summaryrefslogtreecommitdiff
path: root/fs/xfs/libxfs/xfs_dir2_node.c
blob: 9df096cc3c37176aa89b4e7145fcd433368ed3ac (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
/*
 * Copyright (c) 2000-2005 Silicon Graphics, Inc.
 * Copyright (c) 2013 Red Hat, Inc.
 * All Rights Reserved.
 *
 * 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.
 *
 * This program is distributed in the hope that it would 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 the Free Software Foundation,
 * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 */
#include "xfs.h"
#include "xfs_fs.h"
#include "xfs_format.h"
#include "xfs_log_format.h"
#include "xfs_trans_resv.h"
#include "xfs_mount.h"
#include "xfs_da_format.h"
#include "xfs_da_btree.h"
#include "xfs_inode.h"
#include "xfs_bmap.h"
#include "xfs_dir2.h"
#include "xfs_dir2_priv.h"
#include "xfs_error.h"
#include "xfs_trace.h"
#include "xfs_trans.h"
#include "xfs_buf_item.h"
#include "xfs_cksum.h"
#include "xfs_log.h"

/*
 * Function declarations.
 */
static int xfs_dir2_leafn_add(struct xfs_buf *bp, xfs_da_args_t *args,
			      int index);
static void xfs_dir2_leafn_rebalance(xfs_da_state_t *state,
				     xfs_da_state_blk_t *blk1,
				     xfs_da_state_blk_t *blk2);
static int xfs_dir2_leafn_remove(xfs_da_args_t *args, struct xfs_buf *bp,
				 int index, xfs_da_state_blk_t *dblk,
				 int *rval);
static int xfs_dir2_node_addname_int(xfs_da_args_t *args,
				     xfs_da_state_blk_t *fblk);

/*
 * Check internal consistency of a leafn block.
 */
#ifdef DEBUG
static xfs_failaddr_t
xfs_dir3_leafn_check(
	struct xfs_inode	*dp,
	struct xfs_buf		*bp)
{
	struct xfs_dir2_leaf	*leaf = bp->b_addr;
	struct xfs_dir3_icleaf_hdr leafhdr;

	dp->d_ops->leaf_hdr_from_disk(&leafhdr, leaf);

	if (leafhdr.magic == XFS_DIR3_LEAFN_MAGIC) {
		struct xfs_dir3_leaf_hdr *leaf3 = bp->b_addr;
		if (be64_to_cpu(leaf3->info.blkno) != bp->b_bn)
			return __this_address;
	} else if (leafhdr.magic != XFS_DIR2_LEAFN_MAGIC)
		return __this_address;

	return xfs_dir3_leaf_check_int(dp->i_mount, dp, &leafhdr, leaf);
}

static inline void
xfs_dir3_leaf_check(
	struct xfs_inode	*dp,
	struct xfs_buf		*bp)
{
	xfs_failaddr_t		fa;

	fa = xfs_dir3_leafn_check(dp, bp);
	if (!fa)
		return;
	xfs_corruption_error(__func__, XFS_ERRLEVEL_LOW, dp->i_mount,
			bp->b_addr, __FILE__, __LINE__, fa);
	ASSERT(0);
}
#else
#define	xfs_dir3_leaf_check(dp, bp)
#endif

static xfs_failaddr_t
xfs_dir3_free_verify(
	struct xfs_buf		*bp)
{
	struct xfs_mount	*mp = bp->b_target->bt_mount;
	struct xfs_dir2_free_hdr *hdr = bp->b_addr;

	if (xfs_sb_version_hascrc(&mp->m_sb)) {
		struct xfs_dir3_blk_hdr *hdr3 = bp->b_addr;

		if (hdr3->magic != cpu_to_be32(XFS_DIR3_FREE_MAGIC))
			return __this_address;
		if (!uuid_equal(&hdr3->uuid, &mp->m_sb.sb_meta_uuid))
			return __this_address;
		if (be64_to_cpu(hdr3->blkno) != bp->b_bn)
			return __this_address;
		if (!xfs_log_check_lsn(mp, be64_to_cpu(hdr3->lsn)))
			return __this_address;
	} else {
		if (hdr->magic != cpu_to_be32(XFS_DIR2_FREE_MAGIC))
			return __this_address;
	}

	/* XXX: should bounds check the xfs_dir3_icfree_hdr here */

	return NULL;
}

static void
xfs_dir3_free_read_verify(
	struct xfs_buf	*bp)
{
	struct xfs_mount	*mp = bp->b_target->bt_mount;
	xfs_failaddr_t		fa;

	if (xfs_sb_version_hascrc(&mp->m_sb) &&
	    !xfs_buf_verify_cksum(bp, XFS_DIR3_FREE_CRC_OFF))
		xfs_verifier_error(bp, -EFSBADCRC, __this_address);
	else {
		fa = xfs_dir3_free_verify(bp);
		if (fa)
			xfs_verifier_error(bp, -EFSCORRUPTED, fa);
	}
}

static void
xfs_dir3_free_write_verify(
	struct xfs_buf	*bp)
{
	struct xfs_mount	*mp = bp->b_target->bt_mount;
	struct xfs_buf_log_item	*bip = bp->b_log_item;
	struct xfs_dir3_blk_hdr	*hdr3 = bp->b_addr;
	xfs_failaddr_t		fa;

	fa = xfs_dir3_free_verify(bp);
	if (fa) {
		xfs_verifier_error(bp, -EFSCORRUPTED, fa);
		return;
	}

	if (!xfs_sb_version_hascrc(&mp->m_sb))
		return;

	if (bip)
		hdr3->lsn = cpu_to_be64(bip->bli_item.li_lsn);

	xfs_buf_update_cksum(bp, XFS_DIR3_FREE_CRC_OFF);
}

const struct xfs_buf_ops xfs_dir3_free_buf_ops = {
	.name = "xfs_dir3_free",
	.verify_read = xfs_dir3_free_read_verify,
	.verify_write = xfs_dir3_free_write_verify,
	.verify_struct = xfs_dir3_free_verify,
};

/* Everything ok in the free block header? */
static xfs_failaddr_t
xfs_dir3_free_header_check(
	struct xfs_inode	*dp,
	xfs_dablk_t		fbno,
	struct xfs_buf		*bp)
{
	struct xfs_mount	*mp = dp->i_mount;
	unsigned int		firstdb;
	int			maxbests;

	maxbests = dp->d_ops->free_max_bests(mp->m_dir_geo);
	firstdb = (xfs_dir2_da_to_db(mp->m_dir_geo, fbno) -
		   xfs_dir2_byte_to_db(mp->m_dir_geo, XFS_DIR2_FREE_OFFSET)) *
			maxbests;
	if (xfs_sb_version_hascrc(&mp->m_sb)) {
		struct xfs_dir3_free_hdr *hdr3 = bp->b_addr;

		if (be32_to_cpu(hdr3->firstdb) != firstdb)
			return __this_address;
		if (be32_to_cpu(hdr3->nvalid) > maxbests)
			return __this_address;
		if (be32_to_cpu(hdr3->nvalid) < be32_to_cpu(hdr3->nused))
			return __this_address;
	} else {
		struct xfs_dir2_free_hdr *hdr = bp->b_addr;

		if (be32_to_cpu(hdr->firstdb) != firstdb)
			return __this_address;
		if (be32_to_cpu(hdr->nvalid) > maxbests)
			return __this_address;
		if (be32_to_cpu(hdr->nvalid) < be32_to_cpu(hdr->nused))
			return __this_address;
	}
	return NULL;
}

static int
__xfs_dir3_free_read(
	struct xfs_trans	*tp,
	struct xfs_inode	*dp,
	xfs_dablk_t		fbno,
	xfs_daddr_t		mappedbno,
	struct xfs_buf		**bpp)
{
	xfs_failaddr_t		fa;
	int			err;

	err = xfs_da_read_buf(tp, dp, fbno, mappedbno, bpp,
				XFS_DATA_FORK, &xfs_dir3_free_buf_ops);
	if (err || !*bpp)
		return err;

	/* Check things that we can't do in the verifier. */
	fa = xfs_dir3_free_header_check(dp, fbno, *bpp);
	if (fa) {
		xfs_verifier_error(*bpp, -EFSCORRUPTED, fa);
		xfs_trans_brelse(tp, *bpp);
		return -EFSCORRUPTED;
	}

	/* try read returns without an error or *bpp if it lands in a hole */
	if (tp)
		xfs_trans_buf_set_type(tp, *bpp, XFS_BLFT_DIR_FREE_BUF);

	return 0;
}

int
xfs_dir2_free_read(
	struct xfs_trans	*tp,
	struct xfs_inode	*dp,
	xfs_dablk_t		fbno,
	struct xfs_buf		**bpp)
{
	return __xfs_dir3_free_read(tp, dp, fbno, -1, bpp);
}

static int
xfs_dir2_free_try_read(
	struct xfs_trans	*tp,
	struct xfs_inode	*dp,
	xfs_dablk_t		fbno,
	struct xfs_buf		**bpp)
{
	return __xfs_dir3_free_read(tp, dp, fbno, -2, bpp);
}

static int
xfs_dir3_free_get_buf(
	xfs_da_args_t		*args,
	xfs_dir2_db_t		fbno,
	struct xfs_buf		**bpp)
{
	struct xfs_trans	*tp = args->trans;
	struct xfs_inode	*dp = args->dp;
	struct xfs_mount	*mp = dp->i_mount;
	struct xfs_buf		*bp;
	int			error;
	struct xfs_dir3_icfree_hdr hdr;

	error = xfs_da_get_buf(tp, dp, xfs_dir2_db_to_da(args->geo, fbno),
				   -1, &bp, XFS_DATA_FORK);
	if (error)
		return error;

	xfs_trans_buf_set_type(tp, bp, XFS_BLFT_DIR_FREE_BUF);
	bp->b_ops = &xfs_dir3_free_buf_ops;

	/*
	 * Initialize the new block to be empty, and remember
	 * its first slot as our empty slot.
	 */
	memset(bp->b_addr, 0, sizeof(struct xfs_dir3_free_hdr));
	memset(&hdr, 0, sizeof(hdr));

	if (xfs_sb_version_hascrc(&mp->m_sb)) {
		struct xfs_dir3_free_hdr *hdr3 = bp->b_addr;

		hdr.magic = XFS_DIR3_FREE_MAGIC;

		hdr3->hdr.blkno = cpu_to_be64(bp->b_bn);
		hdr3->hdr.owner = cpu_to_be64(dp->i_ino);
		uuid_copy(&hdr3->hdr.uuid, &mp->m_sb.sb_meta_uuid);
	} else
		hdr.magic = XFS_DIR2_FREE_MAGIC;
	dp->d_ops->free_hdr_to_disk(bp->b_addr, &hdr);
	*bpp = bp;
	return 0;
}

