summaryrefslogtreecommitdiff
path: root/fs/jfs/jfs_dmap.c
blob: adb9f05093b2f2aee803c3f4116f34d4afca8cf1 (plain)
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
/*
 *   Copyright (C) International Business Machines Corp., 2000-2004
 *
 *   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; either version 2 of the License, or 
 *   (at your option) any later version.
 * 
 *   This program is distributed in the hope that it will be useful,
 *   but WITHOUT ANY WARRANTY;  without even the implied warranty of
 *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See
 *   the GNU General Public License for more details.
 *
 *   You should have received a copy of the GNU General Public License
 *   along with this program;  if not, write to the Free Software 
 *   Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
 */

#include <linux/fs.h>
#include "jfs_incore.h"
#include "jfs_superblock.h"
#include "jfs_dmap.h"
#include "jfs_imap.h"
#include "jfs_lock.h"
#include "jfs_metapage.h"
#include "jfs_debug.h"

/*
 *	SERIALIZATION of the Block Allocation Map.
 *
 *	the working state of the block allocation map is accessed in
 *	two directions:
 *	
 *	1) allocation and free requests that start at the dmap
 *	   level and move up through the dmap control pages (i.e.
 *	   the vast majority of requests).
 * 
 * 	2) allocation requests that start at dmap control page
 *	   level and work down towards the dmaps.
 *	
 *	the serialization scheme used here is as follows. 
 *
 *	requests which start at the bottom are serialized against each 
 *	other through buffers and each requests holds onto its buffers 
 *	as it works it way up from a single dmap to the required level 
 *	of dmap control page.
 *	requests that start at the top are serialized against each other
 *	and request that start from the bottom by the multiple read/single
 *	write inode lock of the bmap inode. requests starting at the top
 *	take this lock in write mode while request starting at the bottom
 *	take the lock in read mode.  a single top-down request may proceed
 *	exclusively while multiple bottoms-up requests may proceed 
 * 	simultaneously (under the protection of busy buffers).
 *	
 *	in addition to information found in dmaps and dmap control pages,
 *	the working state of the block allocation map also includes read/
 *	write information maintained in the bmap descriptor (i.e. total
 *	free block count, allocation group level free block counts).
 *	a single exclusive lock (BMAP_LOCK) is used to guard this information
 *	in the face of multiple-bottoms up requests.
 *	(lock ordering: IREAD_LOCK, BMAP_LOCK);
 *	
 *	accesses to the persistent state of the block allocation map (limited
 *	to the persistent bitmaps in dmaps) is guarded by (busy) buffers.
 */

#define BMAP_LOCK_INIT(bmp)	init_MUTEX(&bmp->db_bmaplock)
#define BMAP_LOCK(bmp)		down(&bmp->db_bmaplock)
#define BMAP_UNLOCK(bmp)	up(&bmp->db_bmaplock)

/*
 * forward references
 */
static void dbAllocBits(struct bmap * bmp, struct dmap * dp, s64 blkno,
			int nblocks);
static void dbSplit(dmtree_t * tp, int leafno, int splitsz, int newval);
static int dbBackSplit(dmtree_t * tp, int leafno);
static int dbJoin(dmtree_t * tp, int leafno, int newval);
static void dbAdjTree(dmtree_t * tp, int leafno, int newval);
static int dbAdjCtl(struct bmap * bmp, s64 blkno, int newval, int alloc,
		    int level);
static int dbAllocAny(struct bmap * bmp, s64 nblocks, int l2nb, s64 * results);
static int dbAllocNext(struct bmap * bmp, struct dmap * dp, s64 blkno,
		       int nblocks);
static int dbAllocNear(struct bmap * bmp, struct dmap * dp, s64 blkno,
		       int nblocks,
		       int l2nb, s64 * results);
static int dbAllocDmap(struct bmap * bmp, struct dmap * dp, s64 blkno,
		       int nblocks);
static int dbAllocDmapLev(struct bmap * bmp, struct dmap * dp, int nblocks,
			  int l2nb,
			  s64 * results);
static int dbAllocAG(struct bmap * bmp, int agno, s64 nblocks, int l2nb,
		     s64 * results);
static int dbAllocCtl(struct bmap * bmp, s64 nblocks, int l2nb, s64 blkno,
		      s64 * results);
static int dbExtend(struct inode *ip, s64 blkno, s64 nblocks, s64 addnblocks);
static int dbFindBits(u32 word, int l2nb);
static int dbFindCtl(struct bmap * bmp, int l2nb, int level, s64 * blkno);
static int dbFindLeaf(dmtree_t * tp, int l2nb, int *leafidx);
static int dbFreeBits(struct bmap * bmp, struct dmap * dp, s64 blkno,
		      int nblocks);
static int dbFreeDmap(struct bmap * bmp, struct dmap * dp, s64 blkno,
		      int nblocks);
static int dbMaxBud(u8 * cp);
s64 dbMapFileSizeToMapSize(struct inode *ipbmap);
static int blkstol2(s64 nb);

static int cntlz(u32 value);
static int cnttz(u32 word);

static int dbAllocDmapBU(struct bmap * bmp, struct dmap * dp, s64 blkno,
			 int nblocks);
static int dbInitDmap(struct dmap * dp, s64 blkno, int nblocks);
static int dbInitDmapTree(struct dmap * dp);
static int dbInitTree(struct dmaptree * dtp);
static int dbInitDmapCtl(struct dmapctl * dcp, int level, int i);
static int dbGetL2AGSize(s64 nblocks);

/*
 *	buddy table
 *
 * table used for determining buddy sizes within characters of 
 * dmap bitmap words.  the characters themselves serve as indexes
 * into the table, with the table elements yielding the maximum
 * binary buddy of free bits within the character.
 */
static const s8 budtab[256] = {
	3, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
	2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
	2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
	2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
	2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
	2, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0,
	2, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0,
	2, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0,
	2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
	2, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0,
	2, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0,
	2, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0,
	2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
	2, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0,
	2, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0,
	2, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, -1
};


/*
 * NAME:    	dbMount()
 *
 * FUNCTION:	initializate the block allocation map.
 *
 *		memory is allocated for the in-core bmap descriptor and
 *		the in-core descriptor is initialized from disk.
 *
 * PARAMETERS:
 *      ipbmap	-  pointer to in-core inode for the block map.
 *
 * RETURN VALUES:
 *      0	- success
 *      -ENOMEM	- insufficient memory
 *      -EIO	- i/o error
 */
int dbMount(struct inode *ipbmap)
{
	struct bmap *bmp;
	struct dbmap_disk *dbmp_le;
	struct metapage *mp;
	int i;

	/*
	 * allocate/initialize the in-memory bmap descriptor
	 */
	/* allocate memory for the in-memory bmap descriptor */
	bmp = kmalloc(sizeof(struct bmap), GFP_KERNEL);
	if (bmp == NULL)
		return -ENOMEM;

	/* read the on-disk bmap descriptor. */
	mp = read_metapage(ipbmap,
			   BMAPBLKNO << JFS_SBI(ipbmap->i_sb)->l2nbperpage,
			   PSIZE, 0);
	if (mp == NULL) {
		kfree(bmp);
		return -EIO;
	}

	/* copy the on-disk bmap descriptor to its in-memory version. */
	dbmp_le = (struct dbmap_disk *) mp->data;
	bmp->db_mapsize = le64_to_cpu(dbmp_le->dn_mapsize);
	bmp->db_nfree = le64_to_cpu(dbmp_le->dn_nfree);
	bmp->db_l2nbperpage = le32_to_cpu(dbmp_le->dn_l2nbperpage);
	bmp->db_numag = le32_to_cpu(dbmp_le->dn_numag);
	bmp->db_maxlevel = le32_to_cpu(dbmp_le->dn_maxlevel);
	bmp->db_maxag = le32_to_cpu(dbmp_le->dn_maxag);
	bmp->db_agpref = le32_to_cpu(dbmp_le->dn_agpref);
	bmp->db_aglevel = le32_to_cpu(dbmp_le->dn_aglevel);
	bmp->db_agheigth = le32_to_cpu(dbmp_le->dn_agheigth);
	bmp->db_agwidth = le32_to_cpu(dbmp_le->dn_agwidth);
	bmp->db_agstart = le32_to_cpu(dbmp_le->dn_agstart);
	bmp->db_agl2size = le32_to_cpu(dbmp_le->dn_agl2size);
	for (i = 0; i < MAXAG; i++)
		bmp->db_agfree[i] = le64_to_cpu(dbmp_le->dn_agfree[i]);
	bmp->db_agsize = le64_to_cpu(dbmp_le->dn_agsize);
	bmp->db_maxfreebud = dbmp_le->dn_maxfreebud;

	/* release the buffer. */
	release_metapage(mp);

	/* bind the bmap inode and the bmap descriptor to each other. */
	bmp->db_ipbmap = ipbmap;
	JFS_SBI(ipbmap->i_sb)->bmap = bmp;

	memset(bmp->db_active, 0, sizeof(bmp->db_active));

	/*
	 * allocate/initialize the bmap lock
	 */
	BMAP_LOCK_INIT(bmp);

	return (0);
}


/*
 * NAME:    	dbUnmount()
 *
 * FUNCTION:	terminate the block allocation map in preparation for
 *		file system unmount.
 *
 * 		the in-core bmap descriptor is written to disk and
 *		the memory for this descriptor is freed.
 *
 * PARAMETERS:
 *      ipbmap	-  pointer to in-core inode for the block map.
 *
 * RETURN VALUES:
 *      0	- success
 *      -EIO	- i/o error
 */
int dbUnmount(struct inode *ipbmap, int mounterror)
{
	struct bmap *bmp = JFS_SBI(ipbmap->i_sb)->bmap;

	if (!(mounterror || isReadOnly(ipbmap)))
		dbSync(ipbmap);

	/*
	 * Invalidate the page cache buffers
	 */
	truncate_inode_pages(ipbmap->i_mapping, 0);

	/* free the memory for the in-memory bmap. */
	kfree(bmp);

	return (0);
}

/*
 *	dbSync()
 */
int dbSync(struct inode *ipbmap)
{
	struct dbmap_disk *dbmp_le;
	struct bmap *bmp = JFS_SBI(ipbmap->i_sb)->bmap;
	struct metapage *mp;
	int i;

	/*
	 * write bmap global control page
	 */
	/* get the buffer for the on-disk bmap descriptor. */
	mp = read_metapage(ipbmap,
			   BMAPBLKNO << JFS_SBI(ipbmap->i_sb)->l2nbperpage,
			   PSIZE, 0);
	if (mp == NULL) {
		jfs_err("dbSync: read_metapage failed!");
		return -EIO;
	}
	/* copy the in-memory version of the bmap to the on-disk version */
	dbmp_le = (struct dbmap_disk *) mp->data;
	dbmp_le->dn_mapsize = cpu_to_le64(bmp->db_mapsize);
	dbmp_le->dn_nfree = cpu_to_le64(bmp->db_nfree);
	dbmp_le->dn_l2nbperpage = cpu_to_le32(bmp->db_l2nbperpage);
	dbmp_le->dn_numag = cpu_to_le32(bmp->db_numag);
	dbmp_le->dn_maxlevel = cpu_to_le32(bmp->db_maxlevel);
	dbmp_le->dn_maxag = cpu_to_le32(bmp->db_maxag);
	dbmp_le->dn_agpref = cpu_to_le32(bmp->db_agpref);
	dbmp_le->dn_aglevel = cpu_to_le32(bmp->db_aglevel);
	dbmp_le->dn_agheigth = cpu_to_le32(bmp->db_agheigth);
	dbmp_le->dn_agwidth = cpu_to_le32(bmp->db_agwidth);
	dbmp_le->dn_agstart = cpu_to_le32(bmp->db_agstart);
	dbmp_le->dn_agl2size = cpu_to_le32(bmp->db_agl2size);
	for (i = 0; i < MAXAG; i++)
		dbmp_le->dn_agfree[i] = cpu_to_le64(bmp->db_agfree[i]);
	dbmp_le->dn_agsize = cpu_to_le64(bmp->db_agsize);
	dbmp_le->dn_maxfreebud = bmp->db_maxfreebud;

	/* write the buffer */
	write_metapage(mp);

	/*
	 * write out dirty pages of bmap
	 */
	filemap_fdatawrite(ipbmap->i_mapping);
	filemap_fdatawait(ipbmap->i_mapping);

	diWriteSpecial(ipbmap, 0);

	return (0);
}


/*
 * NAME:    	dbFree()
 *
 * FUNCTION:	free the specified block range from the working block
 *		allocation map.
 *
 *		the blocks will be free from the working map one dmap
 *		at a time.
 *
 * PARAMETERS:
 *      ip	-  pointer to in-core inode;
 *      blkno	-  starting block number to be freed.
 *      nblocks	-  number of blocks to be freed.
 *
 * RETURN VALUES:
 *      0	- success
 *      -EIO	- i/o error
 */
int dbFree(struct inode *ip, s64 blkno, s64 nblocks)
{
	struct metapage *mp;
	struct dmap *dp;
	int nb, rc;
	s64 lblkno, rem;
	struct inode *ipbmap = JFS_SBI(ip->i_sb)->ipbmap;
	struct bmap *bmp = JFS_SBI(ip->i_sb)->bmap;

	IREAD_LOCK(ipbmap);

	/* block to be freed better be within the mapsize. */
	if (unlikely((blkno == 0) || (blkno + nblocks > bmp->db_mapsize))) {
		IREAD_UNLOCK(ipbmap);
		printk(KERN_ERR "blkno = %Lx, nblocks = %Lx\n",
		       (unsigned long long) blkno,
		       (unsigned long long) nblocks);
		jfs_error(ip->i_sb,
			  "dbFree: block to be freed is outside the map");
		return -EIO;
	}

	/*
	 * free the blocks a dmap at a time.
	 */
	mp = NULL;
	for (rem = nblocks; rem > 0; rem -= nb, blkno += nb) {
		/* release previous dmap if any */
		if (mp) {
			write_metapage(mp);
		}

		/* get the buffer for the current dmap. */
		lblkno = BLKTODMAP(blkno, bmp->db_l2nbperpage);
		mp = read_metapage(ipbmap, lblkno, PSIZE, 0);
		if (mp == NULL) {
			IREAD_UNLOCK(ipbmap);
			return -EIO;
		}
		dp = (struct dmap *) mp->data;

		/* determine the number of blocks to be freed from
		 * this dmap.
		 */
		nb = min(rem, BPERDMAP - (blkno & (BPERDMAP - 1)));

		/* free the blocks. */
		if ((rc = dbFreeDmap(bmp, dp, blkno, nb))) {
			jfs_error(ip->i_sb, "dbFree: error in block map\n");
			release_metapage(mp);
			IREAD_UNLOCK(ipbmap);
			return (rc);
		}
	}

	/* write the last buffer. */
	write_metapage(mp);

	IREAD_UNLOCK(ipbmap);

	return (0);
}


/*
 * NAME:	dbUpdatePMap()
 *
 * FUNCTION:    update the allocation state (free or allocate) of the
 *		specified block range in the persistent block allocation map.
 *		
 *		the blocks will be updated in the persistent map one
 *		dmap at a time.
 *
 * PARAMETERS:
 *      ipbmap	-  pointer to in-core inode for the block map.
 *      free	- TRUE if block range is to be freed from the persistent
 *		  map; FALSE if it is to   be allocated.
 *      blkno	-  starting block number of the range.
 *      nblocks	-  number of contiguous blocks in the range.
 *      tblk	-  transaction block;
 *
 * RETURN VALUES:
 *      0	- success
 *      -EIO	- i/o error
 */
int
dbUpdatePMap(struct inode *ipbmap,
	     int free, s64 blkno, s64 nblocks, struct tblock * tblk)
{
	int nblks, dbitno, wbitno, rbits;
	int word, nbits, nwords;
	struct bmap *bmp = JFS_SBI(ipbmap->i_sb)->bmap;
	s64 lblkno, rem, lastlblkno;
	u32 mask;
	struct dmap *dp;
	struct metapage *mp;
	struct jfs_log *log;
	int lsn, difft, diffp;
	unsigned long flags;

	/* the blocks better be within the mapsize. */
	if (blkno + nblocks > bmp->db_mapsize) {
		printk(KERN_ERR "blkno = %Lx, nblocks = %Lx\n",
		       (unsigned long long) blkno,
		       (unsigned long long) nblocks);
		jfs_error(ipbmap->i_sb,
			  "dbUpdatePMap: blocks are outside the map");
		return -EIO;
	}

