summaryrefslogtreecommitdiffstats
path: root/fs/reiserfs/stree.c
blob: 00fd879c4a2a4f90c993d901f373f8ceef64c829 (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
/*
 *  Copyright 2000 by Hans Reiser, licensing governed by reiserfs/README
 */

/*
 *  Written by Anatoly P. Pinchuk pap@namesys.botik.ru
 *  Programm System Institute
 *  Pereslavl-Zalessky Russia
 */

/*
 *  This file contains functions dealing with S+tree
 *
 * B_IS_IN_TREE
 * copy_item_head
 * comp_short_keys
 * comp_keys
 * comp_short_le_keys
 * le_key2cpu_key
 * comp_le_keys
 * bin_search
 * get_lkey
 * get_rkey
 * key_in_buffer
 * decrement_bcount
 * reiserfs_check_path
 * pathrelse_and_restore
 * pathrelse
 * search_by_key_reada
 * search_by_key
 * search_for_position_by_key
 * comp_items
 * prepare_for_direct_item
 * prepare_for_direntry_item
 * prepare_for_delete_or_cut
 * calc_deleted_bytes_number
 * init_tb_struct
 * padd_item
 * reiserfs_delete_item
 * reiserfs_delete_solid_item
 * reiserfs_delete_object
 * maybe_indirect_to_direct
 * indirect_to_direct_roll_back
 * reiserfs_cut_from_item
 * truncate_directory
 * reiserfs_do_truncate
 * reiserfs_paste_into_item
 * reiserfs_insert_item
 */

#include <linux/time.h>
#include <linux/string.h>
#include <linux/pagemap.h>
#include <linux/reiserfs_fs.h>
#include <linux/buffer_head.h>
#include <linux/quotaops.h>

/* Does the buffer contain a disk block which is in the tree. */
inline int B_IS_IN_TREE(const struct buffer_head *p_s_bh)
{

	RFALSE(B_LEVEL(p_s_bh) > MAX_HEIGHT,
	       "PAP-1010: block (%b) has too big level (%z)", p_s_bh, p_s_bh);

	return (B_LEVEL(p_s_bh) != FREE_LEVEL);
}

//
// to gets item head in le form
//
inline void copy_item_head(struct item_head *p_v_to,
			   const struct item_head *p_v_from)
{
	memcpy(p_v_to, p_v_from, IH_SIZE);
}

/* k1 is pointer to on-disk structure which is stored in little-endian
   form. k2 is pointer to cpu variable. For key of items of the same
   object this returns 0.
   Returns: -1 if key1 < key2
   0 if key1 == key2
   1 if key1 > key2 */
inline int comp_short_keys(const struct reiserfs_key *le_key,
			   const struct cpu_key *cpu_key)
{
	__u32 n;
	n = le32_to_cpu(le_key->k_dir_id);
	if (n < cpu_key->on_disk_key.k_dir_id)
		return -1;
	if (n > cpu_key->on_disk_key.k_dir_id)
		return 1;
	n = le32_to_cpu(le_key->k_objectid);
	if (n < cpu_key->on_disk_key.k_objectid)
		return -1;
	if (n > cpu_key->on_disk_key.k_objectid)
		return 1;
	return 0;
}

/* k1 is pointer to on-disk structure which is stored in little-endian
   form. k2 is pointer to cpu variable.
   Compare keys using all 4 key fields.
   Returns: -1 if key1 < key2 0
   if key1 = key2 1 if key1 > key2 */
static inline int comp_keys(const struct reiserfs_key *le_key,
			    const struct cpu_key *cpu_key)
{
	int retval;

	retval = comp_short_keys(le_key, cpu_key);
	if (retval)
		return retval;
	if (le_key_k_offset(le_key_version(le_key), le_key) <
	    cpu_key_k_offset(cpu_key))
		return -1;
	if (le_key_k_offset(le_key_version(le_key), le_key) >
	    cpu_key_k_offset(cpu_key))
		return 1;

	if (cpu_key->key_length == 3)
		return 0;

	/* this part is needed only when tail conversion is in progress */
	if (le_key_k_type(le_key_version(le_key), le_key) <
	    cpu_key_k_type(cpu_key))
		return -1;

	if (le_key_k_type(le_key_version(le_key), le_key) >
	    cpu_key_k_type(cpu_key))
		return 1;

	return 0;
}

inline int comp_short_le_keys(const struct reiserfs_key *key1,
			      const struct reiserfs_key *key2)
{
	__u32 *p_s_1_u32, *p_s_2_u32;
	int n_key_length = REISERFS_SHORT_KEY_LEN;

	p_s_1_u32 = (__u32 *) key1;
	p_s_2_u32 = (__u32 *) key2;
	for (; n_key_length--; ++p_s_1_u32, ++p_s_2_u32) {
		if (le32_to_cpu(*p_s_1_u32) < le32_to_cpu(*p_s_2_u32))
			return -1;
		if (le32_to_cpu(*p_s_1_u32) > le32_to_cpu(*p_s_2_u32))
			return 1;
	}
	return 0;
}

inline void le_key2cpu_key(struct cpu_key *to, const struct reiserfs_key *from)
{
	int version;
	to->on_disk_key.k_dir_id = le32_to_cpu(from->k_dir_id);
	to->on_disk_key.k_objectid = le32_to_cpu(from->k_objectid);

	// find out version of the key
	version = le_key_version(from);
	to->version = version;
	to->on_disk_key.k_offset = le_key_k_offset(version, from);
	to->on_disk_key.k_type = le_key_k_type(version, from);
}

// this does not say which one is bigger, it only returns 1 if keys
// are not equal, 0 otherwise
inline int comp_le_keys(const struct reiserfs_key *k1,
			const struct reiserfs_key *k2)
{
	return memcmp(k1, k2, sizeof(struct reiserfs_key));
}

/**************************************************************************
 *  Binary search toolkit function                                        *
 *  Search for an item in the array by the item key                       *
 *  Returns:    1 if found,  0 if not found;                              *
 *        *p_n_pos = number of the searched element if found, else the    *
 *        number of the first element that is larger than p_v_key.        *
 **************************************************************************/
/* For those not familiar with binary search: n_lbound is the leftmost item that it
 could be, n_rbound the rightmost item that it could be.  We examine the item
 halfway between n_lbound and n_rbound, and that tells us either that we can increase
 n_lbound, or decrease n_rbound, or that we have found it, or if n_lbound <= n_rbound that
 there are no possible items, and we have not found it. With each examination we
 cut the number of possible items it could be by one more than half rounded down,
 or we find it. */
static inline int bin_search(const void *p_v_key,	/* Key to search for.                   */
			     const void *p_v_base,	/* First item in the array.             */
			     int p_n_num,	/* Number of items in the array.        */
			     int p_n_width,	/* Item size in the array.
						   searched. Lest the reader be
						   confused, note that this is crafted
						   as a general function, and when it
						   is applied specifically to the array
						   of item headers in a node, p_n_width
						   is actually the item header size not
						   the item size.                      */
			     int *p_n_pos	/* Number of the searched for element. */
    )
{
	int n_rbound, n_lbound, n_j;

	for (n_j = ((n_rbound = p_n_num - 1) + (n_lbound = 0)) / 2;
	     n_lbound <= n_rbound; n_j = (n_rbound + n_lbound) / 2)
		switch (comp_keys
			((struct reiserfs_key *)((char *)p_v_base +
						 n_j * p_n_width),
			 (struct cpu_key *)p_v_key)) {
		case -1:
			n_lbound = n_j + 1;
			continue;
		case 1:
			n_rbound = n_j - 1;
			continue;
		case 0:
			*p_n_pos = n_j;
			return ITEM_FOUND;	/* Key found in the array.  */
		}

	/* bin_search did not find given key, it returns position of key,
	   that is minimal and greater than the given one. */
	*p_n_pos = n_lbound;
	return ITEM_NOT_FOUND;
}

#ifdef CONFIG_REISERFS_CHECK
extern struct tree_balance *cur_tb;
#endif

/* Minimal possible key. It is never in the tree. */
const struct reiserfs_key MIN_KEY = { 0, 0, {{0, 0},} };

/* Maximal possible key. It is never in the tree. */
static const struct reiserfs_key MAX_KEY = {
	__constant_cpu_to_le32(0xffffffff),
	__constant_cpu_to_le32(0xffffffff),
	{{__constant_cpu_to_le32(0xffffffff),
	  __constant_cpu_to_le32(0xffffffff)},}
};

/* Get delimiting key of the buffer by looking for it in the buffers in the path, starting from the bottom
   of the path, and going upwards.  We must check the path's validity at each step.  If the key is not in
   the path, there is no delimiting key in the tree (buffer is first or last buffer in tree), and in this
   case we return a special key, either MIN_KEY or MAX_KEY. */
static inline const struct reiserfs_key *get_lkey(const struct treepath
						  *p_s_chk_path,
						  const struct super_block
						  *sb)
{
	int n_position, n_path_offset = p_s_chk_path->path_length;
	struct buffer_head *p_s_parent;

	RFALSE(n_path_offset < FIRST_PATH_ELEMENT_OFFSET,
	       "PAP-5010: invalid offset in the path");

	/* While not higher in path than first element. */
	while (n_path_offset-- > FIRST_PATH_ELEMENT_OFFSET) {

		RFALSE(!buffer_uptodate
		       (PATH_OFFSET_PBUFFER(p_s_chk_path, n_path_offset)),
		       "PAP-5020: parent is not uptodate");