/*
 * Log entries from a freespace block.
 */
STATIC void
xfs_dir2_free_log_bests(
	struct xfs_da_args	*args,
	struct xfs_buf		*bp,
	int			first,		/* first entry to log */
	int			last)		/* last entry to log */
{
	xfs_dir2_free_t		*free;		/* freespace structure */
	__be16			*bests;

	free = bp->b_addr;
	bests = args->dp->d_ops->free_bests_p(free);
	ASSERT(free->hdr.magic == cpu_to_be32(XFS_DIR2_FREE_MAGIC) ||
	       free->hdr.magic == cpu_to_be32(XFS_DIR3_FREE_MAGIC));
	xfs_trans_log_buf(args->trans, bp,
		(uint)((char *)&bests[first] - (char *)free),
		(uint)((char *)&bests[last] - (char *)free +
		       sizeof(bests[0]) - 1));
}

/*
 * Log header from a freespace block.
 */
static void
xfs_dir2_free_log_header(
	struct xfs_da_args	*args,
	struct xfs_buf		*bp)
{
#ifdef DEBUG
	xfs_dir2_free_t		*free;		/* freespace structure */

	free = bp->b_addr;
	ASSERT(free->hdr.magic == cpu_to_be32(XFS_DIR2_FREE_MAGIC) ||
	       free->hdr.magic == cpu_to_be32(XFS_DIR3_FREE_MAGIC));
#endif
	xfs_trans_log_buf(args->trans, bp, 0,
			  args->dp->d_ops->free_hdr_size - 1);
}

/*
 * Convert a leaf-format directory to a node-format directory.
 * We need to change the magic number of the leaf block, and copy
 * the freespace table out of the leaf block into its own block.
 */
int						/* error */
xfs_dir2_leaf_to_node(
	xfs_da_args_t		*args,		/* operation arguments */
	struct xfs_buf		*lbp)		/* leaf buffer */
{
	xfs_inode_t		*dp;		/* incore directory inode */
	int			error;		/* error return value */
	struct xfs_buf		*fbp;		/* freespace buffer */
	xfs_dir2_db_t		fdb;		/* freespace block number */
	xfs_dir2_free_t		*free;		/* freespace structure */
	__be16			*from;		/* pointer to freespace entry */
	int			i;		/* leaf freespace index */
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	xfs_dir2_leaf_tail_t	*ltp;		/* leaf tail structure */
	int			n;		/* count of live freespc ents */
	xfs_dir2_data_off_t	off;		/* freespace entry value */
	__be16			*to;		/* pointer to freespace entry */
	xfs_trans_t		*tp;		/* transaction pointer */
	struct xfs_dir3_icfree_hdr freehdr;

	trace_xfs_dir2_leaf_to_node(args);

	dp = args->dp;
	tp = args->trans;
	/*
	 * Add a freespace block to the directory.
	 */
	if ((error = xfs_dir2_grow_inode(args, XFS_DIR2_FREE_SPACE, &fdb))) {
		return error;
	}
	ASSERT(fdb == xfs_dir2_byte_to_db(args->geo, XFS_DIR2_FREE_OFFSET));
	/*
	 * Get the buffer for the new freespace block.
	 */
	error = xfs_dir3_free_get_buf(args, fdb, &fbp);
	if (error)
		return error;

	free = fbp->b_addr;
	dp->d_ops->free_hdr_from_disk(&freehdr, free);
	leaf = lbp->b_addr;
	ltp = xfs_dir2_leaf_tail_p(args->geo, leaf);
	if (be32_to_cpu(ltp->bestcount) >
				(uint)dp->i_d.di_size / args->geo->blksize)
		return -EFSCORRUPTED;

	/*
	 * Copy freespace entries from the leaf block to the new block.
	 * Count active entries.
	 */
	from = xfs_dir2_leaf_bests_p(ltp);
	to = dp->d_ops->free_bests_p(free);
	for (i = n = 0; i < be32_to_cpu(ltp->bestcount); i++, from++, to++) {
		if ((off = be16_to_cpu(*from)) != NULLDATAOFF)
			n++;
		*to = cpu_to_be16(off);
	}

	/*
	 * Now initialize the freespace block header.
	 */
	freehdr.nused = n;
	freehdr.nvalid = be32_to_cpu(ltp->bestcount);

	dp->d_ops->free_hdr_to_disk(fbp->b_addr, &freehdr);
	xfs_dir2_free_log_bests(args, fbp, 0, freehdr.nvalid - 1);
	xfs_dir2_free_log_header(args, fbp);

	/*
	 * Converting the leaf to a leafnode is just a matter of changing the
	 * magic number and the ops. Do the change directly to the buffer as
	 * it's less work (and less code) than decoding the header to host
	 * format and back again.
	 */
	if (leaf->hdr.info.magic == cpu_to_be16(XFS_DIR2_LEAF1_MAGIC))
		leaf->hdr.info.magic = cpu_to_be16(XFS_DIR2_LEAFN_MAGIC);
	else
		leaf->hdr.info.magic = cpu_to_be16(XFS_DIR3_LEAFN_MAGIC);
	lbp->b_ops = &xfs_dir3_leafn_buf_ops;
	xfs_trans_buf_set_type(tp, lbp, XFS_BLFT_DIR_LEAFN_BUF);
	xfs_dir3_leaf_log_header(args, lbp);
	xfs_dir3_leaf_check(dp, lbp);
	return 0;
}

/*
 * Add a leaf entry to a leaf block in a node-form directory.
 * The other work necessary is done from the caller.
 */
static int					/* error */
xfs_dir2_leafn_add(
	struct xfs_buf		*bp,		/* leaf buffer */
	xfs_da_args_t		*args,		/* operation arguments */
	int			index)		/* insertion pt for new entry */
{
	int			compact;	/* compacting stale leaves */
	xfs_inode_t		*dp;		/* incore directory inode */
	int			highstale;	/* next stale entry */
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	xfs_dir2_leaf_entry_t	*lep;		/* leaf entry */
	int			lfloghigh;	/* high leaf entry logging */
	int			lfloglow;	/* low leaf entry logging */
	int			lowstale;	/* previous stale entry */
	struct xfs_dir3_icleaf_hdr leafhdr;
	struct xfs_dir2_leaf_entry *ents;

	trace_xfs_dir2_leafn_add(args, index);

	dp = args->dp;
	leaf = bp->b_addr;
	dp->d_ops->leaf_hdr_from_disk(&leafhdr, leaf);
	ents = dp->d_ops->leaf_ents_p(leaf);

	/*
	 * Quick check just to make sure we are not going to index
	 * into other peoples memory
	 */
	if (index < 0)
		return -EFSCORRUPTED;

	/*
	 * If there are already the maximum number of leaf entries in
	 * the block, if there are no stale entries it won't fit.
	 * Caller will do a split.  If there are stale entries we'll do
	 * a compact.
	 */

	if (leafhdr.count == dp->d_ops->leaf_max_ents(args->geo)) {
		if (!leafhdr.stale)
			return -ENOSPC;
		compact = leafhdr.stale > 1;
	} else
		compact = 0;
	ASSERT(index == 0 || be32_to_cpu(ents[index - 1].hashval) <= args->hashval);
	ASSERT(index == leafhdr.count ||
	       be32_to_cpu(ents[index].hashval) >= args->hashval);

	if (args->op_flags & XFS_DA_OP_JUSTCHECK)
		return 0;

	/*
	 * Compact out all but one stale leaf entry.  Leaves behind
	 * the entry closest to index.
	 */
	if (compact)
		xfs_dir3_leaf_compact_x1(&leafhdr, ents, &index, &lowstale,
					 &highstale, &lfloglow, &lfloghigh);
	else if (leafhdr.stale) {
		/*
		 * Set impossible logging indices for this case.
		 */
		lfloglow = leafhdr.count;
		lfloghigh = -1;
	}

	/*
	 * Insert the new entry, log everything.
	 */
	lep = xfs_dir3_leaf_find_entry(&leafhdr, ents, index, compact, lowstale,
				       highstale, &lfloglow, &lfloghigh);

	lep->hashval = cpu_to_be32(args->hashval);
	lep->address = cpu_to_be32(xfs_dir2_db_off_to_dataptr(args->geo,
				args->blkno, args->index));

	dp->d_ops->leaf_hdr_to_disk(leaf, &leafhdr);
	xfs_dir3_leaf_log_header(args, bp);
	xfs_dir3_leaf_log_ents(args, bp, lfloglow, lfloghigh);
	xfs_dir3_leaf_check(dp, bp);
	return 0;
}

#ifdef DEBUG
static void
xfs_dir2_free_hdr_check(
	struct xfs_inode *dp,
	struct xfs_buf	*bp,
	xfs_dir2_db_t	db)
{
	struct xfs_dir3_icfree_hdr hdr;

	dp->d_ops->free_hdr_from_disk(&hdr, bp->b_addr);

	ASSERT((hdr.firstdb %
		dp->d_ops->free_max_bests(dp->i_mount->m_dir_geo)) == 0);
	ASSERT(hdr.firstdb <= db);
	ASSERT(db < hdr.firstdb + hdr.nvalid);
}
#else
#define xfs_dir2_free_hdr_check(dp, bp, db)
#endif	/* DEBUG */

/*
 * Return the last hash value in the leaf.
 * Stale entries are ok.
 */
xfs_dahash_t					/* hash value */
xfs_dir2_leaf_lasthash(
	struct xfs_inode *dp,
	struct xfs_buf	*bp,			/* leaf buffer */
	int		*count)			/* count of entries in leaf */
{
	struct xfs_dir2_leaf	*leaf = bp->b_addr;
	struct xfs_dir2_leaf_entry *ents;
	struct xfs_dir3_icleaf_hdr leafhdr;

	dp->d_ops->leaf_hdr_from_disk(&leafhdr, leaf);

	ASSERT(leafhdr.magic == XFS_DIR2_LEAFN_MAGIC ||
	       leafhdr.magic == XFS_DIR3_LEAFN_MAGIC ||
	       leafhdr.magic == XFS_DIR2_LEAF1_MAGIC ||
	       leafhdr.magic == XFS_DIR3_LEAF1_MAGIC);

	if (count)
		*count = leafhdr.count;
	if (!leafhdr.count)
		return 0;

	ents = dp->d_ops->leaf_ents_p(leaf);
	return be32_to_cpu(ents[leafhdr.count - 1].hashval);
}