	/* compute delta of transaction lsn from log syncpt */
	lsn = tblk->lsn;
	log = (struct jfs_log *) JFS_SBI(tblk->sb)->log;
	logdiff(difft, lsn, log);

	/*
	 * update the block state a dmap at a time.
	 */
	mp = NULL;
	lastlblkno = 0;
	for (rem = nblocks; rem > 0; rem -= nblks, blkno += nblks) {
		/* get the buffer for the current dmap. */
		lblkno = BLKTODMAP(blkno, bmp->db_l2nbperpage);
		if (lblkno != lastlblkno) {
			if (mp) {
				write_metapage(mp);
			}

			mp = read_metapage(bmp->db_ipbmap, lblkno, PSIZE,
					   0);
			if (mp == NULL)
				return -EIO;
			metapage_wait_for_io(mp);
		}
		dp = (struct dmap *) mp->data;

		/* determine the bit number and word within the dmap of
		 * the starting block.  also determine how many blocks
		 * are to be updated within this dmap.
		 */
		dbitno = blkno & (BPERDMAP - 1);
		word = dbitno >> L2DBWORD;
		nblks = min(rem, (s64)BPERDMAP - dbitno);

		/* update the bits of the dmap words. the first and last
		 * words may only have a subset of their bits updated. if
		 * this is the case, we'll work against that word (i.e.
		 * partial first and/or last) only in a single pass.  a 
		 * single pass will also be used to update all words that
		 * are to have all their bits updated.
		 */
		for (rbits = nblks; rbits > 0;
		     rbits -= nbits, dbitno += nbits) {
			/* determine the bit number within the word and
			 * the number of bits within the word.
			 */
			wbitno = dbitno & (DBWORD - 1);
			nbits = min(rbits, DBWORD - wbitno);

			/* check if only part of the word is to be updated. */
			if (nbits < DBWORD) {
				/* update (free or allocate) the bits
				 * in this word.
				 */
				mask =
				    (ONES << (DBWORD - nbits) >> wbitno);
				if (free)
					dp->pmap[word] &=
					    cpu_to_le32(~mask);
				else
					dp->pmap[word] |=
					    cpu_to_le32(mask);

				word += 1;
			} else {
				/* one or more words are to have all
				 * their bits updated.  determine how
				 * many words and how many bits.
				 */
				nwords = rbits >> L2DBWORD;
				nbits = nwords << L2DBWORD;

				/* update (free or allocate) the bits
				 * in these words.
				 */
				if (free)
					memset(&dp->pmap[word], 0,
					       nwords * 4);
				else
					memset(&dp->pmap[word], (int) ONES,
					       nwords * 4);

				word += nwords;
			}
		}

		/*
		 * update dmap lsn
		 */
		if (lblkno == lastlblkno)
			continue;

		lastlblkno = lblkno;

		if (mp->lsn != 0) {
			/* inherit older/smaller lsn */
			logdiff(diffp, mp->lsn, log);
			LOGSYNC_LOCK(log, flags);
			if (difft < diffp) {
				mp->lsn = lsn;

				/* move bp after tblock in logsync list */
				list_move(&mp->synclist, &tblk->synclist);
			}

			/* inherit younger/larger clsn */
			logdiff(difft, tblk->clsn, log);
			logdiff(diffp, mp->clsn, log);
			if (difft > diffp)
				mp->clsn = tblk->clsn;
			LOGSYNC_UNLOCK(log, flags);
		} else {
			mp->log = log;
			mp->lsn = lsn;

			/* insert bp after tblock in logsync list */
			LOGSYNC_LOCK(log, flags);

			log->count++;
			list_add(&mp->synclist, &tblk->synclist);

			mp->clsn = tblk->clsn;
			LOGSYNC_UNLOCK(log, flags);
		}
	}

	/* write the last buffer. */
	if (mp) {
		write_metapage(mp);
	}

	return (0);
}


/*
 * NAME:	dbNextAG()
 *
 * FUNCTION:    find the preferred allocation group for new allocations.
 *
 *		Within the allocation groups, we maintain a preferred
 *		allocation group which consists of a group with at least
 *		average free space.  It is the preferred group that we target
 *		new inode allocation towards.  The tie-in between inode
 *		allocation and block allocation occurs as we allocate the
 *		first (data) block of an inode and specify the inode (block)
 *		as the allocation hint for this block.
 *
 *		We try to avoid having more than one open file growing in
 *		an allocation group, as this will lead to fragmentation.
 *		This differs from the old OS/2 method of trying to keep
 *		empty ags around for large allocations.
 *
 * PARAMETERS:
 *      ipbmap	-  pointer to in-core inode for the block map.
 *
 * RETURN VALUES:
 *      the preferred allocation group number.
 */
int dbNextAG(struct inode *ipbmap)
{
	s64 avgfree;
	int agpref;
	s64 hwm = 0;
	int i;
	int next_best = -1;
	struct bmap *bmp = JFS_SBI(ipbmap->i_sb)->bmap;

	BMAP_LOCK(bmp);

	/* determine the average number of free blocks within the ags. */
	avgfree = (u32)bmp->db_nfree / bmp->db_numag;

	/*
	 * if the current preferred ag does not have an active allocator
	 * and has at least average freespace, return it
	 */
	agpref = bmp->db_agpref;
	if ((atomic_read(&bmp->db_active[agpref]) == 0) &&
	    (bmp->db_agfree[agpref] >= avgfree))
		goto unlock;

	/* From the last preferred ag, find the next one with at least
	 * average free space.
	 */
	for (i = 0 ; i < bmp->db_numag; i++, agpref++) {
		if (agpref == bmp->db_numag)
			agpref = 0;

		if (atomic_read(&bmp->db_active[agpref]))
			/* open file is currently growing in this ag */
			continue;
		if (bmp->db_agfree[agpref] >= avgfree) {
			/* Return this one */
			bmp->db_agpref = agpref;
			goto unlock;
		} else if (bmp->db_agfree[agpref] > hwm) {
			/* Less than avg. freespace, but best so far */
			hwm = bmp->db_agfree[agpref];
			next_best = agpref;
		}
	}

	/*
	 * If no inactive ag was found with average freespace, use the
	 * next best
	 */
	if (next_best != -1)
		bmp->db_agpref = next_best;
	/* else leave db_agpref unchanged */
unlock:
	BMAP_UNLOCK(bmp);

	/* return the preferred group.
	 */
	return (bmp->db_agpref);
}

/*
 * NAME:	dbAlloc()
 *
 * FUNCTION:    attempt to allocate a specified number of contiguous free
 *		blocks from the working allocation block map.
 *
 *		the block allocation policy uses hints and a multi-step
 *		approach.
 *
 *	  	for allocation requests smaller than the number of blocks
 *		per dmap, we first try to allocate the new blocks
 *		immediately following the hint.  if these blocks are not
 *		available, we try to allocate blocks near the hint.  if
 *		no blocks near the hint are available, we next try to 
 *		allocate within the same dmap as contains the hint.
 *
 *		if no blocks are available in the dmap or the allocation
 *		request is larger than the dmap size, we try to allocate
 *		within the same allocation group as contains the hint. if
 *		this does not succeed, we finally try to allocate anywhere
 *		within the aggregate.
 *
 *		we also try to allocate anywhere within the aggregate for
 *		for allocation requests larger than the allocation group
 *		size or requests that specify no hint value.
 *
 * PARAMETERS:
 *      ip	-  pointer to in-core inode;
 *      hint	- allocation hint.
 *      nblocks	- number of contiguous blocks in the range.
 *      results	- on successful return, set to the starting block number
 *		  of the newly allocated contiguous range.
 *
 * RETURN VALUES:
 *      0	- success
 *      -ENOSPC	- insufficient disk resources
 *      -EIO	- i/o error
 */
int dbAlloc(struct inode *ip, s64 hint, s64 nblocks, s64 * results)
{
	int rc, agno;
	struct inode *ipbmap = JFS_SBI(ip->i_sb)->ipbmap;
	struct bmap *bmp;
	struct metapage *mp;
	s64 lblkno, blkno;
	struct dmap *dp;
	int l2nb;
	s64 mapSize;
	int writers;

	/* assert that nblocks is valid */
	assert(nblocks > 0);

#ifdef _STILL_TO_PORT
	/* DASD limit check                                     F226941 */
	if (OVER_LIMIT(ip, nblocks))
		return -ENOSPC;
#endif				/* _STILL_TO_PORT */

	/* get the log2 number of blocks to be allocated.
	 * if the number of blocks is not a log2 multiple, 
	 * it will be rounded up to the next log2 multiple.
	 */
	l2nb = BLKSTOL2(nblocks);

	bmp = JFS_SBI(ip->i_sb)->bmap;

//retry:        /* serialize w.r.t.extendfs() */
	mapSize = bmp->db_mapsize;

	/* the hint should be within the map */
	if (hint >= mapSize) {
		jfs_error(ip->i_sb, "dbAlloc: the hint is outside the map");
		return -EIO;
	}

	/* if the number of blocks to be allocated is greater than the
	 * allocation group size, try to allocate anywhere.
	 */
	if (l2nb > bmp->db_agl2size) {
		IWRITE_LOCK(ipbmap);

		rc = dbAllocAny(bmp, nblocks, l2nb, results);

		goto write_unlock;
	}

	/*
	 * If no hint, let dbNextAG recommend an allocation group
	 */
	if (hint == 0)
		goto pref_ag;

	/* we would like to allocate close to the hint.  adjust the
	 * hint to the block following the hint since the allocators
	 * will start looking for free space starting at this point.
	 */
	blkno = hint + 1;

	if (blkno >= bmp->db_mapsize)
		goto pref_ag;

	agno = blkno >> bmp->db_agl2size;

	/* check if blkno crosses over into a new allocation group.
	 * if so, check if we should allow allocations within this
	 * allocation group.
	 */
	if ((blkno & (bmp->db_agsize - 1)) == 0)
		/* check if the AG is currenly being written to.
		 * if so, call dbNextAG() to find a non-busy
		 * AG with sufficient free space.
		 */
		if (atomic_read(&bmp->db_active[agno]))
			goto pref_ag;

	/* check if the allocation request size can be satisfied from a
	 * single dmap.  if so, try to allocate from the dmap containing
	 * the hint using a tiered strategy.
	 */
	if (nblocks <= BPERDMAP) {
		IREAD_LOCK(ipbmap);

		/* get the buffer for the dmap containing the hint.
		 */
		rc = -EIO;
		lblkno = BLKTODMAP(blkno, bmp->db_l2nbperpage);
		mp = read_metapage(ipbmap, lblkno, PSIZE, 0);
		if (mp == NULL)
			goto read_unlock;

		dp = (struct dmap *) mp->data;

		/* first, try to satisfy the allocation request with the
		 * blocks beginning at the hint.
		 */
		if ((rc = dbAllocNext(bmp, dp, blkno, (int) nblocks))
		    != -ENOSPC) {
			if (rc == 0) {
				*results = blkno;
				mark_metapage_dirty(mp);
			}

			release_metapage(mp);
			goto read_unlock;
		}

		writers = atomic_read(&bmp->db_active[agno]);
		if ((writers > 1) ||
		    ((writers == 1) && (JFS_IP(ip)->active_ag != agno))) {
			/*
			 * Someone else is writing in this allocation
			 * group.  To avoid fragmenting, try another ag
			 */
			release_metapage(mp);
			IREAD_UNLOCK(ipbmap);
			goto pref_ag;
		}

		/* next, try to satisfy the allocation request with blocks
		 * near the hint.
		 */
		if ((rc =
		     dbAllocNear(bmp, dp, blkno, (int) nblocks, l2nb, results))
		    != -ENOSPC) {
			if (rc == 0)
				mark_metapage_dirty(mp);

			release_metapage(mp);
			goto read_unlock;
		}

		/* try to satisfy the allocation request with blocks within
		 * the same dmap as the hint.
		 */
		if ((rc = dbAllocDmapLev(bmp, dp, (int) nblocks, l2nb, results))
		    != -ENOSPC) {
			if (rc == 0)
				mark_metapage_dirty(mp);

			release_metapage(mp);
			goto read_unlock;
		}

		release_metapage(mp);
		IREAD_UNLOCK(ipbmap);
	}

	/* try to satisfy the allocation request with blocks within
	 * the same allocation group as the hint.
	 */
	IWRITE_LOCK(ipbmap);
	if ((rc = dbAllocAG(bmp, agno, nblocks, l2nb, results)) != -ENOSPC)
		goto write_unlock;

	IWRITE_UNLOCK(ipbmap);


      pref_ag:
	/*
	 * Let dbNextAG recommend a preferred allocation group
	 */
	agno = dbNextAG(ipbmap);
	IWRITE_LOCK(ipbmap);

	/* Try to allocate within this allocation group.  if that fails, try to
	 * allocate anywhere in the map.
	 */
	if ((rc = dbAllocAG(bmp, agno, nblocks, l2nb, results)) == -ENOSPC)
		rc = dbAllocAny(bmp, nblocks, l2nb, results);

      write_unlock:
	IWRITE_UNLOCK(ipbmap);

	return (rc);

      read_unlock:
	IREAD_UNLOCK(ipbmap);

	return (rc);
}

#ifdef _NOTYET
/*
 * NAME:	dbAllocExact()
 *
 * FUNCTION:    try to allocate the requested extent;
 *
 * PARAMETERS:
 *      ip	- pointer to in-core inode;
 *      blkno	- extent address;
 *      nblocks	- extent length;
 *
 * RETURN VALUES:
 *      0	- success
 *      -ENOSPC	- insufficient disk resources
 *      -EIO	- i/o error
 */
int dbAllocExact(struct inode *ip, s64 blkno, int nblocks)
{
	int rc;
	struct inode *ipbmap = JFS_SBI(ip->i_sb)->ipbmap;
	struct bmap *bmp = JFS_SBI(ip->i_sb)->bmap;
	struct dmap *dp;
	s64 lblkno;
	struct metapage *mp;

	IREAD_LOCK(ipbmap);

	/*
	 * validate extent request:
	 *
	 * note: defragfs policy:
	 *  max 64 blocks will be moved.  
	 *  allocation request size must be satisfied from a single dmap.
	 */
	if (nblocks <= 0 || nblocks > BPERDMAP || blkno >= bmp->db_mapsize) {
		IREAD_UNLOCK(ipbmap);
		return -EINVAL;
	}

	if (nblocks > ((s64) 1 << bmp->db_maxfreebud)) {
		/* the free space is no longer available */
		IREAD_UNLOCK(ipbmap);
		return -ENOSPC;
	}

	/* read in the dmap covering the extent */
	lblkno = BLKTODMAP(blkno, bmp->db_l2nbperpage);
	mp = read_metapage(ipbmap, lblkno, PSIZE, 0);
	if (mp == NULL) {
		IREAD_UNLOCK(ipbmap);
		return -EIO;
	}
	dp = (struct dmap *) mp->data;

	/* try to allocate the requested extent */
	rc = dbAllocNext(bmp, dp, blkno, nblocks);

	IREAD_UNLOCK(ipbmap);

	if (rc == 0)
		mark_metapage_dirty(mp);

	release_metapage(mp);

	return (rc);
}
#endif /* _NOTYET */

/*
 * NAME:	dbReAlloc()
 *
 * FUNCTION:    attempt to extend a current allocation by a specified
 *		number of blocks.
 *
 *		this routine attempts to satisfy the allocation request
 *		by first trying to extend the existing allocation in
 *		place by allocating the additional blocks as the blocks
 *		immediately following the current allocation.  if these
 *		blocks are not available, this routine will attempt to
 *		allocate a new set of contiguous blocks large enough
 *		to cover the existing allocation plus the additional
 *		number of blocks required.
 *
 * PARAMETERS:
 *      ip	    -  pointer to in-core inode requiring allocation.
 *      blkno	    -  starting block of the current allocation.
 *      nblocks	    -  number of contiguous blocks within the current
 *		       allocation.
 *      addnblocks  -  number of blocks to add to the allocation.
 *      results	-      on successful return, set to the starting block number
 *		       of the existing allocation if the existing allocation
 *		       was extended in place or to a newly allocated contiguous
 *		       range if the existing allocation could not be extended
 *		       in place.
 *
 * RETURN VALUES:
 *      0	- success
 *      -ENOSPC	- insufficient disk resources
 *      -EIO	- i/o error
 */
int
dbReAlloc(struct inode *ip,
	  s64 blkno, s64 nblocks, s64 addnblocks, s64 * results)
{
	int rc;