		/* Parent at the path is not in the tree now. */
		if (!B_IS_IN_TREE
		    (p_s_parent =
		     PATH_OFFSET_PBUFFER(p_s_chk_path, n_path_offset)))
			return &MAX_KEY;
		/* Check whether position in the parent is correct. */
		if ((n_position =
		     PATH_OFFSET_POSITION(p_s_chk_path,
					  n_path_offset)) >
		    B_NR_ITEMS(p_s_parent))
			return &MAX_KEY;
		/* Check whether parent at the path really points to the child. */
		if (B_N_CHILD_NUM(p_s_parent, n_position) !=
		    PATH_OFFSET_PBUFFER(p_s_chk_path,
					n_path_offset + 1)->b_blocknr)
			return &MAX_KEY;
		/* Return delimiting key if position in the parent is not equal to zero. */
		if (n_position)
			return B_N_PDELIM_KEY(p_s_parent, n_position - 1);
	}
	/* Return MIN_KEY if we are in the root of the buffer tree. */
	if (PATH_OFFSET_PBUFFER(p_s_chk_path, FIRST_PATH_ELEMENT_OFFSET)->
	    b_blocknr == SB_ROOT_BLOCK(sb))
		return &MIN_KEY;
	return &MAX_KEY;
}

/* Get delimiting key of the buffer at the path and its right neighbor. */
inline const struct reiserfs_key *get_rkey(const struct treepath *p_s_chk_path,
					   const struct super_block *sb)
{
	int n_position, n_path_offset = p_s_chk_path->path_length;
	struct buffer_head *p_s_parent;

	RFALSE(n_path_offset < FIRST_PATH_ELEMENT_OFFSET,
	       "PAP-5030: invalid offset in the path");

	while (n_path_offset-- > FIRST_PATH_ELEMENT_OFFSET) {

		RFALSE(!buffer_uptodate
		       (PATH_OFFSET_PBUFFER(p_s_chk_path, n_path_offset)),
		       "PAP-5040: parent is not uptodate");

		/* Parent at the path is not in the tree now. */
		if (!B_IS_IN_TREE
		    (p_s_parent =
		     PATH_OFFSET_PBUFFER(p_s_chk_path, n_path_offset)))
			return &MIN_KEY;
		/* Check whether position in the parent is correct. */
		if ((n_position =
		     PATH_OFFSET_POSITION(p_s_chk_path,
					  n_path_offset)) >
		    B_NR_ITEMS(p_s_parent))
			return &MIN_KEY;
		/* Check whether parent at the path really points to the child. */
		if (B_N_CHILD_NUM(p_s_parent, n_position) !=
		    PATH_OFFSET_PBUFFER(p_s_chk_path,
					n_path_offset + 1)->b_blocknr)
			return &MIN_KEY;
		/* Return delimiting key if position in the parent is not the last one. */
		if (n_position != B_NR_ITEMS(p_s_parent))
			return B_N_PDELIM_KEY(p_s_parent, n_position);
	}
	/* Return MAX_KEY if we are in the root of the buffer tree. */
	if (PATH_OFFSET_PBUFFER(p_s_chk_path, FIRST_PATH_ELEMENT_OFFSET)->
	    b_blocknr == SB_ROOT_BLOCK(sb))
		return &MAX_KEY;
	return &MIN_KEY;
}

/* Check whether a key is contained in the tree rooted from a buffer at a path. */
/* This works by looking at the left and right delimiting keys for the buffer in the last path_element in
   the path.  These delimiting keys are stored at least one level above that buffer in the tree. If the
   buffer is the first or last node in the tree order then one of the delimiting keys may be absent, and in
   this case get_lkey and get_rkey return a special key which is MIN_KEY or MAX_KEY. */
static inline int key_in_buffer(struct treepath *p_s_chk_path,	/* Path which should be checked.  */
				const struct cpu_key *p_s_key,	/* Key which should be checked.   */
				struct super_block *sb	/* Super block pointer.	   */
    )
{

	RFALSE(!p_s_key || p_s_chk_path->path_length < FIRST_PATH_ELEMENT_OFFSET
	       || p_s_chk_path->path_length > MAX_HEIGHT,
	       "PAP-5050: pointer to the key(%p) is NULL or invalid path length(%d)",
	       p_s_key, p_s_chk_path->path_length);
	RFALSE(!PATH_PLAST_BUFFER(p_s_chk_path)->b_bdev,
	       "PAP-5060: device must not be NODEV");

	if (comp_keys(get_lkey(p_s_chk_path, sb), p_s_key) == 1)
		/* left delimiting key is bigger, that the key we look for */
		return 0;
	//  if ( comp_keys(p_s_key, get_rkey(p_s_chk_path, sb)) != -1 )
	if (comp_keys(get_rkey(p_s_chk_path, sb), p_s_key) != 1)
		/* p_s_key must be less than right delimitiing key */
		return 0;
	return 1;
}

int reiserfs_check_path(struct treepath *p)
{
	RFALSE(p->path_length != ILLEGAL_PATH_ELEMENT_OFFSET,
	       "path not properly relsed");
	return 0;
}

/* Drop the reference to each buffer in a path and restore
 * dirty bits clean when preparing the buffer for the log.
 * This version should only be called from fix_nodes() */
void pathrelse_and_restore(struct super_block *sb,
			   struct treepath *p_s_search_path)
{
	int n_path_offset = p_s_search_path->path_length;

	RFALSE(n_path_offset < ILLEGAL_PATH_ELEMENT_OFFSET,
	       "clm-4000: invalid path offset");

	while (n_path_offset > ILLEGAL_PATH_ELEMENT_OFFSET) {
		struct buffer_head *bh;
		bh = PATH_OFFSET_PBUFFER(p_s_search_path, n_path_offset--);
		reiserfs_restore_prepared_buffer(sb, bh);
		brelse(bh);
	}
	p_s_search_path->path_length = ILLEGAL_PATH_ELEMENT_OFFSET;
}

/* Drop the reference to each buffer in a path */
void pathrelse(struct treepath *p_s_search_path)
{
	int n_path_offset = p_s_search_path->path_length;

	RFALSE(n_path_offset < ILLEGAL_PATH_ELEMENT_OFFSET,
	       "PAP-5090: invalid path offset");

	while (n_path_offset > ILLEGAL_PATH_ELEMENT_OFFSET)
		brelse(PATH_OFFSET_PBUFFER(p_s_search_path, n_path_offset--));

	p_s_search_path->path_length = ILLEGAL_PATH_ELEMENT_OFFSET;
}

static int is_leaf(char *buf, int blocksize, struct buffer_head *bh)
{
	struct block_head *blkh;
	struct item_head *ih;
	int used_space;
	int prev_location;
	int i;
	int nr;

	blkh = (struct block_head *)buf;
	if (blkh_level(blkh) != DISK_LEAF_NODE_LEVEL) {
		reiserfs_warning(NULL, "reiserfs-5080",
				 "this should be caught earlier");
		return 0;
	}

	nr = blkh_nr_item(blkh);
	if (nr < 1 || nr > ((blocksize - BLKH_SIZE) / (IH_SIZE + MIN_ITEM_LEN))) {
		/* item number is too big or too small */
		reiserfs_warning(NULL, "reiserfs-5081",
				 "nr_item seems wrong: %z", bh);
		return 0;
	}
	ih = (struct item_head *)(buf + BLKH_SIZE) + nr - 1;
	used_space = BLKH_SIZE + IH_SIZE * nr + (blocksize - ih_location(ih));
	if (used_space != blocksize - blkh_free_space(blkh)) {
		/* free space does not match to calculated amount of use space */
		reiserfs_warning(NULL, "reiserfs-5082",
				 "free space seems wrong: %z", bh);
		return 0;
	}
	// FIXME: it is_leaf will hit performance too much - we may have
	// return 1 here

	/* check tables of item heads */
	ih = (struct item_head *)(buf + BLKH_SIZE);
	prev_location = blocksize;
	for (i = 0; i < nr; i++, ih++) {
		if (le_ih_k_type(ih) == TYPE_ANY) {
			reiserfs_warning(NULL, "reiserfs-5083",
					 "wrong item type for item %h",
					 ih);
			return 0;
		}
		if (ih_location(ih) >= blocksize
		    || ih_location(ih) < IH_SIZE * nr) {
			reiserfs_warning(NULL, "reiserfs-5084",
					 "item location seems wrong: %h",
					 ih);
			return 0;
		}
		if (ih_item_len(ih) < 1
		    || ih_item_len(ih) > MAX_ITEM_LEN(blocksize)) {
			reiserfs_warning(NULL, "reiserfs-5085",
					 "item length seems wrong: %h",
					 ih);
			return 0;
		}
		if (prev_location - ih_location(ih) != ih_item_len(ih)) {
			reiserfs_warning(NULL, "reiserfs-5086",
					 "item location seems wrong "
					 "(second one): %h", ih);
			return 0;
		}
		prev_location = ih_location(ih);
	}

	// one may imagine much more checks
	return 1;
}

/* returns 1 if buf looks like an internal node, 0 otherwise */
static int is_internal(char *buf, int blocksize, struct buffer_head *bh)
{
	struct block_head *blkh;
	int nr;
	int used_space;

	blkh = (struct block_head *)buf;
	nr = blkh_level(blkh);
	if (nr <= DISK_LEAF_NODE_LEVEL || nr > MAX_HEIGHT) {
		/* this level is not possible for internal nodes */
		reiserfs_warning(NULL, "reiserfs-5087",
				 "this should be caught earlier");
		return 0;
	}

	nr = blkh_nr_item(blkh);
	if (nr > (blocksize - BLKH_SIZE - DC_SIZE) / (KEY_SIZE + DC_SIZE)) {
		/* for internal which is not root we might check min number of keys */
		reiserfs_warning(NULL, "reiserfs-5088",
				 "number of key seems wrong: %z", bh);
		return 0;
	}

	used_space = BLKH_SIZE + KEY_SIZE * nr + DC_SIZE * (nr + 1);
	if (used_space != blocksize - blkh_free_space(blkh)) {
		reiserfs_warning(NULL, "reiserfs-5089",
				 "free space seems wrong: %z", bh);
		return 0;
	}
	// one may imagine much more checks
	return 1;
}

