summaryrefslogtreecommitdiffstats
path: root/exfat/libexfat/node.c
blob: 56205fced863c70bc649fe659ee8226c4161c24d (plain) (blame)
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
/*
	node.c (09.10.09)
	exFAT file system implementation library.

	Free exFAT implementation.
	Copyright (C) 2010-2013  Andrew Nayenko

	This program is free software; you can redistribute it and/or modify
	it under the terms of the GNU General Public License as published by
	the Free Software Foundation, either version 2 of the License, or
	(at your option) any later version.

	This program is distributed in the hope that it will be useful,
	but WITHOUT ANY WARRANTY; without even the implied warranty of
	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
	GNU General Public License for more details.

	You should have received a copy of the GNU General Public License along
	with this program; if not, write to the Free Software Foundation, Inc.,
	51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
*/

#include "exfat.h"
#include <errno.h>
#include <string.h>
#include <inttypes.h>

/* on-disk nodes iterator */
struct iterator
{
	cluster_t cluster;
	off64_t offset;
	int contiguous;
	char* chunk;
};

struct exfat_node* exfat_get_node(struct exfat_node* node)
{
	/* if we switch to multi-threaded mode we will need atomic
	   increment here and atomic decrement in exfat_put_node() */
	node->references++;
	return node;
}

void exfat_put_node(struct exfat* ef, struct exfat_node* node)
{
	if (--node->references < 0)
	{
		char buffer[UTF8_BYTES(EXFAT_NAME_MAX) + 1];
		exfat_get_name(node, buffer, sizeof(buffer) - 1);
		exfat_bug("reference counter of `%s' is below zero", buffer);
	}

	if (node->references == 0)
	{
		/* FIXME handle I/O error */
		if (exfat_flush_node(ef, node) != 0)
			exfat_bug("node flush failed");
		if (node->flags & EXFAT_ATTRIB_UNLINKED)
		{
			/* free all clusters and node structure itself */
			exfat_truncate(ef, node, 0, true);
			free(node);
		}
		/* FIXME handle I/O error */
		if (exfat_flush(ef) != 0)
			exfat_bug("flush failed");
	}
}

/**
 * Cluster + offset from the beginning of the directory to absolute offset.
 */
static off64_t co2o(struct exfat* ef, cluster_t cluster, off64_t offset)
{
	return exfat_c2o(ef, cluster) + offset % CLUSTER_SIZE(*ef->sb);
}

static int opendir(struct exfat* ef, const struct exfat_node* dir,
		struct iterator* it)
{
	if (!(dir->flags & EXFAT_ATTRIB_DIR))
		exfat_bug("not a directory");
	it->cluster = dir->start_cluster;
	it->offset = 0;
	it->contiguous = IS_CONTIGUOUS(*dir);
	it->chunk = malloc(CLUSTER_SIZE(*ef->sb));
	if (it->chunk == NULL)
	{
		exfat_error("out of memory");
		return -ENOMEM;
	}
	if (exfat_pread(ef->dev, it->chunk, CLUSTER_SIZE(*ef->sb),
			exfat_c2o(ef, it->cluster)) < 0)
	{
		exfat_error("failed to read directory cluster %#x", it->cluster);
		return -EIO;
	}
	return 0;
}

static void closedir(struct iterator* it)
{
	it->cluster = 0;
	it->offset = 0;
	it->contiguous = 0;
	free(it->chunk);
	it->chunk = NULL;
}

static int fetch_next_entry(struct exfat* ef, const struct exfat_node* parent,
		struct iterator* it)
{
	/* move iterator to the next entry in the directory */
	it->offset += sizeof(struct exfat_entry);
	/* fetch the next cluster if needed */
	if ((it->offset & (CLUSTER_SIZE(*ef->sb) - 1)) == 0)
	{
		/* reached the end of directory; the caller should check this
		   condition too */
		if (it->offset >= parent->size)
			return 0;
		it->cluster = exfat_next_cluster(ef, parent, it->cluster);
		if (CLUSTER_INVALID(it->cluster))
		{
			exfat_error("invalid cluster 0x%x while reading directory",
					it->cluster);
			return 1;
		}
		if (exfat_pread(ef->dev, it->chunk, CLUSTER_SIZE(*ef->sb),
				exfat_c2o(ef, it->cluster)) < 0)
		{
			exfat_error("failed to read the next directory cluster %#x",
					it->cluster);
			return 1;
		}
	}
	return 0;
}

static struct exfat_node* allocate_node(void)
{
	struct exfat_node* node = malloc(sizeof(struct exfat_node));
	if (node == NULL)
	{
		exfat_error("failed to allocate node");
		return NULL;
	}
	memset(node, 0, sizeof(struct exfat_node));
	return node;
}

