aboutsummaryrefslogtreecommitdiffstats
path: root/fs/btrfs/ordered-data.c
blob: 7772f02ba28e6966826c0d897475961ad628c2f5 (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
/*
 * Copyright (C) 2007 Oracle.  All rights reserved.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public
 * License v2 as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public
 * License along with this program; if not, write to the
 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 * Boston, MA 021110-1307, USA.
 */

#include <linux/slab.h>
#include <linux/blkdev.h>
#include <linux/writeback.h>
#include <linux/pagevec.h>
#include "ctree.h"
#include "transaction.h"
#include "btrfs_inode.h"
#include "extent_io.h"

static struct kmem_cache *btrfs_ordered_extent_cache;

static u64 entry_end(struct btrfs_ordered_extent *entry)
{
	if (entry->file_offset + entry->len < entry->file_offset)
		return (u64)-1;
	return entry->file_offset + entry->len;
}

/* returns NULL if the insertion worked, or it returns the node it did find
 * in the tree
 */
static struct rb_node *tree_insert(struct rb_root *root, u64 file_offset,
				   struct rb_node *node)
{
	struct rb_node **p = &root->rb_node;
	struct rb_node *parent = NULL;
	struct btrfs_ordered_extent *entry;

	while (*p) {
		parent = *p;
		entry = rb_entry(parent, struct btrfs_ordered_extent, rb_node);

		if (file_offset < entry->file_offset)
			p = &(*p)->rb_left;
		else if (file_offset >= entry_end(entry))
			p = &(*p)->rb_right;
		else
			return parent;
	}

	rb_link_node(node, parent, p);
	rb_insert_color(node, root);
	return NULL;
}

static void ordered_data_tree_panic(struct inode *inode, int errno,
					       u64 offset)
{
	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
	btrfs_panic(fs_info, errno, "Inconsistency in ordered tree at offset "
		    "%llu\n", (unsigned long long)offset);
}

/*
 * look for a given offset in the tree, and if it can't be found return the
 * first lesser offset
 */
static struct rb_node *__tree_search(struct rb_root *root, u64 file_offset,
				     struct rb_node **prev_ret)
{
	struct rb_node *n = root->rb_node;
	struct rb_node *prev = NULL;
	struct rb_node *test;
	struct btrfs_ordered_extent *entry;
	struct btrfs_ordered_extent *prev_entry = NULL;

	while (n) {
		entry = rb_entry(n, struct btrfs_ordered_extent, rb_node);
		prev = n;
		prev_entry = entry;

		if (file_offset < entry->file_offset)
			n = n->rb_left;
		else if (file_offset >= entry_end(entry))
			n = n->rb_right;
		else
			return n;
	}
	if (!prev_ret)
		return NULL;

	while (prev && file_offset >= entry_end(prev_entry)) {
		test = rb_next(prev);
		if (!test)
			break;
		prev_entry = rb_entry(test, struct btrfs_ordered_extent,
				      rb_node);
		if (file_offset < entry_end(prev_entry))
			break;

		prev = test;
	}
	if (prev)
		prev_entry = rb_entry(prev, struct btrfs_ordered_extent,
				      rb_node);
	while (prev && file_offset < entry_end(prev_entry)) {
		test = rb_prev(prev);
		if (!test)
			break;
		prev_entry = rb_entry(test, struct btrfs_ordered_extent,
				      rb_node);
		prev = test;
	}
	*prev_ret = prev;
	return NULL;
}

/*
 * helper to check if a given offset is inside a given entry
 */
static int offset_in_entry(struct btrfs_ordered_extent *entry, u64 file_offset)
{
	if (file_offset < entry->file_offset ||
	    entry->file_offset + entry->len <= file_offset)
		return 0;
	return 1;
}

static int range_overlaps(struct btrfs_ordered_extent *entry, u64 file_offset,
			  u64 len)
{
	if (file_offset + len <= entry->file_offset ||
	    entry->file_offset + entry->len <= file_offset)
		return 0;
	return 1;
}

