aboutsummaryrefslogtreecommitdiff
path: root/fs/jfs/jfs_logmgr.c
blob: c5cc03bcae6911d4235a8b20a12fb7b7b1d0b307 (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
/*
 *   Copyright (C) International Business Machines Corp., 2000-2004
 *   Portions Copyright (C) Christoph Hellwig, 2001-2002
 *
 *   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
 */

/*
 *	jfs_logmgr.c: log manager
 *
 * for related information, see transaction manager (jfs_txnmgr.c), and
 * recovery manager (jfs_logredo.c).
 *
 * note: for detail, RTFS.
 *
 *	log buffer manager:
 * special purpose buffer manager supporting log i/o requirements.
 * per log serial pageout of logpage
 * queuing i/o requests and redrive i/o at iodone
 * maintain current logpage buffer
 * no caching since append only
 * appropriate jfs buffer cache buffers as needed
 *
 *	group commit:
 * transactions which wrote COMMIT records in the same in-memory
 * log page during the pageout of previous/current log page(s) are
 * committed together by the pageout of the page.
 *
 *	TBD lazy commit:
 * transactions are committed asynchronously when the log page
 * containing it COMMIT is paged out when it becomes full;
 *
 *	serialization:
 * . a per log lock serialize log write.
 * . a per log lock serialize group commit.
 * . a per log lock serialize log open/close;
 *
 *	TBD log integrity:
 * careful-write (ping-pong) of last logpage to recover from crash
 * in overwrite.
 * detection of split (out-of-order) write of physical sectors
 * of last logpage via timestamp at end of each sector
 * with its mirror data array at trailer).
 *
 *	alternatives:
 * lsn - 64-bit monotonically increasing integer vs
 * 32-bit lspn and page eor.
 */

#include <linux/fs.h>
#include <linux/blkdev.h>
#include <linux/interrupt.h>
#include <linux/smp_lock.h>
#include <linux/completion.h>
#include <linux/buffer_head.h>		/* for sync_blockdev() */
#include <linux/bio.h>
#include <linux/suspend.h>
#include <linux/delay.h>
#include "jfs_incore.h"
#include "jfs_filsys.h"
#include "jfs_metapage.h"
#include "jfs_superblock.h"
#include "jfs_txnmgr.h"
#include "jfs_debug.h"


/*
 * lbuf's ready to be redriven.  Protected by log_redrive_lock (jfsIO thread)
 */
static struct lbuf *log_redrive_list;
static DEFINE_SPINLOCK(log_redrive_lock);
DECLARE_WAIT_QUEUE_HEAD(jfs_IO_thread_wait);


/*
 *	log read/write serialization (per log)
 */
#define LOG_LOCK_INIT(log)	init_MUTEX(&(log)->loglock)
#define LOG_LOCK(log)		down(&((log)->loglock))
#define LOG_UNLOCK(log)		up(&((log)->loglock))


/*
 *	log group commit serialization (per log)
 */

#define LOGGC_LOCK_INIT(log)	spin_lock_init(&(log)->gclock)
#define LOGGC_LOCK(log)		spin_lock_irq(&(log)->gclock)
#define LOGGC_UNLOCK(log)	spin_unlock_irq(&(log)->gclock)
#define LOGGC_WAKEUP(tblk)	wake_up_all(&(tblk)->gcwait)

/*
 *	log sync serialization (per log)
 */
#define	LOGSYNC_DELTA(logsize)		min((logsize)/8, 128*LOGPSIZE)
#define	LOGSYNC_BARRIER(logsize)	((logsize)/4)
/*
#define	LOGSYNC_DELTA(logsize)		min((logsize)/4, 256*LOGPSIZE)
#define	LOGSYNC_BARRIER(logsize)	((logsize)/2)
*/


/*
 *	log buffer cache synchronization
 */
static DEFINE_SPINLOCK(jfsLCacheLock);

#define	LCACHE_LOCK(flags)	spin_lock_irqsave(&jfsLCacheLock, flags)
#define	LCACHE_UNLOCK(flags)	spin_unlock_irqrestore(&jfsLCacheLock, flags)

/*
 * See __SLEEP_COND in jfs_locks.h
 */
#define LCACHE_SLEEP_COND(wq, cond, flags)	\
do {						\
	if (cond)				\
		break;				\
	__SLEEP_COND(wq, cond, LCACHE_LOCK(flags), LCACHE_UNLOCK(flags)); \
} while (0)

#define	LCACHE_WAKEUP(event)	wake_up(event)


/*
 *	lbuf buffer cache (lCache) control
 */
/* log buffer manager pageout control (cumulative, inclusive) */
#define	lbmREAD		0x0001
#define	lbmWRITE	0x0002	/* enqueue at tail of write queue;
				 * init pageout if at head of queue;
				 */
#define	lbmRELEASE	0x0004	/* remove from write queue
				 * at completion of pageout;
				 * do not free/recycle it yet:
				 * caller will free it;
				 */
#define	lbmSYNC		0x0008	/* do not return to freelist
				 * when removed from write queue;
				 */
#define lbmFREE		0x0010	/* return to freelist
				 * at completion of pageout;
				 * the buffer may be recycled;
				 */
#define	lbmDONE		0x0020
#define	lbmERROR	0x0040
#define lbmGC		0x0080	/* lbmIODone to perform post-GC processing
				 * of log page
				 */
#define lbmDIRECT	0x0100

/*
 * Global list of active external journals
 */
static LIST_HEAD(jfs_external_logs);
static struct jfs_log *dummy_log = NULL;
static DECLARE_MUTEX(jfs_log_sem);

/*
 * forward references
 */
static int lmWriteRecord(struct jfs_log * log, struct tblock * tblk,
			 struct lrd * lrd, struct tlock * tlck);

static int lmNextPage(struct jfs_log * log);
static int lmLogFileSystem(struct jfs_log * log, struct jfs_sb_info *sbi,
			   int activate);

static int open_inline_log(struct super_block *sb);
static int open_dummy_log(struct super_block *sb);
static int lbmLogInit(struct jfs_log * log);
static void lbmLogShutdown(struct jfs_log * log);
static struct lbuf *lbmAllocate(struct jfs_log * log, int);
static void lbmFree(struct lbuf * bp);
static void lbmfree(struct lbuf * bp);
static int lbmRead(struct jfs_log * log, int pn, struct lbuf ** bpp);
static void lbmWrite(struct jfs_log * log, struct lbuf * bp, int flag, int cant_block);
static void lbmDirectWrite(struct jfs_log * log, struct lbuf * bp, int flag);
static int lbmIOWait(struct lbuf * bp, int flag);
static bio_end_io_t lbmIODone;
static void lbmStartIO(struct lbuf * bp);
static void lmGCwrite(struct jfs_log * log, int cant_block);
static int lmLogSync(struct jfs_log * log, int nosyncwait);



/*
 *	statistics
 */
#ifdef CONFIG_JFS_STATISTICS
static struct lmStat {
	uint commit;		/* # of commit */
	uint pagedone;		/* # of page written */
	uint submitted;		/* # of pages submitted */
	uint full_page;		/* # of full pages submitted */
	uint partial_page;	/* # of partial pages submitted */
} lmStat;
#endif


/*
 * NAME:	lmLog()
 *
 * FUNCTION:	write a log record;
 *
 * PARAMETER:
 *
 * RETURN:	lsn - offset to the next log record to write (end-of-log);
 *		-1  - error;
 *
 * note: todo: log error handler
 */
int lmLog(struct jfs_log * log, struct tblock * tblk, struct lrd * lrd,
	  struct tlock * tlck)
{
	int lsn;
	int diffp, difft;
	struct metapage *mp = NULL;
	unsigned long flags;

	jfs_info("lmLog: log:0x%p tblk:0x%p, lrd:0x%p tlck:0x%p",
		 log, tblk, lrd, tlck);

	LOG_LOCK(log);

	/* log by (out-of-transaction) JFS ? */
	if (tblk == NULL)
		goto writeRecord;

	/* log from page ? */
	if (tlck == NULL ||
	    tlck->type & tlckBTROOT || (mp = tlck->mp) == NULL)
		goto writeRecord;

	/*
	 *      initialize/update page/transaction recovery lsn
	 */
	lsn = log->lsn;

	LOGSYNC_LOCK(log, flags);

	/*
	 * initialize page lsn if first log write of the page
	 */
	if (mp->lsn == 0) {
		mp->log = log;
		mp->lsn = lsn;
		log->count++;

		/* insert page at tail of logsynclist */
		list_add_tail(&mp->synclist, &log->synclist);
	}

	/*
	 *      initialize/update lsn of tblock of the page
	 *
	 * transaction inherits oldest lsn of pages associated
	 * with allocation/deallocation of resources (their
	 * log records are used to reconstruct allocation map
	 * at recovery time: inode for inode allocation map,
	 * B+-tree index of extent descriptors for block
	 * allocation map);
	 * allocation map pages inherit transaction lsn at
	 * commit time to allow forwarding log syncpt past log
	 * records associated with allocation/deallocation of
	 * resources only after persistent map of these map pages
	 * have been updated and propagated to home.
	 */
	/*
	 * initialize transaction lsn:
	 */
	if (tblk->lsn == 0) {
		/* inherit lsn of its first page logged */
		tblk->lsn = mp->lsn;
		log->count++;

		/* insert tblock after the page on logsynclist */
		list_add(&tblk->synclist, &mp->synclist);
	}
	/*
	 * update transaction lsn:
	 */
	else {
		/* inherit oldest/smallest lsn of page */
		logdiff(diffp, mp->lsn, log);
		logdiff(difft, tblk->lsn, log);
		if (diffp < difft) {
			/* update tblock lsn with page lsn */
			tblk->lsn = mp->lsn;