static void init_node_meta1(struct exfat_node* node,
		const struct exfat_entry_meta1* meta1)
{
	node->flags = le16_to_cpu(meta1->attrib);
	node->mtime = exfat_exfat2unix(meta1->mdate, meta1->mtime,
			meta1->mtime_cs);
	/* there is no centiseconds field for atime */
	node->atime = exfat_exfat2unix(meta1->adate, meta1->atime, 0);
}

static void init_node_meta2(struct exfat_node* node,
		const struct exfat_entry_meta2* meta2)
{
	node->size = le64_to_cpu(meta2->size);
	node->start_cluster = le32_to_cpu(meta2->start_cluster);
	node->fptr_cluster = node->start_cluster;
	if (meta2->flags & EXFAT_FLAG_CONTIGUOUS)
		node->flags |= EXFAT_ATTRIB_CONTIGUOUS;
}

static const struct exfat_entry* get_entry_ptr(const struct exfat* ef,
		const struct iterator* it)
{
	return (const struct exfat_entry*)
			(it->chunk + it->offset % CLUSTER_SIZE(*ef->sb));
}

/*
 * Reads one entry in directory at position pointed by iterator and fills
 * node structure.
 */
static int readdir(struct exfat* ef, const struct exfat_node* parent,
		struct exfat_node** node, struct iterator* it)
{
	int rc = -EIO;
	const struct exfat_entry* entry;
	const struct exfat_entry_meta1* meta1;
	const struct exfat_entry_meta2* meta2;
	const struct exfat_entry_name* file_name;
	const struct exfat_entry_upcase* upcase;
	const struct exfat_entry_bitmap* bitmap;
	const struct exfat_entry_label* label;
	uint8_t continuations = 0;
	le16_t* namep = NULL;
	uint16_t reference_checksum = 0;
	uint16_t actual_checksum = 0;
	uint64_t real_size = 0;

	*node = NULL;

	for (;;)
	{
		if (it->offset >= parent->size)
		{
			if (continuations != 0)
			{
				exfat_error("expected %hhu continuations", continuations);
				goto error;
			}
			return -ENOENT; /* that's OK, means end of directory */
		}

		entry = get_entry_ptr(ef, it);
		switch (entry->type)
		{
		case EXFAT_ENTRY_FILE:
			if (continuations != 0)
			{
				exfat_error("expected %hhu continuations before new entry",
						continuations);
				goto error;
			}
			meta1 = (const struct exfat_entry_meta1*) entry;
			continuations = meta1->continuations;
			/* each file entry must have at least 2 continuations:
			   info and name */
			if (continuations < 2)
			{
				exfat_error("too few continuations (%hhu)", continuations);
				goto error;
			}
			if (continuations > 1 +
					DIV_ROUND_UP(EXFAT_NAME_MAX, EXFAT_ENAME_MAX))
			{
				exfat_error("too many continuations (%hhu)", continuations);
				goto error;
			}
			reference_checksum = le16_to_cpu(meta1->checksum);
			actual_checksum = exfat_start_checksum(meta1);
			*node = allocate_node();
			if (*node == NULL)
			{
				rc = -ENOMEM;
				goto error;
			}
			/* new node has zero reference counter */
			(*node)->entry_cluster = it->cluster;
			(*node)->entry_offset = it->offset;
			init_node_meta1(*node, meta1);
			namep = (*node)->name;
			break;

		case EXFAT_ENTRY_FILE_INFO:
			if (continuations < 2)
			{
				exfat_error("unexpected continuation (%hhu)",
						continuations);
				goto error;
			}
			meta2 = (const struct exfat_entry_meta2*) entry;
			if (meta2->flags & ~(EXFAT_FLAG_ALWAYS1 | EXFAT_FLAG_CONTIGUOUS))
			{
				exfat_error("unknown flags in meta2 (0x%hhx)", meta2->flags);
				goto error;
			}
			init_node_meta2(*node, meta2);
			actual_checksum = exfat_add_checksum(entry, actual_checksum);
			real_size = le64_to_cpu(meta2->real_size);
			/* empty files must be marked as non-contiguous */
			if ((*node)->size == 0 && (meta2->flags & EXFAT_FLAG_CONTIGUOUS))
			{
				exfat_error("empty file marked as contiguous (0x%hhx)",
						meta2->flags);
				goto error;
			}
			/* directories must be aligned on at cluster boundary */
			if (((*node)->flags & EXFAT_ATTRIB_DIR) &&
				(*node)->size % CLUSTER_SIZE(*ef->sb) != 0)
			{
				exfat_error("directory has invalid size %"PRIu64" bytes",
						(*node)->size);
				goto error;
			}
			--continuations;
			break;

		case EXFAT_ENTRY_FILE_NAME:
			if (continuations == 0)
			{
				exfat_error("unexpected continuation");
				goto error;
			}
			file_name = (const struct exfat_entry_name*) entry;
			actual_checksum = exfat_add_checksum(entry, actual_checksum);

			memcpy(namep, file_name->name,
					MIN(EXFAT_ENAME_MAX,
						((*node)->name + EXFAT_NAME_MAX - namep)) *
					sizeof(le16_t));
			namep += EXFAT_ENAME_MAX;
			if (--continuations == 0)
			{
				/*
				   There are two fields that contain file size. Maybe they
				   plan to add compression support in the future and one of
				   those fields is visible (uncompressed) size and the other
				   is real (compressed) size. Anyway, currently it looks like
				   exFAT does not support compression and both fields must be
				   equal.

