342
8343
8344
8345
8346
8347
8348
8349
8350
8351
8352
8353
8354
8355
8356
8357
8358
8359
8360
8361
8362
8363
8364
8365
8366
8367
8368
8369
8370
8371
8372
8373
8374
8375
8376
8377
8378
8379
8380
8381
8382
8383
8384
8385
8386
8387
8388
8389
8390
8391
8392
8393
8394
8395
8396
8397
8398
8399
8400
8401
8402
8403
8404
8405
8406
8407
8408
8409
8410
8411
8412
8413
8414
8415
8416
8417
8418
8419
8420
8421
8422
8423
8424
8425
8426
8427
8428
8429
8430
8431
8432
8433
8434
8435
8436
8437
8438
8439
8440
8441
8442
8443
8444
8445
8446
8447
8448
8449
8450
8451
8452
8453
8454
8455
8456
8457
8458
8459
8460
8461
8462
8463
8464
8465
8466
8467
8468
8469
8470
8471
8472
8473
8474
8475
8476
8477
8478
8479
8480
8481
8482
8483
8484
8485
8486
8487
8488
8489
8490
8491
8492
8493
8494
8495
8496
8497
8498
8499
8500
8501
8502
8503
8504
8505
8506
8507
8508
8509
8510
8511
8512
8513
8514
8515
8516
8517
8518
8519
8520
8521
8522
8523
8524
8525
8526
8527
8528
8529
8530
8531
8532
8533
8534
8535
8536
8537
8538
8539
8540
8541
8542
8543
8544
8545
8546
8547
8548
8549
8550
8551
8552
8553
8554
8555
8556
8557
8558
8559
8560
8561
8562
8563
8564
8565
8566
8567
8568
8569
8570
8571
8572
8573
8574
8575
8576
8577
8578
8579
8580
8581
8582
8583
8584
8585
8586
8587
8588
8589
8590
8591
8592
8593
8594
8595
8596
8597
8598
8599
8600
8601
8602
8603
8604
8605
8606
8607
8608
8609
8610
8611
8612
8613
8614
8615
8616
8617
8618
8619
8620
8621
8622
8623
8624
8625
8626
8627
8628
8629
8630
8631
8632
8633
8634
8635
8636
8637
8638
8639
8640
8641
8642
8643
8644
8645
8646
8647
8648
8649
8650
8651
8652
8653
8654
8655
8656
8657
8658
8659
8660
8661
8662
8663
8664
8665
8666
8667
8668
8669
8670
8671
8672
8673
8674
8675
8676
8677
8678
8679
8680
8681
8682
8683
8684
8685
8686
8687
8688
8689
8690
8691
8692
8693
8694
8695
8696
8697
8698
8699
8700
8701
8702
8703
8704
8705
8706
8707
8708
8709
8710
8711
8712
8713
8714
8715
8716
8717
8718
8719
8720
8721
8722
8723
8724
8725
8726
8727
8728
8729
8730
8731
8732
8733
8734
8735
8736
8737
8738
8739
8740
8741
8742
8743
8744
8745
8746
8747
8748
8749
8750
8751
8752
8753
8754
8755
8756
8757
8758
8759
8760
8761
8762
8763
8764
8765
8766
8767
8768
8769
8770
8771
8772
8773
8774
8775
8776
8777
8778
8779
8780
8781
8782
8783
8784
8785
8786
8787
8788
8789
8790
8791
8792
8793
8794
8795
8796
8797
8798
8799
8800
8801
8802
8803
8804
8805
8806
8807
8808
8809
8810
8811
8812
8813
8814
8815
8816
8817
8818
8819
8820
8821
8822
8823
8824
8825
8826
8827
8828
8829
8830
8831
8832
8833
8834
8835
8836
8837
8838
8839
8840
8841
8842
8843
8844
8845
8846
8847
8848
8849
8850
8851
8852
8853
8854
8855
8856
8857
8858
8859
8860
8861
8862
8863
8864
8865
8866
8867
8868
8869
8870
8871
8872
8873
8874
8875
8876
8877
8878
8879
8880
8881
8882
8883
8884
8885
8886
8887
8888
8889
8890
8891
8892
8893
8894
8895
8896
8897
8898
8899
8900
8901
8902
8903
8904
8905
8906
8907
8908
8909
8910
8911
8912
8913
8914
8915
8916
8917
8918
8919
8920
8921
8922
8923
8924
8925
8926
8927
8928
8929
8930
8931
8932
8933
8934
8935
8936
8937
8938
8939
8940
8941
8942
8943
8944
8945
8946
8947
8948
8949
8950
8951
8952
8953
8954
8955
8956
8957
8958
8959
8960
8961
8962
8963
8964
8965
8966
8967
8968
8969
8970
8971
8972
8973
8974
8975
8976
8977
8978
8979
8980
8981
8982
8983
8984
8985
8986
8987
8988
8989
8990
8991
8992
8993
8994
8995
8996
8997
8998
8999
9000
9001
9002
9003
9004
9005
9006
9007
9008
9009
9010
9011
9012
9013
9014
9015
9016
9017
9018
9019
9020
9021
9022
9023
9024
9025
9026
9027
9028
9029
9030
9031
9032
9033
9034
9035
9036
9037
9038
9039
9040
9041
9042
9043
9044
9045
9046
9047
9048
9049
9050
9051
9052
9053
9054
9055
9056
9057
9058
9059
9060
9061
9062
9063
9064
9065
9066
9067
9068
9069
9070
9071
9072
9073
9074
9075
9076
9077
9078
9079
9080
9081
9082
9083
9084
9085
9086
9087
9088
9089
9090
9091
9092
9093
9094
9095
9096
9097
9098
9099
9100
9101
9102
9103
9104
9105
9106
9107
9108
9109
9110
9111
9112
9113
9114
9115
9116
9117
9118
9119
9120
9121
9122
9123
9124
9125
9126
9127
9128
9129
9130
9131
9132
9133
9134
9135
9136
9137
9138
9139
9140
9141
9142
|
/* Copyright (C) 2007 MySQL AB & Sanja Belkin. 2010 Monty Program Ab.
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; version 2 of the License.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
#include "maria_def.h"
#include "trnman.h"
#include "ma_blockrec.h" /* for some constants and in-write hooks */
#include "ma_key_recover.h" /* For some in-write hooks */
#include "ma_checkpoint.h"
#include "ma_servicethread.h"
/*
On Windows, neither my_open() nor mysql_file_sync() work for directories.
Also there is no need to flush filesystem changes ,i.e to sync()
directories.
*/
#ifdef __WIN__
#define sync_dir(A,B) 0
#else
#define sync_dir(A,B) mysql_file_sync(A,B)
#endif
/**
@file
@brief Module which writes and reads to a transaction log
*/
/* 0xFF can never be valid first byte of a chunk */
#define TRANSLOG_FILLER 0xFF
/* number of opened log files in the pagecache (should be at least 2) */
#define OPENED_FILES_NUM 3
#define CACHED_FILES_NUM 5
#define CACHED_FILES_NUM_DIRECT_SEARCH_LIMIT 7
#if CACHED_FILES_NUM > CACHED_FILES_NUM_DIRECT_SEARCH_LIMIT
#include <hash.h>
#include <m_ctype.h>
#endif
/** @brief protects checkpoint_in_progress */
static mysql_mutex_t LOCK_soft_sync;
/** @brief for killing the background checkpoint thread */
static mysql_cond_t COND_soft_sync;
/** @brief control structure for checkpoint background thread */
static MA_SERVICE_THREAD_CONTROL soft_sync_control=
{THREAD_DEAD, FALSE, &LOCK_soft_sync, &COND_soft_sync};
/* transaction log file descriptor */
typedef struct st_translog_file
{
uint32 number;
PAGECACHE_FILE handler;
my_bool was_recovered;
my_bool is_sync;
} TRANSLOG_FILE;
/* records buffer size (should be TRANSLOG_PAGE_SIZE * n) */
#define TRANSLOG_WRITE_BUFFER (1024*1024)
/*
pagecache_read/write/inject() use bmove512() on their buffers so those must
be long-aligned, which we guarantee by using the type below:
*/
typedef union
{
ulonglong dummy;
uchar buffer[TRANSLOG_PAGE_SIZE];
} TRANSLOG_PAGE_SIZE_BUFF;
/* min chunk length */
#define TRANSLOG_MIN_CHUNK 3
/*
Number of buffers used by loghandler
Should be at least 4, because one thread can block up to 2 buffers in
normal circumstances (less then half of one and full other, or just
switched one and other), But if we met end of the file in the middle and
have to switch buffer it will be 3. + 1 buffer for flushing/writing.
We have a bigger number here for higher concurrency and to make division
faster.
The number should be power of 2 to be fast.
*/
#define TRANSLOG_BUFFERS_NO 8
/* number of bytes (+ header) which can be unused on first page in sequence */
#define TRANSLOG_MINCHUNK_CONTENT 1
/* version of log file */
#define TRANSLOG_VERSION_ID 10000 /* 1.00.00 */
#define TRANSLOG_PAGE_FLAGS 6 /* transaction log page flags offset */
/* Maximum length of compressed LSNs (the worst case of whole LSN storing) */
#define COMPRESSED_LSN_MAX_STORE_SIZE (2 + LSN_STORE_SIZE)
#define MAX_NUMBER_OF_LSNS_PER_RECORD 2
/* max lsn calculation for buffer */
#define BUFFER_MAX_LSN(B) \
((B)->last_lsn == LSN_IMPOSSIBLE ? (B)->prev_last_lsn : (B)->last_lsn)
/* log write buffer descriptor */
struct st_translog_buffer
{
/*
Cache for current log. Comes first to be aligned for bmove512() in
pagecache_inject()
*/
uchar buffer[TRANSLOG_WRITE_BUFFER];
/*
Maximum LSN of records which ends in this buffer (or IMPOSSIBLE_LSN
if no LSNs ends here)
*/
LSN last_lsn;
/* last_lsn of previous buffer or IMPOSSIBLE_LSN if it is very first one */
LSN prev_last_lsn;
/* This buffer offset in the file */
TRANSLOG_ADDRESS offset;
/*
Next buffer offset in the file (it is not always offset + size,
in case of flush by LSN it can be offset + size - TRANSLOG_PAGE_SIZE)
*/
TRANSLOG_ADDRESS next_buffer_offset;
/* Previous buffer offset to detect it flush finish */
TRANSLOG_ADDRESS prev_buffer_offset;
/*
If the buffer was forced to close it save value of its horizon
otherwise LSN_IMPOSSIBLE
*/
TRANSLOG_ADDRESS pre_force_close_horizon;
/*
How much is written (or will be written when copy_to_buffer_in_progress
become 0) to this buffer
*/
translog_size_t size;
/*
When moving from one log buffer to another, we write the last of the
previous buffer to file and then move to start using the new log
buffer. In the case of a part filed last page, this page is not moved
to the start of the new buffer but instead we set the 'skip_data'
variable to tell us how much data at the beginning of the buffer is not
relevant.
*/
uint skipped_data;
/* File handler for this buffer */
TRANSLOG_FILE *file;
/* Threads which are waiting for buffer filling/freeing */
mysql_cond_t waiting_filling_buffer;
/* Number of records which are in copy progress */
uint copy_to_buffer_in_progress;
/* list of waiting buffer ready threads */
struct st_my_thread_var *waiting_flush;
/*
If true then previous buffer overlap with this one (due to flush of
loghandler, the last page of that buffer is the same as the first page
of this buffer) and have to be written first (because contain old
content of page which present in both buffers)
*/
my_bool overlay;
uint buffer_no;
/*
Lock for the buffer.
Current buffer also lock the whole handler (if one want lock the handler
one should lock the current buffer).
Buffers are locked only in one direction (with overflow and beginning
from the first buffer). If we keep lock on buffer N we can lock only
buffer N+1 (never N-1).
One thread do not lock more then 2 buffer in a time, so to make dead
lock it should be N thread (where N equal number of buffers) takes one
buffer and try to lock next. But it is impossible because there is only
2 cases when thread take 2 buffers: 1) one thread finishes current
buffer (where horizon is) and start next (to which horizon moves). 2)
flush start from buffer after current (oldest) and go till the current
crabbing by buffer sequence. And there is only one flush in a moment
(they are serialised).
Because of above and number of buffers equal 5 we can't get dead lock (it is
impossible to get all 5 buffers locked simultaneously).
*/
mysql_mutex_t mutex;
/*
Some thread is going to close the buffer and it should be
done only by that thread
*/
my_bool is_closing_buffer;
/*
Version of the buffer increases every time buffer the buffer flushed.
With file and offset it allow detect buffer changes
*/
uint8 ver;
/*
When previous buffer sent to disk it set its address here to allow
to detect when it is done
(we have to keep it in this buffer to lock buffers only in one direction).
*/
TRANSLOG_ADDRESS prev_sent_to_disk;
mysql_cond_t prev_sent_to_disk_cond;
};
struct st_buffer_cursor
{
/* pointer into the buffer */
uchar *ptr;
/* current buffer */
struct st_translog_buffer *buffer;
/* How many bytes we wrote on the current page */
uint16 current_page_fill;
/*
How many times we write the page on the disk during flushing process
(for sector protection).
*/
uint16 write_counter;
/* previous write offset */
uint16 previous_offset;
/* Number of current buffer */
uint8 buffer_no;
/*
True if it is just filling buffer after advancing the pointer to
the horizon.
*/
my_bool chaser;
/*
Is current page of the cursor already finished (sector protection
should be applied if it is needed)
*/
my_bool protected;
};
typedef uint8 dirty_buffer_mask_t;
struct st_translog_descriptor
{
/* *** Parameters of the log handler *** */
/* Page cache for the log reads */
PAGECACHE *pagecache;
uint flags;
/* File open flags */
uint open_flags;
/* max size of one log size (for new logs creation) */
uint32 log_file_max_size;
uint32 server_version;
/* server ID (used for replication) */
uint32 server_id;
/* Loghandler's buffer capacity in case of chunk 2 filling */
uint32 buffer_capacity_chunk_2;
/*
Half of the buffer capacity in case of chunk 2 filling,
used to decide will we write a record in one group or many.
It is written to the variable just to avoid devision every
time we need it.
*/
uint32 half_buffer_capacity_chunk_2;
/* Page overhead calculated by flags (whether CRC is enabled, etc) */
uint16 page_overhead;
/*
Page capacity ("useful load") calculated by flags
(TRANSLOG_PAGE_SIZE - page_overhead-1)
*/
uint16 page_capacity_chunk_2;
/* Path to the directory where we store log store files */
char directory[FN_REFLEN];
/* *** Current state of the log handler *** */
/* list of opened files */
DYNAMIC_ARRAY open_files;
/* min/max number of file in the array */
uint32 max_file, min_file;
/* the opened files list guard */
mysql_rwlock_t open_files_lock;
/*
File descriptor of the directory where we store log files for syncing
it.
*/
File directory_fd;
/* buffers for log writing */
struct st_translog_buffer buffers[TRANSLOG_BUFFERS_NO];
/* Mask where 1 in position N mean that buffer N is not flushed */
dirty_buffer_mask_t dirty_buffer_mask;
/* The above variable protection */
mysql_mutex_t dirty_buffer_mask_lock;
/*
horizon - visible end of the log (here is absolute end of the log:
position where next chunk can start
*/
TRANSLOG_ADDRESS horizon;
/* horizon buffer cursor */
struct st_buffer_cursor bc;
/* maximum LSN of the current (not finished) file */
LSN max_lsn;
/*
Last flushed LSN (protected by log_flush_lock).
Pointers in the log ordered like this:
last_lsn_checked <= flushed <= sent_to_disk <= in_buffers_only <=
max_lsn <= horizon
*/
LSN flushed;
/* Last LSN sent to the disk (but maybe not written yet) */
LSN sent_to_disk;
/* Horizon from which log started after initialization */
TRANSLOG_ADDRESS log_start;
TRANSLOG_ADDRESS previous_flush_horizon;
/* All what is after this address is not sent to disk yet */
TRANSLOG_ADDRESS in_buffers_only;
/* protection of sent_to_disk and in_buffers_only */
mysql_mutex_t sent_to_disk_lock;
/*
Protect flushed (see above) and for flush serialization (will
be removed in v1.5
*/
mysql_mutex_t log_flush_lock;
mysql_cond_t log_flush_cond;
mysql_cond_t new_goal_cond;
/* Protects changing of headers of finished files (max_lsn) */
mysql_mutex_t file_header_lock;
/*
Sorted array (with protection) of files where we started writing process
and so we can't give last LSN yet
*/
mysql_mutex_t unfinished_files_lock;
DYNAMIC_ARRAY unfinished_files;
/*
minimum number of still need file calculeted during last
translog_purge call
*/
uint32 min_need_file;
/* Purger data: minimum file in the log (or 0 if unknown) */
uint32 min_file_number;
/* Protect purger from many calls and it's data */
mysql_mutex_t purger_lock;
/* last low water mark checked */
LSN last_lsn_checked;
/**
Must be set to 0 under loghandler lock every time a new LSN
is generated.
*/
my_bool is_everything_flushed;
/* True when flush pass is in progress */
my_bool flush_in_progress;
/* The flush number (used to distinguish two flushes goes one by one) */
volatile int flush_no;
/* Next flush pass variables */
TRANSLOG_ADDRESS next_pass_max_lsn;
pthread_t max_lsn_requester;
};
static struct st_translog_descriptor log_descriptor;
ulong log_purge_type= TRANSLOG_PURGE_IMMIDIATE;
ulong log_file_size= TRANSLOG_FILE_SIZE;
/* sync() of log files directory mode */
ulong sync_log_dir= TRANSLOG_SYNC_DIR_NEWFILE;
ulong maria_group_commit= TRANSLOG_GCOMMIT_NONE;
ulong maria_group_commit_interval= 0;
/* Marker for end of log */
static uchar end_of_log= 0;
#define END_OF_LOG &end_of_log
/**
Switch for "soft" sync (no real sync() but periodical sync by service
thread)
*/
static volatile my_bool soft_sync= FALSE;
/**
Switch for "hard" group commit mode
*/
static volatile my_bool hard_group_commit= FALSE;
/**
File numbers interval which have to be sync()
*/
static uint32 soft_sync_min= 0;
static uint32 soft_sync_max= 0;
static uint32 soft_need_sync= 1;
/**
stores interval in microseconds
*/
static uint32 group_commit_wait= 0;
enum enum_translog_status translog_status= TRANSLOG_UNINITED;
ulonglong translog_syncs= 0; /* Number of sync()s */
/* time of last flush */
static ulonglong flush_start= 0;
/* chunk types */
#define TRANSLOG_CHUNK_LSN 0x00 /* 0 chunk refer as LSN (head or tail */
#define TRANSLOG_CHUNK_FIXED (1 << 6) /* 1 (pseudo)fixed record (also LSN) */
#define TRANSLOG_CHUNK_NOHDR (2 << 6) /* 2 no head chunk (till page end) */
#define TRANSLOG_CHUNK_LNGTH (3 << 6) /* 3 chunk with chunk length */
#define TRANSLOG_CHUNK_TYPE (3 << 6) /* Mask to get chunk type */
#define TRANSLOG_REC_TYPE 0x3F /* Mask to get record type */
#define TRANSLOG_CHUNK_0_CONT 0x3F /* the type to mark chunk 0 continue */
/* compressed (relative) LSN constants */
#define TRANSLOG_CLSN_LEN_BITS 0xC0 /* Mask to get compressed LSN length */
#include <my_atomic.h>
/* an array that maps id of a MARIA_SHARE to this MARIA_SHARE */
static MARIA_SHARE **id_to_share= NULL;
/* lock for id_to_share */
static my_atomic_rwlock_t LOCK_id_to_share;
static my_bool translog_dummy_callback(uchar *page,
pgcache_page_no_t page_no,
uchar* data_ptr);
static my_bool translog_page_validator(uchar *page,
pgcache_page_no_t page_no,
uchar* data_ptr);
static my_bool translog_get_next_chunk(TRANSLOG_SCANNER_DATA *scanner);
static uint32 translog_first_file(TRANSLOG_ADDRESS horizon, int is_protected);
LSN translog_next_LSN(TRANSLOG_ADDRESS addr, TRANSLOG_ADDRESS horizon);
/*
Initialize log_record_type_descriptors
*/
LOG_DESC log_record_type_descriptor[LOGREC_NUMBER_OF_TYPES];
#ifndef DBUG_OFF
#define translog_buffer_lock_assert_owner(B) \
mysql_mutex_assert_owner(&(B)->mutex)
#define translog_lock_assert_owner() \
mysql_mutex_assert_owner(&log_descriptor.bc.buffer->mutex)
void translog_lock_handler_assert_owner()
{
translog_lock_assert_owner();
}
/**
@brief check the description table validity
@param num how many records should be filled
*/
static uint max_allowed_translog_type= 0;
void check_translog_description_table(int num)
{
int i;
DBUG_ENTER("check_translog_description_table");
DBUG_PRINT("enter", ("last record: %d", num));
DBUG_ASSERT(num > 0);
/* last is reserved for extending the table */
DBUG_ASSERT(num < LOGREC_NUMBER_OF_TYPES - 1);
DBUG_ASSERT(log_record_type_descriptor[0].rclass == LOGRECTYPE_NOT_ALLOWED);
max_allowed_translog_type= num;
for (i= 0; i <= num; i++)
{
DBUG_PRINT("info",
("record type: %d class: %d fixed: %u header: %u LSNs: %u "
"name: %s",
i, log_record_type_descriptor[i].rclass,
(uint)log_record_type_descriptor[i].fixed_length,
(uint)log_record_type_descriptor[i].read_header_len,
(uint)log_record_type_descriptor[i].compressed_LSN,
log_record_type_descriptor[i].name));
switch (log_record_type_descriptor[i].rclass) {
case LOGRECTYPE_NOT_ALLOWED:
DBUG_ASSERT(i == 0);
break;
case LOGRECTYPE_VARIABLE_LENGTH:
DBUG_ASSERT(log_record_type_descriptor[i].fixed_length == 0);
DBUG_ASSERT((log_record_type_descriptor[i].compressed_LSN == 0) ||
((log_record_type_descriptor[i].compressed_LSN == 1) &&
(log_record_type_descriptor[i].read_header_len >=
LSN_STORE_SIZE)) ||
((log_record_type_descriptor[i].compressed_LSN == 2) &&
(log_record_type_descriptor[i].read_header_len >=
LSN_STORE_SIZE * 2)));
break;
case LOGRECTYPE_PSEUDOFIXEDLENGTH:
DBUG_ASSERT(log_record_type_descriptor[i].fixed_length ==
log_record_type_descriptor[i].read_header_len);
DBUG_ASSERT(log_record_type_descriptor[i].compressed_LSN > 0);
DBUG_ASSERT(log_record_type_descriptor[i].compressed_LSN <= 2);
break;
case LOGRECTYPE_FIXEDLENGTH:
DBUG_ASSERT(log_record_type_descriptor[i].fixed_length ==
log_record_type_descriptor[i].read_header_len);
DBUG_ASSERT(log_record_type_descriptor[i].compressed_LSN == 0);
break;
default:
DBUG_ASSERT(0);
}
}
for (i= num + 1; i < LOGREC_NUMBER_OF_TYPES; i++)
{
DBUG_ASSERT(log_record_type_descriptor[i].rclass ==
LOGRECTYPE_NOT_ALLOWED);
}
DBUG_VOID_RETURN;
}
#else
#define translog_buffer_lock_assert_owner(B) {}
#define translog_lock_assert_owner() {}
#endif
static LOG_DESC INIT_LOGREC_RESERVED_FOR_CHUNKS23=
{LOGRECTYPE_NOT_ALLOWED, 0, 0, NULL, NULL, NULL, 0,
"reserved", LOGREC_NOT_LAST_IN_GROUP, NULL, NULL };
static LOG_DESC INIT_LOGREC_REDO_INSERT_ROW_HEAD=
{LOGRECTYPE_VARIABLE_LENGTH, 0,
FILEID_STORE_SIZE + PAGE_STORE_SIZE + DIRPOS_STORE_SIZE, NULL,
write_hook_for_redo, NULL, 0,
"redo_insert_row_head", LOGREC_NOT_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_REDO_INSERT_ROW_TAIL=
{LOGRECTYPE_VARIABLE_LENGTH, 0,
FILEID_STORE_SIZE + PAGE_STORE_SIZE + DIRPOS_STORE_SIZE, NULL,
write_hook_for_redo, NULL, 0,
"redo_insert_row_tail", LOGREC_NOT_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_REDO_NEW_ROW_HEAD=
{LOGRECTYPE_VARIABLE_LENGTH, 0,
FILEID_STORE_SIZE + PAGE_STORE_SIZE + DIRPOS_STORE_SIZE, NULL,
write_hook_for_redo, NULL, 0,
"redo_new_row_head", LOGREC_NOT_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_REDO_NEW_ROW_TAIL=
{LOGRECTYPE_VARIABLE_LENGTH, 0,
FILEID_STORE_SIZE + PAGE_STORE_SIZE + DIRPOS_STORE_SIZE, NULL,
write_hook_for_redo, NULL, 0,
"redo_new_row_tail", LOGREC_NOT_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_REDO_INSERT_ROW_BLOBS=
{LOGRECTYPE_VARIABLE_LENGTH, 0, FILEID_STORE_SIZE, NULL,
write_hook_for_redo, NULL, 0,
"redo_insert_row_blobs", LOGREC_NOT_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_REDO_PURGE_ROW_HEAD=
{LOGRECTYPE_FIXEDLENGTH,
FILEID_STORE_SIZE + PAGE_STORE_SIZE + DIRPOS_STORE_SIZE,
FILEID_STORE_SIZE + PAGE_STORE_SIZE + DIRPOS_STORE_SIZE,
NULL, write_hook_for_redo, NULL, 0,
"redo_purge_row_head", LOGREC_NOT_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_REDO_PURGE_ROW_TAIL=
{LOGRECTYPE_FIXEDLENGTH,
FILEID_STORE_SIZE + PAGE_STORE_SIZE + DIRPOS_STORE_SIZE,
FILEID_STORE_SIZE + PAGE_STORE_SIZE + DIRPOS_STORE_SIZE,
NULL, write_hook_for_redo, NULL, 0,
"redo_purge_row_tail", LOGREC_NOT_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_REDO_FREE_BLOCKS=
{LOGRECTYPE_VARIABLE_LENGTH, 0,
FILEID_STORE_SIZE + PAGERANGE_STORE_SIZE,
NULL, write_hook_for_redo, NULL, 0,
"redo_free_blocks", LOGREC_NOT_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_REDO_FREE_HEAD_OR_TAIL=
{LOGRECTYPE_FIXEDLENGTH,
FILEID_STORE_SIZE + PAGE_STORE_SIZE,
FILEID_STORE_SIZE + PAGE_STORE_SIZE,
NULL, write_hook_for_redo, NULL, 0,
"redo_free_head_or_tail", LOGREC_NOT_LAST_IN_GROUP, NULL, NULL};
/* not yet used; for when we have versioning */
static LOG_DESC INIT_LOGREC_REDO_DELETE_ROW=
{LOGRECTYPE_FIXEDLENGTH, 16, 16, NULL, write_hook_for_redo, NULL, 0,
"redo_delete_row", LOGREC_NOT_LAST_IN_GROUP, NULL, NULL};
/** @todo RECOVERY BUG unused, remove? */
static LOG_DESC INIT_LOGREC_REDO_UPDATE_ROW_HEAD=
{LOGRECTYPE_VARIABLE_LENGTH, 0, 9, NULL, write_hook_for_redo, NULL, 0,
"redo_update_row_head", LOGREC_NOT_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_REDO_INDEX=
{LOGRECTYPE_VARIABLE_LENGTH, 0, 9, NULL, write_hook_for_redo, NULL, 0,
"redo_index", LOGREC_NOT_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_REDO_INDEX_NEW_PAGE=
{LOGRECTYPE_VARIABLE_LENGTH, 0,
FILEID_STORE_SIZE + PAGE_STORE_SIZE * 2 + KEY_NR_STORE_SIZE + 1,
NULL, write_hook_for_redo, NULL, 0,
"redo_index_new_page", LOGREC_NOT_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_REDO_INDEX_FREE_PAGE=
{LOGRECTYPE_FIXEDLENGTH, FILEID_STORE_SIZE + PAGE_STORE_SIZE * 2,
FILEID_STORE_SIZE + PAGE_STORE_SIZE * 2,
NULL, write_hook_for_redo, NULL, 0,
"redo_index_free_page", LOGREC_NOT_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_REDO_UNDELETE_ROW=
{LOGRECTYPE_FIXEDLENGTH, 16, 16, NULL, write_hook_for_redo, NULL, 0,
"redo_undelete_row", LOGREC_NOT_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_CLR_END=
{LOGRECTYPE_VARIABLE_LENGTH, 0, LSN_STORE_SIZE + FILEID_STORE_SIZE +
CLR_TYPE_STORE_SIZE, NULL, write_hook_for_clr_end, NULL, 1,
"clr_end", LOGREC_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_PURGE_END=
{LOGRECTYPE_PSEUDOFIXEDLENGTH, 5, 5, NULL, NULL, NULL, 1,
"purge_end", LOGREC_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_UNDO_ROW_INSERT=
{LOGRECTYPE_VARIABLE_LENGTH, 0,
LSN_STORE_SIZE + FILEID_STORE_SIZE + PAGE_STORE_SIZE + DIRPOS_STORE_SIZE,
NULL, write_hook_for_undo_row_insert, NULL, 1,
"undo_row_insert", LOGREC_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_UNDO_ROW_DELETE=
{LOGRECTYPE_VARIABLE_LENGTH, 0,
LSN_STORE_SIZE + FILEID_STORE_SIZE + PAGE_STORE_SIZE + DIRPOS_STORE_SIZE,
NULL, write_hook_for_undo_row_delete, NULL, 1,
"undo_row_delete", LOGREC_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_UNDO_ROW_UPDATE=
{LOGRECTYPE_VARIABLE_LENGTH, 0,
LSN_STORE_SIZE + FILEID_STORE_SIZE + PAGE_STORE_SIZE + DIRPOS_STORE_SIZE,
NULL, write_hook_for_undo_row_update, NULL, 1,
"undo_row_update", LOGREC_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_UNDO_KEY_INSERT=
{LOGRECTYPE_VARIABLE_LENGTH, 0,
LSN_STORE_SIZE + FILEID_STORE_SIZE + KEY_NR_STORE_SIZE,
NULL, write_hook_for_undo_key_insert, NULL, 1,
"undo_key_insert", LOGREC_LAST_IN_GROUP, NULL, NULL};
/* This will never be in the log, only in the clr */
static LOG_DESC INIT_LOGREC_UNDO_KEY_INSERT_WITH_ROOT=
{LOGRECTYPE_VARIABLE_LENGTH, 0,
LSN_STORE_SIZE + FILEID_STORE_SIZE + KEY_NR_STORE_SIZE + PAGE_STORE_SIZE,
NULL, write_hook_for_undo_key, NULL, 1,
"undo_key_insert_with_root", LOGREC_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_UNDO_KEY_DELETE=
{LOGRECTYPE_VARIABLE_LENGTH, 0,
LSN_STORE_SIZE + FILEID_STORE_SIZE + KEY_NR_STORE_SIZE,
NULL, write_hook_for_undo_key_delete, NULL, 1,
"undo_key_delete", LOGREC_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_UNDO_KEY_DELETE_WITH_ROOT=
{LOGRECTYPE_VARIABLE_LENGTH, 0,
LSN_STORE_SIZE + FILEID_STORE_SIZE + KEY_NR_STORE_SIZE + PAGE_STORE_SIZE,
NULL, write_hook_for_undo_key_delete, NULL, 1,
"undo_key_delete_with_root", LOGREC_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_PREPARE=
{LOGRECTYPE_VARIABLE_LENGTH, 0, 0, NULL, NULL, NULL, 0,
"prepare", LOGREC_IS_GROUP_ITSELF, NULL, NULL};
static LOG_DESC INIT_LOGREC_PREPARE_WITH_UNDO_PURGE=
{LOGRECTYPE_VARIABLE_LENGTH, 0, LSN_STORE_SIZE, NULL, NULL, NULL, 1,
"prepare_with_undo_purge", LOGREC_IS_GROUP_ITSELF, NULL, NULL};
static LOG_DESC INIT_LOGREC_COMMIT=
{LOGRECTYPE_FIXEDLENGTH, 0, 0, NULL,
write_hook_for_commit, NULL, 0, "commit", LOGREC_IS_GROUP_ITSELF, NULL,
NULL};
static LOG_DESC INIT_LOGREC_COMMIT_WITH_UNDO_PURGE=
{LOGRECTYPE_PSEUDOFIXEDLENGTH, 5, 5, NULL, write_hook_for_commit, NULL, 1,
"commit_with_undo_purge", LOGREC_IS_GROUP_ITSELF, NULL, NULL};
static LOG_DESC INIT_LOGREC_CHECKPOINT=
{LOGRECTYPE_VARIABLE_LENGTH, 0, 0, NULL, NULL, NULL, 0,
"checkpoint", LOGREC_IS_GROUP_ITSELF, NULL, NULL};
static LOG_DESC INIT_LOGREC_REDO_CREATE_TABLE=
{LOGRECTYPE_VARIABLE_LENGTH, 0, 1 + 2, NULL, NULL, NULL, 0,
"redo_create_table", LOGREC_IS_GROUP_ITSELF, NULL, NULL};
static LOG_DESC INIT_LOGREC_REDO_RENAME_TABLE=
{LOGRECTYPE_VARIABLE_LENGTH, 0, 0, NULL, NULL, NULL, 0,
"redo_rename_table", LOGREC_IS_GROUP_ITSELF, NULL, NULL};
static LOG_DESC INIT_LOGREC_REDO_DROP_TABLE=
{LOGRECTYPE_VARIABLE_LENGTH, 0, 0, NULL, NULL, NULL, 0,
"redo_drop_table", LOGREC_IS_GROUP_ITSELF, NULL, NULL};
static LOG_DESC INIT_LOGREC_REDO_DELETE_ALL=
{LOGRECTYPE_FIXEDLENGTH, FILEID_STORE_SIZE, FILEID_STORE_SIZE,
NULL, write_hook_for_redo_delete_all, NULL, 0,
"redo_delete_all", LOGREC_IS_GROUP_ITSELF, NULL, NULL};
static LOG_DESC INIT_LOGREC_REDO_REPAIR_TABLE=
{LOGRECTYPE_FIXEDLENGTH, FILEID_STORE_SIZE + 8 + 8, FILEID_STORE_SIZE + 8 + 8,
NULL, NULL, NULL, 0,
"redo_repair_table", LOGREC_IS_GROUP_ITSELF, NULL, NULL};
static LOG_DESC INIT_LOGREC_FILE_ID=
{LOGRECTYPE_VARIABLE_LENGTH, 0, 2, NULL, write_hook_for_file_id, NULL, 0,
"file_id", LOGREC_IS_GROUP_ITSELF, NULL, NULL};
static LOG_DESC INIT_LOGREC_LONG_TRANSACTION_ID=
{LOGRECTYPE_FIXEDLENGTH, 6, 6, NULL, NULL, NULL, 0,
"long_transaction_id", LOGREC_IS_GROUP_ITSELF, NULL, NULL};
static LOG_DESC INIT_LOGREC_INCOMPLETE_LOG=
{LOGRECTYPE_FIXEDLENGTH, FILEID_STORE_SIZE, FILEID_STORE_SIZE,
NULL, NULL, NULL, 0,
"incomplete_log", LOGREC_IS_GROUP_ITSELF, NULL, NULL};
static LOG_DESC INIT_LOGREC_INCOMPLETE_GROUP=
{LOGRECTYPE_FIXEDLENGTH, 0, 0,
NULL, NULL, NULL, 0,
"incomplete_group", LOGREC_IS_GROUP_ITSELF, NULL, NULL};
static LOG_DESC INIT_LOGREC_UNDO_BULK_INSERT=
{LOGRECTYPE_VARIABLE_LENGTH, 0,
LSN_STORE_SIZE + FILEID_STORE_SIZE,
NULL, write_hook_for_undo_bulk_insert, NULL, 1,
"undo_bulk_insert", LOGREC_LAST_IN_GROUP, NULL, NULL};
static LOG_DESC INIT_LOGREC_REDO_BITMAP_NEW_PAGE=
{LOGRECTYPE_FIXEDLENGTH, FILEID_STORE_SIZE + PAGE_STORE_SIZE * 2,
FILEID_STORE_SIZE + PAGE_STORE_SIZE * 2,
NULL, NULL, NULL, 0,
"redo_create_bitmap", LOGREC_IS_GROUP_ITSELF, NULL, NULL};
static LOG_DESC INIT_LOGREC_IMPORTED_TABLE=
{LOGRECTYPE_VARIABLE_LENGTH, 0, 0, NULL, NULL, NULL, 0,
"imported_table", LOGREC_IS_GROUP_ITSELF, NULL, NULL};
static LOG_DESC INIT_LOGREC_DEBUG_INFO=
{LOGRECTYPE_VARIABLE_LENGTH, 0, 0, NULL, NULL, NULL, 0,
"info", LOGREC_IS_GROUP_ITSELF, NULL, NULL};
const myf log_write_flags= MY_WME | MY_NABP | MY_WAIT_IF_FULL;
void translog_table_init()
{
int i;
log_record_type_descriptor[LOGREC_RESERVED_FOR_CHUNKS23]=
INIT_LOGREC_RESERVED_FOR_CHUNKS23;
log_record_type_descriptor[LOGREC_REDO_INSERT_ROW_HEAD]=
INIT_LOGREC_REDO_INSERT_ROW_HEAD;
log_record_type_descriptor[LOGREC_REDO_INSERT_ROW_TAIL]=
INIT_LOGREC_REDO_INSERT_ROW_TAIL;
log_record_type_descriptor[LOGREC_REDO_NEW_ROW_HEAD]=
INIT_LOGREC_REDO_NEW_ROW_HEAD;
log_record_type_descriptor[LOGREC_REDO_NEW_ROW_TAIL]=
INIT_LOGREC_REDO_NEW_ROW_TAIL;
log_record_type_descriptor[LOGREC_REDO_INSERT_ROW_BLOBS]=
INIT_LOGREC_REDO_INSERT_ROW_BLOBS;
log_record_type_descriptor[LOGREC_REDO_PURGE_ROW_HEAD]=
INIT_LOGREC_REDO_PURGE_ROW_HEAD;
log_record_type_descriptor[LOGREC_REDO_PURGE_ROW_TAIL]=
INIT_LOGREC_REDO_PURGE_ROW_TAIL;
log_record_type_descriptor[LOGREC_REDO_FREE_BLOCKS]=
INIT_LOGREC_REDO_FREE_BLOCKS;
log_record_type_descriptor[LOGREC_REDO_FREE_HEAD_OR_TAIL]=
INIT_LOGREC_REDO_FREE_HEAD_OR_TAIL;
log_record_type_descriptor[LOGREC_REDO_DELETE_ROW]=
INIT_LOGREC_REDO_DELETE_ROW;
log_record_type_descriptor[LOGREC_REDO_UPDATE_ROW_HEAD]=
INIT_LOGREC_REDO_UPDATE_ROW_HEAD;
log_record_type_descriptor[LOGREC_REDO_INDEX]=
INIT_LOGREC_REDO_INDEX;
log_record_type_descriptor[LOGREC_REDO_INDEX_NEW_PAGE]=
INIT_LOGREC_REDO_INDEX_NEW_PAGE;
log_record_type_descriptor[LOGREC_REDO_INDEX_FREE_PAGE]=
INIT_LOGREC_REDO_INDEX_FREE_PAGE;
log_record_type_descriptor[LOGREC_REDO_UNDELETE_ROW]=
INIT_LOGREC_REDO_UNDELETE_ROW;
log_record_type_descriptor[LOGREC_CLR_END]=
INIT_LOGREC_CLR_END;
log_record_type_descriptor[LOGREC_PURGE_END]=
INIT_LOGREC_PURGE_END;
log_record_type_descriptor[LOGREC_UNDO_ROW_INSERT]=
INIT_LOGREC_UNDO_ROW_INSERT;
log_record_type_descriptor[LOGREC_UNDO_ROW_DELETE]=
INIT_LOGREC_UNDO_ROW_DELETE;
log_record_type_descriptor[LOGREC_UNDO_ROW_UPDATE]=
INIT_LOGREC_UNDO_ROW_UPDATE;
log_record_type_descriptor[LOGREC_UNDO_KEY_INSERT]=
INIT_LOGREC_UNDO_KEY_INSERT;
log_record_type_descriptor[LOGREC_UNDO_KEY_INSERT_WITH_ROOT]=
INIT_LOGREC_UNDO_KEY_INSERT_WITH_ROOT;
log_record_type_descriptor[LOGREC_UNDO_KEY_DELETE]=
INIT_LOGREC_UNDO_KEY_DELETE;
log_record_type_descriptor[LOGREC_UNDO_KEY_DELETE_WITH_ROOT]=
INIT_LOGREC_UNDO_KEY_DELETE_WITH_ROOT;
log_record_type_descriptor[LOGREC_PREPARE]=
INIT_LOGREC_PREPARE;
log_record_type_descriptor[LOGREC_PREPARE_WITH_UNDO_PURGE]=
INIT_LOGREC_PREPARE_WITH_UNDO_PURGE;
log_record_type_descriptor[LOGREC_COMMIT]=
INIT_LOGREC_COMMIT;
log_record_type_descriptor[LOGREC_COMMIT_WITH_UNDO_PURGE]=
INIT_LOGREC_COMMIT_WITH_UNDO_PURGE;
log_record_type_descriptor[LOGREC_CHECKPOINT]=
INIT_LOGREC_CHECKPOINT;
log_record_type_descriptor[LOGREC_REDO_CREATE_TABLE]=
INIT_LOGREC_REDO_CREATE_TABLE;
log_record_type_descriptor[LOGREC_REDO_RENAME_TABLE]=
INIT_LOGREC_REDO_RENAME_TABLE;
log_record_type_descriptor[LOGREC_REDO_DROP_TABLE]=
INIT_LOGREC_REDO_DROP_TABLE;
log_record_type_descriptor[LOGREC_REDO_DELETE_ALL]=
INIT_LOGREC_REDO_DELETE_ALL;
log_record_type_descriptor[LOGREC_REDO_REPAIR_TABLE]=
INIT_LOGREC_REDO_REPAIR_TABLE;
log_record_type_descriptor[LOGREC_FILE_ID]=
INIT_LOGREC_FILE_ID;
log_record_type_descriptor[LOGREC_LONG_TRANSACTION_ID]=
INIT_LOGREC_LONG_TRANSACTION_ID;
log_record_type_descriptor[LOGREC_INCOMPLETE_LOG]=
INIT_LOGREC_INCOMPLETE_LOG;
log_record_type_descriptor[LOGREC_INCOMPLETE_GROUP]=
INIT_LOGREC_INCOMPLETE_GROUP;
log_record_type_descriptor[LOGREC_UNDO_BULK_INSERT]=
INIT_LOGREC_UNDO_BULK_INSERT;
log_record_type_descriptor[LOGREC_REDO_BITMAP_NEW_PAGE]=
INIT_LOGREC_REDO_BITMAP_NEW_PAGE;
log_record_type_descriptor[LOGREC_IMPORTED_TABLE]=
INIT_LOGREC_IMPORTED_TABLE;
log_record_type_descriptor[LOGREC_DEBUG_INFO]=
INIT_LOGREC_DEBUG_INFO;
for (i= LOGREC_FIRST_FREE; i < LOGREC_NUMBER_OF_TYPES; i++)
log_record_type_descriptor[i].rclass= LOGRECTYPE_NOT_ALLOWED;
#ifndef DBUG_OFF
check_translog_description_table(LOGREC_FIRST_FREE -1);
#endif
}
/* all possible flags page overheads */
static uint page_overhead[TRANSLOG_FLAGS_NUM];
typedef struct st_translog_validator_data
{
TRANSLOG_ADDRESS *addr;
my_bool was_recovered;
} TRANSLOG_VALIDATOR_DATA;
/*
Check cursor/buffer consistence
SYNOPSIS
translog_check_cursor
cursor cursor which will be checked
*/
static void translog_check_cursor(struct st_buffer_cursor *cursor
__attribute__((unused)))
{
DBUG_ASSERT(cursor->chaser ||
((ulong) (cursor->ptr - cursor->buffer->buffer) ==
cursor->buffer->size));
DBUG_ASSERT(cursor->buffer->buffer_no == cursor->buffer_no);
DBUG_ASSERT((cursor->ptr -cursor->buffer->buffer) %TRANSLOG_PAGE_SIZE ==
cursor->current_page_fill % TRANSLOG_PAGE_SIZE);
DBUG_ASSERT(cursor->current_page_fill <= TRANSLOG_PAGE_SIZE);
}
/**
@brief switch the loghandler in read only mode in case of write error
*/
void translog_stop_writing()
{
DBUG_ENTER("translog_stop_writing");
DBUG_PRINT("error", ("errno: %d my_errno: %d", errno, my_errno));
translog_status= (translog_status == TRANSLOG_SHUTDOWN ?