			/* move tblock after page on logsynclist */
			list_move(&tblk->synclist, &mp->synclist);
		}
	}

	LOGSYNC_UNLOCK(log, flags);

	/*
	 *      write the log record
	 */
      writeRecord:
	lsn = lmWriteRecord(log, tblk, lrd, tlck);

	/*
	 * forward log syncpt if log reached next syncpt trigger
	 */
	logdiff(diffp, lsn, log);
	if (diffp >= log->nextsync)
		lsn = lmLogSync(log, 0);

	/* update end-of-log lsn */
	log->lsn = lsn;

	LOG_UNLOCK(log);

	/* return end-of-log address */
	return lsn;
}

/*
 * NAME:	lmWriteRecord()
 *
 * FUNCTION:	move the log record to current log page
 *
 * PARAMETER:	cd	- commit descriptor
 *
 * RETURN:	end-of-log address
 *			
 * serialization: LOG_LOCK() held on entry/exit
 */
static int
lmWriteRecord(struct jfs_log * log, struct tblock * tblk, struct lrd * lrd,
	      struct tlock * tlck)
{
	int lsn = 0;		/* end-of-log address */
	struct lbuf *bp;	/* dst log page buffer */
	struct logpage *lp;	/* dst log page */
	caddr_t dst;		/* destination address in log page */
	int dstoffset;		/* end-of-log offset in log page */
	int freespace;		/* free space in log page */
	caddr_t p;		/* src meta-data page */
	caddr_t src;
	int srclen;
	int nbytes;		/* number of bytes to move */
	int i;
	int len;
	struct linelock *linelock;
	struct lv *lv;
	struct lvd *lvd;
	int l2linesize;

	len = 0;

	/* retrieve destination log page to write */
	bp = (struct lbuf *) log->bp;
	lp = (struct logpage *) bp->l_ldata;
	dstoffset = log->eor;

	/* any log data to write ? */
	if (tlck == NULL)
		goto moveLrd;

	/*
	 *      move log record data
	 */
	/* retrieve source meta-data page to log */
	if (tlck->flag & tlckPAGELOCK) {
		p = (caddr_t) (tlck->mp->data);
		linelock = (struct linelock *) & tlck->lock;
	}
	/* retrieve source in-memory inode to log */
	else if (tlck->flag & tlckINODELOCK) {
		if (tlck->type & tlckDTREE)
			p = (caddr_t) &JFS_IP(tlck->ip)->i_dtroot;
		else
			p = (caddr_t) &JFS_IP(tlck->ip)->i_xtroot;
		linelock = (struct linelock *) & tlck->lock;
	}
#ifdef	_JFS_WIP
	else if (tlck->flag & tlckINLINELOCK) {

		inlinelock = (struct inlinelock *) & tlck;
		p = (caddr_t) & inlinelock->pxd;
		linelock = (struct linelock *) & tlck;
	}
#endif				/* _JFS_WIP */
	else {
		jfs_err("lmWriteRecord: UFO tlck:0x%p", tlck);
		return 0;	/* Probably should trap */
	}
	l2linesize = linelock->l2linesize;

      moveData:
	ASSERT(linelock->index <= linelock->maxcnt);

	lv = linelock->lv;
	for (i = 0; i < linelock->index; i++, lv++) {
		if (lv->length == 0)
			continue;

		/* is page full ? */
		if (dstoffset >= LOGPSIZE - LOGPTLRSIZE) {
			/* page become full: move on to next page */
			lmNextPage(log);

			bp = log->bp;
			lp = (struct logpage *) bp->l_ldata;
			dstoffset = LOGPHDRSIZE;
		}

		/*
		 * move log vector data
		 */
		src = (u8 *) p + (lv->offset << l2linesize);
		srclen = lv->length << l2linesize;
		len += srclen;
		while (srclen > 0) {
			freespace = (LOGPSIZE - LOGPTLRSIZE) - dstoffset;
			nbytes = min(freespace, srclen);
			dst = (caddr_t) lp + dstoffset;
			memcpy(dst, src, nbytes);
			dstoffset += nbytes;

			/* is page not full ? */
			if (dstoffset < LOGPSIZE - LOGPTLRSIZE)
				break;

			/* page become full: move on to next page */
			lmNextPage(log);

			bp = (struct lbuf *) log->bp;
			lp = (struct logpage *) bp->l_ldata;
			dstoffset = LOGPHDRSIZE;

			srclen -= nbytes;
			src += nbytes;
		}

		/*
		 * move log vector descriptor
		 */
		len += 4;
		lvd = (struct lvd *) ((caddr_t) lp + dstoffset);
		lvd->offset = cpu_to_le16(lv->offset);
		lvd->length = cpu_to_le16(lv->length);
		dstoffset += 4;
		jfs_info("lmWriteRecord: lv offset:%d length:%d",
			 lv->offset, lv->length);
	}

	if ((i = linelock->next)) {
		linelock = (struct linelock *) lid_to_tlock(i);
		goto moveData;
	}

	/*
	 *      move log record descriptor
	 */
      moveLrd:
	lrd->length = cpu_to_le16(len);

	src = (caddr_t) lrd;
	srclen = LOGRDSIZE;

	while (srclen > 0) {
		freespace = (LOGPSIZE - LOGPTLRSIZE) - dstoffset;
		nbytes = min(freespace, srclen);
		dst = (caddr_t) lp + dstoffset;
		memcpy(dst, src, nbytes);

		dstoffset += nbytes;
		srclen -= nbytes;

		/* are there more to move than freespace of page ? */
		if (srclen)
			goto pageFull;

		/*
		 * end of log record descriptor
		 */

		/* update last log record eor */
		log->eor = dstoffset;
		bp->l_eor = dstoffset;
		lsn = (log->page << L2LOGPSIZE) + dstoffset;

		if (lrd->type & cpu_to_le16(LOG_COMMIT)) {
			tblk->clsn = lsn;
			jfs_info("wr: tclsn:0x%x, beor:0x%x", tblk->clsn,
				 bp->l_eor);

			INCREMENT(lmStat.commit);	/* # of commit */

			/*
			 * enqueue tblock for group commit:
			 *
			 * enqueue tblock of non-trivial/synchronous COMMIT
			 * at tail of group commit queue
			 * (trivial/asynchronous COMMITs are ignored by
			 * group commit.)
			 */
			LOGGC_LOCK(log);

			/* init tblock gc state */
			tblk->flag = tblkGC_QUEUE;
			tblk->bp = log->bp;
			tblk->pn = log->page;
			tblk->eor = log->eor;

			/* enqueue transaction to commit queue */
			list_add_tail(&tblk->cqueue, &log->cqueue);

			LOGGC_UNLOCK(log);
		}

		jfs_info("lmWriteRecord: lrd:0x%04x bp:0x%p pn:%d eor:0x%x",
			le16_to_cpu(lrd->type), log->bp, log->page, dstoffset);

		/* page not full ? */
		if (dstoffset < LOGPSIZE - LOGPTLRSIZE)
			return lsn;

	      pageFull:
		/* page become full: move on to next page */
		lmNextPage(log);

		bp = (struct lbuf *) log->bp;
		lp = (struct logpage *) bp->l_ldata;
		dstoffset = LOGPHDRSIZE;
		src += nbytes;
	}

	return lsn;
}


/*
 * NAME:	lmNextPage()
 *
 * FUNCTION:	write current page and allocate next page.
 *
 * PARAMETER:	log
 *
 * RETURN:	0
 *			
 * serialization: LOG_LOCK() held on entry/exit
 */
static int lmNextPage(struct jfs_log * log)
{
	struct logpage *lp;
	int lspn;		/* log sequence page number */
	int pn;			/* current page number */
	struct lbuf *bp;
	struct lbuf *nextbp;
	struct tblock *tblk;

	/* get current log page number and log sequence page number */
	pn = log->page;
	bp = log->bp;
	lp = (struct logpage *) bp->l_ldata;
	lspn = le32_to_cpu(lp->h.page);

	LOGGC_LOCK(log);

	/*
	 *      write or queue the full page at the tail of write queue
	 */
	/* get the tail tblk on commit queue */
	if (list_empty(&log->cqueue))
		tblk = NULL;
	else
		tblk = list_entry(log->cqueue.prev, struct tblock, cqueue);

	/* every tblk who has COMMIT record on the current page,
	 * and has not been committed, must be on commit queue
	 * since tblk is queued at commit queueu at the time
	 * of writing its COMMIT record on the page before
	 * page becomes full (even though the tblk thread
	 * who wrote COMMIT record may have been suspended
	 * currently);
	 */

	/* is page bound with outstanding tail tblk ? */
	if (tblk && tblk->pn == pn) {
		/* mark tblk for end-of-page */
		tblk->flag |= tblkGC_EOP;

		if (log->cflag & logGC_PAGEOUT) {
			/* if page is not already on write queue,
			 * just enqueue (no lbmWRITE to prevent redrive)
			 * buffer to wqueue to ensure correct serial order
			 * of the pages since log pages will be added
			 * continuously
			 */
			if (bp->l_wqnext == NULL)
				lbmWrite(log, bp, 0, 0);
		} else {
			/*
			 * No current GC leader, initiate group commit
			 */
			log->cflag |= logGC_PAGEOUT;
			lmGCwrite(log, 0);
		}
	}
	/* page is not bound with outstanding tblk:
	 * init write or mark it to be redriven (lbmWRITE)
	 */
	else {
		/* finalize the page */
		bp->l_ceor = bp->l_eor;
		lp->h.eor = lp->t.eor = cpu_to_le16(bp->l_ceor);
		lbmWrite(log, bp, lbmWRITE | lbmRELEASE | lbmFREE, 0);
	}
	LOGGC_UNLOCK(log);