/*
 * look find the first ordered struct that has this offset, otherwise
 * the first one less than this offset
 */
static inline struct rb_node *tree_search(struct btrfs_ordered_inode_tree *tree,
					  u64 file_offset)
{
	struct rb_root *root = &tree->tree;
	struct rb_node *prev = NULL;
	struct rb_node *ret;
	struct btrfs_ordered_extent *entry;

	if (tree->last) {
		entry = rb_entry(tree->last, struct btrfs_ordered_extent,
				 rb_node);
		if (offset_in_entry(entry, file_offset))
			return tree->last;
	}
	ret = __tree_search(root, file_offset, &prev);
	if (!ret)
		ret = prev;
	if (ret)
		tree->last = ret;
	return ret;
}

/* allocate and add a new ordered_extent into the per-inode tree.
 * file_offset is the logical offset in the file
 *
 * start is the disk block number of an extent already reserved in the
 * extent allocation tree
 *
 * len is the length of the extent
 *
 * The tree is given a single reference on the ordered extent that was
 * inserted.
 */
static int __btrfs_add_ordered_extent(struct inode *inode, u64 file_offset,
				      u64 start, u64 len, u64 disk_len,
				      int type, int dio, int compress_type)
{
	struct btrfs_ordered_inode_tree *tree;
	struct rb_node *node;
	struct btrfs_ordered_extent *entry;

	tree = &BTRFS_I(inode)->ordered_tree;
	entry = kmem_cache_zalloc(btrfs_ordered_extent_cache, GFP_NOFS);
	if (!entry)
		return -ENOMEM;

	entry->file_offset = file_offset;
	entry->start = start;
	entry->len = len;
	entry->disk_len = disk_len;
	entry->bytes_left = len;
	entry->inode = igrab(inode);
	entry->compress_type = compress_type;
	if (type != BTRFS_ORDERED_IO_DONE && type != BTRFS_ORDERED_COMPLETE)
		set_bit(type, &entry->flags);

	if (dio)
		set_bit(BTRFS_ORDERED_DIRECT, &entry->flags);

	/* one ref for the tree */
	atomic_set(&entry->refs, 1);
	init_waitqueue_head(&entry->wait);
	INIT_LIST_HEAD(&entry->list);
	INIT_LIST_HEAD(&entry->root_extent_list);

	trace_btrfs_ordered_extent_add(inode, entry);

	spin_lock_irq(&tree->lock);
	node = tree_insert(&tree->tree, file_offset,
			   &entry->rb_node);
	if (node)
		ordered_data_tree_panic(inode, -EEXIST, file_offset);
	spin_unlock_irq(&tree->lock);

	spin_lock(&BTRFS_I(inode)->root->fs_info->ordered_extent_lock);
	list_add_tail(&entry->root_extent_list,
		      &BTRFS_I(inode)->root->fs_info->ordered_extents);
	spin_unlock(&BTRFS_I(inode)->root->fs_info->ordered_extent_lock);

	return 0;
}

int btrfs_add_ordered_extent(struct inode *inode, u64 file_offset,
			     u64 start, u64 len, u64 disk_len, int type)
{
	return __btrfs_add_ordered_extent(inode, file_offset, start, len,
					  disk_len, type, 0,
					  BTRFS_COMPRESS_NONE);
}

int btrfs_add_ordered_extent_dio(struct inode *inode, u64 file_offset,
				 u64 start, u64 len, u64 disk_len, int type)
{
	return __btrfs_add_ordered_extent(inode, file_offset, start, len,
					  disk_len, type, 1,
					  BTRFS_COMPRESS_NONE);
}

int btrfs_add_ordered_extent_compress(struct inode *inode, u64 file_offset,
				      u64 start, u64 len, u64 disk_len,
				      int type, int compress_type)
{
	return __btrfs_add_ordered_extent(inode, file_offset, start, len,
					  disk_len, type, 0,
					  compress_type);
}

