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
|
/*****************************************************************************
Copyright (c) 1995, 2017, Oracle and/or its affiliates. All Rights Reserved.
Copyright (c) 2014, 2020, MariaDB Corporation.
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.,
51 Franklin Street, Fifth Floor, Boston, MA 02110-1335 USA
*****************************************************************************/
/**************************************************//**
@file fil/fil0fil.cc
The tablespace memory cache
Created 10/25/1995 Heikki Tuuri
*******************************************************/
#include "fil0fil.h"
#include "fil0crypt.h"
#include "btr0btr.h"
#include "buf0buf.h"
#include "dict0boot.h"
#include "dict0dict.h"
#include "dict0load.h"
#include "fsp0file.h"
#include "fsp0fsp.h"
#include "hash0hash.h"
#include "log0log.h"
#include "log0recv.h"
#include "mach0data.h"
#include "mtr0log.h"
#include "os0file.h"
#include "page0zip.h"
#include "row0mysql.h"
#include "row0trunc.h"
#include "srv0start.h"
#include "trx0purge.h"
#include "buf0lru.h"
#include "ibuf0ibuf.h"
#include "os0event.h"
#include "sync0sync.h"
#include "buf0flu.h"
#include "os0api.h"
/** Tries to close a file in the LRU list. The caller must hold the fil_sys
mutex.
@return true if success, false if should retry later; since i/o's
generally complete in < 100 ms, and as InnoDB writes at most 128 pages
from the buffer pool in a batch, and then immediately flushes the
files, there is a good chance that the next time we find a suitable
node from the LRU list.
@param[in] print_info if true, prints information why it
cannot close a file */
static
bool
fil_try_to_close_file_in_LRU(bool print_info);
/** Test if a tablespace file can be renamed to a new filepath by checking
if that the old filepath exists and the new filepath does not exist.
@param[in] old_path old filepath
@param[in] new_path new filepath
@param[in] is_discarded whether the tablespace is discarded
@param[in] replace_new whether to ignore the existence of new_path
@return innodb error code */
static dberr_t
fil_rename_tablespace_check(
const char* old_path,
const char* new_path,
bool is_discarded,
bool replace_new = false);
/** Rename a single-table tablespace.
The tablespace must exist in the memory cache.
@param[in] id tablespace identifier
@param[in] old_path old file name
@param[in] new_name new table name in the
databasename/tablename format
@param[in] new_path_in new file name,
or NULL if it is located in the normal data directory
@return true if success */
static bool
fil_rename_tablespace(
ulint id,
const char* old_path,
const char* new_name,
const char* new_path_in);
/*
IMPLEMENTATION OF THE TABLESPACE MEMORY CACHE
=============================================
The tablespace cache is responsible for providing fast read/write access to
tablespaces and logs of the database. File creation and deletion is done
in other modules which know more of the logic of the operation, however.
A tablespace consists of a chain of files. The size of the files does not
have to be divisible by the database block size, because we may just leave
the last incomplete block unused. When a new file is appended to the
tablespace, the maximum size of the file is also specified. At the moment,
we think that it is best to extend the file to its maximum size already at
the creation of the file, because then we can avoid dynamically extending
the file when more space is needed for the tablespace.
A block's position in the tablespace is specified with a 32-bit unsigned
integer. The files in the chain are thought to be catenated, and the block
corresponding to an address n is the nth block in the catenated file (where
the first block is named the 0th block, and the incomplete block fragments
at the end of files are not taken into account). A tablespace can be extended
by appending a new file at the end of the chain.
Our tablespace concept is similar to the one of Oracle.
To acquire more speed in disk transfers, a technique called disk striping is
sometimes used. This means that logical block addresses are divided in a
round-robin fashion across several disks. Windows NT supports disk striping,
so there we do not need to support it in the database. Disk striping is
implemented in hardware in RAID disks. We conclude that it is not necessary
to implement it in the database. Oracle 7 does not support disk striping,
either.
Another trick used at some database sites is replacing tablespace files by
raw disks, that is, the whole physical disk drive, or a partition of it, is
opened as a single file, and it is accessed through byte offsets calculated
from the start of the disk or the partition. This is recommended in some
books on database tuning to achieve more speed in i/o. Using raw disk
certainly prevents the OS from fragmenting disk space, but it is not clear
if it really adds speed. We measured on the Pentium 100 MHz + NT + NTFS file
system + EIDE Conner disk only a negligible difference in speed when reading
from a file, versus reading from a raw disk.
To have fast access to a tablespace or a log file, we put the data structures
to a hash table. Each tablespace and log file is given an unique 32-bit
identifier.
Some operating systems do not support many open files at the same time,
though NT seems to tolerate at least 900 open files. Therefore, we put the
open files in an LRU-list. If we need to open another file, we may close the
file at the end of the LRU-list. When an i/o-operation is pending on a file,
the file cannot be closed. We take the file nodes with pending i/o-operations
out of the LRU-list and keep a count of pending operations. When an operation
completes, we decrement the count and return the file node to the LRU-list if
the count drops to zero. */
/** Reference to the server data directory. Usually it is the
current working directory ".", but in the MySQL Embedded Server Library
it is an absolute path. */
const char* fil_path_to_mysql_datadir;
/** Common InnoDB file extensions */
const char* dot_ext[] = { "", ".ibd", ".isl", ".cfg" };
/** The number of fsyncs done to the log */
ulint fil_n_log_flushes = 0;
/** Number of pending redo log flushes */
ulint fil_n_pending_log_flushes = 0;
/** Number of pending tablespace flushes */
ulint fil_n_pending_tablespace_flushes = 0;
/** The null file address */
const fil_addr_t fil_addr_null = {FIL_NULL, 0};
/** The tablespace memory cache. This variable is NULL before the module is
initialized. */
fil_system_t fil_system;
/** At this age or older a space/page will be rotated */
UNIV_INTERN extern uint srv_fil_crypt_rotate_key_age;
UNIV_INTERN extern ib_mutex_t fil_crypt_threads_mutex;
/** Determine if user has explicitly disabled fsync(). */
# define fil_buffering_disabled(s) \
((s)->purpose == FIL_TYPE_TABLESPACE \
&& srv_file_flush_method \
== SRV_O_DIRECT_NO_FSYNC)
/** Determine if the space id is a user tablespace id or not.
@param[in] space_id Space ID to check
@return true if it is a user tablespace ID */
inline
bool
fil_is_user_tablespace_id(ulint space_id)
{
return(space_id != TRX_SYS_SPACE
&& space_id != SRV_TMP_SPACE_ID
&& !srv_is_undo_tablespace(space_id));
}
#ifdef UNIV_DEBUG
/** Try fil_validate() every this many times */
# define FIL_VALIDATE_SKIP 17
/******************************************************************//**
Checks the consistency of the tablespace cache some of the time.
@return true if ok or the check was skipped */
static
bool
fil_validate_skip(void)
/*===================*/
{
/** The fil_validate() call skip counter. */
static int fil_validate_count = FIL_VALIDATE_SKIP;
/* We want to reduce the call frequency of the costly fil_validate()
check in debug builds. */
int count = my_atomic_add32_explicit(&fil_validate_count, -1,
MY_MEMORY_ORDER_RELAXED);
if (count > 0) {
return(true);
}
my_atomic_store32_explicit(&fil_validate_count, FIL_VALIDATE_SKIP,
MY_MEMORY_ORDER_RELAXED);
return(fil_validate());
}
#endif /* UNIV_DEBUG */
/********************************************************************//**
Determines if a file node belongs to the least-recently-used list.
@return true if the file belongs to fil_system.LRU mutex. */
UNIV_INLINE
bool
fil_space_belongs_in_lru(
/*=====================*/
const fil_space_t* space) /*!< in: file space */
{
switch (space->purpose) {
case FIL_TYPE_TEMPORARY:
case FIL_TYPE_LOG:
return(false);
case FIL_TYPE_TABLESPACE:
return(fil_is_user_tablespace_id(space->id));
case FIL_TYPE_IMPORT:
return(true);
}
ut_ad(0);
return(false);
}
/********************************************************************//**
NOTE: you must call fil_mutex_enter_and_prepare_for_io() first!
Prepares a file node for i/o. Opens the file if it is closed. Updates the
pending i/o's field in the node and the system appropriately. Takes the node
off the LRU list if it is in the LRU list. The caller must hold the fil_sys
mutex.
@return false if the file can't be opened, otherwise true */
static
bool
fil_node_prepare_for_io(
/*====================*/
fil_node_t* node, /*!< in: file node */
fil_space_t* space); /*!< in: space */
/** Update the data structures when an i/o operation finishes.
@param[in,out] node file node
@param[in] type IO context */
static
void
fil_node_complete_io(fil_node_t* node, const IORequest& type);
/** Reads data from a space to a buffer. Remember that the possible incomplete
blocks at the end of file are ignored: they are not taken into account when
calculating the byte offset within a space.
@param[in] page_id page id
@param[in] page_size page size
@param[in] byte_offset remainder of offset in bytes; in aio this
must be divisible by the OS block size
@param[in] len how many bytes to read; this must not cross a
file boundary; in aio this must be a block size multiple
@param[in,out] buf buffer where to store data read; in aio this
must be appropriately aligned
@return DB_SUCCESS, or DB_TABLESPACE_DELETED if we are trying to do
i/o on a tablespace which does not exist */
UNIV_INLINE
dberr_t
fil_read(
const page_id_t page_id,
const page_size_t& page_size,
ulint byte_offset,
ulint len,
void* buf)
{
return(fil_io(IORequestRead, true, page_id, page_size,
byte_offset, len, buf, NULL));
}
/** Writes data to a space from a buffer. Remember that the possible incomplete
blocks at the end of file are ignored: they are not taken into account when
calculating the byte offset within a space.
@param[in] page_id page id
@param[in] page_size page size
@param[in] byte_offset remainder of offset in bytes; in aio this
must be divisible by the OS block size
@param[in] len how many bytes to write; this must not cross
a file boundary; in aio this must be a block size multiple
@param[in] buf buffer from which to write; in aio this must
be appropriately aligned
@return DB_SUCCESS, or DB_TABLESPACE_DELETED if we are trying to do
i/o on a tablespace which does not exist */
UNIV_INLINE
dberr_t
fil_write(
const page_id_t page_id,
const page_size_t& page_size,
ulint byte_offset,
ulint len,
void* buf)
{
ut_ad(!srv_read_only_mode);
return(fil_io(IORequestWrite, true, page_id, page_size,
byte_offset, len, buf, NULL));
}
/*******************************************************************//**
Returns the table space by a given id, NULL if not found.
It is unsafe to dereference the returned pointer. It is fine to check
for NULL. */
fil_space_t*
fil_space_get_by_id(
/*================*/
ulint id) /*!< in: space id */
{
fil_space_t* space;
ut_ad(fil_system.is_initialised());
ut_ad(mutex_own(&fil_system.mutex));
HASH_SEARCH(hash, fil_system.spaces, id,
fil_space_t*, space,
ut_ad(space->magic_n == FIL_SPACE_MAGIC_N),
space->id == id);
return(space);
}
/** Look up a tablespace.
The caller should hold an InnoDB table lock or a MDL that prevents
the tablespace from being dropped during the operation,
or the caller should be in single-threaded crash recovery mode
(no user connections that could drop tablespaces).
If this is not the case, fil_space_acquire() and fil_space_t::release()
should be used instead.
@param[in] id tablespace ID
@return tablespace, or NULL if not found */
fil_space_t*
fil_space_get(
ulint id)
{
mutex_enter(&fil_system.mutex);
fil_space_t* space = fil_space_get_by_id(id);
mutex_exit(&fil_system.mutex);
ut_ad(space == NULL || space->purpose != FIL_TYPE_LOG);
return(space);
}
/** Returns the latch of a file space.
@param[in] id space id
@param[out] flags tablespace flags
@return latch protecting storage allocation */
rw_lock_t*
fil_space_get_latch(
ulint id,
ulint* flags)
{
fil_space_t* space;
ut_ad(fil_system.is_initialised());
mutex_enter(&fil_system.mutex);
space = fil_space_get_by_id(id);
ut_a(space);
if (flags) {
*flags = space->flags;
}
mutex_exit(&fil_system.mutex);
return(&(space->latch));
}
/** Note that the tablespace has been imported.
Initially, purpose=FIL_TYPE_IMPORT so that no redo log is
written while the space ID is being updated in each page. */
void fil_space_t::set_imported()
{
ut_ad(purpose == FIL_TYPE_IMPORT);
const fil_node_t* node = UT_LIST_GET_FIRST(chain);
atomic_write_supported = node->atomic_write
&& srv_use_atomic_writes
&& my_test_if_atomic_write(node->handle,
int(page_size_t(flags).physical()));
purpose = FIL_TYPE_TABLESPACE;
}
/**********************************************************************//**
Checks if all the file nodes in a space are flushed.
@return true if all are flushed */
static
bool
fil_space_is_flushed(
/*=================*/
fil_space_t* space) /*!< in: space */
{
ut_ad(mutex_own(&fil_system.mutex));
for (const fil_node_t* node = UT_LIST_GET_FIRST(space->chain);
node != NULL;
node = UT_LIST_GET_NEXT(chain, node)) {
if (node->needs_flush) {
ut_ad(!fil_buffering_disabled(space));
return(false);
}
}
return(true);
}
/** Append a file to the chain of files of a space.
@param[in] name file name of a file that is not open
@param[in] handle file handle, or OS_FILE_CLOSED
@param[in] size file size in entire database pages
@param[in] is_raw whether this is a raw device
@param[in] atomic_write true if atomic write could be enabled
@param[in] max_pages maximum number of pages in file,
or ULINT_MAX for unlimited
@return file object */
fil_node_t* fil_space_t::add(const char* name, pfs_os_file_t handle,
ulint size, bool is_raw, bool atomic_write,
ulint max_pages)
{
fil_node_t* node;
ut_ad(name != NULL);
ut_ad(fil_system.is_initialised());
node = reinterpret_cast<fil_node_t*>(ut_zalloc_nokey(sizeof(*node)));
node->handle = handle;
node->name = mem_strdup(name);
ut_a(!is_raw || srv_start_raw_disk_in_use);
node->is_raw_disk = is_raw;
node->size = size;
node->magic_n = FIL_NODE_MAGIC_N;
node->init_size = size;
node->max_size = max_pages;
node->space = this;
node->atomic_write = atomic_write;
mutex_enter(&fil_system.mutex);
this->size += size;
UT_LIST_ADD_LAST(chain, node);
if (node->is_open()) {
fil_system.n_open++;
}
mutex_exit(&fil_system.mutex);
return node;
}
/** Read the first page of a data file.
@param[in] first whether this is the very first read
@return whether the page was found valid */
bool fil_node_t::read_page0(bool first)
{
ut_ad(mutex_own(&fil_system.mutex));
ut_a(space->purpose != FIL_TYPE_LOG);
const page_size_t page_size(space->flags);
const ulint psize = page_size.physical();
os_offset_t size_bytes = os_file_get_size(handle);
ut_a(size_bytes != (os_offset_t) -1);
const ulint min_size = FIL_IBD_FILE_INITIAL_SIZE * psize;
if (size_bytes < min_size) {
ib::error() << "The size of the file " << name
<< " is only " << size_bytes
<< " bytes, should be at least " << min_size;
return false;
}
byte* buf2 = static_cast<byte*>(ut_malloc_nokey(2 * psize));
/* Align the memory for file i/o if we might have O_DIRECT set */
byte* page = static_cast<byte*>(ut_align(buf2, psize));
IORequest request(IORequest::READ);
if (os_file_read(request, handle, page, 0, psize) != DB_SUCCESS) {
ib::error() << "Unable to read first page of file " << name;
ut_free(buf2);
return false;
}
srv_stats.page0_read.add(1);
const ulint space_id = fsp_header_get_space_id(page);
ulint flags = fsp_header_get_flags(page);
const ulint size = fsp_header_get_field(page, FSP_SIZE);
const ulint free_limit = fsp_header_get_field(page, FSP_FREE_LIMIT);
const ulint free_len = flst_get_len(FSP_HEADER_OFFSET + FSP_FREE
+ page);
/* Try to read crypt_data from page 0 if it is not yet read. */
if (!space->crypt_data) {
space->crypt_data = fil_space_read_crypt_data(page_size, page);
}
ut_free(buf2);
if (!fsp_flags_is_valid(flags, space->id)) {
ulint cflags = fsp_flags_convert_from_101(flags);
if (cflags == ULINT_UNDEFINED
|| (cflags ^ space->flags) & ~FSP_FLAGS_MEM_MASK) {
ib::error()
<< "Expected tablespace flags "
<< ib::hex(space->flags)
<< " but found " << ib::hex(flags)
<< " in the file " << name;
return false;
}
flags = cflags;
}
if (UNIV_UNLIKELY(space_id != space->id)) {
ib::error() << "Expected tablespace id " << space->id
<< " but found " << space_id
<< " in the file " << name;
return false;
}
if (first) {
ut_ad(space->id != TRX_SYS_SPACE);
/* Truncate the size to a multiple of extent size. */
ulint mask = psize * FSP_EXTENT_SIZE - 1;
if (size_bytes <= mask) {
/* .ibd files start smaller than an
extent size. Do not truncate valid data. */
} else {
size_bytes &= ~os_offset_t(mask);
}
this->size = ulint(size_bytes / psize);
space->size += this->size;
} else if (space->id != TRX_SYS_SPACE || space->size_in_header) {
/* If this is not the first-time open, do nothing.
For the system tablespace, we always get invoked as
first=false, so we detect the true first-time-open based
on size_in_header and proceed to initiailze the data. */
return true;
}
ut_ad(space->free_limit == 0 || space->free_limit == free_limit);
ut_ad(space->free_len == 0 || space->free_len == free_len);
space->size_in_header = size;
space->free_limit = free_limit;
space->free_len = free_len;
return true;
}
/** Open a file node of a tablespace.
@param[in,out] node File node
@return false if the file can't be opened, otherwise true */
static bool fil_node_open_file(fil_node_t* node)
{
bool success;
bool read_only_mode;
fil_space_t* space = node->space;
ut_ad(mutex_own(&fil_system.mutex));
ut_a(node->n_pending == 0);
ut_a(!node->is_open());
read_only_mode = space->purpose != FIL_TYPE_TEMPORARY
&& srv_read_only_mode;
const bool first_time_open = node->size == 0;
if (first_time_open
|| (space->purpose == FIL_TYPE_TABLESPACE
&& node == UT_LIST_GET_FIRST(space->chain)
&& srv_startup_is_before_trx_rollback_phase
&& !undo::Truncate::was_tablespace_truncated(space->id))) {
/* We do not know the size of the file yet. First we
open the file in the normal mode, no async I/O here,
for simplicity. Then do some checks, and close the
file again. NOTE that we could not use the simple
file read function os_file_read() in Windows to read
from a file opened for async I/O! */
retry:
node->handle = os_file_create(
innodb_data_file_key, node->name,
node->is_raw_disk
? OS_FILE_OPEN_RAW | OS_FILE_ON_ERROR_NO_EXIT
: OS_FILE_OPEN | OS_FILE_ON_ERROR_NO_EXIT,
OS_FILE_AIO, OS_DATA_FILE, read_only_mode, &success);
if (!success) {
/* The following call prints an error message */
ulint err = os_file_get_last_error(true);
if (err == EMFILE + 100) {
if (fil_try_to_close_file_in_LRU(true))
goto retry;
}
ib::warn() << "Cannot open '" << node->name << "'."