	/*
	 *      allocate/initialize next page
	 */
	/* if log wraps, the first data page of log is 2
	 * (0 never used, 1 is superblock).
	 */
	log->page = (pn == log->size - 1) ? 2 : pn + 1;
	log->eor = LOGPHDRSIZE;	/* ? valid page empty/full at logRedo() */

	/* allocate/initialize next log page buffer */
	nextbp = lbmAllocate(log, log->page);
	nextbp->l_eor = log->eor;
	log->bp = nextbp;

	/* initialize next log page */
	lp = (struct logpage *) nextbp->l_ldata;
	lp->h.page = lp->t.page = cpu_to_le32(lspn + 1);
	lp->h.eor = lp->t.eor = cpu_to_le16(LOGPHDRSIZE);

	return 0;
}


/*
 * NAME:	lmGroupCommit()
 *
 * FUNCTION:	group commit
 *	initiate pageout of the pages with COMMIT in the order of
 *	page number - redrive pageout of the page at the head of
 *	pageout queue until full page has been written.
 *
 * RETURN:	
 *
 * NOTE:
 *	LOGGC_LOCK serializes log group commit queue, and
 *	transaction blocks on the commit queue.
 *	N.B. LOG_LOCK is NOT held during lmGroupCommit().
 */
int lmGroupCommit(struct jfs_log * log, struct tblock * tblk)
{
	int rc = 0;

	LOGGC_LOCK(log);

	/* group committed already ? */
	if (tblk->flag & tblkGC_COMMITTED) {
		if (tblk->flag & tblkGC_ERROR)
			rc = -EIO;

		LOGGC_UNLOCK(log);
		return rc;
	}
	jfs_info("lmGroup Commit: tblk = 0x%p, gcrtc = %d", tblk, log->gcrtc);

	if (tblk->xflag & COMMIT_LAZY)
		tblk->flag |= tblkGC_LAZY;

	if ((!(log->cflag & logGC_PAGEOUT)) && (!list_empty(&log->cqueue)) &&
	    (!(tblk->xflag & COMMIT_LAZY) || test_bit(log_FLUSH, &log->flag)
	     || jfs_tlocks_low)) {
		/*
		 * No pageout in progress
		 *
		 * start group commit as its group leader.
		 */
		log->cflag |= logGC_PAGEOUT;

		lmGCwrite(log, 0);
	}

	if (tblk->xflag & COMMIT_LAZY) {
		/*
		 * Lazy transactions can leave now
		 */
		LOGGC_UNLOCK(log);
		return 0;
	}

	/* lmGCwrite gives up LOGGC_LOCK, check again */

	if (tblk->flag & tblkGC_COMMITTED) {
		if (tblk->flag & tblkGC_ERROR)
			rc = -EIO;

		LOGGC_UNLOCK(log);
		return rc;
	}

	/* upcount transaction waiting for completion
	 */
	log->gcrtc++;
	tblk->flag |= tblkGC_READY;

	__SLEEP_COND(tblk->gcwait, (tblk->flag & tblkGC_COMMITTED),
		     LOGGC_LOCK(log), LOGGC_UNLOCK(log));

	/* removed from commit queue */
	if (tblk->flag & tblkGC_ERROR)
		rc = -EIO;

	LOGGC_UNLOCK(log);
	return rc;
}

/*
 * NAME:	lmGCwrite()
 *
 * FUNCTION:	group commit write
 *	initiate write of log page, building a group of all transactions
 *	with commit records on that page.
 *
 * RETURN:	None
 *
 * NOTE:
 *	LOGGC_LOCK must be held by caller.
 *	N.B. LOG_LOCK is NOT held during lmGroupCommit().
 */
static void lmGCwrite(struct jfs_log * log, int cant_write)
{
	struct lbuf *bp;
	struct logpage *lp;
	int gcpn;		/* group commit page number */
	struct tblock *tblk;
	struct tblock *xtblk = NULL;

	/*
	 * build the commit group of a log page
	 *
	 * scan commit queue and make a commit group of all
	 * transactions with COMMIT records on the same log page.
	 */
	/* get the head tblk on the commit queue */
	gcpn = list_entry(log->cqueue.next, struct tblock, cqueue)->pn;

	list_for_each_entry(tblk, &log->cqueue, cqueue) {
		if (tblk->pn != gcpn)
			break;

		xtblk = tblk;

		/* state transition: (QUEUE, READY) -> COMMIT */
		tblk->flag |= tblkGC_COMMIT;
	}
	tblk = xtblk;		/* last tblk of the page */

	/*
	 * pageout to commit transactions on the log page.
	 */
	bp = (struct lbuf *) tblk->bp;
	lp = (struct logpage *) bp->l_ldata;
	/* is page already full ? */
	if (tblk->flag & tblkGC_EOP) {
		/* mark page to free at end of group commit of the page */
		tblk->flag &= ~tblkGC_EOP;
		tblk->flag |= tblkGC_FREE;
		bp->l_ceor = bp->l_eor;
		lp->h.eor = lp->t.eor = cpu_to_le16(bp->l_ceor);
		lbmWrite(log, bp, lbmWRITE | lbmRELEASE | lbmGC,
			 cant_write);
		INCREMENT(lmStat.full_page);
	}
	/* page is not yet full */
	else {
		bp->l_ceor = tblk->eor;	/* ? bp->l_ceor = bp->l_eor; */
		lp->h.eor = lp->t.eor = cpu_to_le16(bp->l_ceor);
		lbmWrite(log, bp, lbmWRITE | lbmGC, cant_write);
		INCREMENT(lmStat.partial_page);
	}
}

/*
 * NAME:	lmPostGC()
 *
 * FUNCTION:	group commit post-processing
 *	Processes transactions after their commit records have been written
 *	to disk, redriving log I/O if necessary.
 *
 * RETURN:	None
 *
 * NOTE:
 *	This routine is called a interrupt time by lbmIODone
 */
static void lmPostGC(struct lbuf * bp)
{
	unsigned long flags;
	struct jfs_log *log = bp->l_log;
	struct logpage *lp;
	struct tblock *tblk, *temp;

	//LOGGC_LOCK(log);
	spin_lock_irqsave(&log->gclock, flags);
	/*
	 * current pageout of group commit completed.
	 *
	 * remove/wakeup transactions from commit queue who were
	 * group committed with the current log page
	 */
	list_for_each_entry_safe(tblk, temp, &log->cqueue, cqueue) {
		if (!(tblk->flag & tblkGC_COMMIT))
			break;
		/* if transaction was marked GC_COMMIT then
		 * it has been shipped in the current pageout
		 * and made it to disk - it is committed.
		 */

		if (bp->l_flag & lbmERROR)
			tblk->flag |= tblkGC_ERROR;

		/* remove it from the commit queue */
		list_del(&tblk->cqueue);
		tblk->flag &= ~tblkGC_QUEUE;

		if (tblk == log->flush_tblk) {
			/* we can stop flushing the log now */
			clear_bit(log_FLUSH, &log->flag);
			log->flush_tblk = NULL;
		}

		jfs_info("lmPostGC: tblk = 0x%p, flag = 0x%x", tblk,
			 tblk->flag);

		if (!(tblk->xflag & COMMIT_FORCE))
			/*
			 * Hand tblk over to lazy commit thread
			 */
			txLazyUnlock(tblk);
		else {
			/* state transition: COMMIT -> COMMITTED */
			tblk->flag |= tblkGC_COMMITTED;

			if (tblk->flag & tblkGC_READY)
				log->gcrtc--;

			LOGGC_WAKEUP(tblk);
		}

		/* was page full before pageout ?
		 * (and this is the last tblk bound with the page)
		 */
		if (tblk->flag & tblkGC_FREE)
			lbmFree(bp);
		/* did page become full after pageout ?
		 * (and this is the last tblk bound with the page)
		 */
		else if (tblk->flag & tblkGC_EOP) {
			/* finalize the page */
			lp = (struct logpage *) bp->l_ldata;
			bp->l_ceor = bp->l_eor;
			lp->h.eor = lp->t.eor = cpu_to_le16(bp->l_eor);
			jfs_info("lmPostGC: calling lbmWrite");
			lbmWrite(log, bp, lbmWRITE | lbmRELEASE | lbmFREE,
				 1);
		}

	}

	/* are there any transactions who have entered lnGroupCommit()
	 * (whose COMMITs are after that of the last log page written.
	 * They are waiting for new group commit (above at (SLEEP 1))
	 * or lazy transactions are on a full (queued) log page,
	 * select the latest ready transaction as new group leader and
	 * wake her up to lead her group.
	 */
	if ((!list_empty(&log->cqueue)) &&
	    ((log->gcrtc > 0) || (tblk->bp->l_wqnext != NULL) ||
	     test_bit(log_FLUSH, &log->flag) || jfs_tlocks_low))
		/*
		 * Call lmGCwrite with new group leader
		 */
		lmGCwrite(log, 1);

	/* no transaction are ready yet (transactions are only just
	 * queued (GC_QUEUE) and not entered for group commit yet).
	 * the first transaction entering group commit
	 * will elect herself as new group leader.
	 */
	else
		log->cflag &= ~logGC_PAGEOUT;

	//LOGGC_UNLOCK(log);
	spin_unlock_irqrestore(&log->gclock, flags);
	return;
}

/*
 * NAME:	lmLogSync()
 *
 * FUNCTION:	write log SYNCPT record for specified log
 *	if new sync address is available
 *	(normally the case if sync() is executed by back-ground
 *	process).
 *	if not, explicitly run jfs_blogsync() to initiate
 *	getting of new sync address.
 *	calculate new value of i_nextsync which determines when
 *	this code is called again.
 *
 * PARAMETERS:	log	- log structure
 * 		nosyncwait - 1 if called asynchronously
 *
 * RETURN:	0
 *			
 * serialization: LOG_LOCK() held on entry/exit
 */
static int lmLogSync(struct jfs_log * log, int nosyncwait)
{
	int logsize;
	int written;		/* written since last syncpt */
	int free;		/* free space left available */
	int delta;		/* additional delta to write normally */
	int more;		/* additional write granted */
	struct lrd lrd;
	int lsn;
	struct logsyncblk *lp;
	struct jfs_sb_info *sbi;
	unsigned long flags;