TRANSLOG_UNINITED :
TRANSLOG_READONLY);
log_descriptor.is_everything_flushed= 1;
log_descriptor.open_flags= O_BINARY | O_RDONLY;
DBUG_ASSERT(0);
DBUG_VOID_RETURN;
}
/*
@brief Get file name of the log by log number
@param file_no Number of the log we want to open
@param path Pointer to buffer where file name will be
stored (must be FN_REFLEN bytes at least)
@return pointer to path
*/
char *translog_filename_by_fileno(uint32 file_no, char *path)
{
char buff[11], *end;
uint length;
DBUG_ENTER("translog_filename_by_fileno");
DBUG_ASSERT(file_no <= 0xfffffff);
/* log_descriptor.directory is already formated */
end= strxmov(path, log_descriptor.directory, "aria_log.0000000", NullS);
length= (uint) (int10_to_str(file_no, buff, 10) - buff);
strmov(end - length +1, buff);
DBUG_PRINT("info", ("Path: '%s' path: 0x%lx", path, (ulong) path));
DBUG_RETURN(path);
}
/**
@brief Create log file with given number without cache
@param file_no Number of the log we want to open
retval -1 error
retval # file descriptor number
*/
static File create_logfile_by_number_no_cache(uint32 file_no)
{
File file;
char path[FN_REFLEN];
DBUG_ENTER("create_logfile_by_number_no_cache");
if (translog_status != TRANSLOG_OK)
DBUG_RETURN(-1);
/* TODO: add O_DIRECT to open flags (when buffer is aligned) */
if ((file= mysql_file_create(key_file_translog,
translog_filename_by_fileno(file_no, path),
0, O_BINARY | O_RDWR, MYF(MY_WME))) < 0)
{
DBUG_PRINT("error", ("Error %d during creating file '%s'", errno, path));
translog_stop_writing();
DBUG_RETURN(-1);
}
if (sync_log_dir >= TRANSLOG_SYNC_DIR_NEWFILE &&
sync_dir(log_descriptor.directory_fd, MYF(MY_WME | MY_IGNORE_BADFD)))
{
DBUG_PRINT("error", ("Error %d during syncing directory '%s'",
errno, log_descriptor.directory));
translog_stop_writing();
DBUG_RETURN(-1);
}
DBUG_PRINT("info", ("File: '%s' handler: %d", path, file));
DBUG_RETURN(file);
}
/**
@brief Open (not create) log file with given number without cache
@param file_no Number of the log we want to open
retval -1 error
retval # file descriptor number
*/
static File open_logfile_by_number_no_cache(uint32 file_no)
{
File file;
char path[FN_REFLEN];
DBUG_ENTER("open_logfile_by_number_no_cache");
/* TODO: add O_DIRECT to open flags (when buffer is aligned) */
/* TODO: use mysql_file_create() */
if ((file= mysql_file_open(key_file_translog,
translog_filename_by_fileno(file_no, path),
log_descriptor.open_flags,
MYF(MY_WME))) < 0)
{
DBUG_PRINT("error", ("Error %d during opening file '%s'", errno, path));
DBUG_RETURN(-1);
}
DBUG_PRINT("info", ("File: '%s' handler: %d", path, file));
DBUG_RETURN(file);
}
/**
@brief get file descriptor by given number using cache
@param file_no Number of the log we want to open
retval # file descriptor
retval NULL file is not opened
*/
static TRANSLOG_FILE *get_logfile_by_number(uint32 file_no)
{
TRANSLOG_FILE *file;
DBUG_ENTER("get_logfile_by_number");
mysql_rwlock_rdlock(&log_descriptor.open_files_lock);
if (log_descriptor.max_file - file_no >=
log_descriptor.open_files.elements)
{
DBUG_PRINT("info", ("File #%u is not opened", file_no));
mysql_rwlock_unlock(&log_descriptor.open_files_lock);
DBUG_RETURN(NULL);
}
DBUG_ASSERT(log_descriptor.max_file - log_descriptor.min_file + 1 ==
log_descriptor.open_files.elements);
DBUG_ASSERT(log_descriptor.max_file >= file_no);
DBUG_ASSERT(log_descriptor.min_file <= file_no);
file= *dynamic_element(&log_descriptor.open_files,
log_descriptor.max_file - file_no, TRANSLOG_FILE **);
mysql_rwlock_unlock(&log_descriptor.open_files_lock);
DBUG_PRINT("info", ("File 0x%lx File no: %lu, File handler: %d",
(ulong)file, (ulong)file_no,
(file ? file->handler.file : -1)));
DBUG_ASSERT(!file || file->number == file_no);
DBUG_RETURN(file);
}
/**
@brief get current file descriptor
retval # file descriptor
*/
static TRANSLOG_FILE *get_current_logfile()
{
TRANSLOG_FILE *file;
DBUG_ENTER("get_current_logfile");
mysql_rwlock_rdlock(&log_descriptor.open_files_lock);
DBUG_PRINT("info", ("max_file: %lu min_file: %lu open_files: %lu",
(ulong) log_descriptor.max_file,
(ulong) log_descriptor.min_file,
(ulong) log_descriptor.open_files.elements));
DBUG_ASSERT(log_descriptor.max_file - log_descriptor.min_file + 1 ==
log_descriptor.open_files.elements);
file= *dynamic_element(&log_descriptor.open_files, 0, TRANSLOG_FILE **);
mysql_rwlock_unlock(&log_descriptor.open_files_lock);
DBUG_RETURN(file);
}
uchar maria_trans_file_magic[]=
{ (uchar) 254, (uchar) 254, (uchar) 11, '\001', 'M', 'A', 'R', 'I', 'A',
'L', 'O', 'G' };
#define LOG_HEADER_DATA_SIZE (sizeof(maria_trans_file_magic) + \
8 + 4 + 4 + 4 + 2 + 3 + \
LSN_STORE_SIZE)
/*
Write log file page header in the just opened new log file
SYNOPSIS
translog_write_file_header();
NOTES
First page is just a marker page; We don't store any real log data in it.
RETURN
0 OK
1 ERROR
*/
static my_bool translog_write_file_header()
{
TRANSLOG_FILE *file;
ulonglong timestamp;
uchar page_buff[TRANSLOG_PAGE_SIZE], *page= page_buff;
my_bool rc;
DBUG_ENTER("translog_write_file_header");
/* file tag */
memcpy(page, maria_trans_file_magic, sizeof(maria_trans_file_magic));
page+= sizeof(maria_trans_file_magic);
/* timestamp */
timestamp= my_hrtime().val;
int8store(page, timestamp);
page+= 8;
/* maria version */
int4store(page, TRANSLOG_VERSION_ID);
page+= 4;
/* mysql version (MYSQL_VERSION_ID) */
int4store(page, log_descriptor.server_version);
page+= 4;
/* server ID */
int4store(page, log_descriptor.server_id);
page+= 4;
/* loghandler page_size */
int2store(page, TRANSLOG_PAGE_SIZE - 1);
page+= 2;
/* file number */
int3store(page, LSN_FILE_NO(log_descriptor.horizon));
page+= 3;
lsn_store(page, LSN_IMPOSSIBLE);
page+= LSN_STORE_SIZE;
memset(page, TRANSLOG_FILLER, sizeof(page_buff) - (page- page_buff));
file= get_current_logfile();
rc= my_pwrite(file->handler.file, page_buff, sizeof(page_buff), 0,
log_write_flags) != 0;
/*
Dropping the flag in such way can make false alarm: signalling than the
file in not sync when it is sync, but the situation is quite rare and
protections with mutexes give much more overhead to the whole engine
*/
file->is_sync= 0;
DBUG_RETURN(rc);
}
/*
@brief write the new LSN on the given file header
@param file The file descriptor
@param lsn That LSN which should be written
@retval 0 OK
@retval 1 Error
*/
static my_bool translog_max_lsn_to_header(File file, LSN lsn)
{
uchar lsn_buff[LSN_STORE_SIZE];
my_bool rc;
DBUG_ENTER("translog_max_lsn_to_header");
DBUG_PRINT("enter", ("File descriptor: %ld "
"lsn: (%lu,0x%lx)",
(long) file,
LSN_IN_PARTS(lsn)));
lsn_store(lsn_buff, lsn);
rc= (my_pwrite(file, lsn_buff,
LSN_STORE_SIZE,
(LOG_HEADER_DATA_SIZE - LSN_STORE_SIZE),
log_write_flags) != 0 ||
mysql_file_sync(file, MYF(MY_WME)) != 0);
/*
We should not increase counter in case of error above, but it is so
unlikely that we can ignore this case
*/
translog_syncs++;
DBUG_RETURN(rc);
}
/*
@brief Extract hander file information from loghandler file page
@param desc header information descriptor to be filled with information
@param page_buff buffer with the page content
*/
void translog_interpret_file_header(LOGHANDLER_FILE_INFO *desc,
uchar *page_buff)
{
uchar *ptr;
ptr= page_buff + sizeof(maria_trans_file_magic);
desc->timestamp= uint8korr(ptr);
ptr+= 8;
desc->maria_version= uint4korr(ptr);
ptr+= 4;
desc->mysql_version= uint4korr(ptr);
ptr+= 4;
desc->server_id= uint4korr(ptr + 4);
ptr+= 4;
desc->page_size= uint2korr(ptr) + 1;
ptr+= 2;
desc->file_number= uint3korr(ptr);
ptr+=3;
desc->max_lsn= lsn_korr(ptr);
}
/*
@brief Read hander file information from loghandler file
@param desc header information descriptor to be filled with information
@param file file descriptor to read
@retval 0 OK
@retval 1 Error
*/
my_bool translog_read_file_header(LOGHANDLER_FILE_INFO *desc, File file)
{
uchar page_buff[LOG_HEADER_DATA_SIZE];
DBUG_ENTER("translog_read_file_header");
if (mysql_file_pread(file, page_buff,
sizeof(page_buff), 0, MYF(MY_FNABP | MY_WME)))
{
DBUG_PRINT("info", ("log read fail error: %d", my_errno));
DBUG_RETURN(1);
}
translog_interpret_file_header(desc, page_buff);
DBUG_PRINT("info", ("timestamp: %llu aria ver: %lu mysql ver: %lu "
"server id %lu page size %lu file number %lu "
"max lsn: (%lu,0x%lx)",
(ulonglong) desc->timestamp,
(ulong) desc->maria_version,
(ulong) desc->mysql_version,
(ulong) desc->server_id,
desc->page_size, (ulong) desc->file_number,
LSN_IN_PARTS(desc->max_lsn)));
DBUG_RETURN(0);
}
/*
@brief set the lsn to the files from_file - to_file if it is greater
then written in the file
@param from_file first file number (min)
@param to_file last file number (max)
@param lsn the lsn for writing
@param is_locked true if current thread locked the log handler
@retval 0 OK
@retval 1 Error
*/
static my_bool translog_set_lsn_for_files(uint32 from_file, uint32 to_file,
LSN lsn, my_bool is_locked)
{
uint32 file;
DBUG_ENTER("translog_set_lsn_for_files");
DBUG_PRINT("enter", ("From: %lu to: %lu lsn: (%lu,0x%lx) locked: %d",
(ulong) from_file, (ulong) to_file,
LSN_IN_PARTS(lsn),
is_locked));
DBUG_ASSERT(from_file <= to_file);
DBUG_ASSERT(from_file > 0); /* we have not file 0 */
/* Checks the current file (not finished yet file) */
if (!is_locked)
translog_lock();
if (to_file == (uint32) LSN_FILE_NO(log_descriptor.horizon))
{
if (likely(cmp_translog_addr(lsn, log_descriptor.max_lsn) > 0))
log_descriptor.max_lsn= lsn;
to_file--;
}
if (!is_locked)
translog_unlock();
/* Checks finished files if they are */
mysql_mutex_lock(&log_descriptor.file_header_lock);
for (file= from_file; file <= to_file; file++)
{
LOGHANDLER_FILE_INFO info;
File fd;
LINT_INIT(info.max_lsn);
fd= open_logfile_by_number_no_cache(file);
if ((fd < 0) ||
((translog_read_file_header(&info, fd) ||
(cmp_translog_addr(lsn, info.max_lsn) > 0 &&
translog_max_lsn_to_header(fd, lsn))) |
mysql_file_close(fd, MYF(MY_WME))))
{
translog_stop_writing();
DBUG_RETURN(1);
}
}
mysql_mutex_unlock(&log_descriptor.file_header_lock);
DBUG_RETURN(0);
}
/* descriptor of file in unfinished_files */
struct st_file_counter
{
uint32 file; /* file number */
uint32 counter; /* counter for started writes */
};
/*
@brief mark file "in progress" (for multi-group records)
@param file log file number
*/
static void translog_mark_file_unfinished(uint32 file)
{
int place, i;
struct st_file_counter fc, *fc_ptr;
DBUG_ENTER("translog_mark_file_unfinished");
DBUG_PRINT("enter", ("file: %lu", (ulong) file));
fc.file= file; fc.counter= 1;
mysql_mutex_lock(&log_descriptor.unfinished_files_lock);
if (log_descriptor.unfinished_files.elements == 0)
{
insert_dynamic(&log_descriptor.unfinished_files, (uchar*) &fc);
DBUG_PRINT("info", ("The first element inserted"));
goto end;
}
for (place= log_descriptor.unfinished_files.elements - 1;
place >= 0;
place--)
{
fc_ptr= dynamic_element(&log_descriptor.unfinished_files,
place, struct st_file_counter *);
if (fc_ptr->file <= file)
break;
}
if (place >= 0 && fc_ptr->file == file)
{
fc_ptr->counter++;
DBUG_PRINT("info", ("counter increased"));
goto end;
}
if (place == (int)log_descriptor.unfinished_files.elements)
{
insert_dynamic(&log_descriptor.unfinished_files, (uchar*) &fc);
DBUG_PRINT("info", ("The last element inserted"));
goto end;
}
/* shift and assign new element */
insert_dynamic(&log_descriptor.unfinished_files,
(uchar*)
dynamic_element(&log_descriptor.unfinished_files,
log_descriptor.unfinished_files.elements- 1,
struct st_file_counter *));
for(i= log_descriptor.unfinished_files.elements - 1; i > place; i--)
{
/* we do not use set_dynamic() to avoid unneeded checks */
memcpy(dynamic_element(&log_descriptor.unfinished_files,
i, struct st_file_counter *),
dynamic_element(&log_descriptor.unfinished_files,
i + 1, struct st_file_counter *),
sizeof(struct st_file_counter));
}
memcpy(dynamic_element(&log_descriptor.unfinished_files,
place + 1, struct st_file_counter *),
&fc, sizeof(struct st_file_counter));
end:
mysql_mutex_unlock(&log_descriptor.unfinished_files_lock);
DBUG_VOID_RETURN;
}
/*
@brief remove file mark "in progress" (for multi-group records)
@param file log file number
*/
static void translog_mark_file_finished(uint32 file)
{
int i;
struct st_file_counter *fc_ptr;
DBUG_ENTER("translog_mark_file_finished");
DBUG_PRINT("enter", ("file: %lu", (ulong) file));
LINT_INIT(fc_ptr);
mysql_mutex_lock(&log_descriptor.unfinished_files_lock);
DBUG_ASSERT(log_descriptor.unfinished_files.elements > 0);
for (i= 0;
i < (int) log_descriptor.unfinished_files.elements;
i++)
{
fc_ptr= dynamic_element(&log_descriptor.unfinished_files,
i, struct st_file_counter *);
if (fc_ptr->file == file)
{
break;
}
}
DBUG_ASSERT(i < (int) log_descriptor.unfinished_files.elements);
if (! --fc_ptr->counter)
delete_dynamic_element(&log_descriptor.unfinished_files, i);
mysql_mutex_unlock(&log_descriptor.unfinished_files_lock);
DBUG_VOID_RETURN;
}
/*
@brief get max LSN of the record which parts stored in this file
@param file file number
@return requested LSN or LSN_IMPOSSIBLE/LSN_ERROR
@retval LSN_IMPOSSIBLE File is still not finished
@retval LSN_ERROR Error opening file
@retval # LSN of the record which parts stored in this file
*/
LSN translog_get_file_max_lsn_stored(uint32 file)
{
uint32 limit= FILENO_IMPOSSIBLE;
DBUG_ENTER("translog_get_file_max_lsn_stored");
DBUG_PRINT("enter", ("file: %lu", (ulong)file));
DBUG_ASSERT(translog_status == TRANSLOG_OK ||
translog_status == TRANSLOG_READONLY);
mysql_mutex_lock(&log_descriptor.unfinished_files_lock);
/* find file with minimum file number "in progress" */
if (log_descriptor.unfinished_files.elements > 0)
{
struct st_file_counter *fc_ptr;
fc_ptr= dynamic_element(&log_descriptor.unfinished_files,
0, struct st_file_counter *);
limit= fc_ptr->file; /* minimal file number "in progress" */
}
mysql_mutex_unlock(&log_descriptor.unfinished_files_lock);
/*
if there is no "in progress file" then unfinished file is in progress
for sure
*/
if (limit == FILENO_IMPOSSIBLE)
{
TRANSLOG_ADDRESS horizon= translog_get_horizon();
limit= LSN_FILE_NO(horizon);
}
if (file >= limit)
{
DBUG_PRINT("info", ("The file in in progress"));
DBUG_RETURN(LSN_IMPOSSIBLE);
}
{
LOGHANDLER_FILE_INFO info;
File fd;
fd= open_logfile_by_number_no_cache(file);
if ((fd < 0) ||
(translog_read_file_header(&info, fd) | mysql_file_close(fd, MYF(MY_WME))))
{
DBUG_PRINT("error", ("Can't read file header"));
DBUG_RETURN(LSN_ERROR);
}
DBUG_PRINT("info", ("Max lsn: (%lu,0x%lx)",
LSN_IN_PARTS(info.max_lsn)));
DBUG_RETURN(info.max_lsn);
}
}
/*
Initialize transaction log file buffer
SYNOPSIS
translog_buffer_init()
buffer The buffer to initialize
num Number of this buffer
RETURN
0 OK
1 Error
*/
static my_bool translog_buffer_init(struct st_translog_buffer *buffer, int num)
{
DBUG_ENTER("translog_buffer_init");
buffer->pre_force_close_horizon=
buffer->prev_last_lsn= buffer->last_lsn=
LSN_IMPOSSIBLE;
DBUG_PRINT("info", ("last_lsn and prev_last_lsn set to 0 buffer: 0x%lx",
(ulong) buffer));
buffer->buffer_no= (uint8) num;
/* This Buffer File */
buffer->file= NULL;
buffer->overlay= 0;
/* cache for current log */
memset(buffer->buffer, TRANSLOG_FILLER, TRANSLOG_WRITE_BUFFER);
/* Buffer size */
buffer->size= 0;
buffer->skipped_data= 0;
/* cond of thread which is waiting for buffer filling */
if (mysql_cond_init(key_TRANSLOG_BUFFER_waiting_filling_buffer,
&buffer->waiting_filling_buffer, 0))
DBUG_RETURN(1);
/* Number of records which are in copy progress */
buffer->copy_to_buffer_in_progress= 0;
/* list of waiting buffer ready threads */
buffer->waiting_flush= 0;
/*
Buffers locked by the following mutex. As far as buffers create logical
circle (after last buffer goes first) it trigger false alarm of deadlock
detect system, so we remove check of deadlock for this buffers. Indeed
all mutex locks concentrated around current buffer except flushing
thread (but it is only one thread). One thread can't take more then
2 buffer locks at once. So deadlock is impossible here.
To prevent false alarm of dead lock detection we switch dead lock
detection for one buffer in the middle of the buffers chain. Excluding
only one of eight buffers from deadlock detection hardly can hide other
possible problems which include this mutexes.
*/
if (mysql_mutex_init(key_TRANSLOG_BUFFER_mutex,
&buffer->mutex, MY_MUTEX_INIT_FAST) ||
mysql_cond_init(key_TRANSLOG_BUFFER_prev_sent_to_disk_cond,
&buffer->prev_sent_to_disk_cond, 0))
DBUG_RETURN(1);
buffer->is_closing_buffer= 0;
buffer->prev_sent_to_disk= LSN_IMPOSSIBLE;
buffer->prev_buffer_offset= LSN_IMPOSSIBLE;
buffer->ver= 0;
DBUG_RETURN(0);
}
/*
@brief close transaction log file by descriptor
@param file pagegecache file descriptor reference
@return Operation status
@retval 0 OK
@retval 1 Error
*/
static my_bool translog_close_log_file(TRANSLOG_FILE *file)
{
int rc= 0;
flush_pagecache_blocks(log_descriptor.pagecache, &file->handler,
FLUSH_RELEASE);
/*
Sync file when we close it
TODO: sync only we have changed the log
*/
if (!file->is_sync)
{
rc= mysql_file_sync(file->handler.file, MYF(MY_WME));
translog_syncs++;
}
rc|= mysql_file_close(file->handler.file, MYF(MY_WME));
my_free(file);
return test(rc);
}
/**
@brief Dummy function for write failure (the log to not use
pagecache writing)
*/
void translog_dummy_write_failure(uchar *data __attribute__((unused)))
{
return;
}
/**
@brief Initializes TRANSLOG_FILE structure
@param file reference on the file to initialize
@param number file number
@param is_sync is file synced on disk
*/
static void translog_file_init(TRANSLOG_FILE *file, uint32 number,
my_bool is_sync)
{
pagecache_file_init(file->handler, &translog_page_validator,
&translog_dummy_callback,
&translog_dummy_write_failure,
maria_flush_log_for_page_none, file);
file->number= number;
file->was_recovered= 0;
file->is_sync= is_sync;
}
/**
@brief Create and fill header of new file.
@note the caller must call it right after it has increased
log_descriptor.horizon to the new file
(log_descriptor.horizon+= LSN_ONE_FILE)
@retval 0 OK
@retval 1 Error
*/
static my_bool translog_create_new_file()
{
TRANSLOG_FILE *file= (TRANSLOG_FILE*)my_malloc(sizeof(TRANSLOG_FILE),
MYF(0));
TRANSLOG_FILE *old= get_current_logfile();
uint32 file_no= LSN_FILE_NO(log_descriptor.horizon);
DBUG_ENTER("translog_create_new_file");
if (file == NULL)
goto error;
/*
Writes max_lsn to the file header before finishing it (there is no need
to lock file header buffer because it is still unfinished file, so only
one thread can finish the file and nobody interested of LSN of current
(unfinished) file, because no one can purge it).
*/
if (translog_max_lsn_to_header(old->handler.file, log_descriptor.max_lsn))
goto error;
mysql_rwlock_wrlock(&log_descriptor.open_files_lock);
DBUG_ASSERT(log_descriptor.max_file - log_descriptor.min_file + 1 ==
log_descriptor.open_files.elements);
DBUG_ASSERT(file_no == log_descriptor.max_file + 1);
if (allocate_dynamic(&log_descriptor.open_files,
log_descriptor.max_file - log_descriptor.min_file + 2))
goto error_lock;
if ((file->handler.file=
create_logfile_by_number_no_cache(file_no)) == -1)
goto error_lock;
translog_file_init(file, file_no, 0);
/* this call just expand the array */
insert_dynamic(&log_descriptor.open_files, (uchar*)&file);
log_descriptor.max_file++;
{
char *start= (char*) dynamic_element(&log_descriptor.open_files, 0,
TRANSLOG_FILE**);
memmove(start + sizeof(TRANSLOG_FILE*), start,
sizeof(TRANSLOG_FILE*) *
(log_descriptor.max_file - log_descriptor.min_file + 1 - 1));
}
/* can't fail we because we expanded array */
set_dynamic(&log_descriptor.open_files, (uchar*)&file, 0);
DBUG_ASSERT(log_descriptor.max_file - log_descriptor.min_file + 1 ==
log_descriptor.open_files.elements);
mysql_rwlock_unlock(&log_descriptor.open_files_lock);
DBUG_PRINT("info", ("file_no: %lu", (ulong)file_no));
if (translog_write_file_header())
DBUG_RETURN(1);
if (ma_control_file_write_and_force(last_checkpoint_lsn, file_no,
max_trid_in_control_file,
recovery_failures))
{
translog_stop_writing();
DBUG_RETURN(1);
}
DBUG_RETURN(0);
error_lock:
mysql_rwlock_unlock(&log_descriptor.open_files_lock);
error:
translog_stop_writing();
DBUG_RETURN(1);
}
/**
@brief Locks the loghandler buffer.
@param buffer This buffer which should be locked
@note See comment before buffer 'mutex' variable.
@retval 0 OK
@retval 1 Error
*/
static void translog_buffer_lock(struct st_translog_buffer *buffer)
{
DBUG_ENTER("translog_buffer_lock");
DBUG_PRINT("enter",
("Lock buffer #%u: (0x%lx)", (uint) buffer->buffer_no,
(ulong) buffer));
mysql_mutex_lock(&buffer->mutex);
DBUG_VOID_RETURN;
}
/*
Unlock the loghandler buffer
SYNOPSIS
translog_buffer_unlock()
buffer This buffer which should be unlocked
RETURN
0 OK
1 Error
*/
static void translog_buffer_unlock(struct st_translog_buffer *buffer)
{
DBUG_ENTER("translog_buffer_unlock");
DBUG_PRINT("enter", ("Unlock buffer... #%u (0x%lx)",
(uint) buffer->buffer_no, (ulong) buffer));
mysql_mutex_unlock(&buffer->mutex);
DBUG_VOID_RETURN;
}
/*
Write a header on the page
SYNOPSIS
translog_new_page_header()
horizon Where to write the page
cursor Where to write the page
NOTE
- space for page header should be checked before
*/
static uchar translog_sector_random;
static void translog_new_page_header(TRANSLOG_ADDRESS *horizon,
struct st_buffer_cursor *cursor)
{
uchar *ptr;
DBUG_ENTER("translog_new_page_header");
DBUG_ASSERT(cursor->ptr);
cursor->protected= 0;
ptr= cursor->ptr;
/* Page number */
int3store(ptr, LSN_OFFSET(*horizon) / TRANSLOG_PAGE_SIZE);
ptr+= 3;
/* File number */
int3store(ptr, LSN_FILE_NO(*horizon));
ptr+= 3;
DBUG_ASSERT(TRANSLOG_PAGE_FLAGS == (ptr - cursor->ptr));
cursor->ptr[TRANSLOG_PAGE_FLAGS]= (uchar) log_descriptor.flags;
ptr++;
if (log_descriptor.flags & TRANSLOG_PAGE_CRC)
{
#ifndef DBUG_OFF
DBUG_PRINT("info", ("write 0x11223344 CRC to (%lu,0x%lx)",
LSN_IN_PARTS(*horizon)));
/* This will be overwritten by real CRC; This is just for debugging */
int4store(ptr, 0x11223344);
#endif
/* CRC will be put when page is finished */
ptr+= CRC_SIZE;
}
if (log_descriptor.flags & TRANSLOG_SECTOR_PROTECTION)
{
/*
translog_sector_randmo works like "random" values producer because
it is enough to have such "random" for this purpose and it will
not interfere with higher level pseudo random value generator
*/
ptr[0]= translog_sector_random++;
ptr+= TRANSLOG_PAGE_SIZE / DISK_DRIVE_SECTOR_SIZE;
}
{
uint len= (ptr - cursor->ptr);
(*horizon)+= len; /* increasing the offset part of the address */
cursor->current_page_fill= len;
if (!cursor->chaser)
cursor->buffer->size+= len;
}
cursor->ptr= ptr;
DBUG_PRINT("info", ("NewP buffer #%u: 0x%lx chaser: %d Size: %lu (%lu) "
"Horizon: (%lu,0x%lx)",
(uint) cursor->buffer->buffer_no, (ulong) cursor->buffer,
cursor->chaser, (ulong) cursor->buffer->size,
(ulong) (cursor->ptr - cursor->buffer->buffer),
LSN_IN_PARTS(*horizon)));
translog_check_cursor(cursor);
DBUG_VOID_RETURN;
}
/*
Put sector protection on the page image
SYNOPSIS
translog_put_sector_protection()
page reference on the page content
cursor cursor of the buffer
NOTES
We put a sector protection on all following sectors on the page,
except the first sector that is protected by page header.
*/
static void translog_put_sector_protection(uchar *page,
struct st_buffer_cursor *cursor)
{
uchar *table= page + log_descriptor.page_overhead -
TRANSLOG_PAGE_SIZE / DISK_DRIVE_SECTOR_SIZE;
uint i, offset;
uint16 last_protected_sector= ((cursor->previous_offset - 1) /
DISK_DRIVE_SECTOR_SIZE);
uint16 start_sector= cursor->previous_offset / DISK_DRIVE_SECTOR_SIZE;
uint8 value= table[0] + cursor->write_counter;
DBUG_ENTER("translog_put_sector_protection");
if (start_sector == 0)
{
/* First sector is protected by file & page numbers in the page header. */
start_sector= 1;
}
DBUG_PRINT("enter", ("Write counter:%u value:%u offset:%u, "
"last protected:%u start sector:%u",
(uint) cursor->write_counter,
(uint) value,
(uint) cursor->previous_offset,
(uint) last_protected_sector, (uint) start_sector));
if (last_protected_sector == start_sector)
{
i= last_protected_sector;
offset= last_protected_sector * DISK_DRIVE_SECTOR_SIZE;
/* restore data, because we modified sector which was protected */
if (offset < cursor->previous_offset)
page[offset]= table[i];
}
for (i= start_sector, offset= start_sector * DISK_DRIVE_SECTOR_SIZE;
i < TRANSLOG_PAGE_SIZE / DISK_DRIVE_SECTOR_SIZE;
i++, (offset+= DISK_DRIVE_SECTOR_SIZE))
{
DBUG_PRINT("info", ("sector:%u offset:%u data 0x%x",
i, offset, (uint) page[offset]));
table[i]= page[offset];
page[offset]= value;
DBUG_PRINT("info", ("sector:%u offset:%u data 0x%x",
i, offset, (uint) page[offset]));
}
DBUG_VOID_RETURN;
}
/*
Calculate CRC32 of given area
SYNOPSIS
translog_crc()
area Pointer of the area beginning
length The Area length
RETURN
CRC32
*/
static uint32 translog_crc(uchar *area, uint length)
{
DBUG_ENTER("translog_crc");
DBUG_RETURN(crc32(0L, (unsigned char*) area, length));
}
/*
Finish current page with zeros
SYNOPSIS
translog_finish_page()
horizon \ horizon & buffer pointers
cursor /
*/
static void translog_finish_page(TRANSLOG_ADDRESS *horizon,
struct st_buffer_cursor *cursor)
{
uint16 left= TRANSLOG_PAGE_SIZE - cursor->current_page_fill;
uchar *page= cursor->ptr - cursor->current_page_fill;
DBUG_ENTER("translog_finish_page");
DBUG_PRINT("enter", ("Buffer: #%u 0x%lx "
"Buffer addr: (%lu,0x%lx) "
"Page addr: (%lu,0x%lx) "
"size:%lu (%lu) Pg:%u left:%u",
(uint) cursor->buffer_no, (ulong) cursor->buffer,
LSN_IN_PARTS(cursor->buffer->offset),
(ulong) LSN_FILE_NO(*horizon),
(ulong) (LSN_OFFSET(*horizon) -
cursor->current_page_fill),
(ulong) cursor->buffer->size,
(ulong) (cursor->ptr -cursor->buffer->buffer),
(uint) cursor->current_page_fill, (uint) left));
DBUG_ASSERT(LSN_FILE_NO(*horizon) == LSN_FILE_NO(cursor->buffer->offset));
translog_check_cursor(cursor);
if (cursor->protected)
{
DBUG_PRINT("info", ("Already protected and finished"));
DBUG_VOID_RETURN;
}
cursor->protected= 1;
DBUG_ASSERT(left < TRANSLOG_PAGE_SIZE);
if (left != 0)
{
DBUG_PRINT("info", ("left: %u", (uint) left));
memset(cursor->ptr, TRANSLOG_FILLER, left);
cursor->ptr+= left;
(*horizon)+= left; /* offset increasing */
if (!cursor->chaser)
cursor->buffer->size+= left;
/* We are finishing the page so reset the counter */
cursor->current_page_fill= 0;
DBUG_PRINT("info", ("Finish Page buffer #%u: 0x%lx "
"chaser: %d Size: %lu (%lu)",
(uint) cursor->buffer->buffer_no,
(ulong) cursor->buffer, cursor->chaser,
(ulong) cursor->buffer->size,
(ulong) (cursor->ptr - cursor->buffer->buffer)));
translog_check_cursor(cursor);
}
/*
When we are finishing the page other thread might not finish the page
header yet (in case if we started from the middle of the page) so we
have to read log_descriptor.flags but not the flags from the page.
*/
if (log_descriptor.flags & TRANSLOG_SECTOR_PROTECTION)
{
translog_put_sector_protection(page, cursor);
DBUG_PRINT("info", ("drop write_counter"));
cursor->write_counter= 0;
cursor->previous_offset= 0;
}
if (log_descriptor.flags & TRANSLOG_PAGE_CRC)
{
uint32 crc= translog_crc(page + log_descriptor.page_overhead,
TRANSLOG_PAGE_SIZE -
log_descriptor.page_overhead);
DBUG_PRINT("info", ("CRC: %lx", (ulong) crc));
/* We have page number, file number and flag before crc */
int4store(page + 3 + 3 + 1, crc);
}
DBUG_VOID_RETURN;
}
/*
@brief Wait until all threads have finished closing this buffer.
@param buffer This buffer should be check
*/
static void translog_wait_for_closing(struct st_translog_buffer *buffer)
{
DBUG_ENTER("translog_wait_for_closing");
DBUG_PRINT("enter", ("Buffer #%u 0x%lx copies in progress: %u "
"is closing %u File: %d size: %lu",
(uint) buffer->buffer_no, (ulong) buffer,
(uint) buffer->copy_to_buffer_in_progress,
(uint) buffer->is_closing_buffer,
(buffer->file ? buffer->file->handler.file : -1),
(ulong) buffer->size));
translog_buffer_lock_assert_owner(buffer);
while (buffer->is_closing_buffer)
{
DBUG_PRINT("info", ("wait for writers... buffer: #%u 0x%lx",
(uint) buffer->buffer_no, (ulong) buffer));
DBUG_ASSERT(buffer->file != NULL);
mysql_cond_wait(&buffer->waiting_filling_buffer, &buffer->mutex);
DBUG_PRINT("info", ("wait for writers done buffer: #%u 0x%lx",
(uint) buffer->buffer_no, (ulong) buffer));
}
DBUG_VOID_RETURN;
}
/*
@brief Wait until all threads have finished filling this buffer.