// make sure that bh contains formatted node of reiserfs tree of
// 'level'-th level
static int is_tree_node(struct buffer_head *bh, int level)
{
	if (B_LEVEL(bh) != level) {
		reiserfs_warning(NULL, "reiserfs-5090", "node level %d does "
				 "not match to the expected one %d",
				 B_LEVEL(bh), level);
		return 0;
	}
	if (level == DISK_LEAF_NODE_LEVEL)
		return is_leaf(bh->b_data, bh->b_size, bh);

	return is_internal(bh->b_data, bh->b_size, bh);
}

#define SEARCH_BY_KEY_READA 16

/* The function is NOT SCHEDULE-SAFE! */
static void search_by_key_reada(struct super_block *s,
				struct buffer_head **bh,
				b_blocknr_t *b, int num)
{
	int i, j;

	for (i = 0; i < num; i++) {
		bh[i] = sb_getblk(s, b[i]);
	}
	for (j = 0; j < i; j++) {
		/*
		 * note, this needs attention if we are getting rid of the BKL
		 * you have to make sure the prepared bit isn't set on this buffer
		 */
		if (!buffer_uptodate(bh[j]))
			ll_rw_block(READA, 1, bh + j);
		brelse(bh[j]);
	}
}

/**************************************************************************
 * Algorithm   SearchByKey                                                *
 *             look for item in the Disk S+Tree by its key                *
 * Input:  sb   -  super block                                            *
 *         p_s_key  - pointer to the key to search                        *
 * Output: ITEM_FOUND, ITEM_NOT_FOUND or IO_ERROR                         *
 *         p_s_search_path - path from the root to the needed leaf        *
 **************************************************************************/

/* This function fills up the path from the root to the leaf as it
   descends the tree looking for the key.  It uses reiserfs_bread to
   try to find buffers in the cache given their block number.  If it
   does not find them in the cache it reads them from disk.  For each
   node search_by_key finds using reiserfs_bread it then uses
   bin_search to look through that node.  bin_search will find the
   position of the block_number of the next node if it is looking
   through an internal node.  If it is looking through a leaf node
   bin_search will find the position of the item which has key either
   equal to given key, or which is the maximal key less than the given
   key.  search_by_key returns a path that must be checked for the
   correctness of the top of the path but need not be checked for the
   correctness of the bottom of the path */
/* The function is NOT SCHEDULE-SAFE! */
int search_by_key(struct super_block *sb, const struct cpu_key *p_s_key,	/* Key to search. */
		  struct treepath *p_s_search_path,/* This structure was
						   allocated and initialized
						   by the calling
						   function. It is filled up
						   by this function.  */
		  int n_stop_level	/* How far down the tree to search. To
					   stop at leaf level - set to
					   DISK_LEAF_NODE_LEVEL */
    )
{
	b_blocknr_t n_block_number;
	int expected_level;
	struct buffer_head *p_s_bh;
	struct path_element *p_s_last_element;
	int n_node_level, n_retval;
	int right_neighbor_of_leaf_node;
	int fs_gen;
	struct buffer_head *reada_bh[SEARCH_BY_KEY_READA];
	b_blocknr_t reada_blocks[SEARCH_BY_KEY_READA];
	int reada_count = 0;

#ifdef CONFIG_REISERFS_CHECK
	int n_repeat_counter = 0;
#endif

	PROC_INFO_INC(sb, search_by_key);

	/* As we add each node to a path we increase its count.  This means that
	   we must be careful to release all nodes in a path before we either
	   discard the path struct or re-use the path struct, as we do here. */

	pathrelse(p_s_search_path);

	right_neighbor_of_leaf_node = 0;

	/* With each iteration of this loop we search through the items in the
	   current node, and calculate the next current node(next path element)
	   for the next iteration of this loop.. */
	n_block_number = SB_ROOT_BLOCK(sb);
	expected_level = -1;
	while (1) {

#ifdef CONFIG_REISERFS_CHECK
		if (!(++n_repeat_counter % 50000))
			reiserfs_warning(sb, "PAP-5100",
					 "%s: there were %d iterations of "
					 "while loop looking for key %K",
					 current->comm, n_repeat_counter,
					 p_s_key);
#endif

		/* prep path to have another element added to it. */
		p_s_last_element =
		    PATH_OFFSET_PELEMENT(p_s_search_path,
					 ++p_s_search_path->path_length);
		fs_gen = get_generation(sb);

		/* Read the next tree node, and set the last element in the path to
		   have a pointer to it. */
		if ((p_s_bh = p_s_last_element->pe_buffer =
		     sb_getblk(sb, n_block_number))) {
			if (!buffer_uptodate(p_s_bh) && reada_count > 1) {
				search_by_key_reada(sb, reada_bh,
						    reada_blocks, reada_count);
			}
			ll_rw_block(READ, 1, &p_s_bh);
			wait_on_buffer(p_s_bh);
			if (!buffer_uptodate(p_s_bh))
				goto io_error;
		} else {
		      io_error:
			p_s_search_path->path_length--;
			pathrelse(p_s_search_path);
			return IO_ERROR;
		}
		reada_count = 0;
		if (expected_level == -1)
			expected_level = SB_TREE_HEIGHT(sb);
		expected_level--;

		/* It is possible that schedule occurred. We must check whether the key
		   to search is still in the tree rooted from the current buffer. If
		   not then repeat search from the root. */
		if (fs_changed(fs_gen, sb) &&
		    (!B_IS_IN_TREE(p_s_bh) ||
		     B_LEVEL(p_s_bh) != expected_level ||
		     !key_in_buffer(p_s_search_path, p_s_key, sb))) {
			PROC_INFO_INC(sb, search_by_key_fs_changed);
			PROC_INFO_INC(sb, search_by_key_restarted);
			PROC_INFO_INC(sb,
				      sbk_restarted[expected_level - 1]);
			pathrelse(p_s_search_path);

			/* Get the root block number so that we can repeat the search
			   starting from the root. */
			n_block_number = SB_ROOT_BLOCK(sb);
			expected_level = -1;
			right_neighbor_of_leaf_node = 0;

			/* repeat search from the root */
			continue;
		}

		/* only check that the key is in the buffer if p_s_key is not
		   equal to the MAX_KEY. Latter case is only possible in
		   "finish_unfinished()" processing during mount. */
		RFALSE(comp_keys(&MAX_KEY, p_s_key) &&
		       !key_in_buffer(p_s_search_path, p_s_key, sb),
		       "PAP-5130: key is not in the buffer");
#ifdef CONFIG_REISERFS_CHECK
		if (cur_tb) {
			print_cur_tb("5140");
			reiserfs_panic(sb, "PAP-5140",
				       "schedule occurred in do_balance!");
		}
#endif

		// make sure, that the node contents look like a node of
		// certain level
		if (!is_tree_node(p_s_bh, expected_level)) {
			reiserfs_error(sb, "vs-5150",
				       "invalid format found in block %ld. "
				       "Fsck?", p_s_bh->b_blocknr);
			pathrelse(p_s_search_path);
			return IO_ERROR;
		}

		/* ok, we have acquired next formatted node in the tree */
		n_node_level = B_LEVEL(p_s_bh);

		PROC_INFO_BH_STAT(sb, p_s_bh, n_node_level - 1);

		RFALSE(n_node_level < n_stop_level,
		       "vs-5152: tree level (%d) is less than stop level (%d)",
		       n_node_level, n_stop_level);

		n_retval = bin_search(p_s_key, B_N_PITEM_HEAD(p_s_bh, 0),
				      B_NR_ITEMS(p_s_bh),
				      (n_node_level ==
				       DISK_LEAF_NODE_LEVEL) ? IH_SIZE :
				      KEY_SIZE,
				      &(p_s_last_element->pe_position));
		if (n_node_level == n_stop_level) {
			return n_retval;
		}

		/* we are not in the stop level */
		if (n_retval == ITEM_FOUND)
			/* item has been found, so we choose the pointer which is to the right of the found one */
			p_s_last_element->pe_position++;

		/* if item was not found we choose the position which is to
		   the left of the found item. This requires no code,
		   bin_search did it already. */

		/* So we have chosen a position in the current node which is
		   an internal node.  Now we calculate child block number by
		   position in the node. */
		n_block_number =
		    B_N_CHILD_NUM(p_s_bh, p_s_last_element->pe_position);

		/* if we are going to read leaf nodes, try for read ahead as well */
		if ((p_s_search_path->reada & PATH_READA) &&
		    n_node_level == DISK_LEAF_NODE_LEVEL + 1) {
			int pos = p_s_last_element->pe_position;
			int limit = B_NR_ITEMS(p_s_bh);
			struct reiserfs_key *le_key;

			if (p_s_search_path->reada & PATH_READA_BACK)
				limit = 0;
			while (reada_count < SEARCH_BY_KEY_READA) {
				if (pos == limit)
					break;
				reada_blocks[reada_count++] =
				    B_N_CHILD_NUM(p_s_bh, pos);
				if (p_s_search_path->reada & PATH_READA_BACK)
					pos--;
				else
					pos++;

				/*
				 * check to make sure we're in the same object
				 */
				le_key = B_N_PDELIM_KEY(p_s_bh, pos);
				if (le32_to_cpu(le_key->k_objectid) !=
				    p_s_key->on_disk_key.k_objectid) {
					break;
				}
			}
		}
	}
}

/* Form the path to an item and position in this item which contains
   file byte defined by p_s_key. If there is no such item
   corresponding to the key, we point the path to the item with
   maximal key less than p_s_key, and *p_n_pos_in_item is set to one
   past the last entry/byte in the item.  If searching for entry in a
   directory item, and it is not found, *p_n_pos_in_item is set to one
   entry more than the entry with maximal key which is less than the
   sought key.