/*
 * Look up a leaf entry for space to add a name in a node-format leaf block.
 * The extrablk in state is a freespace block.
 */
STATIC int
xfs_dir2_leafn_lookup_for_addname(
	struct xfs_buf		*bp,		/* leaf buffer */
	xfs_da_args_t		*args,		/* operation arguments */
	int			*indexp,	/* out: leaf entry index */
	xfs_da_state_t		*state)		/* state to fill in */
{
	struct xfs_buf		*curbp = NULL;	/* current data/free buffer */
	xfs_dir2_db_t		curdb = -1;	/* current data block number */
	xfs_dir2_db_t		curfdb = -1;	/* current free block number */
	xfs_inode_t		*dp;		/* incore directory inode */
	int			error;		/* error return value */
	int			fi;		/* free entry index */
	xfs_dir2_free_t		*free = NULL;	/* free block structure */
	int			index;		/* leaf entry index */
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	int			length;		/* length of new data entry */
	xfs_dir2_leaf_entry_t	*lep;		/* leaf entry */
	xfs_mount_t		*mp;		/* filesystem mount point */
	xfs_dir2_db_t		newdb;		/* new data block number */
	xfs_dir2_db_t		newfdb;		/* new free block number */
	xfs_trans_t		*tp;		/* transaction pointer */
	struct xfs_dir2_leaf_entry *ents;
	struct xfs_dir3_icleaf_hdr leafhdr;

	dp = args->dp;
	tp = args->trans;
	mp = dp->i_mount;
	leaf = bp->b_addr;
	dp->d_ops->leaf_hdr_from_disk(&leafhdr, leaf);
	ents = dp->d_ops->leaf_ents_p(leaf);

	xfs_dir3_leaf_check(dp, bp);
	ASSERT(leafhdr.count > 0);

	/*
	 * Look up the hash value in the leaf entries.
	 */
	index = xfs_dir2_leaf_search_hash(args, bp);
	/*
	 * Do we have a buffer coming in?
	 */
	if (state->extravalid) {
		/* If so, it's a free block buffer, get the block number. */
		curbp = state->extrablk.bp;
		curfdb = state->extrablk.blkno;
		free = curbp->b_addr;
		ASSERT(free->hdr.magic == cpu_to_be32(XFS_DIR2_FREE_MAGIC) ||
		       free->hdr.magic == cpu_to_be32(XFS_DIR3_FREE_MAGIC));
	}
	length = dp->d_ops->data_entsize(args->namelen);
	/*
	 * Loop over leaf entries with the right hash value.
	 */
	for (lep = &ents[index];
	     index < leafhdr.count && be32_to_cpu(lep->hashval) == args->hashval;
	     lep++, index++) {
		/*
		 * Skip stale leaf entries.
		 */
		if (be32_to_cpu(lep->address) == XFS_DIR2_NULL_DATAPTR)
			continue;
		/*
		 * Pull the data block number from the entry.
		 */
		newdb = xfs_dir2_dataptr_to_db(args->geo,
					       be32_to_cpu(lep->address));
		/*
		 * For addname, we're looking for a place to put the new entry.
		 * We want to use a data block with an entry of equal
		 * hash value to ours if there is one with room.
		 *
		 * If this block isn't the data block we already have
		 * in hand, take a look at it.
		 */
		if (newdb != curdb) {
			__be16 *bests;

			curdb = newdb;
			/*
			 * Convert the data block to the free block
			 * holding its freespace information.
			 */
			newfdb = dp->d_ops->db_to_fdb(args->geo, newdb);
			/*
			 * If it's not the one we have in hand, read it in.
			 */
			if (newfdb != curfdb) {
				/*
				 * If we had one before, drop it.
				 */
				if (curbp)
					xfs_trans_brelse(tp, curbp);

				error = xfs_dir2_free_read(tp, dp,
						xfs_dir2_db_to_da(args->geo,
								  newfdb),
						&curbp);
				if (error)
					return error;
				free = curbp->b_addr;

				xfs_dir2_free_hdr_check(dp, curbp, curdb);
			}
			/*
			 * Get the index for our entry.
			 */
			fi = dp->d_ops->db_to_fdindex(args->geo, curdb);
			/*
			 * If it has room, return it.
			 */
			bests = dp->d_ops->free_bests_p(free);
			if (unlikely(bests[fi] == cpu_to_be16(NULLDATAOFF))) {
				XFS_ERROR_REPORT("xfs_dir2_leafn_lookup_int",
							XFS_ERRLEVEL_LOW, mp);
				if (curfdb != newfdb)
					xfs_trans_brelse(tp, curbp);
				return -EFSCORRUPTED;
			}
			curfdb = newfdb;
			if (be16_to_cpu(bests[fi]) >= length)
				goto out;
		}
	}
	/* Didn't find any space */
	fi = -1;
out:
	ASSERT(args->op_flags & XFS_DA_OP_OKNOENT);
	if (curbp) {
		/* Giving back a free block. */
		state->extravalid = 1;
		state->extrablk.bp = curbp;
		state->extrablk.index = fi;
		state->extrablk.blkno = curfdb;

		/*
		 * Important: this magic number is not in the buffer - it's for
		 * buffer type information and therefore only the free/data type
		 * matters here, not whether CRCs are enabled or not.
		 */
		state->extrablk.magic = XFS_DIR2_FREE_MAGIC;
	} else {
		state->extravalid = 0;
	}
	/*
	 * Return the index, that will be the insertion point.
	 */
	*indexp = index;
	return -ENOENT;
}

/*
 * Look up a leaf entry in a node-format leaf block.
 * The extrablk in state a data block.
 */
STATIC int
xfs_dir2_leafn_lookup_for_entry(
	struct xfs_buf		*bp,		/* leaf buffer */
	xfs_da_args_t		*args,		/* operation arguments */
	int			*indexp,	/* out: leaf entry index */
	xfs_da_state_t		*state)		/* state to fill in */
{
	struct xfs_buf		*curbp = NULL;	/* current data/free buffer */
	xfs_dir2_db_t		curdb = -1;	/* current data block number */
	xfs_dir2_data_entry_t	*dep;		/* data block entry */
	xfs_inode_t		*dp;		/* incore directory inode */
	int			error;		/* error return value */
	int			index;		/* leaf entry index */
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	xfs_dir2_leaf_entry_t	*lep;		/* leaf entry */
	xfs_mount_t		*mp;		/* filesystem mount point */
	xfs_dir2_db_t		newdb;		/* new data block number */
	xfs_trans_t		*tp;		/* transaction pointer */
	enum xfs_dacmp		cmp;		/* comparison result */
	struct xfs_dir2_leaf_entry *ents;
	struct xfs_dir3_icleaf_hdr leafhdr;

	dp = args->dp;
	tp = args->trans;
	mp = dp->i_mount;
	leaf = bp->b_addr;
	dp->d_ops->leaf_hdr_from_disk(&leafhdr, leaf);
	ents = dp->d_ops->leaf_ents_p(leaf);

	xfs_dir3_leaf_check(dp, bp);
	ASSERT(leafhdr.count > 0);

	/*
	 * Look up the hash value in the leaf entries.
	 */
	index = xfs_dir2_leaf_search_hash(args, bp);
	/*
	 * Do we have a buffer coming in?
	 */
	if (state->extravalid) {
		curbp = state->extrablk.bp;
		curdb = state->extrablk.blkno;
	}
	/*
	 * Loop over leaf entries with the right hash value.
	 */
	for (lep = &ents[index];
	     index < leafhdr.count && be32_to_cpu(lep->hashval) == args->hashval;
	     lep++, index++) {
		/*
		 * Skip stale leaf entries.
		 */
		if (be32_to_cpu(lep->address) == XFS_DIR2_NULL_DATAPTR)
			continue;
		/*
		 * Pull the data block number from the entry.
		 */
		newdb = xfs_dir2_dataptr_to_db(args->geo,
					       be32_to_cpu(lep->address));
		/*
		 * Not adding a new entry, so we really want to find
		 * the name given to us.
		 *
		 * If it's a different data block, go get it.
		 */
		if (newdb != curdb) {
			/*
			 * If we had a block before that we aren't saving
			 * for a CI name, drop it
			 */
			if (curbp && (args->cmpresult == XFS_CMP_DIFFERENT ||
						curdb != state->extrablk.blkno))
				xfs_trans_brelse(tp, curbp);
			/*
			 * If needing the block that is saved with a CI match,
			 * use it otherwise read in the new data block.
			 */
			if (args->cmpresult != XFS_CMP_DIFFERENT &&
					newdb == state->extrablk.blkno) {
				ASSERT(state->extravalid);
				curbp = state->extrablk.bp;
			} else {
				error = xfs_dir3_data_read(tp, dp,
						xfs_dir2_db_to_da(args->geo,
								  newdb),
						-1, &curbp);
				if (error)
					return error;
			}
			xfs_dir3_data_check(dp, curbp);
			curdb = newdb;
		}
		/*
		 * Point to the data entry.
		 */
		dep = (xfs_dir2_data_entry_t *)((char *)curbp->b_addr +
			xfs_dir2_dataptr_to_off(args->geo,
						be32_to_cpu(lep->address)));
		/*
		 * Compare the entry and if it's an exact match, return
		 * EEXIST immediately. If it's the first case-insensitive
		 * match, store the block & inode number and continue looking.
		 */
		cmp = mp->m_dirnameops->compname(args, dep->name, dep->namelen);
		if (cmp != XFS_CMP_DIFFERENT && cmp != args->cmpresult) {
			/* If there is a CI match block, drop it */
			if (args->cmpresult != XFS_CMP_DIFFERENT &&
						curdb != state->extrablk.blkno)
				xfs_trans_brelse(tp, state->extrablk.bp);
			args->cmpresult = cmp;
			args->inumber = be64_to_cpu(dep->inumber);
			args->filetype = dp->d_ops->data_get_ftype(dep);
			*indexp = index;
			state->extravalid = 1;
			state->extrablk.bp = curbp;
			state->extrablk.blkno = curdb;
			state->extrablk.index = (int)((char *)dep -
							(char *)curbp->b_addr);
			state->extrablk.magic = XFS_DIR2_DATA_MAGIC;
			curbp->b_ops = &xfs_dir3_data_buf_ops;
			xfs_trans_buf_set_type(tp, curbp, XFS_BLFT_DIR_DATA_BUF);
			if (cmp == XFS_CMP_EXACT)
				return -EEXIST;
		}
	}
	ASSERT(index == leafhdr.count || (args->op_flags & XFS_DA_OP_OKNOENT));
	if (curbp) {
		if (args->cmpresult == XFS_CMP_DIFFERENT) {
			/* Giving back last used data block. */
			state->extravalid = 1;
			state->extrablk.bp = curbp;
			state->extrablk.index = -1;
			state->extrablk.blkno = curdb;
			state->extrablk.magic = XFS_DIR2_DATA_MAGIC;
			curbp->b_ops = &xfs_dir3_data_buf_ops;
			xfs_trans_buf_set_type(tp, curbp, XFS_BLFT_DIR_DATA_BUF);
		} else {
			/* If the curbp is not the CI match block, drop it */
			if (state->extrablk.bp != curbp)
				xfs_trans_brelse(tp, curbp);
		}
	} else {
		state->extravalid = 0;
	}
	*indexp = index;
	return -ENOENT;
}