" Have you deleted .ibd files under a"
" running mysqld server?";
return(false);
}
if (!node->read_page0(first_time_open)) {
os_file_close(node->handle);
node->handle = OS_FILE_CLOSED;
return false;
}
} else if (space->purpose == FIL_TYPE_LOG) {
node->handle = os_file_create(
innodb_log_file_key, node->name, OS_FILE_OPEN,
OS_FILE_AIO, OS_LOG_FILE, read_only_mode, &success);
} else {
node->handle = os_file_create(
innodb_data_file_key, node->name,
node->is_raw_disk
? OS_FILE_OPEN_RAW | OS_FILE_ON_ERROR_NO_EXIT
: OS_FILE_OPEN | OS_FILE_ON_ERROR_NO_EXIT,
OS_FILE_AIO, OS_DATA_FILE, read_only_mode, &success);
}
if (space->purpose != FIL_TYPE_LOG) {
/*
For the temporary tablespace and during the
non-redo-logged adjustments in
IMPORT TABLESPACE, we do not care about
the atomicity of writes.
Atomic writes is supported if the file can be used
with atomic_writes (not log file), O_DIRECT is
used (tested in ha_innodb.cc) and the file is
device and file system that supports atomic writes
for the given block size
*/
space->atomic_write_supported
= space->purpose == FIL_TYPE_TEMPORARY
|| space->purpose == FIL_TYPE_IMPORT
|| (node->atomic_write
&& srv_use_atomic_writes
&& my_test_if_atomic_write(
node->handle,
int(page_size_t(space->flags)
.physical())));
}
ut_a(success);
ut_a(node->is_open());
fil_system.n_open++;
if (fil_space_belongs_in_lru(space)) {
/* Put the node to the LRU list */
UT_LIST_ADD_FIRST(fil_system.LRU, node);
}
return(true);
}
/** Close the file handle. */
void fil_node_t::close()
{
bool ret;
ut_ad(mutex_own(&fil_system.mutex));
ut_a(is_open());
ut_a(n_pending == 0);
ut_a(n_pending_flushes == 0);
ut_a(!being_extended);
ut_a(!needs_flush
|| space->purpose == FIL_TYPE_TEMPORARY
|| srv_fast_shutdown == 2
|| !srv_was_started);
ret = os_file_close(handle);
ut_a(ret);
/* printf("Closing file %s\n", name); */
handle = OS_FILE_CLOSED;
ut_ad(!is_open());
ut_a(fil_system.n_open > 0);
fil_system.n_open--;
if (fil_space_belongs_in_lru(space)) {
ut_a(UT_LIST_GET_LEN(fil_system.LRU) > 0);
UT_LIST_REMOVE(fil_system.LRU, this);
}
}
/** Tries to close a file in the LRU list. The caller must hold the fil_sys
mutex.
@return true if success, false if should retry later; since i/o's
generally complete in < 100 ms, and as InnoDB writes at most 128 pages
from the buffer pool in a batch, and then immediately flushes the
files, there is a good chance that the next time we find a suitable
node from the LRU list.
@param[in] print_info if true, prints information why it
cannot close a file*/
static
bool
fil_try_to_close_file_in_LRU(
bool print_info)
{
fil_node_t* node;
ut_ad(mutex_own(&fil_system.mutex));
if (print_info) {
ib::info() << "fil_sys open file LRU len "
<< UT_LIST_GET_LEN(fil_system.LRU);
}
for (node = UT_LIST_GET_LAST(fil_system.LRU);
node != NULL;
node = UT_LIST_GET_PREV(LRU, node)) {
if (!node->needs_flush
&& node->n_pending_flushes == 0
&& !node->being_extended) {
node->close();
return(true);
}
if (!print_info) {
continue;
}
if (node->n_pending_flushes > 0) {
ib::info() << "Cannot close file " << node->name
<< ", because n_pending_flushes "
<< node->n_pending_flushes;
}
if (node->needs_flush) {
ib::warn() << "Cannot close file " << node->name
<< ", because is should be flushed first";
}
if (node->being_extended) {
ib::info() << "Cannot close file " << node->name
<< ", because it is being extended";
}
}
return(false);
}
/** Flush any writes cached by the file system.
@param[in,out] space tablespace
@param[in] metadata whether to update file system metadata */
static void fil_flush_low(fil_space_t* space, bool metadata = false)
{
ut_ad(mutex_own(&fil_system.mutex));
ut_ad(space);
ut_ad(!space->stop_new_ops);
if (fil_buffering_disabled(space)) {
/* No need to flush. User has explicitly disabled
buffering. */
ut_ad(!space->is_in_unflushed_spaces);
ut_ad(fil_space_is_flushed(space));
ut_ad(space->n_pending_flushes == 0);
#ifdef UNIV_DEBUG
for (fil_node_t* node = UT_LIST_GET_FIRST(space->chain);
node != NULL;
node = UT_LIST_GET_NEXT(chain, node)) {
ut_ad(!node->needs_flush);
ut_ad(node->n_pending_flushes == 0);
}
#endif /* UNIV_DEBUG */
if (!metadata) return;
}
/* Prevent dropping of the space while we are flushing */
space->n_pending_flushes++;
for (fil_node_t* node = UT_LIST_GET_FIRST(space->chain);
node != NULL;
node = UT_LIST_GET_NEXT(chain, node)) {
if (!node->needs_flush) {
continue;
}
ut_a(node->is_open());
switch (space->purpose) {
case FIL_TYPE_TEMPORARY:
ut_ad(0); // we already checked for this
/* fall through */
case FIL_TYPE_TABLESPACE:
case FIL_TYPE_IMPORT:
fil_n_pending_tablespace_flushes++;
break;
case FIL_TYPE_LOG:
fil_n_pending_log_flushes++;
fil_n_log_flushes++;
break;
}
#ifdef _WIN32
if (node->is_raw_disk) {
goto skip_flush;
}
#endif /* _WIN32 */
ut_a(node->is_open());
node->n_pending_flushes++;
node->needs_flush = false;
mutex_exit(&fil_system.mutex);
os_file_flush(node->handle);
mutex_enter(&fil_system.mutex);
node->n_pending_flushes--;
#ifdef _WIN32
skip_flush:
#endif /* _WIN32 */
if (!node->needs_flush) {
if (space->is_in_unflushed_spaces
&& fil_space_is_flushed(space)) {
fil_system.unflushed_spaces.remove(*space);
space->is_in_unflushed_spaces = false;
}
}
switch (space->purpose) {
case FIL_TYPE_TEMPORARY:
break;
case FIL_TYPE_TABLESPACE:
case FIL_TYPE_IMPORT:
fil_n_pending_tablespace_flushes--;
continue;
case FIL_TYPE_LOG:
fil_n_pending_log_flushes--;
continue;
}
ut_ad(0);
}
space->n_pending_flushes--;
}
/** Try to extend a tablespace.
@param[in,out] space tablespace to be extended
@param[in,out] node last file of the tablespace
@param[in] size desired size in number of pages
@param[out] success whether the operation succeeded
@return whether the operation should be retried */
static ATTRIBUTE_COLD __attribute__((warn_unused_result, nonnull))
bool
fil_space_extend_must_retry(
fil_space_t* space,
fil_node_t* node,
ulint size,
bool* success)
{
ut_ad(mutex_own(&fil_system.mutex));
ut_ad(UT_LIST_GET_LAST(space->chain) == node);
ut_ad(size >= FIL_IBD_FILE_INITIAL_SIZE);
*success = space->size >= size;
if (*success) {
/* Space already big enough */
return(false);
}
if (node->being_extended) {
/* Another thread is currently extending the file. Wait
for it to finish.
It'd have been better to use event driven mechanism but
the entire module is peppered with polling stuff. */
mutex_exit(&fil_system.mutex);
os_thread_sleep(100000);
return(true);
}
node->being_extended = true;
if (!fil_node_prepare_for_io(node, space)) {
/* The tablespace data file, such as .ibd file, is missing */
node->being_extended = false;
return(false);
}
/* At this point it is safe to release fil_system.mutex. No
other thread can rename, delete, close or extend the file because
we have set the node->being_extended flag. */
mutex_exit(&fil_system.mutex);
ut_ad(size >= space->size);
ulint last_page_no = space->size;
const ulint file_start_page_no = last_page_no - node->size;
/* Determine correct file block size */
if (node->block_size == 0) {
node->block_size = os_file_get_block_size(
node->handle, node->name);
}
const page_size_t pageSize(space->flags);
const ulint page_size = pageSize.physical();
/* fil_read_first_page() expects srv_page_size bytes.
fil_node_open_file() expects at least 4 * srv_page_size bytes.*/
os_offset_t new_size = std::max(
os_offset_t(size - file_start_page_no) * page_size,
os_offset_t(FIL_IBD_FILE_INITIAL_SIZE << srv_page_size_shift));
*success = os_file_set_size(node->name, node->handle, new_size,
FSP_FLAGS_HAS_PAGE_COMPRESSION(space->flags));
os_has_said_disk_full = *success;
if (*success) {
os_file_flush(node->handle);
last_page_no = size;
} else {
/* Let us measure the size of the file
to determine how much we were able to
extend it */
os_offset_t fsize = os_file_get_size(node->handle);
ut_a(fsize != os_offset_t(-1));
last_page_no = ulint(fsize / page_size)
+ file_start_page_no;
}
mutex_enter(&fil_system.mutex);
ut_a(node->being_extended);
node->being_extended = false;
ut_a(last_page_no - file_start_page_no >= node->size);
ulint file_size = last_page_no - file_start_page_no;
space->size += file_size - node->size;
node->size = file_size;
const ulint pages_in_MiB = node->size
& ~ulint((1U << (20U - srv_page_size_shift)) - 1);
fil_node_complete_io(node,IORequestRead);
/* Keep the last data file size info up to date, rounded to
full megabytes */
switch (space->id) {
case TRX_SYS_SPACE:
srv_sys_space.set_last_file_size(pages_in_MiB);
fil_flush_low(space, true);
return(false);
default:
ut_ad(space->purpose == FIL_TYPE_TABLESPACE
|| space->purpose == FIL_TYPE_IMPORT);
if (space->purpose == FIL_TYPE_TABLESPACE
&& !space->is_being_truncated) {
fil_flush_low(space, true);
}
return(false);
case SRV_TMP_SPACE_ID:
ut_ad(space->purpose == FIL_TYPE_TEMPORARY);
srv_tmp_space.set_last_file_size(pages_in_MiB);
return(false);
}
}
/*******************************************************************//**
Reserves the fil_system.mutex and tries to make sure we can open at least one
file while holding it. This should be called before calling
fil_node_prepare_for_io(), because that function may need to open a file. */
static
void
fil_mutex_enter_and_prepare_for_io(
/*===============================*/
ulint space_id) /*!< in: space id */
{
for (ulint count = 0;;) {
mutex_enter(&fil_system.mutex);
if (space_id >= SRV_LOG_SPACE_FIRST_ID) {
/* We keep log files always open. */
break;
}
fil_space_t* space = fil_space_get_by_id(space_id);
if (space == NULL) {
break;
}
fil_node_t* node = UT_LIST_GET_LAST(space->chain);
ut_ad(space->id == 0
|| node == UT_LIST_GET_FIRST(space->chain));
if (space->id == 0) {
/* We keep the system tablespace files always
open; this is important in preventing
deadlocks in this module, as a page read
completion often performs another read from
the insert buffer. The insert buffer is in
tablespace 0, and we cannot end up waiting in
this function. */
} else if (!node || node->is_open()) {
/* If the file is already open, no need to do
anything; if the space does not exist, we handle the
situation in the function which called this
function */
} else {
while (fil_system.n_open >= srv_max_n_open_files) {
/* Too many files are open */
if (fil_try_to_close_file_in_LRU(count > 1)) {
/* No problem */
} else if (count >= 2) {
ib::warn() << "innodb_open_files="
<< srv_max_n_open_files
<< " is exceeded ("
<< fil_system.n_open
<< ") files stay open)";
break;
} else {
mutex_exit(&fil_system.mutex);
os_aio_simulated_wake_handler_threads();
os_thread_sleep(20000);
/* Flush tablespaces so that we can
close modified files in the LRU list */
fil_flush_file_spaces(FIL_TYPE_TABLESPACE);
count++;
mutex_enter(&fil_system.mutex);
continue;
}
}
}
ulint size = space->recv_size;
if (UNIV_UNLIKELY(size != 0)) {
ut_ad(node);
bool success;
if (fil_space_extend_must_retry(space, node, size,
&success)) {
continue;
}
ut_ad(mutex_own(&fil_system.mutex));
/* Crash recovery requires the file extension
to succeed. */
ut_a(success);
/* InnoDB data files cannot shrink. */
ut_a(space->size >= size);
/* There could be multiple concurrent I/O requests for
this tablespace (multiple threads trying to extend
this tablespace).
Also, fil_space_set_recv_size() may have been invoked
again during the file extension while fil_system.mutex
was not being held by us.
Only if space->recv_size matches what we read
originally, reset the field. In this way, a
subsequent I/O request will handle any pending
fil_space_set_recv_size(). */
if (size == space->recv_size) {
space->recv_size = 0;
}
}
break;
}
}
/** Try to extend a tablespace if it is smaller than the specified size.
@param[in,out] space tablespace
@param[in] size desired size in pages
@return whether the tablespace is at least as big as requested */
bool
fil_space_extend(
fil_space_t* space,
ulint size)
{
ut_ad(!srv_read_only_mode || space->purpose == FIL_TYPE_TEMPORARY);
bool success;
do {
fil_mutex_enter_and_prepare_for_io(space->id);
} while (fil_space_extend_must_retry(
space, UT_LIST_GET_LAST(space->chain), size,
&success));
mutex_exit(&fil_system.mutex);
return(success);
}
/** Prepare to free a file node object from a tablespace memory cache.
@param[in,out] node file node
@param[in] space tablespace */
static
void
fil_node_close_to_free(
fil_node_t* node,
fil_space_t* space)
{
ut_ad(mutex_own(&fil_system.mutex));
ut_a(node->magic_n == FIL_NODE_MAGIC_N);
ut_a(node->n_pending == 0);
ut_a(!node->being_extended);
if (node->is_open()) {
/* We fool the assertion in fil_node_t::close() to think
there are no unflushed modifications in the file */
node->needs_flush = false;
if (fil_buffering_disabled(space)) {
ut_ad(!space->is_in_unflushed_spaces);
ut_ad(fil_space_is_flushed(space));
} else if (space->is_in_unflushed_spaces
&& fil_space_is_flushed(space)) {
fil_system.unflushed_spaces.remove(*space);
space->is_in_unflushed_spaces = false;
}
node->close();
}
}
/** Detach a space object from the tablespace memory cache.
Closes the files in the chain but does not delete them.
There must not be any pending i/o's or flushes on the files.
@param[in,out] space tablespace */
static
void
fil_space_detach(
fil_space_t* space)
{
ut_ad(mutex_own(&fil_system.mutex));
HASH_DELETE(fil_space_t, hash, fil_system.spaces, space->id, space);
if (space->is_in_unflushed_spaces) {
ut_ad(!fil_buffering_disabled(space));
fil_system.unflushed_spaces.remove(*space);
space->is_in_unflushed_spaces = false;
}
if (space->is_in_rotation_list) {
fil_system.rotation_list.remove(*space);
space->is_in_rotation_list = false;
}
UT_LIST_REMOVE(fil_system.space_list, space);
ut_a(space->magic_n == FIL_SPACE_MAGIC_N);
ut_a(space->n_pending_flushes == 0);
for (fil_node_t* fil_node = UT_LIST_GET_FIRST(space->chain);
fil_node != NULL;
fil_node = UT_LIST_GET_NEXT(chain, fil_node)) {
fil_node_close_to_free(fil_node, space);
}
if (space == fil_system.sys_space) {
fil_system.sys_space = NULL;
} else if (space == fil_system.temp_space) {
fil_system.temp_space = NULL;
}
}
/** Free a tablespace object on which fil_space_detach() was invoked.
There must not be any pending i/o's or flushes on the files.
@param[in,out] space tablespace */
static
void
fil_space_free_low(
fil_space_t* space)
{
/* The tablespace must not be in fil_system.named_spaces. */
ut_ad(srv_fast_shutdown == 2 || !srv_was_started
|| space->max_lsn == 0);
/* Wait for fil_space_t::release_for_io(); after
fil_space_detach(), the tablespace cannot be found, so
fil_space_acquire_for_io() would return NULL */
while (space->pending_io()) {
os_thread_sleep(100);
}
for (fil_node_t* node = UT_LIST_GET_FIRST(space->chain);
node != NULL; ) {
ut_d(space->size -= node->size);
ut_free(node->name);
fil_node_t* old_node = node;
node = UT_LIST_GET_NEXT(chain, node);
ut_free(old_node);
}
ut_ad(space->size == 0);
rw_lock_free(&space->latch);
fil_space_destroy_crypt_data(&space->crypt_data);
ut_free(space->name);
ut_free(space);
}
/** Frees a space object from the tablespace memory cache.
Closes the files in the chain but does not delete them.
There must not be any pending i/o's or flushes on the files.
@param[in] id tablespace identifier
@param[in] x_latched whether the caller holds X-mode space->latch
@return true if success */
bool
fil_space_free(
ulint id,
bool x_latched)
{
ut_ad(id != TRX_SYS_SPACE);
mutex_enter(&fil_system.mutex);
fil_space_t* space = fil_space_get_by_id(id);
if (space != NULL) {
fil_space_detach(space);
}
mutex_exit(&fil_system.mutex);
if (space != NULL) {
if (x_latched) {
rw_lock_x_unlock(&space->latch);
}
if (!recv_recovery_is_on()) {
log_mutex_enter();
}
ut_ad(log_mutex_own());
if (space->max_lsn != 0) {
ut_d(space->max_lsn = 0);
UT_LIST_REMOVE(fil_system.named_spaces, space);
}
if (!recv_recovery_is_on()) {
log_mutex_exit();
}
fil_space_free_low(space);
}
return(space != NULL);
}
/** Create a space memory object and put it to the fil_system hash table.
Error messages are issued to the server log.