				   There is an exception though: pagefile.sys (its real_size
				   is always 0).
				*/
				if (real_size != (*node)->size)
				{
					char buffer[UTF8_BYTES(EXFAT_NAME_MAX) + 1];

					exfat_get_name(*node, buffer, sizeof(buffer) - 1);
					exfat_error("`%s' real size does not equal to size "
							"(%"PRIu64" != %"PRIu64")", buffer,
							real_size, (*node)->size);
					goto error;
				}
				if (actual_checksum != reference_checksum)
				{
					char buffer[UTF8_BYTES(EXFAT_NAME_MAX) + 1];

					exfat_get_name(*node, buffer, sizeof(buffer) - 1);
					exfat_error("`%s' has invalid checksum (0x%hx != 0x%hx)",
							buffer, actual_checksum, reference_checksum);
					goto error;
				}
				if (fetch_next_entry(ef, parent, it) != 0)
					goto error;
				return 0; /* entry completed */
			}
			break;

		case EXFAT_ENTRY_UPCASE:
			if (ef->upcase != NULL)
				break;
			upcase = (const struct exfat_entry_upcase*) entry;
			if (CLUSTER_INVALID(le32_to_cpu(upcase->start_cluster)))
			{
				exfat_error("invalid cluster 0x%x in upcase table",
						le32_to_cpu(upcase->start_cluster));
				goto error;
			}
			if (le64_to_cpu(upcase->size) == 0 ||
				le64_to_cpu(upcase->size) > 0xffff * sizeof(uint16_t) ||
				le64_to_cpu(upcase->size) % sizeof(uint16_t) != 0)
			{
				exfat_error("bad upcase table size (%"PRIu64" bytes)",
						le64_to_cpu(upcase->size));
				goto error;
			}
			ef->upcase = malloc(le64_to_cpu(upcase->size));
			if (ef->upcase == NULL)
			{
				exfat_error("failed to allocate upcase table (%"PRIu64" bytes)",
						le64_to_cpu(upcase->size));
				rc = -ENOMEM;
				goto error;
			}
			ef->upcase_chars = le64_to_cpu(upcase->size) / sizeof(le16_t);

			if (exfat_pread(ef->dev, ef->upcase, le64_to_cpu(upcase->size),
					exfat_c2o(ef, le32_to_cpu(upcase->start_cluster))) < 0)
			{
				exfat_error("failed to read upper case table "
						"(%"PRIu64" bytes starting at cluster %#x)",
						le64_to_cpu(upcase->size),
						le32_to_cpu(upcase->start_cluster));
				goto error;
			}
			break;

		case EXFAT_ENTRY_BITMAP:
			bitmap = (const struct exfat_entry_bitmap*) entry;
			ef->cmap.start_cluster = le32_to_cpu(bitmap->start_cluster);
			if (CLUSTER_INVALID(ef->cmap.start_cluster))
			{
				exfat_error("invalid cluster 0x%x in clusters bitmap",
						ef->cmap.start_cluster);
				goto error;
			}
			ef->cmap.size = le32_to_cpu(ef->sb->cluster_count) -
				EXFAT_FIRST_DATA_CLUSTER;
			if (le64_to_cpu(bitmap->size) < DIV_ROUND_UP(ef->cmap.size, 8))
			{
				exfat_error("invalid clusters bitmap size: %"PRIu64
						" (expected at least %u)",
						le64_to_cpu(bitmap->size),
						DIV_ROUND_UP(ef->cmap.size, 8));
				goto error;
			}
			/* FIXME bitmap can be rather big, up to 512 MB */
			ef->cmap.chunk_size = ef->cmap.size;
			ef->cmap.chunk = malloc(BMAP_SIZE(ef->cmap.chunk_size));
			if (ef->cmap.chunk == NULL)
			{
				exfat_error("failed to allocate clusters bitmap chunk "
						"(%"PRIu64" bytes)", le64_to_cpu(bitmap->size));
				rc = -ENOMEM;
				goto error;
			}

			if (exfat_pread(ef->dev, ef->cmap.chunk,
					BMAP_SIZE(ef->cmap.chunk_size),
					exfat_c2o(ef, ef->cmap.start_cluster)) < 0)
			{
				exfat_error("failed to read clusters bitmap "
						"(%"PRIu64" bytes starting at cluster %#x)",
						le64_to_cpu(bitmap->size), ef->cmap.start_cluster);
				goto error;
			}
			break;