	/* try to extend the allocation in place.
	 */
	if ((rc = dbExtend(ip, blkno, nblocks, addnblocks)) == 0) {
		*results = blkno;
		return (0);
	} else {
		if (rc != -ENOSPC)
			return (rc);
	}

	/* could not extend the allocation in place, so allocate a
	 * new set of blocks for the entire request (i.e. try to get
	 * a range of contiguous blocks large enough to cover the
	 * existing allocation plus the additional blocks.)
	 */
	return (dbAlloc
		(ip, blkno + nblocks - 1, addnblocks + nblocks, results));
}


/*
 * NAME:	dbExtend()
 *
 * FUNCTION:    attempt to extend a current allocation by a specified
 *		number of blocks.
 *
 *		this routine attempts to satisfy the allocation request
 *		by first trying to extend the existing allocation in
 *		place by allocating the additional blocks as the blocks
 *		immediately following the current allocation.
 *
 * PARAMETERS:
 *      ip	    -  pointer to in-core inode requiring allocation.
 *      blkno	    -  starting block of the current allocation.
 *      nblocks	    -  number of contiguous blocks within the current
 *		       allocation.
 *      addnblocks  -  number of blocks to add to the allocation.
 *
 * RETURN VALUES:
 *      0	- success
 *      -ENOSPC	- insufficient disk resources
 *      -EIO	- i/o error
 */
static int dbExtend(struct inode *ip, s64 blkno, s64 nblocks, s64 addnblocks)
{
	struct jfs_sb_info *sbi = JFS_SBI(ip->i_sb);
	s64 lblkno, lastblkno, extblkno;
	uint rel_block;
	struct metapage *mp;
	struct dmap *dp;
	int rc;
	struct inode *ipbmap = sbi->ipbmap;
	struct bmap *bmp;

	/*
	 * We don't want a non-aligned extent to cross a page boundary
	 */
	if (((rel_block = blkno & (sbi->nbperpage - 1))) &&
	    (rel_block + nblocks + addnblocks > sbi->nbperpage))
		return -ENOSPC;

	/* get the last block of the current allocation */
	lastblkno = blkno + nblocks - 1;

	/* determine the block number of the block following
	 * the existing allocation.
	 */
	extblkno = lastblkno + 1;

	IREAD_LOCK(ipbmap);

	/* better be within the file system */
	bmp = sbi->bmap;
	if (lastblkno < 0 || lastblkno >= bmp->db_mapsize) {
		IREAD_UNLOCK(ipbmap);
		jfs_error(ip->i_sb,
			  "dbExtend: the block is outside the filesystem");
		return -EIO;
	}

	/* we'll attempt to extend the current allocation in place by
	 * allocating the additional blocks as the blocks immediately
	 * following the current allocation.  we only try to extend the
	 * current allocation in place if the number of additional blocks
	 * can fit into a dmap, the last block of the current allocation
	 * is not the last block of the file system, and the start of the
	 * inplace extension is not on an allocation group boundary.
	 */
	if (addnblocks > BPERDMAP || extblkno >= bmp->db_mapsize ||
	    (extblkno & (bmp->db_agsize - 1)) == 0) {
		IREAD_UNLOCK(ipbmap);
		return -ENOSPC;
	}

	/* get the buffer for the dmap containing the first block
	 * of the extension.
	 */
	lblkno = BLKTODMAP(extblkno, bmp->db_l2nbperpage);
	mp = read_metapage(ipbmap, lblkno, PSIZE, 0);
	if (mp == NULL) {
		IREAD_UNLOCK(ipbmap);
		return -EIO;
	}

	dp = (struct dmap *) mp->data;

	/* try to allocate the blocks immediately following the
	 * current allocation.
	 */
	rc = dbAllocNext(bmp, dp, extblkno, (int) addnblocks);

	IREAD_UNLOCK(ipbmap);

	/* were we successful ? */
	if (rc == 0)
		write_metapage(mp);
	else
		/* we were not successful */
		release_metapage(mp);


	return (rc);
}


/*
 * NAME:	dbAllocNext()
 *
 * FUNCTION:    attempt to allocate the blocks of the specified block
 *		range within a dmap.
 *
 * PARAMETERS:
 *      bmp	-  pointer to bmap descriptor
 *      dp	-  pointer to dmap.
 *      blkno	-  starting block number of the range.
 *      nblocks	-  number of contiguous free blocks of the range.
 *
 * RETURN VALUES:
 *      0	- success
 *      -ENOSPC	- insufficient disk resources
 *      -EIO	- i/o error
 *
 * serialization: IREAD_LOCK(ipbmap) held on entry/exit;
 */
static int dbAllocNext(struct bmap * bmp, struct dmap * dp, s64 blkno,
		       int nblocks)
{
	int dbitno, word, rembits, nb, nwords, wbitno, nw;
	int l2size;
	s8 *leaf;
	u32 mask;

	if (dp->tree.leafidx != cpu_to_le32(LEAFIND)) {
		jfs_error(bmp->db_ipbmap->i_sb,
			  "dbAllocNext: Corrupt dmap page");
		return -EIO;
	}

	/* pick up a pointer to the leaves of the dmap tree.
	 */
	leaf = dp->tree.stree + le32_to_cpu(dp->tree.leafidx);

	/* determine the bit number and word within the dmap of the
	 * starting block.
	 */
	dbitno = blkno & (BPERDMAP - 1);
	word = dbitno >> L2DBWORD;

	/* check if the specified block range is contained within
	 * this dmap.
	 */
	if (dbitno + nblocks > BPERDMAP)
		return -ENOSPC;

	/* check if the starting leaf indicates that anything
	 * is free.
	 */
	if (leaf[word] == NOFREE)
		return -ENOSPC;

	/* check the dmaps words corresponding to block range to see
	 * if the block range is free.  not all bits of the first and
	 * last words may be contained within the block range.  if this
	 * is the case, we'll work against those words (i.e. partial first
	 * and/or last) on an individual basis (a single pass) and examine
	 * the actual bits to determine if they are free.  a single pass
	 * will be used for all dmap words fully contained within the
	 * specified range.  within this pass, the leaves of the dmap
	 * tree will be examined to determine if the blocks are free. a
	 * single leaf may describe the free space of multiple dmap
	 * words, so we may visit only a subset of the actual leaves
	 * corresponding to the dmap words of the block range.
	 */
	for (rembits = nblocks; rembits > 0; rembits -= nb, dbitno += nb) {
		/* determine the bit number within the word and
		 * the number of bits within the word.
		 */
		wbitno = dbitno & (DBWORD - 1);
		nb = min(rembits, DBWORD - wbitno);

		/* check if only part of the word is to be examined.
		 */
		if (nb < DBWORD) {
			/* check if the bits are free.
			 */
			mask = (ONES << (DBWORD - nb) >> wbitno);
			if ((mask & ~le32_to_cpu(dp->wmap[word])) != mask)
				return -ENOSPC;

			word += 1;
		} else {
			/* one or more dmap words are fully contained
			 * within the block range.  determine how many
			 * words and how many bits.
			 */
			nwords = rembits >> L2DBWORD;
			nb = nwords << L2DBWORD;

			/* now examine the appropriate leaves to determine
			 * if the blocks are free.
			 */
			while (nwords > 0) {
				/* does the leaf describe any free space ?
				 */
				if (leaf[word] < BUDMIN)
					return -ENOSPC;

				/* determine the l2 number of bits provided
				 * by this leaf.
				 */
				l2size =
				    min((int)leaf[word], NLSTOL2BSZ(nwords));

				/* determine how many words were handled.
				 */
				nw = BUDSIZE(l2size, BUDMIN);

				nwords -= nw;
				word += nw;
			}
		}
	}

	/* allocate the blocks.
	 */
	return (dbAllocDmap(bmp, dp, blkno, nblocks));
}


/*
 * NAME:	dbAllocNear()
 *
 * FUNCTION:    attempt to allocate a number of contiguous free blocks near
 *		a specified block (hint) within a dmap.
 *
 *		starting with the dmap leaf that covers the hint, we'll
 *		check the next four contiguous leaves for sufficient free
 *		space.  if sufficient free space is found, we'll allocate
 *		the desired free space.
 *
 * PARAMETERS:
 *      bmp	-  pointer to bmap descriptor
 *      dp	-  pointer to dmap.
 *      blkno	-  block number to allocate near.
 *      nblocks	-  actual number of contiguous free blocks desired.
 *      l2nb	-  log2 number of contiguous free blocks desired.
 *      results	-  on successful return, set to the starting block number
 *		   of the newly allocated range.
 *
 * RETURN VALUES:
 *      0	- success
 *      -ENOSPC	- insufficient disk resources
 *      -EIO	- i/o error
 *
 * serialization: IREAD_LOCK(ipbmap) held on entry/exit;
 */
static int
dbAllocNear(struct bmap * bmp,
	    struct dmap * dp, s64 blkno, int nblocks, int l2nb, s64 * results)
{
	int word, lword, rc;
	s8 *leaf;

	if (dp->tree.leafidx != cpu_to_le32(LEAFIND)) {
		jfs_error(bmp->db_ipbmap->i_sb,
			  "dbAllocNear: Corrupt dmap page");
		return -EIO;
	}

	leaf = dp->tree.stree + le32_to_cpu(dp->tree.leafidx);

	/* determine the word within the dmap that holds the hint
	 * (i.e. blkno).  also, determine the last word in the dmap
	 * that we'll include in our examination.
	 */
	word = (blkno & (BPERDMAP - 1)) >> L2DBWORD;
	lword = min(word + 4, LPERDMAP);

	/* examine the leaves for sufficient free space.
	 */
	for (; word < lword; word++) {
		/* does the leaf describe sufficient free space ?
		 */
		if (leaf[word] < l2nb)
			continue;

		/* determine the block number within the file system
		 * of the first block described by this dmap word.
		 */
		blkno = le64_to_cpu(dp->start) + (word << L2DBWORD);

		/* if not all bits of the dmap word are free, get the
		 * starting bit number within the dmap word of the required
		 * string of free bits and adjust the block number with the
		 * value.
		 */
		if (leaf[word] < BUDMIN)
			blkno +=
			    dbFindBits(le32_to_cpu(dp->wmap[word]), l2nb);

		/* allocate the blocks.
		 */
		if ((rc = dbAllocDmap(bmp, dp, blkno, nblocks)) == 0)
			*results = blkno;

		return (rc);
	}

	return -ENOSPC;
}


/*
 * NAME:	dbAllocAG()
 *
 * FUNCTION:    attempt to allocate the specified number of contiguous
 *		free blocks within the specified allocation group.
 *
 *		unless the allocation group size is equal to the number
 *		of blocks per dmap, the dmap control pages will be used to
 *		find the required free space, if available.  we start the
 *		search at the highest dmap control page level which
 *		distinctly describes the allocation group's free space
 *		(i.e. the highest level at which the allocation group's
 *		free space is not mixed in with that of any other group).
 *		in addition, we start the search within this level at a
 *		height of the dmapctl dmtree at which the nodes distinctly
 *		describe the allocation group's free space.  at this height,
 *		the allocation group's free space may be represented by 1
 *		or two sub-trees, depending on the allocation group size.
 *		we search the top nodes of these subtrees left to right for
 *		sufficient free space.  if sufficient free space is found,
 *		the subtree is searched to find the leftmost leaf that 
 *		has free space.  once we have made it to the leaf, we
 *		move the search to the next lower level dmap control page
 *		corresponding to this leaf.  we continue down the dmap control
 *		pages until we find the dmap that contains or starts the
 *		sufficient free space and we allocate at this dmap.
 *
 *		if the allocation group size is equal to the dmap size,
 *		we'll start at the dmap corresponding to the allocation
 *		group and attempt the allocation at this level.
 *
 *		the dmap control page search is also not performed if the
 *		allocation group is completely free and we go to the first
 *		dmap of the allocation group to do the allocation.  this is
 *		done because the allocation group may be part (not the first
 *		part) of a larger binary buddy system, causing the dmap
 *		control pages to indicate no free space (NOFREE) within
 *		the allocation group.
 *
 * PARAMETERS:
 *      bmp	-  pointer to bmap descriptor
 *	agno	- allocation group number.
 *      nblocks	-  actual number of contiguous free blocks desired.
 *      l2nb	-  log2 number of contiguous free blocks desired.
 *      results	-  on successful return, set to the starting block number
 *		   of the newly allocated range.
 *
 * RETURN VALUES:
 *      0	- success
 *      -ENOSPC	- insufficient disk resources
 *      -EIO	- i/o error
 *
 * note: IWRITE_LOCK(ipmap) held on entry/exit;
 */
static int
dbAllocAG(struct bmap * bmp, int agno, s64 nblocks, int l2nb, s64 * results)
{
	struct metapage *mp;
	struct dmapctl *dcp;
	int rc, ti, i, k, m, n, agperlev;
	s64 blkno, lblkno;
	int budmin;

	/* allocation request should not be for more than the
	 * allocation group size.
	 */
	if (l2nb > bmp->db_agl2size) {
		jfs_error(bmp->db_ipbmap->i_sb,
			  "dbAllocAG: allocation request is larger than the "
			  "allocation group size");
		return -EIO;
	}

	/* determine the starting block number of the allocation
	 * group.
	 */
	blkno = (s64) agno << bmp->db_agl2size;

	/* check if the allocation group size is the minimum allocation
	 * group size or if the allocation group is completely free. if
	 * the allocation group size is the minimum size of BPERDMAP (i.e.
	 * 1 dmap), there is no need to search the dmap control page (below)
	 * that fully describes the allocation group since the allocation
	 * group is already fully described by a dmap.  in this case, we
	 * just call dbAllocCtl() to search the dmap tree and allocate the
	 * required space if available.  
	 *
	 * if the allocation group is completely free, dbAllocCtl() is
	 * also called to allocate the required space.  this is done for
	 * two reasons.  first, it makes no sense searching the dmap control
	 * pages for free space when we know that free space exists.  second,
	 * the dmap control pages may indicate that the allocation group
	 * has no free space if the allocation group is part (not the first
	 * part) of a larger binary buddy system.
	 */
	if (bmp->db_agsize == BPERDMAP
	    || bmp->db_agfree[agno] == bmp->db_agsize) {
		rc = dbAllocCtl(bmp, nblocks, l2nb, blkno, results);
		if ((rc == -ENOSPC) &&
		    (bmp->db_agfree[agno] == bmp->db_agsize)) {
			printk(KERN_ERR "blkno = %Lx, blocks = %Lx\n",
			       (unsigned long long) blkno,
			       (unsigned long long) nblocks);
			jfs_error(bmp->db_ipbmap->i_sb,
				  "dbAllocAG: dbAllocCtl failed in free AG");
		}
		return (rc);
	}

	/* the buffer for the dmap control page that fully describes the
	 * allocation group.
	 */
	lblkno = BLKTOCTL(blkno, bmp->db_l2nbperpage, bmp->db_aglevel);
	mp = read_metapage(bmp->db_ipbmap, lblkno, PSIZE, 0);
	if (mp == NULL)
		return -EIO;
	dcp = (struct dmapctl *) mp->data;
	budmin = dcp->budmin;

	if (dcp->leafidx != cpu_to_le32(CTLLEAFIND)) {
		jfs_error(bmp->db_ipbmap->i_sb,
			  "dbAllocAG: Corrupt dmapctl page");
		release_metapage(mp);
		return -EIO;
	}

	/* search the subtree(s) of the dmap control page that describes
	 * the allocation group, looking for sufficient free space.  to begin,
	 * determine how many allocation groups are represented in a dmap
	 * control page at the control page level (i.e. L0, L1, L2) that
	 * fully describes an allocation group. next, determine the starting
	 * tree index of this allocation group within the control page.
	 */
	agperlev =
	    (1 << (L2LPERCTL - (bmp->db_agheigth << 1))) / bmp->db_agwidth;
	ti = bmp->db_agstart + bmp->db_agwidth * (agno & (agperlev - 1));

	/* dmap control page trees fan-out by 4 and a single allocation 
	 * group may be described by 1 or 2 subtrees within the ag level
	 * dmap control page, depending upon the ag size. examine the ag's
	 * subtrees for sufficient free space, starting with the leftmost
	 * subtree.
	 */
	for (i = 0; i < bmp->db_agwidth; i++, ti++) {
		/* is there sufficient free space ?
		 */
		if (l2nb > dcp->stree[ti])
			continue;