@param[in] name tablespace name
@param[in] id tablespace identifier
@param[in] flags tablespace flags
@param[in] purpose tablespace purpose
@param[in,out] crypt_data encryption information
@param[in] mode encryption mode
@return pointer to created tablespace, to be filled in with fil_space_t::add()
@retval NULL on failure (such as when the same tablespace exists) */
fil_space_t*
fil_space_create(
const char* name,
ulint id,
ulint flags,
fil_type_t purpose,
fil_space_crypt_t* crypt_data,
fil_encryption_t mode)
{
fil_space_t* space;
ut_ad(fil_system.is_initialised());
ut_ad(fsp_flags_is_valid(flags & ~FSP_FLAGS_MEM_MASK, id));
ut_ad(purpose == FIL_TYPE_LOG
|| srv_page_size == UNIV_PAGE_SIZE_ORIG || flags != 0);
DBUG_EXECUTE_IF("fil_space_create_failure", return(NULL););
mutex_enter(&fil_system.mutex);
space = fil_space_get_by_id(id);
if (space != NULL) {
ib::error() << "Trying to add tablespace '" << name
<< "' with id " << id
<< " to the tablespace memory cache, but tablespace '"
<< space->name << "' already exists in the cache!";
mutex_exit(&fil_system.mutex);
return(NULL);
}
space = static_cast<fil_space_t*>(ut_zalloc_nokey(sizeof(*space)));
space->id = id;
space->name = mem_strdup(name);
UT_LIST_INIT(space->chain, &fil_node_t::chain);
if ((purpose == FIL_TYPE_TABLESPACE || purpose == FIL_TYPE_IMPORT)
&& !recv_recovery_is_on()
&& id > fil_system.max_assigned_id) {
if (!fil_system.space_id_reuse_warned) {
fil_system.space_id_reuse_warned = true;
ib::warn() << "Allocated tablespace ID " << id
<< " for " << name << ", old maximum was "
<< fil_system.max_assigned_id;
}
fil_system.max_assigned_id = id;
}
space->purpose = purpose;
space->flags = flags;
space->magic_n = FIL_SPACE_MAGIC_N;
space->crypt_data = crypt_data;
DBUG_LOG("tablespace",
"Created metadata for " << id << " name " << name);
if (crypt_data) {
DBUG_LOG("crypt",
"Tablespace " << id << " name " << name
<< " encryption " << crypt_data->encryption
<< " key id " << crypt_data->key_id
<< ":" << fil_crypt_get_mode(crypt_data)
<< " " << fil_crypt_get_type(crypt_data));
}
rw_lock_create(fil_space_latch_key, &space->latch, SYNC_FSP);
if (space->purpose == FIL_TYPE_TEMPORARY) {
/* SysTablespace::open_or_create() would pass
size!=0 to fil_space_t::add(), so first_time_open
would not hold in fil_node_open_file(), and we
must assign this manually. We do not care about
the durability or atomicity of writes to the
temporary tablespace files. */
space->atomic_write_supported = true;
}
HASH_INSERT(fil_space_t, hash, fil_system.spaces, id, space);
UT_LIST_ADD_LAST(fil_system.space_list, space);
if (id < SRV_LOG_SPACE_FIRST_ID && id > fil_system.max_assigned_id) {
fil_system.max_assigned_id = id;
}
/* Inform key rotation that there could be something
to do */
if (purpose == FIL_TYPE_TABLESPACE
&& !srv_fil_crypt_rotate_key_age && fil_crypt_threads_event &&
(mode == FIL_ENCRYPTION_ON || mode == FIL_ENCRYPTION_OFF ||
srv_encrypt_tables)) {
/* Key rotation is not enabled, need to inform background
encryption threads. */
fil_system.rotation_list.push_back(*space);
space->is_in_rotation_list = true;
mutex_exit(&fil_system.mutex);
os_event_set(fil_crypt_threads_event);
} else {
mutex_exit(&fil_system.mutex);
}
return(space);
}
/*******************************************************************//**
Assigns a new space id for a new single-table tablespace. This works simply by
incrementing the global counter. If 4 billion id's is not enough, we may need
to recycle id's.
@return true if assigned, false if not */
bool
fil_assign_new_space_id(
/*====================*/
ulint* space_id) /*!< in/out: space id */
{
ulint id;
bool success;
mutex_enter(&fil_system.mutex);
id = *space_id;
if (id < fil_system.max_assigned_id) {
id = fil_system.max_assigned_id;
}
id++;
if (id > (SRV_LOG_SPACE_FIRST_ID / 2) && (id % 1000000UL == 0)) {
ib::warn() << "You are running out of new single-table"
" tablespace id's. Current counter is " << id
<< " and it must not exceed" << SRV_LOG_SPACE_FIRST_ID
<< "! To reset the counter to zero you have to dump"
" all your tables and recreate the whole InnoDB"
" installation.";
}
success = (id < SRV_LOG_SPACE_FIRST_ID);
if (success) {
*space_id = fil_system.max_assigned_id = id;
} else {
ib::warn() << "You have run out of single-table tablespace"
" id's! Current counter is " << id
<< ". To reset the counter to zero"
" you have to dump all your tables and"
" recreate the whole InnoDB installation.";
*space_id = ULINT_UNDEFINED;
}
mutex_exit(&fil_system.mutex);
return(success);
}
/** Trigger a call to fil_node_t::read_page0()
@param[in] id tablespace identifier
@return tablespace
@retval NULL if the tablespace does not exist or cannot be read */
fil_space_t* fil_system_t::read_page0(ulint id)
{
mutex_exit(&mutex);
ut_ad(id != 0);
/* It is possible that the tablespace is dropped while we are
not holding the mutex. */
fil_mutex_enter_and_prepare_for_io(id);
fil_space_t* space = fil_space_get_by_id(id);
if (space == NULL || UT_LIST_GET_LEN(space->chain) == 0) {
return(NULL);
}
/* The following code must change when InnoDB supports
multiple datafiles per tablespace. */
ut_a(1 == UT_LIST_GET_LEN(space->chain));
fil_node_t* node = UT_LIST_GET_FIRST(space->chain);
/* It must be a single-table tablespace and we have not opened
the file yet; the following calls will open it and update the
size fields */
if (!fil_node_prepare_for_io(node, space)) {
/* The single-table tablespace can't be opened,
because the ibd file is missing. */
return(NULL);
}
fil_node_complete_io(node, IORequestRead);
return space;
}
/*******************************************************************//**
Returns a pointer to the fil_space_t that is in the memory cache
associated with a space id. The caller must lock fil_system.mutex.
@return file_space_t pointer, NULL if space not found */
UNIV_INLINE
fil_space_t*
fil_space_get_space(
/*================*/
ulint id) /*!< in: space id */
{
fil_space_t* space = fil_space_get_by_id(id);
if (space == NULL || space->size != 0) {
return(space);
}
switch (space->purpose) {
case FIL_TYPE_LOG:
break;
case FIL_TYPE_TEMPORARY:
case FIL_TYPE_TABLESPACE:
case FIL_TYPE_IMPORT:
space = fil_system.read_page0(id);
}
return(space);
}
/** Set the recovered size of a tablespace in pages.
@param id tablespace ID
@param size recovered size in pages */
UNIV_INTERN
void
fil_space_set_recv_size(ulint id, ulint size)
{
mutex_enter(&fil_system.mutex);
ut_ad(size);
ut_ad(id < SRV_LOG_SPACE_FIRST_ID);
if (fil_space_t* space = fil_space_get_space(id)) {
space->recv_size = size;
}
mutex_exit(&fil_system.mutex);
}
/*******************************************************************//**
Returns the size of the space in pages. The tablespace must be cached in the
memory cache.
@return space size, 0 if space not found */
ulint
fil_space_get_size(
/*===============*/
ulint id) /*!< in: space id */
{
fil_space_t* space;
ulint size;
ut_ad(fil_system.is_initialised());
mutex_enter(&fil_system.mutex);
space = fil_space_get_space(id);
size = space ? space->size : 0;
mutex_exit(&fil_system.mutex);
return(size);
}
/*******************************************************************//**
Returns the flags of the space. The tablespace must be cached
in the memory cache.
@return flags, ULINT_UNDEFINED if space not found */
ulint
fil_space_get_flags(
/*================*/
ulint id) /*!< in: space id */
{
fil_space_t* space;
ulint flags;
ut_ad(fil_system.is_initialised());
mutex_enter(&fil_system.mutex);
space = fil_space_get_space(id);
if (space == NULL) {
mutex_exit(&fil_system.mutex);
return(ULINT_UNDEFINED);
}
flags = space->flags;
mutex_exit(&fil_system.mutex);
return(flags);
}
/** Open each file. Only invoked on fil_system.temp_space.
@return whether all files were opened */
bool fil_space_t::open()
{
ut_ad(fil_system.is_initialised());
mutex_enter(&fil_system.mutex);
ut_ad(this == fil_system.temp_space
|| srv_operation == SRV_OPERATION_BACKUP
|| srv_operation == SRV_OPERATION_RESTORE
|| srv_operation == SRV_OPERATION_RESTORE_DELTA);
for (fil_node_t* node = UT_LIST_GET_FIRST(chain);
node != NULL;
node = UT_LIST_GET_NEXT(chain, node)) {
if (!node->is_open() && !fil_node_open_file(node)) {
mutex_exit(&fil_system.mutex);
return false;
}
}
mutex_exit(&fil_system.mutex);
return true;
}
/** Close each file. Only invoked on fil_system.temp_space. */
void fil_space_t::close()
{
if (!fil_system.is_initialised()) {
return;
}
mutex_enter(&fil_system.mutex);
ut_ad(this == fil_system.temp_space
|| srv_operation == SRV_OPERATION_BACKUP
|| srv_operation == SRV_OPERATION_RESTORE
|| srv_operation == SRV_OPERATION_RESTORE_DELTA);
for (fil_node_t* node = UT_LIST_GET_FIRST(chain);
node != NULL;
node = UT_LIST_GET_NEXT(chain, node)) {
if (node->is_open()) {
node->close();
}
}
mutex_exit(&fil_system.mutex);
}
/** Returns the page size of the space and whether it is compressed or not.
The tablespace must be cached in the memory cache.
@param[in] id space id
@param[out] found true if tablespace was found
@return page size */
const page_size_t
fil_space_get_page_size(
ulint id,
bool* found)
{
const ulint flags = fil_space_get_flags(id);
if (flags == ULINT_UNDEFINED) {
*found = false;
return(univ_page_size);
}
*found = true;
return(page_size_t(flags));
}
void fil_system_t::create(ulint hash_size)
{
ut_ad(this == &fil_system);
ut_ad(!is_initialised());
ut_ad(!(srv_page_size % FSP_EXTENT_SIZE));
ut_ad(srv_page_size);
ut_ad(!spaces);
m_initialised = true;
compile_time_assert(!(UNIV_PAGE_SIZE_MAX % FSP_EXTENT_SIZE_MAX));
compile_time_assert(!(UNIV_PAGE_SIZE_MIN % FSP_EXTENT_SIZE_MIN));
ut_ad(hash_size > 0);
mutex_create(LATCH_ID_FIL_SYSTEM, &mutex);
spaces = hash_create(hash_size);
fil_space_crypt_init();
}
void fil_system_t::close()
{
ut_ad(this == &fil_system);
ut_a(!UT_LIST_GET_LEN(LRU));
ut_a(unflushed_spaces.empty());
ut_a(!UT_LIST_GET_LEN(space_list));
ut_ad(!sys_space);
ut_ad(!temp_space);
if (is_initialised()) {
m_initialised = false;
hash_table_free(spaces);
spaces = NULL;
mutex_free(&mutex);
fil_space_crypt_cleanup();
}
ut_ad(!spaces);
}
/*******************************************************************//**
Opens all log files and system tablespace data files. They stay open until the
database server shutdown. This should be called at a server startup after the
space objects for the log and the system tablespace have been created. The
purpose of this operation is to make sure we never run out of file descriptors
if we need to read from the insert buffer or to write to the log. */
void
fil_open_log_and_system_tablespace_files(void)
/*==========================================*/
{
fil_space_t* space;
mutex_enter(&fil_system.mutex);
for (space = UT_LIST_GET_FIRST(fil_system.space_list);
space != NULL;
space = UT_LIST_GET_NEXT(space_list, space)) {
fil_node_t* node;
if (fil_space_belongs_in_lru(space)) {
continue;
}
for (node = UT_LIST_GET_FIRST(space->chain);
node != NULL;
node = UT_LIST_GET_NEXT(chain, node)) {
if (!node->is_open()) {
if (!fil_node_open_file(node)) {
/* This func is called during server's
startup. If some file of log or system
tablespace is missing, the server
can't start successfully. So we should
assert for it. */
ut_a(0);
}
}
if (srv_max_n_open_files < 10 + fil_system.n_open) {
ib::warn() << "You must raise the value of"
" innodb_open_files in my.cnf!"
" Remember that InnoDB keeps all"
" log files and all system"
" tablespace files open"
" for the whole time mysqld is"
" running, and needs to open also"
" some .ibd files if the"
" file-per-table storage model is used."
" Current open files "
<< fil_system.n_open
<< ", max allowed open files "
<< srv_max_n_open_files
<< ".";
}
}
}
mutex_exit(&fil_system.mutex);
}
/*******************************************************************//**
Closes all open files. There must not be any pending i/o's or not flushed
modifications in the files. */
void
fil_close_all_files(void)
/*=====================*/
{
fil_space_t* space;
/* At shutdown, we should not have any files in this list. */
ut_ad(fil_system.is_initialised());
ut_ad(srv_fast_shutdown == 2
|| !srv_was_started
|| UT_LIST_GET_LEN(fil_system.named_spaces) == 0);
mutex_enter(&fil_system.mutex);
for (space = UT_LIST_GET_FIRST(fil_system.space_list);
space != NULL; ) {
fil_node_t* node;
fil_space_t* prev_space = space;
for (node = UT_LIST_GET_FIRST(space->chain);
node != NULL;
node = UT_LIST_GET_NEXT(chain, node)) {
if (node->is_open()) {
node->close();
}
}
space = UT_LIST_GET_NEXT(space_list, space);
fil_space_detach(prev_space);
fil_space_free_low(prev_space);
}
mutex_exit(&fil_system.mutex);
ut_ad(srv_fast_shutdown == 2
|| !srv_was_started
|| UT_LIST_GET_LEN(fil_system.named_spaces) == 0);
}
/*******************************************************************//**
Closes the redo log files. There must not be any pending i/o's or not
flushed modifications in the files. */
void
fil_close_log_files(
/*================*/
bool free) /*!< in: whether to free the memory object */
{
fil_space_t* space;
mutex_enter(&fil_system.mutex);
space = UT_LIST_GET_FIRST(fil_system.space_list);
while (space != NULL) {
fil_node_t* node;
fil_space_t* prev_space = space;
if (space->purpose != FIL_TYPE_LOG) {
space = UT_LIST_GET_NEXT(space_list, space);
continue;
}
/* Log files are not in the fil_system.named_spaces list. */
ut_ad(space->max_lsn == 0);
for (node = UT_LIST_GET_FIRST(space->chain);
node != NULL;
node = UT_LIST_GET_NEXT(chain, node)) {
if (node->is_open()) {
node->close();
}
}
space = UT_LIST_GET_NEXT(space_list, space);
if (free) {
fil_space_detach(prev_space);
fil_space_free_low(prev_space);
}
}
mutex_exit(&fil_system.mutex);
if (free) {
log_sys.log.close();
}
}
/*******************************************************************//**
Sets the max tablespace id counter if the given number is bigger than the
previous value. */
void
fil_set_max_space_id_if_bigger(
/*===========================*/
ulint max_id) /*!< in: maximum known id */
{
if (max_id >= SRV_LOG_SPACE_FIRST_ID) {
ib::fatal() << "Max tablespace id is too high, " << max_id;
}
mutex_enter(&fil_system.mutex);
if (fil_system.max_assigned_id < max_id) {
fil_system.max_assigned_id = max_id;
}
mutex_exit(&fil_system.mutex);
}
/** Write the flushed LSN to the page header of the first page in the
system tablespace.
@param[in] lsn flushed LSN
@return DB_SUCCESS or error number */
dberr_t
fil_write_flushed_lsn(
lsn_t lsn)
{
byte* buf1;
byte* buf;
dberr_t err = DB_TABLESPACE_NOT_FOUND;
buf1 = static_cast<byte*>(ut_malloc_nokey(2U << srv_page_size_shift));
buf = static_cast<byte*>(ut_align(buf1, srv_page_size));
const page_id_t page_id(TRX_SYS_SPACE, 0);
err = fil_read(page_id, univ_page_size, 0, srv_page_size,
buf);
if (err == DB_SUCCESS) {
mach_write_to_8(buf + FIL_PAGE_FILE_FLUSH_LSN_OR_KEY_VERSION, lsn);
err = fil_write(page_id, univ_page_size, 0,
srv_page_size, buf);
fil_flush_file_spaces(FIL_TYPE_TABLESPACE);
}
ut_free(buf1);
return(err);
}
/** Acquire a tablespace when it could be dropped concurrently.
Used by background threads that do not necessarily hold proper locks
for concurrency control.
@param[in] id tablespace ID
@param[in] silent whether to silently ignore missing tablespaces
@return the tablespace
@retval NULL if missing or being deleted or truncated */
UNIV_INTERN
fil_space_t*
fil_space_acquire_low(ulint id, bool silent)
{
fil_space_t* space;
mutex_enter(&fil_system.mutex);
space = fil_space_get_by_id(id);
if (space == NULL) {
if (!silent) {
ib::warn() << "Trying to access missing"
" tablespace " << id;
}
} else if (space->is_stopping()) {
space = NULL;
} else {
space->acquire();
}
mutex_exit(&fil_system.mutex);
return(space);
}
/** Acquire a tablespace for reading or writing a block,
when it could be dropped concurrently.
@param[in] id tablespace ID
@return the tablespace
@retval NULL if missing */
fil_space_t*
fil_space_acquire_for_io(ulint id)
{
mutex_enter(&fil_system.mutex);
fil_space_t* space = fil_space_get_by_id(id);
if (space) {
space->acquire_for_io();
}
mutex_exit(&fil_system.mutex);
return(space);
}
/********************************************************//**
Creates the database directory for a table if it does not exist yet. */
void
fil_create_directory_for_tablename(
/*===============================*/
const char* name) /*!< in: name in the standard
'databasename/tablename' format */
{
const char* namend;
char* path;
ulint len;
len = strlen(fil_path_to_mysql_datadir);
namend = strchr(name, '/');
ut_a(namend);
path = static_cast<char*>(
ut_malloc_nokey(len + ulint(namend - name) + 2));
memcpy(path, fil_path_to_mysql_datadir, len);
path[len] = '/';
memcpy(path + len + 1, name, ulint(namend - name));
path[len + ulint(namend - name) + 1] = 0;
os_normalize_path(path);
bool success = os_file_create_directory(path, false);
ut_a(success);
ut_free(path);
}
/** Write a log record about an operation on a tablespace file.
@param[in] type MLOG_FILE_NAME or MLOG_FILE_DELETE
or MLOG_FILE_CREATE2 or MLOG_FILE_RENAME2
@param[in] space_id tablespace identifier
@param[in] first_page_no first page number in the file
@param[in] path file path
@param[in] new_path if type is MLOG_FILE_RENAME2, the new name
@param[in] flags if type is MLOG_FILE_CREATE2, the space flags
@param[in,out] mtr mini-transaction */
static
void
fil_op_write_log(
mlog_id_t type,
ulint space_id,
ulint first_page_no,
const char* path,
const char* new_path,
ulint flags,
mtr_t* mtr)
{
byte* log_ptr;
ulint len;
ut_ad(first_page_no == 0 || type == MLOG_FILE_CREATE2);
ut_ad(fsp_flags_is_valid(flags, space_id));
/* fil_name_parse() requires that there be at least one path
separator and that the file path end with ".ibd". */
ut_ad(strchr(path, OS_PATH_SEPARATOR) != NULL);
ut_ad(first_page_no /* trimming an undo tablespace */
|| !strcmp(&path[strlen(path) - strlen(DOT_IBD)], DOT_IBD));
log_ptr = mlog_open(mtr, 11 + 4 + 2 + 1);
if (log_ptr == NULL) {
/* Logging in mtr is switched off during crash recovery:
in that case mlog_open returns NULL */
return;
}
log_ptr = mlog_write_initial_log_record_low(
type, space_id, first_page_no, log_ptr, mtr);
if (type == MLOG_FILE_CREATE2) {
mach_write_to_4(log_ptr, flags);
log_ptr += 4;
}
/* Let us store the strings as null-terminated for easier readability
and handling */
len = strlen(path) + 1;
mach_write_to_2(log_ptr, len);
log_ptr += 2;
mlog_close(mtr, log_ptr);
mlog_catenate_string(
mtr, reinterpret_cast<const byte*>(path), len);
switch (type) {
case MLOG_FILE_RENAME2:
ut_ad(strchr(new_path, OS_PATH_SEPARATOR) != NULL);
len = strlen(new_path) + 1;
log_ptr = mlog_open(mtr, 2 + len);
ut_a(log_ptr);
mach_write_to_2(log_ptr, len);
log_ptr += 2;
mlog_close(mtr, log_ptr);
mlog_catenate_string(
mtr, reinterpret_cast<const byte*>(new_path), len);
break;
case MLOG_FILE_NAME:
case MLOG_FILE_DELETE:
case MLOG_FILE_CREATE2:
break;
default:
ut_ad(0);
}
}
/** Write redo log for renaming a file.