/*
 * Add a struct btrfs_ordered_sum into the list of checksums to be inserted
 * when an ordered extent is finished.  If the list covers more than one
 * ordered extent, it is split across multiples.
 */
void btrfs_add_ordered_sum(struct inode *inode,
			   struct btrfs_ordered_extent *entry,
			   struct btrfs_ordered_sum *sum)
{
	struct btrfs_ordered_inode_tree *tree;

	tree = &BTRFS_I(inode)->ordered_tree;
	spin_lock_irq(&tree->lock);
	list_add_tail(&sum->list, &entry->list);
	spin_unlock_irq(&tree->lock);
}

/*
 * this is used to account for finished IO across a given range
 * of the file.  The IO may span ordered extents.  If
 * a given ordered_extent is completely done, 1 is returned, otherwise
 * 0.
 *
 * test_and_set_bit on a flag in the struct btrfs_ordered_extent is used
 * to make sure this function only returns 1 once for a given ordered extent.
 *
 * file_offset is updated to one byte past the range that is recorded as
 * complete.  This allows you to walk forward in the file.
 */
int btrfs_dec_test_first_ordered_pending(struct inode *inode,
				   struct btrfs_ordered_extent **cached,
				   u64 *file_offset, u64 io_size, int uptodate)
{
	struct btrfs_ordered_inode_tree *tree;
	struct rb_node *node;
	struct btrfs_ordered_extent *entry = NULL;
	int ret;
	unsigned long flags;
	u64 dec_end;
	u64 dec_start;
	u64 to_dec;

	tree = &BTRFS_I(inode)->ordered_tree;
	spin_lock_irqsave(&tree->lock, flags);
	node = tree_search(tree, *file_offset);
	if (!node) {
		ret = 1;
		goto out;
	}

	entry = rb_entry(node, struct btrfs_ordered_extent, rb_node);
	if (!offset_in_entry(entry, *file_offset)) {
		ret = 1;
		goto out;
	}

	dec_start = max(*file_offset, entry->file_offset);
	dec_end = min(*file_offset + io_size, entry->file_offset +
		      entry->len);
	*file_offset = dec_end;
	if (dec_start > dec_end) {
		printk(KERN_CRIT "bad ordering dec_start %llu end %llu\n",
		       (unsigned long long)dec_start,
		       (unsigned long long)dec_end);
	}
	to_dec = dec_end - dec_start;
	if (to_dec > entry->bytes_left) {
		printk(KERN_CRIT "bad ordered accounting left %llu size %llu\n",
		       (unsigned long long)entry->bytes_left,
		       (unsigned long long)to_dec);
	}
	entry->bytes_left -= to_dec;
	if (!uptodate)
		set_bit(BTRFS_ORDERED_IOERR, &entry->flags);

	if (entry->bytes_left == 0)
		ret = test_and_set_bit(BTRFS_ORDERED_IO_DONE, &entry->flags);
	else
		ret = 1;
out:
	if (!ret && cached && entry) {
		*cached = entry;
		atomic_inc(&entry->refs);
	}
	spin_unlock_irqrestore(&tree->lock, flags);
	return ret == 0;
}

/*
 * this is used to account for finished IO across a given range
 * of the file.  The IO should not span ordered extents.  If
 * a given ordered_extent is completely done, 1 is returned, otherwise
 * 0.
 *
 * test_and_set_bit on a flag in the struct btrfs_ordered_extent is used
 * to make sure this function only returns 1 once for a given ordered extent.
 */
int btrfs_dec_test_ordered_pending(struct inode *inode,
				   struct btrfs_ordered_extent **cached,
				   u64 file_offset, u64 io_size, int uptodate)
{
	struct btrfs_ordered_inode_tree *tree;
	struct rb_node *node;
	struct btrfs_ordered_extent *entry = NULL;
	unsigned long flags;
	int ret;