		/* sufficient free space found in a subtree. now search down
		 * the subtree to find the leftmost leaf that describes this
		 * free space.
		 */
		for (k = bmp->db_agheigth; k > 0; k--) {
			for (n = 0, m = (ti << 2) + 1; n < 4; n++) {
				if (l2nb <= dcp->stree[m + n]) {
					ti = m + n;
					break;
				}
			}
			if (n == 4) {
				jfs_error(bmp->db_ipbmap->i_sb,
					  "dbAllocAG: failed descending stree");
				release_metapage(mp);
				return -EIO;
			}
		}

		/* determine the block number within the file system
		 * that corresponds to this leaf.
		 */
		if (bmp->db_aglevel == 2)
			blkno = 0;
		else if (bmp->db_aglevel == 1)
			blkno &= ~(MAXL1SIZE - 1);
		else		/* bmp->db_aglevel == 0 */
			blkno &= ~(MAXL0SIZE - 1);

		blkno +=
		    ((s64) (ti - le32_to_cpu(dcp->leafidx))) << budmin;

		/* release the buffer in preparation for going down
		 * the next level of dmap control pages.
		 */
		release_metapage(mp);

		/* check if we need to continue to search down the lower
		 * level dmap control pages.  we need to if the number of
		 * blocks required is less than maximum number of blocks
		 * described at the next lower level.
		 */
		if (l2nb < budmin) {

			/* search the lower level dmap control pages to get
			 * the starting block number of the the dmap that
			 * contains or starts off the free space.
			 */
			if ((rc =
			     dbFindCtl(bmp, l2nb, bmp->db_aglevel - 1,
				       &blkno))) {
				if (rc == -ENOSPC) {
					jfs_error(bmp->db_ipbmap->i_sb,
						  "dbAllocAG: control page "
						  "inconsistent");
					return -EIO;
				}
				return (rc);
			}
		}

		/* allocate the blocks.
		 */
		rc = dbAllocCtl(bmp, nblocks, l2nb, blkno, results);
		if (rc == -ENOSPC) {
			jfs_error(bmp->db_ipbmap->i_sb,
				  "dbAllocAG: unable to allocate blocks");
			rc = -EIO;
		}
		return (rc);
	}

	/* no space in the allocation group.  release the buffer and
	 * return -ENOSPC.
	 */
	release_metapage(mp);

	return -ENOSPC;
}


/*
 * NAME:	dbAllocAny()
 *
 * FUNCTION:    attempt to allocate the specified number of contiguous
 *		free blocks anywhere in the file system.
 *
 *		dbAllocAny() attempts to find the sufficient free space by
 *		searching down the dmap control pages, starting with the
 *		highest level (i.e. L0, L1, L2) control page.  if free space
 *		large enough to satisfy the desired free space is found, the
 *		desired free space is allocated.
 *
 * PARAMETERS:
 *      bmp	-  pointer to bmap descriptor
 *      nblocks	 -  actual number of contiguous free blocks desired.
 *      l2nb	 -  log2 number of contiguous free blocks desired.
 *      results	-  on successful return, set to the starting block number
 *		   of the newly allocated range.
 *
 * RETURN VALUES:
 *      0	- success
 *      -ENOSPC	- insufficient disk resources
 *      -EIO	- i/o error
 *
 * serialization: IWRITE_LOCK(ipbmap) held on entry/exit;
 */
static int dbAllocAny(struct bmap * bmp, s64 nblocks, int l2nb, s64 * results)
{
	int rc;
	s64 blkno = 0;

	/* starting with the top level dmap control page, search
	 * down the dmap control levels for sufficient free space.
	 * if free space is found, dbFindCtl() returns the starting
	 * block number of the dmap that contains or starts off the
	 * range of free space.
	 */
	if ((rc = dbFindCtl(bmp, l2nb, bmp->db_maxlevel, &blkno)))
		return (rc);

	/* allocate the blocks.
	 */
	rc = dbAllocCtl(bmp, nblocks, l2nb, blkno, results);
	if (rc == -ENOSPC) {
		jfs_error(bmp->db_ipbmap->i_sb,
			  "dbAllocAny: unable to allocate blocks");
		return -EIO;
	}
	return (rc);
}


/*
 * NAME:	dbFindCtl()
 *
 * FUNCTION:    starting at a specified dmap control page level and block
 *		number, search down the dmap control levels for a range of
 *	        contiguous free blocks large enough to satisfy an allocation
 *		request for the specified number of free blocks.
 *
 *		if sufficient contiguous free blocks are found, this routine
 *		returns the starting block number within a dmap page that
 *		contains or starts a range of contiqious free blocks that
 *		is sufficient in size.
 *
 * PARAMETERS:
 *      bmp	-  pointer to bmap descriptor
 *      level	-  starting dmap control page level.
 *      l2nb	-  log2 number of contiguous free blocks desired.
 *      *blkno	-  on entry, starting block number for conducting the search.
 *		   on successful return, the first block within a dmap page
 *		   that contains or starts a range of contiguous free blocks.
 *
 * RETURN VALUES:
 *      0	- success
 *      -ENOSPC	- insufficient disk resources
 *      -EIO	- i/o error
 *
 * serialization: IWRITE_LOCK(ipbmap) held on entry/exit;
 */
static int dbFindCtl(struct bmap * bmp, int l2nb, int level, s64 * blkno)
{
	int rc, leafidx, lev;
	s64 b, lblkno;
	struct dmapctl *dcp;
	int budmin;
	struct metapage *mp;

	/* starting at the specified dmap control page level and block
	 * number, search down the dmap control levels for the starting
	 * block number of a dmap page that contains or starts off 
	 * sufficient free blocks.
	 */
	for (lev = level, b = *blkno; lev >= 0; lev--) {
		/* get the buffer of the dmap control page for the block
		 * number and level (i.e. L0, L1, L2).
		 */
		lblkno = BLKTOCTL(b, bmp->db_l2nbperpage, lev);
		mp = read_metapage(bmp->db_ipbmap, lblkno, PSIZE, 0);
		if (mp == NULL)
			return -EIO;
		dcp = (struct dmapctl *) mp->data;
		budmin = dcp->budmin;

		if (dcp->leafidx != cpu_to_le32(CTLLEAFIND)) {
			jfs_error(bmp->db_ipbmap->i_sb,
				  "dbFindCtl: Corrupt dmapctl page");
			release_metapage(mp);
			return -EIO;
		}

		/* search the tree within the dmap control page for
		 * sufficent free space.  if sufficient free space is found,
		 * dbFindLeaf() returns the index of the leaf at which
		 * free space was found.
		 */
		rc = dbFindLeaf((dmtree_t *) dcp, l2nb, &leafidx);

		/* release the buffer.
		 */
		release_metapage(mp);

		/* space found ?
		 */
		if (rc) {
			if (lev != level) {
				jfs_error(bmp->db_ipbmap->i_sb,
					  "dbFindCtl: dmap inconsistent");
				return -EIO;
			}
			return -ENOSPC;
		}

		/* adjust the block number to reflect the location within
		 * the dmap control page (i.e. the leaf) at which free 
		 * space was found.
		 */
		b += (((s64) leafidx) << budmin);

		/* we stop the search at this dmap control page level if
		 * the number of blocks required is greater than or equal
		 * to the maximum number of blocks described at the next
		 * (lower) level.
		 */
		if (l2nb >= budmin)
			break;
	}

	*blkno = b;
	return (0);
}


/*
 * NAME:	dbAllocCtl()
 *
 * FUNCTION:    attempt to allocate a specified number of contiguous
 *		blocks starting within a specific dmap.  
 *		
 *		this routine is called by higher level routines that search
 *		the dmap control pages above the actual dmaps for contiguous
 *		free space.  the result of successful searches by these
 * 		routines are the starting block numbers within dmaps, with
 *		the dmaps themselves containing the desired contiguous free
 *		space or starting a contiguous free space of desired size
 *		that is made up of the blocks of one or more dmaps. these
 *		calls should not fail due to insufficent resources.
 *
 *		this routine is called in some cases where it is not known
 *		whether it will fail due to insufficient resources.  more
 *		specifically, this occurs when allocating from an allocation
 *		group whose size is equal to the number of blocks per dmap.
 *		in this case, the dmap control pages are not examined prior
 *		to calling this routine (to save pathlength) and the call
 *		might fail.
 *
 *		for a request size that fits within a dmap, this routine relies
 *		upon the dmap's dmtree to find the requested contiguous free
 *		space.  for request sizes that are larger than a dmap, the
 *		requested free space will start at the first block of the
 *		first dmap (i.e. blkno).
 *
 * PARAMETERS:
 *      bmp	-  pointer to bmap descriptor
 *      nblocks	 -  actual number of contiguous free blocks to allocate.
 *      l2nb	 -  log2 number of contiguous free blocks to allocate.
 *      blkno	 -  starting block number of the dmap to start the allocation
 *		    from.
 *      results	-  on successful return, set to the starting block number
 *		   of the newly allocated range.
 *
 * RETURN VALUES:
 *      0	- success
 *      -ENOSPC	- insufficient disk resources
 *      -EIO	- i/o error
 *
 * serialization: IWRITE_LOCK(ipbmap) held on entry/exit;
 */
static int
dbAllocCtl(struct bmap * bmp, s64 nblocks, int l2nb, s64 blkno, s64 * results)
{
	int rc, nb;
	s64 b, lblkno, n;
	struct metapage *mp;
	struct dmap *dp;

	/* check if the allocation request is confined to a single dmap.
	 */
	if (l2nb <= L2BPERDMAP) {
		/* get the buffer for the dmap.
		 */
		lblkno = BLKTODMAP(blkno, bmp->db_l2nbperpage);
		mp = read_metapage(bmp->db_ipbmap, lblkno, PSIZE, 0);
		if (mp == NULL)
			return -EIO;
		dp = (struct dmap *) mp->data;

		/* try to allocate the blocks.
		 */
		rc = dbAllocDmapLev(bmp, dp, (int) nblocks, l2nb, results);
		if (rc == 0)
			mark_metapage_dirty(mp);

		release_metapage(mp);

		return (rc);
	}

	/* allocation request involving multiple dmaps. it must start on
	 * a dmap boundary.
	 */
	assert((blkno & (BPERDMAP - 1)) == 0);

	/* allocate the blocks dmap by dmap.
	 */
	for (n = nblocks, b = blkno; n > 0; n -= nb, b += nb) {
		/* get the buffer for the dmap.
		 */
		lblkno = BLKTODMAP(b, bmp->db_l2nbperpage);
		mp = read_metapage(bmp->db_ipbmap, lblkno, PSIZE, 0);
		if (mp == NULL) {
			rc = -EIO;
			goto backout;
		}
		dp = (struct dmap *) mp->data;

		/* the dmap better be all free.
		 */
		if (dp->tree.stree[ROOT] != L2BPERDMAP) {
			release_metapage(mp);
			jfs_error(bmp->db_ipbmap->i_sb,
				  "dbAllocCtl: the dmap is not all free");
			rc = -EIO;
			goto backout;
		}

		/* determine how many blocks to allocate from this dmap.
		 */
		nb = min(n, (s64)BPERDMAP);

		/* allocate the blocks from the dmap.
		 */
		if ((rc = dbAllocDmap(bmp, dp, b, nb))) {
			release_metapage(mp);
			goto backout;
		}

		/* write the buffer.
		 */
		write_metapage(mp);
	}

	/* set the results (starting block number) and return.
	 */
	*results = blkno;
	return (0);

	/* something failed in handling an allocation request involving
	 * multiple dmaps.  we'll try to clean up by backing out any
	 * allocation that has already happened for this request.  if
	 * we fail in backing out the allocation, we'll mark the file
	 * system to indicate that blocks have been leaked.
	 */
      backout:

	/* try to backout the allocations dmap by dmap.
	 */
	for (n = nblocks - n, b = blkno; n > 0;
	     n -= BPERDMAP, b += BPERDMAP) {
		/* get the buffer for this dmap.
		 */
		lblkno = BLKTODMAP(b, bmp->db_l2nbperpage);
		mp = read_metapage(bmp->db_ipbmap, lblkno, PSIZE, 0);
		if (mp == NULL) {
			/* could not back out.  mark the file system
			 * to indicate that we have leaked blocks.
			 */
			jfs_error(bmp->db_ipbmap->i_sb,
				  "dbAllocCtl: I/O Error: Block Leakage.");
			continue;
		}
		dp = (struct dmap *) mp->data;

		/* free the blocks is this dmap.
		 */
		if (dbFreeDmap(bmp, dp, b, BPERDMAP)) {
			/* could not back out.  mark the file system
			 * to indicate that we have leaked blocks.
			 */
			release_metapage(mp);
			jfs_error(bmp->db_ipbmap->i_sb,
				  "dbAllocCtl: Block Leakage.");
			continue;
		}

		/* write the buffer.
		 */
		write_metapage(mp);
	}

	return (rc);
}


/*
 * NAME:	dbAllocDmapLev()
 *
 * FUNCTION:    attempt to allocate a specified number of contiguous blocks
 *		from a specified dmap.
 *		
 *		this routine checks if the contiguous blocks are available.
 *		if so, nblocks of blocks are allocated; otherwise, ENOSPC is
 *		returned.
 *
 * PARAMETERS:
 *      mp	-  pointer to bmap descriptor
 *      dp	-  pointer to dmap to attempt to allocate blocks from. 
 *      l2nb	-  log2 number of contiguous block desired.
 *      nblocks	-  actual number of contiguous block desired.
 *      results	-  on successful return, set to the starting block number
 *		   of the newly allocated range.
 *
 * RETURN VALUES:
 *      0	- success
 *      -ENOSPC	- insufficient disk resources
 *      -EIO	- i/o error
 *
 * serialization: IREAD_LOCK(ipbmap), e.g., from dbAlloc(), or 
 *	IWRITE_LOCK(ipbmap), e.g., dbAllocCtl(), held on entry/exit;
 */
static int
dbAllocDmapLev(struct bmap * bmp,
	       struct dmap * dp, int nblocks, int l2nb, s64 * results)
{
	s64 blkno;
	int leafidx, rc;

	/* can't be more than a dmaps worth of blocks */
	assert(l2nb <= L2BPERDMAP);

	/* search the tree within the dmap page for sufficient
	 * free space.  if sufficient free space is found, dbFindLeaf()
	 * returns the index of the leaf at which free space was found.
	 */
	if (dbFindLeaf((dmtree_t *) & dp->tree, l2nb, &leafidx))
		return -ENOSPC;

	/* determine the block number within the file system corresponding
	 * to the leaf at which free space was found.
	 */
	blkno = le64_to_cpu(dp->start) + (leafidx << L2DBWORD);

	/* if not all bits of the dmap word are free, get the starting
	 * bit number within the dmap word of the required string of free
	 * bits and adjust the block number with this value.
	 */
	if (dp->tree.stree[leafidx + LEAFIND] < BUDMIN)
		blkno += dbFindBits(le32_to_cpu(dp->wmap[leafidx]), l2nb);

	/* allocate the blocks */
	if ((rc = dbAllocDmap(bmp, dp, blkno, nblocks)) == 0)
		*results = blkno;

	return (rc);
}


/*
 * NAME:	dbAllocDmap()
 *
 * FUNCTION:    adjust the disk allocation map to reflect the allocation
 *		of a specified block range within a dmap.
 *
 *		this routine allocates the specified blocks from the dmap
 *		through a call to dbAllocBits(). if the allocation of the
 *		block range causes the maximum string of free blocks within
 *		the dmap to change (i.e. the value of the root of the dmap's
 *		dmtree), this routine will cause this change to be reflected
 *		up through the appropriate levels of the dmap control pages
 *		by a call to dbAdjCtl() for the L0 dmap control page that
 *		covers this dmap.
 *
 * PARAMETERS:
 *      bmp	-  pointer to bmap descriptor
 *      dp	-  pointer to dmap to allocate the block range from.
 *      blkno	-  starting block number of the block to be allocated.
 *      nblocks	-  number of blocks to be allocated.
 *
 * RETURN VALUES:
 *      0	- success
 *      -EIO	- i/o error
 *
 * serialization: IREAD_LOCK(ipbmap) or IWRITE_LOCK(ipbmap) held on entry/exit;
 */
static int dbAllocDmap(struct bmap * bmp, struct dmap * dp, s64 blkno,
		       int nblocks)
{
	s8 oldroot;
	int rc;