@param[in] space_id tablespace id
@param[in] first_page_no first page number in the file
@param[in] old_name tablespace file name
@param[in] new_name tablespace file name after renaming
@param[in,out] mtr mini-transaction */
static
void
fil_name_write_rename_low(
ulint space_id,
ulint first_page_no,
const char* old_name,
const char* new_name,
mtr_t* mtr)
{
ut_ad(!is_predefined_tablespace(space_id));
fil_op_write_log(
MLOG_FILE_RENAME2,
space_id, first_page_no, old_name, new_name, 0, mtr);
}
/** Write redo log for renaming a file.
@param[in] space_id tablespace id
@param[in] old_name tablespace file name
@param[in] new_name tablespace file name after renaming */
static void
fil_name_write_rename(
ulint space_id,
const char* old_name,
const char* new_name)
{
mtr_t mtr;
mtr.start();
fil_name_write_rename_low(space_id, 0, old_name, new_name, &mtr);
mtr.commit();
log_write_up_to(mtr.commit_lsn(), true);
}
/** Write MLOG_FILE_NAME for a file.
@param[in] space_id tablespace id
@param[in] first_page_no first page number in the file
@param[in] name tablespace file name
@param[in,out] mtr mini-transaction */
static
void
fil_name_write(
ulint space_id,
ulint first_page_no,
const char* name,
mtr_t* mtr)
{
fil_op_write_log(
MLOG_FILE_NAME, space_id, first_page_no, name, NULL, 0, mtr);
}
/** Write MLOG_FILE_NAME for a file.
@param[in] space tablespace
@param[in] first_page_no first page number in the file
@param[in] file tablespace file
@param[in,out] mtr mini-transaction */
static
void
fil_name_write(
const fil_space_t* space,
ulint first_page_no,
const fil_node_t* file,
mtr_t* mtr)
{
fil_name_write(space->id, first_page_no, file->name, mtr);
}
/** Replay a file rename operation if possible.
@param[in] space_id tablespace identifier
@param[in] first_page_no first page number in the file
@param[in] name old file name
@param[in] new_name new file name
@return whether the operation was successfully applied
(the name did not exist, or new_name did not exist and
name was successfully renamed to new_name) */
bool
fil_op_replay_rename(
ulint space_id,
ulint first_page_no,
const char* name,
const char* new_name)
{
ut_ad(first_page_no == 0);
/* In order to replay the rename, the following must hold:
* The new name is not already used.
* A tablespace exists with the old name.
* The space ID for that tablepace matches this log entry.
This will prevent unintended renames during recovery. */
fil_space_t* space = fil_space_get(space_id);
if (space == NULL) {
return(true);
}
const bool name_match
= strcmp(name, UT_LIST_GET_FIRST(space->chain)->name) == 0;
if (!name_match) {
return(true);
}
/* Create the database directory for the new name, if
it does not exist yet */
const char* namend = strrchr(new_name, OS_PATH_SEPARATOR);
ut_a(namend != NULL);
char* dir = static_cast<char*>(
ut_malloc_nokey(ulint(namend - new_name) + 1));
memcpy(dir, new_name, ulint(namend - new_name));
dir[namend - new_name] = '\0';
bool success = os_file_create_directory(dir, false);
ut_a(success);
ulint dirlen = 0;
if (const char* dirend = strrchr(dir, OS_PATH_SEPARATOR)) {
dirlen = ulint(dirend - dir) + 1;
}
ut_free(dir);
/* New path must not exist. */
dberr_t err = fil_rename_tablespace_check(
name, new_name, false);
if (err != DB_SUCCESS) {
ib::error() << " Cannot replay file rename."
" Remove either file and try again.";
return(false);
}
char* new_table = mem_strdupl(
new_name + dirlen,
strlen(new_name + dirlen)
- 4 /* remove ".ibd" */);
ut_ad(new_table[ulint(namend - new_name) - dirlen]
== OS_PATH_SEPARATOR);
#if OS_PATH_SEPARATOR != '/'
new_table[namend - new_name - dirlen] = '/';
#endif
if (!fil_rename_tablespace(
space_id, name, new_table, new_name)) {
ut_error;
}
ut_free(new_table);
return(true);
}
/** File operations for tablespace */
enum fil_operation_t {
FIL_OPERATION_DELETE, /*!< delete a single-table tablespace */
FIL_OPERATION_CLOSE, /*!< close a single-table tablespace */
FIL_OPERATION_TRUNCATE /*!< truncate an undo tablespace */
};
/** Check for pending operations.
@param[in] space tablespace
@param[in] count number of attempts so far
@return 0 if no operations else count + 1. */
static
ulint
fil_check_pending_ops(const fil_space_t* space, ulint count)
{
ut_ad(mutex_own(&fil_system.mutex));
if (space == NULL) {
return 0;
}
if (ulint n_pending_ops = my_atomic_loadlint(&space->n_pending_ops)) {
if (count > 5000) {
ib::warn() << "Trying to close/delete/truncate"
" tablespace '" << space->name
<< "' but there are " << n_pending_ops
<< " pending operations on it.";
}
return(count + 1);
}
return(0);
}
/*******************************************************************//**
Check for pending IO.
@return 0 if no pending else count + 1. */
static
ulint
fil_check_pending_io(
/*=================*/
fil_operation_t operation, /*!< in: File operation */
fil_space_t* space, /*!< in/out: Tablespace to check */
fil_node_t** node, /*!< out: Node in space list */
ulint count) /*!< in: number of attempts so far */
{
ut_ad(mutex_own(&fil_system.mutex));
ut_ad(!space->referenced());
switch (operation) {
case FIL_OPERATION_DELETE:
case FIL_OPERATION_CLOSE:
break;
case FIL_OPERATION_TRUNCATE:
space->is_being_truncated = true;
break;
}
/* The following code must change when InnoDB supports
multiple datafiles per tablespace. */
ut_a(UT_LIST_GET_LEN(space->chain) == 1);
*node = UT_LIST_GET_FIRST(space->chain);
if (space->n_pending_flushes > 0 || (*node)->n_pending > 0) {
ut_a(!(*node)->being_extended);
if (count > 1000) {
ib::warn() << "Trying to delete/close/truncate"
" tablespace '" << space->name
<< "' but there are "
<< space->n_pending_flushes
<< " flushes and " << (*node)->n_pending
<< " pending i/o's on it.";
}
return(count + 1);
}
return(0);
}
/*******************************************************************//**
Check pending operations on a tablespace.
@return DB_SUCCESS or error failure. */
static
dberr_t
fil_check_pending_operations(
/*=========================*/
ulint id, /*!< in: space id */
fil_operation_t operation, /*!< in: File operation */
fil_space_t** space, /*!< out: tablespace instance
in memory */
char** path) /*!< out/own: tablespace path */
{
ulint count = 0;
ut_a(!is_system_tablespace(id));
ut_ad(space);
*space = 0;
mutex_enter(&fil_system.mutex);
fil_space_t* sp = fil_space_get_by_id(id);
if (sp) {
sp->stop_new_ops = true;
if (sp->crypt_data) {
sp->acquire();
mutex_exit(&fil_system.mutex);
fil_space_crypt_close_tablespace(sp);
mutex_enter(&fil_system.mutex);
sp->release();
}
}
/* Check for pending operations. */
do {
sp = fil_space_get_by_id(id);
count = fil_check_pending_ops(sp, count);
mutex_exit(&fil_system.mutex);
if (count > 0) {
os_thread_sleep(20000);
}
mutex_enter(&fil_system.mutex);
} while (count > 0);
/* Check for pending IO. */
for (;;) {
sp = fil_space_get_by_id(id);
if (sp == NULL) {
mutex_exit(&fil_system.mutex);
return(DB_TABLESPACE_NOT_FOUND);
}
fil_node_t* node;
count = fil_check_pending_io(operation, sp, &node, count);
if (count == 0 && path) {
*path = mem_strdup(node->name);
}
mutex_exit(&fil_system.mutex);
if (count == 0) {
break;
}
os_thread_sleep(20000);
mutex_enter(&fil_system.mutex);
}
ut_ad(sp);
*space = sp;
return(DB_SUCCESS);
}
/*******************************************************************//**
Closes a single-table tablespace. The tablespace must be cached in the
memory cache. Free all pages used by the tablespace.
@return DB_SUCCESS or error */
dberr_t
fil_close_tablespace(
/*=================*/
trx_t* trx, /*!< in/out: Transaction covering the close */
ulint id) /*!< in: space id */
{
char* path = 0;
fil_space_t* space = 0;
dberr_t err;
ut_a(!is_system_tablespace(id));
err = fil_check_pending_operations(id, FIL_OPERATION_CLOSE,
&space, &path);
if (err != DB_SUCCESS) {
return(err);
}
ut_a(space);
ut_a(path != 0);
rw_lock_x_lock(&space->latch);
/* Invalidate in the buffer pool all pages belonging to the
tablespace. Since we have set space->stop_new_ops = true, readahead
or ibuf merge can no longer read more pages of this tablespace to the
buffer pool. Thus we can clean the tablespace out of the buffer pool
completely and permanently. The flag stop_new_ops also prevents
fil_flush() from being applied to this tablespace. */
{
FlushObserver observer(space, trx, NULL);
buf_LRU_flush_or_remove_pages(id, &observer);
}
/* If the free is successful, the X lock will be released before
the space memory data structure is freed. */
if (!fil_space_free(id, true)) {
rw_lock_x_unlock(&space->latch);
err = DB_TABLESPACE_NOT_FOUND;
} else {
err = DB_SUCCESS;
}
/* If it is a delete then also delete any generated files, otherwise
when we drop the database the remove directory will fail. */
char* cfg_name = fil_make_filepath(path, NULL, CFG, false);
if (cfg_name != NULL) {
os_file_delete_if_exists(innodb_data_file_key, cfg_name, NULL);
ut_free(cfg_name);
}
ut_free(path);
return(err);
}
/** Determine whether a table can be accessed in operations that are
not (necessarily) protected by meta-data locks.
(Rollback would generally be protected, but rollback of
FOREIGN KEY CASCADE/SET NULL is not protected by meta-data locks
but only by InnoDB table locks, which may be broken by
lock_remove_all_on_table().)
@param[in] table persistent table
checked @return whether the table is accessible */
bool fil_table_accessible(const dict_table_t* table)
{
if (UNIV_UNLIKELY(!table->is_readable() || table->corrupted)) {
return(false);
}
mutex_enter(&fil_system.mutex);
bool accessible = table->space && !table->space->is_stopping();
mutex_exit(&fil_system.mutex);
ut_ad(accessible || dict_table_is_file_per_table(table));
return accessible;
}
/** Delete a tablespace and associated .ibd file.
@param[in] id tablespace identifier
@param[in] if_exists whether to ignore missing tablespace
@return DB_SUCCESS or error */
dberr_t fil_delete_tablespace(ulint id, bool if_exists)
{
char* path = 0;
fil_space_t* space = 0;
ut_a(!is_system_tablespace(id));
dberr_t err = fil_check_pending_operations(
id, FIL_OPERATION_DELETE, &space, &path);
if (err != DB_SUCCESS) {
if (!if_exists) {
ib::error() << "Cannot delete tablespace " << id
<< " because it is not found"
" in the tablespace memory cache.";
}
return(err);
}
ut_a(space);
ut_a(path != 0);
/* IMPORTANT: Because we have set space::stop_new_ops there
can't be any new ibuf merges, reads or flushes. We are here
because node::n_pending was zero above. However, it is still
possible to have pending read and write requests:
A read request can happen because the reader thread has
gone through the ::stop_new_ops check in buf_page_init_for_read()
before the flag was set and has not yet incremented ::n_pending
when we checked it above.
A write request can be issued any time because we don't check
the ::stop_new_ops flag when queueing a block for write.
We deal with pending write requests in the following function
where we'd minimally evict all dirty pages belonging to this
space from the flush_list. Note that if a block is IO-fixed
we'll wait for IO to complete.
To deal with potential read requests, we will check the
::stop_new_ops flag in fil_io(). */
buf_LRU_flush_or_remove_pages(id, NULL);
/* If it is a delete then also delete any generated files, otherwise
when we drop the database the remove directory will fail. */
{
/* Before deleting the file, write a log record about
it, so that InnoDB crash recovery will expect the file
to be gone. */
mtr_t mtr;
mtr_start(&mtr);
fil_op_write_log(MLOG_FILE_DELETE, id, 0, path, NULL, 0, &mtr);
mtr_commit(&mtr);
/* Even if we got killed shortly after deleting the
tablespace file, the record must have already been
written to the redo log. */
log_write_up_to(mtr.commit_lsn(), true);
char* cfg_name = fil_make_filepath(path, NULL, CFG, false);
if (cfg_name != NULL) {
os_file_delete_if_exists(innodb_data_file_key, cfg_name, NULL);
ut_free(cfg_name);
}
}
/* Delete the link file pointing to the ibd file we are deleting. */
if (FSP_FLAGS_HAS_DATA_DIR(space->flags)) {
RemoteDatafile::delete_link_file(space->name);
}
mutex_enter(&fil_system.mutex);
/* Double check the sanity of pending ops after reacquiring
the fil_system::mutex. */
if (const fil_space_t* s = fil_space_get_by_id(id)) {
ut_a(s == space);
ut_a(!space->referenced());
ut_a(UT_LIST_GET_LEN(space->chain) == 1);
fil_node_t* node = UT_LIST_GET_FIRST(space->chain);
ut_a(node->n_pending == 0);
fil_space_detach(space);
mutex_exit(&fil_system.mutex);
log_mutex_enter();
if (space->max_lsn != 0) {
ut_d(space->max_lsn = 0);
UT_LIST_REMOVE(fil_system.named_spaces, space);
}
log_mutex_exit();
fil_space_free_low(space);
if (!os_file_delete(innodb_data_file_key, path)
&& !os_file_delete_if_exists(
innodb_data_file_key, path, NULL)) {
/* Note: This is because we have removed the
tablespace instance from the cache. */
err = DB_IO_ERROR;
}
} else {
mutex_exit(&fil_system.mutex);
err = DB_TABLESPACE_NOT_FOUND;
}
ut_free(path);
return(err);
}
/** Prepare to truncate an undo tablespace.
@param[in] space_id undo tablespace id
@return the tablespace
@retval NULL if tablespace not found */
fil_space_t* fil_truncate_prepare(ulint space_id)
{
/* Stop all I/O on the tablespace and ensure that related
pages are flushed to disk. */
fil_space_t* space;
if (fil_check_pending_operations(space_id, FIL_OPERATION_TRUNCATE,
&space, NULL) != DB_SUCCESS) {
return NULL;
}
ut_ad(space != NULL);
return space;
}
/** Write log about an undo tablespace truncate operation. */
void fil_truncate_log(fil_space_t* space, ulint size, mtr_t* mtr)
{
/* Write a MLOG_FILE_CREATE2 record with the new size, so that
recovery and backup will ignore any preceding redo log records
for writing pages that are after the new end of the tablespace. */
ut_ad(UT_LIST_GET_LEN(space->chain) == 1);
const fil_node_t* file = UT_LIST_GET_FIRST(space->chain);
fil_op_write_log(MLOG_FILE_CREATE2, space->id, size, file->name,
NULL, space->flags & ~FSP_FLAGS_MEM_MASK, mtr);
}
/*******************************************************************//**
Allocates and builds a file name from a path, a table or tablespace name
and a suffix. The string must be freed by caller with ut_free().
@param[in] path NULL or the direcory path or the full path and filename.
@param[in] name NULL if path is full, or Table/Tablespace name
@param[in] suffix NULL or the file extention to use.
@param[in] trim_name true if the last name on the path should be trimmed.
@return own: file name */
char*
fil_make_filepath(
const char* path,
const char* name,
ib_extention ext,
bool trim_name)
{
/* The path may contain the basename of the file, if so we do not
need the name. If the path is NULL, we can use the default path,
but there needs to be a name. */
ut_ad(path != NULL || name != NULL);
/* If we are going to strip a name off the path, there better be a
path and a new name to put back on. */
ut_ad(!trim_name || (path != NULL && name != NULL));
if (path == NULL) {
path = fil_path_to_mysql_datadir;
}
ulint len = 0; /* current length */
ulint path_len = strlen(path);
ulint name_len = (name ? strlen(name) : 0);
const char* suffix = dot_ext[ext];
ulint suffix_len = strlen(suffix);
ulint full_len = path_len + 1 + name_len + suffix_len + 1;
char* full_name = static_cast<char*>(ut_malloc_nokey(full_len));
if (full_name == NULL) {
return NULL;
}
/* If the name is a relative path, do not prepend "./". */
if (path[0] == '.'
&& (path[1] == '\0' || path[1] == OS_PATH_SEPARATOR)
&& name != NULL && name[0] == '.') {
path = NULL;
path_len = 0;
}
if (path != NULL) {
memcpy(full_name, path, path_len);
len = path_len;
full_name[len] = '\0';
os_normalize_path(full_name);
}
if (trim_name) {
/* Find the offset of the last DIR separator and set it to
null in order to strip off the old basename from this path. */
char* last_dir_sep = strrchr(full_name, OS_PATH_SEPARATOR);
if (last_dir_sep) {
last_dir_sep[0] = '\0';
len = strlen(full_name);
}
}
if (name != NULL) {
if (len && full_name[len - 1] != OS_PATH_SEPARATOR) {
/* Add a DIR separator */
full_name[len] = OS_PATH_SEPARATOR;
full_name[++len] = '\0';
}
char* ptr = &full_name[len];
memcpy(ptr, name, name_len);
len += name_len;
full_name[len] = '\0';
os_normalize_path(ptr);
}
/* Make sure that the specified suffix is at the end of the filepath
string provided. This assumes that the suffix starts with '.'.
If the first char of the suffix is found in the filepath at the same
length as the suffix from the end, then we will assume that there is
a previous suffix that needs to be replaced. */
if (suffix != NULL) {
/* Need room for the trailing null byte. */
ut_ad(len < full_len);
if ((len > suffix_len)
&& (full_name[len - suffix_len] == suffix[0])) {
/* Another suffix exists, make it the one requested. */
memcpy(&full_name[len - suffix_len], suffix, suffix_len);
} else {
/* No previous suffix, add it. */
ut_ad(len + suffix_len < full_len);
memcpy(&full_name[len], suffix, suffix_len);
full_name[len + suffix_len] = '\0';
}
}
return(full_name);
}
/** Test if a tablespace file can be renamed to a new filepath by checking
if that the old filepath exists and the new filepath does not exist.