		case EXFAT_ENTRY_LABEL:
			label = (const struct exfat_entry_label*) entry;
			if (label->length > EXFAT_ENAME_MAX)
			{
				exfat_error("too long label (%hhu chars)", label->length);
				goto error;
			}
			if (utf16_to_utf8(ef->label, label->name,
						sizeof(ef->label) - 1, EXFAT_ENAME_MAX) != 0)
				goto error;
			break;

		default:
			if (entry->type & EXFAT_ENTRY_VALID)
			{
				exfat_error("unknown entry type 0x%hhx", entry->type);
				goto error;
			}
			break;
		}

		if (fetch_next_entry(ef, parent, it) != 0)
			goto error;
	}
	/* we never reach here */

error:
	free(*node);
	*node = NULL;
	return rc;
}

int exfat_cache_directory(struct exfat* ef, struct exfat_node* dir)
{
	struct iterator it;
	int rc;
	struct exfat_node* node;
	struct exfat_node* current = NULL;

	if (dir->flags & EXFAT_ATTRIB_CACHED)
		return 0; /* already cached */

	rc = opendir(ef, dir, &it);
	if (rc != 0)
		return rc;
	while ((rc = readdir(ef, dir, &node, &it)) == 0)
	{
		node->parent = dir;
		if (current != NULL)
		{
			current->next = node;
			node->prev = current;
		}
		else
			dir->child = node;

		current = node;
	}
	closedir(&it);

	if (rc != -ENOENT)
	{
		/* rollback */
		for (current = dir->child; current; current = node)
		{
			node = current->next;
			free(current);
		}
		dir->child = NULL;
		return rc;
	}

	dir->flags |= EXFAT_ATTRIB_CACHED;
	return 0;
}

static void tree_attach(struct exfat_node* dir, struct exfat_node* node)
{
	node->parent = dir;
	if (dir->child)
	{
		dir->child->prev = node;
		node->next = dir->child;
	}
	dir->child = node;
}

static void tree_detach(struct exfat_node* node)
{
	if (node->prev)
		node->prev->next = node->next;
	else /* this is the first node in the list */
		node->parent->child = node->next;
	if (node->next)
		node->next->prev = node->prev;
	node->parent = NULL;
	node->prev = NULL;
	node->next = NULL;
}

static void reset_cache(struct exfat* ef, struct exfat_node* node)
{
	while (node->child)
	{
		struct exfat_node* p = node->child;
		reset_cache(ef, p);
		tree_detach(p);
		free(p);
	}
	node->flags &= ~EXFAT_ATTRIB_CACHED;
	if (node->references != 0)
	{
		char buffer[UTF8_BYTES(EXFAT_NAME_MAX) + 1];
		exfat_get_name(node, buffer, sizeof(buffer) - 1);
		exfat_warn("non-zero reference counter (%d) for `%s'",
				node->references, buffer);
	}
	while (node->references)
		exfat_put_node(ef, node);
}

void exfat_reset_cache(struct exfat* ef)
{
	reset_cache(ef, ef->root);
}

static void next_entry(struct exfat* ef, const struct exfat_node* parent,
		cluster_t* cluster, off64_t* offset)
{
	*offset += sizeof(struct exfat_entry);
	if (*offset % CLUSTER_SIZE(*ef->sb) == 0)
		/* next cluster cannot be invalid */
		*cluster = exfat_next_cluster(ef, parent, *cluster);
}

int exfat_flush_node(struct exfat* ef, struct exfat_node* node)
{
	cluster_t cluster;
	off64_t offset;
	off64_t meta1_offset, meta2_offset;
	struct exfat_entry_meta1 meta1;
	struct exfat_entry_meta2 meta2;

	if (!(node->flags & EXFAT_ATTRIB_DIRTY))
		return 0; /* no need to flush */

	if (ef->ro)
		exfat_bug("unable to flush node to read-only FS");

	if (node->parent == NULL)
		return 0; /* do not flush unlinked node */

	cluster = node->entry_cluster;
	offset = node->entry_offset;
	meta1_offset = co2o(ef, cluster, offset);
	next_entry(ef, node->parent, &cluster, &offset);
	meta2_offset = co2o(ef, cluster, offset);

	if (exfat_pread(ef->dev, &meta1, sizeof(meta1), meta1_offset) < 0)
	{
		exfat_error("failed to read meta1 entry on flush");
		return -EIO;
	}
	if (meta1.type != EXFAT_ENTRY_FILE)
		exfat_bug("invalid type of meta1: 0x%hhx", meta1.type);
	meta1.attrib = cpu_to_le16(node->flags);
	exfat_unix2exfat(node->mtime, &meta1.mdate, &meta1.mtime, &meta1.mtime_cs);
	exfat_unix2exfat(node->atime, &meta1.adate, &meta1.atime, NULL);