	/* save the current value of the root (i.e. maximum free string)
	 * of the dmap tree.
	 */
	oldroot = dp->tree.stree[ROOT];

	/* allocate the specified (blocks) bits */
	dbAllocBits(bmp, dp, blkno, nblocks);

	/* if the root has not changed, done. */
	if (dp->tree.stree[ROOT] == oldroot)
		return (0);

	/* root changed. bubble the change up to the dmap control pages.
	 * if the adjustment of the upper level control pages fails,
	 * backout the bit allocation (thus making everything consistent).
	 */
	if ((rc = dbAdjCtl(bmp, blkno, dp->tree.stree[ROOT], 1, 0)))
		dbFreeBits(bmp, dp, blkno, nblocks);

	return (rc);
}


/*
 * NAME:	dbFreeDmap()
 *
 * FUNCTION:    adjust the disk allocation map to reflect the allocation
 *		of a specified block range within a dmap.
 *
 *		this routine frees the specified blocks from the dmap through
 *		a call to dbFreeBits(). if the deallocation of the block range
 *		causes the maximum string of free blocks within the dmap to
 *		change (i.e. the value of the root of the dmap's dmtree), this
 *		routine will cause this change to be reflected up through the
 *	        appropriate levels of the dmap control pages by a call to
 *		dbAdjCtl() for the L0 dmap control page that covers this dmap.
 *
 * PARAMETERS:
 *      bmp	-  pointer to bmap descriptor
 *      dp	-  pointer to dmap to free the block range from.
 *      blkno	-  starting block number of the block to be freed.
 *      nblocks	-  number of blocks to be freed.
 *
 * RETURN VALUES:
 *      0	- success
 *      -EIO	- i/o error
 *
 * serialization: IREAD_LOCK(ipbmap) or IWRITE_LOCK(ipbmap) held on entry/exit;
 */
static int dbFreeDmap(struct bmap * bmp, struct dmap * dp, s64 blkno,
		      int nblocks)
{
	s8 oldroot;
	int rc = 0, word;

	/* save the current value of the root (i.e. maximum free string)
	 * of the dmap tree.
	 */
	oldroot = dp->tree.stree[ROOT];

	/* free the specified (blocks) bits */
	rc = dbFreeBits(bmp, dp, blkno, nblocks);

	/* if error or the root has not changed, done. */
	if (rc || (dp->tree.stree[ROOT] == oldroot))
		return (rc);

	/* root changed. bubble the change up to the dmap control pages.
	 * if the adjustment of the upper level control pages fails,
	 * backout the deallocation. 
	 */
	if ((rc = dbAdjCtl(bmp, blkno, dp->tree.stree[ROOT], 0, 0))) {
		word = (blkno & (BPERDMAP - 1)) >> L2DBWORD;

		/* as part of backing out the deallocation, we will have
		 * to back split the dmap tree if the deallocation caused
		 * the freed blocks to become part of a larger binary buddy
		 * system.
		 */
		if (dp->tree.stree[word] == NOFREE)
			dbBackSplit((dmtree_t *) & dp->tree, word);

		dbAllocBits(bmp, dp, blkno, nblocks);
	}

	return (rc);
}


/*
 * NAME:	dbAllocBits()
 *
 * FUNCTION:    allocate a specified block range from a dmap.
 *
 *		this routine updates the dmap to reflect the working
 *		state allocation of the specified block range. it directly
 *		updates the bits of the working map and causes the adjustment
 *		of the binary buddy system described by the dmap's dmtree
 *		leaves to reflect the bits allocated.  it also causes the
 *		dmap's dmtree, as a whole, to reflect the allocated range.
 *
 * PARAMETERS:
 *      bmp	-  pointer to bmap descriptor
 *      dp	-  pointer to dmap to allocate bits from.
 *      blkno	-  starting block number of the bits to be allocated.
 *      nblocks	-  number of bits to be allocated.
 *
 * RETURN VALUES: none
 *
 * serialization: IREAD_LOCK(ipbmap) or IWRITE_LOCK(ipbmap) held on entry/exit;
 */
static void dbAllocBits(struct bmap * bmp, struct dmap * dp, s64 blkno,
			int nblocks)
{
	int dbitno, word, rembits, nb, nwords, wbitno, nw, agno;
	dmtree_t *tp = (dmtree_t *) & dp->tree;
	int size;
	s8 *leaf;

	/* pick up a pointer to the leaves of the dmap tree */
	leaf = dp->tree.stree + LEAFIND;

	/* determine the bit number and word within the dmap of the
	 * starting block.
	 */
	dbitno = blkno & (BPERDMAP - 1);
	word = dbitno >> L2DBWORD;

	/* block range better be within the dmap */
	assert(dbitno + nblocks <= BPERDMAP);

	/* allocate the bits of the dmap's words corresponding to the block
	 * range. not all bits of the first and last words may be contained
	 * within the block range.  if this is the case, we'll work against
	 * those words (i.e. partial first and/or last) on an individual basis
	 * (a single pass), allocating the bits of interest by hand and
	 * updating the leaf corresponding to the dmap word. a single pass
	 * will be used for all dmap words fully contained within the
	 * specified range.  within this pass, the bits of all fully contained
	 * dmap words will be marked as free in a single shot and the leaves
	 * will be updated. a single leaf may describe the free space of
	 * multiple dmap words, so we may update only a subset of the actual
	 * leaves corresponding to the dmap words of the block range.
	 */
	for (rembits = nblocks; rembits > 0; rembits -= nb, dbitno += nb) {
		/* determine the bit number within the word and
		 * the number of bits within the word.
		 */
		wbitno = dbitno & (DBWORD - 1);
		nb = min(rembits, DBWORD - wbitno);

		/* check if only part of a word is to be allocated.
		 */
		if (nb < DBWORD) {
			/* allocate (set to 1) the appropriate bits within
			 * this dmap word.
			 */
			dp->wmap[word] |= cpu_to_le32(ONES << (DBWORD - nb)
						      >> wbitno);

			/* update the leaf for this dmap word. in addition
			 * to setting the leaf value to the binary buddy max
			 * of the updated dmap word, dbSplit() will split
			 * the binary system of the leaves if need be.
			 */
			dbSplit(tp, word, BUDMIN,
				dbMaxBud((u8 *) & dp->wmap[word]));

			word += 1;
		} else {
			/* one or more dmap words are fully contained
			 * within the block range.  determine how many
			 * words and allocate (set to 1) the bits of these
			 * words.
			 */
			nwords = rembits >> L2DBWORD;
			memset(&dp->wmap[word], (int) ONES, nwords * 4);

			/* determine how many bits.
			 */
			nb = nwords << L2DBWORD;

			/* now update the appropriate leaves to reflect
			 * the allocated words.
			 */
			for (; nwords > 0; nwords -= nw) {
			        if (leaf[word] < BUDMIN) {
					jfs_error(bmp->db_ipbmap->i_sb,
						  "dbAllocBits: leaf page "
						  "corrupt");
					break;
				}

				/* determine what the leaf value should be
				 * updated to as the minimum of the l2 number
				 * of bits being allocated and the l2 number
				 * of bits currently described by this leaf.
				 */
				size = min((int)leaf[word], NLSTOL2BSZ(nwords));

				/* update the leaf to reflect the allocation.
				 * in addition to setting the leaf value to
				 * NOFREE, dbSplit() will split the binary
				 * system of the leaves to reflect the current
				 * allocation (size).
				 */
				dbSplit(tp, word, size, NOFREE);

				/* get the number of dmap words handled */
				nw = BUDSIZE(size, BUDMIN);
				word += nw;
			}
		}
	}

	/* update the free count for this dmap */
	dp->nfree = cpu_to_le32(le32_to_cpu(dp->nfree) - nblocks);

	BMAP_LOCK(bmp);

	/* if this allocation group is completely free,
	 * update the maximum allocation group number if this allocation
	 * group is the new max.
	 */
	agno = blkno >> bmp->db_agl2size;
	if (agno > bmp->db_maxag)
		bmp->db_maxag = agno;

	/* update the free count for the allocation group and map */
	bmp->db_agfree[agno] -= nblocks;
	bmp->db_nfree -= nblocks;

	BMAP_UNLOCK(bmp);
}


/*
 * NAME:	dbFreeBits()
 *
 * FUNCTION:    free a specified block range from a dmap.
 *
 *		this routine updates the dmap to reflect the working
 *		state allocation of the specified block range. it directly
 *		updates the bits of the working map and causes the adjustment
 *		of the binary buddy system described by the dmap's dmtree
 *		leaves to reflect the bits freed.  it also causes the dmap's
 *		dmtree, as a whole, to reflect the deallocated range.
 *
 * PARAMETERS:
 *      bmp	-  pointer to bmap descriptor
 *      dp	-  pointer to dmap to free bits from.
 *      blkno	-  starting block number of the bits to be freed.
 *      nblocks	-  number of bits to be freed.
 *
 * RETURN VALUES: 0 for success
 *
 * serialization: IREAD_LOCK(ipbmap) or IWRITE_LOCK(ipbmap) held on entry/exit;
 */
static int dbFreeBits(struct bmap * bmp, struct dmap * dp, s64 blkno,
		       int nblocks)
{
	int dbitno, word, rembits, nb, nwords, wbitno, nw, agno;
	dmtree_t *tp = (dmtree_t *) & dp->tree;
	int rc = 0;
	int size;

	/* determine the bit number and word within the dmap of the
	 * starting block.
	 */
	dbitno = blkno & (BPERDMAP - 1);
	word = dbitno >> L2DBWORD;

	/* block range better be within the dmap.
	 */
	assert(dbitno + nblocks <= BPERDMAP);

	/* free the bits of the dmaps words corresponding to the block range.
	 * not all bits of the first and last words may be contained within
	 * the block range.  if this is the case, we'll work against those
	 * words (i.e. partial first and/or last) on an individual basis
	 * (a single pass), freeing the bits of interest by hand and updating
	 * the leaf corresponding to the dmap word. a single pass will be used
	 * for all dmap words fully contained within the specified range.  
	 * within this pass, the bits of all fully contained dmap words will
	 * be marked as free in a single shot and the leaves will be updated. a
	 * single leaf may describe the free space of multiple dmap words,
	 * so we may update only a subset of the actual leaves corresponding
	 * to the dmap words of the block range.
	 *
	 * dbJoin() is used to update leaf values and will join the binary
	 * buddy system of the leaves if the new leaf values indicate this
	 * should be done.
	 */
	for (rembits = nblocks; rembits > 0; rembits -= nb, dbitno += nb) {
		/* determine the bit number within the word and
		 * the number of bits within the word.
		 */
		wbitno = dbitno & (DBWORD - 1);
		nb = min(rembits, DBWORD - wbitno);

		/* check if only part of a word is to be freed.
		 */
		if (nb < DBWORD) {
			/* free (zero) the appropriate bits within this
			 * dmap word. 
			 */
			dp->wmap[word] &=
			    cpu_to_le32(~(ONES << (DBWORD - nb)
					  >> wbitno));

			/* update the leaf for this dmap word.
			 */
			rc = dbJoin(tp, word,
				    dbMaxBud((u8 *) & dp->wmap[word]));
			if (rc)
				return rc;

			word += 1;
		} else {
			/* one or more dmap words are fully contained
			 * within the block range.  determine how many
			 * words and free (zero) the bits of these words.
			 */
			nwords = rembits >> L2DBWORD;
			memset(&dp->wmap[word], 0, nwords * 4);

			/* determine how many bits.
			 */
			nb = nwords << L2DBWORD;

			/* now update the appropriate leaves to reflect
			 * the freed words.
			 */
			for (; nwords > 0; nwords -= nw) {
				/* determine what the leaf value should be
				 * updated to as the minimum of the l2 number
				 * of bits being freed and the l2 (max) number
				 * of bits that can be described by this leaf.
				 */
				size =
				    min(LITOL2BSZ
					(word, L2LPERDMAP, BUDMIN),
					NLSTOL2BSZ(nwords));

				/* update the leaf.
				 */
				rc = dbJoin(tp, word, size);
				if (rc)
					return rc;

				/* get the number of dmap words handled.
				 */
				nw = BUDSIZE(size, BUDMIN);
				word += nw;
			}
		}
	}

	/* update the free count for this dmap.
	 */
	dp->nfree = cpu_to_le32(le32_to_cpu(dp->nfree) + nblocks);

	BMAP_LOCK(bmp);

	/* update the free count for the allocation group and 
	 * map.
	 */
	agno = blkno >> bmp->db_agl2size;
	bmp->db_nfree += nblocks;
	bmp->db_agfree[agno] += nblocks;

	/* check if this allocation group is not completely free and
	 * if it is currently the maximum (rightmost) allocation group.
	 * if so, establish the new maximum allocation group number by
	 * searching left for the first allocation group with allocation.
	 */
	if ((bmp->db_agfree[agno] == bmp->db_agsize && agno == bmp->db_maxag) ||
	    (agno == bmp->db_numag - 1 &&
	     bmp->db_agfree[agno] == (bmp-> db_mapsize & (BPERDMAP - 1)))) {
		while (bmp->db_maxag > 0) {
			bmp->db_maxag -= 1;
			if (bmp->db_agfree[bmp->db_maxag] !=
			    bmp->db_agsize)
				break;
		}

		/* re-establish the allocation group preference if the
		 * current preference is right of the maximum allocation
		 * group.
		 */
		if (bmp->db_agpref > bmp->db_maxag)
			bmp->db_agpref = bmp->db_maxag;
	}

	BMAP_UNLOCK(bmp);

	return 0;
}


/*
 * NAME:	dbAdjCtl()
 *
 * FUNCTION:	adjust a dmap control page at a specified level to reflect
 *		the change in a lower level dmap or dmap control page's
 *		maximum string of free blocks (i.e. a change in the root
 *		of the lower level object's dmtree) due to the allocation
 *		or deallocation of a range of blocks with a single dmap.
 *
 *		on entry, this routine is provided with the new value of
 *		the lower level dmap or dmap control page root and the
 *		starting block number of the block range whose allocation
 *		or deallocation resulted in the root change.  this range
 *		is respresented by a single leaf of the current dmapctl
 *		and the leaf will be updated with this value, possibly
 *		causing a binary buddy system within the leaves to be 
 *		split or joined.  the update may also cause the dmapctl's
 *		dmtree to be updated.
 *
 *		if the adjustment of the dmap control page, itself, causes its
 *		root to change, this change will be bubbled up to the next dmap
 *		control level by a recursive call to this routine, specifying
 *		the new root value and the next dmap control page level to
 *		be adjusted.
 * PARAMETERS:
 *      bmp	-  pointer to bmap descriptor
 *      blkno	-  the first block of a block range within a dmap.  it is
 *		   the allocation or deallocation of this block range that
 *		   requires the dmap control page to be adjusted.
 *      newval	-  the new value of the lower level dmap or dmap control
 *		   page root.
 *      alloc	-  TRUE if adjustment is due to an allocation.
 *      level	-  current level of dmap control page (i.e. L0, L1, L2) to
 *		   be adjusted.
 *
 * RETURN VALUES:
 *      0	- success
 *      -EIO	- i/o error
 *
 * serialization: IREAD_LOCK(ipbmap) or IWRITE_LOCK(ipbmap) held on entry/exit;
 */
static int
dbAdjCtl(struct bmap * bmp, s64 blkno, int newval, int alloc, int level)
{
	struct metapage *mp;
	s8 oldroot;
	int oldval;
	s64 lblkno;
	struct dmapctl *dcp;
	int rc, leafno, ti;

	/* get the buffer for the dmap control page for the specified
	 * block number and control page level.
	 */
	lblkno = BLKTOCTL(blkno, bmp->db_l2nbperpage, level);
	mp = read_metapage(bmp->db_ipbmap, lblkno, PSIZE, 0);
	if (mp == NULL)
		return -EIO;
	dcp = (struct dmapctl *) mp->data;

	if (dcp->leafidx != cpu_to_le32(CTLLEAFIND)) {
		jfs_error(bmp->db_ipbmap->i_sb,
			  "dbAdjCtl: Corrupt dmapctl page");
		release_metapage(mp);
		return -EIO;
	}