@param[in] old_path old filepath
@param[in] new_path new filepath
@param[in] is_discarded whether the tablespace is discarded
@param[in] replace_new whether to ignore the existence of new_path
@return innodb error code */
static dberr_t
fil_rename_tablespace_check(
const char* old_path,
const char* new_path,
bool is_discarded,
bool replace_new)
{
bool exists = false;
os_file_type_t ftype;
if (!is_discarded
&& os_file_status(old_path, &exists, &ftype)
&& !exists) {
ib::error() << "Cannot rename '" << old_path
<< "' to '" << new_path
<< "' because the source file"
<< " does not exist.";
return(DB_TABLESPACE_NOT_FOUND);
}
exists = false;
if (os_file_status(new_path, &exists, &ftype) && !exists) {
return DB_SUCCESS;
}
if (!replace_new) {
ib::error() << "Cannot rename '" << old_path
<< "' to '" << new_path
<< "' because the target file exists."
" Remove the target file and try again.";
return(DB_TABLESPACE_EXISTS);
}
/* This must be during the ROLLBACK of TRUNCATE TABLE.
Because InnoDB only allows at most one data dictionary
transaction at a time, and because this incomplete TRUNCATE
would have created a new tablespace file, we must remove
a possibly existing tablespace that is associated with the
new tablespace file. */
retry:
mutex_enter(&fil_system.mutex);
for (fil_space_t* space = UT_LIST_GET_FIRST(fil_system.space_list);
space; space = UT_LIST_GET_NEXT(space_list, space)) {
ulint id = space->id;
if (id && id < SRV_LOG_SPACE_FIRST_ID
&& space->purpose == FIL_TYPE_TABLESPACE
&& !strcmp(new_path,
UT_LIST_GET_FIRST(space->chain)->name)) {
ib::info() << "TRUNCATE rollback: " << id
<< "," << new_path;
mutex_exit(&fil_system.mutex);
dberr_t err = fil_delete_tablespace(id);
if (err != DB_SUCCESS) {
return err;
}
goto retry;
}
}
mutex_exit(&fil_system.mutex);
fil_delete_file(new_path);
return(DB_SUCCESS);
}
dberr_t fil_space_t::rename(const char* name, const char* path, bool log,
bool replace)
{
ut_ad(UT_LIST_GET_LEN(chain) == 1);
ut_ad(!is_system_tablespace(id));
if (log) {
dberr_t err = fil_rename_tablespace_check(
chain.start->name, path, false, replace);
if (err != DB_SUCCESS) {
return(err);
}
fil_name_write_rename(id, chain.start->name, path);
}
return fil_rename_tablespace(id, chain.start->name, name, path)
? DB_SUCCESS : DB_ERROR;
}
/** Rename a single-table tablespace.
The tablespace must exist in the memory cache.
@param[in] id tablespace identifier
@param[in] old_path old file name
@param[in] new_name new table name in the
databasename/tablename format
@param[in] new_path_in new file name,
or NULL if it is located in the normal data directory
@return true if success */
static bool
fil_rename_tablespace(
ulint id,
const char* old_path,
const char* new_name,
const char* new_path_in)
{
fil_space_t* space;
fil_node_t* node;
ut_a(id != 0);
ut_ad(strchr(new_name, '/') != NULL);
mutex_enter(&fil_system.mutex);
space = fil_space_get_by_id(id);
if (space == NULL) {
ib::error() << "Cannot find space id " << id
<< " in the tablespace memory cache, though the file '"
<< old_path
<< "' in a rename operation should have that id.";
mutex_exit(&fil_system.mutex);
return(false);
}
/* The following code must change when InnoDB supports
multiple datafiles per tablespace. */
ut_a(UT_LIST_GET_LEN(space->chain) == 1);
node = UT_LIST_GET_FIRST(space->chain);
space->n_pending_ops++;
mutex_exit(&fil_system.mutex);
char* new_file_name = new_path_in == NULL
? fil_make_filepath(NULL, new_name, IBD, false)
: mem_strdup(new_path_in);
char* old_file_name = node->name;
char* new_space_name = mem_strdup(new_name);
char* old_space_name = space->name;
ut_ad(strchr(old_file_name, OS_PATH_SEPARATOR) != NULL);
ut_ad(strchr(new_file_name, OS_PATH_SEPARATOR) != NULL);
if (!recv_recovery_is_on()) {
fil_name_write_rename(id, old_file_name, new_file_name);
log_mutex_enter();
}
/* log_sys.mutex is above fil_system.mutex in the latching order */
ut_ad(log_mutex_own());
mutex_enter(&fil_system.mutex);
ut_ad(space->n_pending_ops);
space->n_pending_ops--;
ut_ad(space->name == old_space_name);
ut_ad(node->name == old_file_name);
bool success;
DBUG_EXECUTE_IF("fil_rename_tablespace_failure_2",
goto skip_second_rename; );
success = os_file_rename(innodb_data_file_key,
old_file_name,
new_file_name);
DBUG_EXECUTE_IF("fil_rename_tablespace_failure_2",
skip_second_rename:
success = false; );
ut_ad(node->name == old_file_name);
if (success) {
node->name = new_file_name;
}
if (!recv_recovery_is_on()) {
log_mutex_exit();
}
ut_ad(space->name == old_space_name);
if (success) {
space->name = new_space_name;
} else {
/* Because nothing was renamed, we must free the new
names, not the old ones. */
old_file_name = new_file_name;
old_space_name = new_space_name;
}
mutex_exit(&fil_system.mutex);
ut_free(old_file_name);
ut_free(old_space_name);
return(success);
}
/** Create a tablespace file.
@param[in] space_id Tablespace ID
@param[in] name Tablespace name in dbname/tablename format.
@param[in] path Path and filename of the datafile to create.
@param[in] flags Tablespace flags
@param[in] size Initial size of the tablespace file in pages,
must be >= FIL_IBD_FILE_INITIAL_SIZE
@param[in] mode MariaDB encryption mode
@param[in] key_id MariaDB encryption key_id
@param[out] err DB_SUCCESS or error code
@return the created tablespace
@retval NULL on error */
fil_space_t*
fil_ibd_create(
ulint space_id,
const char* name,
const char* path,
ulint flags,
ulint size,
fil_encryption_t mode,
uint32_t key_id,
dberr_t* err)
{
pfs_os_file_t file;
byte* buf2;
byte* page;
bool success;
bool has_data_dir = FSP_FLAGS_HAS_DATA_DIR(flags) != 0;
fil_space_t* space = NULL;
fil_space_crypt_t *crypt_data = NULL;
ut_ad(!is_system_tablespace(space_id));
ut_ad(!srv_read_only_mode);
ut_a(space_id < SRV_LOG_SPACE_FIRST_ID);
ut_a(size >= FIL_IBD_FILE_INITIAL_SIZE);
ut_a(fsp_flags_is_valid(flags & ~FSP_FLAGS_MEM_MASK, space_id));
/* Create the subdirectories in the path, if they are
not there already. */
*err = os_file_create_subdirs_if_needed(path);
if (*err != DB_SUCCESS) {
return NULL;
}
file = os_file_create(
innodb_data_file_key, path,
OS_FILE_CREATE | OS_FILE_ON_ERROR_NO_EXIT,
OS_FILE_NORMAL,
OS_DATA_FILE,
srv_read_only_mode,
&success);
if (!success) {
/* The following call will print an error message */
switch (os_file_get_last_error(true)) {
case OS_FILE_ALREADY_EXISTS:
ib::info() << "The file '" << path << "'"
" already exists though the"
" corresponding table did not exist"
" in the InnoDB data dictionary."
" You can resolve the problem by removing"
" the file.";
*err = DB_TABLESPACE_EXISTS;
break;
case OS_FILE_DISK_FULL:
*err = DB_OUT_OF_FILE_SPACE;
break;
default:
*err = DB_ERROR;
}
ib::error() << "Cannot create file '" << path << "'";
return NULL;
}
const bool is_compressed = FSP_FLAGS_HAS_PAGE_COMPRESSION(flags);
#ifdef _WIN32
if (is_compressed) {
os_file_set_sparse_win32(file);
}
#endif
if (!os_file_set_size(
path, file,
os_offset_t(size) << srv_page_size_shift, is_compressed)) {
*err = DB_OUT_OF_FILE_SPACE;
err_exit:
os_file_close(file);
os_file_delete(innodb_data_file_key, path);
return NULL;
}
bool punch_hole = os_is_sparse_file_supported(file);
ulint block_size = os_file_get_block_size(file, path);
/* We have to write the space id to the file immediately and flush the
file to disk. This is because in crash recovery we must be aware what
tablespaces exist and what are their space id's, so that we can apply
the log records to the right file. It may take quite a while until
buffer pool flush algorithms write anything to the file and flush it to
disk. If we would not write here anything, the file would be filled
with zeros from the call of os_file_set_size(), until a buffer pool
flush would write to it. */
buf2 = static_cast<byte*>(ut_malloc_nokey(3U << srv_page_size_shift));
/* Align the memory for file i/o if we might have O_DIRECT set */
page = static_cast<byte*>(ut_align(buf2, srv_page_size));
memset(page, '\0', srv_page_size);
flags |= FSP_FLAGS_PAGE_SSIZE();
fsp_header_init_fields(page, space_id, flags);
mach_write_to_4(page + FIL_PAGE_ARCH_LOG_NO_OR_SPACE_ID, space_id);
/* Create crypt data if the tablespace is either encrypted or user has
requested it to remain unencrypted. */
if (mode == FIL_ENCRYPTION_ON || mode == FIL_ENCRYPTION_OFF ||
srv_encrypt_tables) {
crypt_data = fil_space_create_crypt_data(mode, key_id);
}
if (crypt_data) {
/* Write crypt data information in page0 while creating
ibd file. */
crypt_data->fill_page0(flags, page);
}
const page_size_t page_size(flags);
IORequest request(IORequest::WRITE);
if (!page_size.is_compressed()) {
buf_flush_init_for_writing(NULL, page, NULL, 0);
*err = os_file_write(
request, path, file, page, 0, page_size.physical());
} else {
page_zip_des_t page_zip;
page_zip_set_size(&page_zip, page_size.physical());
page_zip.data = page + srv_page_size;
#ifdef UNIV_DEBUG
page_zip.m_start =
#endif /* UNIV_DEBUG */
page_zip.m_end = page_zip.m_nonempty =
page_zip.n_blobs = 0;
buf_flush_init_for_writing(NULL, page, &page_zip, 0);
*err = os_file_write(
request, path, file, page_zip.data, 0,
page_size.physical());
}
ut_free(buf2);
if (*err != DB_SUCCESS) {
ib::error()
<< "Could not write the first page to"
<< " tablespace '" << path << "'";
goto err_exit;
}
if (!os_file_flush(file)) {
ib::error() << "File flush of tablespace '"
<< path << "' failed";
*err = DB_ERROR;
goto err_exit;
}
if (has_data_dir) {
/* Make the ISL file if the IBD file is not
in the default location. */
*err = RemoteDatafile::create_link_file(name, path);
if (*err != DB_SUCCESS) {
goto err_exit;
}
}
space = fil_space_create(name, space_id, flags, FIL_TYPE_TABLESPACE,
crypt_data, mode);
if (!space) {
free(crypt_data);
*err = DB_ERROR;
} else {
fil_node_t* file = space->add(path, OS_FILE_CLOSED, size,
false, true);
mtr_t mtr;
mtr.start();
fil_op_write_log(
MLOG_FILE_CREATE2, space_id, 0, file->name,
NULL, space->flags & ~FSP_FLAGS_MEM_MASK, &mtr);
fil_name_write(space, 0, file, &mtr);
mtr.commit();
file->block_size = block_size;
space->punch_hole = punch_hole;
*err = DB_SUCCESS;
}
os_file_close(file);
if (*err != DB_SUCCESS) {
if (has_data_dir) {
RemoteDatafile::delete_link_file(name);
}
os_file_delete(innodb_data_file_key, path);
}
return space;
}
/** Try to open a single-table tablespace and optionally check that the
space id in it is correct. If this does not succeed, print an error message
to the .err log. This function is used to open a tablespace when we start
mysqld after the dictionary has been booted, and also in IMPORT TABLESPACE.
NOTE that we assume this operation is used either at the database startup
or under the protection of the dictionary mutex, so that two users cannot
race here. This operation does not leave the file associated with the
tablespace open, but closes it after we have looked at the space id in it.
If the validate boolean is set, we read the first page of the file and
check that the space id in the file is what we expect. We assume that
this function runs much faster if no check is made, since accessing the
file inode probably is much faster (the OS caches them) than accessing
the first page of the file. This boolean may be initially false, but if
a remote tablespace is found it will be changed to true.
If the fix_dict boolean is set, then it is safe to use an internal SQL
statement to update the dictionary tables if they are incorrect.
@param[in] validate true if we should validate the tablespace
@param[in] fix_dict true if the dictionary is available to be fixed
@param[in] purpose FIL_TYPE_TABLESPACE or FIL_TYPE_TEMPORARY
@param[in] id tablespace ID
@param[in] flags expected FSP_SPACE_FLAGS
@param[in] space_name tablespace name of the datafile
If file-per-table, it is the table name in the databasename/tablename format
@param[in] path_in expected filepath, usually read from dictionary
@param[out] err DB_SUCCESS or error code
@return tablespace
@retval NULL if the tablespace could not be opened */
fil_space_t*
fil_ibd_open(
bool validate,
bool fix_dict,
fil_type_t purpose,
ulint id,
ulint flags,
const table_name_t& tablename,
const char* path_in,
dberr_t* err)
{
mutex_enter(&fil_system.mutex);
if (fil_space_t* space = fil_space_get_by_id(id)) {
if (strcmp(space->name, tablename.m_name)) {
table_name_t space_name;
space_name.m_name = space->name;
ib::error()
<< "Trying to open table " << tablename
<< " with id " << id
<< ", conflicting with " << space_name;
space = NULL;
if (err) *err = DB_TABLESPACE_EXISTS;
} else if (err) *err = DB_SUCCESS;
mutex_exit(&fil_system.mutex);
if (space && validate && !srv_read_only_mode) {
fsp_flags_try_adjust(space,
flags & ~FSP_FLAGS_MEM_MASK);
}
return space;
}
mutex_exit(&fil_system.mutex);
bool dict_filepath_same_as_default = false;
bool link_file_found = false;
bool link_file_is_bad = false;
Datafile df_default; /* default location */
Datafile df_dict; /* dictionary location */
RemoteDatafile df_remote; /* remote location */
ulint tablespaces_found = 0;
ulint valid_tablespaces_found = 0;
ut_ad(!fix_dict || rw_lock_own(&dict_operation_lock, RW_LOCK_X));
ut_ad(!fix_dict || mutex_own(&dict_sys->mutex));
ut_ad(!fix_dict || !srv_read_only_mode);
ut_ad(!fix_dict || srv_log_file_size != 0);
ut_ad(fil_type_is_data(purpose));
/* Table flags can be ULINT_UNDEFINED if
dict_tf_to_fsp_flags_failure is set. */
if (flags == ULINT_UNDEFINED) {
corrupted:
if (err) *err = DB_CORRUPTION;
return NULL;
}
ut_ad(fsp_flags_is_valid(flags & ~FSP_FLAGS_MEM_MASK, id));
df_default.init(tablename.m_name, flags);
df_dict.init(tablename.m_name, flags);
df_remote.init(tablename.m_name, flags);
/* Discover the correct file by looking in three possible locations
while avoiding unecessary effort. */
/* We will always look for an ibd in the default location. */
df_default.make_filepath(NULL, tablename.m_name, IBD);
/* Look for a filepath embedded in an ISL where the default file
would be. */
if (df_remote.open_read_only(true) == DB_SUCCESS) {
ut_ad(df_remote.is_open());
/* Always validate a file opened from an ISL pointer */
validate = true;
++tablespaces_found;
link_file_found = true;
} else if (df_remote.filepath() != NULL) {
/* An ISL file was found but contained a bad filepath in it.
Better validate anything we do find. */
validate = true;
}
/* Attempt to open the tablespace at the dictionary filepath. */
if (path_in) {
if (df_default.same_filepath_as(path_in)) {
dict_filepath_same_as_default = true;
} else {
/* Dict path is not the default path. Always validate
remote files. If default is opened, it was moved. */
validate = true;
df_dict.set_filepath(path_in);
if (df_dict.open_read_only(true) == DB_SUCCESS) {
ut_ad(df_dict.is_open());
++tablespaces_found;
}
}
}
/* Always look for a file at the default location. But don't log
an error if the tablespace is already open in remote or dict. */
ut_a(df_default.filepath());
const bool strict = (tablespaces_found == 0);
if (df_default.open_read_only(strict) == DB_SUCCESS) {
ut_ad(df_default.is_open());
++tablespaces_found;
}
/* Check if multiple locations point to the same file. */
if (tablespaces_found > 1 && df_default.same_as(df_remote)) {
/* A link file was found with the default path in it.
Use the default path and delete the link file. */
--tablespaces_found;
df_remote.delete_link_file();
df_remote.close();
}
if (tablespaces_found > 1 && df_default.same_as(df_dict)) {
--tablespaces_found;
df_dict.close();
}
if (tablespaces_found > 1 && df_remote.same_as(df_dict)) {
--tablespaces_found;
df_dict.close();
}
/* We have now checked all possible tablespace locations and
have a count of how many unique files we found. If things are
normal, we only found 1. */
/* For encrypted tablespace, we need to check the
encryption in header of first page. */
if (!validate && tablespaces_found == 1) {
goto skip_validate;
}
/* Read and validate the first page of these three tablespace
locations, if found. */
valid_tablespaces_found +=
(df_remote.validate_to_dd(id, flags) == DB_SUCCESS);
valid_tablespaces_found +=
(df_default.validate_to_dd(id, flags) == DB_SUCCESS);
valid_tablespaces_found +=
(df_dict.validate_to_dd(id, flags) == DB_SUCCESS);
/* Make sense of these three possible locations.
First, bail out if no tablespace files were found. */
if (valid_tablespaces_found == 0) {
os_file_get_last_error(true);
ib::error() << "Could not find a valid tablespace file for `"
<< tablename << "`. " << TROUBLESHOOT_DATADICT_MSG;
goto corrupted;
}
if (!validate) {
goto skip_validate;
}
/* Do not open any tablespaces if more than one tablespace with
the correct space ID and flags were found. */
if (tablespaces_found > 1) {
ib::error() << "A tablespace for `" << tablename
<< "` has been found in multiple places;";
if (df_default.is_open()) {
ib::error() << "Default location: "
<< df_default.filepath()
<< ", Space ID=" << df_default.space_id()
<< ", Flags=" << df_default.flags();
}
if (df_remote.is_open()) {
ib::error() << "Remote location: "
<< df_remote.filepath()
<< ", Space ID=" << df_remote.space_id()
<< ", Flags=" << df_remote.flags();
}
if (df_dict.is_open()) {
ib::error() << "Dictionary location: "
<< df_dict.filepath()
<< ", Space ID=" << df_dict.space_id()
<< ", Flags=" << df_dict.flags();
}
/* Force-recovery will allow some tablespaces to be
skipped by REDO if there was more than one file found.