	tree = &BTRFS_I(inode)->ordered_tree;
	spin_lock_irqsave(&tree->lock, flags);
	if (cached && *cached) {
		entry = *cached;
		goto have_entry;
	}

	node = tree_search(tree, file_offset);
	if (!node) {
		ret = 1;
		goto out;
	}

	entry = rb_entry(node, struct btrfs_ordered_extent, rb_node);
have_entry:
	if (!offset_in_entry(entry, file_offset)) {
		ret = 1;
		goto out;
	}

	if (io_size > entry->bytes_left) {
		printk(KERN_CRIT "bad ordered accounting left %llu size %llu\n",
		       (unsigned long long)entry->bytes_left,
		       (unsigned long long)io_size);
	}
	entry->bytes_left -= io_size;
	if (!uptodate)
		set_bit(BTRFS_ORDERED_IOERR, &entry->flags);

	if (entry->bytes_left == 0)
		ret = test_and_set_bit(BTRFS_ORDERED_IO_DONE, &entry->flags);
	else
		ret = 1;
out:
	if (!ret && cached && entry) {
		*cached = entry;
		atomic_inc(&entry->refs);
	}
	spin_unlock_irqrestore(&tree->lock, flags);
	return ret == 0;
}

/*
 * used to drop a reference on an ordered extent.  This will free
 * the extent if the last reference is dropped
 */
void btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry)
{
	struct list_head *cur;
	struct btrfs_ordered_sum *sum;

	trace_btrfs_ordered_extent_put(entry->inode, entry);

	if (atomic_dec_and_test(&entry->refs)) {
		if (entry->inode)
			btrfs_add_delayed_iput(entry->inode);
		while (!list_empty(&entry->list)) {
			cur = entry->list.next;
			sum = list_entry(cur, struct btrfs_ordered_sum, list);
			list_del(&sum->list);
			kfree(sum);
		}
		kmem_cache_free(btrfs_ordered_extent_cache, entry);
	}
}

/*
 * remove an ordered extent from the tree.  No references are dropped
 * and waiters are woken up.
 */
void btrfs_remove_ordered_extent(struct inode *inode,
				 struct btrfs_ordered_extent *entry)
{
	struct btrfs_ordered_inode_tree *tree;
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct rb_node *node;

	tree = &BTRFS_I(inode)->ordered_tree;
	spin_lock_irq(&tree->lock);
	node = &entry->rb_node;
	rb_erase(node, &tree->tree);
	tree->last = NULL;
	set_bit(BTRFS_ORDERED_COMPLETE, &entry->flags);
	spin_unlock_irq(&tree->lock);

	spin_lock(&root->fs_info->ordered_extent_lock);
	list_del_init(&entry->root_extent_list);

	trace_btrfs_ordered_extent_remove(inode, entry);

	/*
	 * we have no more ordered extents for this inode and
	 * no dirty pages.  We can safely remove it from the
	 * list of ordered extents
	 */
	if (RB_EMPTY_ROOT(&tree->tree) &&
	    !mapping_tagged(inode->i_mapping, PAGECACHE_TAG_DIRTY)) {
		list_del_init(&BTRFS_I(inode)->ordered_operations);
	}
	spin_unlock(&root->fs_info->ordered_extent_lock);
	wake_up(&entry->wait);
}

/*
 * wait for all the ordered extents in a root.  This is done when balancing
 * space between drives.
 */
void btrfs_wait_ordered_extents(struct btrfs_root *root, int delay_iput)
{
	struct list_head splice;
	struct list_head *cur;
	struct btrfs_ordered_extent *ordered;
	struct inode *inode;

	INIT_LIST_HEAD(&splice);

	spin_lock(&root->fs_info->ordered_extent_lock);
	list_splice_init(&root->fs_info->ordered_extents, &splice);
	while (!list_empty(&splice)) {
		cur = splice.next;
		ordered = list_entry(cur, struct btrfs_ordered_extent,
				     root_extent_list);
		list_del_init(&ordered->root_extent_list);
		atomic_inc(&ordered->refs);