/*
 * Look up a leaf entry in a node-format leaf block.
 * If this is an addname then the extrablk in state is a freespace block,
 * otherwise it's a data block.
 */
int
xfs_dir2_leafn_lookup_int(
	struct xfs_buf		*bp,		/* leaf buffer */
	xfs_da_args_t		*args,		/* operation arguments */
	int			*indexp,	/* out: leaf entry index */
	xfs_da_state_t		*state)		/* state to fill in */
{
	if (args->op_flags & XFS_DA_OP_ADDNAME)
		return xfs_dir2_leafn_lookup_for_addname(bp, args, indexp,
							state);
	return xfs_dir2_leafn_lookup_for_entry(bp, args, indexp, state);
}

/*
 * Move count leaf entries from source to destination leaf.
 * Log entries and headers.  Stale entries are preserved.
 */
static void
xfs_dir3_leafn_moveents(
	xfs_da_args_t			*args,	/* operation arguments */
	struct xfs_buf			*bp_s,	/* source */
	struct xfs_dir3_icleaf_hdr	*shdr,
	struct xfs_dir2_leaf_entry	*sents,
	int				start_s,/* source leaf index */
	struct xfs_buf			*bp_d,	/* destination */
	struct xfs_dir3_icleaf_hdr	*dhdr,
	struct xfs_dir2_leaf_entry	*dents,
	int				start_d,/* destination leaf index */
	int				count)	/* count of leaves to copy */
{
	int				stale;	/* count stale leaves copied */

	trace_xfs_dir2_leafn_moveents(args, start_s, start_d, count);

	/*
	 * Silently return if nothing to do.
	 */
	if (count == 0)
		return;

	/*
	 * If the destination index is not the end of the current
	 * destination leaf entries, open up a hole in the destination
	 * to hold the new entries.
	 */
	if (start_d < dhdr->count) {
		memmove(&dents[start_d + count], &dents[start_d],
			(dhdr->count - start_d) * sizeof(xfs_dir2_leaf_entry_t));
		xfs_dir3_leaf_log_ents(args, bp_d, start_d + count,
				       count + dhdr->count - 1);
	}
	/*
	 * If the source has stale leaves, count the ones in the copy range
	 * so we can update the header correctly.
	 */
	if (shdr->stale) {
		int	i;			/* temp leaf index */

		for (i = start_s, stale = 0; i < start_s + count; i++) {
			if (sents[i].address ==
					cpu_to_be32(XFS_DIR2_NULL_DATAPTR))
				stale++;
		}
	} else
		stale = 0;
	/*
	 * Copy the leaf entries from source to destination.
	 */
	memcpy(&dents[start_d], &sents[start_s],
		count * sizeof(xfs_dir2_leaf_entry_t));
	xfs_dir3_leaf_log_ents(args, bp_d, start_d, start_d + count - 1);

	/*
	 * If there are source entries after the ones we copied,
	 * delete the ones we copied by sliding the next ones down.
	 */
	if (start_s + count < shdr->count) {
		memmove(&sents[start_s], &sents[start_s + count],
			count * sizeof(xfs_dir2_leaf_entry_t));
		xfs_dir3_leaf_log_ents(args, bp_s, start_s, start_s + count - 1);
	}

	/*
	 * Update the headers and log them.
	 */
	shdr->count -= count;
	shdr->stale -= stale;
	dhdr->count += count;
	dhdr->stale += stale;
}

/*
 * Determine the sort order of two leaf blocks.
 * Returns 1 if both are valid and leaf2 should be before leaf1, else 0.
 */
int						/* sort order */
xfs_dir2_leafn_order(
	struct xfs_inode	*dp,
	struct xfs_buf		*leaf1_bp,		/* leaf1 buffer */
	struct xfs_buf		*leaf2_bp)		/* leaf2 buffer */
{
	struct xfs_dir2_leaf	*leaf1 = leaf1_bp->b_addr;
	struct xfs_dir2_leaf	*leaf2 = leaf2_bp->b_addr;
	struct xfs_dir2_leaf_entry *ents1;
	struct xfs_dir2_leaf_entry *ents2;
	struct xfs_dir3_icleaf_hdr hdr1;
	struct xfs_dir3_icleaf_hdr hdr2;

	dp->d_ops->leaf_hdr_from_disk(&hdr1, leaf1);
	dp->d_ops->leaf_hdr_from_disk(&hdr2, leaf2);
	ents1 = dp->d_ops->leaf_ents_p(leaf1);
	ents2 = dp->d_ops->leaf_ents_p(leaf2);

	if (hdr1.count > 0 && hdr2.count > 0 &&
	    (be32_to_cpu(ents2[0].hashval) < be32_to_cpu(ents1[0].hashval) ||
	     be32_to_cpu(ents2[hdr2.count - 1].hashval) <
				be32_to_cpu(ents1[hdr1.count - 1].hashval)))
		return 1;
	return 0;
}

/*
 * Rebalance leaf entries between two leaf blocks.
 * This is actually only called when the second block is new,
 * though the code deals with the general case.
 * A new entry will be inserted in one of the blocks, and that
 * entry is taken into account when balancing.
 */
static void
xfs_dir2_leafn_rebalance(
	xfs_da_state_t		*state,		/* btree cursor */
	xfs_da_state_blk_t	*blk1,		/* first btree block */
	xfs_da_state_blk_t	*blk2)		/* second btree block */
{
	xfs_da_args_t		*args;		/* operation arguments */
	int			count;		/* count (& direction) leaves */
	int			isleft;		/* new goes in left leaf */
	xfs_dir2_leaf_t		*leaf1;		/* first leaf structure */
	xfs_dir2_leaf_t		*leaf2;		/* second leaf structure */
	int			mid;		/* midpoint leaf index */
#if defined(DEBUG) || defined(XFS_WARN)
	int			oldstale;	/* old count of stale leaves */
#endif
	int			oldsum;		/* old total leaf count */
	int			swap;		/* swapped leaf blocks */
	struct xfs_dir2_leaf_entry *ents1;
	struct xfs_dir2_leaf_entry *ents2;
	struct xfs_dir3_icleaf_hdr hdr1;
	struct xfs_dir3_icleaf_hdr hdr2;
	struct xfs_inode	*dp = state->args->dp;

	args = state->args;
	/*
	 * If the block order is wrong, swap the arguments.
	 */
	if ((swap = xfs_dir2_leafn_order(dp, blk1->bp, blk2->bp))) {
		xfs_da_state_blk_t	*tmp;	/* temp for block swap */

		tmp = blk1;
		blk1 = blk2;
		blk2 = tmp;
	}
	leaf1 = blk1->bp->b_addr;
	leaf2 = blk2->bp->b_addr;
	dp->d_ops->leaf_hdr_from_disk(&hdr1, leaf1);
	dp->d_ops->leaf_hdr_from_disk(&hdr2, leaf2);
	ents1 = dp->d_ops->leaf_ents_p(leaf1);
	ents2 = dp->d_ops->leaf_ents_p(leaf2);

	oldsum = hdr1.count + hdr2.count;
#if defined(DEBUG) || defined(XFS_WARN)
	oldstale = hdr1.stale + hdr2.stale;
#endif
	mid = oldsum >> 1;

	/*
	 * If the old leaf count was odd then the new one will be even,
	 * so we need to divide the new count evenly.
	 */
	if (oldsum & 1) {
		xfs_dahash_t	midhash;	/* middle entry hash value */

		if (mid >= hdr1.count)
			midhash = be32_to_cpu(ents2[mid - hdr1.count].hashval);
		else
			midhash = be32_to_cpu(ents1[mid].hashval);
		isleft = args->hashval <= midhash;
	}
	/*
	 * If the old count is even then the new count is odd, so there's
	 * no preferred side for the new entry.
	 * Pick the left one.
	 */
	else
		isleft = 1;
	/*
	 * Calculate moved entry count.  Positive means left-to-right,
	 * negative means right-to-left.  Then move the entries.
	 */
	count = hdr1.count - mid + (isleft == 0);
	if (count > 0)
		xfs_dir3_leafn_moveents(args, blk1->bp, &hdr1, ents1,
					hdr1.count - count, blk2->bp,
					&hdr2, ents2, 0, count);
	else if (count < 0)
		xfs_dir3_leafn_moveents(args, blk2->bp, &hdr2, ents2, 0,
					blk1->bp, &hdr1, ents1,
					hdr1.count, count);

	ASSERT(hdr1.count + hdr2.count == oldsum);
	ASSERT(hdr1.stale + hdr2.stale == oldstale);

	/* log the changes made when moving the entries */
	dp->d_ops->leaf_hdr_to_disk(leaf1, &hdr1);
	dp->d_ops->leaf_hdr_to_disk(leaf2, &hdr2);
	xfs_dir3_leaf_log_header(args, blk1->bp);
	xfs_dir3_leaf_log_header(args, blk2->bp);

	xfs_dir3_leaf_check(dp, blk1->bp);
	xfs_dir3_leaf_check(dp, blk2->bp);

	/*
	 * Mark whether we're inserting into the old or new leaf.
	 */
	if (hdr1.count < hdr2.count)
		state->inleaf = swap;
	else if (hdr1.count > hdr2.count)
		state->inleaf = !swap;
	else
		state->inleaf = swap ^ (blk1->index <= hdr1.count);
	/*
	 * Adjust the expected index for insertion.
	 */
	if (!state->inleaf)
		blk2->index = blk1->index - hdr1.count;