	/* determine the leaf number corresponding to the block and
	 * the index within the dmap control tree.
	 */
	leafno = BLKTOCTLLEAF(blkno, dcp->budmin);
	ti = leafno + le32_to_cpu(dcp->leafidx);

	/* save the current leaf value and the current root level (i.e.
	 * maximum l2 free string described by this dmapctl).
	 */
	oldval = dcp->stree[ti];
	oldroot = dcp->stree[ROOT];

	/* check if this is a control page update for an allocation.
	 * if so, update the leaf to reflect the new leaf value using
	 * dbSplit(); otherwise (deallocation), use dbJoin() to udpate
	 * the leaf with the new value.  in addition to updating the
	 * leaf, dbSplit() will also split the binary buddy system of
	 * the leaves, if required, and bubble new values within the
	 * dmapctl tree, if required.  similarly, dbJoin() will join
	 * the binary buddy system of leaves and bubble new values up
	 * the dmapctl tree as required by the new leaf value.
	 */
	if (alloc) {
		/* check if we are in the middle of a binary buddy
		 * system.  this happens when we are performing the
		 * first allocation out of an allocation group that
		 * is part (not the first part) of a larger binary
		 * buddy system.  if we are in the middle, back split
		 * the system prior to calling dbSplit() which assumes
		 * that it is at the front of a binary buddy system.
		 */
		if (oldval == NOFREE) {
			rc = dbBackSplit((dmtree_t *) dcp, leafno);
			if (rc)
				return rc;
			oldval = dcp->stree[ti];
		}
		dbSplit((dmtree_t *) dcp, leafno, dcp->budmin, newval);
	} else {
		rc = dbJoin((dmtree_t *) dcp, leafno, newval);
		if (rc)
			return rc;
	}

	/* check if the root of the current dmap control page changed due
	 * to the update and if the current dmap control page is not at
	 * the current top level (i.e. L0, L1, L2) of the map.  if so (i.e.
	 * root changed and this is not the top level), call this routine
	 * again (recursion) for the next higher level of the mapping to
	 * reflect the change in root for the current dmap control page.
	 */
	if (dcp->stree[ROOT] != oldroot) {
		/* are we below the top level of the map.  if so,
		 * bubble the root up to the next higher level.
		 */
		if (level < bmp->db_maxlevel) {
			/* bubble up the new root of this dmap control page to
			 * the next level.
			 */
			if ((rc =
			     dbAdjCtl(bmp, blkno, dcp->stree[ROOT], alloc,
				      level + 1))) {
				/* something went wrong in bubbling up the new
				 * root value, so backout the changes to the
				 * current dmap control page.
				 */
				if (alloc) {
					dbJoin((dmtree_t *) dcp, leafno,
					       oldval);
				} else {
					/* the dbJoin() above might have
					 * caused a larger binary buddy system
					 * to form and we may now be in the
					 * middle of it.  if this is the case,
					 * back split the buddies.
					 */
					if (dcp->stree[ti] == NOFREE)
						dbBackSplit((dmtree_t *)
							    dcp, leafno);
					dbSplit((dmtree_t *) dcp, leafno,
						dcp->budmin, oldval);
				}

				/* release the buffer and return the error.
				 */
				release_metapage(mp);
				return (rc);
			}
		} else {
			/* we're at the top level of the map. update
			 * the bmap control page to reflect the size
			 * of the maximum free buddy system.
			 */
			assert(level == bmp->db_maxlevel);
			if (bmp->db_maxfreebud != oldroot) {
				jfs_error(bmp->db_ipbmap->i_sb,
					  "dbAdjCtl: the maximum free buddy is "
					  "not the old root");
			}
			bmp->db_maxfreebud = dcp->stree[ROOT];
		}
	}

	/* write the buffer.
	 */
	write_metapage(mp);

	return (0);
}


/*
 * NAME:	dbSplit()
 *
 * FUNCTION:    update the leaf of a dmtree with a new value, splitting
 *		the leaf from the binary buddy system of the dmtree's
 *		leaves, as required.
 *
 * PARAMETERS:
 *      tp	- pointer to the tree containing the leaf.
 *      leafno	- the number of the leaf to be updated.
 *      splitsz	- the size the binary buddy system starting at the leaf
 *		  must be split to, specified as the log2 number of blocks.
 *      newval	- the new value for the leaf.
 *
 * RETURN VALUES: none
 *
 * serialization: IREAD_LOCK(ipbmap) or IWRITE_LOCK(ipbmap) held on entry/exit;
 */
static void dbSplit(dmtree_t * tp, int leafno, int splitsz, int newval)
{
	int budsz;
	int cursz;
	s8 *leaf = tp->dmt_stree + le32_to_cpu(tp->dmt_leafidx);

	/* check if the leaf needs to be split.
	 */
	if (leaf[leafno] > tp->dmt_budmin) {
		/* the split occurs by cutting the buddy system in half
		 * at the specified leaf until we reach the specified
		 * size.  pick up the starting split size (current size
		 * - 1 in l2) and the corresponding buddy size.
		 */
		cursz = leaf[leafno] - 1;
		budsz = BUDSIZE(cursz, tp->dmt_budmin);

		/* split until we reach the specified size.
		 */
		while (cursz >= splitsz) {
			/* update the buddy's leaf with its new value.
			 */
			dbAdjTree(tp, leafno ^ budsz, cursz);

			/* on to the next size and buddy.
			 */
			cursz -= 1;
			budsz >>= 1;
		}
	}

	/* adjust the dmap tree to reflect the specified leaf's new 
	 * value.
	 */
	dbAdjTree(tp, leafno, newval);
}


/*
 * NAME:	dbBackSplit()
 *
 * FUNCTION:    back split the binary buddy system of dmtree leaves
 *		that hold a specified leaf until the specified leaf
 *		starts its own binary buddy system.
 *
 *		the allocators typically perform allocations at the start
 *		of binary buddy systems and dbSplit() is used to accomplish
 *		any required splits.  in some cases, however, allocation
 *		may occur in the middle of a binary system and requires a
 *		back split, with the split proceeding out from the middle of
 *		the system (less efficient) rather than the start of the
 *		system (more efficient).  the cases in which a back split
 *		is required are rare and are limited to the first allocation
 *		within an allocation group which is a part (not first part)
 *		of a larger binary buddy system and a few exception cases
 *		in which a previous join operation must be backed out.
 *
 * PARAMETERS:
 *      tp	- pointer to the tree containing the leaf.
 *      leafno	- the number of the leaf to be updated.
 *
 * RETURN VALUES: none
 *
 * serialization: IREAD_LOCK(ipbmap) or IWRITE_LOCK(ipbmap) held on entry/exit;
 */
static int dbBackSplit(dmtree_t * tp, int leafno)
{
	int budsz, bud, w, bsz, size;
	int cursz;
	s8 *leaf = tp->dmt_stree + le32_to_cpu(tp->dmt_leafidx);

	/* leaf should be part (not first part) of a binary
	 * buddy system.
	 */
	assert(leaf[leafno] == NOFREE);

	/* the back split is accomplished by iteratively finding the leaf
	 * that starts the buddy system that contains the specified leaf and
	 * splitting that system in two.  this iteration continues until
	 * the specified leaf becomes the start of a buddy system. 
	 *
	 * determine maximum possible l2 size for the specified leaf.
	 */
	size =
	    LITOL2BSZ(leafno, le32_to_cpu(tp->dmt_l2nleafs),
		      tp->dmt_budmin);

	/* determine the number of leaves covered by this size.  this
	 * is the buddy size that we will start with as we search for
	 * the buddy system that contains the specified leaf.
	 */
	budsz = BUDSIZE(size, tp->dmt_budmin);

	/* back split.
	 */
	while (leaf[leafno] == NOFREE) {
		/* find the leftmost buddy leaf.
		 */
		for (w = leafno, bsz = budsz;; bsz <<= 1,
		     w = (w < bud) ? w : bud) {
			if (bsz >= le32_to_cpu(tp->dmt_nleafs)) {
				jfs_err("JFS: block map error in dbBackSplit");
				return -EIO;
			}

			/* determine the buddy.
			 */
			bud = w ^ bsz;

			/* check if this buddy is the start of the system.
			 */
			if (leaf[bud] != NOFREE) {
				/* split the leaf at the start of the
				 * system in two.
				 */
				cursz = leaf[bud] - 1;
				dbSplit(tp, bud, cursz, cursz);
				break;
			}
		}
	}

	if (leaf[leafno] != size) {
		jfs_err("JFS: wrong leaf value in dbBackSplit");
		return -EIO;
	}
	return 0;
}


/*
 * NAME:	dbJoin()
 *
 * FUNCTION:    update the leaf of a dmtree with a new value, joining
 *		the leaf with other leaves of the dmtree into a multi-leaf
 *		binary buddy system, as required.
 *
 * PARAMETERS:
 *      tp	- pointer to the tree containing the leaf.
 *      leafno	- the number of the leaf to be updated.
 *      newval	- the new value for the leaf.
 *
 * RETURN VALUES: none
 */
static int dbJoin(dmtree_t * tp, int leafno, int newval)
{
	int budsz, buddy;
	s8 *leaf;

	/* can the new leaf value require a join with other leaves ?
	 */
	if (newval >= tp->dmt_budmin) {
		/* pickup a pointer to the leaves of the tree.
		 */
		leaf = tp->dmt_stree + le32_to_cpu(tp->dmt_leafidx);

		/* try to join the specified leaf into a large binary
		 * buddy system.  the join proceeds by attempting to join
		 * the specified leafno with its buddy (leaf) at new value.
		 * if the join occurs, we attempt to join the left leaf
		 * of the joined buddies with its buddy at new value + 1.
		 * we continue to join until we find a buddy that cannot be
		 * joined (does not have a value equal to the size of the
		 * last join) or until all leaves have been joined into a
		 * single system.
		 *
		 * get the buddy size (number of words covered) of
		 * the new value.
		 */
		budsz = BUDSIZE(newval, tp->dmt_budmin);

		/* try to join.
		 */
		while (budsz < le32_to_cpu(tp->dmt_nleafs)) {
			/* get the buddy leaf.
			 */
			buddy = leafno ^ budsz;

			/* if the leaf's new value is greater than its
			 * buddy's value, we join no more.
			 */
			if (newval > leaf[buddy])
				break;

			/* It shouldn't be less */
			if (newval < leaf[buddy])
				return -EIO;

			/* check which (leafno or buddy) is the left buddy.
			 * the left buddy gets to claim the blocks resulting
			 * from the join while the right gets to claim none.
			 * the left buddy is also eligable to participate in
			 * a join at the next higher level while the right
			 * is not.
			 *
			 */
			if (leafno < buddy) {
				/* leafno is the left buddy.
				 */
				dbAdjTree(tp, buddy, NOFREE);
			} else {
				/* buddy is the left buddy and becomes
				 * leafno.
				 */
				dbAdjTree(tp, leafno, NOFREE);
				leafno = buddy;
			}

			/* on to try the next join.
			 */
			newval += 1;
			budsz <<= 1;
		}
	}

	/* update the leaf value.
	 */
	dbAdjTree(tp, leafno, newval);

	return 0;
}


/*
 * NAME:	dbAdjTree()
 *
 * FUNCTION:    update a leaf of a dmtree with a new value, adjusting
 *		the dmtree, as required, to reflect the new leaf value.
 *		the combination of any buddies must already be done before
 *		this is called.
 *
 * PARAMETERS:
 *      tp	- pointer to the tree to be adjusted.
 *      leafno	- the number of the leaf to be updated.
 *      newval	- the new value for the leaf.
 *
 * RETURN VALUES: none
 */
static void dbAdjTree(dmtree_t * tp, int leafno, int newval)
{
	int lp, pp, k;
	int max;

	/* pick up the index of the leaf for this leafno.
	 */
	lp = leafno + le32_to_cpu(tp->dmt_leafidx);

	/* is the current value the same as the old value ?  if so,
	 * there is nothing to do.
	 */
	if (tp->dmt_stree[lp] == newval)
		return;

	/* set the new value.
	 */
	tp->dmt_stree[lp] = newval;

	/* bubble the new value up the tree as required.
	 */
	for (k = 0; k < le32_to_cpu(tp->dmt_height); k++) {
		/* get the index of the first leaf of the 4 leaf
		 * group containing the specified leaf (leafno).
		 */
		lp = ((lp - 1) & ~0x03) + 1;

		/* get the index of the parent of this 4 leaf group.
		 */
		pp = (lp - 1) >> 2;

		/* determine the maximum of the 4 leaves.
		 */
		max = TREEMAX(&tp->dmt_stree[lp]);

		/* if the maximum of the 4 is the same as the
		 * parent's value, we're done.
		 */
		if (tp->dmt_stree[pp] == max)
			break;

		/* parent gets new value.
		 */
		tp->dmt_stree[pp] = max;

		/* parent becomes leaf for next go-round.
		 */
		lp = pp;
	}
}


/*
 * NAME:	dbFindLeaf()
 *
 * FUNCTION:    search a dmtree_t for sufficient free blocks, returning
 *		the index of a leaf describing the free blocks if 
 *		sufficient free blocks are found.
 *
 *		the search starts at the top of the dmtree_t tree and
 *		proceeds down the tree to the leftmost leaf with sufficient
 *		free space.
 *
 * PARAMETERS:
 *      tp	- pointer to the tree to be searched.
 *      l2nb	- log2 number of free blocks to search for.
 *	leafidx	- return pointer to be set to the index of the leaf
 *		  describing at least l2nb free blocks if sufficient
 *		  free blocks are found.
 *
 * RETURN VALUES:
 *      0	- success
 *      -ENOSPC	- insufficient free blocks. 
 */
static int dbFindLeaf(dmtree_t * tp, int l2nb, int *leafidx)
{
	int ti, n = 0, k, x = 0;

	/* first check the root of the tree to see if there is
	 * sufficient free space.
	 */
	if (l2nb > tp->dmt_stree[ROOT])
		return -ENOSPC;

	/* sufficient free space available. now search down the tree
	 * starting at the next level for the leftmost leaf that
	 * describes sufficient free space.
	 */
	for (k = le32_to_cpu(tp->dmt_height), ti = 1;
	     k > 0; k--, ti = ((ti + n) << 2) + 1) {
		/* search the four nodes at this level, starting from
		 * the left.
		 */
		for (x = ti, n = 0; n < 4; n++) {
			/* sufficient free space found.  move to the next
			 * level (or quit if this is the last level).
			 */
			if (l2nb <= tp->dmt_stree[x + n])
				break;
		}

		/* better have found something since the higher
		 * levels of the tree said it was here.
		 */
		assert(n < 4);
	}

	/* set the return to the leftmost leaf describing sufficient
	 * free space.
	 */
	*leafidx = x + n - le32_to_cpu(tp->dmt_leafidx);

	return (0);
}


/*
 * NAME:	dbFindBits()
 *
 * FUNCTION:    find a specified number of binary buddy free bits within a
 *		dmap bitmap word value.
 *
 *		this routine searches the bitmap value for (1 << l2nb) free
 *		bits at (1 << l2nb) alignments within the value.
 *
 * PARAMETERS:
 *      word	-  dmap bitmap word value.
 *      l2nb	-  number of free bits specified as a log2 number.
 *
 * RETURN VALUES:
 *      starting bit number of free bits.
 */
static int dbFindBits(u32 word, int l2nb)
{
	int bitno, nb;
	u32 mask;

	/* get the number of bits.
	 */
	nb = 1 << l2nb;
	assert(nb <= DBWORD);

	/* complement the word so we can use a mask (i.e. 0s represent
	 * free bits) and compute the mask.
	 */
	word = ~word;
	mask = ONES << (DBWORD - nb);

	/* scan the word for nb free bits at nb alignments.
	 */
	for (bitno = 0; mask != 0; bitno += nb, mask >>= nb) {
		if ((mask & word) == mask)
			break;
	}

	ASSERT(bitno < 32);

	/* return the bit number.
	 */
	return (bitno);
}


/*
 * NAME:	dbMaxBud(u8 *cp)
 *
 * FUNCTION:    determine the largest binary buddy string of free
 *		bits within 32-bits of the map.
 *
 * PARAMETERS:
 *      cp	-  pointer to the 32-bit value.
 *
 * RETURN VALUES:
 *      largest binary buddy of free bits within a dmap word.
 */
static int dbMaxBud(u8 * cp)
{
	signed char tmp1, tmp2;