		/*
		 * the inode may be getting freed (in sys_unlink path).
		 */
		inode = igrab(ordered->inode);

		spin_unlock(&root->fs_info->ordered_extent_lock);

		if (inode) {
			btrfs_start_ordered_extent(inode, ordered, 1);
			btrfs_put_ordered_extent(ordered);
			if (delay_iput)
				btrfs_add_delayed_iput(inode);
			else
				iput(inode);
		} else {
			btrfs_put_ordered_extent(ordered);
		}

		spin_lock(&root->fs_info->ordered_extent_lock);
	}
	spin_unlock(&root->fs_info->ordered_extent_lock);
}

/*
 * this is used during transaction commit to write all the inodes
 * added to the ordered operation list.  These files must be fully on
 * disk before the transaction commits.
 *
 * we have two modes here, one is to just start the IO via filemap_flush
 * and the other is to wait for all the io.  When we wait, we have an
 * extra check to make sure the ordered operation list really is empty
 * before we return
 */
void btrfs_run_ordered_operations(struct btrfs_root *root, int wait)
{
	struct btrfs_inode *btrfs_inode;
	struct inode *inode;
	struct list_head splice;

	INIT_LIST_HEAD(&splice);

	mutex_lock(&root->fs_info->ordered_operations_mutex);
	spin_lock(&root->fs_info->ordered_extent_lock);
again:
	list_splice_init(&root->fs_info->ordered_operations, &splice);

	while (!list_empty(&splice)) {
		btrfs_inode = list_entry(splice.next, struct btrfs_inode,
				   ordered_operations);

		inode = &btrfs_inode->vfs_inode;

		list_del_init(&btrfs_inode->ordered_operations);

		/*
		 * the inode may be getting freed (in sys_unlink path).
		 */
		inode = igrab(inode);

		if (!wait && inode) {
			list_add_tail(&BTRFS_I(inode)->ordered_operations,
			      &root->fs_info->ordered_operations);
		}
		spin_unlock(&root->fs_info->ordered_extent_lock);

		if (inode) {
			if (wait)
				btrfs_wait_ordered_range(inode, 0, (u64)-1);
			else
				filemap_flush(inode->i_mapping);
			btrfs_add_delayed_iput(inode);
		}

		cond_resched();
		spin_lock(&root->fs_info->ordered_extent_lock);
	}
	if (wait && !list_empty(&root->fs_info->ordered_operations))
		goto again;

	spin_unlock(&root->fs_info->ordered_extent_lock);
	mutex_unlock(&root->fs_info->ordered_operations_mutex);
}

/*
 * Used to start IO or wait for a given ordered extent to finish.
 *
 * If wait is one, this effectively waits on page writeback for all the pages
 * in the extent, and it waits on the io completion code to insert
 * metadata into the btree corresponding to the extent
 */
void btrfs_start_ordered_extent(struct inode *inode,
				       struct btrfs_ordered_extent *entry,
				       int wait)
{
	u64 start = entry->file_offset;
	u64 end = start + entry->len - 1;

	trace_btrfs_ordered_extent_start(inode, entry);

	/*
	 * pages in the range can be dirty, clean or writeback.  We
	 * start IO on any dirty ones so the wait doesn't stall waiting
	 * for the flusher thread to find them
	 */
	if (!test_bit(BTRFS_ORDERED_DIRECT, &entry->flags))
		filemap_fdatawrite_range(inode->i_mapping, start, end);
	if (wait) {
		wait_event(entry->wait, test_bit(BTRFS_ORDERED_COMPLETE,
						 &entry->flags));
	}
}

/*
 * Used to wait on ordered extents across a large range of bytes.
 */
void btrfs_wait_ordered_range(struct inode *inode, u64 start, u64 len)
{
	u64 end;
	u64 orig_end;
	struct btrfs_ordered_extent *ordered;
	int found;

	if (start + len < start) {
		orig_end = INT_LIMIT(loff_t);
	} else {
		orig_end = start + len - 1;
		if (orig_end > INT_LIMIT(loff_t))
			orig_end = INT_LIMIT(loff_t);
	}