	/* push dirty metapages out to disk */
	list_for_each_entry(sbi, &log->sb_list, log_list) {
		filemap_flush(sbi->ipbmap->i_mapping);
		filemap_flush(sbi->ipimap->i_mapping);
		filemap_flush(sbi->direct_inode->i_mapping);
	}

	/*
	 *      forward syncpt
	 */
	/* if last sync is same as last syncpt,
	 * invoke sync point forward processing to update sync.
	 */

	if (log->sync == log->syncpt) {
		LOGSYNC_LOCK(log, flags);
		if (list_empty(&log->synclist))
			log->sync = log->lsn;
		else {
			lp = list_entry(log->synclist.next,
					struct logsyncblk, synclist);
			log->sync = lp->lsn;
		}
		LOGSYNC_UNLOCK(log, flags);

	}

	/* if sync is different from last syncpt,
	 * write a SYNCPT record with syncpt = sync.
	 * reset syncpt = sync
	 */
	if (log->sync != log->syncpt) {
		lrd.logtid = 0;
		lrd.backchain = 0;
		lrd.type = cpu_to_le16(LOG_SYNCPT);
		lrd.length = 0;
		lrd.log.syncpt.sync = cpu_to_le32(log->sync);
		lsn = lmWriteRecord(log, NULL, &lrd, NULL);

		log->syncpt = log->sync;
	} else
		lsn = log->lsn;

	/*
	 *      setup next syncpt trigger (SWAG)
	 */
	logsize = log->logsize;

	logdiff(written, lsn, log);
	free = logsize - written;
	delta = LOGSYNC_DELTA(logsize);
	more = min(free / 2, delta);
	if (more < 2 * LOGPSIZE) {
		jfs_warn("\n ... Log Wrap ... Log Wrap ... Log Wrap ...\n");
		/*
		 *      log wrapping
		 *
		 * option 1 - panic ? No.!
		 * option 2 - shutdown file systems
		 *            associated with log ?
		 * option 3 - extend log ?
		 */
		/*
		 * option 4 - second chance
		 *
		 * mark log wrapped, and continue.
		 * when all active transactions are completed,
		 * mark log vaild for recovery.
		 * if crashed during invalid state, log state
		 * implies invald log, forcing fsck().
		 */
		/* mark log state log wrap in log superblock */
		/* log->state = LOGWRAP; */

		/* reset sync point computation */
		log->syncpt = log->sync = lsn;
		log->nextsync = delta;
	} else
		/* next syncpt trigger = written + more */
		log->nextsync = written + more;

	/* return if lmLogSync() from outside of transaction, e.g., sync() */
	if (nosyncwait)
		return lsn;

	/* if number of bytes written from last sync point is more
	 * than 1/4 of the log size, stop new transactions from
	 * starting until all current transactions are completed
	 * by setting syncbarrier flag.
	 */
	if (written > LOGSYNC_BARRIER(logsize) && logsize > 32 * LOGPSIZE) {
		set_bit(log_SYNCBARRIER, &log->flag);
		jfs_info("log barrier on: lsn=0x%x syncpt=0x%x", lsn,
			 log->syncpt);
		/*
		 * We may have to initiate group commit
		 */
		jfs_flush_journal(log, 0);
	}

	return lsn;
}

/*
 * NAME:	jfs_syncpt
 *
 * FUNCTION:	write log SYNCPT record for specified log
 *
 * PARAMETERS:	log	- log structure
 */
void jfs_syncpt(struct jfs_log *log)
{	LOG_LOCK(log);
	lmLogSync(log, 1);
	LOG_UNLOCK(log);
}

/*
 * NAME:	lmLogOpen()
 *
 * FUNCTION:    open the log on first open;
 *	insert filesystem in the active list of the log.
 *
 * PARAMETER:	ipmnt	- file system mount inode
 *		iplog 	- log inode (out)
 *
 * RETURN:
 *
 * serialization:
 */
int lmLogOpen(struct super_block *sb)
{
	int rc;
	struct block_device *bdev;
	struct jfs_log *log;
	struct jfs_sb_info *sbi = JFS_SBI(sb);

	if (sbi->flag & JFS_NOINTEGRITY)
		return open_dummy_log(sb);
	
	if (sbi->mntflag & JFS_INLINELOG)
		return open_inline_log(sb);

	down(&jfs_log_sem);
	list_for_each_entry(log, &jfs_external_logs, journal_list) {
		if (log->bdev->bd_dev == sbi->logdev) {
			if (memcmp(log->uuid, sbi->loguuid,
				   sizeof(log->uuid))) {
				jfs_warn("wrong uuid on JFS journal\n");
				up(&jfs_log_sem);
				return -EINVAL;
			}
			/*
			 * add file system to log active file system list
			 */
			if ((rc = lmLogFileSystem(log, sbi, 1))) {
				up(&jfs_log_sem);
				return rc;
			}
			goto journal_found;
		}
	}

	if (!(log = kmalloc(sizeof(struct jfs_log), GFP_KERNEL))) {
		up(&jfs_log_sem);
		return -ENOMEM;
	}
	memset(log, 0, sizeof(struct jfs_log));
	INIT_LIST_HEAD(&log->sb_list);
	init_waitqueue_head(&log->syncwait);

	/*
	 *      external log as separate logical volume
	 *
	 * file systems to log may have n-to-1 relationship;
	 */

	bdev = open_by_devnum(sbi->logdev, FMODE_READ|FMODE_WRITE);
	if (IS_ERR(bdev)) {
		rc = -PTR_ERR(bdev);
		goto free;
	}

	if ((rc = bd_claim(bdev, log))) {
		goto close;
	}

	log->bdev = bdev;
	memcpy(log->uuid, sbi->loguuid, sizeof(log->uuid));
	
	/*
	 * initialize log:
	 */
	if ((rc = lmLogInit(log)))
		goto unclaim;

	list_add(&log->journal_list, &jfs_external_logs);

	/*
	 * add file system to log active file system list
	 */
	if ((rc = lmLogFileSystem(log, sbi, 1)))
		goto shutdown;

journal_found:
	LOG_LOCK(log);
	list_add(&sbi->log_list, &log->sb_list);
	sbi->log = log;
	LOG_UNLOCK(log);

	up(&jfs_log_sem);
	return 0;

	/*
	 *      unwind on error
	 */
      shutdown:		/* unwind lbmLogInit() */
	list_del(&log->journal_list);
	lbmLogShutdown(log);

      unclaim:
	bd_release(bdev);

      close:		/* close external log device */
	blkdev_put(bdev);

      free:		/* free log descriptor */
	up(&jfs_log_sem);
	kfree(log);

	jfs_warn("lmLogOpen: exit(%d)", rc);
	return rc;
}

static int open_inline_log(struct super_block *sb)
{
	struct jfs_log *log;
	int rc;

	if (!(log = kmalloc(sizeof(struct jfs_log), GFP_KERNEL)))
		return -ENOMEM;
	memset(log, 0, sizeof(struct jfs_log));
	INIT_LIST_HEAD(&log->sb_list);
	init_waitqueue_head(&log->syncwait);

	set_bit(log_INLINELOG, &log->flag);
	log->bdev = sb->s_bdev;
	log->base = addressPXD(&JFS_SBI(sb)->logpxd);
	log->size = lengthPXD(&JFS_SBI(sb)->logpxd) >>
	    (L2LOGPSIZE - sb->s_blocksize_bits);
	log->l2bsize = sb->s_blocksize_bits;
	ASSERT(L2LOGPSIZE >= sb->s_blocksize_bits);

	/*
	 * initialize log.
	 */
	if ((rc = lmLogInit(log))) {
		kfree(log);
		jfs_warn("lmLogOpen: exit(%d)", rc);
		return rc;
	}

	list_add(&JFS_SBI(sb)->log_list, &log->sb_list);
	JFS_SBI(sb)->log = log;

	return rc;
}

static int open_dummy_log(struct super_block *sb)
{
	int rc;

	down(&jfs_log_sem);
	if (!dummy_log) {
		dummy_log = kmalloc(sizeof(struct jfs_log), GFP_KERNEL);
		if (!dummy_log) {
			up(&jfs_log_sem);
			return -ENOMEM;
		}
		memset(dummy_log, 0, sizeof(struct jfs_log));
		INIT_LIST_HEAD(&dummy_log->sb_list);
		init_waitqueue_head(&dummy_log->syncwait);
		dummy_log->no_integrity = 1;
		/* Make up some stuff */
		dummy_log->base = 0;
		dummy_log->size = 1024;
		rc = lmLogInit(dummy_log);
		if (rc) {
			kfree(dummy_log);
			dummy_log = NULL;
			up(&jfs_log_sem);
			return rc;
		}
	}

	LOG_LOCK(dummy_log);
	list_add(&JFS_SBI(sb)->log_list, &dummy_log->sb_list);
	JFS_SBI(sb)->log = dummy_log;
	LOG_UNLOCK(dummy_log);
	up(&jfs_log_sem);

	return 0;
}

/*
 * NAME:	lmLogInit()
 *
 * FUNCTION:	log initialization at first log open.
 *
 *	logredo() (or logformat()) should have been run previously.
 *	initialize the log from log superblock.
 *	set the log state in the superblock to LOGMOUNT and
 *	write SYNCPT log record.
 *		
 * PARAMETER:	log	- log structure
 *
 * RETURN:	0	- if ok
 *		-EINVAL	- bad log magic number or superblock dirty
 *		error returned from logwait()
 *			
 * serialization: single first open thread
 */
int lmLogInit(struct jfs_log * log)
{
	int rc = 0;
	struct lrd lrd;
	struct logsuper *logsuper;
	struct lbuf *bpsuper;
	struct lbuf *bp;
	struct logpage *lp;
	int lsn = 0;

	jfs_info("lmLogInit: log:0x%p", log);