	/* check if the wmap word is all free. if so, the
	 * free buddy size is BUDMIN.
	 */
	if (*((uint *) cp) == 0)
		return (BUDMIN);

	/* check if the wmap word is half free. if so, the
	 * free buddy size is BUDMIN-1.
	 */
	if (*((u16 *) cp) == 0 || *((u16 *) cp + 1) == 0)
		return (BUDMIN - 1);

	/* not all free or half free. determine the free buddy
	 * size thru table lookup using quarters of the wmap word.
	 */
	tmp1 = max(budtab[cp[2]], budtab[cp[3]]);
	tmp2 = max(budtab[cp[0]], budtab[cp[1]]);
	return (max(tmp1, tmp2));
}


/*
 * NAME:	cnttz(uint word)
 *
 * FUNCTION:    determine the number of trailing zeros within a 32-bit
 *		value.
 *
 * PARAMETERS:
 *      value	-  32-bit value to be examined.
 *
 * RETURN VALUES:
 *      count of trailing zeros
 */
static int cnttz(u32 word)
{
	int n;

	for (n = 0; n < 32; n++, word >>= 1) {
		if (word & 0x01)
			break;
	}

	return (n);
}


/*
 * NAME:	cntlz(u32 value)
 *
 * FUNCTION:    determine the number of leading zeros within a 32-bit
 *		value.
 *
 * PARAMETERS:
 *      value	-  32-bit value to be examined.
 *
 * RETURN VALUES:
 *      count of leading zeros
 */
static int cntlz(u32 value)
{
	int n;

	for (n = 0; n < 32; n++, value <<= 1) {
		if (value & HIGHORDER)
			break;
	}
	return (n);
}


/*
 * NAME:	blkstol2(s64 nb)
 *
 * FUNCTION:	convert a block count to its log2 value. if the block
 *	        count is not a l2 multiple, it is rounded up to the next
 *		larger l2 multiple.
 *
 * PARAMETERS:
 *      nb	-  number of blocks
 *
 * RETURN VALUES:
 *      log2 number of blocks
 */
static int blkstol2(s64 nb)
{
	int l2nb;
	s64 mask;		/* meant to be signed */

	mask = (s64) 1 << (64 - 1);

	/* count the leading bits.
	 */
	for (l2nb = 0; l2nb < 64; l2nb++, mask >>= 1) {
		/* leading bit found.
		 */
		if (nb & mask) {
			/* determine the l2 value.
			 */
			l2nb = (64 - 1) - l2nb;

			/* check if we need to round up.
			 */
			if (~mask & nb)
				l2nb++;

			return (l2nb);
		}
	}
	assert(0);
	return 0;		/* fix compiler warning */
}


/*
 * NAME:    	dbAllocBottomUp()
 *
 * FUNCTION:	alloc the specified block range from the working block
 *		allocation map.
 *
 *		the blocks will be alloc from the working map one dmap
 *		at a time.
 *
 * PARAMETERS:
 *      ip	-  pointer to in-core inode;
 *      blkno	-  starting block number to be freed.
 *      nblocks	-  number of blocks to be freed.
 *
 * RETURN VALUES:
 *      0	- success
 *      -EIO	- i/o error
 */
int dbAllocBottomUp(struct inode *ip, s64 blkno, s64 nblocks)
{
	struct metapage *mp;
	struct dmap *dp;
	int nb, rc;
	s64 lblkno, rem;
	struct inode *ipbmap = JFS_SBI(ip->i_sb)->ipbmap;
	struct bmap *bmp = JFS_SBI(ip->i_sb)->bmap;

	IREAD_LOCK(ipbmap);

	/* block to be allocated better be within the mapsize. */
	ASSERT(nblocks <= bmp->db_mapsize - blkno);

	/*
	 * allocate the blocks a dmap at a time.
	 */
	mp = NULL;
	for (rem = nblocks; rem > 0; rem -= nb, blkno += nb) {
		/* release previous dmap if any */
		if (mp) {
			write_metapage(mp);
		}

		/* get the buffer for the current dmap. */
		lblkno = BLKTODMAP(blkno, bmp->db_l2nbperpage);
		mp = read_metapage(ipbmap, lblkno, PSIZE, 0);
		if (mp == NULL) {
			IREAD_UNLOCK(ipbmap);
			return -EIO;
		}
		dp = (struct dmap *) mp->data;

		/* determine the number of blocks to be allocated from
		 * this dmap.
		 */
		nb = min(rem, BPERDMAP - (blkno & (BPERDMAP - 1)));

		/* allocate the blocks. */
		if ((rc = dbAllocDmapBU(bmp, dp, blkno, nb))) {
			release_metapage(mp);
			IREAD_UNLOCK(ipbmap);
			return (rc);
		}
	}

	/* write the last buffer. */
	write_metapage(mp);

	IREAD_UNLOCK(ipbmap);

	return (0);
}


static int dbAllocDmapBU(struct bmap * bmp, struct dmap * dp, s64 blkno,
			 int nblocks)
{
	int rc;
	int dbitno, word, rembits, nb, nwords, wbitno, agno;
	s8 oldroot, *leaf;
	struct dmaptree *tp = (struct dmaptree *) & dp->tree;

	/* save the current value of the root (i.e. maximum free string)
	 * of the dmap tree.
	 */
	oldroot = tp->stree[ROOT];

	/* pick up a pointer to the leaves of the dmap tree */
	leaf = tp->stree + LEAFIND;

	/* determine the bit number and word within the dmap of the
	 * starting block.
	 */
	dbitno = blkno & (BPERDMAP - 1);
	word = dbitno >> L2DBWORD;

	/* block range better be within the dmap */
	assert(dbitno + nblocks <= BPERDMAP);

	/* allocate the bits of the dmap's words corresponding to the block
	 * range. not all bits of the first and last words may be contained
	 * within the block range.  if this is the case, we'll work against
	 * those words (i.e. partial first and/or last) on an individual basis
	 * (a single pass), allocating the bits of interest by hand and
	 * updating the leaf corresponding to the dmap word. a single pass
	 * will be used for all dmap words fully contained within the
	 * specified range.  within this pass, the bits of all fully contained
	 * dmap words will be marked as free in a single shot and the leaves
	 * will be updated. a single leaf may describe the free space of
	 * multiple dmap words, so we may update only a subset of the actual
	 * leaves corresponding to the dmap words of the block range.
	 */
	for (rembits = nblocks; rembits > 0; rembits -= nb, dbitno += nb) {
		/* determine the bit number within the word and
		 * the number of bits within the word.
		 */
		wbitno = dbitno & (DBWORD - 1);
		nb = min(rembits, DBWORD - wbitno);

		/* check if only part of a word is to be allocated.
		 */
		if (nb < DBWORD) {
			/* allocate (set to 1) the appropriate bits within
			 * this dmap word.
			 */
			dp->wmap[word] |= cpu_to_le32(ONES << (DBWORD - nb)
						      >> wbitno);

			word++;
		} else {
			/* one or more dmap words are fully contained
			 * within the block range.  determine how many
			 * words and allocate (set to 1) the bits of these
			 * words.
			 */
			nwords = rembits >> L2DBWORD;
			memset(&dp->wmap[word], (int) ONES, nwords * 4);

			/* determine how many bits */
			nb = nwords << L2DBWORD;
			word += nwords;
		}
	}

	/* update the free count for this dmap */
	dp->nfree = cpu_to_le32(le32_to_cpu(dp->nfree) - nblocks);

	/* reconstruct summary tree */
	dbInitDmapTree(dp);

	BMAP_LOCK(bmp);

	/* if this allocation group is completely free,
	 * update the highest active allocation group number 
	 * if this allocation group is the new max.
	 */
	agno = blkno >> bmp->db_agl2size;
	if (agno > bmp->db_maxag)
		bmp->db_maxag = agno;

	/* update the free count for the allocation group and map */
	bmp->db_agfree[agno] -= nblocks;
	bmp->db_nfree -= nblocks;

	BMAP_UNLOCK(bmp);

	/* if the root has not changed, done. */
	if (tp->stree[ROOT] == oldroot)
		return (0);

	/* root changed. bubble the change up to the dmap control pages.
	 * if the adjustment of the upper level control pages fails,
	 * backout the bit allocation (thus making everything consistent).
	 */
	if ((rc = dbAdjCtl(bmp, blkno, tp->stree[ROOT], 1, 0)))
		dbFreeBits(bmp, dp, blkno, nblocks);

	return (rc);
}


/*
 * NAME:	dbExtendFS()
 *
 * FUNCTION:	extend bmap from blkno for nblocks;
 * 		dbExtendFS() updates bmap ready for dbAllocBottomUp();
 *
 * L2
 *  |
 *   L1---------------------------------L1
 *    |                                  |
 *     L0---------L0---------L0           L0---------L0---------L0
 *      |          |          |            |          |          |
 *       d0,...,dn  d0,...,dn  d0,...,dn    d0,...,dn  d0,...,dn  d0,.,dm;
 * L2L1L0d0,...,dnL0d0,...,dnL0d0,...,dnL1L0d0,...,dnL0d0,...,dnL0d0,..dm
 *
 * <---old---><----------------------------extend----------------------->   
 */
int dbExtendFS(struct inode *ipbmap, s64 blkno,	s64 nblocks)
{
	struct jfs_sb_info *sbi = JFS_SBI(ipbmap->i_sb);
	int nbperpage = sbi->nbperpage;
	int i, i0 = TRUE, j, j0 = TRUE, k, n;
	s64 newsize;
	s64 p;
	struct metapage *mp, *l2mp, *l1mp = NULL, *l0mp = NULL;
	struct dmapctl *l2dcp, *l1dcp, *l0dcp;
	struct dmap *dp;
	s8 *l0leaf, *l1leaf, *l2leaf;
	struct bmap *bmp = sbi->bmap;
	int agno, l2agsize, oldl2agsize;
	s64 ag_rem;

	newsize = blkno + nblocks;

	jfs_info("dbExtendFS: blkno:%Ld nblocks:%Ld newsize:%Ld",
		 (long long) blkno, (long long) nblocks, (long long) newsize);

	/*
	 *      initialize bmap control page.
	 *
	 * all the data in bmap control page should exclude
	 * the mkfs hidden dmap page.
	 */

	/* update mapsize */
	bmp->db_mapsize = newsize;
	bmp->db_maxlevel = BMAPSZTOLEV(bmp->db_mapsize);

	/* compute new AG size */
	l2agsize = dbGetL2AGSize(newsize);
	oldl2agsize = bmp->db_agl2size;

	bmp->db_agl2size = l2agsize;
	bmp->db_agsize = 1 << l2agsize;

	/* compute new number of AG */
	agno = bmp->db_numag;
	bmp->db_numag = newsize >> l2agsize;
	bmp->db_numag += ((u32) newsize % (u32) bmp->db_agsize) ? 1 : 0;

	/*
	 *      reconfigure db_agfree[] 
	 * from old AG configuration to new AG configuration;
	 *
	 * coalesce contiguous k (newAGSize/oldAGSize) AGs;
	 * i.e., (AGi, ..., AGj) where i = k*n and j = k*(n+1) - 1 to AGn;
	 * note: new AG size = old AG size * (2**x).
	 */
	if (l2agsize == oldl2agsize)
		goto extend;
	k = 1 << (l2agsize - oldl2agsize);
	ag_rem = bmp->db_agfree[0];	/* save agfree[0] */
	for (i = 0, n = 0; i < agno; n++) {
		bmp->db_agfree[n] = 0;	/* init collection point */

		/* coalesce cotiguous k AGs; */
		for (j = 0; j < k && i < agno; j++, i++) {
			/* merge AGi to AGn */
			bmp->db_agfree[n] += bmp->db_agfree[i];
		}
	}
	bmp->db_agfree[0] += ag_rem;	/* restore agfree[0] */

	for (; n < MAXAG; n++)
		bmp->db_agfree[n] = 0;

	/*
	 * update highest active ag number
	 */

	bmp->db_maxag = bmp->db_maxag / k;

	/*
	 *      extend bmap
	 *
	 * update bit maps and corresponding level control pages;
	 * global control page db_nfree, db_agfree[agno], db_maxfreebud;
	 */
      extend:
	/* get L2 page */
	p = BMAPBLKNO + nbperpage;	/* L2 page */
	l2mp = read_metapage(ipbmap, p, PSIZE, 0);
	if (!l2mp) {
		jfs_error(ipbmap->i_sb, "dbExtendFS: L2 page could not be read");
		return -EIO;
	}
	l2dcp = (struct dmapctl *) l2mp->data;

	/* compute start L1 */
	k = blkno >> L2MAXL1SIZE;
	l2leaf = l2dcp->stree + CTLLEAFIND + k;
	p = BLKTOL1(blkno, sbi->l2nbperpage);	/* L1 page */

	/*
	 * extend each L1 in L2
	 */
	for (; k < LPERCTL; k++, p += nbperpage) {
		/* get L1 page */
		if (j0) {
			/* read in L1 page: (blkno & (MAXL1SIZE - 1)) */
			l1mp = read_metapage(ipbmap, p, PSIZE, 0);
			if (l1mp == NULL)
				goto errout;
			l1dcp = (struct dmapctl *) l1mp->data;

			/* compute start L0 */
			j = (blkno & (MAXL1SIZE - 1)) >> L2MAXL0SIZE;
			l1leaf = l1dcp->stree + CTLLEAFIND + j;
			p = BLKTOL0(blkno, sbi->l2nbperpage);
			j0 = FALSE;
		} else {
			/* assign/init L1 page */
			l1mp = get_metapage(ipbmap, p, PSIZE, 0);
			if (l1mp == NULL)
				goto errout;

			l1dcp = (struct dmapctl *) l1mp->data;

			/* compute start L0 */
			j = 0;
			l1leaf = l1dcp->stree + CTLLEAFIND;
			p += nbperpage;	/* 1st L0 of L1.k  */
		}

		/*
		 * extend each L0 in L1
		 */
		for (; j < LPERCTL; j++) {
			/* get L0 page */
			if (i0) {
				/* read in L0 page: (blkno & (MAXL0SIZE - 1)) */

				l0mp = read_metapage(ipbmap, p, PSIZE, 0);
				if (l0mp == NULL)
					goto errout;
				l0dcp = (struct dmapctl *) l0mp->data;

				/* compute start dmap */
				i = (blkno & (MAXL0SIZE - 1)) >>
				    L2BPERDMAP;
				l0leaf = l0dcp->stree + CTLLEAFIND + i;
				p = BLKTODMAP(blkno,
					      sbi->l2nbperpage);
				i0 = FALSE;
			} else {
				/* assign/init L0 page */
				l0mp = get_metapage(ipbmap, p, PSIZE, 0);
				if (l0mp == NULL)
					goto errout;

				l0dcp = (struct dmapctl *) l0mp->data;

				/* compute start dmap */
				i = 0;
				l0leaf = l0dcp->stree + CTLLEAFIND;
				p += nbperpage;	/* 1st dmap of L0.j */
			}

			/*
			 * extend each dmap in L0
			 */
			for (; i < LPERCTL; i++) {
				/*
				 * reconstruct the dmap page, and
				 * initialize corresponding parent L0 leaf
				 */
				if ((n = blkno & (BPERDMAP - 1))) {
					/* read in dmap page: */
					mp = read_metapage(ipbmap, p,
							   PSIZE, 0);
					if (mp == NULL)
						goto errout;
					n = min(nblocks, (s64)BPERDMAP - n);
				} else {
					/* assign/init dmap page */
					mp = read_metapage(ipbmap, p,
							   PSIZE, 0);
					if (mp == NULL)
						goto errout;

					n = min(nblocks, (s64)BPERDMAP);
				}

				dp = (struct dmap *) mp->data;
				*l0leaf = dbInitDmap(dp, blkno, n);

				bmp->db_nfree += n;
				agno = le64_to_cpu(dp->start) >> l2agsize;
				bmp->db_agfree[agno] += n;

				write_metapage(mp);

				l0leaf++;
				p += nbperpage;

				blkno += n;
				nblocks -= n;
				if (nblocks == 0)
					break;
			}	/* for each dmap in a L0 */