	if (exfat_pread(ef->dev, &meta2, sizeof(meta2), meta2_offset) < 0)
	{
		exfat_error("failed to read meta2 entry on flush");
		return -EIO;
	}
	if (meta2.type != EXFAT_ENTRY_FILE_INFO)
		exfat_bug("invalid type of meta2: 0x%hhx", meta2.type);
	meta2.size = meta2.real_size = cpu_to_le64(node->size);
	meta2.start_cluster = cpu_to_le32(node->start_cluster);
	meta2.flags = EXFAT_FLAG_ALWAYS1;
	/* empty files must not be marked as contiguous */
	if (node->size != 0 && IS_CONTIGUOUS(*node))
		meta2.flags |= EXFAT_FLAG_CONTIGUOUS;
	/* name hash remains unchanged, no need to recalculate it */

	meta1.checksum = exfat_calc_checksum(&meta1, &meta2, node->name);

	if (exfat_pwrite(ef->dev, &meta1, sizeof(meta1), meta1_offset) < 0)
	{
		exfat_error("failed to write meta1 entry on flush");
		return -EIO;
	}
	if (exfat_pwrite(ef->dev, &meta2, sizeof(meta2), meta2_offset) < 0)
	{
		exfat_error("failed to write meta2 entry on flush");
		return -EIO;
	}

	node->flags &= ~EXFAT_ATTRIB_DIRTY;
	return 0;
}

static bool erase_entry(struct exfat* ef, struct exfat_node* node)
{
	cluster_t cluster = node->entry_cluster;
	off64_t offset = node->entry_offset;
	int name_entries = DIV_ROUND_UP(utf16_length(node->name), EXFAT_ENAME_MAX);
	uint8_t entry_type;

	entry_type = EXFAT_ENTRY_FILE & ~EXFAT_ENTRY_VALID;
	if (exfat_pwrite(ef->dev, &entry_type, 1, co2o(ef, cluster, offset)) < 0)
	{
		exfat_error("failed to erase meta1 entry");
		return false;
	}

	next_entry(ef, node->parent, &cluster, &offset);
	entry_type = EXFAT_ENTRY_FILE_INFO & ~EXFAT_ENTRY_VALID;
	if (exfat_pwrite(ef->dev, &entry_type, 1, co2o(ef, cluster, offset)) < 0)
	{
		exfat_error("failed to erase meta2 entry");
		return false;
	}

	while (name_entries--)
	{
		next_entry(ef, node->parent, &cluster, &offset);
		entry_type = EXFAT_ENTRY_FILE_NAME & ~EXFAT_ENTRY_VALID;
		if (exfat_pwrite(ef->dev, &entry_type, 1,
				co2o(ef, cluster, offset)) < 0)
		{
			exfat_error("failed to erase name entry");
			return false;
		}
	}
	return true;
}

static int shrink_directory(struct exfat* ef, struct exfat_node* dir,
		off64_t deleted_offset)
{
	const struct exfat_node* node;
	const struct exfat_node* last_node;
	uint64_t entries = 0;
	uint64_t new_size;
	int rc;

	if (!(dir->flags & EXFAT_ATTRIB_DIR))
		exfat_bug("attempted to shrink a file");
	if (!(dir->flags & EXFAT_ATTRIB_CACHED))
		exfat_bug("attempted to shrink uncached directory");

	for (last_node = node = dir->child; node; node = node->next)
	{
		if (deleted_offset < node->entry_offset)
		{
			/* there are other entries after the removed one, no way to shrink
			   this directory */
			return 0;
		}
		if (last_node->entry_offset < node->entry_offset)
			last_node = node;
	}

	if (last_node)
	{
		/* offset of the last entry */
		entries += last_node->entry_offset / sizeof(struct exfat_entry);
		/* two subentries with meta info */
		entries += 2;
		/* subentries with file name */
		entries += DIV_ROUND_UP(utf16_length(last_node->name),
				EXFAT_ENAME_MAX);
	}

	new_size = DIV_ROUND_UP(entries * sizeof(struct exfat_entry),
				 CLUSTER_SIZE(*ef->sb)) * CLUSTER_SIZE(*ef->sb);
	if (new_size == 0) /* directory always has at least 1 cluster */
		new_size = CLUSTER_SIZE(*ef->sb);
	if (new_size == dir->size)
		return 0;
	rc = exfat_truncate(ef, dir, new_size, true);
	if (rc != 0)
		return rc;
	return 0;
}

static int delete(struct exfat* ef, struct exfat_node* node)
{
	struct exfat_node* parent = node->parent;
	off64_t deleted_offset = node->entry_offset;
	int rc;

	exfat_get_node(parent);
	if (!erase_entry(ef, node))
	{
		exfat_put_node(ef, parent);
		return -EIO;
	}
	exfat_update_mtime(parent);
	tree_detach(node);
	rc = shrink_directory(ef, parent, deleted_offset);
	exfat_put_node(ef, parent);
	/* file clusters will be freed when node reference counter becomes 0 */
	node->flags |= EXFAT_ATTRIB_UNLINKED;
	return rc;
}