	/* initialize the group commit serialization lock */
	LOGGC_LOCK_INIT(log);

	/* allocate/initialize the log write serialization lock */
	LOG_LOCK_INIT(log);

	LOGSYNC_LOCK_INIT(log);

	INIT_LIST_HEAD(&log->synclist);

	INIT_LIST_HEAD(&log->cqueue);
	log->flush_tblk = NULL;

	log->count = 0;

	/*
	 * initialize log i/o
	 */
	if ((rc = lbmLogInit(log)))
		return rc;

	if (!test_bit(log_INLINELOG, &log->flag))
		log->l2bsize = L2LOGPSIZE;
	
	/* check for disabled journaling to disk */
	if (log->no_integrity) {
		/*
		 * Journal pages will still be filled.  When the time comes
		 * to actually do the I/O, the write is not done, and the
		 * endio routine is called directly.
		 */
		bp = lbmAllocate(log , 0);
		log->bp = bp;
		bp->l_pn = bp->l_eor = 0;
	} else {
		/*
		 * validate log superblock
		 */
		if ((rc = lbmRead(log, 1, &bpsuper)))
			goto errout10;

		logsuper = (struct logsuper *) bpsuper->l_ldata;

		if (logsuper->magic != cpu_to_le32(LOGMAGIC)) {
			jfs_warn("*** Log Format Error ! ***");
			rc = -EINVAL;
			goto errout20;
		}

		/* logredo() should have been run successfully. */
		if (logsuper->state != cpu_to_le32(LOGREDONE)) {
			jfs_warn("*** Log Is Dirty ! ***");
			rc = -EINVAL;
			goto errout20;
		}

		/* initialize log from log superblock */
		if (test_bit(log_INLINELOG,&log->flag)) {
			if (log->size != le32_to_cpu(logsuper->size)) {
				rc = -EINVAL;
				goto errout20;
			}
			jfs_info("lmLogInit: inline log:0x%p base:0x%Lx "
				 "size:0x%x", log,
				 (unsigned long long) log->base, log->size);
		} else {
			if (memcmp(logsuper->uuid, log->uuid, 16)) {
				jfs_warn("wrong uuid on JFS log device");
				goto errout20;
			}
			log->size = le32_to_cpu(logsuper->size);
			log->l2bsize = le32_to_cpu(logsuper->l2bsize);
			jfs_info("lmLogInit: external log:0x%p base:0x%Lx "
				 "size:0x%x", log,
				 (unsigned long long) log->base, log->size);
		}

		log->page = le32_to_cpu(logsuper->end) / LOGPSIZE;
		log->eor = le32_to_cpu(logsuper->end) - (LOGPSIZE * log->page);

		/*
		 * initialize for log append write mode
		 */
		/* establish current/end-of-log page/buffer */
		if ((rc = lbmRead(log, log->page, &bp)))
			goto errout20;

		lp = (struct logpage *) bp->l_ldata;

		jfs_info("lmLogInit: lsn:0x%x page:%d eor:%d:%d",
			 le32_to_cpu(logsuper->end), log->page, log->eor,
			 le16_to_cpu(lp->h.eor));

		log->bp = bp;
		bp->l_pn = log->page;
		bp->l_eor = log->eor;

		/* if current page is full, move on to next page */
		if (log->eor >= LOGPSIZE - LOGPTLRSIZE)
			lmNextPage(log);

		/*
		 * initialize log syncpoint
		 */
		/*
		 * write the first SYNCPT record with syncpoint = 0
		 * (i.e., log redo up to HERE !);
		 * remove current page from lbm write queue at end of pageout
		 * (to write log superblock update), but do not release to
		 * freelist;
		 */
		lrd.logtid = 0;
		lrd.backchain = 0;
		lrd.type = cpu_to_le16(LOG_SYNCPT);
		lrd.length = 0;
		lrd.log.syncpt.sync = 0;
		lsn = lmWriteRecord(log, NULL, &lrd, NULL);
		bp = log->bp;
		bp->l_ceor = bp->l_eor;
		lp = (struct logpage *) bp->l_ldata;
		lp->h.eor = lp->t.eor = cpu_to_le16(bp->l_eor);
		lbmWrite(log, bp, lbmWRITE | lbmSYNC, 0);
		if ((rc = lbmIOWait(bp, 0)))
			goto errout30;

		/*
		 * update/write superblock
		 */
		logsuper->state = cpu_to_le32(LOGMOUNT);
		log->serial = le32_to_cpu(logsuper->serial) + 1;
		logsuper->serial = cpu_to_le32(log->serial);
		lbmDirectWrite(log, bpsuper, lbmWRITE | lbmRELEASE | lbmSYNC);
		if ((rc = lbmIOWait(bpsuper, lbmFREE)))
			goto errout30;
	}

	/* initialize logsync parameters */
	log->logsize = (log->size - 2) << L2LOGPSIZE;
	log->lsn = lsn;
	log->syncpt = lsn;
	log->sync = log->syncpt;
	log->nextsync = LOGSYNC_DELTA(log->logsize);

	jfs_info("lmLogInit: lsn:0x%x syncpt:0x%x sync:0x%x",
		 log->lsn, log->syncpt, log->sync);

	/*
	 * initialize for lazy/group commit
	 */
	log->clsn = lsn;

	return 0;

	/*
	 *      unwind on error
	 */
      errout30:		/* release log page */
	log->wqueue = NULL;
	bp->l_wqnext = NULL;
	lbmFree(bp);

      errout20:		/* release log superblock */
	lbmFree(bpsuper);

      errout10:		/* unwind lbmLogInit() */
	lbmLogShutdown(log);

	jfs_warn("lmLogInit: exit(%d)", rc);
	return rc;
}


/*
 * NAME:	lmLogClose()
 *
 * FUNCTION:	remove file system <ipmnt> from active list of log <iplog>
 *		and close it on last close.
 *
 * PARAMETER:	sb	- superblock
 *
 * RETURN:	errors from subroutines
 *
 * serialization:
 */
int lmLogClose(struct super_block *sb)
{
	struct jfs_sb_info *sbi = JFS_SBI(sb);
	struct jfs_log *log = sbi->log;
	struct block_device *bdev;
	int rc = 0;

	jfs_info("lmLogClose: log:0x%p", log);

	down(&jfs_log_sem);
	LOG_LOCK(log);
	list_del(&sbi->log_list);
	LOG_UNLOCK(log);
	sbi->log = NULL;

	/*
	 * We need to make sure all of the "written" metapages
	 * actually make it to disk
	 */
	sync_blockdev(sb->s_bdev);

	if (test_bit(log_INLINELOG, &log->flag)) {
		/*
		 *      in-line log in host file system
		 */
		rc = lmLogShutdown(log);
		kfree(log);
		goto out;
	}

	if (!log->no_integrity)
		lmLogFileSystem(log, sbi, 0);

	if (!list_empty(&log->sb_list))
		goto out;

	/*
	 * TODO: ensure that the dummy_log is in a state to allow
	 * lbmLogShutdown to deallocate all the buffers and call
	 * kfree against dummy_log.  For now, leave dummy_log & its
	 * buffers in memory, and resuse if another no-integrity mount
	 * is requested.
	 */
	if (log->no_integrity)
		goto out;

	/*
	 *      external log as separate logical volume
	 */
	list_del(&log->journal_list);
	bdev = log->bdev;
	rc = lmLogShutdown(log);

	bd_release(bdev);
	blkdev_put(bdev);

	kfree(log);

      out:
	up(&jfs_log_sem);
	jfs_info("lmLogClose: exit(%d)", rc);
	return rc;
}


/*
 * NAME:	jfs_flush_journal()
 *
 * FUNCTION:	initiate write of any outstanding transactions to the journal
 *		and optionally wait until they are all written to disk
 *
 *		wait == 0  flush until latest txn is committed, don't wait
 *		wait == 1  flush until latest txn is committed, wait
 *		wait > 1   flush until all txn's are complete, wait
 */
void jfs_flush_journal(struct jfs_log *log, int wait)
{
	int i;
	struct tblock *target = NULL;
	struct jfs_sb_info *sbi;

	/* jfs_write_inode may call us during read-only mount */
	if (!log)
		return;

	jfs_info("jfs_flush_journal: log:0x%p wait=%d", log, wait);

	LOGGC_LOCK(log);

	if (!list_empty(&log->cqueue)) {
		/*
		 * This ensures that we will keep writing to the journal as long
		 * as there are unwritten commit records
		 */
		target = list_entry(log->cqueue.prev, struct tblock, cqueue);

		if (test_bit(log_FLUSH, &log->flag)) {
			/*
			 * We're already flushing.
			 * if flush_tblk is NULL, we are flushing everything,
			 * so leave it that way.  Otherwise, update it to the
			 * latest transaction
			 */
			if (log->flush_tblk)
				log->flush_tblk = target;
		} else {
			/* Only flush until latest transaction is committed */
			log->flush_tblk = target;
			set_bit(log_FLUSH, &log->flag);