	/*
	 * Finally sanity check just to make sure we are not returning a
	 * negative index
	 */
	if (blk2->index < 0) {
		state->inleaf = 1;
		blk2->index = 0;
		xfs_alert(dp->i_mount,
	"%s: picked the wrong leaf? reverting original leaf: blk1->index %d",
			__func__, blk1->index);
	}
}

static int
xfs_dir3_data_block_free(
	xfs_da_args_t		*args,
	struct xfs_dir2_data_hdr *hdr,
	struct xfs_dir2_free	*free,
	xfs_dir2_db_t		fdb,
	int			findex,
	struct xfs_buf		*fbp,
	int			longest)
{
	int			logfree = 0;
	__be16			*bests;
	struct xfs_dir3_icfree_hdr freehdr;
	struct xfs_inode	*dp = args->dp;

	dp->d_ops->free_hdr_from_disk(&freehdr, free);
	bests = dp->d_ops->free_bests_p(free);
	if (hdr) {
		/*
		 * Data block is not empty, just set the free entry to the new
		 * value.
		 */
		bests[findex] = cpu_to_be16(longest);
		xfs_dir2_free_log_bests(args, fbp, findex, findex);
		return 0;
	}

	/* One less used entry in the free table. */
	freehdr.nused--;

	/*
	 * If this was the last entry in the table, we can trim the table size
	 * back.  There might be other entries at the end referring to
	 * non-existent data blocks, get those too.
	 */
	if (findex == freehdr.nvalid - 1) {
		int	i;		/* free entry index */

		for (i = findex - 1; i >= 0; i--) {
			if (bests[i] != cpu_to_be16(NULLDATAOFF))
				break;
		}
		freehdr.nvalid = i + 1;
		logfree = 0;
	} else {
		/* Not the last entry, just punch it out.  */
		bests[findex] = cpu_to_be16(NULLDATAOFF);
		logfree = 1;
	}

	dp->d_ops->free_hdr_to_disk(free, &freehdr);
	xfs_dir2_free_log_header(args, fbp);

	/*
	 * If there are no useful entries left in the block, get rid of the
	 * block if we can.
	 */
	if (!freehdr.nused) {
		int error;

		error = xfs_dir2_shrink_inode(args, fdb, fbp);
		if (error == 0) {
			fbp = NULL;
			logfree = 0;
		} else if (error != -ENOSPC || args->total != 0)
			return error;
		/*
		 * It's possible to get ENOSPC if there is no
		 * space reservation.  In this case some one
		 * else will eventually get rid of this block.
		 */
	}

	/* Log the free entry that changed, unless we got rid of it.  */
	if (logfree)
		xfs_dir2_free_log_bests(args, fbp, findex, findex);
	return 0;
}

/*
 * Remove an entry from a node directory.
 * This removes the leaf entry and the data entry,
 * and updates the free block if necessary.
 */
static int					/* error */
xfs_dir2_leafn_remove(
	xfs_da_args_t		*args,		/* operation arguments */
	struct xfs_buf		*bp,		/* leaf buffer */
	int			index,		/* leaf entry index */
	xfs_da_state_blk_t	*dblk,		/* data block */
	int			*rval)		/* resulting block needs join */
{
	xfs_dir2_data_hdr_t	*hdr;		/* data block header */
	xfs_dir2_db_t		db;		/* data block number */
	struct xfs_buf		*dbp;		/* data block buffer */
	xfs_dir2_data_entry_t	*dep;		/* data block entry */
	xfs_inode_t		*dp;		/* incore directory inode */
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	xfs_dir2_leaf_entry_t	*lep;		/* leaf entry */
	int			longest;	/* longest data free entry */
	int			off;		/* data block entry offset */
	int			needlog;	/* need to log data header */
	int			needscan;	/* need to rescan data frees */
	xfs_trans_t		*tp;		/* transaction pointer */
	struct xfs_dir2_data_free *bf;		/* bestfree table */
	struct xfs_dir3_icleaf_hdr leafhdr;
	struct xfs_dir2_leaf_entry *ents;

	trace_xfs_dir2_leafn_remove(args, index);

	dp = args->dp;
	tp = args->trans;
	leaf = bp->b_addr;
	dp->d_ops->leaf_hdr_from_disk(&leafhdr, leaf);
	ents = dp->d_ops->leaf_ents_p(leaf);

	/*
	 * Point to the entry we're removing.
	 */
	lep = &ents[index];

	/*
	 * Extract the data block and offset from the entry.
	 */
	db = xfs_dir2_dataptr_to_db(args->geo, be32_to_cpu(lep->address));
	ASSERT(dblk->blkno == db);
	off = xfs_dir2_dataptr_to_off(args->geo, be32_to_cpu(lep->address));
	ASSERT(dblk->index == off);

	/*
	 * Kill the leaf entry by marking it stale.
	 * Log the leaf block changes.
	 */
	leafhdr.stale++;
	dp->d_ops->leaf_hdr_to_disk(leaf, &leafhdr);
	xfs_dir3_leaf_log_header(args, bp);

	lep->address = cpu_to_be32(XFS_DIR2_NULL_DATAPTR);
	xfs_dir3_leaf_log_ents(args, bp, index, index);

	/*
	 * Make the data entry free.  Keep track of the longest freespace
	 * in the data block in case it changes.
	 */
	dbp = dblk->bp;
	hdr = dbp->b_addr;
	dep = (xfs_dir2_data_entry_t *)((char *)hdr + off);
	bf = dp->d_ops->data_bestfree_p(hdr);
	longest = be16_to_cpu(bf[0].length);
	needlog = needscan = 0;
	xfs_dir2_data_make_free(args, dbp, off,
		dp->d_ops->data_entsize(dep->namelen), &needlog, &needscan);
	/*
	 * Rescan the data block freespaces for bestfree.
	 * Log the data block header if needed.
	 */
	if (needscan)
		xfs_dir2_data_freescan(dp, hdr, &needlog);
	if (needlog)
		xfs_dir2_data_log_header(args, dbp);
	xfs_dir3_data_check(dp, dbp);
	/*
	 * If the longest data block freespace changes, need to update
	 * the corresponding freeblock entry.
	 */
	if (longest < be16_to_cpu(bf[0].length)) {
		int		error;		/* error return value */
		struct xfs_buf	*fbp;		/* freeblock buffer */
		xfs_dir2_db_t	fdb;		/* freeblock block number */
		int		findex;		/* index in freeblock entries */
		xfs_dir2_free_t	*free;		/* freeblock structure */

		/*
		 * Convert the data block number to a free block,
		 * read in the free block.
		 */
		fdb = dp->d_ops->db_to_fdb(args->geo, db);
		error = xfs_dir2_free_read(tp, dp,
					   xfs_dir2_db_to_da(args->geo, fdb),
					   &fbp);
		if (error)
			return error;
		free = fbp->b_addr;
#ifdef DEBUG
	{
		struct xfs_dir3_icfree_hdr freehdr;
		dp->d_ops->free_hdr_from_disk(&freehdr, free);
		ASSERT(freehdr.firstdb == dp->d_ops->free_max_bests(args->geo) *
			(fdb - xfs_dir2_byte_to_db(args->geo,
						   XFS_DIR2_FREE_OFFSET)));
	}
#endif
		/*
		 * Calculate which entry we need to fix.
		 */
		findex = dp->d_ops->db_to_fdindex(args->geo, db);
		longest = be16_to_cpu(bf[0].length);
		/*
		 * If the data block is now empty we can get rid of it
		 * (usually).
		 */
		if (longest == args->geo->blksize -
			       dp->d_ops->data_entry_offset) {
			/*
			 * Try to punch out the data block.
			 */
			error = xfs_dir2_shrink_inode(args, db, dbp);
			if (error == 0) {
				dblk->bp = NULL;
				hdr = NULL;
			}
			/*
			 * We can get ENOSPC if there's no space reservation.
			 * In this case just drop the buffer and some one else
			 * will eventually get rid of the empty block.
			 */
			else if (!(error == -ENOSPC && args->total == 0))
				return error;
		}
		/*
		 * If we got rid of the data block, we can eliminate that entry
		 * in the free block.
		 */
		error = xfs_dir3_data_block_free(args, hdr, free,
						 fdb, findex, fbp, longest);
		if (error)
			return error;
	}

	xfs_dir3_leaf_check(dp, bp);
	/*
	 * Return indication of whether this leaf block is empty enough
	 * to justify trying to join it with a neighbor.
	 */
	*rval = (dp->d_ops->leaf_hdr_size +
		 (uint)sizeof(ents[0]) * (leafhdr.count - leafhdr.stale)) <
		args->geo->magicpct;
	return 0;
}

/*
 * Split the leaf entries in the old block into old and new blocks.
 */
int						/* error */
xfs_dir2_leafn_split(
	xfs_da_state_t		*state,		/* btree cursor */
	xfs_da_state_blk_t	*oldblk,	/* original block */
	xfs_da_state_blk_t	*newblk)	/* newly created block */
{
	xfs_da_args_t		*args;		/* operation arguments */
	xfs_dablk_t		blkno;		/* new leaf block number */
	int			error;		/* error return value */
	struct xfs_inode	*dp;

	/*
	 * Allocate space for a new leaf node.
	 */
	args = state->args;
	dp = args->dp;
	ASSERT(oldblk->magic == XFS_DIR2_LEAFN_MAGIC);
	error = xfs_da_grow_inode(args, &blkno);
	if (error) {
		return error;
	}
	/*
	 * Initialize the new leaf block.
	 */
	error = xfs_dir3_leaf_get_buf(args, xfs_dir2_da_to_db(args->geo, blkno),
				      &newblk->bp, XFS_DIR2_LEAFN_MAGIC);
	if (error)
		return error;

	newblk->blkno = blkno;
	newblk->magic = XFS_DIR2_LEAFN_MAGIC;
	/*
	 * Rebalance the entries across the two leaves, link the new
	 * block into the leaves.
	 */
	xfs_dir2_leafn_rebalance(state, oldblk, newblk);
	error = xfs_da3_blk_link(state, oldblk, newblk);
	if (error) {
		return error;
	}
	/*
	 * Insert the new entry in the correct block.
	 */
	if (state->inleaf)
		error = xfs_dir2_leafn_add(oldblk->bp, args, oldblk->index);
	else
		error = xfs_dir2_leafn_add(newblk->bp, args, newblk->index);
	/*
	 * Update last hashval in each block since we added the name.
	 */
	oldblk->hashval = xfs_dir2_leaf_lasthash(dp, oldblk->bp, NULL);
	newblk->hashval = xfs_dir2_leaf_lasthash(dp, newblk->bp, NULL);
	xfs_dir3_leaf_check(dp, oldblk->bp);
	xfs_dir3_leaf_check(dp, newblk->bp);
	return error;
}