Unlike during the REDO phase of recovery, we now know
if the tablespace is valid according to the dictionary,
which was not available then. So if we did not force
recovery and there is only one good tablespace, ignore
any bad tablespaces. */
if (valid_tablespaces_found > 1 || srv_force_recovery > 0) {
ib::error() << "Will not open tablespace `"
<< tablename << "`";
/* If the file is not open it cannot be valid. */
ut_ad(df_default.is_open() || !df_default.is_valid());
ut_ad(df_dict.is_open() || !df_dict.is_valid());
ut_ad(df_remote.is_open() || !df_remote.is_valid());
/* Having established that, this is an easy way to
look for corrupted data files. */
if (df_default.is_open() != df_default.is_valid()
|| df_dict.is_open() != df_dict.is_valid()
|| df_remote.is_open() != df_remote.is_valid()) {
goto corrupted;
}
error:
if (err) *err = DB_ERROR;
return NULL;
}
/* There is only one valid tablespace found and we did
not use srv_force_recovery during REDO. Use this one
tablespace and clean up invalid tablespace pointers */
if (df_default.is_open() && !df_default.is_valid()) {
df_default.close();
tablespaces_found--;
}
if (df_dict.is_open() && !df_dict.is_valid()) {
df_dict.close();
/* Leave dict.filepath so that SYS_DATAFILES
can be corrected below. */
tablespaces_found--;
}
if (df_remote.is_open() && !df_remote.is_valid()) {
df_remote.close();
tablespaces_found--;
link_file_is_bad = true;
}
}
/* At this point, there should be only one filepath. */
ut_a(tablespaces_found == 1);
ut_a(valid_tablespaces_found == 1);
/* Only fix the dictionary at startup when there is only one thread.
Calls to dict_load_table() can be done while holding other latches. */
if (!fix_dict) {
goto skip_validate;
}
/* We may need to update what is stored in SYS_DATAFILES or
SYS_TABLESPACES or adjust the link file. Since a failure to
update SYS_TABLESPACES or SYS_DATAFILES does not prevent opening
and using the tablespace either this time or the next, we do not
check the return code or fail to open the tablespace. But if it
fails, dict_update_filepath() will issue a warning to the log. */
if (df_dict.filepath()) {
ut_ad(path_in != NULL);
ut_ad(df_dict.same_filepath_as(path_in));
if (df_remote.is_open()) {
if (!df_remote.same_filepath_as(path_in)) {
dict_update_filepath(id, df_remote.filepath());
}
} else if (df_default.is_open()) {
ut_ad(!dict_filepath_same_as_default);
dict_update_filepath(id, df_default.filepath());
if (link_file_is_bad) {
RemoteDatafile::delete_link_file(
tablename.m_name);
}
} else if (!link_file_found || link_file_is_bad) {
ut_ad(df_dict.is_open());
/* Fix the link file if we got our filepath
from the dictionary but a link file did not
exist or it did not point to a valid file. */
RemoteDatafile::delete_link_file(tablename.m_name);
RemoteDatafile::create_link_file(
tablename.m_name, df_dict.filepath());
}
} else if (df_remote.is_open()) {
if (dict_filepath_same_as_default) {
dict_update_filepath(id, df_remote.filepath());
} else if (path_in == NULL) {
/* SYS_DATAFILES record for this space ID
was not found. */
dict_replace_tablespace_and_filepath(
id, tablename.m_name,
df_remote.filepath(), flags);
}
} else if (df_default.is_open()) {
/* We opened the tablespace in the default location.
SYS_DATAFILES.PATH needs to be updated if it is different
from this default path or if the SYS_DATAFILES.PATH was not
supplied and it should have been. Also update the dictionary
if we found an ISL file (since !df_remote.is_open). Since
path_in is not suppled for file-per-table, we must assume
that it matched the ISL. */
if ((path_in != NULL && !dict_filepath_same_as_default)
|| (path_in == NULL && DICT_TF_HAS_DATA_DIR(flags))
|| df_remote.filepath() != NULL) {
dict_replace_tablespace_and_filepath(
id, tablename.m_name, df_default.filepath(),
flags);
}
}
skip_validate:
const byte* first_page =
df_default.is_open() ? df_default.get_first_page() :
df_dict.is_open() ? df_dict.get_first_page() :
df_remote.get_first_page();
fil_space_crypt_t* crypt_data = first_page
? fil_space_read_crypt_data(page_size_t(flags), first_page)
: NULL;
fil_space_t* space = fil_space_create(
tablename.m_name, id, flags, purpose, crypt_data);
if (!space) {
goto error;
}
/* We do not measure the size of the file, that is why
we pass the 0 below */
space->add(
df_remote.is_open() ? df_remote.filepath() :
df_dict.is_open() ? df_dict.filepath() :
df_default.filepath(), OS_FILE_CLOSED, 0, false, true);
if (validate && purpose != FIL_TYPE_IMPORT && !srv_read_only_mode) {
df_remote.close();
df_dict.close();
df_default.close();
fsp_flags_try_adjust(space, flags & ~FSP_FLAGS_MEM_MASK);
}
if (err) *err = DB_SUCCESS;
return space;
}
/** Looks for a pre-existing fil_space_t with the given tablespace ID
and, if found, returns the name and filepath in newly allocated buffers
that the caller must free.
@param[in] space_id The tablespace ID to search for.
@param[out] name Name of the tablespace found.
@param[out] filepath The filepath of the first datafile for the
tablespace.
@return true if tablespace is found, false if not. */
bool
fil_space_read_name_and_filepath(
ulint space_id,
char** name,
char** filepath)
{
bool success = false;
*name = NULL;
*filepath = NULL;
mutex_enter(&fil_system.mutex);
fil_space_t* space = fil_space_get_by_id(space_id);
if (space != NULL) {
*name = mem_strdup(space->name);
fil_node_t* node = UT_LIST_GET_FIRST(space->chain);
*filepath = mem_strdup(node->name);
success = true;
}
mutex_exit(&fil_system.mutex);
return(success);
}
/** Convert a file name to a tablespace name.
@param[in] filename directory/databasename/tablename.ibd
@return database/tablename string, to be freed with ut_free() */
char*
fil_path_to_space_name(
const char* filename)
{
/* Strip the file name prefix and suffix, leaving
only databasename/tablename. */
ulint filename_len = strlen(filename);
const char* end = filename + filename_len;
#ifdef HAVE_MEMRCHR
const char* tablename = 1 + static_cast<const char*>(
memrchr(filename, OS_PATH_SEPARATOR,
filename_len));
const char* dbname = 1 + static_cast<const char*>(
memrchr(filename, OS_PATH_SEPARATOR,
tablename - filename - 1));
#else /* HAVE_MEMRCHR */
const char* tablename = filename;
const char* dbname = NULL;
while (const char* t = static_cast<const char*>(
memchr(tablename, OS_PATH_SEPARATOR,
ulint(end - tablename)))) {
dbname = tablename;
tablename = t + 1;
}
#endif /* HAVE_MEMRCHR */
ut_ad(dbname != NULL);
ut_ad(tablename > dbname);
ut_ad(tablename < end);
ut_ad(end - tablename > 4);
ut_ad(memcmp(end - 4, DOT_IBD, 4) == 0);
char* name = mem_strdupl(dbname, ulint(end - dbname) - 4);
ut_ad(name[tablename - dbname - 1] == OS_PATH_SEPARATOR);
#if OS_PATH_SEPARATOR != '/'
/* space->name uses '/', not OS_PATH_SEPARATOR. */
name[tablename - dbname - 1] = '/';
#endif
return(name);
}
/** Discover the correct IBD file to open given a remote or missing
filepath from the REDO log. Administrators can move a crashed
database to another location on the same machine and try to recover it.
Remote IBD files might be moved as well to the new location.
The problem with this is that the REDO log contains the old location
which may be still accessible. During recovery, if files are found in
both locations, we can chose on based on these priorities;
1. Default location
2. ISL location
3. REDO location
@param[in] space_id tablespace ID
@param[in] df Datafile object with path from redo
@return true if a valid datafile was found, false if not */
static
bool
fil_ibd_discover(
ulint space_id,
Datafile& df)
{
Datafile df_def_per; /* default file-per-table datafile */
RemoteDatafile df_rem_per; /* remote file-per-table datafile */
/* Look for the datafile in the default location. */
const char* filename = df.filepath();
const char* basename = base_name(filename);
/* If this datafile is file-per-table it will have a schema dir. */
ulint sep_found = 0;
const char* db = basename;
for (; db > filename && sep_found < 2; db--) {
if (db[0] == OS_PATH_SEPARATOR) {
sep_found++;
}
}
if (sep_found == 2) {
db += 2;
df_def_per.init(db, 0);
df_def_per.make_filepath(NULL, db, IBD);
if (df_def_per.open_read_only(false) == DB_SUCCESS
&& df_def_per.validate_for_recovery() == DB_SUCCESS
&& df_def_per.space_id() == space_id) {
df.set_filepath(df_def_per.filepath());
df.open_read_only(false);
return(true);
}
/* Look for a remote file-per-table tablespace. */
switch (srv_operation) {
case SRV_OPERATION_BACKUP:
case SRV_OPERATION_RESTORE_DELTA:
ut_ad(0);
break;
case SRV_OPERATION_RESTORE_ROLLBACK_XA:
case SRV_OPERATION_RESTORE_EXPORT:
case SRV_OPERATION_RESTORE:
break;
case SRV_OPERATION_NORMAL:
df_rem_per.set_name(db);
if (df_rem_per.open_link_file() != DB_SUCCESS) {
break;
}
/* An ISL file was found with contents. */
if (df_rem_per.open_read_only(false) != DB_SUCCESS
|| df_rem_per.validate_for_recovery()
!= DB_SUCCESS) {
/* Assume that this ISL file is intended to
be used. Do not continue looking for another
if this file cannot be opened or is not
a valid IBD file. */
ib::error() << "ISL file '"
<< df_rem_per.link_filepath()
<< "' was found but the linked file '"
<< df_rem_per.filepath()
<< "' could not be opened or is"
" not correct.";
return(false);
}
/* Use this file if it has the space_id from the
MLOG record. */
if (df_rem_per.space_id() == space_id) {
df.set_filepath(df_rem_per.filepath());
df.open_read_only(false);
return(true);
}
/* Since old MLOG records can use the same basename
in multiple CREATE/DROP TABLE sequences, this ISL
file could be pointing to a later version of this
basename.ibd file which has a different space_id.
Keep looking. */
}
}
/* No ISL files were found in the default location. Use the location
given in the redo log. */
if (df.open_read_only(false) == DB_SUCCESS
&& df.validate_for_recovery() == DB_SUCCESS
&& df.space_id() == space_id) {
return(true);
}
/* A datafile was not discovered for the filename given. */
return(false);
}
/** Open an ibd tablespace and add it to the InnoDB data structures.
This is similar to fil_ibd_open() except that it is used while processing
the REDO log, so the data dictionary is not available and very little
validation is done. The tablespace name is extracred from the
dbname/tablename.ibd portion of the filename, which assumes that the file
is a file-per-table tablespace. Any name will do for now. General
tablespace names will be read from the dictionary after it has been
recovered. The tablespace flags are read at this time from the first page
of the file in validate_for_recovery().
@param[in] space_id tablespace ID
@param[in] filename path/to/databasename/tablename.ibd
@param[out] space the tablespace, or NULL on error
@return status of the operation */
enum fil_load_status
fil_ibd_load(
ulint space_id,
const char* filename,
fil_space_t*& space)
{
/* If the a space is already in the file system cache with this
space ID, then there is nothing to do. */
mutex_enter(&fil_system.mutex);
space = fil_space_get_by_id(space_id);
mutex_exit(&fil_system.mutex);
if (space != NULL) {
/* Compare the filename we are trying to open with the
filename from the first node of the tablespace we opened
previously. Fail if it is different. */
fil_node_t* node = UT_LIST_GET_FIRST(space->chain);
if (0 != strcmp(innobase_basename(filename),
innobase_basename(node->name))) {
ib::info()
<< "Ignoring data file '" << filename
<< "' with space ID " << space->id
<< ". Another data file called " << node->name
<< " exists with the same space ID.";
space = NULL;
return(FIL_LOAD_ID_CHANGED);
}
return(FIL_LOAD_OK);
}
if (is_mariabackup_restore()) {
/* Replace absolute DATA DIRECTORY file paths with
short names relative to the backup directory. */
if (const char* name = strrchr(filename, OS_PATH_SEPARATOR)) {
while (--name > filename
&& *name != OS_PATH_SEPARATOR);
if (name > filename) {
filename = name + 1;
}
}
}
Datafile file;
file.set_filepath(filename);
file.open_read_only(false);
if (!file.is_open()) {
/* The file has been moved or it is a remote datafile. */
if (!fil_ibd_discover(space_id, file)
|| !file.is_open()) {
return(FIL_LOAD_NOT_FOUND);
}
}
os_offset_t size;
/* Read and validate the first page of the tablespace.
Assign a tablespace name based on the tablespace type. */
switch (file.validate_for_recovery()) {
os_offset_t minimum_size;
case DB_SUCCESS:
if (file.space_id() != space_id) {
ib::info()
<< "Ignoring data file '"
<< file.filepath()
<< "' with space ID " << file.space_id()
<< ", since the redo log references "
<< file.filepath() << " with space ID "
<< space_id << ".";
return(FIL_LOAD_ID_CHANGED);
}
/* Get and test the file size. */
size = os_file_get_size(file.handle());
/* Every .ibd file is created >= 4 pages in size.
Smaller files cannot be OK. */
minimum_size = os_offset_t(FIL_IBD_FILE_INITIAL_SIZE)
<< srv_page_size_shift;
if (size == static_cast<os_offset_t>(-1)) {
/* The following call prints an error message */
os_file_get_last_error(true);
ib::error() << "Could not measure the size of"
" single-table tablespace file '"
<< file.filepath() << "'";
} else if (size < minimum_size) {
ib::error() << "The size of tablespace file '"
<< file.filepath() << "' is only " << size
<< ", should be at least " << minimum_size
<< "!";
} else {
/* Everything is fine so far. */
break;
}
/* fall through */
case DB_TABLESPACE_EXISTS:
return(FIL_LOAD_INVALID);
default:
return(FIL_LOAD_NOT_FOUND);
}
ut_ad(space == NULL);
/* Adjust the memory-based flags that would normally be set by
dict_tf_to_fsp_flags(). In recovery, we have no data dictionary. */
ulint flags = file.flags();
if (FSP_FLAGS_HAS_PAGE_COMPRESSION(flags)) {
flags |= page_zip_level
<< FSP_FLAGS_MEM_COMPRESSION_LEVEL;
}
const byte* first_page = file.get_first_page();
fil_space_crypt_t* crypt_data = first_page
? fil_space_read_crypt_data(page_size_t(flags), first_page)
: NULL;
space = fil_space_create(
file.name(), space_id, flags, FIL_TYPE_TABLESPACE, crypt_data);
if (space == NULL) {
return(FIL_LOAD_INVALID);
}
ut_ad(space->id == file.space_id());
ut_ad(space->id == space_id);
/* We do not use the size information we have about the file, because
the rounding formula for extents and pages is somewhat complex; we
let fil_node_open() do that task. */
space->add(file.filepath(), OS_FILE_CLOSED, 0, false, false);
return(FIL_LOAD_OK);
}
/***********************************************************************//**
A fault-tolerant function that tries to read the next file name in the
directory. We retry 100 times if os_file_readdir_next_file() returns -1. The
idea is to read as much good data as we can and jump over bad data.
@return 0 if ok, -1 if error even after the retries, 1 if at the end
of the directory */
int
fil_file_readdir_next_file(
/*=======================*/
dberr_t* err, /*!< out: this is set to DB_ERROR if an error
was encountered, otherwise not changed */
const char* dirname,/*!< in: directory name or path */
os_file_dir_t dir, /*!< in: directory stream */
os_file_stat_t* info) /*!< in/out: buffer where the
info is returned */
{
for (ulint i = 0; i < 100; i++) {
int ret = os_file_readdir_next_file(dirname, dir, info);
if (ret != -1) {
return(ret);
}
ib::error() << "os_file_readdir_next_file() returned -1 in"
" directory " << dirname
<< ", crash recovery may have failed"
" for some .ibd files!";
*err = DB_ERROR;
}
return(-1);
}
/** Try to adjust FSP_SPACE_FLAGS if they differ from the expectations.
(Typically when upgrading from MariaDB 10.1.0..10.1.20.)
@param[in,out] space tablespace
@param[in] flags desired tablespace flags */
void fsp_flags_try_adjust(fil_space_t* space, ulint flags)
{
ut_ad(!srv_read_only_mode);
ut_ad(fsp_flags_is_valid(flags, space->id));
if (!space->size && (space->purpose != FIL_TYPE_TABLESPACE
|| !fil_space_get_size(space->id))) {
return;
}
/* This code is executed during server startup while no
connections are allowed. We do not need to protect against
DROP TABLE by fil_space_acquire(). */
mtr_t mtr;
mtr.start();
if (buf_block_t* b = buf_page_get(
page_id_t(space->id, 0), page_size_t(flags),
RW_X_LATCH, &mtr)) {
ulint f = fsp_header_get_flags(b->frame);
/* Suppress the message if only the DATA_DIR flag to differs. */
if ((f ^ flags) & ~(1U << FSP_FLAGS_POS_RESERVED)) {
ib::warn()
<< "adjusting FSP_SPACE_FLAGS of file '"
<< UT_LIST_GET_FIRST(space->chain)->name
<< "' from " << ib::hex(f)
<< " to " << ib::hex(flags);
}
if (f != flags) {
mtr.set_named_space(space);
mlog_write_ulint(FSP_HEADER_OFFSET
+ FSP_SPACE_FLAGS + b->frame,
flags, MLOG_4BYTES, &mtr);
}
}
mtr.commit();
}
/** Determine if a matching tablespace exists in the InnoDB tablespace
memory cache. Note that if we have not done a crash recovery at the database
startup, there may be many tablespaces which are not yet in the memory cache.
@param[in] id Tablespace ID
@param[in] name Tablespace name used in fil_space_create().
@param[in] table_flags table flags
@return the tablespace
@retval NULL if no matching tablespace exists in the memory cache */
fil_space_t*
fil_space_for_table_exists_in_mem(
ulint id,
const char* name,
ulint table_flags)
{
const ulint expected_flags = dict_tf_to_fsp_flags(table_flags);
mutex_enter(&fil_system.mutex);
if (fil_space_t* space = fil_space_get_by_id(id)) {
if ((space->flags ^ expected_flags) & ~FSP_FLAGS_MEM_MASK) {
goto func_exit;
}
if (strcmp(space->name, name)) {
ib::error() << "Table " << name
<< " in InnoDB data dictionary"
" has tablespace id " << id
<< ", but the tablespace"
" with that id has name " << space->name << "."
" Have you deleted or moved .ibd files?";
ib::info() << TROUBLESHOOT_DATADICT_MSG;
goto func_exit;
}
/* Adjust the flags that are in FSP_FLAGS_MEM_MASK.
FSP_SPACE_FLAGS will not be written back here. */
space->flags = expected_flags;
mutex_exit(&fil_system.mutex);
if (!srv_read_only_mode) {
fsp_flags_try_adjust(space, expected_flags
& ~FSP_FLAGS_MEM_MASK);
}
return space;
}
func_exit:
mutex_exit(&fil_system.mutex);
return NULL;
}
/*============================ FILE I/O ================================*/
/********************************************************************//**
NOTE: you must call fil_mutex_enter_and_prepare_for_io() first!
Prepares a file node for i/o. Opens the file if it is closed. Updates the
pending i/o's field in the node and the system appropriately. Takes the node
off the LRU list if it is in the LRU list. The caller must hold the fil_sys
mutex.