			/*
			 * Initiate I/O on outstanding transactions
			 */
			if (!(log->cflag & logGC_PAGEOUT)) {
				log->cflag |= logGC_PAGEOUT;
				lmGCwrite(log, 0);
			}
		}
	}
	if ((wait > 1) || test_bit(log_SYNCBARRIER, &log->flag)) {
		/* Flush until all activity complete */
		set_bit(log_FLUSH, &log->flag);
		log->flush_tblk = NULL;
	}

	if (wait && target && !(target->flag & tblkGC_COMMITTED)) {
		DECLARE_WAITQUEUE(__wait, current);

		add_wait_queue(&target->gcwait, &__wait);
		set_current_state(TASK_UNINTERRUPTIBLE);
		LOGGC_UNLOCK(log);
		schedule();
		current->state = TASK_RUNNING;
		LOGGC_LOCK(log);
		remove_wait_queue(&target->gcwait, &__wait);
	}
	LOGGC_UNLOCK(log);

	if (wait < 2)
		return;

	list_for_each_entry(sbi, &log->sb_list, log_list) {
		filemap_fdatawrite(sbi->ipbmap->i_mapping);
		filemap_fdatawrite(sbi->ipimap->i_mapping);
		filemap_fdatawrite(sbi->direct_inode->i_mapping);
	}

	/*
	 * If there was recent activity, we may need to wait
	 * for the lazycommit thread to catch up
	 */
	if ((!list_empty(&log->cqueue)) || !list_empty(&log->synclist)) {
		for (i = 0; i < 200; i++) {	/* Too much? */
			msleep(250);
			if (list_empty(&log->cqueue) &&
			    list_empty(&log->synclist))
				break;
		}
	}
	assert(list_empty(&log->cqueue));
	if (!list_empty(&log->synclist)) {
		struct logsyncblk *lp;

		list_for_each_entry(lp, &log->synclist, synclist) {
			if (lp->xflag & COMMIT_PAGE) {
				struct metapage *mp = (struct metapage *)lp;
				dump_mem("orphan metapage", lp,
					 sizeof(struct metapage));
				dump_mem("page", mp->page, sizeof(struct page));
			}
			else
				dump_mem("orphan tblock", lp,
					 sizeof(struct tblock));
		}
//		current->state = TASK_INTERRUPTIBLE;
//		schedule();
	}
	//assert(list_empty(&log->synclist));
	clear_bit(log_FLUSH, &log->flag);
}

/*
 * NAME:	lmLogShutdown()
 *
 * FUNCTION:	log shutdown at last LogClose().
 *
 *		write log syncpt record.
 *		update super block to set redone flag to 0.
 *
 * PARAMETER:	log	- log inode
 *
 * RETURN:	0	- success
 *			
 * serialization: single last close thread
 */
int lmLogShutdown(struct jfs_log * log)
{
	int rc;
	struct lrd lrd;
	int lsn;
	struct logsuper *logsuper;
	struct lbuf *bpsuper;
	struct lbuf *bp;
	struct logpage *lp;

	jfs_info("lmLogShutdown: log:0x%p", log);

	jfs_flush_journal(log, 2);

	/*
	 * write the last SYNCPT record with syncpoint = 0
	 * (i.e., log redo up to HERE !)
	 */
	lrd.logtid = 0;
	lrd.backchain = 0;
	lrd.type = cpu_to_le16(LOG_SYNCPT);
	lrd.length = 0;
	lrd.log.syncpt.sync = 0;
	
	lsn = lmWriteRecord(log, NULL, &lrd, NULL);
	bp = log->bp;
	lp = (struct logpage *) bp->l_ldata;
	lp->h.eor = lp->t.eor = cpu_to_le16(bp->l_eor);
	lbmWrite(log, log->bp, lbmWRITE | lbmRELEASE | lbmSYNC, 0);
	lbmIOWait(log->bp, lbmFREE);
	log->bp = NULL;

	/*
	 * synchronous update log superblock
	 * mark log state as shutdown cleanly
	 * (i.e., Log does not need to be replayed).
	 */
	if ((rc = lbmRead(log, 1, &bpsuper)))
		goto out;

	logsuper = (struct logsuper *) bpsuper->l_ldata;
	logsuper->state = cpu_to_le32(LOGREDONE);
	logsuper->end = cpu_to_le32(lsn);
	lbmDirectWrite(log, bpsuper, lbmWRITE | lbmRELEASE | lbmSYNC);
	rc = lbmIOWait(bpsuper, lbmFREE);

	jfs_info("lmLogShutdown: lsn:0x%x page:%d eor:%d",
		 lsn, log->page, log->eor);

      out:    
	/*
	 * shutdown per log i/o
	 */
	lbmLogShutdown(log);

	if (rc) {
		jfs_warn("lmLogShutdown: exit(%d)", rc);
	}
	return rc;
}


/*
 * NAME:	lmLogFileSystem()
 *
 * FUNCTION:	insert (<activate> = true)/remove (<activate> = false)
 *	file system into/from log active file system list.
 *
 * PARAMETE:	log	- pointer to logs inode.
 *		fsdev	- kdev_t of filesystem.
 *		serial  - pointer to returned log serial number
 *		activate - insert/remove device from active list.
 *
 * RETURN:	0	- success
 *		errors returned by vms_iowait().
 */
static int lmLogFileSystem(struct jfs_log * log, struct jfs_sb_info *sbi,
			   int activate)
{
	int rc = 0;
	int i;
	struct logsuper *logsuper;
	struct lbuf *bpsuper;
	char *uuid = sbi->uuid;

	/*
	 * insert/remove file system device to log active file system list.
	 */
	if ((rc = lbmRead(log, 1, &bpsuper)))
		return rc;

	logsuper = (struct logsuper *) bpsuper->l_ldata;
	if (activate) {
		for (i = 0; i < MAX_ACTIVE; i++)
			if (!memcmp(logsuper->active[i].uuid, NULL_UUID, 16)) {
				memcpy(logsuper->active[i].uuid, uuid, 16);
				sbi->aggregate = i;
				break;
			}
		if (i == MAX_ACTIVE) {
			jfs_warn("Too many file systems sharing journal!");
			lbmFree(bpsuper);
			return -EMFILE;	/* Is there a better rc? */
		}
	} else {
		for (i = 0; i < MAX_ACTIVE; i++)
			if (!memcmp(logsuper->active[i].uuid, uuid, 16)) {
				memcpy(logsuper->active[i].uuid, NULL_UUID, 16);
				break;
			}
		if (i == MAX_ACTIVE) {
			jfs_warn("Somebody stomped on the journal!");
			lbmFree(bpsuper);
			return -EIO;
		}
		
	}

	/*
	 * synchronous write log superblock:
	 *
	 * write sidestream bypassing write queue:
	 * at file system mount, log super block is updated for
	 * activation of the file system before any log record
	 * (MOUNT record) of the file system, and at file system
	 * unmount, all meta data for the file system has been
	 * flushed before log super block is updated for deactivation
	 * of the file system.
	 */
	lbmDirectWrite(log, bpsuper, lbmWRITE | lbmRELEASE | lbmSYNC);
	rc = lbmIOWait(bpsuper, lbmFREE);

	return rc;
}

/*
 *		log buffer manager (lbm)
 *		------------------------
 *
 * special purpose buffer manager supporting log i/o requirements.
 *
 * per log write queue:
 * log pageout occurs in serial order by fifo write queue and
 * restricting to a single i/o in pregress at any one time.
 * a circular singly-linked list
 * (log->wrqueue points to the tail, and buffers are linked via
 * bp->wrqueue field), and
 * maintains log page in pageout ot waiting for pageout in serial pageout.
 */

/*
 *	lbmLogInit()
 *
 * initialize per log I/O setup at lmLogInit()
 */
static int lbmLogInit(struct jfs_log * log)
{				/* log inode */
	int i;
	struct lbuf *lbuf;

	jfs_info("lbmLogInit: log:0x%p", log);

	/* initialize current buffer cursor */
	log->bp = NULL;

	/* initialize log device write queue */
	log->wqueue = NULL;

	/*
	 * Each log has its own buffer pages allocated to it.  These are
	 * not managed by the page cache.  This ensures that a transaction
	 * writing to the log does not block trying to allocate a page from
	 * the page cache (for the log).  This would be bad, since page
	 * allocation waits on the kswapd thread that may be committing inodes
	 * which would cause log activity.  Was that clear?  I'm trying to
	 * avoid deadlock here.
	 */
	init_waitqueue_head(&log->free_wait);

	log->lbuf_free = NULL;

	for (i = 0; i < LOGPAGES;) {
		char *buffer;
		uint offset;
		struct page *page;

		buffer = (char *) get_zeroed_page(GFP_KERNEL);
		if (buffer == NULL)
			goto error;
		page = virt_to_page(buffer);
		for (offset = 0; offset < PAGE_SIZE; offset += LOGPSIZE) {
			lbuf = kmalloc(sizeof(struct lbuf), GFP_KERNEL);
			if (lbuf == NULL) {
				if (offset == 0)
					free_page((unsigned long) buffer);
				goto error;
			}
			if (offset) /* we already have one reference */
				get_page(page);
			lbuf->l_offset = offset;
			lbuf->l_ldata = buffer + offset;
			lbuf->l_page = page;
			lbuf->l_log = log;
			init_waitqueue_head(&lbuf->l_ioevent);

			lbuf->l_freelist = log->lbuf_free;
			log->lbuf_free = lbuf;
			i++;
		}
	}

	return (0);

      error:
	lbmLogShutdown(log);
	return -ENOMEM;
}


/*
 *	lbmLogShutdown()
 *
 * finalize per log I/O setup at lmLogShutdown()
 */
static void lbmLogShutdown(struct jfs_log * log)
{
	struct lbuf *lbuf;

	jfs_info("lbmLogShutdown: log:0x%p", log);

	lbuf = log->lbuf_free;
	while (lbuf) {
		struct lbuf *next = lbuf->l_freelist;
		__free_page(lbuf->l_page);
		kfree(lbuf);
		lbuf = next;
	}
}