/*
 * Check a leaf block and its neighbors to see if the block should be
 * collapsed into one or the other neighbor.  Always keep the block
 * with the smaller block number.
 * If the current block is over 50% full, don't try to join it, return 0.
 * If the block is empty, fill in the state structure and return 2.
 * If it can be collapsed, fill in the state structure and return 1.
 * If nothing can be done, return 0.
 */
int						/* error */
xfs_dir2_leafn_toosmall(
	xfs_da_state_t		*state,		/* btree cursor */
	int			*action)	/* resulting action to take */
{
	xfs_da_state_blk_t	*blk;		/* leaf block */
	xfs_dablk_t		blkno;		/* leaf block number */
	struct xfs_buf		*bp;		/* leaf buffer */
	int			bytes;		/* bytes in use */
	int			count;		/* leaf live entry count */
	int			error;		/* error return value */
	int			forward;	/* sibling block direction */
	int			i;		/* sibling counter */
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	int			rval;		/* result from path_shift */
	struct xfs_dir3_icleaf_hdr leafhdr;
	struct xfs_dir2_leaf_entry *ents;
	struct xfs_inode	*dp = state->args->dp;

	/*
	 * Check for the degenerate case of the block being over 50% full.
	 * If so, it's not worth even looking to see if we might be able
	 * to coalesce with a sibling.
	 */
	blk = &state->path.blk[state->path.active - 1];
	leaf = blk->bp->b_addr;
	dp->d_ops->leaf_hdr_from_disk(&leafhdr, leaf);
	ents = dp->d_ops->leaf_ents_p(leaf);
	xfs_dir3_leaf_check(dp, blk->bp);

	count = leafhdr.count - leafhdr.stale;
	bytes = dp->d_ops->leaf_hdr_size + count * sizeof(ents[0]);
	if (bytes > (state->args->geo->blksize >> 1)) {
		/*
		 * Blk over 50%, don't try to join.
		 */
		*action = 0;
		return 0;
	}
	/*
	 * Check for the degenerate case of the block being empty.
	 * If the block is empty, we'll simply delete it, no need to
	 * coalesce it with a sibling block.  We choose (arbitrarily)
	 * to merge with the forward block unless it is NULL.
	 */
	if (count == 0) {
		/*
		 * Make altpath point to the block we want to keep and
		 * path point to the block we want to drop (this one).
		 */
		forward = (leafhdr.forw != 0);
		memcpy(&state->altpath, &state->path, sizeof(state->path));
		error = xfs_da3_path_shift(state, &state->altpath, forward, 0,
			&rval);
		if (error)
			return error;
		*action = rval ? 2 : 0;
		return 0;
	}
	/*
	 * Examine each sibling block to see if we can coalesce with
	 * at least 25% free space to spare.  We need to figure out
	 * whether to merge with the forward or the backward block.
	 * We prefer coalescing with the lower numbered sibling so as
	 * to shrink a directory over time.
	 */
	forward = leafhdr.forw < leafhdr.back;
	for (i = 0, bp = NULL; i < 2; forward = !forward, i++) {
		struct xfs_dir3_icleaf_hdr hdr2;

		blkno = forward ? leafhdr.forw : leafhdr.back;
		if (blkno == 0)
			continue;
		/*
		 * Read the sibling leaf block.
		 */
		error = xfs_dir3_leafn_read(state->args->trans, dp,
					    blkno, -1, &bp);
		if (error)
			return error;

		/*
		 * Count bytes in the two blocks combined.
		 */
		count = leafhdr.count - leafhdr.stale;
		bytes = state->args->geo->blksize -
			(state->args->geo->blksize >> 2);

		leaf = bp->b_addr;
		dp->d_ops->leaf_hdr_from_disk(&hdr2, leaf);
		ents = dp->d_ops->leaf_ents_p(leaf);
		count += hdr2.count - hdr2.stale;
		bytes -= count * sizeof(ents[0]);

		/*
		 * Fits with at least 25% to spare.
		 */
		if (bytes >= 0)
			break;
		xfs_trans_brelse(state->args->trans, bp);
	}
	/*
	 * Didn't like either block, give up.
	 */
	if (i >= 2) {
		*action = 0;
		return 0;
	}

	/*
	 * Make altpath point to the block we want to keep (the lower
	 * numbered block) and path point to the block we want to drop.
	 */
	memcpy(&state->altpath, &state->path, sizeof(state->path));
	if (blkno < blk->blkno)
		error = xfs_da3_path_shift(state, &state->altpath, forward, 0,
			&rval);
	else
		error = xfs_da3_path_shift(state, &state->path, forward, 0,
			&rval);
	if (error) {
		return error;
	}
	*action = rval ? 0 : 1;
	return 0;
}

/*
 * Move all the leaf entries from drop_blk to save_blk.
 * This is done as part of a join operation.
 */
void
xfs_dir2_leafn_unbalance(
	xfs_da_state_t		*state,		/* cursor */
	xfs_da_state_blk_t	*drop_blk,	/* dead block */
	xfs_da_state_blk_t	*save_blk)	/* surviving block */
{
	xfs_da_args_t		*args;		/* operation arguments */
	xfs_dir2_leaf_t		*drop_leaf;	/* dead leaf structure */
	xfs_dir2_leaf_t		*save_leaf;	/* surviving leaf structure */
	struct xfs_dir3_icleaf_hdr savehdr;
	struct xfs_dir3_icleaf_hdr drophdr;
	struct xfs_dir2_leaf_entry *sents;
	struct xfs_dir2_leaf_entry *dents;
	struct xfs_inode	*dp = state->args->dp;

	args = state->args;
	ASSERT(drop_blk->magic == XFS_DIR2_LEAFN_MAGIC);
	ASSERT(save_blk->magic == XFS_DIR2_LEAFN_MAGIC);
	drop_leaf = drop_blk->bp->b_addr;
	save_leaf = save_blk->bp->b_addr;

	dp->d_ops->leaf_hdr_from_disk(&savehdr, save_leaf);
	dp->d_ops->leaf_hdr_from_disk(&drophdr, drop_leaf);
	sents = dp->d_ops->leaf_ents_p(save_leaf);
	dents = dp->d_ops->leaf_ents_p(drop_leaf);

	/*
	 * If there are any stale leaf entries, take this opportunity
	 * to purge them.
	 */
	if (drophdr.stale)
		xfs_dir3_leaf_compact(args, &drophdr, drop_blk->bp);
	if (savehdr.stale)
		xfs_dir3_leaf_compact(args, &savehdr, save_blk->bp);

	/*
	 * Move the entries from drop to the appropriate end of save.
	 */
	drop_blk->hashval = be32_to_cpu(dents[drophdr.count - 1].hashval);
	if (xfs_dir2_leafn_order(dp, save_blk->bp, drop_blk->bp))
		xfs_dir3_leafn_moveents(args, drop_blk->bp, &drophdr, dents, 0,
					save_blk->bp, &savehdr, sents, 0,
					drophdr.count);
	else
		xfs_dir3_leafn_moveents(args, drop_blk->bp, &drophdr, dents, 0,
					save_blk->bp, &savehdr, sents,
					savehdr.count, drophdr.count);
	save_blk->hashval = be32_to_cpu(sents[savehdr.count - 1].hashval);

	/* log the changes made when moving the entries */
	dp->d_ops->leaf_hdr_to_disk(save_leaf, &savehdr);
	dp->d_ops->leaf_hdr_to_disk(drop_leaf, &drophdr);
	xfs_dir3_leaf_log_header(args, save_blk->bp);
	xfs_dir3_leaf_log_header(args, drop_blk->bp);

	xfs_dir3_leaf_check(dp, save_blk->bp);
	xfs_dir3_leaf_check(dp, drop_blk->bp);
}

/*
 * Top-level node form directory addname routine.
 */
int						/* error */
xfs_dir2_node_addname(
	xfs_da_args_t		*args)		/* operation arguments */
{
	xfs_da_state_blk_t	*blk;		/* leaf block for insert */
	int			error;		/* error return value */
	int			rval;		/* sub-return value */
	xfs_da_state_t		*state;		/* btree cursor */

	trace_xfs_dir2_node_addname(args);

	/*
	 * Allocate and initialize the state (btree cursor).
	 */
	state = xfs_da_state_alloc();
	state->args = args;
	state->mp = args->dp->i_mount;
	/*
	 * Look up the name.  We're not supposed to find it, but
	 * this gives us the insertion point.
	 */
	error = xfs_da3_node_lookup_int(state, &rval);
	if (error)
		rval = error;
	if (rval != -ENOENT) {
		goto done;
	}
	/*
	 * Add the data entry to a data block.
	 * Extravalid is set to a freeblock found by lookup.
	 */
	rval = xfs_dir2_node_addname_int(args,
		state->extravalid ? &state->extrablk : NULL);
	if (rval) {
		goto done;
	}
	blk = &state->path.blk[state->path.active - 1];
	ASSERT(blk->magic == XFS_DIR2_LEAFN_MAGIC);
	/*
	 * Add the new leaf entry.
	 */
	rval = xfs_dir2_leafn_add(blk->bp, args, blk->index);
	if (rval == 0) {
		/*
		 * It worked, fix the hash values up the btree.
		 */
		if (!(args->op_flags & XFS_DA_OP_JUSTCHECK))
			xfs_da3_fixhashpath(state, &state->path);
	} else {
		/*
		 * It didn't work, we need to split the leaf block.
		 */
		if (args->total == 0) {
			ASSERT(rval == -ENOSPC);
			goto done;
		}
		/*
		 * Split the leaf block and insert the new entry.
		 */
		rval = xfs_da3_split(state);
	}
done:
	xfs_da_state_free(state);
	return rval;
}