@return false if the file can't be opened, otherwise true */
static
bool
fil_node_prepare_for_io(
/*====================*/
fil_node_t* node, /*!< in: file node */
fil_space_t* space) /*!< in: space */
{
ut_ad(node && space);
ut_ad(mutex_own(&fil_system.mutex));
if (fil_system.n_open > srv_max_n_open_files + 5) {
ib::warn() << "Open files " << fil_system.n_open
<< " exceeds the limit " << srv_max_n_open_files;
}
if (!node->is_open()) {
/* File is closed: open it */
ut_a(node->n_pending == 0);
if (!fil_node_open_file(node)) {
return(false);
}
}
if (node->n_pending == 0 && fil_space_belongs_in_lru(space)) {
/* The node is in the LRU list, remove it */
ut_a(UT_LIST_GET_LEN(fil_system.LRU) > 0);
UT_LIST_REMOVE(fil_system.LRU, node);
}
node->n_pending++;
return(true);
}
/** Update the data structures when an i/o operation finishes.
@param[in,out] node file node
@param[in] type IO context */
static
void
fil_node_complete_io(fil_node_t* node, const IORequest& type)
{
ut_ad(mutex_own(&fil_system.mutex));
ut_a(node->n_pending > 0);
--node->n_pending;
ut_ad(type.validate());
if (type.is_write()) {
ut_ad(!srv_read_only_mode
|| node->space->purpose == FIL_TYPE_TEMPORARY);
if (fil_buffering_disabled(node->space)) {
/* We don't need to keep track of unflushed
changes as user has explicitly disabled
buffering. */
ut_ad(!node->space->is_in_unflushed_spaces);
ut_ad(node->needs_flush == false);
} else {
node->needs_flush = true;
if (!node->space->is_in_unflushed_spaces) {
node->space->is_in_unflushed_spaces = true;
fil_system.unflushed_spaces.push_front(
*node->space);
}
}
}
if (node->n_pending == 0 && fil_space_belongs_in_lru(node->space)) {
/* The node must be put back to the LRU list */
UT_LIST_ADD_FIRST(fil_system.LRU, node);
}
}
/** Report information about an invalid page access. */
static
void
fil_report_invalid_page_access(
ulint block_offset, /*!< in: block offset */
ulint space_id, /*!< in: space id */
const char* space_name, /*!< in: space name */
ulint byte_offset, /*!< in: byte offset */
ulint len, /*!< in: I/O length */
bool is_read) /*!< in: I/O type */
{
ib::fatal()
<< "Trying to " << (is_read ? "read" : "write")
<< " page number " << block_offset << " in"
" space " << space_id << ", space name " << space_name << ","
" which is outside the tablespace bounds. Byte offset "
<< byte_offset << ", len " << len <<
(space_id == 0 && !srv_was_started
? "Please check that the configuration matches"
" the InnoDB system tablespace location (ibdata files)"
: "");
}
/** Reads or writes data. This operation could be asynchronous (aio).
@param[in,out] type IO context
@param[in] sync true if synchronous aio is desired
@param[in] page_id page id
@param[in] page_size page size
@param[in] byte_offset remainder of offset in bytes; in aio this
must be divisible by the OS block size
@param[in] len how many bytes to read or write; this must
not cross a file boundary; in aio this must
be a block size multiple
@param[in,out] buf buffer where to store read data or from where
to write; in aio this must be appropriately
aligned
@param[in] message message for aio handler if non-sync aio
used, else ignored
@param[in] ignore_missing_space true=ignore missing space duging read
@return DB_SUCCESS, DB_TABLESPACE_DELETED or DB_TABLESPACE_TRUNCATED
if we are trying to do i/o on a tablespace which does not exist */
dberr_t
fil_io(
const IORequest& type,
bool sync,
const page_id_t page_id,
const page_size_t& page_size,
ulint byte_offset,
ulint len,
void* buf,
void* message,
bool ignore_missing_space)
{
os_offset_t offset;
IORequest req_type(type);
ut_ad(req_type.validate());
ut_ad(len > 0);
ut_ad(byte_offset < srv_page_size);
ut_ad(!page_size.is_compressed() || byte_offset == 0);
ut_ad(srv_page_size == 1UL << srv_page_size_shift);
compile_time_assert((1U << UNIV_PAGE_SIZE_SHIFT_MAX)
== UNIV_PAGE_SIZE_MAX);
compile_time_assert((1U << UNIV_PAGE_SIZE_SHIFT_MIN)
== UNIV_PAGE_SIZE_MIN);
ut_ad(fil_validate_skip());
/* ibuf bitmap pages must be read in the sync AIO mode: */
ut_ad(recv_no_ibuf_operations
|| req_type.is_write()
|| !ibuf_bitmap_page(page_id, page_size)
|| sync
|| req_type.is_log());
ulint mode;
if (sync) {
mode = OS_AIO_SYNC;
} else if (req_type.is_log()) {
mode = OS_AIO_LOG;
} else if (req_type.is_read()
&& !recv_no_ibuf_operations
&& ibuf_page(page_id, page_size, NULL)) {
mode = OS_AIO_IBUF;
/* Reduce probability of deadlock bugs in connection with ibuf:
do not let the ibuf i/o handler sleep */
req_type.clear_do_not_wake();
} else {
mode = OS_AIO_NORMAL;
}
if (req_type.is_read()) {
srv_stats.data_read.add(len);
} else if (req_type.is_write()) {
ut_ad(!srv_read_only_mode
|| fsp_is_system_temporary(page_id.space()));
srv_stats.data_written.add(len);
}
/* Reserve the fil_system mutex and make sure that we can open at
least one file while holding it, if the file is not already open */
fil_mutex_enter_and_prepare_for_io(page_id.space());
fil_space_t* space = fil_space_get_by_id(page_id.space());
/* If we are deleting a tablespace we don't allow async read operations
on that. However, we do allow write operations and sync read operations. */
if (space == NULL
|| (req_type.is_read()
&& !sync
&& space->stop_new_ops
&& !space->is_being_truncated)) {
mutex_exit(&fil_system.mutex);
if (!req_type.ignore_missing() && !ignore_missing_space) {
ib::error()
<< "Trying to do I/O to a tablespace which"
" does not exist. I/O type: "
<< (req_type.is_read() ? "read" : "write")
<< ", page: " << page_id
<< ", I/O length: " << len << " bytes";
}
return(DB_TABLESPACE_DELETED);
}
ut_ad(mode != OS_AIO_IBUF || fil_type_is_data(space->purpose));
ulint cur_page_no = page_id.page_no();
fil_node_t* node = UT_LIST_GET_FIRST(space->chain);
for (;;) {
if (node == NULL) {
if (req_type.ignore_missing()) {
mutex_exit(&fil_system.mutex);
return(DB_ERROR);
}
fil_report_invalid_page_access(
page_id.page_no(), page_id.space(),
space->name, byte_offset, len,
req_type.is_read());
} else if (fil_is_user_tablespace_id(space->id)
&& node->size == 0) {
/* We do not know the size of a single-table tablespace
before we open the file */
break;
} else if (node->size > cur_page_no) {
/* Found! */
break;
} else {
if (space->id != TRX_SYS_SPACE
&& UT_LIST_GET_LEN(space->chain) == 1
&& (srv_is_tablespace_truncated(space->id)
|| srv_was_tablespace_truncated(space))
&& req_type.is_read()) {
/* Handle page which is outside the truncated
tablespace bounds when recovering from a crash
happened during a truncation */
mutex_exit(&fil_system.mutex);
return(DB_TABLESPACE_TRUNCATED);
}
cur_page_no -= node->size;
node = UT_LIST_GET_NEXT(chain, node);
}
}
/* Open file if closed */
if (!fil_node_prepare_for_io(node, space)) {
if (fil_type_is_data(space->purpose)
&& fil_is_user_tablespace_id(space->id)) {
mutex_exit(&fil_system.mutex);
if (!req_type.ignore_missing()) {
ib::error()
<< "Trying to do I/O to a tablespace"
" which exists without .ibd data file."
" I/O type: "
<< (req_type.is_read()
? "read" : "write")
<< ", page: "
<< page_id_t(page_id.space(),
cur_page_no)
<< ", I/O length: " << len << " bytes";
}
return(DB_TABLESPACE_DELETED);
}
/* The tablespace is for log. Currently, we just assert here
to prevent handling errors along the way fil_io returns.
Also, if the log files are missing, it would be hard to
promise the server can continue running. */
ut_a(0);
}
/* Check that at least the start offset is within the bounds of a
single-table tablespace, including rollback tablespaces. */
if (node->size <= cur_page_no
&& space->id != TRX_SYS_SPACE
&& fil_type_is_data(space->purpose)) {
if (req_type.ignore_missing()) {
/* If we can tolerate the non-existent pages, we
should return with DB_ERROR and let caller decide
what to do. */
fil_node_complete_io(node, req_type);
mutex_exit(&fil_system.mutex);
return(DB_ERROR);
}
fil_report_invalid_page_access(
page_id.page_no(), page_id.space(),
space->name, byte_offset, len, req_type.is_read());
}
/* Now we have made the changes in the data structures of fil_system */
mutex_exit(&fil_system.mutex);
/* Calculate the low 32 bits and the high 32 bits of the file offset */
if (!page_size.is_compressed()) {
offset = ((os_offset_t) cur_page_no
<< srv_page_size_shift) + byte_offset;
ut_a(node->size - cur_page_no
>= ((byte_offset + len + (srv_page_size - 1))
>> srv_page_size_shift));
} else {
ulint size_shift;
switch (page_size.physical()) {
case 1024: size_shift = 10; break;
case 2048: size_shift = 11; break;
case 4096: size_shift = 12; break;
case 8192: size_shift = 13; break;
case 16384: size_shift = 14; break;
case 32768: size_shift = 15; break;
case 65536: size_shift = 16; break;
default: ut_error;
}
offset = ((os_offset_t) cur_page_no << size_shift)
+ byte_offset;
ut_a(node->size - cur_page_no
>= (len + (page_size.physical() - 1))
/ page_size.physical());
}
/* Do AIO */
ut_a(byte_offset % OS_FILE_LOG_BLOCK_SIZE == 0);
ut_a((len % OS_FILE_LOG_BLOCK_SIZE) == 0);
const char* name = node->name == NULL ? space->name : node->name;
req_type.set_fil_node(node);
ut_ad(!req_type.is_write()
|| page_id.space() == SRV_LOG_SPACE_FIRST_ID
|| !fil_is_user_tablespace_id(page_id.space())
|| offset == page_id.page_no() * page_size.physical());
/* Queue the aio request */
dberr_t err = os_aio(
req_type,
mode, name, node->handle, buf, offset, len,
space->purpose != FIL_TYPE_TEMPORARY
&& srv_read_only_mode,
node, message);
/* We an try to recover the page from the double write buffer if
the decompression fails or the page is corrupt. */
ut_a(req_type.is_dblwr_recover() || err == DB_SUCCESS);
if (sync) {
/* The i/o operation is already completed when we return from
os_aio: */
mutex_enter(&fil_system.mutex);
fil_node_complete_io(node, req_type);
mutex_exit(&fil_system.mutex);
ut_ad(fil_validate_skip());
}
return(err);
}
/**********************************************************************//**
Waits for an aio operation to complete. This function is used to write the
handler for completed requests. The aio array of pending requests is divided
into segments (see os0file.cc for more info). The thread specifies which
segment it wants to wait for. */
void
fil_aio_wait(
/*=========*/
ulint segment) /*!< in: the number of the segment in the aio
array to wait for */
{
fil_node_t* node;
IORequest type;
void* message;
ut_ad(fil_validate_skip());
dberr_t err = os_aio_handler(segment, &node, &message, &type);
ut_a(err == DB_SUCCESS);
if (node == NULL) {
ut_ad(srv_shutdown_state == SRV_SHUTDOWN_EXIT_THREADS);
return;
}
srv_set_io_thread_op_info(segment, "complete io for fil node");
mutex_enter(&fil_system.mutex);
fil_node_complete_io(node, type);
const fil_type_t purpose = node->space->purpose;
const ulint space_id= node->space->id;
const bool dblwr = node->space->use_doublewrite();
mutex_exit(&fil_system.mutex);
ut_ad(fil_validate_skip());
/* Do the i/o handling */
/* IMPORTANT: since i/o handling for reads will read also the insert
buffer in tablespace 0, you have to be very careful not to introduce
deadlocks in the i/o system. We keep tablespace 0 data files always
open, and use a special i/o thread to serve insert buffer requests. */
switch (purpose) {
case FIL_TYPE_LOG:
srv_set_io_thread_op_info(segment, "complete io for log");
/* We use synchronous writing of the logs
and can only end up here when writing a log checkpoint! */
ut_a(ptrdiff_t(message) == 1);
/* It was a checkpoint write */
switch (srv_flush_t(srv_file_flush_method)) {
case SRV_O_DSYNC:
case SRV_NOSYNC:
break;
case SRV_FSYNC:
case SRV_LITTLESYNC:
case SRV_O_DIRECT:
case SRV_O_DIRECT_NO_FSYNC:
#ifdef _WIN32
case SRV_ALL_O_DIRECT_FSYNC:
#endif
fil_flush(SRV_LOG_SPACE_FIRST_ID);
}
DBUG_PRINT("ib_log", ("checkpoint info written"));
log_sys.complete_checkpoint();
return;
case FIL_TYPE_TABLESPACE:
case FIL_TYPE_TEMPORARY:
case FIL_TYPE_IMPORT:
srv_set_io_thread_op_info(segment, "complete io for buf page");
/* async single page writes from the dblwr buffer don't have
access to the page */
buf_page_t* bpage = static_cast<buf_page_t*>(message);
if (!bpage) {
return;
}
ulint offset = bpage->id.page_no();
dberr_t err = buf_page_io_complete(bpage, dblwr);
if (err == DB_SUCCESS) {
return;
}
ut_ad(type.is_read());
if (recv_recovery_is_on() && !srv_force_recovery) {
recv_sys->found_corrupt_fs = true;
}
if (fil_space_t* space = fil_space_acquire_for_io(space_id)) {
if (space == node->space) {
ib::error() << "Failed to read file '"
<< node->name
<< "' at offset " << offset
<< ": " << err;
}
space->release_for_io();
}
return;
}
ut_ad(0);
}
/**********************************************************************//**
Flushes to disk possible writes cached by the OS. If the space does not exist
or is being dropped, does not do anything. */
void
fil_flush(
/*======*/
ulint space_id) /*!< in: file space id (this can be a group of
log files or a tablespace of the database) */
{
mutex_enter(&fil_system.mutex);
if (fil_space_t* space = fil_space_get_by_id(space_id)) {
if (space->purpose != FIL_TYPE_TEMPORARY
&& !space->is_stopping()) {
fil_flush_low(space);
}
}
mutex_exit(&fil_system.mutex);
}
/** Flush a tablespace.
@param[in,out] space tablespace to flush */
void
fil_flush(fil_space_t* space)
{
ut_ad(space->pending_io());
ut_ad(space->purpose == FIL_TYPE_TABLESPACE
|| space->purpose == FIL_TYPE_IMPORT);
if (!space->is_stopping()) {
mutex_enter(&fil_system.mutex);
if (!space->is_stopping()) {
fil_flush_low(space);
}
mutex_exit(&fil_system.mutex);
}
}
/** Flush to disk the writes in file spaces of the given type
possibly cached by the OS.
@param[in] purpose FIL_TYPE_TABLESPACE or FIL_TYPE_LOG */
void
fil_flush_file_spaces(
fil_type_t purpose)
{
ulint* space_ids;
ulint n_space_ids;
ut_ad(purpose == FIL_TYPE_TABLESPACE || purpose == FIL_TYPE_LOG);
mutex_enter(&fil_system.mutex);
n_space_ids = fil_system.unflushed_spaces.size();
if (n_space_ids == 0) {
mutex_exit(&fil_system.mutex);
return;
}
space_ids = static_cast<ulint*>(
ut_malloc_nokey(n_space_ids * sizeof(*space_ids)));
n_space_ids = 0;
for (sized_ilist<fil_space_t, unflushed_spaces_tag_t>::iterator it
= fil_system.unflushed_spaces.begin(),
end = fil_system.unflushed_spaces.end();
it != end; ++it) {
if (it->purpose == purpose && !it->is_stopping()) {
space_ids[n_space_ids++] = it->id;
}
}
mutex_exit(&fil_system.mutex);
/* Flush the spaces. It will not hurt to call fil_flush() on
a non-existing space id. */
for (ulint i = 0; i < n_space_ids; i++) {
fil_flush(space_ids[i]);
}
ut_free(space_ids);
}
/** Functor to validate the file node list of a tablespace. */
struct Check {
/** Total size of file nodes visited so far */
ulint size;
/** Total number of open files visited so far */
ulint n_open;
/** Constructor */
Check() : size(0), n_open(0) {}
/** Visit a file node
@param[in] elem file node to visit */
void operator()(const fil_node_t* elem)
{
ut_a(elem->is_open() || !elem->n_pending);
n_open += elem->is_open();
size += elem->size;
}
/** Validate a tablespace.
@param[in] space tablespace to validate
@return number of open file nodes */
static ulint validate(const fil_space_t* space)
{
ut_ad(mutex_own(&fil_system.mutex));
Check check;
ut_list_validate(space->chain, check);
ut_a(space->size == check.size);
switch (space->id) {
case TRX_SYS_SPACE:
ut_ad(fil_system.sys_space == NULL
|| fil_system.sys_space == space);
break;
case SRV_TMP_SPACE_ID:
ut_ad(fil_system.temp_space == NULL
|| fil_system.temp_space == space);
break;
default:
break;
}
return(check.n_open);
}
};
/******************************************************************//**
Checks the consistency of the tablespace cache.
@return true if ok */
bool
fil_validate(void)
/*==============*/
{
fil_space_t* space;
fil_node_t* fil_node;
ulint n_open = 0;
mutex_enter(&fil_system.mutex);
/* Look for spaces in the hash table */
for (ulint i = 0; i < hash_get_n_cells(fil_system.spaces); i++) {
for (space = static_cast<fil_space_t*>(
HASH_GET_FIRST(fil_system.spaces, i));
space != 0;
space = static_cast<fil_space_t*>(
HASH_GET_NEXT(hash, space))) {
n_open += Check::validate(space);
}
}
ut_a(fil_system.n_open == n_open);
ut_list_validate(fil_system.LRU);
for (fil_node = UT_LIST_GET_FIRST(fil_system.LRU);
fil_node != 0;
fil_node = UT_LIST_GET_NEXT(LRU, fil_node)) {
ut_a(fil_node->n_pending == 0);
ut_a(!fil_node->being_extended);
ut_a(fil_node->is_open());
ut_a(fil_space_belongs_in_lru(fil_node->space));
}
mutex_exit(&fil_system.mutex);
return(true);
}
/********************************************************************//**
Returns true if file address is undefined.
@return true if undefined */
bool
fil_addr_is_null(
/*=============*/
fil_addr_t addr) /*!< in: address */
{
return(addr.page == FIL_NULL);
}
/********************************************************************//**
Get the predecessor of a file page.
@return FIL_PAGE_PREV */
ulint
fil_page_get_prev(
/*==============*/
const byte* page) /*!< in: file page */
{
return(mach_read_from_4(page + FIL_PAGE_PREV));
}
/********************************************************************//**
Get the successor of a file page.