@param buffer This buffer should be check
*/
static void translog_wait_for_writers(struct st_translog_buffer *buffer)
{
DBUG_ENTER("translog_wait_for_writers");
DBUG_PRINT("enter", ("Buffer #%u 0x%lx copies in progress: %u "
"is closing %u File: %d size: %lu",
(uint) buffer->buffer_no, (ulong) buffer,
(uint) buffer->copy_to_buffer_in_progress,
(uint) buffer->is_closing_buffer,
(buffer->file ? buffer->file->handler.file : -1),
(ulong) buffer->size));
translog_buffer_lock_assert_owner(buffer);
while (buffer->copy_to_buffer_in_progress)
{
DBUG_PRINT("info", ("wait for writers... buffer: #%u 0x%lx",
(uint) buffer->buffer_no, (ulong) buffer));
DBUG_ASSERT(buffer->file != NULL);
mysql_cond_wait(&buffer->waiting_filling_buffer, &buffer->mutex);
DBUG_PRINT("info", ("wait for writers done buffer: #%u 0x%lx",
(uint) buffer->buffer_no, (ulong) buffer));
}
DBUG_VOID_RETURN;
}
/*
Wait for buffer to become free
SYNOPSIS
translog_wait_for_buffer_free()
buffer The buffer we are waiting for
NOTE
- this buffer should be locked
*/
static void translog_wait_for_buffer_free(struct st_translog_buffer *buffer)
{
TRANSLOG_ADDRESS offset= buffer->offset;
TRANSLOG_FILE *file= buffer->file;
uint8 ver= buffer->ver;
DBUG_ENTER("translog_wait_for_buffer_free");
DBUG_PRINT("enter", ("Buffer #%u 0x%lx copies in progress: %u "
"is closing %u File: %d size: %lu",
(uint) buffer->buffer_no, (ulong) buffer,
(uint) buffer->copy_to_buffer_in_progress,
(uint) buffer->is_closing_buffer,
(buffer->file ? buffer->file->handler.file : -1),
(ulong) buffer->size));
translog_wait_for_writers(buffer);
if (offset != buffer->offset || file != buffer->file || ver != buffer->ver)
DBUG_VOID_RETURN; /* the buffer if already freed */
while (buffer->file != NULL)
{
DBUG_PRINT("info", ("wait for writers... buffer: #%u 0x%lx",
(uint) buffer->buffer_no, (ulong) buffer));
mysql_cond_wait(&buffer->waiting_filling_buffer, &buffer->mutex);
DBUG_PRINT("info", ("wait for writers done. buffer: #%u 0x%lx",
(uint) buffer->buffer_no, (ulong) buffer));
}
DBUG_ASSERT(buffer->copy_to_buffer_in_progress == 0);
DBUG_VOID_RETURN;
}
/*
Initialize the cursor for a buffer
SYNOPSIS
translog_cursor_init()
buffer The buffer
cursor It's cursor
buffer_no Number of buffer
*/
static void translog_cursor_init(struct st_buffer_cursor *cursor,
struct st_translog_buffer *buffer,
uint8 buffer_no)
{
DBUG_ENTER("translog_cursor_init");
cursor->ptr= buffer->buffer;
cursor->buffer= buffer;
cursor->buffer_no= buffer_no;
cursor->current_page_fill= 0;
cursor->chaser= (cursor != &log_descriptor.bc);
cursor->write_counter= 0;
cursor->previous_offset= 0;
cursor->protected= 0;
DBUG_VOID_RETURN;
}
/*
@brief Initialize buffer for the current file, and a cursor for this buffer.
@param buffer The buffer
@param cursor It's cursor
@param buffer_no Number of buffer
*/
static void translog_start_buffer(struct st_translog_buffer *buffer,
struct st_buffer_cursor *cursor,
uint buffer_no)
{
DBUG_ENTER("translog_start_buffer");
DBUG_PRINT("enter",
("Assign buffer: #%u (0x%lx) offset: 0x%lx(%lu)",
(uint) buffer->buffer_no, (ulong) buffer,
(ulong) LSN_OFFSET(log_descriptor.horizon),
(ulong) LSN_OFFSET(log_descriptor.horizon)));
DBUG_ASSERT(buffer_no == buffer->buffer_no);
buffer->pre_force_close_horizon=
buffer->prev_last_lsn= buffer->last_lsn= LSN_IMPOSSIBLE;
DBUG_PRINT("info", ("last_lsn and prev_last_lsn set to 0 buffer: 0x%lx",
(ulong) buffer));
buffer->offset= log_descriptor.horizon;
buffer->next_buffer_offset= LSN_IMPOSSIBLE;
buffer->file= get_current_logfile();
buffer->overlay= 0;
buffer->size= 0;
buffer->skipped_data= 0;
translog_cursor_init(cursor, buffer, buffer_no);
DBUG_PRINT("info", ("file: #%ld (%d) init cursor #%u: 0x%lx "
"chaser: %d Size: %lu (%lu)",
(long) (buffer->file ? buffer->file->number : 0),
(buffer->file ? buffer->file->handler.file : -1),
(uint) cursor->buffer->buffer_no, (ulong) cursor->buffer,
cursor->chaser, (ulong) cursor->buffer->size,
(ulong) (cursor->ptr - cursor->buffer->buffer)));
translog_check_cursor(cursor);
mysql_mutex_lock(&log_descriptor.dirty_buffer_mask_lock);
log_descriptor.dirty_buffer_mask|= (1 << buffer->buffer_no);
mysql_mutex_unlock(&log_descriptor.dirty_buffer_mask_lock);
DBUG_VOID_RETURN;
}
/*
@brief Switch to the next buffer in a chain.
@param horizon \ Pointers on current position in file and buffer
@param cursor /
@param new_file Also start new file
@note
- loghandler should be locked
- after return new and old buffer still are locked
@retval 0 OK
@retval 1 Error
*/
static my_bool translog_buffer_next(TRANSLOG_ADDRESS *horizon,
struct st_buffer_cursor *cursor,
my_bool new_file)
{
uint old_buffer_no= cursor->buffer_no;
uint new_buffer_no= (old_buffer_no + 1) % TRANSLOG_BUFFERS_NO;
struct st_translog_buffer *new_buffer= log_descriptor.buffers + new_buffer_no;
my_bool chasing= cursor->chaser;
DBUG_ENTER("translog_buffer_next");
DBUG_PRINT("info", ("horizon: (%lu,0x%lx) chasing: %d",
LSN_IN_PARTS(log_descriptor.horizon), chasing));
DBUG_ASSERT(cmp_translog_addr(log_descriptor.horizon, *horizon) >= 0);
translog_finish_page(horizon, cursor);
if (!chasing)
{
translog_buffer_lock(new_buffer);
#ifndef DBUG_OFF
{
TRANSLOG_ADDRESS offset= new_buffer->offset;
TRANSLOG_FILE *file= new_buffer->file;
uint8 ver= new_buffer->ver;
translog_lock_assert_owner();
#endif
translog_wait_for_buffer_free(new_buffer);
#ifndef DBUG_OFF
/* We keep the handler locked so nobody can start this new buffer */
DBUG_ASSERT(offset == new_buffer->offset && new_buffer->file == NULL &&
(file == NULL ? ver : (uint8)(ver + 1)) == new_buffer->ver);
}
#endif
}
else
DBUG_ASSERT(new_buffer->file != NULL);
if (new_file)
{
/* move the horizon to the next file and its header page */
(*horizon)+= LSN_ONE_FILE;
(*horizon)= LSN_REPLACE_OFFSET(*horizon, TRANSLOG_PAGE_SIZE);
if (!chasing && translog_create_new_file())
{
DBUG_RETURN(1);
}
}
/* prepare next page */
if (chasing)
translog_cursor_init(cursor, new_buffer, new_buffer_no);
else
{
translog_lock_assert_owner();
translog_start_buffer(new_buffer, cursor, new_buffer_no);
new_buffer->prev_buffer_offset=
log_descriptor.buffers[old_buffer_no].offset;
new_buffer->prev_last_lsn=
BUFFER_MAX_LSN(log_descriptor.buffers + old_buffer_no);
}
log_descriptor.buffers[old_buffer_no].next_buffer_offset= new_buffer->offset;
DBUG_PRINT("info", ("prev_last_lsn set to (%lu,0x%lx) buffer: 0x%lx",
LSN_IN_PARTS(new_buffer->prev_last_lsn),
(ulong) new_buffer));
translog_new_page_header(horizon, cursor);
DBUG_RETURN(0);
}
/*
Sets max LSN sent to file, and address from which data is only in the buffer
SYNOPSIS
translog_set_sent_to_disk()
buffer buffer which we have sent to disk
TODO: use atomic operations if possible (64bit architectures?)
*/
static void translog_set_sent_to_disk(struct st_translog_buffer *buffer)
{
LSN lsn= buffer->last_lsn;
TRANSLOG_ADDRESS in_buffers= buffer->next_buffer_offset;
DBUG_ENTER("translog_set_sent_to_disk");
mysql_mutex_lock(&log_descriptor.sent_to_disk_lock);
DBUG_PRINT("enter", ("lsn: (%lu,0x%lx) in_buffers: (%lu,0x%lx) "
"in_buffers_only: (%lu,0x%lx) start: (%lu,0x%lx) "
"sent_to_disk: (%lu,0x%lx)",
LSN_IN_PARTS(lsn),
LSN_IN_PARTS(in_buffers),
LSN_IN_PARTS(log_descriptor.log_start),
LSN_IN_PARTS(log_descriptor.in_buffers_only),
LSN_IN_PARTS(log_descriptor.sent_to_disk)));
/*
We write sequentially (first part of following assert) but we rewrite
the same page in case we started mysql and shut it down immediately
(second part of the following assert)
*/
DBUG_ASSERT(cmp_translog_addr(lsn, log_descriptor.sent_to_disk) >= 0 ||
cmp_translog_addr(lsn, log_descriptor.log_start) < 0);
log_descriptor.sent_to_disk= lsn;
/* LSN_IMPOSSIBLE == 0 => it will work for very first time */
if (cmp_translog_addr(in_buffers, log_descriptor.in_buffers_only) > 0)
{
log_descriptor.in_buffers_only= in_buffers;
DBUG_PRINT("info", ("set new in_buffers_only"));
}
mysql_mutex_unlock(&log_descriptor.sent_to_disk_lock);
DBUG_VOID_RETURN;
}
/*
Sets address from which data is only in the buffer
SYNOPSIS
translog_set_only_in_buffers()
lsn LSN to assign
in_buffers to assign to in_buffers_only
*/
static void translog_set_only_in_buffers(TRANSLOG_ADDRESS in_buffers)
{
DBUG_ENTER("translog_set_only_in_buffers");
mysql_mutex_lock(&log_descriptor.sent_to_disk_lock);
DBUG_PRINT("enter", ("in_buffers: (%lu,0x%lx) "
"in_buffers_only: (%lu,0x%lx)",
LSN_IN_PARTS(in_buffers),
LSN_IN_PARTS(log_descriptor.in_buffers_only)));
/* LSN_IMPOSSIBLE == 0 => it will work for very first time */
if (cmp_translog_addr(in_buffers, log_descriptor.in_buffers_only) > 0)
{
if (translog_status != TRANSLOG_OK)
DBUG_VOID_RETURN;
log_descriptor.in_buffers_only= in_buffers;
DBUG_PRINT("info", ("set new in_buffers_only"));
}
mysql_mutex_unlock(&log_descriptor.sent_to_disk_lock);
DBUG_VOID_RETURN;
}
/*
Gets address from which data is only in the buffer
SYNOPSIS
translog_only_in_buffers()
RETURN
address from which data is only in the buffer
*/
static TRANSLOG_ADDRESS translog_only_in_buffers()
{
register TRANSLOG_ADDRESS addr;
DBUG_ENTER("translog_only_in_buffers");
mysql_mutex_lock(&log_descriptor.sent_to_disk_lock);
addr= log_descriptor.in_buffers_only;
mysql_mutex_unlock(&log_descriptor.sent_to_disk_lock);
DBUG_RETURN(addr);
}
/*
Get max LSN sent to file
SYNOPSIS
translog_get_sent_to_disk()
RETURN
max LSN send to file
*/
static LSN translog_get_sent_to_disk()
{
register LSN lsn;
DBUG_ENTER("translog_get_sent_to_disk");
mysql_mutex_lock(&log_descriptor.sent_to_disk_lock);
lsn= log_descriptor.sent_to_disk;
DBUG_PRINT("info", ("sent to disk up to (%lu,0x%lx)", LSN_IN_PARTS(lsn)));
mysql_mutex_unlock(&log_descriptor.sent_to_disk_lock);
DBUG_RETURN(lsn);
}
/*
Get first chunk address on the given page
SYNOPSIS
translog_get_first_chunk_offset()
page The page where to find first chunk
RETURN
first chunk offset
*/
static my_bool translog_get_first_chunk_offset(uchar *page)
{
DBUG_ENTER("translog_get_first_chunk_offset");
DBUG_ASSERT(page[TRANSLOG_PAGE_FLAGS] < TRANSLOG_FLAGS_NUM);
DBUG_RETURN(page_overhead[page[TRANSLOG_PAGE_FLAGS]]);
}
/*
Write coded length of record
SYNOPSIS
translog_write_variable_record_1group_code_len
dst Destination buffer pointer
length Length which should be coded
header_len Calculated total header length
*/
static void
translog_write_variable_record_1group_code_len(uchar *dst,
translog_size_t length,
uint16 header_len)
{
switch (header_len) {
case 6: /* (5 + 1) */
DBUG_ASSERT(length <= 250);
*dst= (uint8) length;
return;
case 8: /* (5 + 3) */
DBUG_ASSERT(length <= 0xFFFF);
*dst= 251;
int2store(dst + 1, length);
return;
case 9: /* (5 + 4) */
DBUG_ASSERT(length <= (ulong) 0xFFFFFF);
*dst= 252;
int3store(dst + 1, length);
return;
case 10: /* (5 + 5) */
*dst= 253;
int4store(dst + 1, length);
return;
default:
DBUG_ASSERT(0);
}
return;
}
/*
Decode record data length and advance given pointer to the next field
SYNOPSIS
translog_variable_record_1group_decode_len()
src The pointer to the pointer to the length beginning
RETURN
decoded length
*/
static translog_size_t translog_variable_record_1group_decode_len(uchar **src)
{
uint8 first= (uint8) (**src);
switch (first) {
case 251:
(*src)+= 3;
return (uint2korr((*src) - 2));
case 252:
(*src)+= 4;
return (uint3korr((*src) - 3));
case 253:
(*src)+= 5;
return (uint4korr((*src) - 4));
case 254:
case 255:
DBUG_ASSERT(0); /* reserved for future use */
return (0);
default:
(*src)++;
return (first);
}
}
/*
Get total length of this chunk (not only body)
SYNOPSIS
translog_get_total_chunk_length()
page The page where chunk placed
offset Offset of the chunk on this place
RETURN
total length of the chunk
*/
static uint16 translog_get_total_chunk_length(uchar *page, uint16 offset)
{
DBUG_ENTER("translog_get_total_chunk_length");
switch (page[offset] & TRANSLOG_CHUNK_TYPE) {
case TRANSLOG_CHUNK_LSN:
{
/* 0 chunk referred as LSN (head or tail) */
translog_size_t rec_len;
uchar *start= page + offset;
uchar *ptr= start + 1 + 2; /* chunk type and short trid */
uint16 chunk_len, header_len, page_rest;
DBUG_PRINT("info", ("TRANSLOG_CHUNK_LSN"));
rec_len= translog_variable_record_1group_decode_len(&ptr);
chunk_len= uint2korr(ptr);
header_len= (uint16) (ptr -start) + 2;
DBUG_PRINT("info", ("rec len: %lu chunk len: %u header len: %u",
(ulong) rec_len, (uint) chunk_len, (uint) header_len));
if (chunk_len)
{
DBUG_PRINT("info", ("chunk len: %u + %u = %u",
(uint) header_len, (uint) chunk_len,
(uint) (chunk_len + header_len)));
DBUG_RETURN(chunk_len + header_len);
}
page_rest= TRANSLOG_PAGE_SIZE - offset;
DBUG_PRINT("info", ("page_rest %u", (uint) page_rest));
if (rec_len + header_len < page_rest)
DBUG_RETURN(rec_len + header_len);
DBUG_RETURN(page_rest);
}
case TRANSLOG_CHUNK_FIXED:
{
uchar *ptr;
uint type= page[offset] & TRANSLOG_REC_TYPE;
uint length;
int i;
/* 1 (pseudo)fixed record (also LSN) */
DBUG_PRINT("info", ("TRANSLOG_CHUNK_FIXED"));
DBUG_ASSERT(log_record_type_descriptor[type].rclass ==
LOGRECTYPE_FIXEDLENGTH ||
log_record_type_descriptor[type].rclass ==
LOGRECTYPE_PSEUDOFIXEDLENGTH);
if (log_record_type_descriptor[type].rclass == LOGRECTYPE_FIXEDLENGTH)
{
DBUG_PRINT("info",
("Fixed length: %u",
(uint) (log_record_type_descriptor[type].fixed_length + 3)));
DBUG_RETURN(log_record_type_descriptor[type].fixed_length + 3);
}
ptr= page + offset + 3; /* first compressed LSN */
length= log_record_type_descriptor[type].fixed_length + 3;
for (i= 0; i < log_record_type_descriptor[type].compressed_LSN; i++)
{
/* first 2 bits is length - 2 */
uint len= (((uint8) (*ptr)) >> 6) + 2;
if (ptr[0] == 0 && ((uint8) ptr[1]) == 1)
len+= LSN_STORE_SIZE; /* case of full LSN storing */
ptr+= len;
/* subtract saved bytes */
length-= (LSN_STORE_SIZE - len);
}
DBUG_PRINT("info", ("Pseudo-fixed length: %u", length));
DBUG_RETURN(length);
}
case TRANSLOG_CHUNK_NOHDR:
/* 2 no header chunk (till page end) */
DBUG_PRINT("info", ("TRANSLOG_CHUNK_NOHDR length: %u",
(uint) (TRANSLOG_PAGE_SIZE - offset)));
DBUG_RETURN(TRANSLOG_PAGE_SIZE - offset);
case TRANSLOG_CHUNK_LNGTH: /* 3 chunk with chunk length */
DBUG_PRINT("info", ("TRANSLOG_CHUNK_LNGTH"));
DBUG_ASSERT(TRANSLOG_PAGE_SIZE - offset >= 3);
DBUG_PRINT("info", ("length: %u", uint2korr(page + offset + 1) + 3));
DBUG_RETURN(uint2korr(page + offset + 1) + 3);
default:
DBUG_ASSERT(0);
DBUG_RETURN(0);
}
}
/*
@brief Waits previous buffer flush finish
@param buffer buffer for check
@retval 0 previous buffer flushed and this thread have to flush this one
@retval 1 previous buffer flushed and this buffer flushed by other thread too
*/
my_bool translog_prev_buffer_flush_wait(struct st_translog_buffer *buffer)
{
TRANSLOG_ADDRESS offset= buffer->offset;
TRANSLOG_FILE *file= buffer->file;
uint8 ver= buffer->ver;
DBUG_ENTER("translog_prev_buffer_flush_wait");
DBUG_PRINT("enter", ("buffer: 0x%lx #%u offset: (%lu,0x%lx) "
"prev sent: (%lu,0x%lx) prev offset: (%lu,0x%lx)",
(ulong) buffer, (uint) buffer->buffer_no,
LSN_IN_PARTS(buffer->offset),
LSN_IN_PARTS(buffer->prev_sent_to_disk),
LSN_IN_PARTS(buffer->prev_buffer_offset)));
translog_buffer_lock_assert_owner(buffer);
if (buffer->prev_buffer_offset != buffer->prev_sent_to_disk)
{
do {
mysql_cond_wait(&buffer->prev_sent_to_disk_cond, &buffer->mutex);
if (buffer->file != file || buffer->offset != offset ||
buffer->ver != ver)
DBUG_RETURN(1); /* some the thread flushed the buffer already */
} while(buffer->prev_buffer_offset != buffer->prev_sent_to_disk);
}
DBUG_RETURN(0);
}
/*
Flush given buffer
SYNOPSIS
translog_buffer_flush()
buffer This buffer should be flushed
RETURN
0 OK
1 Error
*/
static my_bool translog_buffer_flush(struct st_translog_buffer *buffer)
{
uint32 i, pg;
TRANSLOG_ADDRESS offset= buffer->offset;
TRANSLOG_FILE *file= buffer->file;
uint8 ver= buffer->ver;
uint skipped_data;
DBUG_ENTER("translog_buffer_flush");
DBUG_PRINT("enter",
("Buffer: #%u 0x%lx file: %d offset: (%lu,0x%lx) size: %lu",
(uint) buffer->buffer_no, (ulong) buffer,
buffer->file->handler.file,
LSN_IN_PARTS(buffer->offset),
(ulong) buffer->size));
translog_buffer_lock_assert_owner(buffer);
if (buffer->file == NULL)
DBUG_RETURN(0);
translog_wait_for_writers(buffer);
if (buffer->file != file || buffer->offset != offset || buffer->ver != ver)
DBUG_RETURN(0); /* some the thread flushed the buffer already */
if (buffer->is_closing_buffer)
{
/* some other flush in progress */
translog_wait_for_closing(buffer);
if (buffer->file != file || buffer->offset != offset || buffer->ver != ver)
DBUG_RETURN(0); /* some the thread flushed the buffer already */
}
if (buffer->overlay && translog_prev_buffer_flush_wait(buffer))
DBUG_RETURN(0); /* some the thread flushed the buffer already */
/*
Send page by page in the pagecache what we are going to write on the
disk
*/
file= buffer->file;
skipped_data= buffer->skipped_data;
DBUG_ASSERT(skipped_data < TRANSLOG_PAGE_SIZE);
for (i= 0, pg= LSN_OFFSET(buffer->offset) / TRANSLOG_PAGE_SIZE;
i < buffer->size;
i+= TRANSLOG_PAGE_SIZE, pg++)
{
TRANSLOG_ADDRESS addr= (buffer->offset + i);
TRANSLOG_VALIDATOR_DATA data;
DBUG_PRINT("info", ("send log form %lu till %lu address: (%lu,0x%lx) "
"page #: %lu buffer size: %lu buffer: 0x%lx",
(ulong) i, (ulong) (i + TRANSLOG_PAGE_SIZE),
LSN_IN_PARTS(addr), (ulong) pg, (ulong) buffer->size,
(ulong) buffer));
data.addr= &addr;
DBUG_ASSERT(log_descriptor.pagecache->block_size == TRANSLOG_PAGE_SIZE);
DBUG_ASSERT(i + TRANSLOG_PAGE_SIZE <= buffer->size);
if (translog_status != TRANSLOG_OK && translog_status != TRANSLOG_SHUTDOWN)
DBUG_RETURN(1);
if (pagecache_write_part(log_descriptor.pagecache,
&file->handler, pg, 3,
buffer->buffer + i,
PAGECACHE_PLAIN_PAGE,
PAGECACHE_LOCK_LEFT_UNLOCKED,
PAGECACHE_PIN_LEFT_UNPINNED,
PAGECACHE_WRITE_DONE, 0,
LSN_IMPOSSIBLE,
skipped_data,
TRANSLOG_PAGE_SIZE - skipped_data))
{
DBUG_PRINT("error",
("Can't write page (%lu,0x%lx) to pagecache, error: %d",
(ulong) buffer->file->number,
(ulong) (LSN_OFFSET(buffer->offset)+ i),
my_errno));
translog_stop_writing();
DBUG_RETURN(1);
}
skipped_data= 0;
}
file->is_sync= 0;
if (my_pwrite(file->handler.file, buffer->buffer + buffer->skipped_data,
buffer->size - buffer->skipped_data,
LSN_OFFSET(buffer->offset) + buffer->skipped_data,
log_write_flags))
{
DBUG_PRINT("error", ("Can't write buffer (%lu,0x%lx) size %lu "
"to the disk (%d)",
(ulong) file->handler.file,
(ulong) LSN_OFFSET(buffer->offset),
(ulong) buffer->size, errno));
translog_stop_writing();
DBUG_RETURN(1);
}
/*
Dropping the flag in such way can make false alarm: signalling than the
file in not sync when it is sync, but the situation is quite rare and
protections with mutexes give much more overhead to the whole engine
*/
file->is_sync= 0;
if (LSN_OFFSET(buffer->last_lsn) != 0) /* if buffer->last_lsn is set */
{
if (translog_prev_buffer_flush_wait(buffer))
DBUG_RETURN(0); /* some the thread flushed the buffer already */
translog_set_sent_to_disk(buffer);
}
else
translog_set_only_in_buffers(buffer->next_buffer_offset);
/* say to next buffer that we are finished */
{
struct st_translog_buffer *next_buffer=
log_descriptor.buffers + ((buffer->buffer_no + 1) % TRANSLOG_BUFFERS_NO);
if (likely(translog_status == TRANSLOG_OK)){
translog_buffer_lock(next_buffer);
next_buffer->prev_sent_to_disk= buffer->offset;
translog_buffer_unlock(next_buffer);
mysql_cond_broadcast(&next_buffer->prev_sent_to_disk_cond);
}
else
{
/*
It is shutdown =>
1) there is only one thread
2) mutexes of other buffers can be destroyed => we can't use them
*/
next_buffer->prev_sent_to_disk= buffer->offset;
}
}
/* Free buffer */
buffer->file= NULL;
buffer->overlay= 0;
buffer->ver++;
mysql_mutex_lock(&log_descriptor.dirty_buffer_mask_lock);
log_descriptor.dirty_buffer_mask&= ~(1 << buffer->buffer_no);
mysql_mutex_unlock(&log_descriptor.dirty_buffer_mask_lock);
mysql_cond_broadcast(&buffer->waiting_filling_buffer);
DBUG_RETURN(0);
}
/*
Recover page with sector protection (wipe out failed chunks)
SYNOPSYS
translog_recover_page_up_to_sector()
page reference on the page
offset offset of failed sector
RETURN
0 OK
1 Error
*/
static my_bool translog_recover_page_up_to_sector(uchar *page, uint16 offset)
{
uint16 chunk_offset= translog_get_first_chunk_offset(page), valid_chunk_end;
DBUG_ENTER("translog_recover_page_up_to_sector");
DBUG_PRINT("enter", ("offset: %u first chunk: %u",
(uint) offset, (uint) chunk_offset));
while (page[chunk_offset] != TRANSLOG_FILLER && chunk_offset < offset)
{
uint16 chunk_length;
if ((chunk_length=
translog_get_total_chunk_length(page, chunk_offset)) == 0)
{
DBUG_PRINT("error", ("cant get chunk length (offset %u)",
(uint) chunk_offset));
DBUG_RETURN(1);
}
DBUG_PRINT("info", ("chunk: offset: %u length %u",
(uint) chunk_offset, (uint) chunk_length));
if (((ulong) chunk_offset) + ((ulong) chunk_length) > TRANSLOG_PAGE_SIZE)
{
DBUG_PRINT("error", ("damaged chunk (offset %u) in trusted area",
(uint) chunk_offset));
DBUG_RETURN(1);
}
chunk_offset+= chunk_length;
}
valid_chunk_end= chunk_offset;
/* end of trusted area - sector parsing */
while (page[chunk_offset] != TRANSLOG_FILLER)
{
uint16 chunk_length;
if ((chunk_length=
translog_get_total_chunk_length(page, chunk_offset)) == 0)
break;
DBUG_PRINT("info", ("chunk: offset: %u length %u",
(uint) chunk_offset, (uint) chunk_length));
if (((ulong) chunk_offset) + ((ulong) chunk_length) >
(uint) (offset + DISK_DRIVE_SECTOR_SIZE))
break;
chunk_offset+= chunk_length;
valid_chunk_end= chunk_offset;
}
DBUG_PRINT("info", ("valid chunk end offset: %u", (uint) valid_chunk_end));
memset(page + valid_chunk_end, TRANSLOG_FILLER,
TRANSLOG_PAGE_SIZE - valid_chunk_end);
DBUG_RETURN(0);
}
/**
@brief Dummy write callback.
*/
static my_bool
translog_dummy_callback(uchar *page __attribute__((unused)),
pgcache_page_no_t page_no __attribute__((unused)),
uchar* data_ptr __attribute__((unused)))
{
return 0;
}
/**
@brief Checks and removes sector protection.
@param page reference on the page content.
@param file transaction log descriptor.
@retvat 0 OK
@retval 1 Error
*/
static my_bool
translog_check_sector_protection(uchar *page, TRANSLOG_FILE *file)
{
uint i, offset;
uchar *table= page + page_overhead[page[TRANSLOG_PAGE_FLAGS]] -
TRANSLOG_PAGE_SIZE / DISK_DRIVE_SECTOR_SIZE;
uint8 current= table[0];
DBUG_ENTER("translog_check_sector_protection");
for (i= 1, offset= DISK_DRIVE_SECTOR_SIZE;
i < TRANSLOG_PAGE_SIZE / DISK_DRIVE_SECTOR_SIZE;
i++, offset+= DISK_DRIVE_SECTOR_SIZE)
{
/*
TODO: add chunk counting for "suspecting" sectors (difference is
more than 1-2), if difference more then present chunks then it is
the problem.
*/
uint8 test= page[offset];
DBUG_PRINT("info", ("sector: #%u offset: %u current: %lx "
"read: 0x%x stored: 0x%x%x",
i, offset, (ulong) current,
(uint) uint2korr(page + offset), (uint) table[i],
(uint) table[i + 1]));
/*
3 is minimal possible record length. So we can have "distance"
between 2 sectors value more then DISK_DRIVE_SECTOR_SIZE / 3
only if it is old value, i.e. the sector was not written.
*/
if (((test < current) &&
((uint)(0xFFL - current + test) > DISK_DRIVE_SECTOR_SIZE / 3)) ||
((test >= current) &&
((uint)(test - current) > DISK_DRIVE_SECTOR_SIZE / 3)))
{
if (translog_recover_page_up_to_sector(page, offset))
DBUG_RETURN(1);
file->was_recovered= 1;
DBUG_RETURN(0);
}
/* Restore value on the page */
page[offset]= table[i];
current= test;
DBUG_PRINT("info", ("sector: #%u offset: %u current: %lx "
"read: 0x%x stored: 0x%x",
i, offset, (ulong) current,
(uint) page[offset], (uint) table[i]));
}
DBUG_RETURN(0);
}
/**
@brief Log page validator (read callback)
@param page The page data to check
@param page_no The page number (<offset>/<page length>)
@param data_ptr Read callback data pointer (pointer to TRANSLOG_FILE)
@todo: add turning loghandler to read-only mode after merging with
that patch.
@retval 0 OK
@retval 1 Error
*/
static my_bool translog_page_validator(uchar *page,
pgcache_page_no_t page_no,
uchar* data_ptr)
{
uint this_page_page_overhead;
uint flags;
uchar *page_pos;
TRANSLOG_FILE *data= (TRANSLOG_FILE *) data_ptr;
#ifndef DBUG_OFF
pgcache_page_no_t offset= page_no * TRANSLOG_PAGE_SIZE;
#endif
DBUG_ENTER("translog_page_validator");
data->was_recovered= 0;
if ((pgcache_page_no_t) uint3korr(page) != page_no ||
(uint32) uint3korr(page + 3) != data->number)
{
DBUG_PRINT("error", ("Page (%lu,0x%lx): "
"page address written in the page is incorrect: "
"File %lu instead of %lu or page %lu instead of %lu",
(ulong) data->number, (ulong) offset,
(ulong) uint3korr(page + 3), (ulong) data->number,
(ulong) uint3korr(page),
(ulong) page_no));
DBUG_RETURN(1);
}
flags= (uint)(page[TRANSLOG_PAGE_FLAGS]);
this_page_page_overhead= page_overhead[flags];
if (flags & ~(TRANSLOG_PAGE_CRC | TRANSLOG_SECTOR_PROTECTION |
TRANSLOG_RECORD_CRC))
{
DBUG_PRINT("error", ("Page (%lu,0x%lx): "
"Garbage in the page flags field detected : %x",
(ulong) data->number, (ulong) offset,
(uint) flags));
DBUG_RETURN(1);
}
page_pos= page + (3 + 3 + 1);
if (flags & TRANSLOG_PAGE_CRC)
{
uint32 crc= translog_crc(page + this_page_page_overhead,
TRANSLOG_PAGE_SIZE -
this_page_page_overhead);
if (crc != uint4korr(page_pos))
{
DBUG_PRINT("error", ("Page (%lu,0x%lx): "
"CRC mismatch: calculated: %lx on the page %lx",
(ulong) data->number, (ulong) offset,
(ulong) crc, (ulong) uint4korr(page_pos)));
DBUG_RETURN(1);
}
page_pos+= CRC_SIZE; /* Skip crc */
}
if (flags & TRANSLOG_SECTOR_PROTECTION &&
translog_check_sector_protection(page, data))
{
DBUG_RETURN(1);
}
DBUG_RETURN(0);
}
/**
@brief Locks the loghandler.
*/
void translog_lock()
{
uint8 current_buffer;
DBUG_ENTER("translog_lock");
/*
Locking the loghandler mean locking current buffer, but it can change
during locking, so we should check it
*/
for (;;)
{
/*
log_descriptor.bc.buffer_no is only one byte so its reading is
an atomic operation
*/
current_buffer= log_descriptor.bc.buffer_no;
translog_buffer_lock(log_descriptor.buffers + current_buffer);
if (log_descriptor.bc.buffer_no == current_buffer)
break;
translog_buffer_unlock(log_descriptor.buffers + current_buffer);
}
DBUG_VOID_RETURN;
}
/*
Unlock the loghandler
SYNOPSIS
translog_unlock()
RETURN
0 OK
1 Error
*/
void translog_unlock()
{
translog_buffer_unlock(log_descriptor.bc.buffer);
}
/**
@brief Get log page by file number and offset of the beginning of the page
@param data validator data, which contains the page address
@param buffer buffer for page placing
(might not be used in some cache implementations)
@param direct_link if it is not NULL then caller can accept direct
link to the page cache
@retval NULL Error
@retval # pointer to the page cache which should be used to read this page
*/
static uchar *translog_get_page(TRANSLOG_VALIDATOR_DATA *data, uchar *buffer,
PAGECACHE_BLOCK_LINK **direct_link)
{
TRANSLOG_ADDRESS addr= *(data->addr), in_buffers;
uint32 file_no= LSN_FILE_NO(addr);
TRANSLOG_FILE *file;
DBUG_ENTER("translog_get_page");
DBUG_PRINT("enter", ("File: %lu Offset: %lu(0x%lx)",
(ulong) file_no,
(ulong) LSN_OFFSET(addr),
(ulong) LSN_OFFSET(addr)));
/* it is really page address */
DBUG_ASSERT(LSN_OFFSET(addr) % TRANSLOG_PAGE_SIZE == 0);
if (direct_link)
*direct_link= NULL;
restart:
in_buffers= translog_only_in_buffers();
DBUG_PRINT("info", ("in_buffers: (%lu,0x%lx)",
LSN_IN_PARTS(in_buffers)));
if (in_buffers != LSN_IMPOSSIBLE &&
cmp_translog_addr(addr, in_buffers) >= 0)
{
translog_lock();
DBUG_ASSERT(cmp_translog_addr(addr, log_descriptor.horizon) < 0);
/* recheck with locked loghandler */
in_buffers= translog_only_in_buffers();
if (cmp_translog_addr(addr, in_buffers) >= 0)
{
uint16 buffer_no= log_descriptor.bc.buffer_no;
#ifndef DBUG_OFF
uint16 buffer_start= buffer_no;
#endif
struct st_translog_buffer *buffer_unlock= log_descriptor.bc.buffer;
struct st_translog_buffer *curr_buffer= log_descriptor.bc.buffer;
for (;;)
{
/*
if the page is in the buffer and it is the last version of the
page (in case of division the page by buffer flush)
*/
if (curr_buffer->file != NULL &&
cmp_translog_addr(addr, curr_buffer->offset) >= 0 &&
cmp_translog_addr(addr,
(curr_buffer->next_buffer_offset ?
curr_buffer->next_buffer_offset:
curr_buffer->offset + curr_buffer->size)) < 0)
{
TRANSLOG_ADDRESS offset= curr_buffer->offset;
TRANSLOG_FILE *fl= curr_buffer->file;
uchar *from, *table= NULL;
int is_last_unfinished_page;
uint last_protected_sector= 0;
uint skipped_data= curr_buffer->skipped_data;
TRANSLOG_FILE file_copy;
uint8 ver= curr_buffer->ver;
translog_wait_for_writers(curr_buffer);
if (offset != curr_buffer->offset || fl != curr_buffer->file ||
ver != curr_buffer->ver)
{
DBUG_ASSERT(buffer_unlock == curr_buffer);
translog_buffer_unlock(buffer_unlock);
goto restart;
}
DBUG_ASSERT(LSN_FILE_NO(addr) == LSN_FILE_NO(curr_buffer->offset));
from= curr_buffer->buffer + (addr - curr_buffer->offset);
if (skipped_data && addr == curr_buffer->offset)
{
/*
We read page part of which is not present in buffer,
so we should read absent part from file (page cache actually)
*/
file= get_logfile_by_number(file_no);
DBUG_ASSERT(file != NULL);
/*
it's ok to not lock the page because:
- The log handler has it's own page cache.
- There is only one thread that can access the log
cache at a time
*/
if (!(buffer= pagecache_read(log_descriptor.pagecache,
&file->handler,
LSN_OFFSET(addr) / TRANSLOG_PAGE_SIZE,
3, buffer,
PAGECACHE_PLAIN_PAGE,
PAGECACHE_LOCK_LEFT_UNLOCKED,
NULL)))
DBUG_RETURN(NULL);
}
else
skipped_data= 0; /* Read after skipped in buffer data */
/*
Now we have correct data in buffer up to 'skipped_data'. The
following memcpy() will move the data from the internal buffer
that was not yet on disk.
*/
memcpy(buffer + skipped_data, from + skipped_data,
TRANSLOG_PAGE_SIZE - skipped_data);
/*
We can use copy then in translog_page_validator() because it
do not put it permanently somewhere.
We have to use copy because after releasing log lock we can't
guaranty that the file still be present (in real life it will be
present but theoretically possible that it will be released
already from last files cache);
*/
file_copy= *(curr_buffer->file);
file_copy.handler.callback_data= (uchar*) &file_copy;
is_last_unfinished_page= ((log_descriptor.bc.buffer ==
curr_buffer) &&
(log_descriptor.bc.ptr >= from) &&
(log_descriptor.bc.ptr <
from + TRANSLOG_PAGE_SIZE));
if (is_last_unfinished_page &&
(buffer[TRANSLOG_PAGE_FLAGS] & TRANSLOG_SECTOR_PROTECTION))
{
last_protected_sector= ((log_descriptor.bc.previous_offset - 1) /
DISK_DRIVE_SECTOR_SIZE);
table= buffer + log_descriptor.page_overhead -
TRANSLOG_PAGE_SIZE / DISK_DRIVE_SECTOR_SIZE;
}
DBUG_ASSERT(buffer_unlock == curr_buffer);
translog_buffer_unlock(buffer_unlock);
if (is_last_unfinished_page)
{
uint i;
/*
This is last unfinished page => we should not check CRC and
remove only that protection which already installed (no need
to check it)
We do not check the flag of sector protection, because if
(buffer[TRANSLOG_PAGE_FLAGS] & TRANSLOG_SECTOR_PROTECTION) is
not set then last_protected_sector will be 0 so following loop
will be never executed
*/
DBUG_PRINT("info", ("This is last unfinished page, "
"last protected sector %u",
last_protected_sector));
for (i= 1; i <= last_protected_sector; i++)
{
uint offset= i * DISK_DRIVE_SECTOR_SIZE;
DBUG_PRINT("info", ("Sector %u: 0x%02x <- 0x%02x",
i, buffer[offset],
table[i]));
buffer[offset]= table[i];
}
}
else
{
/*
This IF should be true because we use in-memory data which
supposed to be correct.
*/
if (translog_page_validator(buffer,
LSN_OFFSET(addr) / TRANSLOG_PAGE_SIZE,
(uchar*) &file_copy))
{
DBUG_ASSERT(0);
buffer= NULL;
}
}
DBUG_RETURN(buffer);
}
buffer_no= (buffer_no + 1) % TRANSLOG_BUFFERS_NO;
curr_buffer= log_descriptor.buffers + buffer_no;
translog_buffer_lock(curr_buffer);
translog_buffer_unlock(buffer_unlock);
buffer_unlock= curr_buffer;
/* we can't make a full circle */
DBUG_ASSERT(buffer_start != buffer_no);
}
}
translog_unlock();
}
file= get_logfile_by_number(file_no);
DBUG_ASSERT(file != NULL);
buffer= pagecache_read(log_descriptor.pagecache, &file->handler,
LSN_OFFSET(addr) / TRANSLOG_PAGE_SIZE,
3, (direct_link ? NULL : buffer),
PAGECACHE_PLAIN_PAGE,
(direct_link ?