	/* start IO across the range first to instantiate any delalloc
	 * extents
	 */
	filemap_fdatawrite_range(inode->i_mapping, start, orig_end);

	/*
	 * So with compression we will find and lock a dirty page and clear the
	 * first one as dirty, setup an async extent, and immediately return
	 * with the entire range locked but with nobody actually marked with
	 * writeback.  So we can't just filemap_write_and_wait_range() and
	 * expect it to work since it will just kick off a thread to do the
	 * actual work.  So we need to call filemap_fdatawrite_range _again_
	 * since it will wait on the page lock, which won't be unlocked until
	 * after the pages have been marked as writeback and so we're good to go
	 * from there.  We have to do this otherwise we'll miss the ordered
	 * extents and that results in badness.  Please Josef, do not think you
	 * know better and pull this out at some point in the future, it is
	 * right and you are wrong.
	 */
	if (test_bit(BTRFS_INODE_HAS_ASYNC_EXTENT,
		     &BTRFS_I(inode)->runtime_flags))
		filemap_fdatawrite_range(inode->i_mapping, start, orig_end);

	filemap_fdatawait_range(inode->i_mapping, start, orig_end);

	end = orig_end;
	found = 0;
	while (1) {
		ordered = btrfs_lookup_first_ordered_extent(inode, end);
		if (!ordered)
			break;
		if (ordered->file_offset > orig_end) {
			btrfs_put_ordered_extent(ordered);
			break;
		}
		if (ordered->file_offset + ordered->len < start) {
			btrfs_put_ordered_extent(ordered);
			break;
		}
		found++;
		btrfs_start_ordered_extent(inode, ordered, 1);
		end = ordered->file_offset;
		btrfs_put_ordered_extent(ordered);
		if (end == 0 || end == start)
			break;
		end--;
	}
}

/*
 * find an ordered extent corresponding to file_offset.  return NULL if
 * nothing is found, otherwise take a reference on the extent and return it
 */
struct btrfs_ordered_extent *btrfs_lookup_ordered_extent(struct inode *inode,
							 u64 file_offset)
{
	struct btrfs_ordered_inode_tree *tree;
	struct rb_node *node;
	struct btrfs_ordered_extent *entry = NULL;

	tree = &BTRFS_I(inode)->ordered_tree;
	spin_lock_irq(&tree->lock);
	node = tree_search(tree, file_offset);
	if (!node)
		goto out;

	entry = rb_entry(node, struct btrfs_ordered_extent, rb_node);
	if (!offset_in_entry(entry, file_offset))
		entry = NULL;
	if (entry)
		atomic_inc(&entry->refs);
out:
	spin_unlock_irq(&tree->lock);
	return entry;
}

/* Since the DIO code tries to lock a wide area we need to look for any ordered
 * extents that exist in the range, rather than just the start of the range.
 */
struct btrfs_ordered_extent *btrfs_lookup_ordered_range(struct inode *inode,
							u64 file_offset,
							u64 len)
{
	struct btrfs_ordered_inode_tree *tree;
	struct rb_node *node;
	struct btrfs_ordered_extent *entry = NULL;

	tree = &BTRFS_I(inode)->ordered_tree;
	spin_lock_irq(&tree->lock);
	node = tree_search(tree, file_offset);
	if (!node) {
		node = tree_search(tree, file_offset + len);
		if (!node)
			goto out;
	}

	while (1) {
		entry = rb_entry(node, struct btrfs_ordered_extent, rb_node);
		if (range_overlaps(entry, file_offset, len))
			break;

		if (entry->file_offset >= file_offset + len) {
			entry = NULL;
			break;
		}
		entry = NULL;
		node = rb_next(node);
		if (!node)
			break;
	}
out:
	if (entry)
		atomic_inc(&entry->refs);
	spin_unlock_irq(&tree->lock);
	return entry;
}