   Note that if there is no entry in this same node which is one more,
   then we point to an imaginary entry.  for direct items, the
   position is in units of bytes, for indirect items the position is
   in units of blocknr entries, for directory items the position is in
   units of directory entries.  */

/* The function is NOT SCHEDULE-SAFE! */
int search_for_position_by_key(struct super_block *sb,	/* Pointer to the super block.          */
			       const struct cpu_key *p_cpu_key,	/* Key to search (cpu variable)         */
			       struct treepath *p_s_search_path	/* Filled up by this function.          */
    )
{
	struct item_head *p_le_ih;	/* pointer to on-disk structure */
	int n_blk_size;
	loff_t item_offset, offset;
	struct reiserfs_dir_entry de;
	int retval;

	/* If searching for directory entry. */
	if (is_direntry_cpu_key(p_cpu_key))
		return search_by_entry_key(sb, p_cpu_key, p_s_search_path,
					   &de);

	/* If not searching for directory entry. */

	/* If item is found. */
	retval = search_item(sb, p_cpu_key, p_s_search_path);
	if (retval == IO_ERROR)
		return retval;
	if (retval == ITEM_FOUND) {

		RFALSE(!ih_item_len
		       (B_N_PITEM_HEAD
			(PATH_PLAST_BUFFER(p_s_search_path),
			 PATH_LAST_POSITION(p_s_search_path))),
		       "PAP-5165: item length equals zero");

		pos_in_item(p_s_search_path) = 0;
		return POSITION_FOUND;
	}

	RFALSE(!PATH_LAST_POSITION(p_s_search_path),
	       "PAP-5170: position equals zero");

	/* Item is not found. Set path to the previous item. */
	p_le_ih =
	    B_N_PITEM_HEAD(PATH_PLAST_BUFFER(p_s_search_path),
			   --PATH_LAST_POSITION(p_s_search_path));
	n_blk_size = sb->s_blocksize;

	if (comp_short_keys(&(p_le_ih->ih_key), p_cpu_key)) {
		return FILE_NOT_FOUND;
	}
	// FIXME: quite ugly this far

	item_offset = le_ih_k_offset(p_le_ih);
	offset = cpu_key_k_offset(p_cpu_key);

	/* Needed byte is contained in the item pointed to by the path. */
	if (item_offset <= offset &&
	    item_offset + op_bytes_number(p_le_ih, n_blk_size) > offset) {
		pos_in_item(p_s_search_path) = offset - item_offset;
		if (is_indirect_le_ih(p_le_ih)) {
			pos_in_item(p_s_search_path) /= n_blk_size;
		}
		return POSITION_FOUND;
	}

	/* Needed byte is not contained in the item pointed to by the
	   path. Set pos_in_item out of the item. */
	if (is_indirect_le_ih(p_le_ih))
		pos_in_item(p_s_search_path) =
		    ih_item_len(p_le_ih) / UNFM_P_SIZE;
	else
		pos_in_item(p_s_search_path) = ih_item_len(p_le_ih);

	return POSITION_NOT_FOUND;
}

/* Compare given item and item pointed to by the path. */
int comp_items(const struct item_head *stored_ih, const struct treepath *p_s_path)
{
	struct buffer_head *p_s_bh;
	struct item_head *ih;

	/* Last buffer at the path is not in the tree. */
	if (!B_IS_IN_TREE(p_s_bh = PATH_PLAST_BUFFER(p_s_path)))
		return 1;

	/* Last path position is invalid. */
	if (PATH_LAST_POSITION(p_s_path) >= B_NR_ITEMS(p_s_bh))
		return 1;

	/* we need only to know, whether it is the same item */
	ih = get_ih(p_s_path);
	return memcmp(stored_ih, ih, IH_SIZE);
}

/* unformatted nodes are not logged anymore, ever.  This is safe
** now
*/
#define held_by_others(bh) (atomic_read(&(bh)->b_count) > 1)

// block can not be forgotten as it is in I/O or held by someone
#define block_in_use(bh) (buffer_locked(bh) || (held_by_others(bh)))

// prepare for delete or cut of direct item
static inline int prepare_for_direct_item(struct treepath *path,
					  struct item_head *le_ih,
					  struct inode *inode,
					  loff_t new_file_length, int *cut_size)
{
	loff_t round_len;

	if (new_file_length == max_reiserfs_offset(inode)) {
		/* item has to be deleted */
		*cut_size = -(IH_SIZE + ih_item_len(le_ih));
		return M_DELETE;
	}
	// new file gets truncated
	if (get_inode_item_key_version(inode) == KEY_FORMAT_3_6) {
		//
		round_len = ROUND_UP(new_file_length);
		/* this was n_new_file_length < le_ih ... */
		if (round_len < le_ih_k_offset(le_ih)) {
			*cut_size = -(IH_SIZE + ih_item_len(le_ih));
			return M_DELETE;	/* Delete this item. */
		}
		/* Calculate first position and size for cutting from item. */
		pos_in_item(path) = round_len - (le_ih_k_offset(le_ih) - 1);
		*cut_size = -(ih_item_len(le_ih) - pos_in_item(path));

		return M_CUT;	/* Cut from this item. */
	}

	// old file: items may have any length

	if (new_file_length < le_ih_k_offset(le_ih)) {
		*cut_size = -(IH_SIZE + ih_item_len(le_ih));
		return M_DELETE;	/* Delete this item. */
	}
	/* Calculate first position and size for cutting from item. */
	*cut_size = -(ih_item_len(le_ih) -
		      (pos_in_item(path) =
		       new_file_length + 1 - le_ih_k_offset(le_ih)));
	return M_CUT;		/* Cut from this item. */
}

static inline int prepare_for_direntry_item(struct treepath *path,
					    struct item_head *le_ih,
					    struct inode *inode,
					    loff_t new_file_length,
					    int *cut_size)
{
	if (le_ih_k_offset(le_ih) == DOT_OFFSET &&
	    new_file_length == max_reiserfs_offset(inode)) {
		RFALSE(ih_entry_count(le_ih) != 2,
		       "PAP-5220: incorrect empty directory item (%h)", le_ih);
		*cut_size = -(IH_SIZE + ih_item_len(le_ih));
		return M_DELETE;	/* Delete the directory item containing "." and ".." entry. */
	}

	if (ih_entry_count(le_ih) == 1) {
		/* Delete the directory item such as there is one record only
		   in this item */
		*cut_size = -(IH_SIZE + ih_item_len(le_ih));
		return M_DELETE;
	}

	/* Cut one record from the directory item. */
	*cut_size =
	    -(DEH_SIZE +
	      entry_length(get_last_bh(path), le_ih, pos_in_item(path)));
	return M_CUT;
}

#define JOURNAL_FOR_FREE_BLOCK_AND_UPDATE_SD (2 * JOURNAL_PER_BALANCE_CNT + 1)

/*  If the path points to a directory or direct item, calculate mode and the size cut, for balance.
    If the path points to an indirect item, remove some number of its unformatted nodes.
    In case of file truncate calculate whether this item must be deleted/truncated or last
    unformatted node of this item will be converted to a direct item.
    This function returns a determination of what balance mode the calling function should employ. */
static char prepare_for_delete_or_cut(struct reiserfs_transaction_handle *th, struct inode *inode, struct treepath *p_s_path, const struct cpu_key *p_s_item_key, int *p_n_removed,	/* Number of unformatted nodes which were removed
																						   from end of the file. */
				      int *p_n_cut_size, unsigned long long n_new_file_length	/* MAX_KEY_OFFSET in case of delete. */
    )
{
	struct super_block *sb = inode->i_sb;
	struct item_head *p_le_ih = PATH_PITEM_HEAD(p_s_path);
	struct buffer_head *p_s_bh = PATH_PLAST_BUFFER(p_s_path);

	BUG_ON(!th->t_trans_id);

	/* Stat_data item. */
	if (is_statdata_le_ih(p_le_ih)) {

		RFALSE(n_new_file_length != max_reiserfs_offset(inode),
		       "PAP-5210: mode must be M_DELETE");

		*p_n_cut_size = -(IH_SIZE + ih_item_len(p_le_ih));
		return M_DELETE;
	}

	/* Directory item. */
	if (is_direntry_le_ih(p_le_ih))
		return prepare_for_direntry_item(p_s_path, p_le_ih, inode,
						 n_new_file_length,
						 p_n_cut_size);

	/* Direct item. */
	if (is_direct_le_ih(p_le_ih))
		return prepare_for_direct_item(p_s_path, p_le_ih, inode,
					       n_new_file_length, p_n_cut_size);

	/* Case of an indirect item. */
	{
	    int blk_size = sb->s_blocksize;
	    struct item_head s_ih;
	    int need_re_search;
	    int delete = 0;
	    int result = M_CUT;
	    int pos = 0;

	    if ( n_new_file_length == max_reiserfs_offset (inode) ) {
		/* prepare_for_delete_or_cut() is called by
		 * reiserfs_delete_item() */
		n_new_file_length = 0;
		delete = 1;
	    }

	    do {
		need_re_search = 0;
		*p_n_cut_size = 0;
		p_s_bh = PATH_PLAST_BUFFER(p_s_path);
		copy_item_head(&s_ih, PATH_PITEM_HEAD(p_s_path));
		pos = I_UNFM_NUM(&s_ih);

		while (le_ih_k_offset (&s_ih) + (pos - 1) * blk_size > n_new_file_length) {
		    __le32 *unfm;
		    __u32 block;

		    /* Each unformatted block deletion may involve one additional
		     * bitmap block into the transaction, thereby the initial
		     * journal space reservation might not be enough. */
		    if (!delete && (*p_n_cut_size) != 0 &&
			reiserfs_transaction_free_space(th) < JOURNAL_FOR_FREE_BLOCK_AND_UPDATE_SD) {
			break;
		    }

		    unfm = (__le32 *)B_I_PITEM(p_s_bh, &s_ih) + pos - 1;
		    block = get_block_num(unfm, 0);

		    if (block != 0) {
			reiserfs_prepare_for_journal(sb, p_s_bh, 1);
			put_block_num(unfm, 0, 0);
			journal_mark_dirty (th, sb, p_s_bh);
			reiserfs_free_block(th, inode, block, 1);
		    }

		    cond_resched();

		    if (item_moved (&s_ih, p_s_path))  {
			need_re_search = 1;
			break;
		    }

		    pos --;
		    (*p_n_removed) ++;
		    (*p_n_cut_size) -= UNFM_P_SIZE;

		    if (pos == 0) {
			(*p_n_cut_size) -= IH_SIZE;
			result = M_DELETE;
			break;
		    }
		}
		/* a trick.  If the buffer has been logged, this will do nothing.  If
		** we've broken the loop without logging it, it will restore the
		** buffer */
		reiserfs_restore_prepared_buffer(sb, p_s_bh);
	    } while (need_re_search &&
		     search_for_position_by_key(sb, p_s_item_key, p_s_path) == POSITION_FOUND);
	    pos_in_item(p_s_path) = pos * UNFM_P_SIZE;

	    if (*p_n_cut_size == 0) {
		/* Nothing were cut. maybe convert last unformatted node to the
		 * direct item? */
		result = M_CONVERT;
	    }
	    return result;
	}
}