PAGECACHE_LOCK_READ :
PAGECACHE_LOCK_LEFT_UNLOCKED),
direct_link);
DBUG_PRINT("info", ("Direct link is assigned to : 0x%lx * 0x%lx",
(ulong) direct_link,
(ulong)(direct_link ? *direct_link : NULL)));
data->was_recovered= file->was_recovered;
DBUG_RETURN(buffer);
}
/**
@brief free direct log page link
@param direct_link the direct log page link to be freed
*/
static void translog_free_link(PAGECACHE_BLOCK_LINK *direct_link)
{
DBUG_ENTER("translog_free_link");
DBUG_PRINT("info", ("Direct link: 0x%lx",
(ulong) direct_link));
if (direct_link)
pagecache_unlock_by_link(log_descriptor.pagecache, direct_link,
PAGECACHE_LOCK_READ_UNLOCK, PAGECACHE_UNPIN,
LSN_IMPOSSIBLE, LSN_IMPOSSIBLE, 0, FALSE);
DBUG_VOID_RETURN;
}
/**
@brief Finds last full page of the given log file.
@param addr address structure to fill with data, which contain
file number of the log file
@param last_page_ok Result of the check whether last page OK.
(for now only we check only that file length
divisible on page length).
@param no_errors suppress messages about non-critical errors
@retval 0 OK
@retval 1 Error
*/
static my_bool translog_get_last_page_addr(TRANSLOG_ADDRESS *addr,
my_bool *last_page_ok,
my_bool no_errors)
{
char path[FN_REFLEN];
uint32 rec_offset;
my_off_t file_size;
uint32 file_no= LSN_FILE_NO(*addr);
TRANSLOG_FILE *file;
#ifndef DBUG_OFF
char buff[21];
#endif
DBUG_ENTER("translog_get_last_page_addr");
if (likely((file= get_logfile_by_number(file_no)) != NULL))
{
/*
This function used only during initialization of loghandler or in
scanner (which mean we need read that part of the log), so the
requested log file have to be opened and can't be freed after
returning pointer on it (file_size).
*/
file_size= mysql_file_seek(file->handler.file, 0, SEEK_END, MYF(0));
}
else
{
/*
This branch is used only during very early initialization
when files are not opened.
*/
File fd;
if ((fd= mysql_file_open(key_file_translog,
translog_filename_by_fileno(file_no, path),
O_RDONLY, (no_errors ? MYF(0) : MYF(MY_WME)))) < 0)
{
my_errno= errno;
DBUG_PRINT("error", ("Error %d during opening file #%d",
errno, file_no));
DBUG_RETURN(1);
}
file_size= mysql_file_seek(fd, 0, SEEK_END, MYF(0));
mysql_file_close(fd, MYF(0));
}
DBUG_PRINT("info", ("File size: %s", llstr(file_size, buff)));
if (file_size == MY_FILEPOS_ERROR)
DBUG_RETURN(1);
DBUG_ASSERT(file_size < ULL(0xffffffff));
if (((uint32)file_size) > TRANSLOG_PAGE_SIZE)
{
rec_offset= (((((uint32)file_size) / TRANSLOG_PAGE_SIZE) - 1) *
TRANSLOG_PAGE_SIZE);
*last_page_ok= (((uint32)file_size) == rec_offset + TRANSLOG_PAGE_SIZE);
}
else
{
*last_page_ok= 0;
rec_offset= 0;
}
*addr= MAKE_LSN(file_no, rec_offset);
DBUG_PRINT("info", ("Last page: 0x%lx ok: %d", (ulong) rec_offset,
*last_page_ok));
DBUG_RETURN(0);
}
/**
@brief Get number bytes for record length storing
@param length Record length which will be encoded
@return 1,3,4,5 - number of bytes to store given length
*/
static uint translog_variable_record_length_bytes(translog_size_t length)
{
if (length < 250)
return 1;
if (length < 0xFFFF)
return 3;
if (length < (ulong) 0xFFFFFF)
return 4;
return 5;
}
/**
@brief Gets header of this chunk.
@param chunk The pointer to the chunk beginning
@retval # total length of the chunk
@retval 0 Error
*/
static uint16 translog_get_chunk_header_length(uchar *chunk)
{
DBUG_ENTER("translog_get_chunk_header_length");
switch (*chunk & TRANSLOG_CHUNK_TYPE) {
case TRANSLOG_CHUNK_LSN:
{
/* 0 chunk referred as LSN (head or tail) */
translog_size_t rec_len;
uchar *start= chunk;
uchar *ptr= start + 1 + 2;
uint16 chunk_len, header_len;
DBUG_PRINT("info", ("TRANSLOG_CHUNK_LSN"));
rec_len= translog_variable_record_1group_decode_len(&ptr);
chunk_len= uint2korr(ptr);
header_len= (uint16) (ptr - start) +2;
DBUG_PRINT("info", ("rec len: %lu chunk len: %u header len: %u",
(ulong) rec_len, (uint) chunk_len, (uint) header_len));
if (chunk_len)
{
/* TODO: fine header end */
/*
The last chunk of multi-group record can be base for it header
calculation (we skip to the first group to read the header) so if we
stuck here something is wrong.
*/
DBUG_ASSERT(0);
DBUG_RETURN(0); /* Keep compiler happy */
}
DBUG_RETURN(header_len);
}
case TRANSLOG_CHUNK_FIXED:
{
/* 1 (pseudo)fixed record (also LSN) */
DBUG_PRINT("info", ("TRANSLOG_CHUNK_FIXED = 3"));
DBUG_RETURN(3);
}
case TRANSLOG_CHUNK_NOHDR:
/* 2 no header chunk (till page end) */
DBUG_PRINT("info", ("TRANSLOG_CHUNK_NOHDR = 1"));
DBUG_RETURN(1);
break;
case TRANSLOG_CHUNK_LNGTH:
/* 3 chunk with chunk length */
DBUG_PRINT("info", ("TRANSLOG_CHUNK_LNGTH = 3"));
DBUG_RETURN(3);
break;
default:
DBUG_ASSERT(0);
DBUG_RETURN(0); /* Keep compiler happy */
}
}
/**
@brief Truncate the log to the given address. Used during the startup if the
end of log if corrupted.
@param addr new horizon
@retval 0 OK
@retval 1 Error
*/
static my_bool translog_truncate_log(TRANSLOG_ADDRESS addr)
{
uchar *page;
TRANSLOG_ADDRESS current_page;
uint32 next_page_offset, page_rest;
uint32 i;
File fd;
int rc;
TRANSLOG_VALIDATOR_DATA data;
char path[FN_REFLEN];
uchar page_buff[TRANSLOG_PAGE_SIZE];
DBUG_ENTER("translog_truncate_log");
/* TODO: write warning to the client */
DBUG_PRINT("warning", ("removing all records from (%lu,0x%lx) "
"till (%lu,0x%lx)",
LSN_IN_PARTS(addr),
LSN_IN_PARTS(log_descriptor.horizon)));
DBUG_ASSERT(cmp_translog_addr(addr, log_descriptor.horizon) < 0);
/* remove files between the address and horizon */
for (i= LSN_FILE_NO(addr) + 1; i <= LSN_FILE_NO(log_descriptor.horizon); i++)
if (mysql_file_delete(key_file_translog,
translog_filename_by_fileno(i, path), MYF(MY_WME)))
{
translog_unlock();
DBUG_RETURN(1);
}
/* truncate the last file up to the last page */
next_page_offset= LSN_OFFSET(addr);
next_page_offset= (next_page_offset -
((next_page_offset - 1) % TRANSLOG_PAGE_SIZE + 1) +
TRANSLOG_PAGE_SIZE);
page_rest= next_page_offset - LSN_OFFSET(addr);
memset(page_buff, TRANSLOG_FILLER, page_rest);
rc= ((fd= open_logfile_by_number_no_cache(LSN_FILE_NO(addr))) < 0 ||
((mysql_file_chsize(fd, next_page_offset, TRANSLOG_FILLER, MYF(MY_WME)) ||
(page_rest && my_pwrite(fd, page_buff, page_rest, LSN_OFFSET(addr),
log_write_flags)) ||
mysql_file_sync(fd, MYF(MY_WME)))));
translog_syncs++;
rc|= (fd > 0 && mysql_file_close(fd, MYF(MY_WME)));
if (sync_log_dir >= TRANSLOG_SYNC_DIR_ALWAYS)
{
rc|= sync_dir(log_descriptor.directory_fd, MYF(MY_WME | MY_IGNORE_BADFD));
translog_syncs++;
}
if (rc)
DBUG_RETURN(1);
/* fix the horizon */
log_descriptor.horizon= addr;
/* fix the buffer data */
current_page= MAKE_LSN(LSN_FILE_NO(addr), (next_page_offset -
TRANSLOG_PAGE_SIZE));
data.addr= ¤t_page;
if ((page= translog_get_page(&data, log_descriptor.buffers->buffer, NULL)) ==
NULL)
DBUG_RETURN(1);
if (page != log_descriptor.buffers->buffer)
memcpy(log_descriptor.buffers->buffer, page, TRANSLOG_PAGE_SIZE);
log_descriptor.bc.buffer->offset= current_page;
log_descriptor.bc.buffer->size= LSN_OFFSET(addr) - LSN_OFFSET(current_page);
log_descriptor.bc.ptr=
log_descriptor.buffers->buffer + log_descriptor.bc.buffer->size;
log_descriptor.bc.current_page_fill= log_descriptor.bc.buffer->size;
DBUG_RETURN(0);
}
/**
Applies function 'callback' to all files (in a directory) which
name looks like a log's name (aria_log.[0-9]{7}).
If 'callback' returns TRUE this interrupts the walk and returns
TRUE. Otherwise FALSE is returned after processing all log files.
It cannot just use log_descriptor.directory because that may not yet have
been initialized.
@param directory directory to scan
@param callback function to apply; is passed directory and base
name of found file
*/
my_bool translog_walk_filenames(const char *directory,
my_bool (*callback)(const char *,
const char *))
{
MY_DIR *dirp;
uint i;
my_bool rc= FALSE;
/* Finds and removes transaction log files */
if (!(dirp = my_dir(directory, MYF(MY_DONT_SORT))))
return FALSE;
for (i= 0; i < dirp->number_off_files; i++)
{
char *file= dirp->dir_entry[i].name;
if (strncmp(file, "aria_log.", 10) == 0 &&
file[10] >= '0' && file[10] <= '9' &&
file[11] >= '0' && file[11] <= '9' &&
file[12] >= '0' && file[12] <= '9' &&
file[13] >= '0' && file[13] <= '9' &&
file[14] >= '0' && file[14] <= '9' &&
file[15] >= '0' && file[15] <= '9' &&
file[16] >= '0' && file[16] <= '9' &&
file[17] >= '0' && file[17] <= '9' &&
file[18] == '\0' && (*callback)(directory, file))
{
rc= TRUE;
break;
}
}
my_dirend(dirp);
return rc;
}
/**
@brief Fills table of dependence length of page header from page flags
*/
void translog_fill_overhead_table()
{
uint i;
for (i= 0; i < TRANSLOG_FLAGS_NUM; i++)
{
page_overhead[i]= 7;
if (i & TRANSLOG_PAGE_CRC)
page_overhead[i]+= CRC_SIZE;
if (i & TRANSLOG_SECTOR_PROTECTION)
page_overhead[i]+= TRANSLOG_PAGE_SIZE /
DISK_DRIVE_SECTOR_SIZE;
}
}
/**
Callback to find first log in directory.
*/
static my_bool translog_callback_search_first(const char *directory
__attribute__((unused)),
const char *filename
__attribute__((unused)))
{
return TRUE;
}
/**
@brief Checks that chunk is LSN one
@param type type of the chunk
@retval 1 the chunk is LNS
@retval 0 the chunk is not LSN
*/
static my_bool translog_is_LSN_chunk(uchar type)
{
DBUG_ENTER("translog_is_LSN_chunk");
DBUG_PRINT("info", ("byte: %x chunk type: %u record type: %u",
type, type >> 6, type & TRANSLOG_REC_TYPE));
DBUG_RETURN(((type & TRANSLOG_CHUNK_TYPE) == TRANSLOG_CHUNK_FIXED) ||
(((type & TRANSLOG_CHUNK_TYPE) == TRANSLOG_CHUNK_LSN) &&
((type & TRANSLOG_REC_TYPE)) != TRANSLOG_CHUNK_0_CONT));
}
/**
@brief Initialize transaction log
@param directory Directory where log files are put
@param log_file_max_size max size of one log size (for new logs creation)
@param server_version version of MySQL server (MYSQL_VERSION_ID)
@param server_id server ID (replication & Co)
@param pagecache Page cache for the log reads
@param flags flags (TRANSLOG_PAGE_CRC, TRANSLOG_SECTOR_PROTECTION
TRANSLOG_RECORD_CRC)
@param read_only Put transaction log in read-only mode
@param init_table_func function to initialize record descriptors table
@param no_errors suppress messages about non-critical errors
@todo
Free used resources in case of error.
@retval 0 OK
@retval 1 Error
*/
my_bool translog_init_with_table(const char *directory,
uint32 log_file_max_size,
uint32 server_version,
uint32 server_id, PAGECACHE *pagecache,
uint flags, my_bool readonly,
void (*init_table_func)(),
my_bool no_errors)
{
int i;
int old_log_was_recovered= 0, logs_found= 0;
uint old_flags= flags;
uint32 start_file_num= 1;
TRANSLOG_ADDRESS sure_page, last_page, last_valid_page, checkpoint_lsn;
my_bool version_changed= 0;
DBUG_ENTER("translog_init_with_table");
translog_syncs= 0;
flush_start= 0;
id_to_share= NULL;
log_descriptor.directory_fd= -1;
log_descriptor.is_everything_flushed= 1;
log_descriptor.flush_in_progress= 0;
log_descriptor.flush_no= 0;
log_descriptor.next_pass_max_lsn= LSN_IMPOSSIBLE;
/* Normally in Aria this this calls translog_table_init() */
(*init_table_func)();
compile_time_assert(sizeof(log_descriptor.dirty_buffer_mask) * 8 >=
TRANSLOG_BUFFERS_NO);
log_descriptor.dirty_buffer_mask= 0;
if (readonly)
log_descriptor.open_flags= O_BINARY | O_RDONLY;
else
log_descriptor.open_flags= O_BINARY | O_RDWR;
if (mysql_mutex_init(key_TRANSLOG_BUFFER_mutex,
&log_descriptor.sent_to_disk_lock, MY_MUTEX_INIT_FAST) ||
mysql_mutex_init(key_TRANSLOG_DESCRIPTOR_file_header_lock,
&log_descriptor.file_header_lock, MY_MUTEX_INIT_FAST) ||
mysql_mutex_init(key_TRANSLOG_DESCRIPTOR_unfinished_files_lock,
&log_descriptor.unfinished_files_lock, MY_MUTEX_INIT_FAST) ||
mysql_mutex_init(key_TRANSLOG_DESCRIPTOR_purger_lock,
&log_descriptor.purger_lock, MY_MUTEX_INIT_FAST) ||
mysql_mutex_init(key_TRANSLOG_DESCRIPTOR_log_flush_lock,
&log_descriptor.log_flush_lock, MY_MUTEX_INIT_FAST) ||
mysql_mutex_init(key_TRANSLOG_DESCRIPTOR_dirty_buffer_mask_lock,
&log_descriptor.dirty_buffer_mask_lock, MY_MUTEX_INIT_FAST) ||
mysql_cond_init(key_TRANSLOG_DESCRIPTOR_log_flush_cond,
&log_descriptor.log_flush_cond, 0) ||
mysql_cond_init(key_TRANSLOG_DESCRIPTOR_new_goal_cond,
&log_descriptor.new_goal_cond, 0) ||
mysql_rwlock_init(key_TRANSLOG_DESCRIPTOR_open_files_lock,
&log_descriptor.open_files_lock) ||
my_init_dynamic_array(&log_descriptor.open_files,
sizeof(TRANSLOG_FILE*), 10, 10) ||
my_init_dynamic_array(&log_descriptor.unfinished_files,
sizeof(struct st_file_counter),
10, 10))
goto err;
log_descriptor.min_need_file= 0;
log_descriptor.min_file_number= 0;
log_descriptor.last_lsn_checked= LSN_IMPOSSIBLE;
/* Directory to store files */
unpack_dirname(log_descriptor.directory, directory);
#ifndef __WIN__
if ((log_descriptor.directory_fd= my_open(log_descriptor.directory,
O_RDONLY, MYF(MY_WME))) < 0)
{
my_errno= errno;
DBUG_PRINT("error", ("Error %d during opening directory '%s'",
errno, log_descriptor.directory));
goto err;
}
#endif
log_descriptor.in_buffers_only= LSN_IMPOSSIBLE;
DBUG_ASSERT(log_file_max_size % TRANSLOG_PAGE_SIZE == 0 &&
log_file_max_size >= TRANSLOG_MIN_FILE_SIZE);
/* max size of one log size (for new logs creation) */
log_file_size= log_descriptor.log_file_max_size=
log_file_max_size;
/* server version */
log_descriptor.server_version= server_version;
/* server ID */
log_descriptor.server_id= server_id;
/* Page cache for the log reads */
log_descriptor.pagecache= pagecache;
/* Flags */
DBUG_ASSERT((flags &
~(TRANSLOG_PAGE_CRC | TRANSLOG_SECTOR_PROTECTION |
TRANSLOG_RECORD_CRC)) == 0);
log_descriptor.flags= flags;
translog_fill_overhead_table();
log_descriptor.page_overhead= page_overhead[flags];
log_descriptor.page_capacity_chunk_2=
TRANSLOG_PAGE_SIZE - log_descriptor.page_overhead - 1;
compile_time_assert(TRANSLOG_WRITE_BUFFER % TRANSLOG_PAGE_SIZE == 0);
log_descriptor.buffer_capacity_chunk_2=
(TRANSLOG_WRITE_BUFFER / TRANSLOG_PAGE_SIZE) *
log_descriptor.page_capacity_chunk_2;
log_descriptor.half_buffer_capacity_chunk_2=
log_descriptor.buffer_capacity_chunk_2 / 2;
DBUG_PRINT("info",
("Overhead: %u pc2: %u bc2: %u, bc2/2: %u",
log_descriptor.page_overhead,
log_descriptor.page_capacity_chunk_2,
log_descriptor.buffer_capacity_chunk_2,
log_descriptor.half_buffer_capacity_chunk_2));
/* Just to init it somehow (hack for bootstrap)*/
{
TRANSLOG_FILE *file= 0;
log_descriptor.min_file = log_descriptor.max_file= 1;
insert_dynamic(&log_descriptor.open_files, (uchar *)&file);
translog_start_buffer(log_descriptor.buffers, &log_descriptor.bc, 0);
pop_dynamic(&log_descriptor.open_files);
}
/* Buffers for log writing */
for (i= 0; i < TRANSLOG_BUFFERS_NO; i++)
{
if (translog_buffer_init(log_descriptor.buffers + i, i))
goto err;
DBUG_PRINT("info", ("translog_buffer buffer #%u: 0x%lx",
i, (ulong) log_descriptor.buffers + i));
}
/*
last_logno and last_checkpoint_lsn were set in
ma_control_file_create_or_open()
*/
logs_found= (last_logno != FILENO_IMPOSSIBLE);
translog_status= (readonly ? TRANSLOG_READONLY : TRANSLOG_OK);
checkpoint_lsn= last_checkpoint_lsn;
if (logs_found)
{
my_bool pageok;
DBUG_PRINT("info", ("log found..."));
/*
TODO: scan directory for aria_log.XXXXXXXX files and find
highest XXXXXXXX & set logs_found
TODO: check that last checkpoint within present log addresses space
find the log end
*/
if (LSN_FILE_NO(last_checkpoint_lsn) == FILENO_IMPOSSIBLE)
{
DBUG_ASSERT(LSN_OFFSET(last_checkpoint_lsn) == 0);
/* only last log needs to be checked */
sure_page= MAKE_LSN(last_logno, TRANSLOG_PAGE_SIZE);
}
else
{
sure_page= last_checkpoint_lsn;
DBUG_ASSERT(LSN_OFFSET(sure_page) % TRANSLOG_PAGE_SIZE != 0);
sure_page-= LSN_OFFSET(sure_page) % TRANSLOG_PAGE_SIZE;
}
/* Set horizon to the beginning of the last file first */
log_descriptor.horizon= last_page= MAKE_LSN(last_logno, 0);
if (translog_get_last_page_addr(&last_page, &pageok, no_errors))
{
if (!translog_walk_filenames(log_descriptor.directory,
&translog_callback_search_first))
{
/*
Files was deleted, just start from the next log number, so that
existing tables are in the past.
*/
start_file_num= last_logno + 1;
checkpoint_lsn= LSN_IMPOSSIBLE; /* no log so no checkpoint */
logs_found= 0;
}
else
goto err;
}
else if (LSN_OFFSET(last_page) == 0)
{
if (LSN_FILE_NO(last_page) == 1)
{
logs_found= 0; /* file #1 has no pages */
DBUG_PRINT("info", ("log found. But is is empty => no log assumed"));
}
else
{
last_page-= LSN_ONE_FILE;
if (translog_get_last_page_addr(&last_page, &pageok, 0))
goto err;
}
}
if (logs_found)
{
uint32 i;
log_descriptor.min_file= translog_first_file(log_descriptor.horizon, 1);
log_descriptor.max_file= last_logno;
/* Open all files */
if (allocate_dynamic(&log_descriptor.open_files,
log_descriptor.max_file -
log_descriptor.min_file + 1))
goto err;
for (i = log_descriptor.max_file; i >= log_descriptor.min_file; i--)
{
/*
We can't allocate all file together because they will be freed
one by one
*/
TRANSLOG_FILE *file= (TRANSLOG_FILE *)my_malloc(sizeof(TRANSLOG_FILE),
MYF(0));
compile_time_assert(MY_FILEPOS_ERROR > ULL(0xffffffff));
if (file == NULL ||
(file->handler.file=
open_logfile_by_number_no_cache(i)) < 0 ||
mysql_file_seek(file->handler.file, 0, SEEK_END, MYF(0)) >=
ULL(0xffffffff))
{
int j;
for (j= i - log_descriptor.min_file - 1; j > 0; j--)
{
TRANSLOG_FILE *el=
*dynamic_element(&log_descriptor.open_files, j,
TRANSLOG_FILE **);
mysql_file_close(el->handler.file, MYF(MY_WME));
my_free(el);
}
if (file)
{
free(file);
goto err;
}
else
goto err;
}
translog_file_init(file, i, 1);
/* we allocated space so it can't fail */
insert_dynamic(&log_descriptor.open_files, (uchar *)&file);
}
DBUG_ASSERT(log_descriptor.max_file - log_descriptor.min_file + 1 ==
log_descriptor.open_files.elements);
}
}
else if (readonly)
{
/* There is no logs and there is read-only mode => nothing to read */
DBUG_PRINT("error", ("No logs and read-only mode"));
goto err;
}
if (logs_found)
{
TRANSLOG_ADDRESS current_page= sure_page;
my_bool pageok;
DBUG_PRINT("info", ("The log is really present"));
DBUG_ASSERT(sure_page <= last_page);
/* TODO: check page size */
last_valid_page= LSN_IMPOSSIBLE;
/*
Scans and validate pages. We need it to show "outside" only for sure
valid part of the log. If the log was damaged then fixed we have to
cut off damaged part before some other process start write something
in the log.
*/
do
{
TRANSLOG_ADDRESS current_file_last_page;
current_file_last_page= current_page;
if (translog_get_last_page_addr(¤t_file_last_page, &pageok, 0))
goto err;
if (!pageok)
{
DBUG_PRINT("error", ("File %lu have no complete last page",
(ulong) LSN_FILE_NO(current_file_last_page)));
old_log_was_recovered= 1;
/* This file is not written till the end so it should be last */
last_page= current_file_last_page;
/* TODO: issue warning */
}
do
{
TRANSLOG_VALIDATOR_DATA data;
TRANSLOG_PAGE_SIZE_BUFF psize_buff;
uchar *page;
data.addr= ¤t_page;
if ((page= translog_get_page(&data, psize_buff.buffer, NULL)) == NULL)
goto err;
if (data.was_recovered)
{
DBUG_PRINT("error", ("file no: %lu (%d) "
"rec_offset: 0x%lx (%lu) (%d)",
(ulong) LSN_FILE_NO(current_page),
(uint3korr(page + 3) !=
LSN_FILE_NO(current_page)),
(ulong) LSN_OFFSET(current_page),
(ulong) (LSN_OFFSET(current_page) /
TRANSLOG_PAGE_SIZE),
(uint3korr(page) !=
LSN_OFFSET(current_page) /
TRANSLOG_PAGE_SIZE)));
old_log_was_recovered= 1;
break;
}
old_flags= page[TRANSLOG_PAGE_FLAGS];
last_valid_page= current_page;
current_page+= TRANSLOG_PAGE_SIZE; /* increase offset */
} while (current_page <= current_file_last_page);
current_page+= LSN_ONE_FILE;
current_page= LSN_REPLACE_OFFSET(current_page, TRANSLOG_PAGE_SIZE);
} while (LSN_FILE_NO(current_page) <= LSN_FILE_NO(last_page) &&
!old_log_was_recovered);
if (last_valid_page == LSN_IMPOSSIBLE)
{
/* Panic!!! Even page which should be valid is invalid */
/* TODO: issue error */
goto err;
}
DBUG_PRINT("info", ("Last valid page is in file: %lu "
"offset: %lu (0x%lx) "
"Logs found: %d was recovered: %d "
"flags match: %d",
(ulong) LSN_FILE_NO(last_valid_page),
(ulong) LSN_OFFSET(last_valid_page),
(ulong) LSN_OFFSET(last_valid_page),
logs_found, old_log_was_recovered,
(old_flags == flags)));
/* TODO: check server ID */
if (logs_found && !old_log_was_recovered && old_flags == flags)
{
TRANSLOG_VALIDATOR_DATA data;
TRANSLOG_PAGE_SIZE_BUFF psize_buff;
uchar *page;
uint16 chunk_offset;
data.addr= &last_valid_page;
/* continue old log */
DBUG_ASSERT(LSN_FILE_NO(last_valid_page)==
LSN_FILE_NO(log_descriptor.horizon));
if ((page= translog_get_page(&data, psize_buff.buffer, NULL)) == NULL ||
(chunk_offset= translog_get_first_chunk_offset(page)) == 0)
goto err;
/* Puts filled part of old page in the buffer */
log_descriptor.horizon= last_valid_page;
translog_start_buffer(log_descriptor.buffers, &log_descriptor.bc, 0);
/*
Free space if filled with TRANSLOG_FILLER and first uchar of
real chunk can't be TRANSLOG_FILLER
*/
while (chunk_offset < TRANSLOG_PAGE_SIZE &&
page[chunk_offset] != TRANSLOG_FILLER)
{
uint16 chunk_length;
if ((chunk_length=
translog_get_total_chunk_length(page, chunk_offset)) == 0)
goto err;
DBUG_PRINT("info", ("chunk: offset: %u length: %u",
(uint) chunk_offset, (uint) chunk_length));
chunk_offset+= chunk_length;
/* chunk can't cross the page border */
DBUG_ASSERT(chunk_offset <= TRANSLOG_PAGE_SIZE);
}
memcpy(log_descriptor.buffers->buffer, page, chunk_offset);
log_descriptor.bc.buffer->size+= chunk_offset;
log_descriptor.bc.ptr+= chunk_offset;
log_descriptor.bc.current_page_fill= chunk_offset;
log_descriptor.horizon= LSN_REPLACE_OFFSET(log_descriptor.horizon,
(chunk_offset +
LSN_OFFSET(last_valid_page)));
DBUG_PRINT("info", ("Move Page #%u: 0x%lx chaser: %d Size: %lu (%lu)",
(uint) log_descriptor.bc.buffer_no,
(ulong) log_descriptor.bc.buffer,
log_descriptor.bc.chaser,
(ulong) log_descriptor.bc.buffer->size,
(ulong) (log_descriptor.bc.ptr - log_descriptor.bc.
buffer->buffer)));
translog_check_cursor(&log_descriptor.bc);
}
if (!old_log_was_recovered && old_flags == flags)
{
LOGHANDLER_FILE_INFO info;
/*
Accessing &log_descriptor.open_files without mutex is safe
because it is initialization
*/
if (translog_read_file_header(&info,
(*dynamic_element(&log_descriptor.
open_files,
0, TRANSLOG_FILE **))->
handler.file))
goto err;
version_changed= (info.maria_version != TRANSLOG_VERSION_ID);
}
}
DBUG_PRINT("info", ("Logs found: %d was recovered: %d",
logs_found, old_log_was_recovered));
if (!logs_found)
{
TRANSLOG_FILE *file= (TRANSLOG_FILE*)my_malloc(sizeof(TRANSLOG_FILE),
MYF(0));
DBUG_PRINT("info", ("The log is not found => we will create new log"));
if (file == NULL)
goto err;
/* Start new log system from scratch */
log_descriptor.horizon= MAKE_LSN(start_file_num,
TRANSLOG_PAGE_SIZE); /* header page */
if ((file->handler.file=
create_logfile_by_number_no_cache(start_file_num)) == -1)
goto err;
translog_file_init(file, start_file_num, 0);
if (insert_dynamic(&log_descriptor.open_files, (uchar*)&file))
goto err;
log_descriptor.min_file= log_descriptor.max_file= start_file_num;
if (translog_write_file_header())
goto err;
DBUG_ASSERT(log_descriptor.max_file - log_descriptor.min_file + 1 ==
log_descriptor.open_files.elements);
if (ma_control_file_write_and_force(checkpoint_lsn, start_file_num,
max_trid_in_control_file,
recovery_failures))
goto err;
/* assign buffer 0 */
translog_start_buffer(log_descriptor.buffers, &log_descriptor.bc, 0);
translog_new_page_header(&log_descriptor.horizon, &log_descriptor.bc);
}
else if ((old_log_was_recovered || old_flags != flags || version_changed) &&
!readonly)
{
/* leave the damaged file untouched */
log_descriptor.horizon+= LSN_ONE_FILE;
/* header page */
log_descriptor.horizon= LSN_REPLACE_OFFSET(log_descriptor.horizon,
TRANSLOG_PAGE_SIZE);
if (translog_create_new_file())
goto err;
/*
Buffer system left untouched after recovery => we should init it
(starting from buffer 0)
*/
translog_start_buffer(log_descriptor.buffers, &log_descriptor.bc, 0);
translog_new_page_header(&log_descriptor.horizon, &log_descriptor.bc);
}
/* all LSNs that are on disk are flushed */
log_descriptor.log_start= log_descriptor.sent_to_disk=
log_descriptor.flushed= log_descriptor.horizon;
log_descriptor.in_buffers_only= log_descriptor.bc.buffer->offset;
log_descriptor.max_lsn= LSN_IMPOSSIBLE; /* set to 0 */
/*
Now 'flushed' is set to 'horizon' value, but 'horizon' is (potentially)
address of the next LSN and we want indicate that all LSNs that are
already on the disk are flushed so we need decrease horizon on 1 (we are
sure that there is no LSN on the disk which is greater then 'flushed'
and there will not be LSN created that is equal or less then the value
of the 'flushed').
*/
log_descriptor.flushed--; /* offset decreased */
log_descriptor.sent_to_disk--; /* offset decreased */
/*
Log records will refer to a MARIA_SHARE by a unique 2-byte id; set up
structures for generating 2-byte ids:
*/
my_atomic_rwlock_init(&LOCK_id_to_share);
id_to_share= (MARIA_SHARE **) my_malloc(SHARE_ID_MAX * sizeof(MARIA_SHARE*),
MYF(MY_WME | MY_ZEROFILL));
if (unlikely(!id_to_share))
goto err;
id_to_share--; /* min id is 1 */
/* Check the last LSN record integrity */
if (logs_found)
{
TRANSLOG_SCANNER_DATA scanner;
TRANSLOG_ADDRESS page_addr;
LSN last_lsn= LSN_IMPOSSIBLE;
/*
take very last page address and try to find LSN record on it
if it fail take address of previous page and so on
*/
page_addr= (log_descriptor.horizon -
((log_descriptor.horizon - 1) % TRANSLOG_PAGE_SIZE + 1));
if (translog_scanner_init(page_addr, 1, &scanner, 1))
goto err;
scanner.page_offset= page_overhead[scanner.page[TRANSLOG_PAGE_FLAGS]];
for (;;)
{
uint chunk_1byte;
chunk_1byte= scanner.page[scanner.page_offset];
while (!translog_is_LSN_chunk(chunk_1byte) &&
scanner.page != END_OF_LOG &&
scanner.page[scanner.page_offset] != TRANSLOG_FILLER &&
scanner.page_addr == page_addr)
{
if (translog_get_next_chunk(&scanner))
{
translog_destroy_scanner(&scanner);
goto err;
}
if (scanner.page != END_OF_LOG)
chunk_1byte= scanner.page[scanner.page_offset];
}
if (translog_is_LSN_chunk(chunk_1byte))
{
last_lsn= scanner.page_addr + scanner.page_offset;
if (translog_get_next_chunk(&scanner))
{
translog_destroy_scanner(&scanner);
goto err;
}
if (scanner.page == END_OF_LOG)
break; /* it was the last record */
chunk_1byte= scanner.page[scanner.page_offset];
continue; /* try to find other record on this page */
}
if (last_lsn != LSN_IMPOSSIBLE)
break; /* there is no more records on the page */
/* We have to make step back */
if (unlikely(LSN_OFFSET(page_addr) == TRANSLOG_PAGE_SIZE))
{
uint32 file_no= LSN_FILE_NO(page_addr);
my_bool last_page_ok;
/* it is beginning of the current file */
if (unlikely(file_no == 1))
{
/*
It is beginning of the log => there is no LSNs in the log =>
There is no harm in leaving it "as-is".
*/
log_descriptor.previous_flush_horizon= log_descriptor.horizon;
DBUG_PRINT("info", ("previous_flush_horizon: (%lu,0x%lx)",
LSN_IN_PARTS(log_descriptor.
previous_flush_horizon)));
DBUG_RETURN(0);
}
file_no--;
page_addr= MAKE_LSN(file_no, TRANSLOG_PAGE_SIZE);
translog_get_last_page_addr(&page_addr, &last_page_ok, 0);
/* page should be OK as it is not the last file */
DBUG_ASSERT(last_page_ok);
}
else
{
page_addr-= TRANSLOG_PAGE_SIZE;
}
translog_destroy_scanner(&scanner);
if (translog_scanner_init(page_addr, 1, &scanner, 1))
goto err;
scanner.page_offset= page_overhead[scanner.page[TRANSLOG_PAGE_FLAGS]];
}
translog_destroy_scanner(&scanner);
/* Now scanner points to the last LSN chunk, lets check it */
{
TRANSLOG_HEADER_BUFFER rec;
translog_size_t rec_len;
int len;
uchar buffer[1];
DBUG_PRINT("info", ("going to check the last found record (%lu,0x%lx)",
LSN_IN_PARTS(last_lsn)));
len=
translog_read_record_header(last_lsn, &rec);
if (unlikely (len == RECHEADER_READ_ERROR ||
len == RECHEADER_READ_EOF))
{
DBUG_PRINT("error", ("unexpected end of log or record during "
"reading record header: (%lu,0x%lx) len: %d",
LSN_IN_PARTS(last_lsn), len));
if (readonly)
log_descriptor.log_start= log_descriptor.horizon= last_lsn;
else if (translog_truncate_log(last_lsn))
{
translog_free_record_header(&rec);
goto err;
}
}
else
{
DBUG_ASSERT(last_lsn == rec.lsn);
if (likely(rec.record_length != 0))
{
/*
Reading the last byte of record will trigger scanning all
record chunks for now
*/
rec_len= translog_read_record(rec.lsn, rec.record_length - 1, 1,
buffer, NULL);
if (rec_len != 1)
{
DBUG_PRINT("error", ("unexpected end of log or record during "
"reading record body: (%lu,0x%lx) len: %d",
LSN_IN_PARTS(rec.lsn),
len));
if (readonly)
log_descriptor.log_start= log_descriptor.horizon= last_lsn;
else if (translog_truncate_log(last_lsn))
{
translog_free_record_header(&rec);
goto err;
}
}
}
}
translog_free_record_header(&rec);
}
}
log_descriptor.previous_flush_horizon= log_descriptor.horizon;
DBUG_PRINT("info", ("previous_flush_horizon: (%lu,0x%lx)",
LSN_IN_PARTS(log_descriptor.previous_flush_horizon)));
DBUG_RETURN(0);
err:
ma_message_no_user(0, "log initialization failed");
DBUG_RETURN(1);
}
/*
@brief Free transaction log file buffer.
@param buffer_no The buffer to free
*/
static void translog_buffer_destroy(struct st_translog_buffer *buffer)
{
DBUG_ENTER("translog_buffer_destroy");
DBUG_PRINT("enter",
("Buffer #%u: 0x%lx file: %d offset: (%lu,0x%lx) size: %lu",
(uint) buffer->buffer_no, (ulong) buffer,
(buffer->file ? buffer->file->handler.file : -1),
LSN_IN_PARTS(buffer->offset),
(ulong) buffer->size));
if (buffer->file != NULL)
{
/*
We ignore errors here, because we can't do something about it
(it is shutting down)
We also have to take the locks even if there can't be any other
threads running, because translog_buffer_flush()
requires that we have the buffer locked.
*/
translog_buffer_lock(buffer);
translog_buffer_flush(buffer);
translog_buffer_unlock(buffer);
}
DBUG_PRINT("info", ("Destroy mutex: 0x%lx", (ulong) &buffer->mutex));
mysql_mutex_destroy(&buffer->mutex);
mysql_cond_destroy(&buffer->waiting_filling_buffer);
DBUG_VOID_RETURN;
}
/*
Free log handler resources
SYNOPSIS
translog_destroy()
*/
void translog_destroy()
{
TRANSLOG_FILE **file;
uint i;
uint8 current_buffer;
DBUG_ENTER("translog_destroy");
DBUG_ASSERT(translog_status == TRANSLOG_OK ||
translog_status == TRANSLOG_READONLY);
translog_lock();
current_buffer= log_descriptor.bc.buffer_no;
translog_status= (translog_status == TRANSLOG_READONLY ?
TRANSLOG_UNINITED :
TRANSLOG_SHUTDOWN);
if (log_descriptor.bc.buffer->file != NULL)
translog_finish_page(&log_descriptor.horizon, &log_descriptor.bc);
translog_unlock();
for (i= 0; i < TRANSLOG_BUFFERS_NO; i++)
{
struct st_translog_buffer *buffer= (log_descriptor.buffers +
((i + current_buffer + 1) %
TRANSLOG_BUFFERS_NO));
translog_buffer_destroy(buffer);
}
translog_status= TRANSLOG_UNINITED;
/* close files */
while ((file= (TRANSLOG_FILE **)pop_dynamic(&log_descriptor.open_files)))
translog_close_log_file(*file);
mysql_mutex_destroy(&log_descriptor.sent_to_disk_lock);
mysql_mutex_destroy(&log_descriptor.file_header_lock);
mysql_mutex_destroy(&log_descriptor.unfinished_files_lock);
mysql_mutex_destroy(&log_descriptor.purger_lock);
mysql_mutex_destroy(&log_descriptor.log_flush_lock);
mysql_mutex_destroy(&log_descriptor.dirty_buffer_mask_lock);
mysql_cond_destroy(&log_descriptor.log_flush_cond);
mysql_cond_destroy(&log_descriptor.new_goal_cond);
mysql_rwlock_destroy(&log_descriptor.open_files_lock);
delete_dynamic(&log_descriptor.open_files);
delete_dynamic(&log_descriptor.unfinished_files);
if (log_descriptor.directory_fd >= 0)
mysql_file_close(log_descriptor.directory_fd, MYF(MY_WME));
my_atomic_rwlock_destroy(&LOCK_id_to_share);
if (id_to_share != NULL)
my_free(id_to_share + 1);
DBUG_VOID_RETURN;
}
/*
@brief Starts new page.
@param horizon \ Position in file and buffer where we are
@param cursor /
@param prev_buffer Buffer which should be flushed will be assigned here.
This is always set (to NULL if nothing to flush).
@note We do not want to flush the buffer immediately because we want to
let caller of this function first advance 'horizon' pointer and unlock the
loghandler and only then flush the log which can take some time.