/*
 * lookup and return any extent before 'file_offset'.  NULL is returned
 * if none is found
 */
struct btrfs_ordered_extent *
btrfs_lookup_first_ordered_extent(struct inode *inode, u64 file_offset)
{
	struct btrfs_ordered_inode_tree *tree;
	struct rb_node *node;
	struct btrfs_ordered_extent *entry = NULL;

	tree = &BTRFS_I(inode)->ordered_tree;
	spin_lock_irq(&tree->lock);
	node = tree_search(tree, file_offset);
	if (!node)
		goto out;

	entry = rb_entry(node, struct btrfs_ordered_extent, rb_node);
	atomic_inc(&entry->refs);
out:
	spin_unlock_irq(&tree->lock);
	return entry;
}

/*
 * After an extent is done, call this to conditionally update the on disk
 * i_size.  i_size is updated to cover any fully written part of the file.
 */
int btrfs_ordered_update_i_size(struct inode *inode, u64 offset,
				struct btrfs_ordered_extent *ordered)
{
	struct btrfs_ordered_inode_tree *tree = &BTRFS_I(inode)->ordered_tree;
	u64 disk_i_size;
	u64 new_i_size;
	u64 i_size = i_size_read(inode);
	struct rb_node *node;
	struct rb_node *prev = NULL;
	struct btrfs_ordered_extent *test;
	int ret = 1;

	if (ordered)
		offset = entry_end(ordered);
	else
		offset = ALIGN(offset, BTRFS_I(inode)->root->sectorsize);

	spin_lock_irq(&tree->lock);
	disk_i_size = BTRFS_I(inode)->disk_i_size;

	/* truncate file */
	if (disk_i_size > i_size) {
		BTRFS_I(inode)->disk_i_size = i_size;
		ret = 0;
		goto out;
	}

	/*
	 * if the disk i_size is already at the inode->i_size, or
	 * this ordered extent is inside the disk i_size, we're done
	 */
	if (disk_i_size == i_size || offset <= disk_i_size) {
		goto out;
	}

	/*
	 * walk backward from this ordered extent to disk_i_size.
	 * if we find an ordered extent then we can't update disk i_size
	 * yet
	 */
	if (ordered) {
		node = rb_prev(&ordered->rb_node);
	} else {
		prev = tree_search(tree, offset);
		/*
		 * we insert file extents without involving ordered struct,
		 * so there should be no ordered struct cover this offset
		 */
		if (prev) {
			test = rb_entry(prev, struct btrfs_ordered_extent,
					rb_node);
			BUG_ON(offset_in_entry(test, offset));
		}
		node = prev;
	}
	for (; node; node = rb_prev(node)) {
		test = rb_entry(node, struct btrfs_ordered_extent, rb_node);

		/* We treat this entry as if it doesnt exist */
		if (test_bit(BTRFS_ORDERED_UPDATED_ISIZE, &test->flags))
			continue;
		if (test->file_offset + test->len <= disk_i_size)
			break;
		if (test->file_offset >= i_size)
			break;
		if (test->file_offset >= disk_i_size) {
			/*
			 * we don't update disk_i_size now, so record this
			 * undealt i_size. Or we will not know the real
			 * i_size.
			 */
			if (test->outstanding_isize < offset)
				test->outstanding_isize = offset;
			if (ordered &&
			    ordered->outstanding_isize >
			    test->outstanding_isize)
				test->outstanding_isize =
						ordered->outstanding_isize;
			goto out;
		}
	}
	new_i_size = min_t(u64, offset, i_size);