/*
 * Add the data entry for a node-format directory name addition.
 * The leaf entry is added in xfs_dir2_leafn_add.
 * We may enter with a freespace block that the lookup found.
 */
static int					/* error */
xfs_dir2_node_addname_int(
	xfs_da_args_t		*args,		/* operation arguments */
	xfs_da_state_blk_t	*fblk)		/* optional freespace block */
{
	xfs_dir2_data_hdr_t	*hdr;		/* data block header */
	xfs_dir2_db_t		dbno;		/* data block number */
	struct xfs_buf		*dbp;		/* data block buffer */
	xfs_dir2_data_entry_t	*dep;		/* data entry pointer */
	xfs_inode_t		*dp;		/* incore directory inode */
	xfs_dir2_data_unused_t	*dup;		/* data unused entry pointer */
	int			error;		/* error return value */
	xfs_dir2_db_t		fbno;		/* freespace block number */
	struct xfs_buf		*fbp;		/* freespace buffer */
	int			findex;		/* freespace entry index */
	xfs_dir2_free_t		*free=NULL;	/* freespace block structure */
	xfs_dir2_db_t		ifbno;		/* initial freespace block no */
	xfs_dir2_db_t		lastfbno=0;	/* highest freespace block no */
	int			length;		/* length of the new entry */
	int			logfree;	/* need to log free entry */
	xfs_mount_t		*mp;		/* filesystem mount point */
	int			needlog;	/* need to log data header */
	int			needscan;	/* need to rescan data frees */
	__be16			*tagp;		/* data entry tag pointer */
	xfs_trans_t		*tp;		/* transaction pointer */
	__be16			*bests;
	struct xfs_dir3_icfree_hdr freehdr;
	struct xfs_dir2_data_free *bf;
	xfs_dir2_data_aoff_t	aoff;

	dp = args->dp;
	mp = dp->i_mount;
	tp = args->trans;
	length = dp->d_ops->data_entsize(args->namelen);
	/*
	 * If we came in with a freespace block that means that lookup
	 * found an entry with our hash value.  This is the freespace
	 * block for that data entry.
	 */
	if (fblk) {
		fbp = fblk->bp;
		/*
		 * Remember initial freespace block number.
		 */
		ifbno = fblk->blkno;
		free = fbp->b_addr;
		findex = fblk->index;
		bests = dp->d_ops->free_bests_p(free);
		dp->d_ops->free_hdr_from_disk(&freehdr, free);

		/*
		 * This means the free entry showed that the data block had
		 * space for our entry, so we remembered it.
		 * Use that data block.
		 */
		if (findex >= 0) {
			ASSERT(findex < freehdr.nvalid);
			ASSERT(be16_to_cpu(bests[findex]) != NULLDATAOFF);
			ASSERT(be16_to_cpu(bests[findex]) >= length);
			dbno = freehdr.firstdb + findex;
		} else {
			/*
			 * The data block looked at didn't have enough room.
			 * We'll start at the beginning of the freespace entries.
			 */
			dbno = -1;
			findex = 0;
		}
	} else {
		/*
		 * Didn't come in with a freespace block, so no data block.
		 */
		ifbno = dbno = -1;
		fbp = NULL;
		findex = 0;
	}

	/*
	 * If we don't have a data block yet, we're going to scan the
	 * freespace blocks looking for one.  Figure out what the
	 * highest freespace block number is.
	 */
	if (dbno == -1) {
		xfs_fileoff_t	fo;		/* freespace block number */

		if ((error = xfs_bmap_last_offset(dp, &fo, XFS_DATA_FORK)))
			return error;
		lastfbno = xfs_dir2_da_to_db(args->geo, (xfs_dablk_t)fo);
		fbno = ifbno;
	}
	/*
	 * While we haven't identified a data block, search the freeblock
	 * data for a good data block.  If we find a null freeblock entry,
	 * indicating a hole in the data blocks, remember that.
	 */
	while (dbno == -1) {
		/*
		 * If we don't have a freeblock in hand, get the next one.
		 */
		if (fbp == NULL) {
			/*
			 * Happens the first time through unless lookup gave
			 * us a freespace block to start with.
			 */
			if (++fbno == 0)
				fbno = xfs_dir2_byte_to_db(args->geo,
							XFS_DIR2_FREE_OFFSET);
			/*
			 * If it's ifbno we already looked at it.
			 */
			if (fbno == ifbno)
				fbno++;
			/*
			 * If it's off the end we're done.
			 */
			if (fbno >= lastfbno)
				break;
			/*
			 * Read the block.  There can be holes in the
			 * freespace blocks, so this might not succeed.
			 * This should be really rare, so there's no reason
			 * to avoid it.
			 */
			error = xfs_dir2_free_try_read(tp, dp,
					xfs_dir2_db_to_da(args->geo, fbno),
					&fbp);
			if (error)
				return error;
			if (!fbp)
				continue;
			free = fbp->b_addr;
			findex = 0;
		}
		/*
		 * Look at the current free entry.  Is it good enough?
		 *
		 * The bests initialisation should be where the bufer is read in
		 * the above branch. But gcc is too stupid to realise that bests
		 * and the freehdr are actually initialised if they are placed
		 * there, so we have to do it here to avoid warnings. Blech.
		 */
		bests = dp->d_ops->free_bests_p(free);
		dp->d_ops->free_hdr_from_disk(&freehdr, free);
		if (be16_to_cpu(bests[findex]) != NULLDATAOFF &&
		    be16_to_cpu(bests[findex]) >= length)
			dbno = freehdr.firstdb + findex;
		else {
			/*
			 * Are we done with the freeblock?
			 */
			if (++findex == freehdr.nvalid) {
				/*
				 * Drop the block.
				 */
				xfs_trans_brelse(tp, fbp);
				fbp = NULL;
				if (fblk && fblk->bp)
					fblk->bp = NULL;
			}
		}
	}
	/*
	 * If we don't have a data block, we need to allocate one and make
	 * the freespace entries refer to it.
	 */
	if (unlikely(dbno == -1)) {
		/*
		 * Not allowed to allocate, return failure.
		 */
		if ((args->op_flags & XFS_DA_OP_JUSTCHECK) || args->total == 0)
			return -ENOSPC;

		/*
		 * Allocate and initialize the new data block.
		 */
		if (unlikely((error = xfs_dir2_grow_inode(args,
							 XFS_DIR2_DATA_SPACE,
							 &dbno)) ||
		    (error = xfs_dir3_data_init(args, dbno, &dbp))))
			return error;

		/*
		 * If (somehow) we have a freespace block, get rid of it.
		 */
		if (fbp)
			xfs_trans_brelse(tp, fbp);
		if (fblk && fblk->bp)
			fblk->bp = NULL;

		/*
		 * Get the freespace block corresponding to the data block
		 * that was just allocated.
		 */
		fbno = dp->d_ops->db_to_fdb(args->geo, dbno);
		error = xfs_dir2_free_try_read(tp, dp,
				       xfs_dir2_db_to_da(args->geo, fbno),
				       &fbp);
		if (error)
			return error;

		/*
		 * If there wasn't a freespace block, the read will
		 * return a NULL fbp.  Allocate and initialize a new one.
		 */
		if (!fbp) {
			error = xfs_dir2_grow_inode(args, XFS_DIR2_FREE_SPACE,
						    &fbno);
			if (error)
				return error;

			if (dp->d_ops->db_to_fdb(args->geo, dbno) != fbno) {
				xfs_alert(mp,
"%s: dir ino %llu needed freesp block %lld for data block %lld, got %lld ifbno %llu lastfbno %d",
					__func__, (unsigned long long)dp->i_ino,
					(long long)dp->d_ops->db_to_fdb(
								args->geo, dbno),
					(long long)dbno, (long long)fbno,
					(unsigned long long)ifbno, lastfbno);
				if (fblk) {
					xfs_alert(mp,
				" fblk "PTR_FMT" blkno %llu index %d magic 0x%x",
						fblk,
						(unsigned long long)fblk->blkno,
						fblk->index,
						fblk->magic);
				} else {
					xfs_alert(mp, " ... fblk is NULL");
				}
				XFS_ERROR_REPORT("xfs_dir2_node_addname_int",
						 XFS_ERRLEVEL_LOW, mp);
				return -EFSCORRUPTED;
			}

			/*
			 * Get a buffer for the new block.
			 */
			error = xfs_dir3_free_get_buf(args, fbno, &fbp);
			if (error)
				return error;
			free = fbp->b_addr;
			bests = dp->d_ops->free_bests_p(free);
			dp->d_ops->free_hdr_from_disk(&freehdr, free);

			/*
			 * Remember the first slot as our empty slot.
			 */
			freehdr.firstdb =
				(fbno - xfs_dir2_byte_to_db(args->geo,
							XFS_DIR2_FREE_OFFSET)) *
					dp->d_ops->free_max_bests(args->geo);
		} else {
			free = fbp->b_addr;
			bests = dp->d_ops->free_bests_p(free);
			dp->d_ops->free_hdr_from_disk(&freehdr, free);
		}

		/*
		 * Set the freespace block index from the data block number.
		 */
		findex = dp->d_ops->db_to_fdindex(args->geo, dbno);
		/*
		 * If it's after the end of the current entries in the
		 * freespace block, extend that table.
		 */
		if (findex >= freehdr.nvalid) {
			ASSERT(findex < dp->d_ops->free_max_bests(args->geo));
			freehdr.nvalid = findex + 1;
			/*
			 * Tag new entry so nused will go up.
			 */
			bests[findex] = cpu_to_be16(NULLDATAOFF);
		}
		/*
		 * If this entry was for an empty data block
		 * (this should always be true) then update the header.
		 */
		if (bests[findex] == cpu_to_be16(NULLDATAOFF)) {
			freehdr.nused++;
			dp->d_ops->free_hdr_to_disk(fbp->b_addr, &freehdr);
			xfs_dir2_free_log_header(args, fbp);
		}
		/*
		 * Update the real value in the table.
		 * We haven't allocated the data entry yet so this will
		 * change again.
		 */
		hdr = dbp->b_addr;
		bf = dp->d_ops->data_bestfree_p(hdr);
		bests[findex] = bf[0].length;
		logfree = 1;
	}
	/*
	 * We had a data block so we don't have to make a new one.
	 */
	else {
		/*
		 * If just checking, we succeeded.
		 */
		if (args->op_flags & XFS_DA_OP_JUSTCHECK)
			return 0;