@retval 0 OK
@retval 1 Error
*/
static my_bool translog_page_next(TRANSLOG_ADDRESS *horizon,
struct st_buffer_cursor *cursor,
struct st_translog_buffer **prev_buffer)
{
struct st_translog_buffer *buffer= cursor->buffer;
DBUG_ENTER("translog_page_next");
*prev_buffer= NULL;
if ((cursor->ptr + TRANSLOG_PAGE_SIZE >
cursor->buffer->buffer + TRANSLOG_WRITE_BUFFER) ||
(LSN_OFFSET(*horizon) >
log_descriptor.log_file_max_size - TRANSLOG_PAGE_SIZE))
{
DBUG_PRINT("info", ("Switch to next buffer Buffer Size: %lu (%lu) => %d "
"File size: %lu max: %lu => %d",
(ulong) cursor->buffer->size,
(ulong) (cursor->ptr - cursor->buffer->buffer),
(cursor->ptr + TRANSLOG_PAGE_SIZE >
cursor->buffer->buffer + TRANSLOG_WRITE_BUFFER),
(ulong) LSN_OFFSET(*horizon),
(ulong) log_descriptor.log_file_max_size,
(LSN_OFFSET(*horizon) >
(log_descriptor.log_file_max_size -
TRANSLOG_PAGE_SIZE))));
if (translog_buffer_next(horizon, cursor,
LSN_OFFSET(*horizon) >
(log_descriptor.log_file_max_size -
TRANSLOG_PAGE_SIZE)))
DBUG_RETURN(1);
*prev_buffer= buffer;
DBUG_PRINT("info", ("Buffer #%u (0x%lu): have to be flushed",
(uint) buffer->buffer_no, (ulong) buffer));
}
else
{
DBUG_PRINT("info", ("Use the same buffer #%u (0x%lu): "
"Buffer Size: %lu (%lu)",
(uint) buffer->buffer_no,
(ulong) buffer,
(ulong) cursor->buffer->size,
(ulong) (cursor->ptr - cursor->buffer->buffer)));
translog_finish_page(horizon, cursor);
translog_new_page_header(horizon, cursor);
}
DBUG_RETURN(0);
}
/*
Write data of given length to the current page
SYNOPSIS
translog_write_data_on_page()
horizon \ Pointers on file and buffer
cursor /
length IN length of the chunk
buffer buffer with data
RETURN
0 OK
1 Error
*/
static my_bool translog_write_data_on_page(TRANSLOG_ADDRESS *horizon,
struct st_buffer_cursor *cursor,
translog_size_t length,
uchar *buffer)
{
DBUG_ENTER("translog_write_data_on_page");
DBUG_PRINT("enter", ("Chunk length: %lu Page size %u",
(ulong) length, (uint) cursor->current_page_fill));
DBUG_ASSERT(length > 0);
DBUG_ASSERT(length + cursor->current_page_fill <= TRANSLOG_PAGE_SIZE);
DBUG_ASSERT(length + cursor->ptr <= cursor->buffer->buffer +
TRANSLOG_WRITE_BUFFER);
memcpy(cursor->ptr, buffer, length);
cursor->ptr+= length;
(*horizon)+= length; /* adds offset */
cursor->current_page_fill+= length;
if (!cursor->chaser)
cursor->buffer->size+= length;
DBUG_PRINT("info", ("Write data buffer #%u: 0x%lx "
"chaser: %d Size: %lu (%lu)",
(uint) cursor->buffer->buffer_no, (ulong) cursor->buffer,
cursor->chaser, (ulong) cursor->buffer->size,
(ulong) (cursor->ptr - cursor->buffer->buffer)));
translog_check_cursor(cursor);
DBUG_RETURN(0);
}
/*
Write data from parts of given length to the current page
SYNOPSIS
translog_write_parts_on_page()
horizon \ Pointers on file and buffer
cursor /
length IN length of the chunk
parts IN/OUT chunk source
RETURN
0 OK
1 Error
*/
static my_bool translog_write_parts_on_page(TRANSLOG_ADDRESS *horizon,
struct st_buffer_cursor *cursor,
translog_size_t length,
struct st_translog_parts *parts)
{
translog_size_t left= length;
uint cur= (uint) parts->current;
DBUG_ENTER("translog_write_parts_on_page");
DBUG_PRINT("enter", ("Chunk length: %lu parts: %u of %u. Page size: %u "
"Buffer size: %lu (%lu)",
(ulong) length,
(uint) (cur + 1), (uint) parts->elements,
(uint) cursor->current_page_fill,
(ulong) cursor->buffer->size,
(ulong) (cursor->ptr - cursor->buffer->buffer)));
DBUG_ASSERT(length > 0);
DBUG_ASSERT(length + cursor->current_page_fill <= TRANSLOG_PAGE_SIZE);
DBUG_ASSERT(length + cursor->ptr <= cursor->buffer->buffer +
TRANSLOG_WRITE_BUFFER);
do
{
translog_size_t len;
LEX_CUSTRING *part;
const uchar *buff;
DBUG_ASSERT(cur < parts->elements);
part= parts->parts + cur;
buff= part->str;
DBUG_PRINT("info", ("Part: %u Length: %lu left: %lu buff: 0x%lx",
(uint) (cur + 1), (ulong) part->length, (ulong) left,
(ulong) buff));
if (part->length > left)
{
/* we should write less then the current part */
len= left;
part->length-= len;
part->str+= len;
DBUG_PRINT("info", ("Set new part: %u Length: %lu",
(uint) (cur + 1), (ulong) part->length));
}
else
{
len= (translog_size_t) part->length;
cur++;
DBUG_PRINT("info", ("moved to next part (len: %lu)", (ulong) len));
}
DBUG_PRINT("info", ("copy: 0x%lx <- 0x%lx %u",
(ulong) cursor->ptr, (ulong)buff, (uint)len));
if (likely(len))
{
memcpy(cursor->ptr, buff, len);
left-= len;
cursor->ptr+= len;
}
} while (left);
DBUG_PRINT("info", ("Horizon: (%lu,0x%lx) Length %lu(0x%lx)",
LSN_IN_PARTS(*horizon),
(ulong) length, (ulong) length));
parts->current= cur;
(*horizon)+= length; /* offset increasing */
cursor->current_page_fill+= length;
if (!cursor->chaser)
cursor->buffer->size+= length;
/*
We do not not updating parts->total_record_length here because it is
need only before writing record to have total length
*/
DBUG_PRINT("info", ("Write parts buffer #%u: 0x%lx "
"chaser: %d Size: %lu (%lu) "
"Horizon: (%lu,0x%lx) buff offset: 0x%lx",
(uint) cursor->buffer->buffer_no, (ulong) cursor->buffer,
cursor->chaser, (ulong) cursor->buffer->size,
(ulong) (cursor->ptr - cursor->buffer->buffer),
LSN_IN_PARTS(*horizon),
(ulong) (LSN_OFFSET(cursor->buffer->offset) +
cursor->buffer->size)));
translog_check_cursor(cursor);
DBUG_RETURN(0);
}
/*
Put 1 group chunk type 0 header into parts array
SYNOPSIS
translog_write_variable_record_1group_header()
parts Descriptor of record source parts
type The log record type
short_trid Short transaction ID or 0 if it has no sense
header_length Calculated header length of chunk type 0
chunk0_header Buffer for the chunk header writing
*/
static void
translog_write_variable_record_1group_header(struct st_translog_parts *parts,
enum translog_record_type type,
SHORT_TRANSACTION_ID short_trid,
uint16 header_length,
uchar *chunk0_header)
{
LEX_CUSTRING *part;
DBUG_ASSERT(parts->current != 0); /* first part is left for header */
part= parts->parts + (--parts->current);
parts->total_record_length+= (translog_size_t) (part->length= header_length);
part->str= chunk0_header;
/* puts chunk type */
*chunk0_header= (uchar) (type | TRANSLOG_CHUNK_LSN);
int2store(chunk0_header + 1, short_trid);
/* puts record length */
translog_write_variable_record_1group_code_len(chunk0_header + 3,
parts->record_length,
header_length);
/* puts 0 as chunk length which indicate 1 group record */
int2store(chunk0_header + header_length - 2, 0);
}
/*
Increase number of writers for this buffer
SYNOPSIS
translog_buffer_increase_writers()
buffer target buffer
*/
static inline void
translog_buffer_increase_writers(struct st_translog_buffer *buffer)
{
DBUG_ENTER("translog_buffer_increase_writers");
translog_buffer_lock_assert_owner(buffer);
buffer->copy_to_buffer_in_progress++;
DBUG_PRINT("info", ("copy_to_buffer_in_progress. Buffer #%u 0x%lx progress: %d",
(uint) buffer->buffer_no, (ulong) buffer,
buffer->copy_to_buffer_in_progress));
DBUG_VOID_RETURN;
}
/*
Decrease number of writers for this buffer
SYNOPSIS
translog_buffer_decrease_writers()
buffer target buffer
*/
static void translog_buffer_decrease_writers(struct st_translog_buffer *buffer)
{
DBUG_ENTER("translog_buffer_decrease_writers");
translog_buffer_lock_assert_owner(buffer);
buffer->copy_to_buffer_in_progress--;
DBUG_PRINT("info",
("copy_to_buffer_in_progress. Buffer #%u 0x%lx progress: %d",
(uint) buffer->buffer_no, (ulong) buffer,
buffer->copy_to_buffer_in_progress));
if (buffer->copy_to_buffer_in_progress == 0)
mysql_cond_broadcast(&buffer->waiting_filling_buffer);
DBUG_VOID_RETURN;
}
/**
@brief Skip to the next page for chaser (thread which advanced horizon
pointer and now feeling the buffer)
@param horizon \ Pointers on file position and buffer
@param cursor /
@retval 1 OK
@retval 0 Error
*/
static my_bool translog_chaser_page_next(TRANSLOG_ADDRESS *horizon,
struct st_buffer_cursor *cursor)
{
struct st_translog_buffer *buffer_to_flush;
my_bool rc;
DBUG_ENTER("translog_chaser_page_next");
DBUG_ASSERT(cursor->chaser);
rc= translog_page_next(horizon, cursor, &buffer_to_flush);
if (buffer_to_flush != NULL)
{
translog_buffer_lock(buffer_to_flush);
translog_buffer_decrease_writers(buffer_to_flush);
if (!rc)
rc= translog_buffer_flush(buffer_to_flush);
translog_buffer_unlock(buffer_to_flush);
}
DBUG_RETURN(rc);
}
/*
Put chunk 2 from new page beginning
SYNOPSIS
translog_write_variable_record_chunk2_page()
parts Descriptor of record source parts
horizon \ Pointers on file position and buffer
cursor /
RETURN
0 OK
1 Error
*/
static my_bool
translog_write_variable_record_chunk2_page(struct st_translog_parts *parts,
TRANSLOG_ADDRESS *horizon,
struct st_buffer_cursor *cursor)
{
uchar chunk2_header[1];
DBUG_ENTER("translog_write_variable_record_chunk2_page");
chunk2_header[0]= TRANSLOG_CHUNK_NOHDR;
if (translog_chaser_page_next(horizon, cursor))
DBUG_RETURN(1);
/* Puts chunk type */
translog_write_data_on_page(horizon, cursor, 1, chunk2_header);
/* Puts chunk body */
translog_write_parts_on_page(horizon, cursor,
log_descriptor.page_capacity_chunk_2, parts);
DBUG_RETURN(0);
}
/*
Put chunk 3 of requested length in the buffer from new page beginning
SYNOPSIS
translog_write_variable_record_chunk3_page()
parts Descriptor of record source parts
length Length of this chunk
horizon \ Pointers on file position and buffer
cursor /
RETURN
0 OK
1 Error
*/
static my_bool
translog_write_variable_record_chunk3_page(struct st_translog_parts *parts,
uint16 length,
TRANSLOG_ADDRESS *horizon,
struct st_buffer_cursor *cursor)
{
LEX_CUSTRING *part;
uchar chunk3_header[1 + 2];
DBUG_ENTER("translog_write_variable_record_chunk3_page");
if (translog_chaser_page_next(horizon, cursor))
DBUG_RETURN(1);
if (length == 0)
{
/* It was call to write page header only (no data for chunk 3) */
DBUG_PRINT("info", ("It is a call to make page header only"));
DBUG_RETURN(0);
}
DBUG_ASSERT(parts->current != 0); /* first part is left for header */
part= parts->parts + (--parts->current);
parts->total_record_length+= (translog_size_t) (part->length= 1 + 2);
part->str= chunk3_header;
/* Puts chunk type */
*chunk3_header= (uchar) (TRANSLOG_CHUNK_LNGTH);
/* Puts chunk length */
int2store(chunk3_header + 1, length);
translog_write_parts_on_page(horizon, cursor, length + 1 + 2, parts);
DBUG_RETURN(0);
}
/*
Move log pointer (horizon) on given number pages starting from next page,
and given offset on the last page
SYNOPSIS
translog_advance_pointer()
pages Number of full pages starting from the next one
last_page_data Plus this data on the last page
RETURN
0 OK
1 Error
*/
static my_bool translog_advance_pointer(int pages, uint16 last_page_data)
{
translog_size_t last_page_offset= (log_descriptor.page_overhead +
last_page_data);
translog_size_t offset= (TRANSLOG_PAGE_SIZE -
log_descriptor.bc.current_page_fill +
pages * TRANSLOG_PAGE_SIZE + last_page_offset);
translog_size_t buffer_end_offset, file_end_offset, min_offset;
DBUG_ENTER("translog_advance_pointer");
DBUG_PRINT("enter", ("Pointer: (%lu, 0x%lx) + %u + %u pages + %u + %u",
LSN_IN_PARTS(log_descriptor.horizon),
(uint) (TRANSLOG_PAGE_SIZE -
log_descriptor.bc.current_page_fill),
pages, (uint) log_descriptor.page_overhead,
(uint) last_page_data));
translog_lock_assert_owner();
if (pages == -1)
{
/*
It is special case when we advance the pointer on the same page.
It can happened when we write last part of multi-group record.
*/
DBUG_ASSERT(last_page_data + log_descriptor.bc.current_page_fill <=
TRANSLOG_PAGE_SIZE);
offset= last_page_data;
last_page_offset= log_descriptor.bc.current_page_fill + last_page_data;
goto end;
}
DBUG_PRINT("info", ("last_page_offset %lu", (ulong) last_page_offset));
DBUG_ASSERT(last_page_offset <= TRANSLOG_PAGE_SIZE);
/*
The loop will be executed 1-3 times. Usually we advance the
pointer to fill only the current buffer (if we have more then 1/2 of
buffer free or 2 buffers (rest of current and all next). In case of
really huge record end where we write last group with "table of
content" of all groups and ignore buffer borders we can occupy
3 buffers.
*/
for (;;)
{
uint8 new_buffer_no;
struct st_translog_buffer *new_buffer;
struct st_translog_buffer *old_buffer;
buffer_end_offset= TRANSLOG_WRITE_BUFFER - log_descriptor.bc.buffer->size;
if (likely(log_descriptor.log_file_max_size >=
LSN_OFFSET(log_descriptor.horizon)))
file_end_offset= (log_descriptor.log_file_max_size -
LSN_OFFSET(log_descriptor.horizon));
else
{
/*
We already have written more then current file limit allow,
So we will finish this page and start new file
*/
file_end_offset= (TRANSLOG_PAGE_SIZE -
log_descriptor.bc.current_page_fill);
}
DBUG_PRINT("info", ("offset: %lu buffer_end_offs: %lu, "
"file_end_offs: %lu",
(ulong) offset, (ulong) buffer_end_offset,
(ulong) file_end_offset));
DBUG_PRINT("info", ("Buff #%u %u (0x%lx) offset 0x%lx + size 0x%lx = "
"0x%lx (0x%lx)",
(uint) log_descriptor.bc.buffer->buffer_no,
(uint) log_descriptor.bc.buffer_no,
(ulong) log_descriptor.bc.buffer,
(ulong) LSN_OFFSET(log_descriptor.bc.buffer->offset),
(ulong) log_descriptor.bc.buffer->size,
(ulong) (LSN_OFFSET(log_descriptor.bc.buffer->offset) +
log_descriptor.bc.buffer->size),
(ulong) LSN_OFFSET(log_descriptor.horizon)));
DBUG_ASSERT(LSN_OFFSET(log_descriptor.bc.buffer->offset) +
log_descriptor.bc.buffer->size ==
LSN_OFFSET(log_descriptor.horizon));
if (offset <= buffer_end_offset && offset <= file_end_offset)
break;
old_buffer= log_descriptor.bc.buffer;
new_buffer_no= (log_descriptor.bc.buffer_no + 1) % TRANSLOG_BUFFERS_NO;
new_buffer= log_descriptor.buffers + new_buffer_no;
translog_buffer_lock(new_buffer);
#ifndef DBUG_OFF
{
TRANSLOG_ADDRESS offset= new_buffer->offset;
TRANSLOG_FILE *file= new_buffer->file;
uint8 ver= new_buffer->ver;
translog_lock_assert_owner();
#endif
translog_wait_for_buffer_free(new_buffer);
#ifndef DBUG_OFF
/* We keep the handler locked so nobody can start this new buffer */
DBUG_ASSERT(offset == new_buffer->offset && new_buffer->file == NULL &&
(file == NULL ? ver : (uint8)(ver + 1)) == new_buffer->ver);
}
#endif
min_offset= min(buffer_end_offset, file_end_offset);
/* TODO: check is it ptr or size enough */
log_descriptor.bc.buffer->size+= min_offset;
log_descriptor.bc.ptr+= min_offset;
DBUG_PRINT("info", ("NewP buffer #%u: 0x%lx chaser: %d Size: %lu (%lu)",
(uint) log_descriptor.bc.buffer->buffer_no,
(ulong) log_descriptor.bc.buffer,
log_descriptor.bc.chaser,
(ulong) log_descriptor.bc.buffer->size,
(ulong) (log_descriptor.bc.ptr -log_descriptor.bc.
buffer->buffer)));
DBUG_ASSERT((ulong) (log_descriptor.bc.ptr -
log_descriptor.bc.buffer->buffer) ==
log_descriptor.bc.buffer->size);
DBUG_ASSERT(log_descriptor.bc.buffer->buffer_no ==
log_descriptor.bc.buffer_no);
translog_buffer_increase_writers(log_descriptor.bc.buffer);
if (file_end_offset <= buffer_end_offset)
{
log_descriptor.horizon+= LSN_ONE_FILE;
log_descriptor.horizon= LSN_REPLACE_OFFSET(log_descriptor.horizon,
TRANSLOG_PAGE_SIZE);
DBUG_PRINT("info", ("New file: %lu",
(ulong) LSN_FILE_NO(log_descriptor.horizon)));
if (translog_create_new_file())
{
DBUG_RETURN(1);
}
}
else
{
DBUG_PRINT("info", ("The same file"));
log_descriptor.horizon+= min_offset; /* offset increasing */
}
translog_start_buffer(new_buffer, &log_descriptor.bc, new_buffer_no);
old_buffer->next_buffer_offset= new_buffer->offset;
new_buffer->prev_buffer_offset= old_buffer->offset;
translog_buffer_unlock(old_buffer);
offset-= min_offset;
}
DBUG_PRINT("info", ("drop write_counter"));
log_descriptor.bc.write_counter= 0;
log_descriptor.bc.previous_offset= 0;
end:
log_descriptor.bc.ptr+= offset;
log_descriptor.bc.buffer->size+= offset;
translog_buffer_increase_writers(log_descriptor.bc.buffer);
log_descriptor.horizon+= offset; /* offset increasing */
log_descriptor.bc.current_page_fill= last_page_offset;
DBUG_PRINT("info", ("NewP buffer #%u: 0x%lx chaser: %d Size: %lu (%lu) "
"offset: %u last page: %u",
(uint) log_descriptor.bc.buffer->buffer_no,
(ulong) log_descriptor.bc.buffer,
log_descriptor.bc.chaser,
(ulong) log_descriptor.bc.buffer->size,
(ulong) (log_descriptor.bc.ptr -
log_descriptor.bc.buffer->
buffer), (uint) offset,
(uint) last_page_offset));
DBUG_PRINT("info",
("pointer moved to: (%lu, 0x%lx)",
LSN_IN_PARTS(log_descriptor.horizon)));
translog_check_cursor(&log_descriptor.bc);
log_descriptor.bc.protected= 0;
DBUG_RETURN(0);
}
/*
Get page rest
SYNOPSIS
translog_get_current_page_rest()
NOTE loghandler should be locked
RETURN
number of bytes left on the current page
*/
static uint translog_get_current_page_rest()
{
return (TRANSLOG_PAGE_SIZE - log_descriptor.bc.current_page_fill);
}
/*
Get buffer rest in full pages
SYNOPSIS
translog_get_current_buffer_rest()
NOTE loghandler should be locked
RETURN
number of full pages left on the current buffer
*/
static uint translog_get_current_buffer_rest()
{
return ((log_descriptor.bc.buffer->buffer + TRANSLOG_WRITE_BUFFER -
log_descriptor.bc.ptr) /
TRANSLOG_PAGE_SIZE);
}
/*
Calculate possible group size without first (current) page
SYNOPSIS
translog_get_current_group_size()
NOTE loghandler should be locked
RETURN
group size without first (current) page
*/
static translog_size_t translog_get_current_group_size()
{
/* buffer rest in full pages */
translog_size_t buffer_rest= translog_get_current_buffer_rest();
DBUG_ENTER("translog_get_current_group_size");
DBUG_PRINT("info", ("buffer_rest in pages: %u", buffer_rest));
buffer_rest*= log_descriptor.page_capacity_chunk_2;
/* in case of only half of buffer free we can write this and next buffer */
if (buffer_rest < log_descriptor.half_buffer_capacity_chunk_2)
{
DBUG_PRINT("info", ("buffer_rest: %lu -> add %lu",
(ulong) buffer_rest,
(ulong) log_descriptor.buffer_capacity_chunk_2));
buffer_rest+= log_descriptor.buffer_capacity_chunk_2;
}
DBUG_PRINT("info", ("buffer_rest: %lu", (ulong) buffer_rest));
DBUG_RETURN(buffer_rest);
}
static inline void set_lsn(LSN *lsn, LSN value)
{
DBUG_ENTER("set_lsn");
translog_lock_assert_owner();
*lsn= value;
/* we generate LSN so something is not flushed in log */
log_descriptor.is_everything_flushed= 0;
DBUG_PRINT("info", ("new LSN appeared: (%lu,0x%lx)", LSN_IN_PARTS(value)));
DBUG_VOID_RETURN;
}
/**
@brief Write variable record in 1 group.
@param lsn LSN of the record will be written here
@param type the log record type
@param short_trid Short transaction ID or 0 if it has no sense
@param parts Descriptor of record source parts
@param buffer_to_flush Buffer which have to be flushed if it is not 0
@param header_length Calculated header length of chunk type 0
@param trn Transaction structure pointer for hooks by
record log type, for short_id
@param hook_arg Argument which will be passed to pre-write and
in-write hooks of this record.
@note
We must have a translog_lock() when entering this function
We must have buffer_to_flush locked (if not null)
@return Operation status
@retval 0 OK
@retval 1 Error
*/
static my_bool
translog_write_variable_record_1group(LSN *lsn,
enum translog_record_type type,
MARIA_HA *tbl_info,
SHORT_TRANSACTION_ID short_trid,
struct st_translog_parts *parts,
struct st_translog_buffer
*buffer_to_flush, uint16 header_length,
TRN *trn, void *hook_arg)
{
TRANSLOG_ADDRESS horizon;
struct st_buffer_cursor cursor;
int rc= 0;
uint i;
translog_size_t record_rest, full_pages, first_page;
uint additional_chunk3_page= 0;
uchar chunk0_header[1 + 2 + 5 + 2];
DBUG_ENTER("translog_write_variable_record_1group");
translog_lock_assert_owner();
if (buffer_to_flush)
translog_buffer_lock_assert_owner(buffer_to_flush);
set_lsn(lsn, horizon= log_descriptor.horizon);
if (translog_set_lsn_for_files(LSN_FILE_NO(*lsn), LSN_FILE_NO(*lsn),
*lsn, TRUE) ||
(log_record_type_descriptor[type].inwrite_hook &&
(*log_record_type_descriptor[type].inwrite_hook)(type, trn, tbl_info,
lsn, hook_arg)))
{
translog_unlock();
DBUG_RETURN(1);
}
cursor= log_descriptor.bc;
cursor.chaser= 1;
/* Advance pointer to be able unlock the loghandler */
first_page= translog_get_current_page_rest();
record_rest= parts->record_length - (first_page - header_length);
full_pages= record_rest / log_descriptor.page_capacity_chunk_2;
record_rest= (record_rest % log_descriptor.page_capacity_chunk_2);
if (record_rest + 1 == log_descriptor.page_capacity_chunk_2)
{
DBUG_PRINT("info", ("2 chunks type 3 is needed"));
/* We will write 2 chunks type 3 at the end of this group */
additional_chunk3_page= 1;
record_rest= 1;
}
DBUG_PRINT("info", ("first_page: %u (%u) full_pages: %u (%lu) "
"additional: %u (%u) rest %u = %u",
first_page, first_page - header_length,
full_pages,
(ulong) full_pages *
log_descriptor.page_capacity_chunk_2,
additional_chunk3_page,
additional_chunk3_page *
(log_descriptor.page_capacity_chunk_2 - 1),
record_rest, parts->record_length));
/* record_rest + 3 is chunk type 3 overhead + record_rest */
rc|= translog_advance_pointer((int)(full_pages + additional_chunk3_page),
(record_rest ? record_rest + 3 : 0));
log_descriptor.bc.buffer->last_lsn= *lsn;
DBUG_PRINT("info", ("last_lsn set to (%lu,0x%lx) buffer: 0x%lx",
LSN_IN_PARTS(log_descriptor.bc.buffer->last_lsn),
(ulong) log_descriptor.bc.buffer));
translog_unlock();
/*
Check if we switched buffer and need process it (current buffer is
unlocked already => we will not delay other threads
*/
if (buffer_to_flush != NULL)
{
if (!rc)
rc= translog_buffer_flush(buffer_to_flush);
translog_buffer_unlock(buffer_to_flush);
}
if (rc)
DBUG_RETURN(1);
translog_write_variable_record_1group_header(parts, type, short_trid,
header_length, chunk0_header);
/* fill the pages */
translog_write_parts_on_page(&horizon, &cursor, first_page, parts);
DBUG_PRINT("info", ("absolute horizon: (%lu,0x%lx) local: (%lu,0x%lx)",
LSN_IN_PARTS(log_descriptor.horizon),
LSN_IN_PARTS(horizon)));
for (i= 0; i < full_pages; i++)
{
if (translog_write_variable_record_chunk2_page(parts, &horizon, &cursor))
DBUG_RETURN(1);
DBUG_PRINT("info", ("absolute horizon: (%lu,0x%lx) local: (%lu,0x%lx)",
LSN_IN_PARTS(log_descriptor.horizon),
LSN_IN_PARTS(horizon)));
}
if (additional_chunk3_page)
{
if (translog_write_variable_record_chunk3_page(parts,
log_descriptor.
page_capacity_chunk_2 - 2,
&horizon, &cursor))
DBUG_RETURN(1);
DBUG_PRINT("info", ("absolute horizon: (%lu,0x%lx) local: (%lu,0x%lx)",
LSN_IN_PARTS(log_descriptor.horizon),
LSN_IN_PARTS(horizon)));
DBUG_ASSERT(cursor.current_page_fill == TRANSLOG_PAGE_SIZE);
}
if (translog_write_variable_record_chunk3_page(parts,
record_rest,
&horizon, &cursor))
DBUG_RETURN(1);
DBUG_PRINT("info", ("absolute horizon: (%lu,0x%lx) local: (%lu,0x%lx)",
(ulong) LSN_FILE_NO(log_descriptor.horizon),
(ulong) LSN_OFFSET(log_descriptor.horizon),
(ulong) LSN_FILE_NO(horizon),
(ulong) LSN_OFFSET(horizon)));
translog_buffer_lock(cursor.buffer);
translog_buffer_decrease_writers(cursor.buffer);
translog_buffer_unlock(cursor.buffer);
DBUG_RETURN(rc);
}
/**
@brief Write variable record in 1 chunk.
@param lsn LSN of the record will be written here
@param type the log record type
@param short_trid Short transaction ID or 0 if it has no sense
@param parts Descriptor of record source parts
@param buffer_to_flush Buffer which have to be flushed if it is not 0
@param header_length Calculated header length of chunk type 0
@param trn Transaction structure pointer for hooks by
record log type, for short_id
@param hook_arg Argument which will be passed to pre-write and
in-write hooks of this record.
@note
We must have a translog_lock() when entering this function
We must have buffer_to_flush locked (if not null)
@return Operation status
@retval 0 OK
@retval 1 Error
*/
static my_bool
translog_write_variable_record_1chunk(LSN *lsn,
enum translog_record_type type,
MARIA_HA *tbl_info,
SHORT_TRANSACTION_ID short_trid,
struct st_translog_parts *parts,
struct st_translog_buffer
*buffer_to_flush, uint16 header_length,
TRN *trn, void *hook_arg)
{
int rc;
uchar chunk0_header[1 + 2 + 5 + 2];
DBUG_ENTER("translog_write_variable_record_1chunk");
translog_lock_assert_owner();
if (buffer_to_flush)
translog_buffer_lock_assert_owner(buffer_to_flush);
translog_write_variable_record_1group_header(parts, type, short_trid,
header_length, chunk0_header);
set_lsn(lsn, log_descriptor.horizon);
if (translog_set_lsn_for_files(LSN_FILE_NO(*lsn), LSN_FILE_NO(*lsn),
*lsn, TRUE) ||
(log_record_type_descriptor[type].inwrite_hook &&
(*log_record_type_descriptor[type].inwrite_hook)(type, trn, tbl_info,
lsn, hook_arg)))
{
translog_unlock();
DBUG_RETURN(1);
}
rc= translog_write_parts_on_page(&log_descriptor.horizon,
&log_descriptor.bc,
parts->total_record_length, parts);
log_descriptor.bc.buffer->last_lsn= *lsn;
DBUG_PRINT("info", ("last_lsn set to (%lu,0x%lx) buffer: 0x%lx",
LSN_IN_PARTS(log_descriptor.bc.buffer->last_lsn),
(ulong) log_descriptor.bc.buffer));
translog_unlock();
/*
check if we switched buffer and need process it (current buffer is
unlocked already => we will not delay other threads
*/
if (buffer_to_flush != NULL)
{
if (!rc)
rc= translog_buffer_flush(buffer_to_flush);
translog_buffer_unlock(buffer_to_flush);
}
DBUG_RETURN(rc);
}
/*
@brief Calculates and write LSN difference (compressed LSN).
@param base_lsn LSN from which we calculate difference
@param lsn LSN for codding
@param dst Result will be written to dst[-pack_length] .. dst[-1]
@note To store an LSN in a compact way we will use the following compression:
If a log record has LSN1, and it contains the LSN2 as a back reference,
Instead of LSN2 we write LSN1-LSN2, encoded as:
two bits the number N (see below)
14 bits
N bytes
That is, LSN is encoded in 2..5 bytes, and the number of bytes minus 2
is stored in the first two bits.
@note function made to write the result in backward direction with no
special sense or tricks both directions are equal in complicity
@retval # pointer on coded LSN
*/
static uchar *translog_put_LSN_diff(LSN base_lsn, LSN lsn, uchar *dst)
{
uint64 diff;
DBUG_ENTER("translog_put_LSN_diff");
DBUG_PRINT("enter", ("Base: (%lu,0x%lx) val: (%lu,0x%lx) dst: 0x%lx",
LSN_IN_PARTS(base_lsn), LSN_IN_PARTS(lsn),
(ulong) dst));
DBUG_ASSERT(base_lsn > lsn);
diff= base_lsn - lsn;
DBUG_PRINT("info", ("Diff: 0x%llx", (ulonglong) diff));
if (diff <= 0x3FFF)
{
dst-= 2;
/*
Note we store this high uchar first to ensure that first uchar has
0 in the 3 upper bits.
*/
dst[0]= (uchar)(diff >> 8);
dst[1]= (uchar)(diff & 0xFF);
}
else if (diff <= 0x3FFFFFL)
{
dst-= 3;
dst[0]= (uchar)(0x40 | (diff >> 16));
int2store(dst + 1, diff & 0xFFFF);
}
else if (diff <= 0x3FFFFFFFL)
{
dst-= 4;
dst[0]= (uchar)(0x80 | (diff >> 24));
int3store(dst + 1, diff & 0xFFFFFFL);
}
else if (diff <= LL(0x3FFFFFFFFF))
{
dst-= 5;
dst[0]= (uchar)(0xC0 | (diff >> 32));
int4store(dst + 1, diff & 0xFFFFFFFFL);
}
else
{
/*
It is full LSN after special 1 diff (which is impossible
in real life)
*/
dst-= 2 + LSN_STORE_SIZE;
dst[0]= 0;
dst[1]= 1;
lsn_store(dst + 2, lsn);
}
DBUG_PRINT("info", ("new dst: 0x%lx", (ulong) dst));
DBUG_RETURN(dst);
}
/*
Get LSN from LSN-difference (compressed LSN)
SYNOPSIS
translog_get_LSN_from_diff()
base_lsn LSN from which we calculate difference
src pointer to coded lsn
dst pointer to buffer where to write 7byte LSN
NOTE:
To store an LSN in a compact way we will use the following compression:
If a log record has LSN1, and it contains the lSN2 as a back reference,
Instead of LSN2 we write LSN1-LSN2, encoded as:
two bits the number N (see below)
14 bits
N bytes
That is, LSN is encoded in 2..5 bytes, and the number of bytes minus 2
is stored in the first two bits.
RETURN
pointer to buffer after decoded LSN
*/
static uchar *translog_get_LSN_from_diff(LSN base_lsn, uchar *src, uchar *dst)
{
LSN lsn;
uint32 diff;
uint32 first_byte;
uint32 file_no, rec_offset;
uint8 code;
DBUG_ENTER("translog_get_LSN_from_diff");
DBUG_PRINT("enter", ("Base: (%lu,0x%lx) src: 0x%lx dst 0x%lx",
LSN_IN_PARTS(base_lsn), (ulong) src, (ulong) dst));
first_byte= *((uint8*) src);
code= first_byte >> 6; /* Length is in 2 most significant bits */
first_byte&= 0x3F;
src++; /* Skip length + encode */
file_no= LSN_FILE_NO(base_lsn); /* Assume relative */
DBUG_PRINT("info", ("code: %u first byte: %lu",
(uint) code, (ulong) first_byte));
switch (code) {
case 0:
if (first_byte == 0 && *((uint8*)src) == 1)
{
/*
It is full LSN after special 1 diff (which is impossible
in real life)
*/
memcpy(dst, src + 1, LSN_STORE_SIZE);
DBUG_PRINT("info", ("Special case of full LSN, new src: 0x%lx",
(ulong) (src + 1 + LSN_STORE_SIZE)));
DBUG_RETURN(src + 1 + LSN_STORE_SIZE);
}
rec_offset= LSN_OFFSET(base_lsn) - ((first_byte << 8) + *((uint8*)src));
break;
case 1:
diff= uint2korr(src);
rec_offset= LSN_OFFSET(base_lsn) - ((first_byte << 16) + diff);
break;
case 2:
diff= uint3korr(src);
rec_offset= LSN_OFFSET(base_lsn) - ((first_byte << 24) + diff);
break;
case 3:
{
ulonglong base_offset= LSN_OFFSET(base_lsn);
diff= uint4korr(src);
if (diff > LSN_OFFSET(base_lsn))
{
/* take 1 from file offset */
first_byte++;
base_offset+= LL(0x100000000);
}
file_no= LSN_FILE_NO(base_lsn) - first_byte;
DBUG_ASSERT(base_offset - diff <= UINT_MAX);
rec_offset= (uint32)(base_offset - diff);
break;
}
default:
DBUG_ASSERT(0);
DBUG_RETURN(NULL);
}
lsn= MAKE_LSN(file_no, rec_offset);
src+= code + 1;
lsn_store(dst, lsn);
DBUG_PRINT("info", ("new src: 0x%lx", (ulong) src));
DBUG_RETURN(src);
}
/**
@brief Encodes relative LSNs listed in the parameters.
@param parts Parts list with encoded LSN(s)
@param base_lsn LSN which is base for encoding
@param lsns number of LSN(s) to encode
@param compressed_LSNs buffer which can be used for storing compressed LSN(s)
*/
static void translog_relative_LSN_encode(struct st_translog_parts *parts,
LSN base_lsn,
uint lsns, uchar *compressed_LSNs)
{
LEX_CUSTRING *part;
uint lsns_len= lsns * LSN_STORE_SIZE;
uchar buffer_src[MAX_NUMBER_OF_LSNS_PER_RECORD * LSN_STORE_SIZE];
uchar *buffer= buffer_src;
const uchar *cbuffer;
DBUG_ENTER("translog_relative_LSN_encode");
DBUG_ASSERT(parts->current != 0);
part= parts->parts + parts->current;
/* collect all LSN(s) in one chunk if it (they) is (are) divided */
if (part->length < lsns_len)
{
uint copied= part->length;
LEX_CUSTRING *next_part;
DBUG_PRINT("info", ("Using buffer: 0x%lx", (ulong) compressed_LSNs));
memcpy(buffer, part->str, part->length);
next_part= parts->parts + parts->current + 1;
do
{
DBUG_ASSERT(next_part < parts->parts + parts->elements);
if ((next_part->length + copied) < lsns_len)
{
memcpy(buffer + copied, next_part->str,
next_part->length);
copied+= next_part->length;
next_part->length= 0; next_part->str= 0;
/* delete_dynamic_element(&parts->parts, parts->current + 1); */
next_part++;
parts->current++;
part= parts->parts + parts->current;
}
else
{
uint len= lsns_len - copied;
memcpy(buffer + copied, next_part->str, len);
copied= lsns_len;
next_part->str+= len;
next_part->length-= len;
}
} while (copied < lsns_len);
cbuffer= buffer;
}
else
{
cbuffer= part->str;
part->str+= lsns_len;
part->length-= lsns_len;
parts->current--;
part= parts->parts + parts->current;
}
{
/* Compress */
LSN ref;
int economy;
const uchar *src_ptr;
uchar *dst_ptr= compressed_LSNs + (MAX_NUMBER_OF_LSNS_PER_RECORD *
COMPRESSED_LSN_MAX_STORE_SIZE);
/*
We write the result in backward direction with no special sense or
tricks both directions are equal in complicity
*/
for (src_ptr= cbuffer + lsns_len - LSN_STORE_SIZE;
src_ptr >= (const uchar*)cbuffer;
src_ptr-= LSN_STORE_SIZE)
{
ref= lsn_korr(src_ptr);
dst_ptr= translog_put_LSN_diff(base_lsn, ref, dst_ptr);
}
part->length= (uint)((compressed_LSNs +
(MAX_NUMBER_OF_LSNS_PER_RECORD *
COMPRESSED_LSN_MAX_STORE_SIZE)) -
dst_ptr);
parts->record_length-= (economy= lsns_len - part->length);
DBUG_PRINT("info", ("new length of LSNs: %lu economy: %d",
(ulong)part->length, economy));
parts->total_record_length-= economy;
part->str= dst_ptr;
}
DBUG_VOID_RETURN;
}
/**
@brief Write multi-group variable-size record.
@param lsn LSN of the record will be written here
@param type the log record type
@param short_trid Short transaction ID or 0 if it has no sense
@param parts Descriptor of record source parts
@param buffer_to_flush Buffer which have to be flushed if it is not 0
@param header_length Header length calculated for 1 group
@param buffer_rest Beginning from which we plan to write in full pages
@param trn Transaction structure pointer for hooks by
record log type, for short_id
@param hook_arg Argument which will be passed to pre-write and
in-write hooks of this record.
@note
We must have a translog_lock() when entering this function
We must have buffer_to_flush locked (if not null)
buffer_to_flush should *NOT* be locked when calling this function.