/* Calculate number of bytes which will be deleted or cut during balance */
static int calc_deleted_bytes_number(struct tree_balance *p_s_tb, char c_mode)
{
	int n_del_size;
	struct item_head *p_le_ih = PATH_PITEM_HEAD(p_s_tb->tb_path);

	if (is_statdata_le_ih(p_le_ih))
		return 0;

	n_del_size =
	    (c_mode ==
	     M_DELETE) ? ih_item_len(p_le_ih) : -p_s_tb->insert_size[0];
	if (is_direntry_le_ih(p_le_ih)) {
		// return EMPTY_DIR_SIZE; /* We delete emty directoris only. */
		// we can't use EMPTY_DIR_SIZE, as old format dirs have a different
		// empty size.  ick. FIXME, is this right?
		//
		return n_del_size;
	}

	if (is_indirect_le_ih(p_le_ih))
		n_del_size = (n_del_size / UNFM_P_SIZE) * (PATH_PLAST_BUFFER(p_s_tb->tb_path)->b_size);	// - get_ih_free_space (p_le_ih);
	return n_del_size;
}

static void init_tb_struct(struct reiserfs_transaction_handle *th,
			   struct tree_balance *p_s_tb,
			   struct super_block *sb,
			   struct treepath *p_s_path, int n_size)
{

	BUG_ON(!th->t_trans_id);

	memset(p_s_tb, '\0', sizeof(struct tree_balance));
	p_s_tb->transaction_handle = th;
	p_s_tb->tb_sb = sb;
	p_s_tb->tb_path = p_s_path;
	PATH_OFFSET_PBUFFER(p_s_path, ILLEGAL_PATH_ELEMENT_OFFSET) = NULL;
	PATH_OFFSET_POSITION(p_s_path, ILLEGAL_PATH_ELEMENT_OFFSET) = 0;
	p_s_tb->insert_size[0] = n_size;
}

void padd_item(char *item, int total_length, int length)
{
	int i;

	for (i = total_length; i > length;)
		item[--i] = 0;
}

#ifdef REISERQUOTA_DEBUG
char key2type(struct reiserfs_key *ih)
{
	if (is_direntry_le_key(2, ih))
		return 'd';
	if (is_direct_le_key(2, ih))
		return 'D';
	if (is_indirect_le_key(2, ih))
		return 'i';
	if (is_statdata_le_key(2, ih))
		return 's';
	return 'u';
}

char head2type(struct item_head *ih)
{
	if (is_direntry_le_ih(ih))
		return 'd';
	if (is_direct_le_ih(ih))
		return 'D';
	if (is_indirect_le_ih(ih))
		return 'i';
	if (is_statdata_le_ih(ih))
		return 's';
	return 'u';
}
#endif

/* Delete object item. */
int reiserfs_delete_item(struct reiserfs_transaction_handle *th, struct treepath *p_s_path,	/* Path to the deleted item. */
			 const struct cpu_key *p_s_item_key,	/* Key to search for the deleted item.  */
			 struct inode *p_s_inode,	/* inode is here just to update i_blocks and quotas */
			 struct buffer_head *p_s_un_bh)
{				/* NULL or unformatted node pointer.    */
	struct super_block *sb = p_s_inode->i_sb;
	struct tree_balance s_del_balance;
	struct item_head s_ih;
	struct item_head *q_ih;
	int quota_cut_bytes;
	int n_ret_value, n_del_size, n_removed;

#ifdef CONFIG_REISERFS_CHECK
	char c_mode;
	int n_iter = 0;
#endif

	BUG_ON(!th->t_trans_id);

	init_tb_struct(th, &s_del_balance, sb, p_s_path,
		       0 /*size is unknown */ );

	while (1) {
		n_removed = 0;

#ifdef CONFIG_REISERFS_CHECK
		n_iter++;
		c_mode =
#endif
		    prepare_for_delete_or_cut(th, p_s_inode, p_s_path,
					      p_s_item_key, &n_removed,
					      &n_del_size,
					      max_reiserfs_offset(p_s_inode));

		RFALSE(c_mode != M_DELETE, "PAP-5320: mode must be M_DELETE");

		copy_item_head(&s_ih, PATH_PITEM_HEAD(p_s_path));
		s_del_balance.insert_size[0] = n_del_size;

		n_ret_value = fix_nodes(M_DELETE, &s_del_balance, NULL, NULL);
		if (n_ret_value != REPEAT_SEARCH)
			break;

		PROC_INFO_INC(sb, delete_item_restarted);

		// file system changed, repeat search
		n_ret_value =
		    search_for_position_by_key(sb, p_s_item_key, p_s_path);
		if (n_ret_value == IO_ERROR)
			break;
		if (n_ret_value == FILE_NOT_FOUND) {
			reiserfs_warning(sb, "vs-5340",
					 "no items of the file %K found",
					 p_s_item_key);
			break;
		}
	}			/* while (1) */

	if (n_ret_value != CARRY_ON) {
		unfix_nodes(&s_del_balance);
		return 0;
	}
	// reiserfs_delete_item returns item length when success
	n_ret_value = calc_deleted_bytes_number(&s_del_balance, M_DELETE);
	q_ih = get_ih(p_s_path);
	quota_cut_bytes = ih_item_len(q_ih);

	/* hack so the quota code doesn't have to guess if the file
	 ** has a tail.  On tail insert, we allocate quota for 1 unformatted node.
	 ** We test the offset because the tail might have been
	 ** split into multiple items, and we only want to decrement for
	 ** the unfm node once
	 */
	if (!S_ISLNK(p_s_inode->i_mode) && is_direct_le_ih(q_ih)) {
		if ((le_ih_k_offset(q_ih) & (sb->s_blocksize - 1)) == 1) {
			quota_cut_bytes = sb->s_blocksize + UNFM_P_SIZE;
		} else {
			quota_cut_bytes = 0;
		}
	}

	if (p_s_un_bh) {
		int off;
		char *data;

		/* We are in direct2indirect conversion, so move tail contents
		   to the unformatted node */
		/* note, we do the copy before preparing the buffer because we
		 ** don't care about the contents of the unformatted node yet.
		 ** the only thing we really care about is the direct item's data
		 ** is in the unformatted node.
		 **
		 ** Otherwise, we would have to call reiserfs_prepare_for_journal on
		 ** the unformatted node, which might schedule, meaning we'd have to
		 ** loop all the way back up to the start of the while loop.
		 **
		 ** The unformatted node must be dirtied later on.  We can't be
		 ** sure here if the entire tail has been deleted yet.
		 **
		 ** p_s_un_bh is from the page cache (all unformatted nodes are
		 ** from the page cache) and might be a highmem page.  So, we
		 ** can't use p_s_un_bh->b_data.
		 ** -clm
		 */

		data = kmap_atomic(p_s_un_bh->b_page, KM_USER0);
		off = ((le_ih_k_offset(&s_ih) - 1) & (PAGE_CACHE_SIZE - 1));
		memcpy(data + off,
		       B_I_PITEM(PATH_PLAST_BUFFER(p_s_path), &s_ih),
		       n_ret_value);
		kunmap_atomic(data, KM_USER0);
	}
	/* Perform balancing after all resources have been collected at once. */
	do_balance(&s_del_balance, NULL, NULL, M_DELETE);

#ifdef REISERQUOTA_DEBUG
	reiserfs_debug(sb, REISERFS_DEBUG_CODE,
		       "reiserquota delete_item(): freeing %u, id=%u type=%c",
		       quota_cut_bytes, p_s_inode->i_uid, head2type(&s_ih));
#endif
	DQUOT_FREE_SPACE_NODIRTY(p_s_inode, quota_cut_bytes);

	/* Return deleted body length */
	return n_ret_value;
}

/* Summary Of Mechanisms For Handling Collisions Between Processes:

 deletion of the body of the object is performed by iput(), with the
 result that if multiple processes are operating on a file, the
 deletion of the body of the file is deferred until the last process
 that has an open inode performs its iput().

 writes and truncates are protected from collisions by use of
 semaphores.

 creates, linking, and mknod are protected from collisions with other
 processes by making the reiserfs_add_entry() the last step in the
 creation, and then rolling back all changes if there was a collision.
 - Hans
*/

/* this deletes item which never gets split */
void reiserfs_delete_solid_item(struct reiserfs_transaction_handle *th,
				struct inode *inode, struct reiserfs_key *key)
{
	struct tree_balance tb;
	INITIALIZE_PATH(path);
	int item_len = 0;
	int tb_init = 0;
	struct cpu_key cpu_key;
	int retval;
	int quota_cut_bytes = 0;