/*
 *	lbmAllocate()
 *
 * allocate an empty log buffer
 */
static struct lbuf *lbmAllocate(struct jfs_log * log, int pn)
{
	struct lbuf *bp;
	unsigned long flags;

	/*
	 * recycle from log buffer freelist if any
	 */
	LCACHE_LOCK(flags);
	LCACHE_SLEEP_COND(log->free_wait, (bp = log->lbuf_free), flags);
	log->lbuf_free = bp->l_freelist;
	LCACHE_UNLOCK(flags);

	bp->l_flag = 0;

	bp->l_wqnext = NULL;
	bp->l_freelist = NULL;

	bp->l_pn = pn;
	bp->l_blkno = log->base + (pn << (L2LOGPSIZE - log->l2bsize));
	bp->l_ceor = 0;

	return bp;
}


/*
 *	lbmFree()
 *
 * release a log buffer to freelist
 */
static void lbmFree(struct lbuf * bp)
{
	unsigned long flags;

	LCACHE_LOCK(flags);

	lbmfree(bp);

	LCACHE_UNLOCK(flags);
}

static void lbmfree(struct lbuf * bp)
{
	struct jfs_log *log = bp->l_log;

	assert(bp->l_wqnext == NULL);

	/*
	 * return the buffer to head of freelist
	 */
	bp->l_freelist = log->lbuf_free;
	log->lbuf_free = bp;

	wake_up(&log->free_wait);
	return;
}


/*
 * NAME:	lbmRedrive
 *
 * FUNCTION:	add a log buffer to the the log redrive list
 *
 * PARAMETER:
 *     bp	- log buffer
 *
 * NOTES:
 *	Takes log_redrive_lock.
 */
static inline void lbmRedrive(struct lbuf *bp)
{
	unsigned long flags;

	spin_lock_irqsave(&log_redrive_lock, flags);
	bp->l_redrive_next = log_redrive_list;
	log_redrive_list = bp;
	spin_unlock_irqrestore(&log_redrive_lock, flags);

	wake_up(&jfs_IO_thread_wait);
}


/*
 *	lbmRead()
 */
static int lbmRead(struct jfs_log * log, int pn, struct lbuf ** bpp)
{
	struct bio *bio;
	struct lbuf *bp;

	/*
	 * allocate a log buffer
	 */
	*bpp = bp = lbmAllocate(log, pn);
	jfs_info("lbmRead: bp:0x%p pn:0x%x", bp, pn);

	bp->l_flag |= lbmREAD;

	bio = bio_alloc(GFP_NOFS, 1);

	bio->bi_sector = bp->l_blkno << (log->l2bsize - 9);
	bio->bi_bdev = log->bdev;
	bio->bi_io_vec[0].bv_page = bp->l_page;
	bio->bi_io_vec[0].bv_len = LOGPSIZE;
	bio->bi_io_vec[0].bv_offset = bp->l_offset;

	bio->bi_vcnt = 1;
	bio->bi_idx = 0;
	bio->bi_size = LOGPSIZE;

	bio->bi_end_io = lbmIODone;
	bio->bi_private = bp;
	submit_bio(READ_SYNC, bio);

	wait_event(bp->l_ioevent, (bp->l_flag != lbmREAD));

	return 0;
}


/*
 *	lbmWrite()
 *
 * buffer at head of pageout queue stays after completion of
 * partial-page pageout and redriven by explicit initiation of
 * pageout by caller until full-page pageout is completed and
 * released.
 *
 * device driver i/o done redrives pageout of new buffer at
 * head of pageout queue when current buffer at head of pageout
 * queue is released at the completion of its full-page pageout.
 *
 * LOGGC_LOCK() serializes lbmWrite() by lmNextPage() and lmGroupCommit().
 * LCACHE_LOCK() serializes xflag between lbmWrite() and lbmIODone()
 */
static void lbmWrite(struct jfs_log * log, struct lbuf * bp, int flag,
		     int cant_block)
{
	struct lbuf *tail;
	unsigned long flags;

	jfs_info("lbmWrite: bp:0x%p flag:0x%x pn:0x%x", bp, flag, bp->l_pn);

	/* map the logical block address to physical block address */
	bp->l_blkno =
	    log->base + (bp->l_pn << (L2LOGPSIZE - log->l2bsize));

	LCACHE_LOCK(flags);		/* disable+lock */

	/*
	 * initialize buffer for device driver
	 */
	bp->l_flag = flag;

	/*
	 *      insert bp at tail of write queue associated with log
	 *
	 * (request is either for bp already/currently at head of queue
	 * or new bp to be inserted at tail)
	 */
	tail = log->wqueue;

	/* is buffer not already on write queue ? */
	if (bp->l_wqnext == NULL) {
		/* insert at tail of wqueue */
		if (tail == NULL) {
			log->wqueue = bp;
			bp->l_wqnext = bp;
		} else {
			log->wqueue = bp;
			bp->l_wqnext = tail->l_wqnext;
			tail->l_wqnext = bp;
		}

		tail = bp;
	}

	/* is buffer at head of wqueue and for write ? */
	if ((bp != tail->l_wqnext) || !(flag & lbmWRITE)) {
		LCACHE_UNLOCK(flags);	/* unlock+enable */
		return;
	}

	LCACHE_UNLOCK(flags);	/* unlock+enable */

	if (cant_block)
		lbmRedrive(bp);
	else if (flag & lbmSYNC)
		lbmStartIO(bp);
	else {
		LOGGC_UNLOCK(log);
		lbmStartIO(bp);
		LOGGC_LOCK(log);
	}
}


/*
 *	lbmDirectWrite()
 *
 * initiate pageout bypassing write queue for sidestream
 * (e.g., log superblock) write;
 */
static void lbmDirectWrite(struct jfs_log * log, struct lbuf * bp, int flag)
{
	jfs_info("lbmDirectWrite: bp:0x%p flag:0x%x pn:0x%x",
		 bp, flag, bp->l_pn);

	/*
	 * initialize buffer for device driver
	 */
	bp->l_flag = flag | lbmDIRECT;

	/* map the logical block address to physical block address */
	bp->l_blkno =
	    log->base + (bp->l_pn << (L2LOGPSIZE - log->l2bsize));

	/*
	 *      initiate pageout of the page
	 */
	lbmStartIO(bp);
}


/*
 * NAME:	lbmStartIO()
 *
 * FUNCTION:	Interface to DD strategy routine
 *
 * RETURN:      none
 *
 * serialization: LCACHE_LOCK() is NOT held during log i/o;
 */
static void lbmStartIO(struct lbuf * bp)
{
	struct bio *bio;
	struct jfs_log *log = bp->l_log;

	jfs_info("lbmStartIO\n");

	bio = bio_alloc(GFP_NOFS, 1);
	bio->bi_sector = bp->l_blkno << (log->l2bsize - 9);
	bio->bi_bdev = log->bdev;
	bio->bi_io_vec[0].bv_page = bp->l_page;
	bio->bi_io_vec[0].bv_len = LOGPSIZE;
	bio->bi_io_vec[0].bv_offset = bp->l_offset;

	bio->bi_vcnt = 1;
	bio->bi_idx = 0;
	bio->bi_size = LOGPSIZE;

	bio->bi_end_io = lbmIODone;
	bio->bi_private = bp;

	/* check if journaling to disk has been disabled */
	if (log->no_integrity) {
		bio->bi_size = 0;
		lbmIODone(bio, 0, 0);
	} else {
		submit_bio(WRITE_SYNC, bio);
		INCREMENT(lmStat.submitted);
	}
}


/*
 *	lbmIOWait()
 */
static int lbmIOWait(struct lbuf * bp, int flag)
{
	unsigned long flags;
	int rc = 0;

	jfs_info("lbmIOWait1: bp:0x%p flag:0x%x:0x%x", bp, bp->l_flag, flag);

	LCACHE_LOCK(flags);		/* disable+lock */

	LCACHE_SLEEP_COND(bp->l_ioevent, (bp->l_flag & lbmDONE), flags);

	rc = (bp->l_flag & lbmERROR) ? -EIO : 0;

	if (flag & lbmFREE)
		lbmfree(bp);

	LCACHE_UNLOCK(flags);	/* unlock+enable */

	jfs_info("lbmIOWait2: bp:0x%p flag:0x%x:0x%x", bp, bp->l_flag, flag);
	return rc;
}

/*
 *	lbmIODone()
 *
 * executed at INTIODONE level
 */
static int lbmIODone(struct bio *bio, unsigned int bytes_done, int error)
{
	struct lbuf *bp = bio->bi_private;
	struct lbuf *nextbp, *tail;
	struct jfs_log *log;
	unsigned long flags;

	if (bio->bi_size)
		return 1;

	/*
	 * get back jfs buffer bound to the i/o buffer
	 */
	jfs_info("lbmIODone: bp:0x%p flag:0x%x", bp, bp->l_flag);

	LCACHE_LOCK(flags);		/* disable+lock */

	bp->l_flag |= lbmDONE;

	if (!test_bit(BIO_UPTODATE, &bio->bi_flags)) {
		bp->l_flag |= lbmERROR;

		jfs_err("lbmIODone: I/O error in JFS log");
	}

	bio_put(bio);

	/*
	 *      pagein completion
	 */
	if (bp->l_flag & lbmREAD) {
		bp->l_flag &= ~lbmREAD;

		LCACHE_UNLOCK(flags);	/* unlock+enable */

		/* wakeup I/O initiator */
		LCACHE_WAKEUP(&bp->l_ioevent);

		return 0;
	}

	/*
	 *      pageout completion
	 *
	 * the bp at the head of write queue has completed pageout.
	 *
	 * if single-commit/full-page pageout, remove the current buffer
	 * from head of pageout queue, and redrive pageout with
	 * the new buffer at head of pageout queue;
	 * otherwise, the partial-page pageout buffer stays at
	 * the head of pageout queue to be redriven for pageout
	 * by lmGroupCommit() until full-page pageout is completed.
	 */
	bp->l_flag &= ~lbmWRITE;
	INCREMENT(lmStat.pagedone);