(This is note is here as this is different from most other
translog_write...() functions which require the buffer to be locked)
@return Operation status
@retval 0 OK
@retval 1 Error
*/
static my_bool
translog_write_variable_record_mgroup(LSN *lsn,
enum translog_record_type type,
MARIA_HA *tbl_info,
SHORT_TRANSACTION_ID short_trid,
struct st_translog_parts *parts,
struct st_translog_buffer
*buffer_to_flush,
uint16 header_length,
translog_size_t buffer_rest,
TRN *trn, void *hook_arg)
{
TRANSLOG_ADDRESS horizon;
struct st_buffer_cursor cursor;
int rc= 0;
uint i, chunk2_page, full_pages;
uint curr_group= 0;
translog_size_t record_rest, first_page, chunk3_pages, chunk0_pages= 1;
translog_size_t done= 0;
struct st_translog_group_descriptor group;
DYNAMIC_ARRAY groups;
uint16 chunk3_size;
uint16 page_capacity= log_descriptor.page_capacity_chunk_2 + 1;
uint16 last_page_capacity;
my_bool new_page_before_chunk0= 1, first_chunk0= 1;
uchar chunk0_header[1 + 2 + 5 + 2 + 2], group_desc[7 + 1];
uchar chunk2_header[1];
uint header_fixed_part= header_length + 2;
uint groups_per_page= (page_capacity - header_fixed_part) / (7 + 1);
uint file_of_the_first_group;
int pages_to_skip;
struct st_translog_buffer *buffer_of_last_lsn;
DBUG_ENTER("translog_write_variable_record_mgroup");
translog_lock_assert_owner();
chunk2_header[0]= TRANSLOG_CHUNK_NOHDR;
if (my_init_dynamic_array(&groups,
sizeof(struct st_translog_group_descriptor),
10, 10))
{
translog_unlock();
DBUG_PRINT("error", ("init array failed"));
DBUG_RETURN(1);
}
first_page= translog_get_current_page_rest();
record_rest= parts->record_length - (first_page - 1);
DBUG_PRINT("info", ("Record Rest: %lu", (ulong) record_rest));
if (record_rest < buffer_rest)
{
/*
The record (group 1 type) is larger than the free space on the page
- we need to split it in two. But when we split it in two, the first
part is big enough to hold all the data of the record (because the
header of the first part of the split is smaller than the header of
the record as a whole when it takes only one chunk)
*/
DBUG_PRINT("info", ("too many free space because changing header"));
buffer_rest-= log_descriptor.page_capacity_chunk_2;
DBUG_ASSERT(record_rest >= buffer_rest);
}
file_of_the_first_group= LSN_FILE_NO(log_descriptor.horizon);
translog_mark_file_unfinished(file_of_the_first_group);
do
{
group.addr= horizon= log_descriptor.horizon;
cursor= log_descriptor.bc;
cursor.chaser= 1;
if ((full_pages= buffer_rest / log_descriptor.page_capacity_chunk_2) > 255)
{
/* sizeof(uint8) == 256 is max number of chunk in multi-chunks group */
full_pages= 255;
buffer_rest= full_pages * log_descriptor.page_capacity_chunk_2;
}
/*
group chunks =
full pages + first page (which actually can be full, too).
But here we assign number of chunks - 1
*/
group.num= full_pages;
if (insert_dynamic(&groups, (uchar*) &group))
{
DBUG_PRINT("error", ("insert into array failed"));
goto err_unlock;
}
DBUG_PRINT("info", ("chunk: #%u first_page: %u (%u) "
"full_pages: %lu (%lu) "
"Left %lu",
groups.elements,
first_page, first_page - 1,
(ulong) full_pages,
(ulong) (full_pages *
log_descriptor.page_capacity_chunk_2),
(ulong)(parts->record_length - (first_page - 1 +
buffer_rest) -
done)));
rc|= translog_advance_pointer((int)full_pages, 0);
translog_unlock();
if (buffer_to_flush != NULL)
{
translog_buffer_decrease_writers(buffer_to_flush);
if (!rc)
rc= translog_buffer_flush(buffer_to_flush);
translog_buffer_unlock(buffer_to_flush);
buffer_to_flush= NULL;
}
if (rc)
{
DBUG_PRINT("error", ("flush of unlock buffer failed"));
goto err;
}
translog_write_data_on_page(&horizon, &cursor, 1, chunk2_header);
translog_write_parts_on_page(&horizon, &cursor, first_page - 1, parts);
DBUG_PRINT("info", ("absolute horizon: (%lu,0x%lx) local: (%lu,0x%lx) "
"Left %lu",
LSN_IN_PARTS(log_descriptor.horizon),
LSN_IN_PARTS(horizon),
(ulong) (parts->record_length - (first_page - 1) -
done)));
for (i= 0; i < full_pages; i++)
{
if (translog_write_variable_record_chunk2_page(parts, &horizon, &cursor))
goto err;
DBUG_PRINT("info", ("absolute horizon: (%lu,0x%lx) "
"local: (%lu,0x%lx) "
"Left: %lu",
LSN_IN_PARTS(log_descriptor.horizon),
LSN_IN_PARTS(horizon),
(ulong) (parts->record_length - (first_page - 1) -
i * log_descriptor.page_capacity_chunk_2 -
done)));
}
done+= (first_page - 1 + buffer_rest);
if (translog_chaser_page_next(&horizon, &cursor))
{
DBUG_PRINT("error", ("flush of unlock buffer failed"));
goto err;
}
translog_buffer_lock(cursor.buffer);
translog_buffer_decrease_writers(cursor.buffer);
translog_buffer_unlock(cursor.buffer);
translog_lock();
/* Check that we have place for chunk type 2 */
first_page= translog_get_current_page_rest();
if (first_page <= 1)
{
if (translog_page_next(&log_descriptor.horizon, &log_descriptor.bc,
&buffer_to_flush))
goto err_unlock;
first_page= translog_get_current_page_rest();
}
buffer_rest= translog_get_current_group_size();
} while ((translog_size_t)(first_page + buffer_rest) <
(translog_size_t)(parts->record_length - done));
group.addr= horizon= log_descriptor.horizon;
cursor= log_descriptor.bc;
cursor.chaser= 1;
group.num= 0; /* 0 because it does not matter */
if (insert_dynamic(&groups, (uchar*) &group))
{
DBUG_PRINT("error", ("insert into array failed"));
goto err_unlock;
}
record_rest= parts->record_length - done;
DBUG_PRINT("info", ("Record rest: %lu", (ulong) record_rest));
if (first_page > record_rest + 1)
{
/*
We have not so much data to fill all first page
(no speaking about full pages)
so it will be:
<chunk0 <data>>
or
<chunk0>...<chunk0><chunk0 <data>>
or
<chunk3 <data>><chunk0>...<chunk0><chunk0 <possible data of 1 byte>>
*/
chunk2_page= full_pages= 0;
last_page_capacity= first_page;
pages_to_skip= -1;
}
else
{
/*
We will have:
<chunk2 <data>>...<chunk2 <data>><chunk0 <data>>
or
<chunk2 <data>>...<chunk2 <data>><chunk0>...<chunk0><chunk0 <data>>
or
<chunk3 <data>><chunk0>...<chunk0><chunk0 <possible data of 1 byte>>
*/
chunk2_page= 1;
record_rest-= (first_page - 1);
pages_to_skip= full_pages=
record_rest / log_descriptor.page_capacity_chunk_2;
record_rest= (record_rest % log_descriptor.page_capacity_chunk_2);
last_page_capacity= page_capacity;
}
chunk3_size= 0;
chunk3_pages= 0;
if (last_page_capacity > record_rest + 1 && record_rest != 0)
{
if (last_page_capacity >
record_rest + header_fixed_part + groups.elements * (7 + 1))
{
/* 1 record of type 0 */
chunk3_pages= 0;
}
else
{
pages_to_skip++;
chunk3_pages= 1;
if (record_rest + 2 == last_page_capacity)
{
chunk3_size= record_rest - 1;
record_rest= 1;
}
else
{
chunk3_size= record_rest;
record_rest= 0;
}
}
}
/*
A first non-full page will hold type 0 chunk only if it fit in it with
all its headers
*/
while (page_capacity <
record_rest + header_fixed_part +
(groups.elements - groups_per_page * (chunk0_pages - 1)) * (7 + 1))
chunk0_pages++;
DBUG_PRINT("info", ("chunk0_pages: %u groups %u groups per full page: %u "
"Group on last page: %u",
chunk0_pages, groups.elements,
groups_per_page,
(groups.elements -
((page_capacity - header_fixed_part) / (7 + 1)) *
(chunk0_pages - 1))));
DBUG_PRINT("info", ("first_page: %u chunk2: %u full_pages: %u (%lu) "
"chunk3: %u (%u) rest: %u",
first_page,
chunk2_page, full_pages,
(ulong) full_pages *
log_descriptor.page_capacity_chunk_2,
chunk3_pages, (uint) chunk3_size, (uint) record_rest));
rc= translog_advance_pointer(pages_to_skip + (int)(chunk0_pages - 1),
record_rest + header_fixed_part +
(groups.elements -
((page_capacity -
header_fixed_part) / (7 + 1)) *
(chunk0_pages - 1)) * (7 + 1));
buffer_of_last_lsn= log_descriptor.bc.buffer;
translog_unlock();
if (buffer_to_flush != NULL)
{
translog_buffer_decrease_writers(buffer_to_flush);
if (!rc)
rc= translog_buffer_flush(buffer_to_flush);
translog_buffer_unlock(buffer_to_flush);
buffer_to_flush= NULL;
}
if (rc)
{
DBUG_PRINT("error", ("flush of unlock buffer failed"));
goto err;
}
if (rc)
goto err;
if (chunk2_page)
{
DBUG_PRINT("info", ("chunk 2 to finish first page"));
translog_write_data_on_page(&horizon, &cursor, 1, chunk2_header);
translog_write_parts_on_page(&horizon, &cursor, first_page - 1, parts);
DBUG_PRINT("info", ("absolute horizon: (%lu,0x%lx) local: (%lu,0x%lx) "
"Left: %lu",
LSN_IN_PARTS(log_descriptor.horizon),
LSN_IN_PARTS(horizon),
(ulong) (parts->record_length - (first_page - 1) -
done)));
}
else if (chunk3_pages)
{
uchar chunk3_header[3];
DBUG_PRINT("info", ("chunk 3"));
DBUG_ASSERT(full_pages == 0);
chunk3_pages= 0;
chunk3_header[0]= TRANSLOG_CHUNK_LNGTH;
int2store(chunk3_header + 1, chunk3_size);
translog_write_data_on_page(&horizon, &cursor, 3, chunk3_header);
translog_write_parts_on_page(&horizon, &cursor, chunk3_size, parts);
DBUG_PRINT("info", ("absolute horizon: (%lu,0x%lx) local: (%lu,0x%lx) "
"Left: %lu",
LSN_IN_PARTS(log_descriptor.horizon),
LSN_IN_PARTS(horizon),
(ulong) (parts->record_length - chunk3_size - done)));
}
else
{
DBUG_PRINT("info", ("no new_page_before_chunk0"));
new_page_before_chunk0= 0;
}
for (i= 0; i < full_pages; i++)
{
DBUG_ASSERT(chunk2_page != 0);
if (translog_write_variable_record_chunk2_page(parts, &horizon, &cursor))
goto err;
DBUG_PRINT("info", ("absolute horizon: (%lu,0x%lx) local: (%lu,0x%lx) "
"Left: %lu",
LSN_IN_PARTS(log_descriptor.horizon),
LSN_IN_PARTS(horizon),
(ulong) (parts->record_length - (first_page - 1) -
i * log_descriptor.page_capacity_chunk_2 -
done)));
}
if (chunk3_pages &&
translog_write_variable_record_chunk3_page(parts,
chunk3_size,
&horizon, &cursor))
goto err;
DBUG_PRINT("info", ("absolute horizon: (%lu,0x%lx) local: (%lu,0x%lx)",
LSN_IN_PARTS(log_descriptor.horizon),
LSN_IN_PARTS(horizon)));
*chunk0_header= (uchar) (type | TRANSLOG_CHUNK_LSN);
int2store(chunk0_header + 1, short_trid);
translog_write_variable_record_1group_code_len(chunk0_header + 3,
parts->record_length,
header_length);
do
{
int limit;
if (new_page_before_chunk0 &&
translog_chaser_page_next(&horizon, &cursor))
{
DBUG_PRINT("error", ("flush of unlock buffer failed"));
goto err;
}
new_page_before_chunk0= 1;
if (first_chunk0)
{
first_chunk0= 0;
/*
We can drop "log_descriptor.is_everything_flushed" earlier when have
lock on loghandler and assign initial value of "horizon" variable or
before unlocking loghandler (because we will increase writers
counter on the buffer and every thread which wanted flush the buffer
will wait till we finish with it). But IMHO better here take short
lock and do not bother other threads with waiting.
*/
translog_lock();
set_lsn(lsn, horizon);
buffer_of_last_lsn->last_lsn= *lsn;
DBUG_PRINT("info", ("last_lsn set to (%lu,0x%lx) buffer: 0x%lx",
LSN_IN_PARTS(buffer_of_last_lsn->last_lsn),
(ulong) buffer_of_last_lsn));
if (log_record_type_descriptor[type].inwrite_hook &&
(*log_record_type_descriptor[type].inwrite_hook) (type, trn,
tbl_info,
lsn, hook_arg))
goto err_unlock;
translog_unlock();
}
/*
A first non-full page will hold type 0 chunk only if it fit in it with
all its headers => the fist page is full or number of groups less then
possible number of full page.
*/
limit= (groups_per_page < groups.elements - curr_group ?
groups_per_page : groups.elements - curr_group);
DBUG_PRINT("info", ("Groups: %u curr: %u limit: %u",
(uint) groups.elements, (uint) curr_group,
(uint) limit));
if (chunk0_pages == 1)
{
DBUG_PRINT("info", ("chunk_len: 2 + %u * (7+1) + %u = %u",
(uint) limit, (uint) record_rest,
(uint) (2 + limit * (7 + 1) + record_rest)));
int2store(chunk0_header + header_length - 2,
2 + limit * (7 + 1) + record_rest);
}
else
{
DBUG_PRINT("info", ("chunk_len: 2 + %u * (7+1) = %u",
(uint) limit, (uint) (2 + limit * (7 + 1))));
int2store(chunk0_header + header_length - 2, 2 + limit * (7 + 1));
}
int2store(chunk0_header + header_length, groups.elements - curr_group);
translog_write_data_on_page(&horizon, &cursor, header_fixed_part,
chunk0_header);
for (i= curr_group; i < limit + curr_group; i++)
{
struct st_translog_group_descriptor *grp_ptr;
grp_ptr= dynamic_element(&groups, i,
struct st_translog_group_descriptor *);
lsn_store(group_desc, grp_ptr->addr);
group_desc[7]= grp_ptr->num;
translog_write_data_on_page(&horizon, &cursor, (7 + 1), group_desc);
}
if (chunk0_pages == 1 && record_rest != 0)
translog_write_parts_on_page(&horizon, &cursor, record_rest, parts);
chunk0_pages--;
curr_group+= limit;
/* put special type to indicate that it is not LSN chunk */
*chunk0_header= (uchar) (TRANSLOG_CHUNK_LSN | TRANSLOG_CHUNK_0_CONT);
} while (chunk0_pages != 0);
translog_buffer_lock(cursor.buffer);
translog_buffer_decrease_writers(cursor.buffer);
translog_buffer_unlock(cursor.buffer);
rc= 0;
if (translog_set_lsn_for_files(file_of_the_first_group, LSN_FILE_NO(*lsn),
*lsn, FALSE))
goto err;
translog_mark_file_finished(file_of_the_first_group);
delete_dynamic(&groups);
DBUG_RETURN(rc);
err_unlock:
translog_unlock();
err:
if (buffer_to_flush != NULL)
{
/* This is to prevent locking buffer forever in case of error */
translog_buffer_decrease_writers(buffer_to_flush);
if (!rc)
rc= translog_buffer_flush(buffer_to_flush);
translog_buffer_unlock(buffer_to_flush);
buffer_to_flush= NULL;
}
translog_mark_file_finished(file_of_the_first_group);
delete_dynamic(&groups);
DBUG_RETURN(1);
}
/**
@brief Write the variable length log record.
@param lsn LSN of the record will be written here
@param type the log record type
@param short_trid Short transaction ID or 0 if it has no sense
@param parts Descriptor of record source parts
@param trn Transaction structure pointer for hooks by
record log type, for short_id
@param hook_arg Argument which will be passed to pre-write and
in-write hooks of this record.
@return Operation status
@retval 0 OK
@retval 1 Error
*/
static my_bool translog_write_variable_record(LSN *lsn,
enum translog_record_type type,
MARIA_HA *tbl_info,
SHORT_TRANSACTION_ID short_trid,
struct st_translog_parts *parts,
TRN *trn, void *hook_arg)
{
struct st_translog_buffer *buffer_to_flush= NULL;
uint header_length1= 1 + 2 + 2 +
translog_variable_record_length_bytes(parts->record_length);
ulong buffer_rest;
uint page_rest;
/* Max number of such LSNs per record is 2 */
uchar compressed_LSNs[MAX_NUMBER_OF_LSNS_PER_RECORD *
COMPRESSED_LSN_MAX_STORE_SIZE];
my_bool res;
DBUG_ENTER("translog_write_variable_record");
translog_lock();
DBUG_PRINT("info", ("horizon: (%lu,0x%lx)",
LSN_IN_PARTS(log_descriptor.horizon)));
page_rest= TRANSLOG_PAGE_SIZE - log_descriptor.bc.current_page_fill;
DBUG_PRINT("info", ("header length: %u page_rest: %u",
header_length1, page_rest));
/*
header and part which we should read have to fit in one chunk
TODO: allow to divide readable header
*/
if (page_rest <
(header_length1 + log_record_type_descriptor[type].read_header_len))
{
DBUG_PRINT("info",
("Next page, size: %u header: %u + %u",
log_descriptor.bc.current_page_fill,
header_length1,
log_record_type_descriptor[type].read_header_len));
translog_page_next(&log_descriptor.horizon, &log_descriptor.bc,
&buffer_to_flush);
/* Chunk 2 header is 1 byte, so full page capacity will be one uchar more */
page_rest= log_descriptor.page_capacity_chunk_2 + 1;
DBUG_PRINT("info", ("page_rest: %u", page_rest));
}
/*
To minimize compressed size we will compress always relative to
very first chunk address (log_descriptor.horizon for now)
*/
if (log_record_type_descriptor[type].compressed_LSN > 0)
{
translog_relative_LSN_encode(parts, log_descriptor.horizon,
log_record_type_descriptor[type].
compressed_LSN, compressed_LSNs);
/* recalculate header length after compression */
header_length1= 1 + 2 + 2 +
translog_variable_record_length_bytes(parts->record_length);
DBUG_PRINT("info", ("after compressing LSN(s) header length: %u "
"record length: %lu",
header_length1, (ulong)parts->record_length));
}
/* TODO: check space on current page for header + few bytes */
if (page_rest >= parts->record_length + header_length1)
{
/* following function makes translog_unlock(); */
res= translog_write_variable_record_1chunk(lsn, type, tbl_info,
short_trid,
parts, buffer_to_flush,
header_length1, trn, hook_arg);
DBUG_RETURN(res);
}
buffer_rest= translog_get_current_group_size();
if (buffer_rest >= parts->record_length + header_length1 - page_rest)
{
/* following function makes translog_unlock(); */
res= translog_write_variable_record_1group(lsn, type, tbl_info,
short_trid,
parts, buffer_to_flush,
header_length1, trn, hook_arg);
DBUG_RETURN(res);
}
/* following function makes translog_unlock(); */
res= translog_write_variable_record_mgroup(lsn, type, tbl_info,
short_trid,
parts, buffer_to_flush,
header_length1,
buffer_rest, trn, hook_arg);
DBUG_RETURN(res);
}
/**
@brief Write the fixed and pseudo-fixed log record.
@param lsn LSN of the record will be written here
@param type the log record type
@param short_trid Short transaction ID or 0 if it has no sense
@param parts Descriptor of record source parts
@param trn Transaction structure pointer for hooks by
record log type, for short_id
@param hook_arg Argument which will be passed to pre-write and
in-write hooks of this record.
@return Operation status
@retval 0 OK
@retval 1 Error
*/
static my_bool translog_write_fixed_record(LSN *lsn,
enum translog_record_type type,
MARIA_HA *tbl_info,
SHORT_TRANSACTION_ID short_trid,
struct st_translog_parts *parts,
TRN *trn, void *hook_arg)
{
struct st_translog_buffer *buffer_to_flush= NULL;
uchar chunk1_header[1 + 2];
/* Max number of such LSNs per record is 2 */
uchar compressed_LSNs[MAX_NUMBER_OF_LSNS_PER_RECORD *
COMPRESSED_LSN_MAX_STORE_SIZE];
LEX_CUSTRING *part;
int rc= 1;
DBUG_ENTER("translog_write_fixed_record");
DBUG_ASSERT((log_record_type_descriptor[type].rclass ==
LOGRECTYPE_FIXEDLENGTH &&
parts->record_length ==
log_record_type_descriptor[type].fixed_length) ||
(log_record_type_descriptor[type].rclass ==
LOGRECTYPE_PSEUDOFIXEDLENGTH &&
parts->record_length ==
log_record_type_descriptor[type].fixed_length));
translog_lock();
DBUG_PRINT("info", ("horizon: (%lu,0x%lx)",
LSN_IN_PARTS(log_descriptor.horizon)));
DBUG_ASSERT(log_descriptor.bc.current_page_fill <= TRANSLOG_PAGE_SIZE);
DBUG_PRINT("info",
("Page size: %u record: %u next cond: %d",
log_descriptor.bc.current_page_fill,
(parts->record_length +
log_record_type_descriptor[type].compressed_LSN * 2 + 3),
((((uint) log_descriptor.bc.current_page_fill) +
(parts->record_length +
log_record_type_descriptor[type].compressed_LSN * 2 + 3)) >
TRANSLOG_PAGE_SIZE)));
/*
check that there is enough place on current page.
NOTE: compressing may increase page LSN size on two bytes for every LSN
*/
if ((((uint) log_descriptor.bc.current_page_fill) +
(parts->record_length +
log_record_type_descriptor[type].compressed_LSN * 2 + 3)) >
TRANSLOG_PAGE_SIZE)
{
DBUG_PRINT("info", ("Next page"));
if (translog_page_next(&log_descriptor.horizon, &log_descriptor.bc,
&buffer_to_flush))
goto err; /* rc == 1 */
if (buffer_to_flush)
translog_buffer_lock_assert_owner(buffer_to_flush);
}
set_lsn(lsn, log_descriptor.horizon);
if (translog_set_lsn_for_files(LSN_FILE_NO(*lsn), LSN_FILE_NO(*lsn),
*lsn, TRUE) ||
(log_record_type_descriptor[type].inwrite_hook &&
(*log_record_type_descriptor[type].inwrite_hook)(type, trn, tbl_info,
lsn, hook_arg)))
goto err;
/* compress LSNs */
if (log_record_type_descriptor[type].rclass ==
LOGRECTYPE_PSEUDOFIXEDLENGTH)
{
DBUG_ASSERT(log_record_type_descriptor[type].compressed_LSN > 0);
translog_relative_LSN_encode(parts, *lsn,
log_record_type_descriptor[type].
compressed_LSN, compressed_LSNs);
}
/*
Write the whole record at once (we know that there is enough place on
the destination page)
*/
DBUG_ASSERT(parts->current != 0); /* first part is left for header */
part= parts->parts + (--parts->current);
parts->total_record_length+= (translog_size_t) (part->length= 1 + 2);
part->str= chunk1_header;
*chunk1_header= (uchar) (type | TRANSLOG_CHUNK_FIXED);
int2store(chunk1_header + 1, short_trid);
rc= translog_write_parts_on_page(&log_descriptor.horizon,
&log_descriptor.bc,
parts->total_record_length, parts);
log_descriptor.bc.buffer->last_lsn= *lsn;
DBUG_PRINT("info", ("last_lsn set to (%lu,0x%lx) buffer: 0x%lx",
LSN_IN_PARTS(log_descriptor.bc.buffer->last_lsn),
(ulong) log_descriptor.bc.buffer));
err:
translog_unlock();
/*
check if we switched buffer and need process it (current buffer is
unlocked already => we will not delay other threads
*/
if (buffer_to_flush != NULL)
{
if (!rc)
rc= translog_buffer_flush(buffer_to_flush);
translog_buffer_unlock(buffer_to_flush);
}
DBUG_RETURN(rc);
}
/**
@brief Writes the log record
If share has no 2-byte-id yet, gives an id to the share and logs
LOGREC_FILE_ID. If transaction has not logged LOGREC_LONG_TRANSACTION_ID
yet, logs it.
@param lsn LSN of the record will be written here
@param type the log record type
@param trn Transaction structure pointer for hooks by
record log type, for short_id
@param tbl_info MARIA_HA of table or NULL
@param rec_len record length or 0 (count it)
@param part_no number of parts or 0 (count it)
@param parts_data zero ended (in case of number of parts is 0)
array of LEX_STRINGs (parts), first
TRANSLOG_INTERNAL_PARTS positions in the log
should be unused (need for loghandler)
@param store_share_id if tbl_info!=NULL then share's id will
automatically be stored in the two first bytes
pointed (so pointer is assumed to be !=NULL)
@param hook_arg argument which will be passed to pre-write and
in-write hooks of this record.
@return Operation status
@retval 0 OK
@retval 1 Error
*/
my_bool translog_write_record(LSN *lsn,
enum translog_record_type type,
TRN *trn, MARIA_HA *tbl_info,
translog_size_t rec_len,
uint part_no,
LEX_CUSTRING *parts_data,
uchar *store_share_id,
void *hook_arg)
{
struct st_translog_parts parts;
LEX_CUSTRING *part;
int rc;
uint short_trid= trn->short_id;
DBUG_ENTER("translog_write_record");
DBUG_PRINT("enter", ("type: %u (%s) ShortTrID: %u rec_len: %lu",
(uint) type, log_record_type_descriptor[type].name,
(uint) short_trid, (ulong) rec_len));
DBUG_ASSERT(translog_status == TRANSLOG_OK ||
translog_status == TRANSLOG_READONLY);
DBUG_ASSERT(type != 0);
DBUG_ASSERT((uint)type <= max_allowed_translog_type);
if (unlikely(translog_status != TRANSLOG_OK))
{
DBUG_PRINT("error", ("Transaction log is write protected"));
DBUG_RETURN(1);
}
if (tbl_info && type != LOGREC_FILE_ID)
{
MARIA_SHARE *share= tbl_info->s;
DBUG_ASSERT(share->now_transactional);
if (unlikely(share->id == 0))
{
/*
First log write for this MARIA_SHARE; give it a short id.
When the lock manager is enabled and needs a short id, it should be
assigned in the lock manager (because row locks will be taken before
log records are written; for example SELECT FOR UPDATE takes locks but
writes no log record.
*/
if (unlikely(translog_assign_id_to_share(tbl_info, trn)))
DBUG_RETURN(1);
}
fileid_store(store_share_id, share->id);
}
if (unlikely(!(trn->first_undo_lsn & TRANSACTION_LOGGED_LONG_ID)))
{
LSN dummy_lsn;
LEX_CUSTRING log_array[TRANSLOG_INTERNAL_PARTS + 1];
uchar log_data[6];
DBUG_ASSERT(trn->undo_lsn == LSN_IMPOSSIBLE);
int6store(log_data, trn->trid);
log_array[TRANSLOG_INTERNAL_PARTS + 0].str= log_data;
log_array[TRANSLOG_INTERNAL_PARTS + 0].length= sizeof(log_data);
trn->first_undo_lsn|= TRANSACTION_LOGGED_LONG_ID; /* no recursion */
if (unlikely(translog_write_record(&dummy_lsn, LOGREC_LONG_TRANSACTION_ID,
trn, NULL, sizeof(log_data),
sizeof(log_array)/sizeof(log_array[0]),
log_array, NULL, NULL)))
DBUG_RETURN(1);
}
parts.parts= parts_data;
/* count parts if they are not counted by upper level */
if (part_no == 0)
{
for (part_no= TRANSLOG_INTERNAL_PARTS;
parts_data[part_no].length != 0;
part_no++);
}
parts.elements= part_no;
parts.current= TRANSLOG_INTERNAL_PARTS;
/* clear TRANSLOG_INTERNAL_PARTS */
compile_time_assert(TRANSLOG_INTERNAL_PARTS != 0);
parts_data[0].str= 0;
parts_data[0].length= 0;
/* count length of the record */
if (rec_len == 0)
{
for(part= parts_data + TRANSLOG_INTERNAL_PARTS;\
part < parts_data + part_no;
part++)
{
rec_len+= (translog_size_t) part->length;
}
}
parts.record_length= rec_len;
#ifndef DBUG_OFF
{
uint i;
uint len= 0;
#ifdef HAVE_valgrind
ha_checksum checksum= 0;
#endif
for (i= TRANSLOG_INTERNAL_PARTS; i < part_no; i++)
{
#ifdef HAVE_valgrind
/* Find unitialized bytes early */
checksum+= my_checksum(checksum, parts_data[i].str,
parts_data[i].length);
#endif
len+= parts_data[i].length;
}
DBUG_ASSERT(len == rec_len);
}
#endif
/*
Start total_record_length from record_length then overhead will
be add
*/
parts.total_record_length= parts.record_length;
DBUG_PRINT("info", ("record length: %lu", (ulong) parts.record_length));
/* process this parts */
if (!(rc= (log_record_type_descriptor[type].prewrite_hook &&
(*log_record_type_descriptor[type].prewrite_hook)(type, trn,
tbl_info,
hook_arg))))
{
switch (log_record_type_descriptor[type].rclass) {
case LOGRECTYPE_VARIABLE_LENGTH:
rc= translog_write_variable_record(lsn, type, tbl_info,
short_trid, &parts, trn, hook_arg);
break;
case LOGRECTYPE_PSEUDOFIXEDLENGTH:
case LOGRECTYPE_FIXEDLENGTH:
rc= translog_write_fixed_record(lsn, type, tbl_info,
short_trid, &parts, trn, hook_arg);
break;
case LOGRECTYPE_NOT_ALLOWED:
DBUG_ASSERT(0);
default:
DBUG_ASSERT(0);
rc= 1;
}
}
DBUG_PRINT("info", ("LSN: (%lu,0x%lx)", LSN_IN_PARTS(*lsn)));
DBUG_RETURN(rc);
}
/*
Decode compressed (relative) LSN(s)
SYNOPSIS
translog_relative_lsn_decode()
base_lsn LSN for encoding
src Decode LSN(s) from here
dst Put decoded LSNs here
lsns number of LSN(s)
RETURN
position in sources after decoded LSN(s)
*/
static uchar *translog_relative_LSN_decode(LSN base_lsn,
uchar *src, uchar *dst, uint lsns)
{
uint i;
for (i= 0; i < lsns; i++, dst+= LSN_STORE_SIZE)
{
src= translog_get_LSN_from_diff(base_lsn, src, dst);
}
return src;
}
/**
@brief Get header of fixed/pseudo length record and call hook for
it processing
@param page Pointer to the buffer with page where LSN chunk is
placed
@param page_offset Offset of the first chunk in the page
@param buff Buffer to be filled with header data
@return Length of header or operation status
@retval # number of bytes in TRANSLOG_HEADER_BUFFER::header where
stored decoded part of the header
*/
static int translog_fixed_length_header(uchar *page,
translog_size_t page_offset,
TRANSLOG_HEADER_BUFFER *buff)
{
struct st_log_record_type_descriptor *desc=
log_record_type_descriptor + buff->type;
uchar *src= page + page_offset + 3;
uchar *dst= buff->header;
uchar *start= src;
int lsns= desc->compressed_LSN;
uint length= desc->fixed_length;
DBUG_ENTER("translog_fixed_length_header");
buff->record_length= length;
if (desc->rclass == LOGRECTYPE_PSEUDOFIXEDLENGTH)
{
DBUG_ASSERT(lsns > 0);
src= translog_relative_LSN_decode(buff->lsn, src, dst, lsns);
lsns*= LSN_STORE_SIZE;
dst+= lsns;
length-= lsns;
buff->compressed_LSN_economy= (lsns - (int) (src - start));
}
else
buff->compressed_LSN_economy= 0;
memcpy(dst, src, length);
buff->non_header_data_start_offset= (uint16) (page_offset +
((src + length) -
(page + page_offset)));
buff->non_header_data_len= 0;
DBUG_RETURN(buff->record_length);
}
/*
Free resources used by TRANSLOG_HEADER_BUFFER
SYNOPSIS
translog_free_record_header();
*/
void translog_free_record_header(TRANSLOG_HEADER_BUFFER *buff)
{
DBUG_ENTER("translog_free_record_header");
if (buff->groups_no != 0)
{
my_free(buff->groups);
buff->groups_no= 0;
}
DBUG_VOID_RETURN;
}
/**
@brief Returns the current horizon at the end of the current log
@return Horizon
@retval LSN_ERROR error
@retvar # Horizon
*/
TRANSLOG_ADDRESS translog_get_horizon()
{
TRANSLOG_ADDRESS res;
DBUG_ASSERT(translog_status == TRANSLOG_OK ||
translog_status == TRANSLOG_READONLY);
translog_lock();
res= log_descriptor.horizon;
translog_unlock();
return res;
}
/**
@brief Returns the current horizon at the end of the current log, caller is
assumed to already hold the lock
@return Horizon
@retval LSN_ERROR error
@retvar # Horizon
*/
TRANSLOG_ADDRESS translog_get_horizon_no_lock()
{
DBUG_ASSERT(translog_status == TRANSLOG_OK ||
translog_status == TRANSLOG_READONLY);
translog_lock_assert_owner();
return log_descriptor.horizon;
}
/*
Set last page in the scanner data structure
SYNOPSIS
translog_scanner_set_last_page()
scanner Information about current chunk during scanning
RETURN
0 OK
1 Error
*/
static my_bool translog_scanner_set_last_page(TRANSLOG_SCANNER_DATA *scanner)
{
my_bool page_ok;
if (LSN_FILE_NO(scanner->page_addr) == LSN_FILE_NO(scanner->horizon))
{
/* It is last file => we can easy find last page address by horizon */
uint pagegrest= LSN_OFFSET(scanner->horizon) % TRANSLOG_PAGE_SIZE;
scanner->last_file_page= (scanner->horizon -
(pagegrest ? pagegrest : TRANSLOG_PAGE_SIZE));
return (0);
}
scanner->last_file_page= scanner->page_addr;
return (translog_get_last_page_addr(&scanner->last_file_page, &page_ok, 0));
}
/**
@brief Get page from page cache according to requested method
@param scanner The scanner data
@return operation status
@retval 0 OK
@retval 1 Error
*/
static my_bool
translog_scanner_get_page(TRANSLOG_SCANNER_DATA *scanner)
{
TRANSLOG_VALIDATOR_DATA data;
DBUG_ENTER("translog_scanner_get_page");
data.addr= &scanner->page_addr;
data.was_recovered= 0;
DBUG_RETURN((scanner->page=
translog_get_page(&data, scanner->buffer,
(scanner->use_direct_link ?
&scanner->direct_link :
NULL))) ==
NULL);
}
/**
@brief Initialize reader scanner.
@param lsn LSN with which it have to be inited
@param fixed_horizon true if it is OK do not read records which was written
after scanning beginning
@param scanner scanner which have to be inited
@param use_direct prefer using direct lings from page handler
where it is possible.