int exfat_unlink(struct exfat* ef, struct exfat_node* node)
{
	if (node->flags & EXFAT_ATTRIB_DIR)
		return -EISDIR;
	return delete(ef, node);
}

int exfat_rmdir(struct exfat* ef, struct exfat_node* node)
{
	if (!(node->flags & EXFAT_ATTRIB_DIR))
		return -ENOTDIR;
	/* check that directory is empty */
	exfat_cache_directory(ef, node);
	if (node->child)
		return -ENOTEMPTY;
	return delete(ef, node);
}

static int grow_directory(struct exfat* ef, struct exfat_node* dir,
		uint64_t asize, uint32_t difference)
{
	return exfat_truncate(ef, dir,
			DIV_ROUND_UP(asize + difference, CLUSTER_SIZE(*ef->sb))
				* CLUSTER_SIZE(*ef->sb), true);
}

static int find_slot(struct exfat* ef, struct exfat_node* dir,
		cluster_t* cluster, off64_t* offset, int subentries)
{
	struct iterator it;
	int rc;
	const struct exfat_entry* entry;
	int contiguous = 0;

	rc = opendir(ef, dir, &it);
	if (rc != 0)
		return rc;
	for (;;)
	{
		if (contiguous == 0)
		{
			*cluster = it.cluster;
			*offset = it.offset;
		}
		entry = get_entry_ptr(ef, &it);
		if (entry->type & EXFAT_ENTRY_VALID)
			contiguous = 0;
		else
			contiguous++;
		if (contiguous == subentries)
			break;	/* suitable slot is found */
		if (it.offset + sizeof(struct exfat_entry) >= dir->size)
		{
			rc = grow_directory(ef, dir, dir->size,
					(subentries - contiguous) * sizeof(struct exfat_entry));
			if (rc != 0)
			{
				closedir(&it);
				return rc;
			}
		}
		if (fetch_next_entry(ef, dir, &it) != 0)
		{
			closedir(&it);
			return -EIO;
		}
	}
	closedir(&it);
	return 0;
}

static int write_entry(struct exfat* ef, struct exfat_node* dir,
		const le16_t* name, cluster_t cluster, off64_t offset, uint16_t attrib)
{
	struct exfat_node* node;
	struct exfat_entry_meta1 meta1;
	struct exfat_entry_meta2 meta2;
	const size_t name_length = utf16_length(name);
	const int name_entries = DIV_ROUND_UP(name_length, EXFAT_ENAME_MAX);
	int i;

	node = allocate_node();
	if (node == NULL)
		return -ENOMEM;
	node->entry_cluster = cluster;
	node->entry_offset = offset;
	memcpy(node->name, name, name_length * sizeof(le16_t));

	memset(&meta1, 0, sizeof(meta1));
	meta1.type = EXFAT_ENTRY_FILE;
	meta1.continuations = 1 + name_entries;
	meta1.attrib = cpu_to_le16(attrib);
	exfat_unix2exfat(time(NULL), &meta1.crdate, &meta1.crtime,
			&meta1.crtime_cs);
	meta1.adate = meta1.mdate = meta1.crdate;
	meta1.atime = meta1.mtime = meta1.crtime;
	meta1.mtime_cs = meta1.crtime_cs; /* there is no atime_cs */

	memset(&meta2, 0, sizeof(meta2));
	meta2.type = EXFAT_ENTRY_FILE_INFO;
	meta2.flags = EXFAT_FLAG_ALWAYS1;
	meta2.name_length = name_length;
	meta2.name_hash = exfat_calc_name_hash(ef, node->name);
	meta2.start_cluster = cpu_to_le32(EXFAT_CLUSTER_FREE);

	meta1.checksum = exfat_calc_checksum(&meta1, &meta2, node->name);

	if (exfat_pwrite(ef->dev, &meta1, sizeof(meta1),
			co2o(ef, cluster, offset)) < 0)
	{
		exfat_error("failed to write meta1 entry");
		return -EIO;
	}
	next_entry(ef, dir, &cluster, &offset);
	if (exfat_pwrite(ef->dev, &meta2, sizeof(meta2),
			co2o(ef, cluster, offset)) < 0)
	{
		exfat_error("failed to write meta2 entry");
		return -EIO;
	}
	for (i = 0; i < name_entries; i++)
	{
		struct exfat_entry_name name_entry = {EXFAT_ENTRY_FILE_NAME, 0};
		memcpy(name_entry.name, node->name + i * EXFAT_ENAME_MAX,
				MIN(EXFAT_ENAME_MAX, EXFAT_NAME_MAX - i * EXFAT_ENAME_MAX) *
				sizeof(le16_t));
		next_entry(ef, dir, &cluster, &offset);
		if (exfat_pwrite(ef->dev, &name_entry, sizeof(name_entry),
				co2o(ef, cluster, offset)) < 0)
		{
			exfat_error("failed to write name entry");
			return -EIO;
		}
	}