	BUG_ON(!th->t_trans_id);

	le_key2cpu_key(&cpu_key, key);

	while (1) {
		retval = search_item(th->t_super, &cpu_key, &path);
		if (retval == IO_ERROR) {
			reiserfs_error(th->t_super, "vs-5350",
				       "i/o failure occurred trying "
				       "to delete %K", &cpu_key);
			break;
		}
		if (retval != ITEM_FOUND) {
			pathrelse(&path);
			// No need for a warning, if there is just no free space to insert '..' item into the newly-created subdir
			if (!
			    ((unsigned long long)
			     GET_HASH_VALUE(le_key_k_offset
					    (le_key_version(key), key)) == 0
			     && (unsigned long long)
			     GET_GENERATION_NUMBER(le_key_k_offset
						   (le_key_version(key),
						    key)) == 1))
				reiserfs_warning(th->t_super, "vs-5355",
						 "%k not found", key);
			break;
		}
		if (!tb_init) {
			tb_init = 1;
			item_len = ih_item_len(PATH_PITEM_HEAD(&path));
			init_tb_struct(th, &tb, th->t_super, &path,
				       -(IH_SIZE + item_len));
		}
		quota_cut_bytes = ih_item_len(PATH_PITEM_HEAD(&path));

		retval = fix_nodes(M_DELETE, &tb, NULL, NULL);
		if (retval == REPEAT_SEARCH) {
			PROC_INFO_INC(th->t_super, delete_solid_item_restarted);
			continue;
		}

		if (retval == CARRY_ON) {
			do_balance(&tb, NULL, NULL, M_DELETE);
			if (inode) {	/* Should we count quota for item? (we don't count quotas for save-links) */
#ifdef REISERQUOTA_DEBUG
				reiserfs_debug(th->t_super, REISERFS_DEBUG_CODE,
					       "reiserquota delete_solid_item(): freeing %u id=%u type=%c",
					       quota_cut_bytes, inode->i_uid,
					       key2type(key));
#endif
				DQUOT_FREE_SPACE_NODIRTY(inode,
							 quota_cut_bytes);
			}
			break;
		}
		// IO_ERROR, NO_DISK_SPACE, etc
		reiserfs_warning(th->t_super, "vs-5360",
				 "could not delete %K due to fix_nodes failure",
				 &cpu_key);
		unfix_nodes(&tb);
		break;
	}

	reiserfs_check_path(&path);
}

int reiserfs_delete_object(struct reiserfs_transaction_handle *th,
			   struct inode *inode)
{
	int err;
	inode->i_size = 0;
	BUG_ON(!th->t_trans_id);

	/* for directory this deletes item containing "." and ".." */
	err =
	    reiserfs_do_truncate(th, inode, NULL, 0 /*no timestamp updates */ );
	if (err)
		return err;

#if defined( USE_INODE_GENERATION_COUNTER )
	if (!old_format_only(th->t_super)) {
		__le32 *inode_generation;

		inode_generation =
		    &REISERFS_SB(th->t_super)->s_rs->s_inode_generation;
		le32_add_cpu(inode_generation, 1);
	}
/* USE_INODE_GENERATION_COUNTER */
#endif
	reiserfs_delete_solid_item(th, inode, INODE_PKEY(inode));

	return err;
}

static void unmap_buffers(struct page *page, loff_t pos)
{
	struct buffer_head *bh;
	struct buffer_head *head;
	struct buffer_head *next;
	unsigned long tail_index;
	unsigned long cur_index;

	if (page) {
		if (page_has_buffers(page)) {
			tail_index = pos & (PAGE_CACHE_SIZE - 1);
			cur_index = 0;
			head = page_buffers(page);
			bh = head;
			do {
				next = bh->b_this_page;

				/* we want to unmap the buffers that contain the tail, and
				 ** all the buffers after it (since the tail must be at the
				 ** end of the file).  We don't want to unmap file data
				 ** before the tail, since it might be dirty and waiting to
				 ** reach disk
				 */
				cur_index += bh->b_size;
				if (cur_index > tail_index) {
					reiserfs_unmap_buffer(bh);
				}
				bh = next;
			} while (bh != head);
		}
	}
}

static int maybe_indirect_to_direct(struct reiserfs_transaction_handle *th,
				    struct inode *p_s_inode,
				    struct page *page,
				    struct treepath *p_s_path,
				    const struct cpu_key *p_s_item_key,
				    loff_t n_new_file_size, char *p_c_mode)
{
	struct super_block *sb = p_s_inode->i_sb;
	int n_block_size = sb->s_blocksize;
	int cut_bytes;
	BUG_ON(!th->t_trans_id);
	BUG_ON(n_new_file_size != p_s_inode->i_size);

	/* the page being sent in could be NULL if there was an i/o error
	 ** reading in the last block.  The user will hit problems trying to
	 ** read the file, but for now we just skip the indirect2direct
	 */
	if (atomic_read(&p_s_inode->i_count) > 1 ||
	    !tail_has_to_be_packed(p_s_inode) ||
	    !page || (REISERFS_I(p_s_inode)->i_flags & i_nopack_mask)) {
		/* leave tail in an unformatted node */
		*p_c_mode = M_SKIP_BALANCING;
		cut_bytes =
		    n_block_size - (n_new_file_size & (n_block_size - 1));
		pathrelse(p_s_path);
		return cut_bytes;
	}
	/* Permorm the conversion to a direct_item. */
	/*return indirect_to_direct (p_s_inode, p_s_path, p_s_item_key, n_new_file_size, p_c_mode); */
	return indirect2direct(th, p_s_inode, page, p_s_path, p_s_item_key,
			       n_new_file_size, p_c_mode);
}

/* we did indirect_to_direct conversion. And we have inserted direct
   item successesfully, but there were no disk space to cut unfm
   pointer being converted. Therefore we have to delete inserted
   direct item(s) */
static void indirect_to_direct_roll_back(struct reiserfs_transaction_handle *th,
					 struct inode *inode, struct treepath *path)
{
	struct cpu_key tail_key;
	int tail_len;
	int removed;
	BUG_ON(!th->t_trans_id);

	make_cpu_key(&tail_key, inode, inode->i_size + 1, TYPE_DIRECT, 4);	// !!!!
	tail_key.key_length = 4;

	tail_len =
	    (cpu_key_k_offset(&tail_key) & (inode->i_sb->s_blocksize - 1)) - 1;
	while (tail_len) {
		/* look for the last byte of the tail */
		if (search_for_position_by_key(inode->i_sb, &tail_key, path) ==
		    POSITION_NOT_FOUND)
			reiserfs_panic(inode->i_sb, "vs-5615",
				       "found invalid item");
		RFALSE(path->pos_in_item !=
		       ih_item_len(PATH_PITEM_HEAD(path)) - 1,
		       "vs-5616: appended bytes found");
		PATH_LAST_POSITION(path)--;

		removed =
		    reiserfs_delete_item(th, path, &tail_key, inode,
					 NULL /*unbh not needed */ );
		RFALSE(removed <= 0
		       || removed > tail_len,
		       "vs-5617: there was tail %d bytes, removed item length %d bytes",
		       tail_len, removed);
		tail_len -= removed;
		set_cpu_key_k_offset(&tail_key,
				     cpu_key_k_offset(&tail_key) - removed);
	}
	reiserfs_warning(inode->i_sb, "reiserfs-5091", "indirect_to_direct "
			 "conversion has been rolled back due to "
			 "lack of disk space");
	//mark_file_without_tail (inode);
	mark_inode_dirty(inode);
}

/* (Truncate or cut entry) or delete object item. Returns < 0 on failure */
int reiserfs_cut_from_item(struct reiserfs_transaction_handle *th,
			   struct treepath *p_s_path,
			   struct cpu_key *p_s_item_key,
			   struct inode *p_s_inode,
			   struct page *page, loff_t n_new_file_size)
{
	struct super_block *sb = p_s_inode->i_sb;
	/* Every function which is going to call do_balance must first
	   create a tree_balance structure.  Then it must fill up this
	   structure by using the init_tb_struct and fix_nodes functions.
	   After that we can make tree balancing. */
	struct tree_balance s_cut_balance;
	struct item_head *p_le_ih;
	int n_cut_size = 0,	/* Amount to be cut. */
	    n_ret_value = CARRY_ON, n_removed = 0,	/* Number of the removed unformatted nodes. */
	    n_is_inode_locked = 0;
	char c_mode;		/* Mode of the balance. */
	int retval2 = -1;
	int quota_cut_bytes;
	loff_t tail_pos = 0;

	BUG_ON(!th->t_trans_id);

	init_tb_struct(th, &s_cut_balance, p_s_inode->i_sb, p_s_path,
		       n_cut_size);

	/* Repeat this loop until we either cut the item without needing
	   to balance, or we fix_nodes without schedule occurring */
	while (1) {
		/* Determine the balance mode, position of the first byte to
		   be cut, and size to be cut.  In case of the indirect item
		   free unformatted nodes which are pointed to by the cut
		   pointers. */

		c_mode =
		    prepare_for_delete_or_cut(th, p_s_inode, p_s_path,
					      p_s_item_key, &n_removed,
					      &n_cut_size, n_new_file_size);
		if (c_mode == M_CONVERT) {
			/* convert last unformatted node to direct item or leave
			   tail in the unformatted node */
			RFALSE(n_ret_value != CARRY_ON,
			       "PAP-5570: can not convert twice");

			n_ret_value =
			    maybe_indirect_to_direct(th, p_s_inode, page,
						     p_s_path, p_s_item_key,
						     n_new_file_size, &c_mode);
			if (c_mode == M_SKIP_BALANCING)
				/* tail has been left in the unformatted node */
				return n_ret_value;

			n_is_inode_locked = 1;

			/* removing of last unformatted node will change value we
			   have to return to truncate. Save it */
			retval2 = n_ret_value;
			/*retval2 = sb->s_blocksize - (n_new_file_size & (sb->s_blocksize - 1)); */

			/* So, we have performed the first part of the conversion:
			   inserting the new direct item.  Now we are removing the
			   last unformatted node pointer. Set key to search for
			   it. */
			set_cpu_key_k_type(p_s_item_key, TYPE_INDIRECT);
			p_s_item_key->key_length = 4;
			n_new_file_size -=
			    (n_new_file_size & (sb->s_blocksize - 1));
			tail_pos = n_new_file_size;
			set_cpu_key_k_offset(p_s_item_key, n_new_file_size + 1);
			if (search_for_position_by_key
			    (sb, p_s_item_key,
			     p_s_path) == POSITION_NOT_FOUND) {
				print_block(PATH_PLAST_BUFFER(p_s_path), 3,
					    PATH_LAST_POSITION(p_s_path) - 1,
					    PATH_LAST_POSITION(p_s_path) + 1);
				reiserfs_panic(sb, "PAP-5580", "item to "
					       "convert does not exist (%K)",
					       p_s_item_key);
			}
			continue;
		}
		if (n_cut_size == 0) {
			pathrelse(p_s_path);
			return 0;
		}

		s_cut_balance.insert_size[0] = n_cut_size;

		n_ret_value = fix_nodes(c_mode, &s_cut_balance, NULL, NULL);
		if (n_ret_value != REPEAT_SEARCH)
			break;