	/* update committed lsn */
	log = bp->l_log;
	log->clsn = (bp->l_pn << L2LOGPSIZE) + bp->l_ceor;

	if (bp->l_flag & lbmDIRECT) {
		LCACHE_WAKEUP(&bp->l_ioevent);
		LCACHE_UNLOCK(flags);
		return 0;
	}

	tail = log->wqueue;

	/* single element queue */
	if (bp == tail) {
		/* remove head buffer of full-page pageout
		 * from log device write queue
		 */
		if (bp->l_flag & lbmRELEASE) {
			log->wqueue = NULL;
			bp->l_wqnext = NULL;
		}
	}
	/* multi element queue */
	else {
		/* remove head buffer of full-page pageout
		 * from log device write queue
		 */
		if (bp->l_flag & lbmRELEASE) {
			nextbp = tail->l_wqnext = bp->l_wqnext;
			bp->l_wqnext = NULL;

			/*
			 * redrive pageout of next page at head of write queue:
			 * redrive next page without any bound tblk
			 * (i.e., page w/o any COMMIT records), or
			 * first page of new group commit which has been
			 * queued after current page (subsequent pageout
			 * is performed synchronously, except page without
			 * any COMMITs) by lmGroupCommit() as indicated
			 * by lbmWRITE flag;
			 */
			if (nextbp->l_flag & lbmWRITE) {
				/*
				 * We can't do the I/O at interrupt time.
				 * The jfsIO thread can do it
				 */
				lbmRedrive(nextbp);
			}
		}
	}

	/*
	 *      synchronous pageout:
	 *
	 * buffer has not necessarily been removed from write queue
	 * (e.g., synchronous write of partial-page with COMMIT):
	 * leave buffer for i/o initiator to dispose
	 */
	if (bp->l_flag & lbmSYNC) {
		LCACHE_UNLOCK(flags);	/* unlock+enable */

		/* wakeup I/O initiator */
		LCACHE_WAKEUP(&bp->l_ioevent);
	}

	/*
	 *      Group Commit pageout:
	 */
	else if (bp->l_flag & lbmGC) {
		LCACHE_UNLOCK(flags);
		lmPostGC(bp);
	}

	/*
	 *      asynchronous pageout:
	 *
	 * buffer must have been removed from write queue:
	 * insert buffer at head of freelist where it can be recycled
	 */
	else {
		assert(bp->l_flag & lbmRELEASE);
		assert(bp->l_flag & lbmFREE);
		lbmfree(bp);

		LCACHE_UNLOCK(flags);	/* unlock+enable */
	}

	return 0;
}

int jfsIOWait(void *arg)
{
	struct lbuf *bp;

	daemonize("jfsIO");

	complete(&jfsIOwait);

	do {
		DECLARE_WAITQUEUE(wq, current);

		spin_lock_irq(&log_redrive_lock);
		while ((bp = log_redrive_list) != 0) {
			log_redrive_list = bp->l_redrive_next;
			bp->l_redrive_next = NULL;
			spin_unlock_irq(&log_redrive_lock);
			lbmStartIO(bp);
			spin_lock_irq(&log_redrive_lock);
		}
		if (current->flags & PF_FREEZE) {
			spin_unlock_irq(&log_redrive_lock);
			refrigerator(PF_FREEZE);
		} else {
			add_wait_queue(&jfs_IO_thread_wait, &wq);
			set_current_state(TASK_INTERRUPTIBLE);
			spin_unlock_irq(&log_redrive_lock);
			schedule();
			current->state = TASK_RUNNING;
			remove_wait_queue(&jfs_IO_thread_wait, &wq);
		}
	} while (!jfs_stop_threads);

	jfs_info("jfsIOWait being killed!");
	complete_and_exit(&jfsIOwait, 0);
}

/*
 * NAME:	lmLogFormat()/jfs_logform()
 *
 * FUNCTION:	format file system log
 *
 * PARAMETERS:
 *      log	- volume log
 *	logAddress - start address of log space in FS block
 *	logSize	- length of log space in FS block;
 *
 * RETURN:	0	- success
 *		-EIO	- i/o error
 *
 * XXX: We're synchronously writing one page at a time.  This needs to
 *	be improved by writing multiple pages at once.
 */
int lmLogFormat(struct jfs_log *log, s64 logAddress, int logSize)
{
	int rc = -EIO;
	struct jfs_sb_info *sbi;
	struct logsuper *logsuper;
	struct logpage *lp;
	int lspn;		/* log sequence page number */
	struct lrd *lrd_ptr;
	int npages = 0;
	struct lbuf *bp;

	jfs_info("lmLogFormat: logAddress:%Ld logSize:%d",
		 (long long)logAddress, logSize);

	sbi = list_entry(log->sb_list.next, struct jfs_sb_info, log_list);

	/* allocate a log buffer */
	bp = lbmAllocate(log, 1);

	npages = logSize >> sbi->l2nbperpage;

	/*
	 *      log space:
	 *
	 * page 0 - reserved;
	 * page 1 - log superblock;
	 * page 2 - log data page: A SYNC log record is written
	 *          into this page at logform time;
	 * pages 3-N - log data page: set to empty log data pages;
	 */
	/*
	 *      init log superblock: log page 1
	 */
	logsuper = (struct logsuper *) bp->l_ldata;

	logsuper->magic = cpu_to_le32(LOGMAGIC);
	logsuper->version = cpu_to_le32(LOGVERSION);
	logsuper->state = cpu_to_le32(LOGREDONE);
	logsuper->flag = cpu_to_le32(sbi->mntflag);	/* ? */
	logsuper->size = cpu_to_le32(npages);
	logsuper->bsize = cpu_to_le32(sbi->bsize);
	logsuper->l2bsize = cpu_to_le32(sbi->l2bsize);
	logsuper->end = cpu_to_le32(2 * LOGPSIZE + LOGPHDRSIZE + LOGRDSIZE);

	bp->l_flag = lbmWRITE | lbmSYNC | lbmDIRECT;
	bp->l_blkno = logAddress + sbi->nbperpage;
	lbmStartIO(bp);
	if ((rc = lbmIOWait(bp, 0)))
		goto exit;

	/*
	 *      init pages 2 to npages-1 as log data pages:
	 *
	 * log page sequence number (lpsn) initialization:
	 *
	 * pn:   0     1     2     3                 n-1
	 *       +-----+-----+=====+=====+===.....===+=====+
	 * lspn:             N-1   0     1           N-2
	 *                   <--- N page circular file ---->
	 *
	 * the N (= npages-2) data pages of the log is maintained as
	 * a circular file for the log records;
	 * lpsn grows by 1 monotonically as each log page is written
	 * to the circular file of the log;
	 * and setLogpage() will not reset the page number even if
	 * the eor is equal to LOGPHDRSIZE. In order for binary search
	 * still work in find log end process, we have to simulate the
	 * log wrap situation at the log format time.
	 * The 1st log page written will have the highest lpsn. Then
	 * the succeeding log pages will have ascending order of
	 * the lspn starting from 0, ... (N-2)
	 */
	lp = (struct logpage *) bp->l_ldata;
	/*
	 * initialize 1st log page to be written: lpsn = N - 1,
	 * write a SYNCPT log record is written to this page
	 */
	lp->h.page = lp->t.page = cpu_to_le32(npages - 3);
	lp->h.eor = lp->t.eor = cpu_to_le16(LOGPHDRSIZE + LOGRDSIZE);

	lrd_ptr = (struct lrd *) &lp->data;
	lrd_ptr->logtid = 0;
	lrd_ptr->backchain = 0;
	lrd_ptr->type = cpu_to_le16(LOG_SYNCPT);
	lrd_ptr->length = 0;
	lrd_ptr->log.syncpt.sync = 0;

	bp->l_blkno += sbi->nbperpage;
	bp->l_flag = lbmWRITE | lbmSYNC | lbmDIRECT;
	lbmStartIO(bp);
	if ((rc = lbmIOWait(bp, 0)))
		goto exit;

	/*
	 *      initialize succeeding log pages: lpsn = 0, 1, ..., (N-2)
	 */
	for (lspn = 0; lspn < npages - 3; lspn++) {
		lp->h.page = lp->t.page = cpu_to_le32(lspn);
		lp->h.eor = lp->t.eor = cpu_to_le16(LOGPHDRSIZE);

		bp->l_blkno += sbi->nbperpage;
		bp->l_flag = lbmWRITE | lbmSYNC | lbmDIRECT;
		lbmStartIO(bp);
		if ((rc = lbmIOWait(bp, 0)))
			goto exit;
	}

	rc = 0;
exit:
	/*
	 *      finalize log
	 */
	/* release the buffer */
	lbmFree(bp);

	return rc;
}

#ifdef CONFIG_JFS_STATISTICS
int jfs_lmstats_read(char *buffer, char **start, off_t offset, int length,
		      int *eof, void *data)
{
	int len = 0;
	off_t begin;

	len += sprintf(buffer,
		       "JFS Logmgr stats\n"
		       "================\n"
		       "commits = %d\n"
		       "writes submitted = %d\n"
		       "writes completed = %d\n"
		       "full pages submitted = %d\n"
		       "partial pages submitted = %d\n",
		       lmStat.commit,
		       lmStat.submitted,
		       lmStat.pagedone,
		       lmStat.full_page,
		       lmStat.partial_page);

	begin = offset;
	*start = buffer + begin;
	len -= begin;

	if (len > length)
		len = length;
	else
		*eof = 1;

	if (len < 0)
		len = 0;

	return len;
}
#endif /* CONFIG_JFS_STATISTICS */