@note If direct link was used translog_destroy_scanner should be
called after it using
@return status of the operation
@retval 0 OK
@retval 1 Error
*/
my_bool translog_scanner_init(LSN lsn,
my_bool fixed_horizon,
TRANSLOG_SCANNER_DATA *scanner,
my_bool use_direct)
{
TRANSLOG_VALIDATOR_DATA data;
DBUG_ENTER("translog_scanner_init");
DBUG_PRINT("enter", ("Scanner: 0x%lx LSN: (%lu,0x%lx)",
(ulong) scanner, LSN_IN_PARTS(lsn)));
DBUG_ASSERT(translog_status == TRANSLOG_OK ||
translog_status == TRANSLOG_READONLY);
data.addr= &scanner->page_addr;
data.was_recovered= 0;
scanner->page_offset= LSN_OFFSET(lsn) % TRANSLOG_PAGE_SIZE;
scanner->fixed_horizon= fixed_horizon;
scanner->use_direct_link= use_direct;
scanner->direct_link= NULL;
scanner->horizon= translog_get_horizon();
DBUG_PRINT("info", ("horizon: (%lu,0x%lx)", LSN_IN_PARTS(scanner->horizon)));
/* lsn < horizon */
DBUG_ASSERT(lsn <= scanner->horizon);
scanner->page_addr= lsn;
scanner->page_addr-= scanner->page_offset; /*decrease offset */
if (translog_scanner_set_last_page(scanner))
DBUG_RETURN(1);
if (translog_scanner_get_page(scanner))
DBUG_RETURN(1);
DBUG_RETURN(0);
}
/**
@brief Destroy scanner object;
@param scanner The scanner object to destroy
*/
void translog_destroy_scanner(TRANSLOG_SCANNER_DATA *scanner)
{
DBUG_ENTER("translog_destroy_scanner");
DBUG_PRINT("enter", ("Scanner: 0x%lx", (ulong)scanner));
translog_free_link(scanner->direct_link);
DBUG_VOID_RETURN;
}
/*
Checks End of the Log
SYNOPSIS
translog_scanner_eol()
scanner Information about current chunk during scanning
RETURN
1 End of the Log
0 OK
*/
static my_bool translog_scanner_eol(TRANSLOG_SCANNER_DATA *scanner)
{
DBUG_ENTER("translog_scanner_eol");
DBUG_PRINT("enter",
("Horizon: (%lu, 0x%lx) Current: (%lu, 0x%lx+0x%x=0x%lx)",
LSN_IN_PARTS(scanner->horizon),
LSN_IN_PARTS(scanner->page_addr),
(uint) scanner->page_offset,
(ulong) (LSN_OFFSET(scanner->page_addr) + scanner->page_offset)));
if (scanner->horizon > (scanner->page_addr +
scanner->page_offset))
{
DBUG_PRINT("info", ("Horizon is not reached"));
DBUG_RETURN(0);
}
if (scanner->fixed_horizon)
{
DBUG_PRINT("info", ("Horizon is fixed and reached"));
DBUG_RETURN(1);
}
scanner->horizon= translog_get_horizon();
DBUG_PRINT("info",
("Horizon is re-read, EOL: %d",
scanner->horizon <= (scanner->page_addr +
scanner->page_offset)));
DBUG_RETURN(scanner->horizon <= (scanner->page_addr +
scanner->page_offset));
}
/**
@brief Cheks End of the Page
@param scanner Information about current chunk during scanning
@retval 1 End of the Page
@retval 0 OK
*/
static my_bool translog_scanner_eop(TRANSLOG_SCANNER_DATA *scanner)
{
DBUG_ENTER("translog_scanner_eop");
DBUG_RETURN(scanner->page_offset >= TRANSLOG_PAGE_SIZE ||
scanner->page[scanner->page_offset] == TRANSLOG_FILLER);
}
/**
@brief Checks End of the File (i.e. we are scanning last page, which do not
mean end of this page)
@param scanner Information about current chunk during scanning
@retval 1 End of the File
@retval 0 OK
*/
static my_bool translog_scanner_eof(TRANSLOG_SCANNER_DATA *scanner)
{
DBUG_ENTER("translog_scanner_eof");
DBUG_ASSERT(LSN_FILE_NO(scanner->page_addr) ==
LSN_FILE_NO(scanner->last_file_page));
DBUG_PRINT("enter", ("curr Page: 0x%lx last page: 0x%lx "
"normal EOF: %d",
(ulong) LSN_OFFSET(scanner->page_addr),
(ulong) LSN_OFFSET(scanner->last_file_page),
LSN_OFFSET(scanner->page_addr) ==
LSN_OFFSET(scanner->last_file_page)));
/*
TODO: detect damaged file EOF,
TODO: issue warning if damaged file EOF detected
*/
DBUG_RETURN(scanner->page_addr ==
scanner->last_file_page);
}
/*
Move scanner to the next chunk
SYNOPSIS
translog_get_next_chunk()
scanner Information about current chunk during scanning
RETURN
0 OK
1 Error
*/
static my_bool
translog_get_next_chunk(TRANSLOG_SCANNER_DATA *scanner)
{
uint16 len;
DBUG_ENTER("translog_get_next_chunk");
if (translog_scanner_eop(scanner))
len= TRANSLOG_PAGE_SIZE - scanner->page_offset;
else if ((len= translog_get_total_chunk_length(scanner->page,
scanner->page_offset)) == 0)
DBUG_RETURN(1);
scanner->page_offset+= len;
if (translog_scanner_eol(scanner))
{
scanner->page= END_OF_LOG;
scanner->page_offset= 0;
DBUG_RETURN(0);
}
if (translog_scanner_eop(scanner))
{
/* before reading next page we should unpin current one if it was pinned */
translog_free_link(scanner->direct_link);
if (translog_scanner_eof(scanner))
{
DBUG_PRINT("info", ("horizon: (%lu,0x%lx) pageaddr: (%lu,0x%lx)",
LSN_IN_PARTS(scanner->horizon),
LSN_IN_PARTS(scanner->page_addr)));
/* if it is log end it have to be caught before */
DBUG_ASSERT(LSN_FILE_NO(scanner->horizon) >
LSN_FILE_NO(scanner->page_addr));
scanner->page_addr+= LSN_ONE_FILE;
scanner->page_addr= LSN_REPLACE_OFFSET(scanner->page_addr,
TRANSLOG_PAGE_SIZE);
if (translog_scanner_set_last_page(scanner))
DBUG_RETURN(1);
}
else
{
scanner->page_addr+= TRANSLOG_PAGE_SIZE; /* offset increased */
}
if (translog_scanner_get_page(scanner))
DBUG_RETURN(1);
scanner->page_offset= translog_get_first_chunk_offset(scanner->page);
if (translog_scanner_eol(scanner))
{
scanner->page= END_OF_LOG;
scanner->page_offset= 0;
DBUG_RETURN(0);
}
DBUG_ASSERT(scanner->page[scanner->page_offset] != TRANSLOG_FILLER);
}
DBUG_RETURN(0);
}
/**
@brief Get header of variable length record and call hook for it processing
@param page Pointer to the buffer with page where LSN chunk is
placed
@param page_offset Offset of the first chunk in the page
@param buff Buffer to be filled with header data
@param scanner If present should be moved to the header page if
it differ from LSN page
@return Length of header or operation status
@retval RECHEADER_READ_ERROR error
@retval RECHEADER_READ_EOF End of the log reached during the read
@retval # number of bytes in
TRANSLOG_HEADER_BUFFER::header where
stored decoded part of the header
*/
static int
translog_variable_length_header(uchar *page, translog_size_t page_offset,
TRANSLOG_HEADER_BUFFER *buff,
TRANSLOG_SCANNER_DATA *scanner)
{
struct st_log_record_type_descriptor *desc= (log_record_type_descriptor +
buff->type);
uchar *src= page + page_offset + 1 + 2;
uchar *dst= buff->header;
LSN base_lsn;
uint lsns= desc->compressed_LSN;
uint16 chunk_len;
uint16 length= desc->read_header_len;
uint16 buffer_length= length;
uint16 body_len;
int rc;
TRANSLOG_SCANNER_DATA internal_scanner;
DBUG_ENTER("translog_variable_length_header");
buff->record_length= translog_variable_record_1group_decode_len(&src);
chunk_len= uint2korr(src);
DBUG_PRINT("info", ("rec len: %lu chunk len: %u length: %u bufflen: %u",
(ulong) buff->record_length, (uint) chunk_len,
(uint) length, (uint) buffer_length));
if (chunk_len == 0)
{
uint16 page_rest;
DBUG_PRINT("info", ("1 group"));
src+= 2;
page_rest= (uint16) (TRANSLOG_PAGE_SIZE - (src - page));
base_lsn= buff->lsn;
body_len= min(page_rest, buff->record_length);
}
else
{
uint grp_no, curr;
uint header_to_skip;
uint16 page_rest;
DBUG_PRINT("info", ("multi-group"));
grp_no= buff->groups_no= uint2korr(src + 2);
if (!(buff->groups=
(TRANSLOG_GROUP*) my_malloc(sizeof(TRANSLOG_GROUP) * grp_no,
MYF(0))))
DBUG_RETURN(RECHEADER_READ_ERROR);
DBUG_PRINT("info", ("Groups: %u", (uint) grp_no));
src+= (2 + 2);
page_rest= (uint16) (TRANSLOG_PAGE_SIZE - (src - page));
curr= 0;
header_to_skip= src - (page + page_offset);
buff->chunk0_pages= 0;
for (;;)
{
uint i, read_length= grp_no;
buff->chunk0_pages++;
if (page_rest < grp_no * (7 + 1))
read_length= page_rest / (7 + 1);
DBUG_PRINT("info", ("Read chunk0 page#%u read: %u left: %u "
"start from: %u",
buff->chunk0_pages, read_length, grp_no, curr));
for (i= 0; i < read_length; i++, curr++)
{
DBUG_ASSERT(curr < buff->groups_no);
buff->groups[curr].addr= lsn_korr(src + i * (7 + 1));
buff->groups[curr].num= src[i * (7 + 1) + 7];
DBUG_PRINT("info", ("group #%u (%lu,0x%lx) chunks: %u",
curr,
LSN_IN_PARTS(buff->groups[curr].addr),
(uint) buff->groups[curr].num));
}
grp_no-= read_length;
if (grp_no == 0)
{
if (scanner)
{
buff->chunk0_data_addr= scanner->page_addr;
/* offset increased */
buff->chunk0_data_addr+= (page_offset + header_to_skip +
read_length * (7 + 1));
}
else
{
buff->chunk0_data_addr= buff->lsn;
/* offset increased */
buff->chunk0_data_addr+= (header_to_skip + read_length * (7 + 1));
}
buff->chunk0_data_len= chunk_len - 2 - read_length * (7 + 1);
DBUG_PRINT("info", ("Data address: (%lu,0x%lx) len: %u",
LSN_IN_PARTS(buff->chunk0_data_addr),
buff->chunk0_data_len));
break;
}
if (scanner == NULL)
{
DBUG_PRINT("info", ("use internal scanner for header reading"));
scanner= &internal_scanner;
if (translog_scanner_init(buff->lsn, 1, scanner, 0))
{
rc= RECHEADER_READ_ERROR;
goto exit_and_free;
}
}
if (translog_get_next_chunk(scanner))
{
if (scanner == &internal_scanner)
translog_destroy_scanner(scanner);
rc= RECHEADER_READ_ERROR;
goto exit_and_free;
}
if (scanner->page == END_OF_LOG)
{
if (scanner == &internal_scanner)
translog_destroy_scanner(scanner);
rc= RECHEADER_READ_EOF;
goto exit_and_free;
}
page= scanner->page;
page_offset= scanner->page_offset;
src= page + page_offset + header_to_skip;
chunk_len= uint2korr(src - 2 - 2);
DBUG_PRINT("info", ("Chunk len: %u", (uint) chunk_len));
page_rest= (uint16) (TRANSLOG_PAGE_SIZE - (src - page));
}
if (scanner == NULL)
{
DBUG_PRINT("info", ("use internal scanner"));
scanner= &internal_scanner;
}
else
{
translog_destroy_scanner(scanner);
}
base_lsn= buff->groups[0].addr;
translog_scanner_init(base_lsn, 1, scanner, scanner == &internal_scanner);
/* first group chunk is always chunk type 2 */
page= scanner->page;
page_offset= scanner->page_offset;
src= page + page_offset + 1;
page_rest= (uint16) (TRANSLOG_PAGE_SIZE - (src - page));
body_len= page_rest;
if (scanner == &internal_scanner)
translog_destroy_scanner(scanner);
}
if (lsns)
{
uchar *start= src;
src= translog_relative_LSN_decode(base_lsn, src, dst, lsns);
lsns*= LSN_STORE_SIZE;
dst+= lsns;
length-= lsns;
buff->record_length+= (buff->compressed_LSN_economy=
(int) (lsns - (src - start)));
DBUG_PRINT("info", ("lsns: %u length: %u economy: %d new length: %lu",
lsns / LSN_STORE_SIZE, (uint) length,
(int) buff->compressed_LSN_economy,
(ulong) buff->record_length));
body_len-= (uint16) (src - start);
}
else
buff->compressed_LSN_economy= 0;
DBUG_ASSERT(body_len >= length);
body_len-= length;
memcpy(dst, src, length);
buff->non_header_data_start_offset= (uint16) (src + length - page);
buff->non_header_data_len= body_len;
DBUG_PRINT("info", ("non_header_data_start_offset: %u len: %u buffer: %u",
buff->non_header_data_start_offset,
buff->non_header_data_len, buffer_length));
DBUG_RETURN(buffer_length);
exit_and_free:
my_free(buff->groups);
buff->groups_no= 0; /* prevent try to use of buff->groups */
DBUG_RETURN(rc);
}
/**
@brief Read record header from the given buffer
@param page page content buffer
@param page_offset offset of the chunk in the page
@param buff destination buffer
@param scanner If this is set the scanner will be moved to the
record header page (differ from LSN page in case of
multi-group records)
@return Length of header or operation status
@retval RECHEADER_READ_ERROR error
@retval # number of bytes in
TRANSLOG_HEADER_BUFFER::header where
stored decoded part of the header
*/
int translog_read_record_header_from_buffer(uchar *page,
uint16 page_offset,
TRANSLOG_HEADER_BUFFER *buff,
TRANSLOG_SCANNER_DATA *scanner)
{
translog_size_t res;
DBUG_ENTER("translog_read_record_header_from_buffer");
DBUG_PRINT("info", ("page byte: 0x%x offset: %u",
(uint) page[page_offset], (uint) page_offset));
DBUG_ASSERT(translog_is_LSN_chunk(page[page_offset]));
DBUG_ASSERT(translog_status == TRANSLOG_OK ||
translog_status == TRANSLOG_READONLY);
buff->type= (page[page_offset] & TRANSLOG_REC_TYPE);
buff->short_trid= uint2korr(page + page_offset + 1);
DBUG_PRINT("info", ("Type %u, Short TrID %u, LSN (%lu,0x%lx)",
(uint) buff->type, (uint)buff->short_trid,
LSN_IN_PARTS(buff->lsn)));
/* Read required bytes from the header and call hook */
switch (log_record_type_descriptor[buff->type].rclass) {
case LOGRECTYPE_VARIABLE_LENGTH:
res= translog_variable_length_header(page, page_offset, buff,
scanner);
break;
case LOGRECTYPE_PSEUDOFIXEDLENGTH:
case LOGRECTYPE_FIXEDLENGTH:
res= translog_fixed_length_header(page, page_offset, buff);
break;
default:
DBUG_ASSERT(0); /* we read some junk (got no LSN) */
res= RECHEADER_READ_ERROR;
}
DBUG_RETURN(res);
}
/**
@brief Read record header and some fixed part of a record (the part depend
on record type).
@param lsn log record serial number (address of the record)
@param buff log record header buffer
@note Some type of record can be read completely by this call
@note "Decoded" header stored in TRANSLOG_HEADER_BUFFER::header (relative
LSN can be translated to absolute one), some fields can be added (like
actual header length in the record if the header has variable length)
@return Length of header or operation status
@retval RECHEADER_READ_ERROR error
@retval # number of bytes in
TRANSLOG_HEADER_BUFFER::header where
stored decoded part of the header
*/
int translog_read_record_header(LSN lsn, TRANSLOG_HEADER_BUFFER *buff)
{
TRANSLOG_PAGE_SIZE_BUFF psize_buff;
uchar *page;
translog_size_t res, page_offset= LSN_OFFSET(lsn) % TRANSLOG_PAGE_SIZE;
PAGECACHE_BLOCK_LINK *direct_link;
TRANSLOG_ADDRESS addr;
TRANSLOG_VALIDATOR_DATA data;
DBUG_ENTER("translog_read_record_header");
DBUG_PRINT("enter", ("LSN: (%lu,0x%lx)", LSN_IN_PARTS(lsn)));
DBUG_ASSERT(LSN_OFFSET(lsn) % TRANSLOG_PAGE_SIZE != 0);
DBUG_ASSERT(translog_status == TRANSLOG_OK ||
translog_status == TRANSLOG_READONLY);
buff->lsn= lsn;
buff->groups_no= 0;
data.addr= &addr;
data.was_recovered= 0;
addr= lsn;
addr-= page_offset; /* offset decreasing */
res= (!(page= translog_get_page(&data, psize_buff.buffer, &direct_link))) ?
RECHEADER_READ_ERROR :
translog_read_record_header_from_buffer(page, page_offset, buff, 0);
translog_free_link(direct_link);
DBUG_RETURN(res);
}
/**
@brief Read record header and some fixed part of a record (the part depend
on record type).
@param scan scanner position to read
@param buff log record header buffer
@param move_scanner request to move scanner to the header position
@note Some type of record can be read completely by this call
@note "Decoded" header stored in TRANSLOG_HEADER_BUFFER::header (relative
LSN can be translated to absolute one), some fields can be added (like
actual header length in the record if the header has variable length)
@return Length of header or operation status
@retval RECHEADER_READ_ERROR error
@retval # number of bytes in
TRANSLOG_HEADER_BUFFER::header where stored
decoded part of the header
*/
int translog_read_record_header_scan(TRANSLOG_SCANNER_DATA *scanner,
TRANSLOG_HEADER_BUFFER *buff,
my_bool move_scanner)
{
translog_size_t res;
DBUG_ENTER("translog_read_record_header_scan");
DBUG_PRINT("enter", ("Scanner: Cur: (%lu,0x%lx) Hrz: (%lu,0x%lx) "
"Lst: (%lu,0x%lx) Offset: %u(%x) fixed %d",
LSN_IN_PARTS(scanner->page_addr),
LSN_IN_PARTS(scanner->horizon),
LSN_IN_PARTS(scanner->last_file_page),
(uint) scanner->page_offset,
(uint) scanner->page_offset, scanner->fixed_horizon));
DBUG_ASSERT(translog_status == TRANSLOG_OK ||
translog_status == TRANSLOG_READONLY);
buff->groups_no= 0;
buff->lsn= scanner->page_addr;
buff->lsn+= scanner->page_offset; /* offset increasing */
res= translog_read_record_header_from_buffer(scanner->page,
scanner->page_offset,
buff,
(move_scanner ?
scanner : 0));
DBUG_RETURN(res);
}
/**
@brief Read record header and some fixed part of the next record (the part
depend on record type).
@param scanner data for scanning if lsn is NULL scanner data
will be used for continue scanning.
The scanner can be NULL.
@param buff log record header buffer
@return Length of header or operation status
@retval RECHEADER_READ_ERROR error
@retval RECHEADER_READ_EOF EOF
@retval # number of bytes in
TRANSLOG_HEADER_BUFFER::header where
stored decoded part of the header
*/
int translog_read_next_record_header(TRANSLOG_SCANNER_DATA *scanner,
TRANSLOG_HEADER_BUFFER *buff)
{
translog_size_t res;
DBUG_ENTER("translog_read_next_record_header");
buff->groups_no= 0; /* to be sure that we will free it right */
DBUG_PRINT("enter", ("scanner: 0x%lx", (ulong) scanner));
DBUG_PRINT("info", ("Scanner: Cur: (%lu,0x%lx) Hrz: (%lu,0x%lx) "
"Lst: (%lu,0x%lx) Offset: %u(%x) fixed: %d",
LSN_IN_PARTS(scanner->page_addr),
LSN_IN_PARTS(scanner->horizon),
LSN_IN_PARTS(scanner->last_file_page),
(uint) scanner->page_offset,
(uint) scanner->page_offset, scanner->fixed_horizon));
DBUG_ASSERT(translog_status == TRANSLOG_OK ||
translog_status == TRANSLOG_READONLY);
do
{
if (translog_get_next_chunk(scanner))
DBUG_RETURN(RECHEADER_READ_ERROR);
if (scanner->page == END_OF_LOG)
{
DBUG_PRINT("info", ("End of file from the scanner"));
/* Last record was read */
buff->lsn= LSN_IMPOSSIBLE;
DBUG_RETURN(RECHEADER_READ_EOF);
}
DBUG_PRINT("info", ("Page: (%lu,0x%lx) offset: %lu byte: %x",
LSN_IN_PARTS(scanner->page_addr),
(ulong) scanner->page_offset,
(uint) scanner->page[scanner->page_offset]));
} while (!translog_is_LSN_chunk(scanner->page[scanner->page_offset]) &&
scanner->page[scanner->page_offset] != TRANSLOG_FILLER);
if (scanner->page[scanner->page_offset] == TRANSLOG_FILLER)
{
DBUG_PRINT("info", ("End of file"));
/* Last record was read */
buff->lsn= LSN_IMPOSSIBLE;
/* Return 'end of log' marker */
res= RECHEADER_READ_EOF;
}
else
res= translog_read_record_header_scan(scanner, buff, 0);
DBUG_RETURN(res);
}
/*
Moves record data reader to the next chunk and fill the data reader
information about that chunk.
SYNOPSIS
translog_record_read_next_chunk()
data data cursor
RETURN
0 OK
1 Error
*/
static my_bool translog_record_read_next_chunk(TRANSLOG_READER_DATA *data)
{
translog_size_t new_current_offset= data->current_offset + data->chunk_size;
uint16 chunk_header_len, chunk_len;
uint8 type;
DBUG_ENTER("translog_record_read_next_chunk");
if (data->eor)
{
DBUG_PRINT("info", ("end of the record flag set"));
DBUG_RETURN(1);
}
if (data->header.groups_no &&
data->header.groups_no - 1 != data->current_group &&
data->header.groups[data->current_group].num == data->current_chunk)
{
/* Goto next group */
data->current_group++;
data->current_chunk= 0;
DBUG_PRINT("info", ("skip to group: #%u", data->current_group));
translog_destroy_scanner(&data->scanner);
translog_scanner_init(data->header.groups[data->current_group].addr,
1, &data->scanner, 1);
}
else
{
data->current_chunk++;
if (translog_get_next_chunk(&data->scanner))
DBUG_RETURN(1);
if (data->scanner.page == END_OF_LOG)
{
/*
Actually it should not happened, but we want to quit nicely in case
of a truncated log
*/
DBUG_RETURN(1);
}
}
type= data->scanner.page[data->scanner.page_offset] & TRANSLOG_CHUNK_TYPE;
if (type == TRANSLOG_CHUNK_LSN && data->header.groups_no)
{
DBUG_PRINT("info",
("Last chunk: data len: %u offset: %u group: %u of %u",
data->header.chunk0_data_len, data->scanner.page_offset,
data->current_group, data->header.groups_no - 1));
DBUG_ASSERT(data->header.groups_no - 1 == data->current_group);
DBUG_ASSERT(data->header.lsn ==
data->scanner.page_addr + data->scanner.page_offset);
translog_destroy_scanner(&data->scanner);
translog_scanner_init(data->header.chunk0_data_addr, 1, &data->scanner, 1);
data->chunk_size= data->header.chunk0_data_len;
data->body_offset= data->scanner.page_offset;
data->current_offset= new_current_offset;
data->eor= 1;
DBUG_RETURN(0);
}
if (type == TRANSLOG_CHUNK_LSN || type == TRANSLOG_CHUNK_FIXED)
{
data->eor= 1;
DBUG_RETURN(1); /* End of record */
}
chunk_header_len=
translog_get_chunk_header_length(data->scanner.page +
data->scanner.page_offset);
chunk_len= translog_get_total_chunk_length(data->scanner.page,
data->scanner.page_offset);
data->chunk_size= chunk_len - chunk_header_len;
data->body_offset= data->scanner.page_offset + chunk_header_len;
data->current_offset= new_current_offset;
DBUG_PRINT("info", ("grp: %u chunk: %u body_offset: %u chunk_size: %u "
"current_offset: %lu",
(uint) data->current_group,
(uint) data->current_chunk,
(uint) data->body_offset,
(uint) data->chunk_size, (ulong) data->current_offset));
DBUG_RETURN(0);
}
/*
Initialize record reader data from LSN
SYNOPSIS
translog_init_reader_data()
lsn reference to LSN we should start from
data reader data to initialize
RETURN
0 OK
1 Error
*/
static my_bool translog_init_reader_data(LSN lsn,
TRANSLOG_READER_DATA *data)
{
int read_header;
DBUG_ENTER("translog_init_reader_data");
if (translog_scanner_init(lsn, 1, &data->scanner, 1) ||
((read_header=
translog_read_record_header_scan(&data->scanner, &data->header, 1))
== RECHEADER_READ_ERROR))
DBUG_RETURN(1);
data->read_header= read_header;
data->body_offset= data->header.non_header_data_start_offset;
data->chunk_size= data->header.non_header_data_len;
data->current_offset= data->read_header;
data->current_group= 0;
data->current_chunk= 0;
data->eor= 0;
DBUG_PRINT("info", ("read_header: %u "
"body_offset: %u chunk_size: %u current_offset: %lu",
(uint) data->read_header,
(uint) data->body_offset,
(uint) data->chunk_size, (ulong) data->current_offset));
DBUG_RETURN(0);
}
/**
@brief Destroy reader data object
*/
static void translog_destroy_reader_data(TRANSLOG_READER_DATA *data)
{
translog_destroy_scanner(&data->scanner);
translog_free_record_header(&data->header);
}
/*
Read a part of the record.
SYNOPSIS
translog_read_record_header()
lsn log record serial number (address of the record)
offset From the beginning of the record beginning (read
by translog_read_record_header).
length Length of record part which have to be read.
buffer Buffer where to read the record part (have to be at
least 'length' bytes length)
RETURN
length of data actually read
*/
translog_size_t translog_read_record(LSN lsn,
translog_size_t offset,
translog_size_t length,
uchar *buffer,
TRANSLOG_READER_DATA *data)
{
translog_size_t requested_length= length;
translog_size_t end= offset + length;
TRANSLOG_READER_DATA internal_data;
DBUG_ENTER("translog_read_record");
DBUG_ASSERT(translog_status == TRANSLOG_OK ||
translog_status == TRANSLOG_READONLY);
if (data == NULL)
{
DBUG_ASSERT(lsn != LSN_IMPOSSIBLE);
data= &internal_data;
}
if (lsn ||
(offset < data->current_offset &&
!(offset < data->read_header && offset + length < data->read_header)))
{
if (translog_init_reader_data(lsn, data))
DBUG_RETURN(0);
}
DBUG_PRINT("info", ("Offset: %lu length: %lu "
"Scanner: Cur: (%lu,0x%lx) Hrz: (%lu,0x%lx) "
"Lst: (%lu,0x%lx) Offset: %u(%x) fixed: %d",
(ulong) offset, (ulong) length,
LSN_IN_PARTS(data->scanner.page_addr),
LSN_IN_PARTS(data->scanner.horizon),
LSN_IN_PARTS(data->scanner.last_file_page),
(uint) data->scanner.page_offset,
(uint) data->scanner.page_offset,
data->scanner.fixed_horizon));
if (offset < data->read_header)
{
uint16 len= min(data->read_header, end) - offset;
DBUG_PRINT("info",
("enter header offset: %lu length: %lu",
(ulong) offset, (ulong) length));
memcpy(buffer, data->header.header + offset, len);
length-= len;
if (length == 0)
{
translog_destroy_reader_data(data);
DBUG_RETURN(requested_length);
}
offset+= len;
buffer+= len;
DBUG_PRINT("info",
("len: %u offset: %lu curr: %lu length: %lu",
len, (ulong) offset, (ulong) data->current_offset,
(ulong) length));
}
/* TODO: find first page which we should read by offset */
/* read the record chunk by chunk */
for(;;)
{
uint page_end= data->current_offset + data->chunk_size;
DBUG_PRINT("info",
("enter body offset: %lu curr: %lu "
"length: %lu page_end: %lu",
(ulong) offset, (ulong) data->current_offset, (ulong) length,
(ulong) page_end));
if (offset < page_end)
{
uint len= page_end - offset;
set_if_smaller(len, length); /* in case we read beyond record's end */
DBUG_ASSERT(offset >= data->current_offset);
memcpy(buffer,
data->scanner.page + data->body_offset +
(offset - data->current_offset), len);
length-= len;
if (length == 0)
{
translog_destroy_reader_data(data);
DBUG_RETURN(requested_length);
}
offset+= len;
buffer+= len;
DBUG_PRINT("info",
("len: %u offset: %lu curr: %lu length: %lu",
len, (ulong) offset, (ulong) data->current_offset,
(ulong) length));
}
if (translog_record_read_next_chunk(data))
{
translog_destroy_reader_data(data);
DBUG_RETURN(requested_length - length);
}
}
}
/*
@brief Force skipping to the next buffer
@todo Do not copy old page content if all page protections are switched off
(because we do not need calculate something or change old parts of the page)
*/
static void translog_force_current_buffer_to_finish()
{
TRANSLOG_ADDRESS new_buff_beginning;
uint16 old_buffer_no= log_descriptor.bc.buffer_no;
uint16 new_buffer_no= (old_buffer_no + 1) % TRANSLOG_BUFFERS_NO;
struct st_translog_buffer *new_buffer= (log_descriptor.buffers +
new_buffer_no);
struct st_translog_buffer *old_buffer= log_descriptor.bc.buffer;
uchar *data= log_descriptor.bc.ptr - log_descriptor.bc.current_page_fill;
uint16 left= TRANSLOG_PAGE_SIZE - log_descriptor.bc.current_page_fill;
uint16 current_page_fill, write_counter, previous_offset;
DBUG_ENTER("translog_force_current_buffer_to_finish");
DBUG_PRINT("enter", ("Buffer #%u 0x%lx "
"Buffer addr: (%lu,0x%lx) "
"Page addr: (%lu,0x%lx) "
"size: %lu (%lu) Pg: %u left: %u in progress %u",
(uint) old_buffer_no,
(ulong) old_buffer,
LSN_IN_PARTS(old_buffer->offset),
(ulong) LSN_FILE_NO(log_descriptor.horizon),
(ulong) (LSN_OFFSET(log_descriptor.horizon) -
log_descriptor.bc.current_page_fill),
(ulong) old_buffer->size,
(ulong) (log_descriptor.bc.ptr -log_descriptor.bc.
buffer->buffer),
(uint) log_descriptor.bc.current_page_fill,
(uint) left,
(uint) old_buffer->
copy_to_buffer_in_progress));
translog_lock_assert_owner();
LINT_INIT(current_page_fill);
new_buff_beginning= old_buffer->offset;
new_buff_beginning+= old_buffer->size; /* increase offset */
DBUG_ASSERT(log_descriptor.bc.ptr !=NULL);
DBUG_ASSERT(LSN_FILE_NO(log_descriptor.horizon) ==
LSN_FILE_NO(old_buffer->offset));
translog_check_cursor(&log_descriptor.bc);
DBUG_ASSERT(left < TRANSLOG_PAGE_SIZE);
if (left)
{
/*
TODO: if 'left' is so small that can't hold any other record
then do not move the page
*/
DBUG_PRINT("info", ("left: %u", (uint) left));
old_buffer->pre_force_close_horizon=
old_buffer->offset + old_buffer->size;
/* decrease offset */
new_buff_beginning-= log_descriptor.bc.current_page_fill;
current_page_fill= log_descriptor.bc.current_page_fill;
memset(log_descriptor.bc.ptr, TRANSLOG_FILLER, left);
old_buffer->size+= left;
DBUG_PRINT("info", ("Finish Page buffer #%u: 0x%lx "
"Size: %lu",
(uint) old_buffer->buffer_no,
(ulong) old_buffer,
(ulong) old_buffer->size));
DBUG_ASSERT(old_buffer->buffer_no ==
log_descriptor.bc.buffer_no);
}
else
{
log_descriptor.bc.current_page_fill= 0;
}
translog_buffer_lock(new_buffer);
#ifndef DBUG_OFF
{
TRANSLOG_ADDRESS offset= new_buffer->offset;
TRANSLOG_FILE *file= new_buffer->file;
uint8 ver= new_buffer->ver;
translog_lock_assert_owner();
#endif
translog_wait_for_buffer_free(new_buffer);
#ifndef DBUG_OFF
/* We keep the handler locked so nobody can start this new buffer */
DBUG_ASSERT(offset == new_buffer->offset && new_buffer->file == NULL &&
(file == NULL ? ver : (uint8)(ver + 1)) == new_buffer->ver);
}
#endif
write_counter= log_descriptor.bc.write_counter;
previous_offset= log_descriptor.bc.previous_offset;
translog_start_buffer(new_buffer, &log_descriptor.bc, new_buffer_no);
/* Fix buffer offset (which was incorrectly set to horizon) */
log_descriptor.bc.buffer->offset= new_buff_beginning;
log_descriptor.bc.write_counter= write_counter;
log_descriptor.bc.previous_offset= previous_offset;
new_buffer->prev_last_lsn= BUFFER_MAX_LSN(old_buffer);
DBUG_PRINT("info", ("prev_last_lsn set to (%lu,0x%lx) buffer: 0x%lx",
LSN_IN_PARTS(new_buffer->prev_last_lsn),
(ulong) new_buffer));
/*
Advances this log pointer, increases writers and let other threads to
write to the log while we process old page content
*/
if (left)
{
log_descriptor.bc.ptr+= current_page_fill;
log_descriptor.bc.buffer->size= log_descriptor.bc.current_page_fill=
current_page_fill;
new_buffer->overlay= 1;
}
else
translog_new_page_header(&log_descriptor.horizon, &log_descriptor.bc);
translog_buffer_increase_writers(new_buffer);
translog_buffer_unlock(new_buffer);
/*
We have to wait until all writers finish before start changing the
pages by applying protection and copying the page content in the
new buffer.
*/
#ifndef DBUG_OFF
{
TRANSLOG_ADDRESS offset= old_buffer->offset;
TRANSLOG_FILE *file= old_buffer->file;
uint8 ver= old_buffer->ver;
#endif
/*
Now only one thread can flush log (buffer can flush many threads but
log flush log flush where this function is used can do only one thread)
so no other thread can set is_closing_buffer.
*/
DBUG_ASSERT(!old_buffer->is_closing_buffer);
old_buffer->is_closing_buffer= 1; /* Other flushes will wait */
DBUG_PRINT("enter", ("Buffer #%u 0x%lx is_closing_buffer set",
(uint) old_buffer->buffer_no, (ulong) old_buffer));
translog_wait_for_writers(old_buffer);
#ifndef DBUG_OFF
/* We blocked flushing this buffer so the buffer should not changed */
DBUG_ASSERT(offset == old_buffer->offset && file == old_buffer->file &&
ver == old_buffer->ver);
}
#endif
if (log_descriptor.flags & TRANSLOG_SECTOR_PROTECTION)
{
translog_put_sector_protection(data, &log_descriptor.bc);
if (left)
{
log_descriptor.bc.write_counter++;
log_descriptor.bc.previous_offset= current_page_fill;
}
else
{
DBUG_PRINT("info", ("drop write_counter"));
log_descriptor.bc.write_counter= 0;
log_descriptor.bc.previous_offset= 0;
}
}
if (log_descriptor.flags & TRANSLOG_PAGE_CRC)
{
uint32 crc= translog_crc(data + log_descriptor.page_overhead,
TRANSLOG_PAGE_SIZE -
log_descriptor.page_overhead);
DBUG_PRINT("info", ("CRC: 0x%lx", (ulong) crc));
int4store(data + 3 + 3 + 1, crc);
}
old_buffer->is_closing_buffer= 0;
DBUG_PRINT("enter", ("Buffer #%u 0x%lx is_closing_buffer cleared",
(uint) old_buffer->buffer_no, (ulong) old_buffer));
mysql_cond_broadcast(&old_buffer->waiting_filling_buffer);
if (left)
{
if (log_descriptor.flags &
(TRANSLOG_PAGE_CRC | TRANSLOG_SECTOR_PROTECTION))
memcpy(new_buffer->buffer, data, current_page_fill);
else
{
/*
This page header does not change if we add more data to the page so
we can not copy it and will not overwrite later
*/
new_buffer->skipped_data= current_page_fill;
TRASH_ALLOC(new_buffer->buffer, current_page_fill);
DBUG_ASSERT(new_buffer->skipped_data < TRANSLOG_PAGE_SIZE);
}
}
old_buffer->next_buffer_offset= new_buffer->offset;
translog_buffer_lock(new_buffer);
new_buffer->prev_buffer_offset= old_buffer->offset;
translog_buffer_decrease_writers(new_buffer);
translog_buffer_unlock(new_buffer);
DBUG_VOID_RETURN;
}
/**
@brief Waits while given lsn will be flushed
@param lsn log record serial number up to which (inclusive)
the log has to be flushed
*/
void translog_flush_wait_for_end(LSN lsn)
{
DBUG_ENTER("translog_flush_wait_for_end");
DBUG_PRINT("enter", ("LSN: (%lu,0x%lx)", LSN_IN_PARTS(lsn)));
mysql_mutex_assert_owner(&log_descriptor.log_flush_lock);
while (cmp_translog_addr(log_descriptor.flushed, lsn) < 0)
mysql_cond_wait(&log_descriptor.log_flush_cond,
&log_descriptor.log_flush_lock);
DBUG_VOID_RETURN;
}
/**
@brief Sets goal for the next flush pass and waits for this pass end.
@param lsn log record serial number up to which (inclusive)
the log has to be flushed
*/
void translog_flush_set_new_goal_and_wait(TRANSLOG_ADDRESS lsn)
{
int flush_no= log_descriptor.flush_no;
DBUG_ENTER("translog_flush_set_new_goal_and_wait");
DBUG_PRINT("enter", ("LSN: (%lu,0x%lx)", LSN_IN_PARTS(lsn)));
mysql_mutex_assert_owner(&log_descriptor.log_flush_lock);
if (cmp_translog_addr(lsn, log_descriptor.next_pass_max_lsn) > 0)
{
log_descriptor.next_pass_max_lsn= lsn;
log_descriptor.max_lsn_requester= pthread_self();
mysql_cond_broadcast(&log_descriptor.new_goal_cond);
}
while (flush_no == log_descriptor.flush_no)
{
mysql_cond_wait(&log_descriptor.log_flush_cond,
&log_descriptor.log_flush_lock);
}
DBUG_VOID_RETURN;
}
/**
@brief sync() range of files (inclusive) and directory (by request)
@param min min internal file number to flush
@param max max internal file number to flush
@param sync_dir need sync directory
return Operation status
@retval 0 OK
@retval 1 Error
*/
static my_bool translog_sync_files(uint32 min, uint32 max,
my_bool sync_dir)
{
uint fn;
my_bool rc= 0;
ulonglong flush_interval;
DBUG_ENTER("translog_sync_files");
DBUG_PRINT("info", ("min: %lu max: %lu sync dir: %d",
(ulong) min, (ulong) max, (int) sync_dir));
DBUG_ASSERT(min <= max);
flush_interval= group_commit_wait;
if (flush_interval)
flush_start= microsecond_interval_timer();
for (fn= min; fn <= max; fn++)
{
TRANSLOG_FILE *file= get_logfile_by_number(fn);
DBUG_ASSERT(file != NULL);
if (!file->is_sync)
{
if (mysql_file_sync(file->handler.file, MYF(MY_WME)))
{
rc= 1;
translog_stop_writing();
DBUG_RETURN(rc);
}
translog_syncs++;
file->is_sync= 1;
}
}
if (sync_dir)
{
if (!(rc= sync_dir(log_descriptor.directory_fd,
MYF(MY_WME | MY_IGNORE_BADFD))))
translog_syncs++;
}
DBUG_RETURN(rc);
}
/*
@brief Flushes buffers with LSNs in them less or equal address <lsn>
@param lsn address up to which all LSNs should be flushed,
can be reset to real last LSN address
@parem sent_to_disk returns 'sent to disk' position
@param flush_horizon returns horizon of the flush
@note About terminology see comment to translog_flush().
*/
void translog_flush_buffers(TRANSLOG_ADDRESS *lsn,
TRANSLOG_ADDRESS *sent_to_disk,
TRANSLOG_ADDRESS *flush_horizon)
{
dirty_buffer_mask_t dirty_buffer_mask;
uint i;
uint8 last_buffer_no, start_buffer_no;
DBUG_ENTER("translog_flush_buffers");
LINT_INIT(last_buffer_no);
/*
We will recheck information when will lock buffers one by
one so we can use unprotected read here (this is just for
speed up buffers processing)
*/
dirty_buffer_mask= log_descriptor.dirty_buffer_mask;
DBUG_PRINT("info", ("Dirty buffer mask: %lx current buffer: %u",
(ulong) dirty_buffer_mask,
(uint) log_descriptor.bc.buffer_no));
for (i= (log_descriptor.bc.buffer_no + 1) % TRANSLOG_BUFFERS_NO;
i != log_descriptor.bc.buffer_no && !(dirty_buffer_mask & (1 << i));
i= (i + 1) % TRANSLOG_BUFFERS_NO) {}
start_buffer_no= i;
DBUG_PRINT("info",
("start from: %u current: %u prev last lsn: (%lu,0x%lx)",
(uint) start_buffer_no, (uint) log_descriptor.bc.buffer_no,
LSN_IN_PARTS(log_descriptor.bc.buffer->prev_last_lsn)));
/*
if LSN up to which we have to flush bigger then maximum LSN of previous
buffer and at least one LSN was saved in the current buffer (last_lsn !=
LSN_IMPOSSIBLE) then we have to close the current buffer.
*/
if (cmp_translog_addr(*lsn, log_descriptor.bc.buffer->prev_last_lsn) > 0 &&
log_descriptor.bc.buffer->last_lsn != LSN_IMPOSSIBLE)
{
struct st_translog_buffer *buffer= log_descriptor.bc.buffer;
*lsn= log_descriptor.bc.buffer->last_lsn; /* fix lsn if it was horizon */
DBUG_PRINT("info", ("LSN to flush fixed to last lsn: (%lu,0x%lx)",
LSN_IN_PARTS(*lsn)));
last_buffer_no= log_descriptor.bc.buffer_no;
log_descriptor.is_everything_flushed= 1;
translog_force_current_buffer_to_finish();
translog_buffer_unlock(buffer);
}
else if (log_descriptor.bc.buffer->prev_last_lsn != LSN_IMPOSSIBLE)
{
/* fix lsn if it was horizon */
*lsn= log_descriptor.bc.buffer->prev_last_lsn;
DBUG_PRINT("info", ("LSN to flush fixed to prev last lsn: (%lu,0x%lx)",
LSN_IN_PARTS(*lsn)));
last_buffer_no= ((log_descriptor.bc.buffer_no + TRANSLOG_BUFFERS_NO -1) %
TRANSLOG_BUFFERS_NO);
translog_unlock();
}
else if (log_descriptor.bc.buffer->last_lsn == LSN_IMPOSSIBLE)
{
DBUG_PRINT("info", ("There is no LSNs yet generated => do nothing"));
translog_unlock();
DBUG_VOID_RETURN;
}
/* flush buffers */
*sent_to_disk= translog_get_sent_to_disk();
if (cmp_translog_addr(*lsn, *sent_to_disk) > 0)
{
DBUG_PRINT("info", ("Start buffer #: %u last buffer #: %u",
(uint) start_buffer_no, (uint) last_buffer_no));
last_buffer_no= (last_buffer_no + 1) % TRANSLOG_BUFFERS_NO;
i= start_buffer_no;
do
{
struct st_translog_buffer *buffer= log_descriptor.buffers + i;
translog_buffer_lock(buffer);
DBUG_PRINT("info", ("Check buffer: 0x%lx #: %u "
"prev last LSN: (%lu,0x%lx) "
"last LSN: (%lu,0x%lx) status: %s",
(ulong)(buffer),
(uint) i,
LSN_IN_PARTS(buffer->prev_last_lsn),
LSN_IN_PARTS(buffer->last_lsn),
(buffer->file ?
"dirty" : "closed")));
if (buffer->prev_last_lsn <= *lsn &&
buffer->file != NULL)
{
DBUG_ASSERT(*flush_horizon <= buffer->offset + buffer->size);
*flush_horizon= (buffer->pre_force_close_horizon != LSN_IMPOSSIBLE ?
buffer->pre_force_close_horizon :
buffer->offset + buffer->size);
/* pre_force_close_horizon is reset during new buffer start */
DBUG_PRINT("info", ("flush_horizon: (%lu,0x%lx)",
LSN_IN_PARTS(*flush_horizon)));
DBUG_ASSERT(*flush_horizon <= log_descriptor.horizon);
translog_buffer_flush(buffer);
}
translog_buffer_unlock(buffer);
i= (i + 1) % TRANSLOG_BUFFERS_NO;
} while (i != last_buffer_no);
*sent_to_disk= translog_get_sent_to_disk();
}
DBUG_VOID_RETURN;
}
/**
@brief Flush the log up to given LSN (included)
@param lsn log record serial number up to which (inclusive)
the log has to be flushed
@return Operation status
@retval 0 OK
@retval 1 Error
@note
- Non group commit logic: Commits made in passes. Thread which started
flush first is performing actual flush, other threads sets new goal (LSN)
of the next pass (if it is maximum) and waits for the pass end or just
wait for the pass end.