	init_node_meta1(node, &meta1);
	init_node_meta2(node, &meta2);

	tree_attach(dir, node);
	exfat_update_mtime(dir);
	return 0;
}

static int create(struct exfat* ef, const char* path, uint16_t attrib)
{
	struct exfat_node* dir;
	struct exfat_node* existing;
	cluster_t cluster = EXFAT_CLUSTER_BAD;
	off64_t offset = -1;
	le16_t name[EXFAT_NAME_MAX + 1];
	int rc;

	rc = exfat_split(ef, &dir, &existing, name, path);
	if (rc != 0)
		return rc;
	if (existing != NULL)
	{
		exfat_put_node(ef, existing);
		exfat_put_node(ef, dir);
		return -EEXIST;
	}

	rc = find_slot(ef, dir, &cluster, &offset,
			2 + DIV_ROUND_UP(utf16_length(name), EXFAT_ENAME_MAX));
	if (rc != 0)
	{
		exfat_put_node(ef, dir);
		return rc;
	}
	rc = write_entry(ef, dir, name, cluster, offset, attrib);
	exfat_put_node(ef, dir);
	return rc;
}

int exfat_mknod(struct exfat* ef, const char* path)
{
	return create(ef, path, EXFAT_ATTRIB_ARCH);
}

int exfat_mkdir(struct exfat* ef, const char* path)
{
	int rc;
	struct exfat_node* node;

	rc = create(ef, path, EXFAT_ATTRIB_ARCH | EXFAT_ATTRIB_DIR);
	if (rc != 0)
		return rc;
	rc = exfat_lookup(ef, &node, path);
	if (rc != 0)
		return 0;
	/* directories always have at least one cluster */
	rc = exfat_truncate(ef, node, CLUSTER_SIZE(*ef->sb), true);
	if (rc != 0)
	{
		delete(ef, node);
		exfat_put_node(ef, node);
		return rc;
	}
	exfat_put_node(ef, node);
	return 0;
}

static int rename_entry(struct exfat* ef, struct exfat_node* dir,
		struct exfat_node* node, const le16_t* name, cluster_t new_cluster,
		off64_t new_offset)
{
	struct exfat_entry_meta1 meta1;
	struct exfat_entry_meta2 meta2;
	cluster_t old_cluster = node->entry_cluster;
	off64_t old_offset = node->entry_offset;
	const size_t name_length = utf16_length(name);
	const int name_entries = DIV_ROUND_UP(name_length, EXFAT_ENAME_MAX);
	int i;

	if (exfat_pread(ef->dev, &meta1, sizeof(meta1),
			co2o(ef, old_cluster, old_offset)) < 0)
	{
		exfat_error("failed to read meta1 entry on rename");
		return -EIO;
	}
	next_entry(ef, node->parent, &old_cluster, &old_offset);
	if (exfat_pread(ef->dev, &meta2, sizeof(meta2),
			co2o(ef, old_cluster, old_offset)) < 0)
	{
		exfat_error("failed to read meta2 entry on rename");
		return -EIO;
	}
	meta1.continuations = 1 + name_entries;
	meta2.name_hash = exfat_calc_name_hash(ef, name);
	meta2.name_length = name_length;
	meta1.checksum = exfat_calc_checksum(&meta1, &meta2, name);

	if (!erase_entry(ef, node))
		return -EIO;

	node->entry_cluster = new_cluster;
	node->entry_offset = new_offset;

	if (exfat_pwrite(ef->dev, &meta1, sizeof(meta1),
			co2o(ef, new_cluster, new_offset)) < 0)
	{
		exfat_error("failed to write meta1 entry on rename");
		return -EIO;
	}
	next_entry(ef, dir, &new_cluster, &new_offset);
	if (exfat_pwrite(ef->dev, &meta2, sizeof(meta2),
			co2o(ef, new_cluster, new_offset)) < 0)
	{
		exfat_error("failed to write meta2 entry on rename");
		return -EIO;
	}

	for (i = 0; i < name_entries; i++)
	{
		struct exfat_entry_name name_entry = {EXFAT_ENTRY_FILE_NAME, 0};
		memcpy(name_entry.name, name + i * EXFAT_ENAME_MAX,
				EXFAT_ENAME_MAX * sizeof(le16_t));
		next_entry(ef, dir, &new_cluster, &new_offset);
		if (exfat_pwrite(ef->dev, &name_entry, sizeof(name_entry),
				co2o(ef, new_cluster, new_offset)) < 0)
		{
			exfat_error("failed to write name entry on rename");
			return -EIO;
		}
	}

	memcpy(node->name, name, (EXFAT_NAME_MAX + 1) * sizeof(le16_t));
	tree_detach(node);
	tree_attach(dir, node);
	return 0;
}

int exfat_rename(struct exfat* ef, const char* old_path, const char* new_path)
{
	struct exfat_node* node;
	struct exfat_node* existing;
	struct exfat_node* dir;
	cluster_t cluster = EXFAT_CLUSTER_BAD;
	off64_t offset = -1;
	le16_t name[EXFAT_NAME_MAX + 1];
	int rc;

	rc = exfat_lookup(ef, &node, old_path);
	if (rc != 0)
		return rc;

	rc = exfat_split(ef, &dir, &existing, name, new_path);
	if (rc != 0)
	{
		exfat_put_node(ef, node);
		return rc;
	}