@return FIL_PAGE_NEXT */
ulint
fil_page_get_next(
/*==============*/
const byte* page) /*!< in: file page */
{
return(mach_read_from_4(page + FIL_PAGE_NEXT));
}
/*********************************************************************//**
Sets the file page type. */
void
fil_page_set_type(
/*==============*/
byte* page, /*!< in/out: file page */
ulint type) /*!< in: type */
{
ut_ad(page);
mach_write_to_2(page + FIL_PAGE_TYPE, type);
}
/********************************************************************//**
Delete the tablespace file and any related files like .cfg.
This should not be called for temporary tables.
@param[in] ibd_filepath File path of the IBD tablespace */
void
fil_delete_file(
/*============*/
const char* ibd_filepath)
{
/* Force a delete of any stale .ibd files that are lying around. */
ib::info() << "Deleting " << ibd_filepath;
os_file_delete_if_exists(innodb_data_file_key, ibd_filepath, NULL);
char* cfg_filepath = fil_make_filepath(
ibd_filepath, NULL, CFG, false);
if (cfg_filepath != NULL) {
os_file_delete_if_exists(
innodb_data_file_key, cfg_filepath, NULL);
ut_free(cfg_filepath);
}
}
/** Generate redo log for swapping two .ibd files
@param[in] old_table old table
@param[in] new_table new table
@param[in] tmp_name temporary table name
@param[in,out] mtr mini-transaction
@return innodb error code */
dberr_t
fil_mtr_rename_log(
const dict_table_t* old_table,
const dict_table_t* new_table,
const char* tmp_name,
mtr_t* mtr)
{
ut_ad(old_table->space != fil_system.temp_space);
ut_ad(new_table->space != fil_system.temp_space);
ut_ad(old_table->space->id == old_table->space_id);
ut_ad(new_table->space->id == new_table->space_id);
/* If neither table is file-per-table,
there will be no renaming of files. */
if (!old_table->space_id && !new_table->space_id) {
return(DB_SUCCESS);
}
const bool has_data_dir = DICT_TF_HAS_DATA_DIR(old_table->flags);
if (old_table->space_id) {
char* tmp_path = fil_make_filepath(
has_data_dir ? old_table->data_dir_path : NULL,
tmp_name, IBD, has_data_dir);
if (tmp_path == NULL) {
return(DB_OUT_OF_MEMORY);
}
const char* old_path = old_table->space->chain.start->name;
/* Temp filepath must not exist. */
dberr_t err = fil_rename_tablespace_check(
old_path, tmp_path, !old_table->space);
if (err != DB_SUCCESS) {
ut_free(tmp_path);
return(err);
}
fil_name_write_rename_low(
old_table->space_id, 0, old_path, tmp_path, mtr);
ut_free(tmp_path);
}
if (new_table->space_id) {
const char* new_path = new_table->space->chain.start->name;
char* old_path = fil_make_filepath(
has_data_dir ? old_table->data_dir_path : NULL,
old_table->name.m_name, IBD, has_data_dir);
/* Destination filepath must not exist unless this ALTER
TABLE starts and ends with a file_per-table tablespace. */
if (!old_table->space_id) {
dberr_t err = fil_rename_tablespace_check(
new_path, old_path, !new_table->space);
if (err != DB_SUCCESS) {
ut_free(old_path);
return(err);
}
}
fil_name_write_rename_low(
new_table->space_id, 0, new_path, old_path, mtr);
ut_free(old_path);
}
return DB_SUCCESS;
}
#ifdef UNIV_DEBUG
/** Check that a tablespace is valid for mtr_commit().
@param[in] space persistent tablespace that has been changed */
static
void
fil_space_validate_for_mtr_commit(
const fil_space_t* space)
{
ut_ad(!mutex_own(&fil_system.mutex));
ut_ad(space != NULL);
ut_ad(space->purpose == FIL_TYPE_TABLESPACE);
ut_ad(!is_predefined_tablespace(space->id));
/* We are serving mtr_commit(). While there is an active
mini-transaction, we should have !space->stop_new_ops. This is
guaranteed by meta-data locks or transactional locks, or
dict_operation_lock (X-lock in DROP, S-lock in purge).
However, a file I/O thread can invoke change buffer merge
while fil_check_pending_operations() is waiting for operations
to quiesce. This is not a problem, because
ibuf_merge_or_delete_for_page() would call
fil_space_acquire() before mtr_start() and
fil_space_t::release() after mtr_commit(). This is why
n_pending_ops should not be zero if stop_new_ops is set. */
ut_ad(!space->stop_new_ops
|| space->is_being_truncated /* fil_truncate_prepare() */
|| space->referenced());
}
#endif /* UNIV_DEBUG */
/** Write a MLOG_FILE_NAME record for a persistent tablespace.
@param[in] space tablespace
@param[in,out] mtr mini-transaction */
static
void
fil_names_write(
const fil_space_t* space,
mtr_t* mtr)
{
ut_ad(UT_LIST_GET_LEN(space->chain) == 1);
fil_name_write(space, 0, UT_LIST_GET_FIRST(space->chain), mtr);
}
/** Note that a non-predefined persistent tablespace has been modified
by redo log.
@param[in,out] space tablespace */
void
fil_names_dirty(
fil_space_t* space)
{
ut_ad(log_mutex_own());
ut_ad(recv_recovery_is_on());
ut_ad(log_sys.lsn != 0);
ut_ad(space->max_lsn == 0);
ut_d(fil_space_validate_for_mtr_commit(space));
UT_LIST_ADD_LAST(fil_system.named_spaces, space);
space->max_lsn = log_sys.lsn;
}
/** Write MLOG_FILE_NAME records when a non-predefined persistent
tablespace was modified for the first time since the latest
fil_names_clear().
@param[in,out] space tablespace
@param[in,out] mtr mini-transaction */
void
fil_names_dirty_and_write(
fil_space_t* space,
mtr_t* mtr)
{
ut_ad(log_mutex_own());
ut_d(fil_space_validate_for_mtr_commit(space));
ut_ad(space->max_lsn == log_sys.lsn);
UT_LIST_ADD_LAST(fil_system.named_spaces, space);
fil_names_write(space, mtr);
DBUG_EXECUTE_IF("fil_names_write_bogus",
{
char bogus_name[] = "./test/bogus file.ibd";
os_normalize_path(bogus_name);
fil_name_write(
SRV_LOG_SPACE_FIRST_ID, 0,
bogus_name, mtr);
});
}
/** On a log checkpoint, reset fil_names_dirty_and_write() flags
and write out MLOG_FILE_NAME and MLOG_CHECKPOINT if needed.
@param[in] lsn checkpoint LSN
@param[in] do_write whether to always write MLOG_CHECKPOINT
@return whether anything was written to the redo log
@retval false if no flags were set and nothing written
@retval true if anything was written to the redo log */
bool
fil_names_clear(
lsn_t lsn,
bool do_write)
{
mtr_t mtr;
ulint mtr_checkpoint_size = LOG_CHECKPOINT_FREE_PER_THREAD;
DBUG_EXECUTE_IF(
"increase_mtr_checkpoint_size",
mtr_checkpoint_size = 75 * 1024;
);
ut_ad(log_mutex_own());
if (log_sys.append_on_checkpoint) {
mtr_write_log(log_sys.append_on_checkpoint);
do_write = true;
}
mtr.start();
for (fil_space_t* space = UT_LIST_GET_FIRST(fil_system.named_spaces);
space != NULL; ) {
fil_space_t* next = UT_LIST_GET_NEXT(named_spaces, space);
ut_ad(space->max_lsn > 0);
if (space->max_lsn < lsn) {
/* The tablespace was last dirtied before the
checkpoint LSN. Remove it from the list, so
that if the tablespace is not going to be
modified any more, subsequent checkpoints will
avoid calling fil_names_write() on it. */
space->max_lsn = 0;
UT_LIST_REMOVE(fil_system.named_spaces, space);
}
/* max_lsn is the last LSN where fil_names_dirty_and_write()
was called. If we kept track of "min_lsn" (the first LSN
where max_lsn turned nonzero), we could avoid the
fil_names_write() call if min_lsn > lsn. */
fil_names_write(space, &mtr);
do_write = true;
const mtr_buf_t* mtr_log = mtr_get_log(&mtr);
/** If the mtr buffer size exceeds the size of
LOG_CHECKPOINT_FREE_PER_THREAD then commit the multi record
mini-transaction, start the new mini-transaction to
avoid the parsing buffer overflow error during recovery. */
if (mtr_log->size() > mtr_checkpoint_size) {
ut_ad(mtr_log->size() < (RECV_PARSING_BUF_SIZE / 2));
mtr.commit_checkpoint(lsn, false);
mtr.start();
}
space = next;
}
if (do_write) {
mtr.commit_checkpoint(lsn, true);
} else {
ut_ad(!mtr.has_modifications());
}
return(do_write);
}
/** Truncate a single-table tablespace. The tablespace must be cached
in the memory cache.
@param space_id space id
@param dir_path directory path
@param tablename the table name in the usual
databasename/tablename format of InnoDB
@param flags tablespace flags
@param trunc_to_default truncate to default size if tablespace
is being newly re-initialized.
@return DB_SUCCESS or error */
dberr_t
truncate_t::truncate(
/*=================*/
ulint space_id,
const char* dir_path,
const char* tablename,
ulint flags,
bool trunc_to_default)
{
dberr_t err = DB_SUCCESS;
char* path;
ut_a(!is_system_tablespace(space_id));
if (FSP_FLAGS_HAS_DATA_DIR(flags)) {
ut_ad(dir_path != NULL);
path = fil_make_filepath(dir_path, tablename, IBD, true);
} else {
path = fil_make_filepath(NULL, tablename, IBD, false);
}
if (path == NULL) {
return(DB_OUT_OF_MEMORY);
}
mutex_enter(&fil_system.mutex);
fil_space_t* space = fil_space_get_by_id(space_id);
/* The following code must change when InnoDB supports
multiple datafiles per tablespace. */
ut_a(UT_LIST_GET_LEN(space->chain) == 1);
fil_node_t* node = UT_LIST_GET_FIRST(space->chain);
if (trunc_to_default) {
space->size = node->size = FIL_IBD_FILE_INITIAL_SIZE;
}
const bool already_open = node->is_open();
if (!already_open) {
bool ret;
node->handle = os_file_create_simple_no_error_handling(
innodb_data_file_key, path, OS_FILE_OPEN,
OS_FILE_READ_WRITE,
space->purpose != FIL_TYPE_TEMPORARY
&& srv_read_only_mode, &ret);
if (!ret) {
ib::error() << "Failed to open tablespace file "
<< path << ".";
ut_free(path);
return(DB_ERROR);
}
ut_a(node->is_open());
}
os_offset_t trunc_size = trunc_to_default
? FIL_IBD_FILE_INITIAL_SIZE
: space->size;
const bool success = os_file_truncate(
path, node->handle, trunc_size << srv_page_size_shift);
if (!success) {
ib::error() << "Cannot truncate file " << path
<< " in TRUNCATE TABLESPACE.";
err = DB_ERROR;
}
space->stop_new_ops = false;
/* If we opened the file in this function, close it. */
if (!already_open) {
bool closed = os_file_close(node->handle);
if (!closed) {
ib::error() << "Failed to close tablespace file "
<< path << ".";
err = DB_ERROR;
} else {
node->handle = OS_FILE_CLOSED;
}
}
mutex_exit(&fil_system.mutex);
ut_free(path);
return(err);
}
/* Unit Tests */
#ifdef UNIV_ENABLE_UNIT_TEST_MAKE_FILEPATH
#define MF fil_make_filepath
#define DISPLAY ib::info() << path
void
test_make_filepath()
{
char* path;
const char* long_path =
"this/is/a/very/long/path/including/a/very/"
"looooooooooooooooooooooooooooooooooooooooooooooooo"
"oooooooooooooooooooooooooooooooooooooooooooooooooo"
"oooooooooooooooooooooooooooooooooooooooooooooooooo"
"oooooooooooooooooooooooooooooooooooooooooooooooooo"
"oooooooooooooooooooooooooooooooooooooooooooooooooo"
"oooooooooooooooooooooooooooooooooooooooooooooooooo"
"oooooooooooooooooooooooooooooooooooooooooooooooooo"
"oooooooooooooooooooooooooooooooooooooooooooooooooo"
"oooooooooooooooooooooooooooooooooooooooooooooooooo"
"oooooooooooooooooooooooooooooooooooooooooooooooong"
"/folder/name";
path = MF("/this/is/a/path/with/a/filename", NULL, IBD, false); DISPLAY;
path = MF("/this/is/a/path/with/a/filename", NULL, ISL, false); DISPLAY;
path = MF("/this/is/a/path/with/a/filename", NULL, CFG, false); DISPLAY;
path = MF("/this/is/a/path/with/a/filename.ibd", NULL, IBD, false); DISPLAY;
path = MF("/this/is/a/path/with/a/filename.ibd", NULL, IBD, false); DISPLAY;
path = MF("/this/is/a/path/with/a/filename.dat", NULL, IBD, false); DISPLAY;
path = MF(NULL, "tablespacename", NO_EXT, false); DISPLAY;
path = MF(NULL, "tablespacename", IBD, false); DISPLAY;
path = MF(NULL, "dbname/tablespacename", NO_EXT, false); DISPLAY;
path = MF(NULL, "dbname/tablespacename", IBD, false); DISPLAY;
path = MF(NULL, "dbname/tablespacename", ISL, false); DISPLAY;
path = MF(NULL, "dbname/tablespacename", CFG, false); DISPLAY;
path = MF(NULL, "dbname\\tablespacename", NO_EXT, false); DISPLAY;
path = MF(NULL, "dbname\\tablespacename", IBD, false); DISPLAY;
path = MF("/this/is/a/path", "dbname/tablespacename", IBD, false); DISPLAY;
path = MF("/this/is/a/path", "dbname/tablespacename", IBD, true); DISPLAY;
path = MF("./this/is/a/path", "dbname/tablespacename.ibd", IBD, true); DISPLAY;
path = MF("this\\is\\a\\path", "dbname/tablespacename", IBD, true); DISPLAY;
path = MF("/this/is/a/path", "dbname\\tablespacename", IBD, true); DISPLAY;
path = MF(long_path, NULL, IBD, false); DISPLAY;
path = MF(long_path, "tablespacename", IBD, false); DISPLAY;
path = MF(long_path, "tablespacename", IBD, true); DISPLAY;
}
#endif /* UNIV_ENABLE_UNIT_TEST_MAKE_FILEPATH */
/* @} */
/** Return the next fil_space_t.
Once started, the caller must keep calling this until it returns NULL.
fil_space_t::acquire() and fil_space_t::release() are invoked here which
blocks a concurrent operation from dropping the tablespace.
@param[in] prev_space Pointer to the previous fil_space_t.
If NULL, use the first fil_space_t on fil_system.space_list.
@return pointer to the next fil_space_t.
@retval NULL if this was the last*/
fil_space_t*
fil_space_next(fil_space_t* prev_space)
{
fil_space_t* space=prev_space;
mutex_enter(&fil_system.mutex);
if (!space) {
space = UT_LIST_GET_FIRST(fil_system.space_list);
} else {
ut_a(space->referenced());
/* Move on to the next fil_space_t */
space->release();
space = UT_LIST_GET_NEXT(space_list, space);
}
/* Skip spaces that are being created by
fil_ibd_create(), or dropped, or !tablespace. */
while (space != NULL
&& (UT_LIST_GET_LEN(space->chain) == 0
|| space->is_stopping()
|| space->purpose != FIL_TYPE_TABLESPACE)) {
space = UT_LIST_GET_NEXT(space_list, space);
}
if (space != NULL) {
space->acquire();
}
mutex_exit(&fil_system.mutex);
return(space);
}
/**
Remove space from key rotation list if there are no more
pending operations.
@param[in,out] space Tablespace */
static
void
fil_space_remove_from_keyrotation(fil_space_t* space)
{
ut_ad(mutex_own(&fil_system.mutex));
ut_ad(space);
if (!space->referenced() && space->is_in_rotation_list) {
space->is_in_rotation_list = false;
ut_a(!fil_system.rotation_list.empty());
fil_system.rotation_list.remove(*space);
}
}
/** Return the next fil_space_t from key rotation list.
Once started, the caller must keep calling this until it returns NULL.
fil_space_t::acquire() and fil_space_t::release() are invoked here which
blocks a concurrent operation from dropping the tablespace.
@param[in] prev_space Pointer to the previous fil_space_t.
If NULL, use the first fil_space_t on fil_system.space_list.
@param[in] recheck recheck of the tablespace is needed or
still encryption thread does write page0 for it
@param[in] key_version key version of the key state thread
@return pointer to the next fil_space_t.
@retval NULL if this was the last */
fil_space_t *fil_system_t::keyrotate_next(fil_space_t *prev_space,
bool recheck, uint key_version)
{
mutex_enter(&fil_system.mutex);
/* If one of the encryption threads already started the encryption
of the table then don't remove the unencrypted spaces from rotation list
If there is a change in innodb_encrypt_tables variables value then
don't remove the last processed tablespace from the rotation list. */
const bool remove= (!recheck || prev_space->crypt_data) &&
!key_version == !srv_encrypt_tables;
sized_ilist<fil_space_t, rotation_list_tag_t>::iterator it=
prev_space ? prev_space : fil_system.rotation_list.end();
if (it == fil_system.rotation_list.end())
it= fil_system.rotation_list.begin();
else
{
/* Move on to the next fil_space_t */
prev_space->release();
++it;
while (it != fil_system.rotation_list.end() &&
(UT_LIST_GET_LEN(it->chain) == 0 || it->is_stopping()))
++it;
if (remove)
fil_space_remove_from_keyrotation(prev_space);
}
fil_space_t *space= it == fil_system.rotation_list.end() ? NULL : &*it;
if (space)
space->acquire();
mutex_exit(&fil_system.mutex);
return space;
}
/** Determine the block size of the data file.
@param[in] space tablespace
@param[in] offset page number
@return block size */
UNIV_INTERN
ulint
fil_space_get_block_size(const fil_space_t* space, unsigned offset)
{
ulint block_size = 512;
for (fil_node_t* node = UT_LIST_GET_FIRST(space->chain);
node != NULL;
node = UT_LIST_GET_NEXT(chain, node)) {
block_size = node->block_size;
if (node->size > offset) {
ut_ad(node->size <= 0xFFFFFFFFU);
break;
}
offset -= static_cast<unsigned>(node->size);
}
/* Currently supporting block size up to 4K,
fall back to default if bigger requested. */
if (block_size > 4096) {
block_size = 512;
}
return block_size;
}
/*******************************************************************//**
Returns the table space by a given id, NULL if not found. */
fil_space_t*
fil_space_found_by_id(
/*==================*/
ulint id) /*!< in: space id */
{
fil_space_t* space = NULL;
mutex_enter(&fil_system.mutex);
space = fil_space_get_by_id(id);
/* Not found if space is being deleted */
if (space && space->stop_new_ops) {
space = NULL;
}
mutex_exit(&fil_system.mutex);
return space;
}
/**
Get should we punch hole to tablespace.
@param[in] node File node
@return true, if punch hole should be tried, false if not. */
bool
fil_node_should_punch_hole(
const fil_node_t* node)
{
return (node->space->punch_hole);
}
/**
Set punch hole to tablespace to given value.
@param[in] node File node
@param[in] val value to be set. */
void
fil_space_set_punch_hole(
fil_node_t* node,
bool val)
{
node->space->punch_hole = val;
}
|