- If hard group commit enabled and rate set to zero:
The first thread sends all changed buffers to disk. This is repeated
as long as there are new LSNs added. The process can not loop
forever because we have limited number of threads and they will wait
for the data to be synced.
Pseudo code:
do
send changed buffers to disk
while new_goal
sync
- If hard group commit switched ON and less than rate microseconds has
passed from last sync, then after buffers have been sent to disk
wait until rate microseconds has passed since last sync, do sync and return.
This ensures that if we call sync infrequently we don't do any waits.
- If soft group commit enabled everything works as with 'non group commit'
but the thread doesn't do any real sync(). If rate is not zero the
sync() will be performed by a service thread with the given rate
when needed (new LSN appears).
@note Terminology:
'sent to disk' means written to disk but not sync()ed,
'flushed' mean sent to disk and synced().
*/
my_bool translog_flush(TRANSLOG_ADDRESS lsn)
{
struct timespec abstime;
ulonglong flush_interval;
ulonglong time_spent;
LSN sent_to_disk= LSN_IMPOSSIBLE;
TRANSLOG_ADDRESS flush_horizon;
my_bool rc= 0;
my_bool hgroup_commit_at_start;
DBUG_ENTER("translog_flush");
DBUG_PRINT("enter", ("Flush up to LSN: (%lu,0x%lx)", LSN_IN_PARTS(lsn)));
DBUG_ASSERT(translog_status == TRANSLOG_OK ||
translog_status == TRANSLOG_READONLY);
LINT_INIT(sent_to_disk);
LINT_INIT(flush_interval);
mysql_mutex_lock(&log_descriptor.log_flush_lock);
DBUG_PRINT("info", ("Everything is flushed up to (%lu,0x%lx)",
LSN_IN_PARTS(log_descriptor.flushed)));
if (cmp_translog_addr(log_descriptor.flushed, lsn) >= 0)
{
mysql_mutex_unlock(&log_descriptor.log_flush_lock);
DBUG_RETURN(0);
}
if (log_descriptor.flush_in_progress)
{
translog_lock();
/* fix lsn if it was horizon */
if (cmp_translog_addr(lsn, log_descriptor.bc.buffer->last_lsn) > 0)
lsn= BUFFER_MAX_LSN(log_descriptor.bc.buffer);
translog_unlock();
translog_flush_set_new_goal_and_wait(lsn);
if (!pthread_equal(log_descriptor.max_lsn_requester, pthread_self()))
{
/*
translog_flush_wait_for_end() release log_flush_lock while is
waiting then acquire it again
*/
translog_flush_wait_for_end(lsn);
mysql_mutex_unlock(&log_descriptor.log_flush_lock);
DBUG_RETURN(0);
}
log_descriptor.next_pass_max_lsn= LSN_IMPOSSIBLE;
}
log_descriptor.flush_in_progress= 1;
flush_horizon= log_descriptor.previous_flush_horizon;
DBUG_PRINT("info", ("flush_in_progress is set, flush_horizon: (%lu,0x%lx)",
LSN_IN_PARTS(flush_horizon)));
mysql_mutex_unlock(&log_descriptor.log_flush_lock);
hgroup_commit_at_start= hard_group_commit;
if (hgroup_commit_at_start)
flush_interval= group_commit_wait;
translog_lock();
if (log_descriptor.is_everything_flushed)
{
DBUG_PRINT("info", ("everything is flushed"));
translog_unlock();
mysql_mutex_lock(&log_descriptor.log_flush_lock);
goto out;
}
for (;;)
{
/* Following function flushes buffers and makes translog_unlock() */
translog_flush_buffers(&lsn, &sent_to_disk, &flush_horizon);
if (!hgroup_commit_at_start)
break; /* flush pass is ended */
retest:
/*
We do not check time here because mysql_mutex_lock rarely takes
a lot of time so we can sacrifice a bit precision to performance
(taking into account that microsecond_interval_timer() might be
expensive call).
*/
if (flush_interval == 0)
break; /* flush pass is ended */
mysql_mutex_lock(&log_descriptor.log_flush_lock);
if (log_descriptor.next_pass_max_lsn == LSN_IMPOSSIBLE)
{
if (flush_interval == 0 ||
(time_spent= (microsecond_interval_timer() - flush_start)) >=
flush_interval)
{
mysql_mutex_unlock(&log_descriptor.log_flush_lock);
break;
}
DBUG_PRINT("info", ("flush waits: %llu interval: %llu spent: %llu",
flush_interval - time_spent,
flush_interval, time_spent));
/* wait time or next goal */
set_timespec_nsec(abstime, flush_interval - time_spent);
mysql_cond_timedwait(&log_descriptor.new_goal_cond,
&log_descriptor.log_flush_lock,
&abstime);
mysql_mutex_unlock(&log_descriptor.log_flush_lock);
DBUG_PRINT("info", ("retest conditions"));
goto retest;
}
/* take next goal */
lsn= log_descriptor.next_pass_max_lsn;
log_descriptor.next_pass_max_lsn= LSN_IMPOSSIBLE;
/* prevent other thread from continue */
log_descriptor.max_lsn_requester= pthread_self();
DBUG_PRINT("info", ("flush took next goal: (%lu,0x%lx)",
LSN_IN_PARTS(lsn)));
mysql_mutex_unlock(&log_descriptor.log_flush_lock);
/* next flush pass */
DBUG_PRINT("info", ("next flush pass"));
translog_lock();
}
/*
sync() files from previous flush till current one
*/
if (!soft_sync || hgroup_commit_at_start)
{
if ((rc=
translog_sync_files(LSN_FILE_NO(log_descriptor.flushed),
LSN_FILE_NO(lsn),
sync_log_dir >= TRANSLOG_SYNC_DIR_ALWAYS &&
(LSN_FILE_NO(log_descriptor.
previous_flush_horizon) !=
LSN_FILE_NO(flush_horizon) ||
(LSN_OFFSET(log_descriptor.
previous_flush_horizon) /
TRANSLOG_PAGE_SIZE) !=
(LSN_OFFSET(flush_horizon) /
TRANSLOG_PAGE_SIZE)))))
{
sent_to_disk= LSN_IMPOSSIBLE;
mysql_mutex_lock(&log_descriptor.log_flush_lock);
goto out;
}
/* keep values for soft sync() and forced sync() actual */
{
uint32 fileno= LSN_FILE_NO(lsn);
soft_sync_min= fileno;
soft_sync_max= fileno;
}
}
else
{
soft_sync_max= LSN_FILE_NO(lsn);
soft_need_sync= 1;
}
DBUG_ASSERT(flush_horizon <= log_descriptor.horizon);
mysql_mutex_lock(&log_descriptor.log_flush_lock);
log_descriptor.previous_flush_horizon= flush_horizon;
out:
if (sent_to_disk != LSN_IMPOSSIBLE)
log_descriptor.flushed= sent_to_disk;
log_descriptor.flush_in_progress= 0;
log_descriptor.flush_no++;
DBUG_PRINT("info", ("flush_in_progress is dropped"));
mysql_mutex_unlock(&log_descriptor.log_flush_lock);
mysql_cond_broadcast(&log_descriptor.log_flush_cond);
DBUG_RETURN(rc);
}
/**
@brief Gives a 2-byte-id to MARIA_SHARE and logs this fact
If a MARIA_SHARE does not yet have a 2-byte-id (unique over all currently
open MARIA_SHAREs), give it one and record this assignment in the log
(LOGREC_FILE_ID log record).
@param tbl_info table
@param trn calling transaction
@return Operation status
@retval 0 OK
@retval 1 Error
@note Can be called even if share already has an id (then will do nothing)
*/
int translog_assign_id_to_share(MARIA_HA *tbl_info, TRN *trn)
{
uint16 id;
MARIA_SHARE *share= tbl_info->s;
/*
If you give an id to a non-BLOCK_RECORD table, you also need to release
this id somewhere. Then you can change the assertion.
*/
DBUG_ASSERT(share->data_file_type == BLOCK_RECORD);
/* re-check under mutex to avoid having 2 ids for the same share */
mysql_mutex_lock(&share->intern_lock);
if (unlikely(share->id == 0))
{
LSN lsn;
LEX_CUSTRING log_array[TRANSLOG_INTERNAL_PARTS + 2];
uchar log_data[FILEID_STORE_SIZE];
/* Inspired by set_short_trid() of trnman.c */
uint i= share->kfile.file % SHARE_ID_MAX + 1;
id= 0;
do
{
my_atomic_rwlock_wrlock(&LOCK_id_to_share);
for ( ; i <= SHARE_ID_MAX ; i++) /* the range is [1..SHARE_ID_MAX] */
{
void *tmp= NULL;
if (id_to_share[i] == NULL &&
my_atomic_casptr((void **)&id_to_share[i], &tmp, share))
{
id= (uint16) i;
break;
}
}
my_atomic_rwlock_wrunlock(&LOCK_id_to_share);
i= 1; /* scan the whole array */
} while (id == 0);
DBUG_PRINT("info", ("id_to_share: 0x%lx -> %u", (ulong)share, id));
fileid_store(log_data, id);
log_array[TRANSLOG_INTERNAL_PARTS + 0].str= log_data;
log_array[TRANSLOG_INTERNAL_PARTS + 0].length= sizeof(log_data);
/*
open_file_name is an unresolved name (symlinks are not resolved, datadir
is not realpath-ed, etc) which is good: the log can be moved to another
directory and continue working.
*/
log_array[TRANSLOG_INTERNAL_PARTS + 1].str=
(uchar *)share->open_file_name.str;
log_array[TRANSLOG_INTERNAL_PARTS + 1].length=
share->open_file_name.length + 1;
/*
We can't unlock share->intern_lock before the log entry is written to
ensure no one uses the id before it's logged.
*/
if (unlikely(translog_write_record(&lsn, LOGREC_FILE_ID, trn, tbl_info,
(translog_size_t)
(sizeof(log_data) +
log_array[TRANSLOG_INTERNAL_PARTS +
1].length),
sizeof(log_array)/sizeof(log_array[0]),
log_array, NULL, NULL)))
{
mysql_mutex_unlock(&share->intern_lock);
return 1;
}
/*
Now when translog record is done, we can set share->id.
If we set it before, then translog_write_record may pick up the id
before it's written to the log.
*/
share->id= id;
share->state.logrec_file_id= lsn;
}
mysql_mutex_unlock(&share->intern_lock);
return 0;
}
/**
@brief Recycles a MARIA_SHARE's short id.
@param share table
@note Must be called only if share has an id (i.e. id != 0)
*/
void translog_deassign_id_from_share(MARIA_SHARE *share)
{
DBUG_PRINT("info", ("id_to_share: 0x%lx id %u -> 0",
(ulong)share, share->id));
/*
We don't need any mutex as we are called only when closing the last
instance of the table or at the end of REPAIR: no writes can be
happening. But a Checkpoint may be reading share->id, so we require this
mutex:
*/
mysql_mutex_assert_owner(&share->intern_lock);
my_atomic_rwlock_rdlock(&LOCK_id_to_share);
my_atomic_storeptr((void **)&id_to_share[share->id], 0);
my_atomic_rwlock_rdunlock(&LOCK_id_to_share);
share->id= 0;
/* useless but safety: */
share->lsn_of_file_id= LSN_IMPOSSIBLE;
}
void translog_assign_id_to_share_from_recovery(MARIA_SHARE *share,
uint16 id)
{
DBUG_ASSERT(maria_in_recovery && !maria_multi_threaded);
DBUG_ASSERT(share->data_file_type == BLOCK_RECORD);
DBUG_ASSERT(share->id == 0);
DBUG_ASSERT(id_to_share[id] == NULL);
id_to_share[share->id= id]= share;
}
/**
@brief check if such log file exists
@param file_no number of the file to test
@retval 0 no such file
@retval 1 there is file with such number
*/
my_bool translog_is_file(uint file_no)
{
MY_STAT stat_buff;
char path[FN_REFLEN];
return (test(mysql_file_stat(key_file_translog,
translog_filename_by_fileno(file_no, path),
&stat_buff, MYF(0))));
}
/**
@brief returns minimum log file number
@param horizon the end of the log
@param is_protected true if it is under purge_log protection
@retval minimum file number
@retval 0 no files found
*/
static uint32 translog_first_file(TRANSLOG_ADDRESS horizon, int is_protected)
{
uint min_file= 0, max_file;
DBUG_ENTER("translog_first_file");
if (!is_protected)
mysql_mutex_lock(&log_descriptor.purger_lock);
if (log_descriptor.min_file_number &&
translog_is_file(log_descriptor.min_file_number))
{
DBUG_PRINT("info", ("cached %lu",
(ulong) log_descriptor.min_file_number));
if (!is_protected)
mysql_mutex_unlock(&log_descriptor.purger_lock);
DBUG_RETURN(log_descriptor.min_file_number);
}
max_file= LSN_FILE_NO(horizon);
/* binary search for last file */
while (min_file != max_file && min_file != (max_file - 1))
{
uint test= (min_file + max_file) / 2;
DBUG_PRINT("info", ("min_file: %u test: %u max_file: %u",
min_file, test, max_file));
if (test == max_file)
test--;
if (translog_is_file(test))
max_file= test;
else
min_file= test;
}
log_descriptor.min_file_number= max_file;
if (!is_protected)
mysql_mutex_unlock(&log_descriptor.purger_lock);
DBUG_PRINT("info", ("first file :%lu", (ulong) max_file));
DBUG_ASSERT(max_file >= 1);
DBUG_RETURN(max_file);
}
/**
@brief returns the most close LSN higher the given chunk address
@param addr the chunk address to start from
@param horizon the horizon if it is known or LSN_IMPOSSIBLE
@retval LSN_ERROR Error
@retval LSN_IMPOSSIBLE no LSNs after the address
@retval # LSN of the most close LSN higher the given chunk address
*/
LSN translog_next_LSN(TRANSLOG_ADDRESS addr, TRANSLOG_ADDRESS horizon)
{
TRANSLOG_SCANNER_DATA scanner;
LSN result;
DBUG_ENTER("translog_next_LSN");
if (horizon == LSN_IMPOSSIBLE)
horizon= translog_get_horizon();
if (addr == horizon)
DBUG_RETURN(LSN_IMPOSSIBLE);
translog_scanner_init(addr, 0, &scanner, 1);
/*
addr can point not to a chunk beginning but page end so next
page beginning.
*/
if (addr % TRANSLOG_PAGE_SIZE == 0)
{
/*
We are emulating the page end which cased such horizon value to
trigger translog_scanner_eop().
We can't just increase addr on page header overhead because it
can be file end so we allow translog_get_next_chunk() to skip
to the next page in correct way
*/
scanner.page_addr-= TRANSLOG_PAGE_SIZE;
scanner.page_offset= TRANSLOG_PAGE_SIZE;
#ifndef DBUG_OFF
scanner.page= NULL; /* prevent using incorrect page content */
#endif
}
/* addr can point not to a chunk beginning but to a page end */
if (translog_scanner_eop(&scanner))
{
if (translog_get_next_chunk(&scanner))
{
result= LSN_ERROR;
goto out;
}
if (scanner.page == END_OF_LOG)
{
result= LSN_IMPOSSIBLE;
goto out;
}
}
while (!translog_is_LSN_chunk(scanner.page[scanner.page_offset]) &&
scanner.page[scanner.page_offset] != TRANSLOG_FILLER)
{
if (translog_get_next_chunk(&scanner))
{
result= LSN_ERROR;
goto out;
}
if (scanner.page == END_OF_LOG)
{
result= LSN_IMPOSSIBLE;
goto out;
}
}
if (scanner.page[scanner.page_offset] == TRANSLOG_FILLER)
result= LSN_IMPOSSIBLE; /* reached page filler */
else
result= scanner.page_addr + scanner.page_offset;
out:
translog_destroy_scanner(&scanner);
DBUG_RETURN(result);
}
/**
@brief returns the LSN of the first record starting in this log
@retval LSN_ERROR Error
@retval LSN_IMPOSSIBLE no log or the log is empty
@retval # LSN of the first record
*/
LSN translog_first_lsn_in_log()
{
TRANSLOG_ADDRESS addr, horizon= translog_get_horizon();
TRANSLOG_VALIDATOR_DATA data;
uint file;
uint16 chunk_offset;
uchar *page;
DBUG_ENTER("translog_first_lsn_in_log");
DBUG_PRINT("info", ("Horizon: (%lu,0x%lx)", LSN_IN_PARTS(horizon)));
DBUG_ASSERT(translog_status == TRANSLOG_OK ||
translog_status == TRANSLOG_READONLY);
if (!(file= translog_first_file(horizon, 0)))
{
/* log has no records yet */
DBUG_RETURN(LSN_IMPOSSIBLE);
}
addr= MAKE_LSN(file, TRANSLOG_PAGE_SIZE); /* the first page of the file */
data.addr= &addr;
{
TRANSLOG_PAGE_SIZE_BUFF psize_buff;
if ((page= translog_get_page(&data, psize_buff.buffer, NULL)) == NULL ||
(chunk_offset= translog_get_first_chunk_offset(page)) == 0)
DBUG_RETURN(LSN_ERROR);
}
addr+= chunk_offset;
DBUG_RETURN(translog_next_LSN(addr, horizon));
}
/**
@brief Returns theoretical first LSN if first log is present
@retval LSN_ERROR Error
@retval LSN_IMPOSSIBLE no log
@retval # LSN of the first record
*/
LSN translog_first_theoretical_lsn()
{
TRANSLOG_ADDRESS addr= translog_get_horizon();
TRANSLOG_PAGE_SIZE_BUFF psize_buff;
uchar *page;
TRANSLOG_VALIDATOR_DATA data;
DBUG_ENTER("translog_first_theoretical_lsn");
DBUG_PRINT("info", ("Horizon: (%lu,0x%lx)", LSN_IN_PARTS(addr)));
DBUG_ASSERT(translog_status == TRANSLOG_OK ||
translog_status == TRANSLOG_READONLY);
if (!translog_is_file(1))
DBUG_RETURN(LSN_IMPOSSIBLE);
if (addr == MAKE_LSN(1, TRANSLOG_PAGE_SIZE))
{
/* log has no records yet */
DBUG_RETURN(MAKE_LSN(1, TRANSLOG_PAGE_SIZE +
log_descriptor.page_overhead));
}
addr= MAKE_LSN(1, TRANSLOG_PAGE_SIZE); /* the first page of the file */
data.addr= &addr;
if ((page= translog_get_page(&data, psize_buff.buffer, NULL)) == NULL)
DBUG_RETURN(LSN_ERROR);
DBUG_RETURN(MAKE_LSN(1, TRANSLOG_PAGE_SIZE +
page_overhead[page[TRANSLOG_PAGE_FLAGS]]));
}
/**
@brief Checks given low water mark and purge files if it is need
@param low the last (minimum) address which is need
@retval 0 OK
@retval 1 Error
*/
my_bool translog_purge(TRANSLOG_ADDRESS low)
{
uint32 last_need_file= LSN_FILE_NO(low);
uint32 min_unsync;
int soft;
TRANSLOG_ADDRESS horizon= translog_get_horizon();
int rc= 0;
DBUG_ENTER("translog_purge");
DBUG_PRINT("enter", ("low: (%lu,0x%lx)", LSN_IN_PARTS(low)));
DBUG_ASSERT(translog_status == TRANSLOG_OK ||
translog_status == TRANSLOG_READONLY);
soft= soft_sync;
min_unsync= soft_sync_min;
DBUG_PRINT("info", ("min_unsync: %lu", (ulong) min_unsync));
if (soft && min_unsync < last_need_file)
{
last_need_file= min_unsync;
DBUG_PRINT("info", ("last_need_file set to %lu", (ulong)last_need_file));
}
mysql_mutex_lock(&log_descriptor.purger_lock);
DBUG_PRINT("info", ("last_lsn_checked file: %lu:",
(ulong) log_descriptor.last_lsn_checked));
if (LSN_FILE_NO(log_descriptor.last_lsn_checked) < last_need_file)
{
uint32 i;
uint32 min_file= translog_first_file(horizon, 1);
DBUG_ASSERT(min_file != 0); /* log is already started */
DBUG_PRINT("info", ("min_file: %lu:",(ulong) min_file));
for(i= min_file; i < last_need_file && rc == 0; i++)
{
LSN lsn= translog_get_file_max_lsn_stored(i);
if (lsn == LSN_IMPOSSIBLE)
break; /* files are still in writing */
if (lsn == LSN_ERROR)
{
rc= 1;
break;
}
if (cmp_translog_addr(lsn, low) >= 0)
break;
DBUG_PRINT("info", ("purge file %lu", (ulong) i));
/* remove file descriptor from the cache */
/*
log_descriptor.min_file can be changed only here during execution
and the function is serialized, so we can access it without problems
*/
if (i >= log_descriptor.min_file)
{
TRANSLOG_FILE *file;
mysql_rwlock_wrlock(&log_descriptor.open_files_lock);
DBUG_ASSERT(log_descriptor.max_file - log_descriptor.min_file + 1 ==
log_descriptor.open_files.elements);
DBUG_ASSERT(log_descriptor.min_file == i);
file= *((TRANSLOG_FILE **)pop_dynamic(&log_descriptor.open_files));
DBUG_PRINT("info", ("Files : %d", log_descriptor.open_files.elements));
DBUG_ASSERT(i == file->number);
log_descriptor.min_file++;
DBUG_ASSERT(log_descriptor.max_file - log_descriptor.min_file + 1 ==
log_descriptor.open_files.elements);
mysql_rwlock_unlock(&log_descriptor.open_files_lock);
translog_close_log_file(file);
}
if (log_purge_type == TRANSLOG_PURGE_IMMIDIATE)
{
char path[FN_REFLEN], *file_name;
file_name= translog_filename_by_fileno(i, path);
rc= test(mysql_file_delete(key_file_translog,
file_name, MYF(MY_WME)));
}
}
if (unlikely(rc == 1))
log_descriptor.min_need_file= 0; /* impossible value */
else
log_descriptor.min_need_file= i;
}
mysql_mutex_unlock(&log_descriptor.purger_lock);
DBUG_RETURN(rc);
}
/**
@brief Purges files by stored min need file in case of
"ondemend" purge type
@note This function do real work only if it is "ondemend" purge type
and translog_purge() was called at least once and last time without
errors
@retval 0 OK
@retval 1 Error
*/
my_bool translog_purge_at_flush()
{
uint32 i, min_file;
int rc= 0;
DBUG_ENTER("translog_purge_at_flush");
DBUG_ASSERT(translog_status == TRANSLOG_OK ||
translog_status == TRANSLOG_READONLY);
if (unlikely(translog_status == TRANSLOG_READONLY))
{
DBUG_PRINT("info", ("The log is read only => exit"));
DBUG_RETURN(0);
}
if (log_purge_type != TRANSLOG_PURGE_ONDEMAND)
{
DBUG_PRINT("info", ("It is not \"at_flush\" => exit"));
DBUG_RETURN(0);
}
mysql_mutex_lock(&log_descriptor.purger_lock);
if (unlikely(log_descriptor.min_need_file == 0))
{
DBUG_PRINT("info", ("No info about min need file => exit"));
mysql_mutex_unlock(&log_descriptor.purger_lock);
DBUG_RETURN(0);
}
min_file= translog_first_file(translog_get_horizon(), 1);
DBUG_ASSERT(min_file != 0); /* log is already started */
for(i= min_file; i < log_descriptor.min_need_file && rc == 0; i++)
{
char path[FN_REFLEN], *file_name;
DBUG_PRINT("info", ("purge file %lu\n", (ulong) i));
file_name= translog_filename_by_fileno(i, path);
rc= test(mysql_file_delete(key_file_translog,
file_name, MYF(MY_WME)));
}
mysql_mutex_unlock(&log_descriptor.purger_lock);
DBUG_RETURN(rc);
}
/**
@brief Gets min file number
@param horizon the end of the log
@retval minimum file number
@retval 0 no files found
*/
uint32 translog_get_first_file(TRANSLOG_ADDRESS horizon)
{
return translog_first_file(horizon, 0);
}
/**
@brief Gets min file number which is needed
@retval minimum file number
@retval 0 unknown
*/
uint32 translog_get_first_needed_file()
{
uint32 file_no;
mysql_mutex_lock(&log_descriptor.purger_lock);
file_no= log_descriptor.min_need_file;
mysql_mutex_unlock(&log_descriptor.purger_lock);
return file_no;
}
/**
@brief Gets transaction log file size
@return transaction log file size
*/
uint32 translog_get_file_size()
{
uint32 res;
translog_lock();
res= log_descriptor.log_file_max_size;
translog_unlock();
return (res);
}
/**
@brief Sets transaction log file size
@return Returns actually set transaction log size
*/
void translog_set_file_size(uint32 size)
{
struct st_translog_buffer *old_buffer= NULL;
DBUG_ENTER("translog_set_file_size");
translog_lock();
DBUG_PRINT("enter", ("Size: %lu", (ulong) size));
DBUG_ASSERT(size % TRANSLOG_PAGE_SIZE == 0 &&
size >= TRANSLOG_MIN_FILE_SIZE);
log_descriptor.log_file_max_size= size;
/* if current file longer then finish it*/
if (LSN_OFFSET(log_descriptor.horizon) >= log_descriptor.log_file_max_size)
{
old_buffer= log_descriptor.bc.buffer;
translog_buffer_next(&log_descriptor.horizon, &log_descriptor.bc, 1);
translog_buffer_unlock(old_buffer);
}
translog_unlock();
if (old_buffer)
{
translog_buffer_lock(old_buffer);
translog_buffer_flush(old_buffer);
translog_buffer_unlock(old_buffer);
}
DBUG_VOID_RETURN;
}
/**
Write debug information to log if we EXTRA_DEBUG is enabled
*/
my_bool translog_log_debug_info(TRN *trn __attribute__((unused)),
enum translog_debug_info_type type
__attribute__((unused)),
uchar *info __attribute__((unused)),
size_t length __attribute__((unused)))
{
#ifdef EXTRA_DEBUG
LEX_CUSTRING log_array[TRANSLOG_INTERNAL_PARTS + 2];
uchar debug_type;
LSN lsn;
if (!trn)
{
/*
We can't log the current transaction because we don't have
an active transaction. Use a temporary transaction object instead
*/
trn= &dummy_transaction_object;
}
debug_type= (uchar) type;
log_array[TRANSLOG_INTERNAL_PARTS + 0].str= &debug_type;
log_array[TRANSLOG_INTERNAL_PARTS + 0].length= 1;
log_array[TRANSLOG_INTERNAL_PARTS + 1].str= info;
log_array[TRANSLOG_INTERNAL_PARTS + 1].length= length;
return translog_write_record(&lsn, LOGREC_DEBUG_INFO,
trn, NULL,
(translog_size_t) (1+ length),
sizeof(log_array)/sizeof(log_array[0]),
log_array, NULL, NULL);
#else
return 0;
#endif
}
/**
Sets soft sync mode
@param mode TRUE if we need switch soft sync on else off
*/
void translog_soft_sync(my_bool mode)
{
soft_sync= mode;
}
/**
Sets hard group commit
@param mode TRUE if we need switch hard group commit on else off
*/
void translog_hard_group_commit(my_bool mode)
{
hard_group_commit= mode;
}
/**
@brief forced log sync (used when we are switching modes)
*/
void translog_sync()
{
uint32 max= get_current_logfile()->number;
uint32 min;
DBUG_ENTER("ma_translog_sync");
min= soft_sync_min;
if (!min)
min= max;
translog_sync_files(min, max, sync_log_dir >= TRANSLOG_SYNC_DIR_ALWAYS);
DBUG_VOID_RETURN;
}
/**
@brief set rate for group commit
@param interval interval to set.
@note We use this function with additional variable because have to
restart service thread with new value which we can't make inside changing
variable routine (update_maria_group_commit_interval)
*/
void translog_set_group_commit_interval(uint32 interval)
{
DBUG_ENTER("translog_set_group_commit_interval");
group_commit_wait= interval;
DBUG_PRINT("info", ("wait: %llu",
(ulonglong)group_commit_wait));
DBUG_VOID_RETURN;
}
/**
@brief syncing service thread
*/
static pthread_handler_t
ma_soft_sync_background( void *arg __attribute__((unused)))
{
my_thread_init();
{
DBUG_ENTER("ma_soft_sync_background");
for(;;)
{
ulonglong prev_loop= microsecond_interval_timer();
ulonglong time, sleep;
uint32 min, max, sync_request;
min= soft_sync_min;
max= soft_sync_max;
sync_request= soft_need_sync;
soft_sync_min= max;
soft_need_sync= 0;
sleep= group_commit_wait;
if (sync_request)
translog_sync_files(min, max, FALSE);
time= microsecond_interval_timer() - prev_loop;
if (time > sleep)
sleep= 0;
else
sleep-= time;
if (my_service_thread_sleep(&soft_sync_control, sleep))
break;
}
my_service_thread_signal_end(&soft_sync_control);
my_thread_end();
DBUG_RETURN(0);
}
}
/**
@brief Starts syncing thread
*/
int translog_soft_sync_start(void)
{
pthread_t th;
int res= 0;
uint32 min, max;
DBUG_ENTER("translog_soft_sync_start");
/* check and init variables */
min= soft_sync_min;
max= soft_sync_max;
if (!max)
soft_sync_max= max= get_current_logfile()->number;
if (!min)
soft_sync_min= max;
soft_need_sync= 1;
if (!(res= ma_service_thread_control_init(&soft_sync_control)))
if (!(res= mysql_thread_create(key_thread_soft_sync,
&th, NULL, ma_soft_sync_background, NULL)))
soft_sync_control.status= THREAD_RUNNING;
DBUG_RETURN(res);
}
/**
@brief Stops syncing thread
*/
void translog_soft_sync_end(void)
{
DBUG_ENTER("translog_soft_sync_end");
if (soft_sync_control.inited)
{
ma_service_thread_control_end(&soft_sync_control);
}
DBUG_VOID_RETURN;
}
/**
@brief Dump information about file header page.
*/
static void dump_header_page(uchar *buff)
{
LOGHANDLER_FILE_INFO desc;
char strbuff[21];
translog_interpret_file_header(&desc, buff);
printf(" This can be header page:\n"
" Timestamp: %s\n"
" Aria log version: %lu\n"
" Server version: %lu\n"
" Server id %lu\n"
" Page size %lu\n",
llstr(desc.timestamp, strbuff),
desc.maria_version,
desc.mysql_version,
desc.server_id,
desc.page_size);
if (desc.page_size != TRANSLOG_PAGE_SIZE)
printf(" WARNING: page size is not equal compiled in one %lu!!!\n",
(ulong) TRANSLOG_PAGE_SIZE);
printf(" File number %lu\n"
" Max lsn: (%lu,0x%lx)\n",
desc.file_number,
LSN_IN_PARTS(desc.max_lsn));
}
static const char *record_class_string[]=
{
"LOGRECTYPE_NOT_ALLOWED",
"LOGRECTYPE_VARIABLE_LENGTH",
"LOGRECTYPE_PSEUDOFIXEDLENGTH",
"LOGRECTYPE_FIXEDLENGTH"
};
/**
@brief dump information about transaction log chunk
@param buffer reference to the whole page
@param ptr pointer to the chunk
@reval # reference to the next chunk
@retval NULL can't interpret data
*/
static uchar *dump_chunk(uchar *buffer, uchar *ptr)
{
uint length;
if (*ptr == TRANSLOG_FILLER)
{
printf(" Filler till the page end\n");
for (; ptr < buffer + TRANSLOG_PAGE_SIZE; ptr++)
{
if (*ptr != TRANSLOG_FILLER)
{
printf(" WARNING: non filler character met before page end "
"(page + 0x%04x: 0x%02x) (stop interpretation)!!!",
(uint) (ptr - buffer), (uint) ptr[0]);
return NULL;
}
}
return ptr;
}
if (*ptr == 0 || *ptr == 0xFF)
{
printf(" WARNING: chunk can't start from 0x0 "
"(stop interpretation)!!!\n");
return NULL;
}
switch (ptr[0] & TRANSLOG_CHUNK_TYPE) {
case TRANSLOG_CHUNK_LSN:
printf(" LSN chunk type 0 (variable length)\n");
if (likely((ptr[0] & TRANSLOG_REC_TYPE) != TRANSLOG_CHUNK_0_CONT))
{
printf(" Record type %u: %s record class %s compressed LSNs: %u\n",
ptr[0] & TRANSLOG_REC_TYPE,
(log_record_type_descriptor[ptr[0] & TRANSLOG_REC_TYPE].name ?
log_record_type_descriptor[ptr[0] & TRANSLOG_REC_TYPE].name :
"NULL"),
record_class_string[log_record_type_descriptor[ptr[0] &
TRANSLOG_REC_TYPE].
rclass],
log_record_type_descriptor[ptr[0] & TRANSLOG_REC_TYPE].
compressed_LSN);
if (log_record_type_descriptor[ptr[0] & TRANSLOG_REC_TYPE].rclass !=
LOGRECTYPE_VARIABLE_LENGTH)
{
printf(" WARNING: this record class here can't be used "
"(stop interpretation)!!!\n");
break;
}
}
else
printf(" Continuation of previous chunk 0 header \n");
printf(" Short transaction id: %u\n", (uint) uint2korr(ptr + 1));
{
uchar *hdr_ptr= ptr + 1 + 2; /* chunk type and short trid */
uint16 chunk_len;
printf (" Record length: %lu\n",
(ulong) translog_variable_record_1group_decode_len(&hdr_ptr));
chunk_len= uint2korr(hdr_ptr);
if (chunk_len == 0)
printf (" It is 1 group record (chunk length == 0)\n");
else
{
uint16 groups, i;
printf (" Chunk length %u\n", (uint) chunk_len);
groups= uint2korr(hdr_ptr + 2);
hdr_ptr+= 4;
printf (" Number of groups left to the end %u:\n", (uint) groups);
for(i= 0;
i < groups && hdr_ptr < buffer + TRANSLOG_PAGE_SIZE;
i++, hdr_ptr+= LSN_STORE_SIZE + 1)
{
TRANSLOG_ADDRESS gpr_addr= lsn_korr(hdr_ptr);
uint pages= hdr_ptr[LSN_STORE_SIZE];
printf (" Group +#%u: (%lu,0x%lx) pages: %u\n",
(uint) i, LSN_IN_PARTS(gpr_addr), pages);
}
}
}
break;
case TRANSLOG_CHUNK_FIXED:
printf(" LSN chunk type 1 (fixed size)\n");
printf(" Record type %u: %s record class %s compressed LSNs: %u\n",
ptr[0] & TRANSLOG_REC_TYPE,
(log_record_type_descriptor[ptr[0] & TRANSLOG_REC_TYPE].name ?
log_record_type_descriptor[ptr[0] & TRANSLOG_REC_TYPE].name :
"NULL"),
record_class_string[log_record_type_descriptor[ptr[0] &
TRANSLOG_REC_TYPE].
rclass],
log_record_type_descriptor[ptr[0] & TRANSLOG_REC_TYPE].
compressed_LSN);
if (log_record_type_descriptor[ptr[0] & TRANSLOG_REC_TYPE].rclass !=
LOGRECTYPE_PSEUDOFIXEDLENGTH &&
log_record_type_descriptor[ptr[0] & TRANSLOG_REC_TYPE].rclass !=
LOGRECTYPE_FIXEDLENGTH)
{
printf(" WARNING: this record class here can't be used "
"(stop interpretation)!!!\n");
}
printf(" Short transaction id: %u\n", (uint) uint2korr(ptr + 1));
break;
case TRANSLOG_CHUNK_NOHDR:
printf(" No header chunk type 2(till the end of the page)\n");
if (ptr[0] & TRANSLOG_REC_TYPE)
{
printf(" WARNING: chunk header content record type: 0x%02x "
"(dtop interpretation)!!!",
(uint) ptr[0]);
return NULL;
}
break;
case TRANSLOG_CHUNK_LNGTH:
printf(" Chunk with length type 3\n");
if (ptr[0] & TRANSLOG_REC_TYPE)
{
printf(" WARNING: chunk header content record type: 0x%02x "
"(dtop interpretation)!!!",
(uint) ptr[0]);
return NULL;
}
break;
}
{
intptr offset= ptr - buffer;
DBUG_ASSERT(offset >= 0 && offset <= UINT_MAX16);
length= translog_get_total_chunk_length(buffer, (uint16)offset);
}
printf(" Length %u\n", length);
ptr+= length;
return ptr;
}
/**
@brief Dump information about page with data.
*/
static void dump_datapage(uchar *buffer, File handler)
{
uchar *ptr;
ulong offset;
uint32 page, file;
uint header_len;
printf(" Page: %ld File number: %ld\n",
(ulong) (page= uint3korr(buffer)),
(ulong) (file= uint3korr(buffer + 3)));
if (page == 0)
printf(" WARNING: page == 0!!!\n");
if (file == 0)
printf(" WARNING: file == 0!!!\n");
offset= page * TRANSLOG_PAGE_SIZE;
printf(" Flags (0x%x):\n", (uint) buffer[TRANSLOG_PAGE_FLAGS]);
if (buffer[TRANSLOG_PAGE_FLAGS])
{
if (buffer[TRANSLOG_PAGE_FLAGS] & TRANSLOG_PAGE_CRC)
printf(" Page CRC\n");
if (buffer[TRANSLOG_PAGE_FLAGS] & TRANSLOG_SECTOR_PROTECTION)
printf(" Sector protection\n");
if (buffer[TRANSLOG_PAGE_FLAGS] & TRANSLOG_RECORD_CRC)
printf(" Record CRC (WARNING: not yet implemented!!!)\n");
if (buffer[TRANSLOG_PAGE_FLAGS] & ~(TRANSLOG_PAGE_CRC |
TRANSLOG_SECTOR_PROTECTION |
TRANSLOG_RECORD_CRC))
{
printf(" WARNING: unknown flags (stop interpretation)!!!\n");
return;
}
}
else
printf(" No flags\n");
printf(" Page header length: %u\n",
(header_len= page_overhead[buffer[TRANSLOG_PAGE_FLAGS]]));
if (buffer[TRANSLOG_PAGE_FLAGS] & TRANSLOG_RECORD_CRC)
{
uint32 crc= uint4korr(buffer + TRANSLOG_PAGE_FLAGS + 1);
uint32 ccrc;
printf (" Page CRC 0x%04lx\n", (ulong) crc);
ccrc= translog_crc(buffer + header_len, TRANSLOG_PAGE_SIZE - header_len);
if (crc != ccrc)
printf(" WARNING: calculated CRC: 0x%04lx!!!\n", (ulong) ccrc);
}
if (buffer[TRANSLOG_PAGE_FLAGS] & TRANSLOG_SECTOR_PROTECTION)
{
TRANSLOG_FILE tfile;
{
uchar *table= buffer + header_len -
TRANSLOG_PAGE_SIZE / DISK_DRIVE_SECTOR_SIZE;
uint i;
printf(" Sector protection current value: 0x%02x\n", (uint) table[0]);
for (i= 1; i < TRANSLOG_PAGE_SIZE / DISK_DRIVE_SECTOR_SIZE; i++)
{
printf(" Sector protection in sector: 0x%02x saved value 0x%02x\n",
(uint)buffer[i * DISK_DRIVE_SECTOR_SIZE],
(uint)table[i]);
}
}
tfile.number= file;
tfile.handler.file= handler;
pagecache_file_init(tfile.handler, NULL, NULL, NULL, NULL, NULL);
tfile.was_recovered= 0;
tfile.is_sync= 1;
if (translog_check_sector_protection(buffer, &tfile))
printf(" WARNING: sector protection found problems!!!\n");
}
ptr= buffer + header_len;
while (ptr && ptr < buffer + TRANSLOG_PAGE_SIZE)
{
printf(" Chunk (%lu,0x%lx):\n",
(ulong)file, (ulong) offset + (ptr - buffer));
ptr= dump_chunk(buffer, ptr);
}
}
/**
@brief Dump information about page.
*/
void dump_page(uchar *buffer, File handler)
{
if (strncmp((char*)maria_trans_file_magic, (char*)buffer,
sizeof(maria_trans_file_magic)) == 0)
{
dump_header_page(buffer);
}
dump_datapage(buffer, handler);
}
|