	/* check that target is not a subdirectory of the source */
	if (node->flags & EXFAT_ATTRIB_DIR)
	{
		struct exfat_node* p;

		for (p = dir; p; p = p->parent)
			if (node == p)
			{
				if (existing != NULL)
					exfat_put_node(ef, existing);
				exfat_put_node(ef, dir);
				exfat_put_node(ef, node);
				return -EINVAL;
			}
	}

	if (existing != NULL)
	{
		/* remove target if it's not the same node as source */
		if (existing != node)
		{
			if (existing->flags & EXFAT_ATTRIB_DIR)
			{
				if (node->flags & EXFAT_ATTRIB_DIR)
					rc = exfat_rmdir(ef, existing);
				else
					rc = -ENOTDIR;
			}
			else
			{
				if (!(node->flags & EXFAT_ATTRIB_DIR))
					rc = exfat_unlink(ef, existing);
				else
					rc = -EISDIR;
			}
			exfat_put_node(ef, existing);
			if (rc != 0)
			{
				exfat_put_node(ef, dir);
				exfat_put_node(ef, node);
				return rc;
			}
		}
		else
			exfat_put_node(ef, existing);
	}

	rc = find_slot(ef, dir, &cluster, &offset,
			2 + DIV_ROUND_UP(utf16_length(name), EXFAT_ENAME_MAX));
	if (rc != 0)
	{
		exfat_put_node(ef, dir);
		exfat_put_node(ef, node);
		return rc;
	}
	rc = rename_entry(ef, dir, node, name, cluster, offset);
	exfat_put_node(ef, dir);
	exfat_put_node(ef, node);
	return 0;
}

void exfat_utimes(struct exfat_node* node, const struct timespec tv[2])
{
	node->atime = tv[0].tv_sec;
	node->mtime = tv[1].tv_sec;
	node->flags |= EXFAT_ATTRIB_DIRTY;
}

void exfat_update_atime(struct exfat_node* node)
{
	node->atime = time(NULL);
	node->flags |= EXFAT_ATTRIB_DIRTY;
}

void exfat_update_mtime(struct exfat_node* node)
{
	node->mtime = time(NULL);
	node->flags |= EXFAT_ATTRIB_DIRTY;
}

const char* exfat_get_label(struct exfat* ef)
{
	return ef->label;
}

static int find_label(struct exfat* ef, cluster_t* cluster, off64_t* offset)
{
	struct iterator it;
	int rc;

	rc = opendir(ef, ef->root, &it);
	if (rc != 0)
		return rc;

	for (;;)
	{
		if (it.offset >= ef->root->size)
		{
			closedir(&it);
			return -ENOENT;
		}

		if (get_entry_ptr(ef, &it)->type == EXFAT_ENTRY_LABEL)
		{
			*cluster = it.cluster;
			*offset = it.offset;
			closedir(&it);
			return 0;
		}

		if (fetch_next_entry(ef, ef->root, &it) != 0)
		{
			closedir(&it);
			return -EIO;
		}
	}
}

int exfat_set_label(struct exfat* ef, const char* label)
{
	le16_t label_utf16[EXFAT_ENAME_MAX + 1];
	int rc;
	cluster_t cluster;
	off64_t offset;
	struct exfat_entry_label entry;

	memset(label_utf16, 0, sizeof(label_utf16));
	rc = utf8_to_utf16(label_utf16, label, EXFAT_ENAME_MAX, strlen(label));
	if (rc != 0)
		return rc;

	rc = find_label(ef, &cluster, &offset);
	if (rc == -ENOENT)
		rc = find_slot(ef, ef->root, &cluster, &offset, 1);
	if (rc != 0)
		return rc;

	entry.type = EXFAT_ENTRY_LABEL;
	entry.length = utf16_length(label_utf16);
	memcpy(entry.name, label_utf16, sizeof(entry.name));
	if (entry.length == 0)
		entry.type ^= EXFAT_ENTRY_VALID;

	if (exfat_pwrite(ef->dev, &entry, sizeof(struct exfat_entry_label),
			co2o(ef, cluster, offset)) < 0)
	{
		exfat_error("failed to write label entry");
		return -EIO;
	}
	strcpy(ef->label, label);
	return 0;
}