			/*
			 * build current L0 page from its leaves, and 
			 * initialize corresponding parent L1 leaf
			 */
			*l1leaf = dbInitDmapCtl(l0dcp, 0, ++i);
			write_metapage(l0mp);
			l0mp = NULL;

			if (nblocks)
				l1leaf++;	/* continue for next L0 */
			else {
				/* more than 1 L0 ? */
				if (j > 0)
					break;	/* build L1 page */
				else {
					/* summarize in global bmap page */
					bmp->db_maxfreebud = *l1leaf;
					release_metapage(l1mp);
					release_metapage(l2mp);
					goto finalize;
				}
			}
		}		/* for each L0 in a L1 */

		/*
		 * build current L1 page from its leaves, and 
		 * initialize corresponding parent L2 leaf
		 */
		*l2leaf = dbInitDmapCtl(l1dcp, 1, ++j);
		write_metapage(l1mp);
		l1mp = NULL;

		if (nblocks)
			l2leaf++;	/* continue for next L1 */
		else {
			/* more than 1 L1 ? */
			if (k > 0)
				break;	/* build L2 page */
			else {
				/* summarize in global bmap page */
				bmp->db_maxfreebud = *l2leaf;
				release_metapage(l2mp);
				goto finalize;
			}
		}
	}			/* for each L1 in a L2 */

	jfs_error(ipbmap->i_sb,
		  "dbExtendFS: function has not returned as expected");
errout:
	if (l0mp)
		release_metapage(l0mp);
	if (l1mp)
		release_metapage(l1mp);
	release_metapage(l2mp);
	return -EIO;

	/*
	 *      finalize bmap control page
	 */
finalize:

	return 0;
}


/*
 *	dbFinalizeBmap()
 */
void dbFinalizeBmap(struct inode *ipbmap)
{
	struct bmap *bmp = JFS_SBI(ipbmap->i_sb)->bmap;
	int actags, inactags, l2nl;
	s64 ag_rem, actfree, inactfree, avgfree;
	int i, n;

	/*
	 *      finalize bmap control page
	 */
//finalize:
	/* 
	 * compute db_agpref: preferred ag to allocate from
	 * (the leftmost ag with average free space in it);
	 */
//agpref:
	/* get the number of active ags and inacitve ags */
	actags = bmp->db_maxag + 1;
	inactags = bmp->db_numag - actags;
	ag_rem = bmp->db_mapsize & (bmp->db_agsize - 1);	/* ??? */

	/* determine how many blocks are in the inactive allocation
	 * groups. in doing this, we must account for the fact that
	 * the rightmost group might be a partial group (i.e. file
	 * system size is not a multiple of the group size).
	 */
	inactfree = (inactags && ag_rem) ?
	    ((inactags - 1) << bmp->db_agl2size) + ag_rem
	    : inactags << bmp->db_agl2size;

	/* determine how many free blocks are in the active
	 * allocation groups plus the average number of free blocks
	 * within the active ags.
	 */
	actfree = bmp->db_nfree - inactfree;
	avgfree = (u32) actfree / (u32) actags;

	/* if the preferred allocation group has not average free space.
	 * re-establish the preferred group as the leftmost
	 * group with average free space.
	 */
	if (bmp->db_agfree[bmp->db_agpref] < avgfree) {
		for (bmp->db_agpref = 0; bmp->db_agpref < actags;
		     bmp->db_agpref++) {
			if (bmp->db_agfree[bmp->db_agpref] >= avgfree)
				break;
		}
		if (bmp->db_agpref >= bmp->db_numag) {
			jfs_error(ipbmap->i_sb,
				  "cannot find ag with average freespace");
		}
	}

	/*
	 * compute db_aglevel, db_agheigth, db_width, db_agstart:
	 * an ag is covered in aglevel dmapctl summary tree, 
	 * at agheight level height (from leaf) with agwidth number of nodes 
	 * each, which starts at agstart index node of the smmary tree node 
	 * array;
	 */
	bmp->db_aglevel = BMAPSZTOLEV(bmp->db_agsize);
	l2nl =
	    bmp->db_agl2size - (L2BPERDMAP + bmp->db_aglevel * L2LPERCTL);
	bmp->db_agheigth = l2nl >> 1;
	bmp->db_agwidth = 1 << (l2nl - (bmp->db_agheigth << 1));
	for (i = 5 - bmp->db_agheigth, bmp->db_agstart = 0, n = 1; i > 0;
	     i--) {
		bmp->db_agstart += n;
		n <<= 2;
	}

}


/*
 * NAME:	dbInitDmap()/ujfs_idmap_page()
 *                                                                    
 * FUNCTION:	initialize working/persistent bitmap of the dmap page
 *		for the specified number of blocks:
 *                                                                    
 *		at entry, the bitmaps had been initialized as free (ZEROS);
 *		The number of blocks will only account for the actually 
 *		existing blocks. Blocks which don't actually exist in 
 *		the aggregate will be marked as allocated (ONES);
 *
 * PARAMETERS:
 *	dp	- pointer to page of map
 *	nblocks	- number of blocks this page
 *
 * RETURNS: NONE
 */
static int dbInitDmap(struct dmap * dp, s64 Blkno, int nblocks)
{
	int blkno, w, b, r, nw, nb, i;

	/* starting block number within the dmap */
	blkno = Blkno & (BPERDMAP - 1);

	if (blkno == 0) {
		dp->nblocks = dp->nfree = cpu_to_le32(nblocks);
		dp->start = cpu_to_le64(Blkno);

		if (nblocks == BPERDMAP) {
			memset(&dp->wmap[0], 0, LPERDMAP * 4);
			memset(&dp->pmap[0], 0, LPERDMAP * 4);
			goto initTree;
		}
	} else {
		dp->nblocks =
		    cpu_to_le32(le32_to_cpu(dp->nblocks) + nblocks);
		dp->nfree = cpu_to_le32(le32_to_cpu(dp->nfree) + nblocks);
	}

	/* word number containing start block number */
	w = blkno >> L2DBWORD;

	/*
	 * free the bits corresponding to the block range (ZEROS):
	 * note: not all bits of the first and last words may be contained 
	 * within the block range.
	 */
	for (r = nblocks; r > 0; r -= nb, blkno += nb) {
		/* number of bits preceding range to be freed in the word */
		b = blkno & (DBWORD - 1);
		/* number of bits to free in the word */
		nb = min(r, DBWORD - b);

		/* is partial word to be freed ? */
		if (nb < DBWORD) {
			/* free (set to 0) from the bitmap word */
			dp->wmap[w] &= cpu_to_le32(~(ONES << (DBWORD - nb)
						     >> b));
			dp->pmap[w] &= cpu_to_le32(~(ONES << (DBWORD - nb)
						     >> b));

			/* skip the word freed */
			w++;
		} else {
			/* free (set to 0) contiguous bitmap words */
			nw = r >> L2DBWORD;
			memset(&dp->wmap[w], 0, nw * 4);
			memset(&dp->pmap[w], 0, nw * 4);

			/* skip the words freed */
			nb = nw << L2DBWORD;
			w += nw;
		}
	}

	/*
	 * mark bits following the range to be freed (non-existing 
	 * blocks) as allocated (ONES)
	 */

	if (blkno == BPERDMAP)
		goto initTree;

	/* the first word beyond the end of existing blocks */
	w = blkno >> L2DBWORD;

	/* does nblocks fall on a 32-bit boundary ? */
	b = blkno & (DBWORD - 1);
	if (b) {
		/* mark a partial word allocated */
		dp->wmap[w] = dp->pmap[w] = cpu_to_le32(ONES >> b);
		w++;
	}

	/* set the rest of the words in the page to allocated (ONES) */
	for (i = w; i < LPERDMAP; i++)
		dp->pmap[i] = dp->wmap[i] = cpu_to_le32(ONES);

	/*
	 * init tree
	 */
      initTree:
	return (dbInitDmapTree(dp));
}


/*
 * NAME:	dbInitDmapTree()/ujfs_complete_dmap()
 *                                                                    
 * FUNCTION:	initialize summary tree of the specified dmap:
 *
 *		at entry, bitmap of the dmap has been initialized;
 *                                                                    
 * PARAMETERS:
 *	dp	- dmap to complete
 *	blkno	- starting block number for this dmap
 *	treemax	- will be filled in with max free for this dmap
 *
 * RETURNS:	max free string at the root of the tree
 */
static int dbInitDmapTree(struct dmap * dp)
{
	struct dmaptree *tp;
	s8 *cp;
	int i;

	/* init fixed info of tree */
	tp = &dp->tree;
	tp->nleafs = cpu_to_le32(LPERDMAP);
	tp->l2nleafs = cpu_to_le32(L2LPERDMAP);
	tp->leafidx = cpu_to_le32(LEAFIND);
	tp->height = cpu_to_le32(4);
	tp->budmin = BUDMIN;

	/* init each leaf from corresponding wmap word:
	 * note: leaf is set to NOFREE(-1) if all blocks of corresponding
	 * bitmap word are allocated. 
	 */
	cp = tp->stree + le32_to_cpu(tp->leafidx);
	for (i = 0; i < LPERDMAP; i++)
		*cp++ = dbMaxBud((u8 *) & dp->wmap[i]);

	/* build the dmap's binary buddy summary tree */
	return (dbInitTree(tp));
}


/*
 * NAME:	dbInitTree()/ujfs_adjtree()
 *                                                                    
 * FUNCTION:	initialize binary buddy summary tree of a dmap or dmapctl.
 *
 *		at entry, the leaves of the tree has been initialized 
 *		from corresponding bitmap word or root of summary tree
 *		of the child control page;
 *		configure binary buddy system at the leaf level, then
 *		bubble up the values of the leaf nodes up the tree.
 *
 * PARAMETERS:
 *	cp	- Pointer to the root of the tree
 *	l2leaves- Number of leaf nodes as a power of 2
 *	l2min	- Number of blocks that can be covered by a leaf
 *		  as a power of 2
 *
 * RETURNS: max free string at the root of the tree
 */
static int dbInitTree(struct dmaptree * dtp)
{
	int l2max, l2free, bsize, nextb, i;
	int child, parent, nparent;
	s8 *tp, *cp, *cp1;

	tp = dtp->stree;

	/* Determine the maximum free string possible for the leaves */
	l2max = le32_to_cpu(dtp->l2nleafs) + dtp->budmin;

	/*
	 * configure the leaf levevl into binary buddy system
	 *
	 * Try to combine buddies starting with a buddy size of 1 
	 * (i.e. two leaves). At a buddy size of 1 two buddy leaves 
	 * can be combined if both buddies have a maximum free of l2min; 
	 * the combination will result in the left-most buddy leaf having 
	 * a maximum free of l2min+1.  
	 * After processing all buddies for a given size, process buddies 
	 * at the next higher buddy size (i.e. current size * 2) and 
	 * the next maximum free (current free + 1).  
	 * This continues until the maximum possible buddy combination 
	 * yields maximum free.
	 */
	for (l2free = dtp->budmin, bsize = 1; l2free < l2max;
	     l2free++, bsize = nextb) {
		/* get next buddy size == current buddy pair size */
		nextb = bsize << 1;

		/* scan each adjacent buddy pair at current buddy size */
		for (i = 0, cp = tp + le32_to_cpu(dtp->leafidx);
		     i < le32_to_cpu(dtp->nleafs);
		     i += nextb, cp += nextb) {
			/* coalesce if both adjacent buddies are max free */
			if (*cp == l2free && *(cp + bsize) == l2free) {
				*cp = l2free + 1;	/* left take right */
				*(cp + bsize) = -1;	/* right give left */
			}
		}
	}

	/*
	 * bubble summary information of leaves up the tree.
	 *
	 * Starting at the leaf node level, the four nodes described by
	 * the higher level parent node are compared for a maximum free and 
	 * this maximum becomes the value of the parent node.  
	 * when all lower level nodes are processed in this fashion then 
	 * move up to the next level (parent becomes a lower level node) and 
	 * continue the process for that level.
	 */
	for (child = le32_to_cpu(dtp->leafidx),
	     nparent = le32_to_cpu(dtp->nleafs) >> 2;
	     nparent > 0; nparent >>= 2, child = parent) {
		/* get index of 1st node of parent level */
		parent = (child - 1) >> 2;

		/* set the value of the parent node as the maximum 
		 * of the four nodes of the current level.
		 */
		for (i = 0, cp = tp + child, cp1 = tp + parent;
		     i < nparent; i++, cp += 4, cp1++)
			*cp1 = TREEMAX(cp);
	}

	return (*tp);
}


/*
 *	dbInitDmapCtl()
 *
 * function: initialize dmapctl page
 */
static int dbInitDmapCtl(struct dmapctl * dcp, int level, int i)
{				/* start leaf index not covered by range */
	s8 *cp;

	dcp->nleafs = cpu_to_le32(LPERCTL);
	dcp->l2nleafs = cpu_to_le32(L2LPERCTL);
	dcp->leafidx = cpu_to_le32(CTLLEAFIND);
	dcp->height = cpu_to_le32(5);
	dcp->budmin = L2BPERDMAP + L2LPERCTL * level;

	/*
	 * initialize the leaves of current level that were not covered 
	 * by the specified input block range (i.e. the leaves have no 
	 * low level dmapctl or dmap).
	 */
	cp = &dcp->stree[CTLLEAFIND + i];
	for (; i < LPERCTL; i++)
		*cp++ = NOFREE;

	/* build the dmap's binary buddy summary tree */
	return (dbInitTree((struct dmaptree *) dcp));
}


/*
 * NAME:	dbGetL2AGSize()/ujfs_getagl2size()
 *                                                                    
 * FUNCTION:	Determine log2(allocation group size) from aggregate size
 *                                                                    
 * PARAMETERS:
 *	nblocks	- Number of blocks in aggregate
 *
 * RETURNS: log2(allocation group size) in aggregate blocks
 */
static int dbGetL2AGSize(s64 nblocks)
{
	s64 sz;
	s64 m;
	int l2sz;

	if (nblocks < BPERDMAP * MAXAG)
		return (L2BPERDMAP);

	/* round up aggregate size to power of 2 */
	m = ((u64) 1 << (64 - 1));
	for (l2sz = 64; l2sz >= 0; l2sz--, m >>= 1) {
		if (m & nblocks)
			break;
	}

	sz = (s64) 1 << l2sz;
	if (sz < nblocks)
		l2sz += 1;

	/* agsize = roundupSize/max_number_of_ag */
	return (l2sz - L2MAXAG);
}


/*
 * NAME:	dbMapFileSizeToMapSize()
 *                                                                    
 * FUNCTION:	compute number of blocks the block allocation map file 
 *		can cover from the map file size;
 *
 * RETURNS:	Number of blocks which can be covered by this block map file;
 */

/*
 * maximum number of map pages at each level including control pages
 */
#define MAXL0PAGES	(1 + LPERCTL)
#define MAXL1PAGES	(1 + LPERCTL * MAXL0PAGES)
#define MAXL2PAGES	(1 + LPERCTL * MAXL1PAGES)

/*
 * convert number of map pages to the zero origin top dmapctl level
 */
#define BMAPPGTOLEV(npages)	\
	(((npages) <= 3 + MAXL0PAGES) ? 0 \
       : ((npages) <= 2 + MAXL1PAGES) ? 1 : 2)

s64 dbMapFileSizeToMapSize(struct inode * ipbmap)
{
	struct super_block *sb = ipbmap->i_sb;
	s64 nblocks;
	s64 npages, ndmaps;
	int level, i;
	int complete, factor;

	nblocks = ipbmap->i_size >> JFS_SBI(sb)->l2bsize;
	npages = nblocks >> JFS_SBI(sb)->l2nbperpage;
	level = BMAPPGTOLEV(npages);

	/* At each level, accumulate the number of dmap pages covered by 
	 * the number of full child levels below it;
	 * repeat for the last incomplete child level.
	 */
	ndmaps = 0;
	npages--;		/* skip the first global control page */
	/* skip higher level control pages above top level covered by map */
	npages -= (2 - level);
	npages--;		/* skip top level's control page */
	for (i = level; i >= 0; i--) {
		factor =
		    (i == 2) ? MAXL1PAGES : ((i == 1) ? MAXL0PAGES : 1);
		complete = (u32) npages / factor;
		ndmaps += complete * ((i == 2) ? LPERCTL * LPERCTL
				      : ((i == 1) ? LPERCTL : 1));

		/* pages in last/incomplete child */
		npages = (u32) npages % factor;
		/* skip incomplete child's level control page */
		npages--;
	}

	/* convert the number of dmaps into the number of blocks 
	 * which can be covered by the dmaps;
	 */
	nblocks = ndmaps << L2BPERDMAP;

	return (nblocks);
}