		PROC_INFO_INC(sb, cut_from_item_restarted);

		n_ret_value =
		    search_for_position_by_key(sb, p_s_item_key, p_s_path);
		if (n_ret_value == POSITION_FOUND)
			continue;

		reiserfs_warning(sb, "PAP-5610", "item %K not found",
				 p_s_item_key);
		unfix_nodes(&s_cut_balance);
		return (n_ret_value == IO_ERROR) ? -EIO : -ENOENT;
	}			/* while */

	// check fix_nodes results (IO_ERROR or NO_DISK_SPACE)
	if (n_ret_value != CARRY_ON) {
		if (n_is_inode_locked) {
			// FIXME: this seems to be not needed: we are always able
			// to cut item
			indirect_to_direct_roll_back(th, p_s_inode, p_s_path);
		}
		if (n_ret_value == NO_DISK_SPACE)
			reiserfs_warning(sb, "reiserfs-5092",
					 "NO_DISK_SPACE");
		unfix_nodes(&s_cut_balance);
		return -EIO;
	}

	/* go ahead and perform balancing */

	RFALSE(c_mode == M_PASTE || c_mode == M_INSERT, "invalid mode");

	/* Calculate number of bytes that need to be cut from the item. */
	quota_cut_bytes =
	    (c_mode ==
	     M_DELETE) ? ih_item_len(get_ih(p_s_path)) : -s_cut_balance.
	    insert_size[0];
	if (retval2 == -1)
		n_ret_value = calc_deleted_bytes_number(&s_cut_balance, c_mode);
	else
		n_ret_value = retval2;

	/* For direct items, we only change the quota when deleting the last
	 ** item.
	 */
	p_le_ih = PATH_PITEM_HEAD(s_cut_balance.tb_path);
	if (!S_ISLNK(p_s_inode->i_mode) && is_direct_le_ih(p_le_ih)) {
		if (c_mode == M_DELETE &&
		    (le_ih_k_offset(p_le_ih) & (sb->s_blocksize - 1)) ==
		    1) {
			// FIXME: this is to keep 3.5 happy
			REISERFS_I(p_s_inode)->i_first_direct_byte = U32_MAX;
			quota_cut_bytes = sb->s_blocksize + UNFM_P_SIZE;
		} else {
			quota_cut_bytes = 0;
		}
	}
#ifdef CONFIG_REISERFS_CHECK
	if (n_is_inode_locked) {
		struct item_head *le_ih =
		    PATH_PITEM_HEAD(s_cut_balance.tb_path);
		/* we are going to complete indirect2direct conversion. Make
		   sure, that we exactly remove last unformatted node pointer
		   of the item */
		if (!is_indirect_le_ih(le_ih))
			reiserfs_panic(sb, "vs-5652",
				       "item must be indirect %h", le_ih);

		if (c_mode == M_DELETE && ih_item_len(le_ih) != UNFM_P_SIZE)
			reiserfs_panic(sb, "vs-5653", "completing "
				       "indirect2direct conversion indirect "
				       "item %h being deleted must be of "
				       "4 byte long", le_ih);

		if (c_mode == M_CUT
		    && s_cut_balance.insert_size[0] != -UNFM_P_SIZE) {
			reiserfs_panic(sb, "vs-5654", "can not complete "
				       "indirect2direct conversion of %h "
				       "(CUT, insert_size==%d)",
				       le_ih, s_cut_balance.insert_size[0]);
		}
		/* it would be useful to make sure, that right neighboring
		   item is direct item of this file */
	}
#endif

	do_balance(&s_cut_balance, NULL, NULL, c_mode);
	if (n_is_inode_locked) {
		/* we've done an indirect->direct conversion.  when the data block
		 ** was freed, it was removed from the list of blocks that must
		 ** be flushed before the transaction commits, make sure to
		 ** unmap and invalidate it
		 */
		unmap_buffers(page, tail_pos);
		REISERFS_I(p_s_inode)->i_flags &= ~i_pack_on_close_mask;
	}
#ifdef REISERQUOTA_DEBUG
	reiserfs_debug(p_s_inode->i_sb, REISERFS_DEBUG_CODE,
		       "reiserquota cut_from_item(): freeing %u id=%u type=%c",
		       quota_cut_bytes, p_s_inode->i_uid, '?');
#endif
	DQUOT_FREE_SPACE_NODIRTY(p_s_inode, quota_cut_bytes);
	return n_ret_value;
}

static void truncate_directory(struct reiserfs_transaction_handle *th,
			       struct inode *inode)
{
	BUG_ON(!th->t_trans_id);
	if (inode->i_nlink)
		reiserfs_error(inode->i_sb, "vs-5655", "link count != 0");

	set_le_key_k_offset(KEY_FORMAT_3_5, INODE_PKEY(inode), DOT_OFFSET);
	set_le_key_k_type(KEY_FORMAT_3_5, INODE_PKEY(inode), TYPE_DIRENTRY);
	reiserfs_delete_solid_item(th, inode, INODE_PKEY(inode));
	reiserfs_update_sd(th, inode);
	set_le_key_k_offset(KEY_FORMAT_3_5, INODE_PKEY(inode), SD_OFFSET);
	set_le_key_k_type(KEY_FORMAT_3_5, INODE_PKEY(inode), TYPE_STAT_DATA);
}

/* Truncate file to the new size. Note, this must be called with a transaction
   already started */
int reiserfs_do_truncate(struct reiserfs_transaction_handle *th, struct inode *p_s_inode,	/* ->i_size contains new
												   size */
			 struct page *page,	/* up to date for last block */
			 int update_timestamps	/* when it is called by
						   file_release to convert
						   the tail - no timestamps
						   should be updated */
    )
{
	INITIALIZE_PATH(s_search_path);	/* Path to the current object item. */
	struct item_head *p_le_ih;	/* Pointer to an item header. */
	struct cpu_key s_item_key;	/* Key to search for a previous file item. */
	loff_t n_file_size,	/* Old file size. */
	 n_new_file_size;	/* New file size. */
	int n_deleted;		/* Number of deleted or truncated bytes. */
	int retval;
	int err = 0;

	BUG_ON(!th->t_trans_id);
	if (!
	    (S_ISREG(p_s_inode->i_mode) || S_ISDIR(p_s_inode->i_mode)
	     || S_ISLNK(p_s_inode->i_mode)))
		return 0;

	if (S_ISDIR(p_s_inode->i_mode)) {
		// deletion of directory - no need to update timestamps
		truncate_directory(th, p_s_inode);
		return 0;
	}

	/* Get new file size. */
	n_new_file_size = p_s_inode->i_size;

	// FIXME: note, that key type is unimportant here
	make_cpu_key(&s_item_key, p_s_inode, max_reiserfs_offset(p_s_inode),
		     TYPE_DIRECT, 3);

	retval =
	    search_for_position_by_key(p_s_inode->i_sb, &s_item_key,
				       &s_search_path);
	if (retval == IO_ERROR) {
		reiserfs_error(p_s_inode->i_sb, "vs-5657",
			       "i/o failure occurred trying to truncate %K",
			       &s_item_key);
		err = -EIO;
		goto out;
	}
	if (retval == POSITION_FOUND || retval == FILE_NOT_FOUND) {
		reiserfs_error(p_s_inode->i_sb, "PAP-5660",
			       "wrong result %d of search for %K", retval,
			       &s_item_key);

		err = -EIO;
		goto out;
	}

	s_search_path.pos_in_item--;

	/* Get real file size (total length of all file items) */
	p_le_ih = PATH_PITEM_HEAD(&s_search_path);
	if (is_statdata_le_ih(p_le_ih))
		n_file_size = 0;
	else {
		loff_t offset = le_ih_k_offset(p_le_ih);
		int bytes =
		    op_bytes_number(p_le_ih, p_s_inode->i_sb->s_blocksize);

		/* this may mismatch with real file size: if last direct item
		   had no padding zeros and last unformatted node had no free
		   space, this file would have this file size */
		n_file_size = offset + bytes - 1;
	}
	/*
	 * are we doing a full truncate or delete, if so
	 * kick in the reada code
	 */
	if (n_new_file_size == 0)
		s_search_path.reada = PATH_READA | PATH_READA_BACK;

	if (n_file_size == 0 || n_file_size < n_new_file_size) {
		goto update_and_out;
	}

	/* Update key to search for the last file item. */
	set_cpu_key_k_offset(&s_item_key, n_file_size);

	do {
		/* Cut or delete file item. */
		n_deleted =
		    reiserfs_cut_from_item(th, &s_search_path, &s_item_key,
					   p_s_inode, page, n_new_file_size);
		if (n_deleted < 0) {
			reiserfs_warning(p_s_inode->i_sb, "vs-5665",
					 "reiserfs_cut_from_item failed");
			reiserfs_check_path(&s_search_path);
			return 0;
		}

		RFALSE(n_deleted > n_file_size,
		       "PAP-5670: reiserfs_cut_from_item: too many bytes deleted: deleted %d, file_size %lu, item_key %K",
		       n_deleted, n_file_size, &s_item_key);

		/* Change key to search the last file item. */
		n_file_size -= n_deleted;

		set_cpu_key_k_offset(&s_item_key, n_file_size);

		/* While there are bytes to truncate and previous file item is presented in the tree. */