	/*
	 * Some ordered extents may completed before the current one, and
	 * we hold the real i_size in ->outstanding_isize.
	 */
	if (ordered && ordered->outstanding_isize > new_i_size)
		new_i_size = min_t(u64, ordered->outstanding_isize, i_size);
	BTRFS_I(inode)->disk_i_size = new_i_size;
	ret = 0;
out:
	/*
	 * We need to do this because we can't remove ordered extents until
	 * after the i_disk_size has been updated and then the inode has been
	 * updated to reflect the change, so we need to tell anybody who finds
	 * this ordered extent that we've already done all the real work, we
	 * just haven't completed all the other work.
	 */
	if (ordered)
		set_bit(BTRFS_ORDERED_UPDATED_ISIZE, &ordered->flags);
	spin_unlock_irq(&tree->lock);
	return ret;
}

/*
 * search the ordered extents for one corresponding to 'offset' and
 * try to find a checksum.  This is used because we allow pages to
 * be reclaimed before their checksum is actually put into the btree
 */
int btrfs_find_ordered_sum(struct inode *inode, u64 offset, u64 disk_bytenr,
			   u32 *sum)
{
	struct btrfs_ordered_sum *ordered_sum;
	struct btrfs_sector_sum *sector_sums;
	struct btrfs_ordered_extent *ordered;
	struct btrfs_ordered_inode_tree *tree = &BTRFS_I(inode)->ordered_tree;
	unsigned long num_sectors;
	unsigned long i;
	u32 sectorsize = BTRFS_I(inode)->root->sectorsize;
	int ret = 1;

	ordered = btrfs_lookup_ordered_extent(inode, offset);
	if (!ordered)
		return 1;

	spin_lock_irq(&tree->lock);
	list_for_each_entry_reverse(ordered_sum, &ordered->list, list) {
		if (disk_bytenr >= ordered_sum->bytenr) {
			num_sectors = ordered_sum->len / sectorsize;
			sector_sums = ordered_sum->sums;
			for (i = 0; i < num_sectors; i++) {
				if (sector_sums[i].bytenr == disk_bytenr) {
					*sum = sector_sums[i].sum;
					ret = 0;
					goto out;
				}
			}
		}
	}
out:
	spin_unlock_irq(&tree->lock);
	btrfs_put_ordered_extent(ordered);
	return ret;
}


/*
 * add a given inode to the list of inodes that must be fully on
 * disk before a transaction commit finishes.
 *
 * This basically gives us the ext3 style data=ordered mode, and it is mostly
 * used to make sure renamed files are fully on disk.
 *
 * It is a noop if the inode is already fully on disk.
 *
 * If trans is not null, we'll do a friendly check for a transaction that
 * is already flushing things and force the IO down ourselves.
 */
void btrfs_add_ordered_operation(struct btrfs_trans_handle *trans,
				 struct btrfs_root *root, struct inode *inode)
{
	u64 last_mod;

	last_mod = max(BTRFS_I(inode)->generation, BTRFS_I(inode)->last_trans);

	/*
	 * if this file hasn't been changed since the last transaction
	 * commit, we can safely return without doing anything
	 */
	if (last_mod < root->fs_info->last_trans_committed)
		return;

	/*
	 * the transaction is already committing.  Just start the IO and
	 * don't bother with all of this list nonsense
	 */
	if (trans && root->fs_info->running_transaction->blocked) {
		btrfs_wait_ordered_range(inode, 0, (u64)-1);
		return;
	}

	spin_lock(&root->fs_info->ordered_extent_lock);
	if (list_empty(&BTRFS_I(inode)->ordered_operations)) {
		list_add_tail(&BTRFS_I(inode)->ordered_operations,
			      &root->fs_info->ordered_operations);
	}
	spin_unlock(&root->fs_info->ordered_extent_lock);
}

int __init ordered_data_init(void)
{
	btrfs_ordered_extent_cache = kmem_cache_create("btrfs_ordered_extent",
				     sizeof(struct btrfs_ordered_extent), 0,
				     SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD,
				     NULL);
	if (!btrfs_ordered_extent_cache)
		return -ENOMEM;
	return 0;
}

void ordered_data_exit(void)
{
	if (btrfs_ordered_extent_cache)
		kmem_cache_destroy(btrfs_ordered_extent_cache);
}