		/*
		 * Read the data block in.
		 */
		error = xfs_dir3_data_read(tp, dp,
					   xfs_dir2_db_to_da(args->geo, dbno),
					   -1, &dbp);
		if (error)
			return error;
		hdr = dbp->b_addr;
		bf = dp->d_ops->data_bestfree_p(hdr);
		logfree = 0;
	}
	ASSERT(be16_to_cpu(bf[0].length) >= length);
	/*
	 * Point to the existing unused space.
	 */
	dup = (xfs_dir2_data_unused_t *)
	      ((char *)hdr + be16_to_cpu(bf[0].offset));
	needscan = needlog = 0;
	/*
	 * Mark the first part of the unused space, inuse for us.
	 */
	aoff = (xfs_dir2_data_aoff_t)((char *)dup - (char *)hdr);
	error = xfs_dir2_data_use_free(args, dbp, dup, aoff, length,
			&needlog, &needscan);
	if (error) {
		xfs_trans_brelse(tp, dbp);
		return error;
	}
	/*
	 * Fill in the new entry and log it.
	 */
	dep = (xfs_dir2_data_entry_t *)dup;
	dep->inumber = cpu_to_be64(args->inumber);
	dep->namelen = args->namelen;
	memcpy(dep->name, args->name, dep->namelen);
	dp->d_ops->data_put_ftype(dep, args->filetype);
	tagp = dp->d_ops->data_entry_tag_p(dep);
	*tagp = cpu_to_be16((char *)dep - (char *)hdr);
	xfs_dir2_data_log_entry(args, dbp, dep);
	/*
	 * Rescan the block for bestfree if needed.
	 */
	if (needscan)
		xfs_dir2_data_freescan(dp, hdr, &needlog);
	/*
	 * Log the data block header if needed.
	 */
	if (needlog)
		xfs_dir2_data_log_header(args, dbp);
	/*
	 * If the freespace entry is now wrong, update it.
	 */
	bests = dp->d_ops->free_bests_p(free); /* gcc is so stupid */
	if (be16_to_cpu(bests[findex]) != be16_to_cpu(bf[0].length)) {
		bests[findex] = bf[0].length;
		logfree = 1;
	}
	/*
	 * Log the freespace entry if needed.
	 */
	if (logfree)
		xfs_dir2_free_log_bests(args, fbp, findex, findex);
	/*
	 * Return the data block and offset in args, then drop the data block.
	 */
	args->blkno = (xfs_dablk_t)dbno;
	args->index = be16_to_cpu(*tagp);
	return 0;
}

/*
 * Lookup an entry in a node-format directory.
 * All the real work happens in xfs_da3_node_lookup_int.
 * The only real output is the inode number of the entry.
 */
int						/* error */
xfs_dir2_node_lookup(
	xfs_da_args_t	*args)			/* operation arguments */
{
	int		error;			/* error return value */
	int		i;			/* btree level */
	int		rval;			/* operation return value */
	xfs_da_state_t	*state;			/* btree cursor */

	trace_xfs_dir2_node_lookup(args);

	/*
	 * Allocate and initialize the btree cursor.
	 */
	state = xfs_da_state_alloc();
	state->args = args;
	state->mp = args->dp->i_mount;
	/*
	 * Fill in the path to the entry in the cursor.
	 */
	error = xfs_da3_node_lookup_int(state, &rval);
	if (error)
		rval = error;
	else if (rval == -ENOENT && args->cmpresult == XFS_CMP_CASE) {
		/* If a CI match, dup the actual name and return -EEXIST */
		xfs_dir2_data_entry_t	*dep;

		dep = (xfs_dir2_data_entry_t *)
			((char *)state->extrablk.bp->b_addr +
						 state->extrablk.index);
		rval = xfs_dir_cilookup_result(args, dep->name, dep->namelen);
	}
	/*
	 * Release the btree blocks and leaf block.
	 */
	for (i = 0; i < state->path.active; i++) {
		xfs_trans_brelse(args->trans, state->path.blk[i].bp);
		state->path.blk[i].bp = NULL;
	}
	/*
	 * Release the data block if we have it.
	 */
	if (state->extravalid && state->extrablk.bp) {
		xfs_trans_brelse(args->trans, state->extrablk.bp);
		state->extrablk.bp = NULL;
	}
	xfs_da_state_free(state);
	return rval;
}

/*
 * Remove an entry from a node-format directory.
 */
int						/* error */
xfs_dir2_node_removename(
	struct xfs_da_args	*args)		/* operation arguments */
{
	struct xfs_da_state_blk	*blk;		/* leaf block */
	int			error;		/* error return value */
	int			rval;		/* operation return value */
	struct xfs_da_state	*state;		/* btree cursor */

	trace_xfs_dir2_node_removename(args);

	/*
	 * Allocate and initialize the btree cursor.
	 */
	state = xfs_da_state_alloc();
	state->args = args;
	state->mp = args->dp->i_mount;

	/* Look up the entry we're deleting, set up the cursor. */
	error = xfs_da3_node_lookup_int(state, &rval);
	if (error)
		goto out_free;

	/* Didn't find it, upper layer screwed up. */
	if (rval != -EEXIST) {
		error = rval;
		goto out_free;
	}

	blk = &state->path.blk[state->path.active - 1];
	ASSERT(blk->magic == XFS_DIR2_LEAFN_MAGIC);
	ASSERT(state->extravalid);
	/*
	 * Remove the leaf and data entries.
	 * Extrablk refers to the data block.
	 */
	error = xfs_dir2_leafn_remove(args, blk->bp, blk->index,
		&state->extrablk, &rval);
	if (error)
		goto out_free;
	/*
	 * Fix the hash values up the btree.
	 */
	xfs_da3_fixhashpath(state, &state->path);
	/*
	 * If we need to join leaf blocks, do it.
	 */
	if (rval && state->path.active > 1)
		error = xfs_da3_join(state);
	/*
	 * If no errors so far, try conversion to leaf format.
	 */
	if (!error)
		error = xfs_dir2_node_to_leaf(state);
out_free:
	xfs_da_state_free(state);
	return error;
}

/*
 * Replace an entry's inode number in a node-format directory.
 */
int						/* error */
xfs_dir2_node_replace(
	xfs_da_args_t		*args)		/* operation arguments */
{
	xfs_da_state_blk_t	*blk;		/* leaf block */
	xfs_dir2_data_hdr_t	*hdr;		/* data block header */
	xfs_dir2_data_entry_t	*dep;		/* data entry changed */
	int			error;		/* error return value */
	int			i;		/* btree level */
	xfs_ino_t		inum;		/* new inode number */
	int			ftype;		/* new file type */
	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
	xfs_dir2_leaf_entry_t	*lep;		/* leaf entry being changed */
	int			rval;		/* internal return value */
	xfs_da_state_t		*state;		/* btree cursor */

	trace_xfs_dir2_node_replace(args);

	/*
	 * Allocate and initialize the btree cursor.
	 */
	state = xfs_da_state_alloc();
	state->args = args;
	state->mp = args->dp->i_mount;

	/*
	 * We have to save new inode number and ftype since
	 * xfs_da3_node_lookup_int() is going to overwrite them
	 */
	inum = args->inumber;
	ftype = args->filetype;

	/*
	 * Lookup the entry to change in the btree.
	 */
	error = xfs_da3_node_lookup_int(state, &rval);
	if (error) {
		rval = error;
	}
	/*
	 * It should be found, since the vnodeops layer has looked it up
	 * and locked it.  But paranoia is good.
	 */
	if (rval == -EEXIST) {
		struct xfs_dir2_leaf_entry *ents;
		/*
		 * Find the leaf entry.
		 */
		blk = &state->path.blk[state->path.active - 1];
		ASSERT(blk->magic == XFS_DIR2_LEAFN_MAGIC);
		leaf = blk->bp->b_addr;
		ents = args->dp->d_ops->leaf_ents_p(leaf);
		lep = &ents[blk->index];
		ASSERT(state->extravalid);
		/*
		 * Point to the data entry.
		 */
		hdr = state->extrablk.bp->b_addr;
		ASSERT(hdr->magic == cpu_to_be32(XFS_DIR2_DATA_MAGIC) ||
		       hdr->magic == cpu_to_be32(XFS_DIR3_DATA_MAGIC));
		dep = (xfs_dir2_data_entry_t *)
		      ((char *)hdr +
		       xfs_dir2_dataptr_to_off(args->geo,
					       be32_to_cpu(lep->address)));
		ASSERT(inum != be64_to_cpu(dep->inumber));
		/*
		 * Fill in the new inode number and log the entry.
		 */
		dep->inumber = cpu_to_be64(inum);
		args->dp->d_ops->data_put_ftype(dep, ftype);
		xfs_dir2_data_log_entry(args, state->extrablk.bp, dep);
		rval = 0;
	}
	/*
	 * Didn't find it, and we're holding a data block.  Drop it.
	 */
	else if (state->extravalid) {
		xfs_trans_brelse(args->trans, state->extrablk.bp);
		state->extrablk.bp = NULL;
	}
	/*
	 * Release all the buffers in the cursor.
	 */
	for (i = 0; i < state->path.active; i++) {
		xfs_trans_brelse(args->trans, state->path.blk[i].bp);
		state->path.blk[i].bp = NULL;
	}
	xfs_da_state_free(state);
	return rval;
}

/*
 * Trim off a trailing empty freespace block.
 * Return (in rvalp) 1 if we did it, 0 if not.
 */
int						/* error */
xfs_dir2_node_trim_free(
	xfs_da_args_t		*args,		/* operation arguments */
	xfs_fileoff_t		fo,		/* free block number */
	int			*rvalp)		/* out: did something */
{
	struct xfs_buf		*bp;		/* freespace buffer */
	xfs_inode_t		*dp;		/* incore directory inode */
	int			error;		/* error return code */
	xfs_dir2_free_t		*free;		/* freespace structure */
	xfs_trans_t		*tp;		/* transaction pointer */
	struct xfs_dir3_icfree_hdr freehdr;

	dp = args->dp;
	tp = args->trans;

	*rvalp = 0;

	/*
	 * Read the freespace block.
	 */
	error = xfs_dir2_free_try_read(tp, dp, fo, &bp);
	if (error)
		return error;
	/*
	 * There can be holes in freespace.  If fo is a hole, there's
	 * nothing to do.
	 */
	if (!bp)
		return 0;
	free = bp->b_addr;
	dp->d_ops->free_hdr_from_disk(&freehdr, free);

	/*
	 * If there are used entries, there's nothing to do.
	 */
	if (freehdr.nused > 0) {
		xfs_trans_brelse(tp, bp);
		return 0;
	}
	/*
	 * Blow the block away.
	 */
	error = xfs_dir2_shrink_inode(args,
			xfs_dir2_da_to_db(args->geo, (xfs_dablk_t)fo), bp);
	if (error) {
		/*
		 * Can't fail with ENOSPC since that only happens with no
		 * space reservation, when breaking up an extent into two
		 * pieces.  This is the last block of an extent.
		 */
		ASSERT(error != -ENOSPC);
		xfs_trans_brelse(tp, bp);
		return error;
	}
	/*
	 * Return that we succeeded.
	 */
	*rvalp = 1;
	return 0;
}