		/*
		 ** This loop could take a really long time, and could log
		 ** many more blocks than a transaction can hold.  So, we do a polite
		 ** journal end here, and if the transaction needs ending, we make
		 ** sure the file is consistent before ending the current trans
		 ** and starting a new one
		 */
		if (journal_transaction_should_end(th, 0) ||
		    reiserfs_transaction_free_space(th) <= JOURNAL_FOR_FREE_BLOCK_AND_UPDATE_SD) {
			int orig_len_alloc = th->t_blocks_allocated;
			pathrelse(&s_search_path);

			if (update_timestamps) {
				p_s_inode->i_mtime = p_s_inode->i_ctime =
				    CURRENT_TIME_SEC;
			}
			reiserfs_update_sd(th, p_s_inode);

			err = journal_end(th, p_s_inode->i_sb, orig_len_alloc);
			if (err)
				goto out;
			err = journal_begin(th, p_s_inode->i_sb,
					    JOURNAL_FOR_FREE_BLOCK_AND_UPDATE_SD + JOURNAL_PER_BALANCE_CNT * 4) ;
			if (err)
				goto out;
			reiserfs_update_inode_transaction(p_s_inode);
		}
	} while (n_file_size > ROUND_UP(n_new_file_size) &&
		 search_for_position_by_key(p_s_inode->i_sb, &s_item_key,
					    &s_search_path) == POSITION_FOUND);

	RFALSE(n_file_size > ROUND_UP(n_new_file_size),
	       "PAP-5680: truncate did not finish: new_file_size %Ld, current %Ld, oid %d",
	       n_new_file_size, n_file_size, s_item_key.on_disk_key.k_objectid);

      update_and_out:
	if (update_timestamps) {
		// this is truncate, not file closing
		p_s_inode->i_mtime = p_s_inode->i_ctime = CURRENT_TIME_SEC;
	}
	reiserfs_update_sd(th, p_s_inode);

      out:
	pathrelse(&s_search_path);
	return err;
}

#ifdef CONFIG_REISERFS_CHECK
// this makes sure, that we __append__, not overwrite or add holes
static void check_research_for_paste(struct treepath *path,
				     const struct cpu_key *p_s_key)
{
	struct item_head *found_ih = get_ih(path);

	if (is_direct_le_ih(found_ih)) {
		if (le_ih_k_offset(found_ih) +
		    op_bytes_number(found_ih,
				    get_last_bh(path)->b_size) !=
		    cpu_key_k_offset(p_s_key)
		    || op_bytes_number(found_ih,
				       get_last_bh(path)->b_size) !=
		    pos_in_item(path))
			reiserfs_panic(NULL, "PAP-5720", "found direct item "
				       "%h or position (%d) does not match "
				       "to key %K", found_ih,
				       pos_in_item(path), p_s_key);
	}
	if (is_indirect_le_ih(found_ih)) {
		if (le_ih_k_offset(found_ih) +
		    op_bytes_number(found_ih,
				    get_last_bh(path)->b_size) !=
		    cpu_key_k_offset(p_s_key)
		    || I_UNFM_NUM(found_ih) != pos_in_item(path)
		    || get_ih_free_space(found_ih) != 0)
			reiserfs_panic(NULL, "PAP-5730", "found indirect "
				       "item (%h) or position (%d) does not "
				       "match to key (%K)",
				       found_ih, pos_in_item(path), p_s_key);
	}
}
#endif				/* config reiserfs check */

/* Paste bytes to the existing item. Returns bytes number pasted into the item. */
int reiserfs_paste_into_item(struct reiserfs_transaction_handle *th, struct treepath *p_s_search_path,	/* Path to the pasted item.          */
			     const struct cpu_key *p_s_key,	/* Key to search for the needed item. */
			     struct inode *inode,	/* Inode item belongs to */
			     const char *p_c_body,	/* Pointer to the bytes to paste.    */
			     int n_pasted_size)
{				/* Size of pasted bytes.             */
	struct tree_balance s_paste_balance;
	int retval;
	int fs_gen;

	BUG_ON(!th->t_trans_id);

	fs_gen = get_generation(inode->i_sb);

#ifdef REISERQUOTA_DEBUG
	reiserfs_debug(inode->i_sb, REISERFS_DEBUG_CODE,
		       "reiserquota paste_into_item(): allocating %u id=%u type=%c",
		       n_pasted_size, inode->i_uid,
		       key2type(&(p_s_key->on_disk_key)));
#endif

	if (DQUOT_ALLOC_SPACE_NODIRTY(inode, n_pasted_size)) {
		pathrelse(p_s_search_path);
		return -EDQUOT;
	}
	init_tb_struct(th, &s_paste_balance, th->t_super, p_s_search_path,
		       n_pasted_size);
#ifdef DISPLACE_NEW_PACKING_LOCALITIES
	s_paste_balance.key = p_s_key->on_disk_key;
#endif

	/* DQUOT_* can schedule, must check before the fix_nodes */
	if (fs_changed(fs_gen, inode->i_sb)) {
		goto search_again;
	}

	while ((retval =
		fix_nodes(M_PASTE, &s_paste_balance, NULL,
			  p_c_body)) == REPEAT_SEARCH) {
	      search_again:
		/* file system changed while we were in the fix_nodes */
		PROC_INFO_INC(th->t_super, paste_into_item_restarted);
		retval =
		    search_for_position_by_key(th->t_super, p_s_key,
					       p_s_search_path);
		if (retval == IO_ERROR) {
			retval = -EIO;
			goto error_out;
		}
		if (retval == POSITION_FOUND) {
			reiserfs_warning(inode->i_sb, "PAP-5710",
					 "entry or pasted byte (%K) exists",
					 p_s_key);
			retval = -EEXIST;
			goto error_out;
		}
#ifdef CONFIG_REISERFS_CHECK
		check_research_for_paste(p_s_search_path, p_s_key);
#endif
	}

	/* Perform balancing after all resources are collected by fix_nodes, and
	   accessing them will not risk triggering schedule. */
	if (retval == CARRY_ON) {
		do_balance(&s_paste_balance, NULL /*ih */ , p_c_body, M_PASTE);
		return 0;
	}
	retval = (retval == NO_DISK_SPACE) ? -ENOSPC : -EIO;
      error_out:
	/* this also releases the path */
	unfix_nodes(&s_paste_balance);
#ifdef REISERQUOTA_DEBUG
	reiserfs_debug(inode->i_sb, REISERFS_DEBUG_CODE,
		       "reiserquota paste_into_item(): freeing %u id=%u type=%c",
		       n_pasted_size, inode->i_uid,
		       key2type(&(p_s_key->on_disk_key)));
#endif
	DQUOT_FREE_SPACE_NODIRTY(inode, n_pasted_size);
	return retval;
}

/* Insert new item into the buffer at the path. */
int reiserfs_insert_item(struct reiserfs_transaction_handle *th, struct treepath *p_s_path,	/* Path to the inserteded item.         */
			 const struct cpu_key *key, struct item_head *p_s_ih,	/* Pointer to the item header to insert. */
			 struct inode *inode, const char *p_c_body)
{				/* Pointer to the bytes to insert.      */
	struct tree_balance s_ins_balance;
	int retval;
	int fs_gen = 0;
	int quota_bytes = 0;

	BUG_ON(!th->t_trans_id);

	if (inode) {		/* Do we count quotas for item? */
		fs_gen = get_generation(inode->i_sb);
		quota_bytes = ih_item_len(p_s_ih);

		/* hack so the quota code doesn't have to guess if the file has
		 ** a tail, links are always tails, so there's no guessing needed
		 */
		if (!S_ISLNK(inode->i_mode) && is_direct_le_ih(p_s_ih)) {
			quota_bytes = inode->i_sb->s_blocksize + UNFM_P_SIZE;
		}
#ifdef REISERQUOTA_DEBUG
		reiserfs_debug(inode->i_sb, REISERFS_DEBUG_CODE,
			       "reiserquota insert_item(): allocating %u id=%u type=%c",
			       quota_bytes, inode->i_uid, head2type(p_s_ih));
#endif
		/* We can't dirty inode here. It would be immediately written but
		 * appropriate stat item isn't inserted yet... */
		if (DQUOT_ALLOC_SPACE_NODIRTY(inode, quota_bytes)) {
			pathrelse(p_s_path);
			return -EDQUOT;
		}
	}
	init_tb_struct(th, &s_ins_balance, th->t_super, p_s_path,
		       IH_SIZE + ih_item_len(p_s_ih));
#ifdef DISPLACE_NEW_PACKING_LOCALITIES
	s_ins_balance.key = key->on_disk_key;
#endif
	/* DQUOT_* can schedule, must check to be sure calling fix_nodes is safe */
	if (inode && fs_changed(fs_gen, inode->i_sb)) {
		goto search_again;
	}

	while ((retval =
		fix_nodes(M_INSERT, &s_ins_balance, p_s_ih,
			  p_c_body)) == REPEAT_SEARCH) {
	      search_again:
		/* file system changed while we were in the fix_nodes */
		PROC_INFO_INC(th->t_super, insert_item_restarted);
		retval = search_item(th->t_super, key, p_s_path);
		if (retval == IO_ERROR) {
			retval = -EIO;
			goto error_out;
		}
		if (retval == ITEM_FOUND) {
			reiserfs_warning(th->t_super, "PAP-5760",
					 "key %K already exists in the tree",
					 key);
			retval = -EEXIST;
			goto error_out;
		}
	}

	/* make balancing after all resources will be collected at a time */
	if (retval == CARRY_ON) {
		do_balance(&s_ins_balance, p_s_ih, p_c_body, M_INSERT);
		return 0;
	}

	retval = (retval == NO_DISK_SPACE) ? -ENOSPC : -EIO;
      error_out:
	/* also releases the path */
	unfix_nodes(&s_ins_balance);
#ifdef REISERQUOTA_DEBUG
	reiserfs_debug(th->t_super, REISERFS_DEBUG_CODE,
		       "reiserquota insert_item(): freeing %u id=%u type=%c",
		       quota_bytes, inode->i_uid, head2type(p_s_ih));
#endif
	if (inode)
		DQUOT_FREE_SPACE_NODIRTY(inode, quota_bytes);
	return retval;
}
OpenPOWER on IntegriCloud